CONSTRUCTION HEALTH AND SAFETY (H&S)
PERFORMANCE IMPROVEMENT- A CLIENTCENTRED MODEL
INNOCENT MUSONDA
2012
CONSTRUCTION HEALTH AND SAFETY (H&S)
PERFORMANCE IMPROVEMENT- A CLIENT-CENTRED
MODEL
A thesis presented
by
Innocent Musonda
to
THE FACULTY OF ENGINEERING AND BUILT
ENVIRONMENT
in fulfilment of the requirements for the degree of
DOCTOR OF PHILOSOPHY
in the subject of
ENGINEERING MANAGEMENT
UNIVERSITY OF JOHANNESBURG, JOHANNESBURG,
SOUTH AFRICA
March 2012
PROMOTER: PROF J.H.C. PRETORIUS
CO-PROMOTER: PROF T. C. HAUPT
ii
DEDICATION
This thesis is dedicated to my two daughters Nkatya and Nkumbu
Musonda. It is also dedicated to mum, for having endured so much for us,
may this work bring joy to you. Your sacrifice was worth it. It is also
dedicated to construction workers, who continue to endure the harsh
construction environments.
iii
DECLARATION
I, Innocent Musonda, declare that “construction health and safety performance
improvement – a client–centred model” is the result of my own work, except to the
extent indicated. All sources that I have used or quoted have been duly acknowledged
by means of complete references. The thesis is submitted in fulfilment of the
requirements of the degree Doctor of Philosophy in Engineering Management and has
never been presented anywhere else for a similar purpose.
March 2012
______________________
Innocent Musonda
Date
iv
ACKNOWLEDGEMENTS
I am deeply indebted to my Lord Jesus Christ without whom, accomplishing this task
could not have been possible. Ebenezer
I would also like to recognise and acknowledge the following individuals and
organisations for having contributed in various ways to the successful completion of
this research

Professors Pretorius and Haupt for having spent their time to offer professional
advice, insight and motivation until the task was completed;

Professor Haupt for having facilitated financial assistance from the National
Research Fund (NRF). I was able to have study breaks, collect data, and attend
national and international conferences as a result of this financial assistance;

The University of Johannesburg for the financial assistance and office
facilities;

My wife, Patricia for having provided that extra support that I needed most of
the time;

Mr Ferdinand Fester, my head of department, for the support and numerous
time off so that I was able to conduct research;

SACPCMP, 2010 second year University of Johannesburg Building students,
Chiwala Maipambe, Kauzya Siwale, Martin Bwalya and Bwalya Mukuka for
assisting with data collection;

Statkon for data input and analysis. My special gratitude goes to Jurgen
Becker for the countless hours spent working with me on structural equation
modelling;

Mr Harrison Njamu & his wife, and Pastor Evans Chinthalima & his wife, for
their encouragement, support and prayers;

My work colleagues; Clinton, Justus, Kauzya, Didi, Maphefo, Nazeem and
Ferdi for their encouragement and for making this task easier and enjoyable;

My family, “abena munkonge” for their prayers and for having provided the
motivation that I needed to push on; and

Heidi, for critically and professionally editing the language content;
v
EXECUTIVE SUMMARY
Health and safety (H&S) is still a problem in the construction industry. Both locally in
Southern Africa and internationally, the construction industry continues to lag behind
other industries in terms of H&S performance. As a result, there has been a
considerable amount of research and interest on H&S performance improvement.
However, there seems to be little research of note focusing on the role and
contribution of construction clients to H&S performance improvement.
This research project investigated and modelled H&S performance on the role and
contribution of clients to H&S. The primary aim of the research was to model how
improved H&S performance could be achieved if clients and/or owners were
involved.
Literature was reviewed on the status of H&S performance in the construction
industry, the significance of the industry and on the current trends in H&S
performance improvement. Empirical data was collected through a Delphi study and
through a field questionnaire survey. Analysis of results from the Delphi study was
done with Microsoft Excel to output descriptive statistics. These descriptive statistics
informed on whether the experts had reached consensus. Analysis of results from the
field questionnaire was done with EQS version 6.1 and MPlus version 6.0. Both of
these are structural equation modelling software.
A conceptual client centred H&S performance improvement model was based on the
theory developed from literature review findings and the Delphi study. The
questionnaire survey was conducted for the purpose of validating the conceptual
model. The questionnaire survey was conducted in Gaborone, Botswana and
Johannesburg, South Africa. Respondents were asked to report on current
construction projects that they were part of as a professional. Respondents were asked
to indicate their assessment of client H&S culture, contractor, designer and project
H&S performance.
Results from the investigation pertained to three broad areas. The first results related
to theory on H&S performance and performance improvement. Findings were that
H&S performance improvement was unlikely if H&S culture was not improved.
vi
Further findings were that clients could provide the needed impetus to improve H&S
performance.
The second set of results pertained to the Delphi study. Findings from this study were
that the external environment, made up of the political, economic, social, legislative,
professional bodies and technology, and had influence on client H&S culture. In
addition, client H&S culture had influence on contractor, designer and project H&S
performance. Further findings from literature and the Delphi study indicated that the
client H&S culture could be a six factor model defined by leadership, involvement,
procedures, commitment, competence and communication. These factors were
collectively referred to as the LIP+3C model of client H&S culture.
The third set of results pertained to a field questionnaire survey. Generally, the
findings were that the hypotheses that client H&S culture had influence on contractor,
designer and project H&S performance, could not be rejected. Furthermore, the
hypothesis that the external environment had influence on client H&S culture could
also not be rejected. Structural equation modelling results on the model’s goodnessof-fit and statistical significance of parameter estimates met the cut-off criteria for the
hypothesised model’s fit to the sample data. However, the factors of client H&S
culture found to be of significant influence, statistically, were client commitment,
procedures and communication.
The current study established through a Delphi and a questionnaire survey, that
construction client had influence on contractor, designer and the overall project H&S
performance. Furthermore, the study established that client H&S culture was
influenced by the external environment factors of legislation, economic, professional
bodies, and materials and technology. Findings also showed that the client H&S
culture construct fit the six factor model. The six factors were leadership,
involvement, procedures, commitment, communication and competence (LIP+3C).
Based on the obtained results the client centred model seemed to be tenable in the
current sample of construction professionals. The manner in which the observed
variables reflect the client centred structural model instilled confidence in the
construct validity of the structural model. However, due to the idiosyncratic dataset, it
remains to be seen if the proposed theoretical model replicates to other datasets. If this
vii
is indeed the case, the current research makes a significant contribution towards
understanding H&S in the construction industry. Indicator variables could then be
used as check items to determine and characterise client H&S culture as well as
project H&S performance.
Therefore the study recommended that the construction industry should focus on
improving H&S culture in the industry. Particularly, improving client H&S culture
was seen as critical to improving project H&S performance. Consequently, it was
recommended that all key stakeholders should be made aware of the variables that
define project H&S performance so that planning, organizing, monitoring and control
could be possible. Improving H&S culture and performance was found to be a factor
of undertaking these activities.
viii
LIST OF ABBREVIATIONS
ANOVA
Analysis of Variance
BFTU
Botswana Federation of Trade Unions
BLS
Bureau of Labor Statistics
BP
British Petroleum
BSc
Bachelor of Science degree
CFA
Confirmatory Factor Analysis
CI
Confidence Interval
CIB
International Council for Research and Innovation in Building
and Construction
CIB W099
CIB Working Commission on Safety and
Health in
Construction
CIDB
Construction Industry Development Board (South Africa)
CDM
Construction Design and Management regulations
CFI
Comparative Factor Index
CIOB
Chartered Institute of Building
CRC
Construction Research Centre
CSO
Central Statistics Office (Botswana)
DOETE
Department of Energy, Trade and Enterprises (Ireland)
DSc
Doctor of Science degree
EFA
Exploratory Factor Analysis
EU
European Union
FV
Full Variable model
GDP
Gross Domestic Product
GLM
General Linear Modelling
H&S
Health and Safety
HSA
Health and Safety Authority
HSE
Health and Safety Executive
IAEA
International Atomic Energy Agency
IET
Institution of Engineering and Technology
ILO
International Labour Organisation
INSAG
International Nuclear Safety Action Group
IOSH
Institution of Occupational Safety and Health Institution o
ix
IQD
Inter-Quartile Deviation
LIP+3C
Acronym standing for Leadership, Involvement, Procedures,
Commitment, Communication and Competence
LM
Lagrange Multiplier
MAD
Average (mean) absolute deviation
MBA-GP
Master Builders Association - Gauteng province
MSc
Master of Science Degree
NRF
National Research Foundation
RML
Robust Maximum Likelihood
RMSEA
Root Mean Square Error of Approximation
SACPCMP
South African Council of Project and Construction
Management Professions
SEM
Structural Equation Modelling
SPSS
Statistical Package for Social Sciences
SRMR
Standardised Root Mean square Residual
TLI
Tucker Lewis Index
TRIR
Total Recordable Injury Rate
UJ
University of Johannesburg
UK
United Kingdom
UMIST
University of Manchester Institute of Science and Technology
USA
United States of America
USD
United States Dollar
ZAR
South African Rand
x
TABLE OF CONTENTS
Contents
Page No
DEDICATION............................................................................................................ iii
DECLARATION......................................................................................................... iv
ACKNOWLEDGEMENTS ........................................................................................ v
EXECUTIVE SUMMARY ........................................................................................ vi
TABLE OF CONTENTS ........................................................................................... xi
LIST OF TABLES .................................................................................................... xvi
LIST OF FIGURES ............................................................................................... xviii
LIST OF APPENDICES ........................................................................................... xx
DEFINITION OF TERMS AND CONCEPTS USED IN THE THESIS ............ xxi
CHAPTER ONE .......................................................................................................... 1
INTRODUCTION........................................................................................................ 1
1.1
BACKGROUND ............................................................................................. 1
1.2
THE PROBLEM STATEMENT .................................................................... 7
1.3
THE STUDY ................................................................................................... 8
1.3.1
The general objectives of the study ........................................................ 8
1.3.2
Methods .................................................................................................. 8
1.3.3
Results .................................................................................................. 10
1.3.4
Limitation and delimitation of the study .............................................. 11
1.3.5
Assumptions ......................................................................................... 12
1.3.6
Importance of the study ........................................................................ 12
1.3.7
Research organisation .......................................................................... 13
1.3.8
Ethical consideration during the study ................................................. 15
1.3.9
Resources ............................................................................................. 15
1.4
CHAPTER SUMMARY ............................................................................... 16
CHAPTER TWO ....................................................................................................... 17
THE CONSTRUCTION INDUSTRY AND ITS HEALTH AND SAFETY
PERFORMANCE ...................................................................................................... 17
2.1
INTRODUCTION ......................................................................................... 17
2.2
THE CONSTRUCTION INDUSTRY .......................................................... 19
xi
2.2.1
Complexity of the construction industry .............................................. 19
2.2.2
Contribution of clients to industry performance .................................. 21
2.2.3
Contribution of lack of integration to performance ............................. 22
2.2.4
Construction workers’ challenges ........................................................ 23
2.3.
H&S IN CONSTRUCTION INDUSTRY .................................................... 24
2.4.
H&S PERFORMANCE OF THE CONSTRUCTION INDUSTRY ............ 27
2.4.1
H&S Performance at International level .............................................. 27
2.5
H&S IN DEVELOPING COUNTRIES & SOUTH AFRICA ...................... 31
2.6
CHAPTER SUMMARY ............................................................................... 33
CHAPTER THREE ................................................................................................... 35
H&S PERFORMANCE IMPROVEMENT AND CULTURE .............................. 35
3.1
INTRODUCTION ......................................................................................... 35
3.1.1
Performance improvement ................................................................... 35
3.1.2
Measuring H&S performance improvement ........................................ 37
3.2
TECHNIQUES TO IMPROVE H&S PERFORMANCE ............................. 38
3.2.1
Influence network technique ................................................................ 38
3.2.2
Climate / Perception surveys ................................................................ 41
3.2.3
Behavioural audits ................................................................................ 42
3.2.4
Conformance to H&S systems ............................................................. 42
3.2.5
Developing a H&S culture ................................................................... 43
3.2.6
Improvement through design ............................................................... 44
3.3
CULTURE .................................................................................................... 45
3.3.1
Why H&S culture? ............................................................................... 46
3.3.2
What is culture? .................................................................................... 50
3.3.3
Climate and culture .............................................................................. 56
3.3.4
What are the factors of H&S culture? .................................................. 58
3.4
EXTERNAL ENVIRONMENT’S INFLUENCE ON H&S CULTURE ..... 66
3.4.1
Legislative ............................................................................................ 66
3.4.2
Economic .............................................................................................. 70
3.4.3
Political influence ................................................................................. 72
3.5
CHAPTER SUMMARY ............................................................................... 73
CHAPTER FOUR ...................................................................................................... 74
INFLUENCE OF CONSTRUCTION CLIENTS ON PROJECT H&S
PERFORMANCE ...................................................................................................... 74
4.1
CLIENTS’ ROLE IN H&S PERFORMANCE ............................................. 74
4.2
ROLE OF CLIENTS IN DESIGNING FOR H&S ....................................... 80
xii
4.3
CHAPTER SUMMARY ............................................................................... 85
CHAPTER FIVE ....................................................................................................... 86
RESEARCH METHODS .......................................................................................... 86
5.1
INTRODUCTION ......................................................................................... 86
5.2
RESEARCH DESIGN .................................................................................. 87
5.3
METHODS.................................................................................................... 89
5.3.1
Literature review .................................................................................. 91
5.3.2
Delphi method ...................................................................................... 92
5.3.3
Questionnaire survey .......................................................................... 111
5.4
CHAPTER SUMMARY ............................................................................. 132
CHAPTER SIX ........................................................................................................ 133
RESULTS FROM THE DELPHI STUDY ............................................................ 133
6.1
INTRODUCTION ....................................................................................... 133
6.2
FINDINGS FROM THE DELPHI STUDY................................................ 134
6.2.1
Impact of the external environment on client H&S culture – D1 ...... 134
6.2.2
Client influence on H&S consideration in the project lifecycle - D2 137
6.2.3
Impact of client H&S culture on contractor H&S performance - D3 140
6.2.4
Impact of client H&S culture on designer H&S performance - D4 ... 142
6.2.5
Direct client’s influence on project H&S performance - D5.............. 153
6.3
DISCUSSION OF DELPHI RESULTS ...................................................... 154
6.3.1
Objective D1 ...................................................................................... 154
6.3.2
Objective D2 ...................................................................................... 156
6.3.3
Objective D3 ...................................................................................... 157
6.3.4
Objective D4 ...................................................................................... 159
6.3.5
Objective D5 ...................................................................................... 160
6.4
CHAPTER SUMMARY ............................................................................. 161
CHAPTER SEVEN .................................................................................................. 162
THE CONCEPTUAL CLIENT-CENTRED CONSTRUCTION H&S
PERFORMANCE IMPROVEMENT MODEL.................................................... 162
7.1
INTRODUCTION ....................................................................................... 162
7.2
CLIENT-CENTRED H&S IMPROVEMENT MODEL ............................ 166
7.3
CHAPTER SUMMARY ............................................................................. 167
CHAPTER EIGHT .................................................................................................. 168
QUESTIONNAIRE SURVEY RESULTS ............................................................. 168
xiii
8.1
INTRODUCTION ....................................................................................... 168
8.2
STATISTICS ON SEM ASSUMPTIONS .................................................. 170
8.2.1
Outliers and Missing data ................................................................... 170
8.2.2
Distribution characteristics of the data ............................................... 170
8.2.3
Identifiability of the model ................................................................. 171
8.3
FIT STATISTICS ON MEASUREMENT MODELS (CFA)..................... 173
8.3.1
Client H&S culture ............................................................................. 173
8.3.2
Project H&S Performance .................................................................. 186
8.3.3
Contractor H&S performance ............................................................ 193
8.3.4
Designers H&S Performance ............................................................. 199
8.3.5
External environment ......................................................................... 206
8.4
CONSTRUCT VALIDITY OF THE SEM MODEL .................................. 212
8.5
FIT STATISTICS ON THE STRUCTURAL MODEL .............................. 212
8.5.1
Model 1.0 – External influence on client H&S culture ...................... 213
8.5.2
Model 2.0 – Client’s influence on project H&S performance ........... 220
8.5
CHAPTER SUMMARY ............................................................................. 233
CHAPTER NINE ..................................................................................................... 235
DISCUSSION OF RESULTS ................................................................................. 235
9.1
QUESTIONNAIRE SURVEY RESULTS ................................................. 235
9.1.1
External environment’s influence on client H&S culture .................. 235
9.1.2
Influence of client on contractor H&S performance (H2) ................. 238
9.1.3
Influence of client on designer H&S performance (H3) .................... 241
9.1.4
Influence of client on project H&S performance (H4, H7 and H8) ... 243
9.1.5
Influence of contractor on project H&S performance (H5) ............... 245
9.1.6
Influence of designer on project H&S performance (H6) .................. 246
9.2
QUESTIONNAIRE AND DELPHI SURVEY RESULTS ........................ 246
9.2.1
Direct external environment’s influence on client H&S culture ........ 246
9.2.2
Direct client’s influence on contractor H&S performance ................. 247
9.2.3
Direct client’s influence on designer H&S performance ................... 248
9.2.4
Direct client’s influence on project H&S performance ...................... 248
9.3
CHAPTER SUMMARY ............................................................................. 249
CHAPTER TEN ....................................................................................................... 250
CONCLUSIONS ...................................................................................................... 250
10.1
10.1.1
CONCLUSION ON THE OVERALL STUDY ...................................... 250
Objective GO1 .................................................................................... 250
xiv
10.1.2
Objective GO2 .................................................................................... 251
10.1.3
Objective GO3 .................................................................................... 251
10.1.4
Objective GO4 .................................................................................... 251
10.1.5
Objective GO5 .................................................................................... 252
10.2
CONTRIBUTION AND VALUE OF THE RESEARCH ...................... 252
10.2.1
Methodological ................................................................................... 252
10.2.2
Theoretical .......................................................................................... 253
10.2.3
Practical .............................................................................................. 254
10.3
IMPLICATIONS FOR PRACTICE ........................................................ 256
10.4
RECOMMENDATIONS......................................................................... 256
10.4.1.
Methodological ................................................................................... 256
10.4.2
Theoretical .......................................................................................... 257
10.4.3
Practical .............................................................................................. 258
10.5
LIMITATIONS ....................................................................................... 259
10.6
SUGGESTIONS FOR FURTHER RESEARCH .................................... 259
10.7
CHAPTER SUMMARY AND CONCLUSION ..................................... 260
REFERENCES ......................................................................................................... 262
APPENDIX A ........................................................................................................... 286
APPENDIX B ........................................................................................................... 287
APPENDIX C ........................................................................................................... 288
APPENDIX D ........................................................................................................... 302
APPENDIX E ........................................................................................................... 321
APPENDIX F ........................................................................................................... 322
APPENDIX G ........................................................................................................... 334
APPENDIX H ........................................................................................................... 335
APPENDIX J ............................................................................................................ 343
xv
LIST OF TABLES
Table 1: Incidence rate of accidents at work in EU-15 ................................................ 28
Table 2: Accidents in South Africa's construction industry ........................................ 32
Table 3: Fatality rate in South Africa .......................................................................... 33
Table 4: Definitions of culture ..................................................................................... 53
Table 5: Common factors of H&S culture ................................................................... 62
Table 6: Indicators of factors of H&S Culture............................................................. 67
Table 7: Research procedure followed in current study .............................................. 90
Table 8: Delphi panel composition by Regions ........................................................... 95
Table 9: Delphi panel composition by highest qualifications ...................................... 96
Table 10: Delphi panel composition by number of years of experience ..................... 97
Table 11: Number of publications by Delphi panel members ..................................... 98
Table 12: Calculation of Likelihood and Impact significance ................................... 106
Table 13: Impact significance and severity rating scale ............................................ 107
Table 14: Likelihood rating scale .............................................................................. 107
Table 15: MAD, range and percentage agreeing values ............................................ 109
Table 16: Cut-off criteria of fit statistics.................................................................... 128
Table 17: Respondents profile ................................................................................... 131
Table 18: Contract value of projects reported on in this study .................................. 131
Table 19: Project type reported on in this study ........................................................ 131
Table 20: Client organisations represented in the survey .......................................... 132
Table 21: Contractor selection method reported by respondents ............................... 132
Table 22: Influence level of various project stakeholders ......................................... 154
Table 23: Univariate and Mardia’s normalised Multivariate estimates ..................... 172
Table 24: Postulated model variables for client culture............................................. 176
Table 25: Average absolute residuals for measurement models ................................ 177
Table 26: Robust fit indexes for client H&S culture construct .................................. 179
Table 27: Parameter estimates of client H&S culture measurement model .............. 181
Table 28: Factor loading and Z-statistics of client model.......................................... 182
Table 29: Correlations between factors of client H&S culture .................................. 185
Table 30: Reliability and construct validity of client H&S culture model ................ 185
Table 31: Indicator variables for the project H&S Performance construct................ 187
Table 32: Robust fit indexes for project H&S Performance construct ...................... 189
xvi
Table 33: Factor loading and Z-statistic of Project H&S performance model .......... 190
Table 34: Reliability and construct validity of Project H&S performance model ..... 192
Table 35: Postulated contractor H&S performance measurement model .................. 194
Table 36: Robust fit indexes for Contractor H&S Performance construct ................ 196
Table 37: Factor loading and Z-statistics of contractor H&S performance model .... 197
Table 38: Correlations of factors of contractor H&S performance model ................ 198
Table 39: Reliability and construct validity of contractor H&S performance model 199
Table 40: Postulated designer H&S Performance model .......................................... 202
Table 41: Robust fit indexes for designer H&S performance construct .................... 203
Table 42: Factor loading and Z-statistics of designer H&S performance model ...... 204
Table 43: Correlations of factors of designer H&S performance model ................... 204
Table 44: Reliability and construct validity of designer H&S performance model .. 205
Table 45: Postulated external environment influence model ..................................... 206
Table 46: Residual values for external environment influence model ...................... 208
Table 47: Robust fit indexes for external environment influence construct .............. 209
Table 48: Factor loading and Z-statistics of the external environment model .......... 210
Table 49: Reliability and construct validity of the environment factor ..................... 211
Table 50: Reliability and construct validity of the client H&S culture model .......... 214
Table 51: Robust fit indexes for SEM Model 1.0 ...................................................... 218
Table 52: Model 1.0 factor loadings and Z-statistic .................................................. 219
Table 53: Robust fit indexes for SEM Model 2.0 ...................................................... 224
Table 54: Parameter estimates and test statistic for model 2.0 .................................. 228
Table 55: Specific indirect effects of client H&S culture .......................................... 232
Table 56: Total indirect effects of client H&S culture .............................................. 232
xvii
LIST OF FIGURES
Figure 1: Fatal injuries in all sectors in UK ................................................................. 29
Figure 2: Fatal injuries in UK’s construction industry ................................................ 30
Figure 3: Fatal accidents in USA ................................................................................. 31
Figure 4 : Influence Network Technique ..................................................................... 39
Figure 5: Relationship between interventions and number of incidents ...................... 47
Figure 6: H&S culture ladder ....................................................................................... 57
Figure 7: H&S culture framework ............................................................................... 58
Figure 8: Research design outline ................................................................................ 91
Figure 9: Delphi panel members’ contribution to publications ................................... 98
Figure 10: Delphi process .......................................................................................... 102
Figure 11: Plot of MAD and range to determine consensus ...................................... 109
Figure 12: Percent of panellists with similar inclination on an issue ........................ 109
Figure 13: External environment construct................................................................ 116
Figure 14: Client H&S culture measurement model .................................................. 117
Figure 15: Contractor H&S performance measurement model ................................. 118
Figure 16: Designers H&S Performance measurement model .................................. 118
Figure 17: Project H&S performance measurement model ....................................... 119
Figure 18: Questionnaire survey procedure ............................................................... 124
Figure 19: Impact significance of external environment factors to client culture ..... 136
Figure 20: Client likelihood to implement H&S elements ........................................ 136
Figure 21: Client likelihood to implement H&S elements ........................................ 137
Figure 22: Impact significance of factors of client H&S culture ............................... 138
Figure 23: Likelihood of H&S consideration due to client’s influence ..................... 139
Figure 24: Impact significance of client factors on contractor H&S performance .... 140
Figure 25: Contractor likelihood to implement H&S elements ................................. 141
Figure 26: Impact significance of client factors on designer H&S performance ...... 144
Figure 27: Overall impact significance of client H&S culture factors ...................... 145
Figure 28: Designer likelihood to implement H&S elements due to client influence146
Figure 29: Designer & contractor likelihood to implement H&S elements .............. 147
Figure 30: Impact of client factors on designer and contractor H&S performance ... 148
xviii
Figure 31: Impact significance of factors of designer H&S culture .......................... 149
Figure 32: Impact of client and designer H&S culture on H&S consideration ......... 150
Figure 33: Likelihood of H&S consideration due to designers' influence ................. 152
Figure 34: Likelihood of H&S consideration due to client and designer influence .. 152
Figure 35: Theorised client centred H&S performance improvement ....................... 167
Figure 36: Hypothesised model-Client influence on project H&S performance ....... 168
Figure 37: Environmental influence on client H&S culture ...................................... 169
Figure 38: Theorised six factor client H&S culture model ........................................ 174
Figure 39: Contractor H&S performance construct ................................................... 195
Figure 40: Designer H&S performance measurement model .................................... 200
Figure 41: External environment construct................................................................ 207
Figure 42: Model 1.0 - External environment influence on client H&S culture ....... 216
Figure 43: Model 2 - Client influence on project H&S performance ........................ 221
Figure 44: Finalised model 2.0 for client H&S culture influence.............................. 233
Figure 45: Finalised Client-centred H&S performance improvement model ............ 234
xix
LIST OF APPENDICES
APPENDIX A
Invitation letter to participate in a Delphi study
APPENDIX B
Request to experts to submit their curriculum vitae
APPENDIX C
Delphi instructions for round 1 and questionnaire
APPENDIX D
Delphi instructions for round 2 and an example of completed
questionnaire showing group medians
APPENDIX E
Delphi instructions for round 3
APPENDIX F
Introduction letter and questionnaire for industry survey
APPENDIX G
Delphi Mean Absolute Deviations (MAD)
APPENDIX H
Model 2.0 covariance matrix
APPENDIX J
Model 1.0 covariance matrix
xx
DEFINITION OF TERMS AND CONCEPTS USED IN THE
THESIS
Client
Client is the party for whom construction work is performed and is also referred to as
the owner. The two terms have been used interchangeably to mean one and the same
person (Hoonakker, Loushine, Carayon, Kallman, Kapp & Smith, 2005:467; Huang &
Hinze, 2006a:164; Tang, Qiang, Duffield, Young & Lu, 2008:458; Ulang, Gibb &
Anumba, 2009:2). It refers to the person or entity that procures the services of a
contractor to construct a facility and enters into an agreement with a contractor to
construct a facility. Clients can be once-off or repeated procurers of work.
Climate
Climate is the “descriptive measure reflecting a group’s perceptions of the
organisational atmosphere and the way things are done within a group or
organisation” (Flin, Mearns, O’Connor and Bryden, 2000:178). H&S climate reflects
employees’ perceptions about the organisation’s H&S. It is a snapshot that describes
the way things are done (Choudhry, Fang & Mohamed, 2007:208). It is also referred
to as the indicator of the prevailing culture in an organisation (INSAG, 1991:4; Flin et
al.,, 2000:178; IOSH, 2004:7; Wamuziri, 2006:169)
Culture
Culture is referred to as the characteristic set of beliefs, values and attitudes held and
shared by all members of a group or organisation which influences and can be seen
from the behaviour patterns evidenced through observation or description of what
goes on by those that are part of the organisation. H&S Culture is therefore a subset of
the organisational culture (Cooper, 2000:113; Hudson, 2007:698; Wiegmann, Zhang,
Thaden, Sharma & Mitchell, 2002:11; Wamuziri, 2006:167)
Delphi
The Delphi Technique is a structured group interaction that works through rounds of
opinion collection and feedback from the researcher. Each round is composed of a
written survey conducted through administration of a questionnaire either
electronically or hard copy, followed by feedback to the respondents by the researcher
of the statistical scores for each survey question. After each round the respondents are
xxi
surveyed again to determine whether their opinions have shifted after seeing the
statistical results from the prior rounds. As a result of the process there is typically a
convergence of opinion. This is achieved usually after three or four rounds and a
stabilised opinion emerges. Group opinion on each question may reflect agreement,
disagreement or some of each (Pivo, 2008:23)
Designer
The term designer comes from the word design. To design in relation to any structure
means any hard copy or electronic drawing, design detail, design instruction, scope of
works document or specification relating to the structure. A designer is
1. Anyone who specifies or alters a design, or who specifies the use of a
particular method of work or material such as a quantity surveyor who insists
on a certain material;
2. Architects, building designers, geotechnical engineers, civil structural
engineers, building surveyors, building services engineers, landscape
architects and all other design practitioners contributing to or having overall
responsibility for any part of the design (Queensland Government, 2007:8).
The term also refers to those individuals that arrange for people under their control to
prepare designs for a structure such as project managers (Griffiths & Griffiths,
2011:65).
Health and Safety (H&S)
Health refers to occupational health and is defined as the wellness of workers relative
to their occupation or place of work (ILO, 1983:1491). Safety, on the other hand, is
the prevention of risk to the health of workers, and visitors, a work place and to
property (ILO, 1983:1491). It is freedom from unacceptable risk (Kinnersley &
Roelen, 2007:32).
Impact significance
According to the Oxford dictionary, impact is described as a noticeable effect or
influence (2001:447) whereas significance is the importance of something or large
enough to have an effect or be noticed (2001:838). Impact significance therefore, is
xxii
defined as a measure of the level of effect of an influencing factor. It describes
whether the impact is large enough to have an effect or be noticed. In this study, the
magnitude of the impact significance is defined by an ordinal scale of 0 to 10, with ‘0’
being low impact significance and ‘10’ being critical impact significance.
H&S commitment
H&S commitment, especially by clients or contractors management, is shown when
they undertake or perform inter alia the following:
1. Make useful regular visits to construction sites on H&S matters;
2. Discuss H&S matters with staff;
3. Spend time and money on H&S;
4. Do not tolerate violations of H&S procedures and actively try to improve
systems so as to discourage violations (Loughborough & UMIST, 2004:23).
It is also shown through participation in the organisation’s H&S Management
processes, with the allocation of H&S responsibilities throughout the project
organisation (Lingard, Blismas, Cooke & Cooper, 2009:134).
Competence
Competence is a performance shaping factor which influences capacity. Competence
is also associated with knowledge, conferred by training and instruction of personnel
and by their self-education (INSAG, 1991:5).
H&S communication
H&S communication is achieved when information is circulated to all stakeholders
about risks, ensuring that mechanisms are established to convey and record H&S risks
throughout the project life cycle facilitating bottom-up communication of H&S issues.
It is also a consultative process that enables worker participation in the making of
decisions that impact on H&S (Lingard et al., 2009:135). Communication can be
achieved through good, clear, concise and relevant written materials such as letters,
specifications, and posters. It can also be achieved through good briefings on current
issues day to day and in formal H&S meetings, listening and providing feedback.
xxiii
H&S involvement
H&S involvement is shown through the extent to which stakeholders get personally
involved in critical H&S activities, their presence and contribution to H&S meetings,
and planning sessions. It is also categorised by the extent to which there is good
communication about H&S issues from and between all stakeholders (Human
Engineering, 2003:14).
H&S leadership
H&S leadership is demonstrated through the incorporation of H&S considerations at
every level of decision making, through having a policy on H&S, active monitoring,
control and coordination of all stakeholders on a project. Leadership is further shown
in the statements made concerning H&S, the questions, the items on the agenda and in
the criteria used to evaluate individuals or organisations or plant (Simon & Frazee,
2005).
H&S procedures
H&S procedures are also referred to as rules. They include systems, policies and
plans, procedures relating to monitoring, reporting, and analysis of H&S risks
(Human Engineering, 2003:4).
Lagging indicators
Lagging indicators are measures of H&S performance (Lingard et al., 2009:136;
IOSH, 2010:7) that are based on accidents, incidents or compensation cost records on
construction projects (Flin et al., 2000:177; Choudhry et al., 2007:1006). They are
also referred to as trailing indicators and are linked to the outcome of an accident
(Toellner, 2001:42). They are reactive in nature.
Leading indicators
Leading indicators are metrics associated with measurable system or individual
behaviours linked to accident prevention (Toellner, 2001:44). Leading indicators
therefore include measures such as H&S audits or the measurement of H&S climate
and are described as predictive measures (Flin et al., 2000:177). The H&S climate is
also referred to as a “leading indicator” (Seo, Torabi, Blair, & Ellis, 2004:429). They
are proactive in nature.
xxiv
H&S Performance
H&S performance entails complying with H&S standards, procedures, policies, goals
and or regulations as well as participation in H&S programmes or initiatives. It is
determined by knowledge, skill and motivation of an individual (Griffin & Neal,
2000:349). H&S performance measurement can be done by evaluating several
indicators such as the H&S culture (Arezes & Miguel, 2003:25).
Severity
Severity is referred to as the consequence of an event (de Leur & Sayed, 2003:713). In
terms of a hazard or harm, it is the seriousness of the harm that could result from
contact or exposure to the hazard. Therefore, severity is the degree of magnitude
associated with an incident or event (Hallowell, 2008:9).
Severity is therefore the degree of magnitude of consequence of an incident or event
happening and its units may be a subjective measure of negative impact. This
subjective measure could be rated as negligible, minor, moderate, major or critical.
xxv
CHAPTER ONE
INTRODUCTION
1.1
BACKGROUND
Construction health and safety (H&S) performance improvement has in recent years
become a priority and therefore has gained industry-wide attention (Cheung, Cheung
& Suen, 2004:159; Hamalainen, Saarela & Takala, 2009:125). This development has
been in part due to the poor H&S performance of the industry, introduction of major
pieces of legislation coupled with increased personal responsibility and accountability
of senior managers and organisations for H&S (Fitzgerald, 2005:324). Other reasons
include a need to develop a good or better image of the construction industry (Misnan
& Mohamed, 2007:401) and in some ways to address the H&S record which in
comparison to many industries is undesirable. For larger multi-national organisations,
the need for H&S improvement has been a corporate social responsibility issue and a
driver to improve their H&S performance (Smallman & John, 2001:237).
Apart from the legislative pressure, the debate concerning the personal responsibility
that senior managers should bear for their organisations on H&S failures has resulted
in most organisations placing focus on H&S improvement (Fitzgerald, 2005:324).
Cost is a further reason why more and more organisations are beginning to focus on
H&S (Hamalainen, Saarela & Takala, 2009:125).
However, improving H&S performance in the construction industry has proved to be
somewhat challenging partly due to the industry’s complex character. Despite the
complexity of this industry, H&S performance improvement remains a crucial issue
and its importance or need has been demonstrated in numerous studies (Smallman &
John, 2001; ILO, 2003; Hoonakker et al., 2005; Lee, Halpin & Chang, 2006).
It is economically important that H&S should be improved in the construction
industry. Poor H&S performance is costly and can impact negatively on an industry
and indeed on an economy. It is estimated that the costs of accidents account for about
1
four Percent of the global Gross Domestic Product (GDP), (ILO, 2003:15). Egan
(1998:15) estimates that accidents can account for about three to six Percent of total
construction project costs. Studies conducted in Europe among members of the
European Union in 2002 on costs of accidents, revealed that as a percentage of the
GDP they could be as high as eight-and-a-half Percent (Karjalainen, 2004:3). In the
United Kingdom (UK), the Health and Safety Executive (HSE) report of 2008/09,
indicated that 1.2 million people who had worked during that year were suffering
from an illness, both long- standing and new cases caused or worsened by their
current or past work, equating to 3,900 per 100, 000 people or three-and-nine-tenths
Percent (HSE, 2010a:2). In terms of costs, occupational ill health and injury
accounted for almost three Percent (Wright, 2007:161). In South Africa, it was
estimated that occupational injuries and diseases accounted for about three-and-a-half
Percent of GDP (Republic of South Africa, 2003:4). In Botswana, the Botswana
Federation of Trade Unions (BFTU) estimates that occupational injuries and fatalities
account for over three Percent of GDP (BFTU, 2007:14). H&S performance
improvement is therefore a fundamental issue because it is aimed at eliminating or
reducing the risk of accidents and its severity in the construction industry.
The construction industry is an important industrial sector of a national economy. In
highlighting the importance of the construction industry to national economies, Egan
(1998:6) argued that the construction industry was a pillar of the economy. Murie
(2007:3) reported that globally, the construction industry is a USD three trillion (ZAR
21 Trillion) giant industry. It is estimated that the construction industry accounts for
about ten Percent of the global GDP (Murie, 2007:5). In Botswana, the construction
industry contributes about seven Percent to the National GDP (World Bank,
2008:online) and about five Percent in South Africa (CIDB, 2004:5).
Furthermore, the construction industry is one of the largest employers globally.
According to Murie (2007:5), the construction industry employs about 180 million
people or about seven Percent of global employment. In South Africa the construction
industry employed about seven-and-nine-tenths Percent of the total labour force
(Statistics South Africa, 2011:Viii). In Botswana, the construction industry was the
third largest employer (CSO, 2009:6). The construction industry is clearly an
important sector of most economies including that of South Africa and Botswana.
2
Because of the importance of the construction industry, improving H&S performance
in the industry is also as compelling as the other highlighted reasons.
Despite the industry’s importance, its image in terms of H&S can only be described as
poor. The image of the construction industry has partly been spoilt by poor H&S
performance. It is reported that workers in the construction industry are more at risk
of an accident, ill health and or even a fatality at work place than other manufacturing
based industries (Loughborough & UMIST, 2003:vii; Hoonakker et al., 2005:461).
Likewise, generally construction sites are still one of the most dangerous workplaces
because of the high incidence of accidents (Teo et al., 2005:329; Kines et al.,
2007:53).
In the construction industry, the risk of a fatality is at least five times more likely than
in other manufacturing based industries (Sawacha et al., 1999:309; Loughborough &
UMIST, 2003:6). According to Bomel (2001:0.5) the construction industry is a
hazardous environment where workers have direct exposure to heights, forces, and
power. Workers face these risks every day of their working lives. Of great concern,
therefore, is the exposure of workers to hazards that are in construction projects.
Consequently, H&S performance improvement has to do with the elimination of
hazards that workers are constantly exposed to. Apart from these hazards,
construction workers have also been subjected to various challenging situations. The
construction industry is a physically demanding work-place requiring resilience and
strength (Bomel, 2001:2.3). Coupled with the physically demanding nature of the
industry, the construction industry offers low status, low pay, short term employment,
unregistered, informal and hazardous jobs in a highly fragmented industry (Murie,
2007:5; Kulchartchai & Hadikusumo, 2010:47). Furthermore, construction workers
are faced with exploitative employment practices, and hardships experienced by
especially migrant workers (Murie, 2007:5).
Therefore, H&S performance improvement in the industry should be a priority. It is
warranted that research should be encouraged on this matter given the importance and
dangerous nature of the industry. In addition, much research on H&S performance
improvement in the industry is justified to improve conditions in which construction
workers operate.
3
Consequently, a number of studies have been conducted on the subject of H&S
performance improvement. However most studies on H&S performance have tended
to focus on understanding the causal factors underlying construction accidents, such
as the studies conducted by Mansingh & Haupt (2008), Bomel (2001), and
Loughborough University & UMIST (2003). Other studies have focused on
addressing H&S at the construction stage and on issues such as the use of incentives
to improve contractor performance (Tang et al. 2008), and designers’ roles and
responsibilities (Kartam, Flood & Koushki, 2000). Furthermore, studies that address
procedures and systems at the construction stage as well as behavioural issues
surrounding workers have been conducted (Goodrum & Gangwar, 2004; Cameron &
Duff, 2007).
In the recent past various studies have been conducted on improving H&S
performance through improving the H&S culture of mostly contracting organisations
(Dingsdag et al., 2006; Chinda & Mohamed, 2008; Zhou, Fang & Wang, 2008). The
studies on H&S culture have been complimented by studies on behaviour based H&S
performance (Duff, Robertson, Phillips & Cooper, 1994; Lingard & Rowlinson, 1997;
Petersen, 2000; Keil Centre, 2000; Cooper, 2009) and H&S climate (Zhou, Fang &
Mohamed, 2011).
Despite these numerous studies on H&S performance improvement in the
construction industry, few studies have addressed the role,
contribution,
responsibilities and influence of construction clients on H&S performance. There is
little evidence that studies have been conducted on the role of construction clients
apart from the one conducted by Huang and Hinze (Huang & Hinze, 2006a:164;
Lingard et al., 2009:132). They analysed project characteristics, owners’ safety
requirements in contract documents, owners’ selection of safe contractors and
owners’ participation in safety management and related these factors to H&S
performance. H&S performance was expressed in terms of the total recordable injury
rate (TRIR) which is the total number of Occupational Safety and Health Authority
(OSHA) recordable injuries for every 200,000 worker hours (Huang & Hinze,
2006a:167). The TRIR records were provided by respondents on the projects they
investigated. Owner representatives were interviewed in these construction projects.
4
Specifically Huang & Hinze (2006a) tried to establish whether H&S performance was
influenced by the
1. size of the construction project, labour arrangements, or type of the project;
2. selection criteria of contractors;
3. provision of contractual safety requirements; and
4. owner involvement through participation in H&S recognition programmes,
monitoring safety performance, funding H&S initiatives, investigating
accidents and conducting or requiring H&S training.
Projects where respondents indicated that the owner was involved were compared
with records of TRIR obtained for the particular project. It is interesting to note that
the average TRIR for all the contracting organisations selected for participation was
below the industry median value of 7.80 recorded in 2001(Huang & Hinze,
2006a:169). The median TRIR of the selected contracting organisations was 1.48 and
therefore were considered to be relatively safe. It was concluded that owners could
positively influence project H&S performance because the TRIR was consistently
below 1.50 for projects where owners had a safe contractor selection criterion,
provided contractual H&S requirements and was involved in H&S management. It
was found that owners’ management commitment, H&S observation programmes,
personal accountability, safety communication, implementation of H&S programmes,
safety inspections, H&S and constructability reviews of designs and H&S culture
were important ways to improve H&S performance.
However, Huang & Hinze’s findings may not be conclusive because many authors
have highlighted the limitations in using accident records as H&S performance
indicators, from which the TRIR they used is calculated. The TRIR is a lagging
indicator and lagging indicators are problematic indicators of H&S performance
(HSE, 2001:5; Toellner, 2001:42; Jafri, Ahmad & Kamsah, 2005:707). Consequently
there has been an increasing move away from relying only on lagging indicators such
as the TRIR (Flin et al., 2000:177; Svedung & Rasmussen, 2002:398; Galvin,
2005:254; Huang & Hinze, 2006a:169; Choudhry, et al., 2007:209). Therefore, this
limitation demands an alternative approach to investigate clients’ influence on H&S
5
performance. Furthermore, it is necessary to investigate other factors such as the
influence of client H&S commitment, H&S observation programme, personal
accountability, communication, implementation of the H&S programmes, H&S
inspections, H&S reviews of designs and H&S culture.
Therefore this study presents an evaluation of the impact of client H&S culture on
project H&S performance. Client H&S culture has been hypothesized as an
independent variable having influence on the dependent variable of project H&S
performance. In this study, H&S performance outcomes have been taken to be factors
of H&S culture since H&S culture can be used to measure performance
(Guldenmund, 2000:254; Arezes & Miguel, 2003:21; Wamuziri, 2006:169). Good
H&S performance is reflected by a positive H&S culture and H&S culture can be
determined by assessing H&S climate of a construction project. The influence of
client H&S culture defined by six factors including the aspects of management
commitment, H&S
observation
programmes, personal
accountability, H&S
communication, implementation of H&S programmes, H&S inspections, and design
reviews for H&S was evaluated. In addition to these, the influence of the leadership
and competence factor was also evaluated.
The approach used in this study was aimed at evaluating the influence of factors of
client H&S culture namely: leadership, involvement, procedures, commitment,
communication and competence on project H&S performance. The following
procedure was adopted, namely:
1. A review of literature on H&S in general and H&S performance improvement
in particular. The review of literature on H&S roles and responsibilities of key
stakeholders in construction was conducted;
2. A Delphi study to explore the influence of various key stakeholders on H&S
performance. A conceptual model was theorised based on the literature review
and Delphi findings;
3. A questionnaire survey among construction professionals working on current
or on projects that had just been completed. Construction professionals include
persons working for major stakeholders of the construction industry, namely:
6
contractors, clients and designers. The survey instrument was based on the
hypothesised relationships between variables in the conceptual model.
4. Client H&S culture was analysed and related to project H&S performance;
5. Conclusions were made based on the findings from the questionnaire survey.
Findings relate to the goodness-of-fit of the hypothesised conceptual model in
relation to the sample data and how statistically significant these relations
were. A final client-centred H&S performance improvement model was
thereafter developed.
1.2
THE PROBLEM STATEMENT
H&S performance improvement is a very important issue because of the human
considerations, economic benefits that result from a better performance and the
importance of the construction industry to the national economy. The H&S
performance of the industry has consistently been poor and consequently the
construction industry has developed an undesirable H&S image compared to other
industrial sectors.
Previous studies on H&S performance improvement have focused on issues to do
with mainly contracting and design organisations, H&S management systems and
procedures and more recently with theories of H&S culture and climate. A review of
literature seems to suggest that only one study has investigated the role and influence
of construction clients. Moreover H&S in the construction industry is increasingly of
major concern and lags behind other industries.
Therefore the problem that has been addressed in this study may be stated as follows:
Given that contractors and designers may not be relied on to drive H&S performance
improvement without the involvement of clients, the limited research into the overall
impact and influence of the direct and holistic active involvement of clients in H&S
performance, and the absence of a client-centred construction H&S performance
model the achievement of incremental construction H&S performance improvement
is unlikely.
7
1.3
THE STUDY
1.3.1
The general objectives of the study
The primary objectives of the study will be to inter alia
GO1. establish the status of H&S in the construction industry, the role of clients,
designers and the top management of contractors in H&S performance;
GO2. examine current trends in H&S performance improvement;
GO3. establish the role and impact of construction clients and their potential
contribution to and influence on project H&S performance;
GO4. develop a client-centred model for H&S performance improvement in the
construction industry; and
GO5. validate the conceptualised client-centred model for H&S performance
improvement by comparing the Delphi and literature review outcome with the
field questionnaire survey outcome.
Objectives GO1 to GO2 will be achieved through a review of literature on H&S
performance improvement. The review will provide a theoretical framework of what
the trends are in H&S performance improvement and what the status of H&S is in the
construction industry. The general objectives GO3 and GO4 will be achieved by
conducting a Delphi study. The Delphi method will be explained in detail in the
methods section and the reasons why this method was opted to other methods will
also be explained. The final general objective GO5 will be achieved by conducting a
field questionnaire survey and modelling of the results using structural equation
modelling. Software packages namely, EQS and MPlus will be used for structural
equation modelling.
1.3.2
Methods
A mixed method of qualitative and quantitative data collection was used. For the
qualitative part, a Delphi study was conducted and for the quantitative method, a field
questionnaire survey was used.
8
Data needed and means of obtaining it (empirical measures)
Literature on H&S performance improvement was reviewed to provide a background
to the study. Various sources were reviewed including books, articles in accredited
Journals, published and unpublished works such as dissertations and the web based
publications on the subject.
Two methods were used to collect empirical data. These methods were the Delphi and
field questionnaire survey methods. A detailed description of these methods will be
presented later in the text.
In the Delphi method, the data that needed to be collected was the ratings of impact of
factors of client H&S culture on construction project H&S performance. In addition,
ratings were also made on the likelihood of H&S elements or programmes being
implemented on a construction project as a result of client’s influence. This data was
obtained through the use of questionnaires. Experts were asked to complete the
questionnaires and reach consensus on the rated likelihoods and severity of various
H&S factors. The process involved a three round iterative process with the main aim
of getting experts to reach consensus on the questions raised in the questionnaires.
Experts were also encouraged to give reasons for their dissenting views.
In the questionnaire survey, the data that needed to be collected was the evidence of
factors of client H&S culture namely: leadership, involvement, procedures,
commitment, communication and competence. In addition, data regarding H&S
performance at project level was needed. The H&S performance was determined from
perceptions of respondents who included all construction project professionals such as
engineers, quantity surveyors, architects, safety officers and project managers. This
data was obtained with questionnaires. These were completed by construction
professionals working on on-going projects or those projects that had been completed
within the two years prior to this study.
Data source
In the Delphi study, data regarding the ratings on the impact of factors of client H&S
culture and the likelihood of H&S elements being implemented on a construction
project as a result of clients’ influence was obtained from the expert panel.
9
On the other hand, data from the questionnaire survey was obtained from construction
professionals working either for construction clients, designers or contracting
organisations. The data related to on-going construction projects or projects that had
been completed within the two years prior to this study.
Data analysis
Data obtained from the Delphi was analysed with Microsoft EXCEL, a spread-sheet
software. The output from the analysis was a set of descriptive statistics such as
means, median, standard deviations and derivatives of these statistics.
As for data obtained through the field questionnaire survey, structural equation
modelling (SEM) was utilised using the EQS and MPlus software packages. Outputs
from the analysis were univariate and multivariate descriptive statistics as well as
measures of goodness-of-fit of the hypothesised model. Other outputs included
measures of statistical significance of parameter estimates.
1.3.3
Results
Efforts directed at H&S performance improvement require an understanding of the
major influences on H&S. Results of the study related to the relationship between the
exogenous variable, client H&S culture, and the endogenous variable, project H&S
culture. These results will be presented as graphs and tables of values describing the
extent of the client H&S culture’s influence on project H&S performance.
Using data from both the Delphi and the questionnaire survey, a client-centred model
for H&S performance improvement was conceptualised and validated. This model
will be presented as the final output of the study.
Delphi specific objectives
The specific objectives for conducting a Delphi study were to establish the influence
and impact of the:
D1. external environment factors on client H&S performance;
D2. client H&S culture on H&S consideration (hazard identification, risk analysis,
assessment and mitigation) throughout the project life cycle;
10
D3. client H&S culture on contractor H&S performance (top management);
D4. client H&S culture on designer H&S performance; and
D5. client H&S culture on the overall project H&S performance.
The main outputs from the Delphi study will be the identification of the factors of
H&S culture with significant influence and a conceptual model defining client,
designer and contractor interrelationships.
The conceptual model will be validated by results from a field questionnaire survey.
Specific objectives of the field questionnaire survey
The specific objectives for conducting the questionnaire survey and thereby satisfy
the general objective GO6 of validating the conceptual model were to:
Q1. identify external environment factors that had a higher influence on client
H&S culture;
Q2. establish clients’ influence on designer H&S performance;
Q3. establish clients’ influence on contractor H&S performance
Q4. determine Clients’ influence on the overall construction project H&S
performance; and
Q5. determine the goodness–of–fit of the hypothesized client-centred-model to the
sample data.
1.3.4
Limitation and delimitation of the study
The study will focus on the influence of client H&S culture on construction project
H&S performance as well as on contractor and designer H&S performance.
A causal analysis between the exogenous variable, client H&S culture and the
endogenous variable, project H&S performance will be established through structural
equation modelling using the software EQS version 6.1 and MPlus version 6.0.
11
The study was restricted to the South African and Botswana’s construction industries.
Questionnaire surveys were conducted in Johannesburg and Gaborone. Information
was obtained regarding construction projects that were currently on-going and or had
been completed two years prior to the study.
1.3.5
Assumptions
The following assumptions were made namely:
1. H&S was managed to a varying degree on construction projects;
2. all construction professionals working on a particular project were aware of all
processes, procedures and were privy to contract information relating to the
construction projects that they were reporting on;
3. respondents were capable of responding to the research instruments as they
were considered to be knowledgeable individuals;
4. questionnaires were completed honestly within the boundaries of the
knowledge of the respondents;
5. H&S performance on construction projects was influenced by external factors
that include clients, the external environmental factors of legislative,
economics, politics, technology and social; and
6. some clients were involved in H&S management in construction projects.
1.3.6
Importance of the study
A gap exists in literature on the impact of client H&S culture on project H&S
performance. This study will therefore contribute to existing knowledge by
establishing the impact of client H&S culture on H&S performance. In addition,
instead of using lagging indicators such as the TRIR as indicators of H&S
performance, factors of H&S culture which are considered to be leading indicators
were evaluated as outcome variables. The study also used an innovative mixed
methodology of Delphi and structural equation modelling to analyse and model
client’s influence on H&S performance.
Therefore the study will add new knowledge on the impact of client H&S culture on
project H&S performance. The innovative methods and H&S performance measures
used in the study will also contribute to the existing body of knowledge on H&S
performance improvement.
12
In addition to the empirical study, a critical review of literature on H&S performance
improvement will expand existing knowledge by providing a synthesised literature
that will be useful for improving H&S performance in the industry.
H&S performance in the construction industry is poor and is of great concern.
Research conducted in Botswana by Musonda & Smallwood (2008) revealed that the
level of H&S awareness in the construction industry was low, H&S legislation was
not complied with, top management of contractors was not committed to H&S
implementation, there was a lack of H&S management systems, procedures, and
protocol, and clients and designers did not participate in the implementation of H&S.
Consequently, the construction industry has a poor image compared to other
industries. It has been described as hazardous. Its workers are constantly exposed to
these hazards. Workers also face added challenges due to being involved in physically
demanding jobs which are low paying and of low status. The construction industry is
dirty, difficult and dangerous (Kikwasi, 2008:56).
Despite the poor image associated with the construction industry, the construction
industry has also been described as a pillar to national economies (Egan, 1998:6). The
construction industry contributes about ten Percent to global GDP and about five
Percent to South Africa’s GDP. In Botswana the construction industry contributes
about seven Percent to the National GDP. The construction industry is also the third
largest employer both in Botswana and South Africa. Addressing H&S performance
in this industry is therefore logical for an industry that is so important and a major
employer.
1.3.7
Research organisation
This study is organised as follows:
Introduction (chapter 1)
This chapter presents information on the background to the study, a definition of
terms and key concepts that have been used in this study as well as the main research
problem. The chapter also presents a general description of the study stating the aim,
importance and objectives of the study. In addition, the assumptions that were made
in the study are also described. Furthermore, a description of methods that were used
13
to conduct this study, including the results that were obtained and ethical
considerations are discussed.
Literature review (chapters 2 to 4)
These chapters present the review of literature that informed this study. The literature
related to the H&S in construction, the status of H&S in construction, the concept of
H&S performance improvement, current trends on H&S performance improvement
strategies suggested in literature, the nature of the construction industry and the
concept of H&S culture. The clients’ role and H&S culture is also discussed.
Research methods (chapter 5)
This chapter contains a detailed description of the methods and the tools used to
collect data for this study. In addition, this chapter describes the participants to this
study as well as a detailed description of the results, analysis of the results and how
results are presented in the findings section. The above is done for every method of
data collection used in this study namely: the Delphi and field questionnaire survey
methods. Finally, the research design is also described in this chapter.
Chapter 5 also presents findings regarding the nature of projects that were reported on
by the respondents in the questionnaire survey as well as the profile of respondents to
the questionnaire survey.
Findings from the Delphi study (chapter 6)
This chapter discusses findings from the Delphi Method. These findings from the
Delphi study are presented relative to each Delphi objective. In addition, a discussion
and interpretation of the Delphi results is also presented at the end of this chapter.
Conceptual model -Client centred H&S performance improvement (chapter 7)
Chapter seven is a discussion of findings from the review of literature and the Delphi
study. This discussion forms the basis of the conceptual model’s theory. The
hypothesised client centred model for H&S performance improvement is presented in
this chapter. Chapter 7 describes the Client-centred H&S performance improvement
model in detail. It describes the variables of the model as well as the interrelationships between all variables of the model.
14
Findings from the questionnaire survey (chapter 8)
Chapter eight contains findings from the questionnaire survey. Findings from this
survey are discussed relative to each questionnaire survey objective.
Discussion of questionnaire results (chapter 9)
This chapter covers discussion, analysis and interpretation of the results obtained from
the questionnaire. Discussions on the goodness-of-fit of the postulated client-centred
model for H&S performance improvement are presented in this chapter.
Conclusion and recommendations (chapter 10)
This chapter concludes the study and contains recommendations based on the
conclusions drawn from the study. Recommendations for future research are also
made in this chapter. As part of the recommendations, a description of the finalised
client-centred H&S performance improvement model is done.
1.3.8
Ethical consideration during the study
Ethical issues were a key consideration in undertaking this study. The principle of
voluntary participation was upheld. People were therefore not coerced into
participating in the research but had to give their voluntary consent. Great effort was
made to help protect the privacy of research participants by ensuring confidentiality in
not making available identifying information to anyone who was not directly involved
in the study. Confidentiality was further enhanced by keeping participants anonymous
throughout the study.
The study was also subject to independent review by the promoters to help protect all
participants and the researcher against potential legal implications of neglecting to
address important ethical issues and also to uphold integrity, honesty and quality
assurance.
1.3.9
Resources
Financial assistance for conducting this study came mainly from the management of
the University of Johannesburg. The University paid for all tuition fees. Further
assistance came from a National Research Fund (NRF) research grant on H&S of
Professor Haupt as grant holder.
15
In addition, this study benefitted from the exposure gained from attending national
and international conferences. Other resources included easy access to library
facilities at the University of Johannesburg and its large database of accredited
academic journals. It was not going to be possible to conduct the study without these
resources.
Furthermore, the teaching load of the candidate was reduced especially in the final
year. This permitted for enough time to be devoted to the study.
1.4
CHAPTER SUMMARY
Chapter one provided a background to the current study. The motivation for the study
was that Health and safety (H&S) is still a problem in the construction industry. Both
locally in Southern Africa and internationally, the construction industry continues to
lag behind other industries in terms of H&S performance. This problem is
compounded also by the situation were contractors and designers may not be relied on
to drive H&S performance improvement. Consequently it was theorized that H&S
performance improvement was unlikely without the involvement of clients, and
therefore there was need to develop a client-centred construction H&S performance
model in order to achieve an incremental construction H&S performance
improvement.
Both qualitative and quantitative methods namely, the Delphi and structural equation
modeling method were chosen as the research methods. The importance of conducting
the study was because a gap exists in literature on the influence of client H&S culture
on project H&S performance. The study will therefore contribute to existing
knowledge by establishing the impact of client influence on H&S performance and
further develop a client-centred H&S performance improvement model. Therefore the
first step in conducting the research was to gain an understanding of the construction
industry and its H&S performance. This information will now be presented in chapter
two.
16
CHAPTER TWO
THE CONSTRUCTION INDUSTRY AND ITS HEALTH AND
SAFETY PERFORMANCE
2.1
INTRODUCTION
Health and safety (H&S) performance improvement in the construction industry has
received considerable attention in recent years. For many large construction
organisations it is a top priority (Choudhry, Fang, & Mohamed, 2007:207). This has
been in part due to the introduction and the pressure from the legislative environment
(Mitropoulos, Abdelhamid & Howell, 2005:816), coupled with increased personal
responsibility of senior managers and organisations for H&S (Fitzgerald, 2005:324), a
need to develop a good or better image of the construction industry (Misnan &
Mohammed, 2007:401) and in certain ways to address the H&S record which in
comparison to many industries is undesirable (Mohamed, 2002:375; Loughborough &
UMIST, 2003:6; Behm, 2005:590; Haslam, Hide, Gibb, Gyi, Pavitt, Atkinson & Duff,
2005:401; Kulchartchai & Hadikusumo, 2010:45). For larger multi-national
organisations, the need for H&S improvement is a corporate social responsibility
issue and therefore corporate organisations are working at improving their H&S
performance (Smallman, 2001:407).
To help synthesise recent research on the subject of H&S improvement and aid in a
broader understanding of the subject, an elucidation of the problem of H&S in the
construction industry has been made and is presented in this section. A review of
relevant literature has been conducted on H&S performance of the industry and the
impact of accidents in the construction industry. Further, a review has been conducted
on the H&S performance improvement methods that have been proposed in order to
establish the trends in the industry.
There are many proposed methods and approaches for H&S performance
improvement. Many of these approaches have been suggested to be used in
conjunction with other strategies. Some of these include:
17

designing for construction worker H&S (Hetherington, 1995; Behm, 2005;
Gambatese, Behm & Hinze, 2005; Hecker, Gambatese & Weinstein, 2005;
Weinstein, Gambatese & Hecker, 2005; Kinnersley & Roelen, 2007);

continual improvement of H&S Management Systems (Chua & Goh, 2004),

addressing H&S culture (Molenaar, Brown, Caile & Smith, 2002; Parker,
Lawrie & Hudson, 2006; Molenaar, Park & Washington, 2009);

use of incentives and disincentives (Tang, Qiang, Duffield, Young & Lu,
2008);

multi-stakeholder involvement (Suraji, Sulaiman, Mahyuddin & Mohamed,
2006; Lingard, Blismas, Cooke & Cooper, 2009), and

behaviour based H&S (Duff et al., 1994; The Keil Centre, 2003; Salem,
Lothlikar, Genaidy, & Abdelhamid, 2007).
However these studies tended to focus on specific aspects of H&S improvement and
did not necessarily approach the problem of H&S performance in a holistic way.
Culture and particularly the H&S culture of an organisation and by extension, the
industry has been identified and is seen to be the only way through which H&S
performance improvement is going to be realised because it is at the core of some of
the major accidents and incidents (Gadd & Collins, 2002:2; Wiegmann, Zhang,
Thaden, Sharma & Mitchell, 2002:3; National Commission on the BP deep-water
Horizon Oil Spill and offshore drilling, 2011:ix). There is a general agreement that
H&S culture can influence or has an impact on H&S performance (Wamuziri,
2006:168; Choudhry et al., 2007:207; Fernandez-Muniz, Peon, & Ordas, 2007:636).
Despite this agreement by various researchers on the potential of H&S culture to
improve H&S performance, there is no consensus on the definition of culture and how
culture can be measured. Of the 19 definitions reviewed in this study, 12 of them have
differing definitions (INSAG, 1991; Hudson, 1999; Gadd & Collins, 2002; Molenaar
et al., 2002; Wiegmann et al., 2002; IOSH, 2004; Fitzgerald, 2005; Dingsdag et al.,
2006; Fernandez-Muniz et al., 2007; Misnan et al., 2007, Molenaar, et al., 2009). As a
result it seems as though it is this lack of consensus that has caused in part a lack of
full utilisation of the concept to improving H&S performance in the construction
industry.
18
Therefore in addition to reviewing literature regarding H&S performance in the
construction industry, a review has also been conducted of the concept of H&S
culture and its influence on H&S performance.
The rest of this section presents a discussion of literature regarding the construction
industry, H&S in the construction industry, H&S performance in the industry, impact
of accidents on the industry and the concept of H&S performance improvement.
Furthermore a review and synthesis of literature on H&S performance improvement
and the concept of H&S culture is presented.
2.2
THE CONSTRUCTION INDUSTRY
2.2.1
Complexity of the construction industry
Performance improvement in the construction industry is made difficult by the nature
of the industry. The construction industry is a complex industry. Teo, Ling & Chong
(2005:329) observed that construction sites were generally complex and unsafe
because of the extensive use of sophisticated plants, equipment, modern methods of
construction, multidisciplinary and multitasked aspects of its project work force. It is
this complex nature that shapes the industry’s way of functioning and its performance
(Sawacha, Naoum & Fong, 1999:309; Dubois & Gadde, 2001:2) which according to
many leaves much to be desired. Tzortzopoulos, Sexton & Cooper (2005:471) argue
that the complexity of design and construction is the primary reason for the difficulty
in sustaining significant improvements in the industry. The industry’s complexity is
described as follows, namely
“The physical substance of a house is a pile of materials assembled from
widely scattered sources. They undergo different kinds of and degrees of
processing in large number of places, require many types of handling over
periods that vary greatly in length, and uses the services of a multitude of
people organised into many different sorts of business entity (Dubois &
Gadde, 2001:2)”
Constructing a single building requires procurement of many types of materials from
many places and suppliers, involves many different types of organisations and is
conducted in different kind of environments. The construction industry is complex
and unique and is different in many ways from other manufacturing industries. The
19
construction industry has unique characteristics that compound its complexity and
pose a challenge to performance improvement such as the following, namely:

Temporary employment (Pellicer & Molenaar, 2009:44);

work location for any group of workers often changing (Riley & Clare-Brown,
2001:150; McDonald, Lipscomb, Bondy & Glazner, 2009:53);

Temporary work sites where workers are employed by different employers but
working alongside each other ( Chan & Chan, 2004:203, Pellicer & Molenaar,
2009:44; Misnan et al. 2008:1902);

An industry comprised mostly of small employers (Pellicer & Molenaar,
2009:44). For example, in the United Kingdom, 98% of the registered
companies, employed 24 or less workers in their companies (Dainty, Briscoe
& Millet, 2001:163);

A culturally diverse workforce with some workers coming from foreign
countries and other workers working as self-employed independent contractors
(Bomel, 2001:2.4);

Large numbers of people are employed in this industry and have to combine a
diverse range of skills to complete a project (Bomel, 2001:2.4; Dainty, et al.,
2001:163; Riley & Clare-Brown, 2001:150; Pellicer & Molenaar, 2009:44);

A large number of subcontractors (Bomel, 2001:2.4; Riley & Clare-Brown,
2001:150; Pellicer & Molenaar, 2009:44);

Construction projects with short periods (Bomel, 2001:2.4; Dainty et al.,
2001:163; Riley & Clare-Brown, 2001:150);

Sites evolving as construction proceeds thereby changing the hazards that
workers face week by week (Bomel, 2001:2.3; Riley & Clare-Brown,
2001:150);

A fragmented industry (Egan, 1998:8; Chan et al., 2004:203). Dainty et
al.,(2001:163) argue that the proliferation in subcontracting has further
complicated the situation through the increased fragmentation of the
production process;

An industry subjected to cyclical economic downturns (Egan, 1998:9; Dainty,
et al., 2001:163), and
20

An industry with a low and unreliable rate of profitability (Egan, 1998:7;
Pellicer & Molenaar, 2009:44).
In addition, the construction industry is largely operationalised through a professional
system. In a professional system, the requirement is to have designs fully specified at
tender stage. However assumptions regarding the competence of designers in the
technical details of a wide range of construction technologies and the ability of the
client to keep requirements fixed over a period of time, compromise the effectiveness
of the professional system (Winch, 2000:142). The reality is that designs are rarely
fully specified. The professional system arguably has led to developing complex
contracts that enabled changes to be negotiated. The professional system became over
time “a risk-shedding rather than a risk sharing system” (Winch, 2000:145). This
professional system as a result introduced further complexity in the industry.
Further complexity arises from decisions made on costs as well. Construction which
operates according to short term contracts in a very competitive environment tends to
drive decisions toward the lowest cost bidder and the immediate cost is usually
favoured over long term quality or H&S for that matter. This has resulted in the
stagnation of the industry.
The industry is also by its nature involving change. Various activities in the
construction industry are subject to a constantly changing working environment
(Bomel, 2001:0.5; Riley & Clare-Brown, 2001:150; Suraji, Sulaiman, Mahyuddin &
Mohamed 2006:52). This characteristic creates an unusual working environment
when compared to other work places such as factories, offices, process plants and
even farms which in comparison to the construction industry have experienced
tremendous improvement in all areas.
2.2.2
Contribution of clients to industry performance
The types of clients that the industry serves also do not help the state of the industry.
These include domestic and small commercial clients who generally have no
knowledge or understanding of construction or the construction process itself.
However the industry also has major clients whose reputation and repeated need for
construction ensure that they understand the construction process. The client has been
key to the improvement or lack of it of the construction industry (Bomel, 2001:5.3).
21
For instance, the developer as a client presents another situation where the aim is to
convert a piece of land into a high value structure which can be sold for profit. As a
result, there is no regard for the in-between process including H&S consideration.
Notwithstanding the private developer, Egan (1998:7) contends that too many clients
in the industry are non-discriminatory and still equate price with cost, selecting
designers and constructors almost exclusively on the basis of the tendered price. This
tendency is seen as one of the greatest barriers to improvement of the industry. Clients
almost exclusively focus on cost, time and quality. As a result, many aspects of
production are compromised and in a highly competitive market, the successful
tenderer will frequently be the one that pays the lowest wages, does not provide safety
equipment or have insurance coverage for accidents, and has the largest proportion of
informal workers (Cotton, Sohail, & Scott, 2005:21; Wells & Hawkins, nd:3), for
whom no tax or social security is paid, and who are not covered in practice by any
legal or social protection. This lowest-price culture in competitive bidding is
incompatible with H&S and definitely with the quality of products (Murie, 2007:7).
2.2.3
Contribution of lack of integration to performance
The construction industry suffers from the problem of lack of cooperation and
integration. Egan (1998:13) observed that the construction industry was basically an
industry typically dealing with the project process as a series of sequential and largely
separate operations undertaken by individual designers, contractors and suppliers who
have no stake in the long term success of the product and therefore have no
commitment to it. Therefore, changing this culture is fundamental if performance
improvement is to be realised. Egan (1998:19) rightly argue that the rationale behind
development of an integrated process is that the efficiency of project delivery is
presently constrained by the largely separated processes through which they are
generally planned, designed and constructed. These processes reflect the fragmented
structure of the industry and sustain a contractual and confrontational culture (Egan
1998:19). Riley & Clare-Brown, (2001:150) also contend that the industry is typified
by conflict and an adversarial attitude between all parties in the project.
Furthermore, the construction industry has always been bifurcated between an
established high performance and high cost sector and a less skilled, lower cost sector.
22
As a result, organisations have different outlooks and aspirations and therefore
compromising performance.
In the construction industry, there is little focus on other important issues such as
cooperation, integration and partnering that have resulted in improvements in other
industries. The saying ‘time is money’, is taken quite literally in the construction
industry. The driving force in construction projects is the schedule as capital
investment cannot be realised until construction is complete (Bomel, 2001:0.5).
2.2.4
Construction workers’ challenges
The construction industry is a physically demanding workplace (Pungvongsanuraks &
Chinda, 2010:178). As a result, it requires resilience and strength and consequently
has developed as a male dominated regime with the associated macho culture (Bomel,
2001:0.5; Mullen, 2004:283). In addition, the construction industry worldwide,
principally offers low status, low-paid, short-term, unregistered, informal, and
hazardous jobs in a highly fragmented industry (Murie, 2007:5; Kulchartchai &
Hadikusumo, 2010:47). Many workers, in particular emigrants, are faced with
exploitative employment practices, hardship, and hazards (Fitzgerald & Howarth,
2009:2). The construction industry is also responsible for far more than its share of
occupational accidents and work-related ill health (Hola, 2007:255; Kulchartchai &
Hadikusumo, 2010:45). Construction is a hazardous occupation (Ng et al., 2005:2;
Chen, Lu & Huang, 2011:398). For almost all key risks, chemicals, dusts, manual
handling, physical hazards, and psychosocial hazards, exposures are routine and
excessive. Poor welfare facilities compound the hazards and the notion that only the
tough survive seem to be a predominant culture in the industry (Murie; 2007:5).
The construction industry is not attractive and because the work in the industry has a
low social status, its problems also have low visibility and their resolution a low
priority. For example, workers’ health conditions are often not recognised, but if they
are recognised, they are not diagnosed and if finally they are diagnosed, they are not
attributed to work and are rarely compensated or effectively treated (Murie, 2007:6).
The conditions in the industry further encourage a culture for rapid hire and fire in
addition to the low status jobs that they offer. This rapid hire and fire scenario creates
23
uncertainty for workers and a pressure to perform. The pressure to perform
inadvertently results in poor H&S performance (Bomel, 2001:05).
Protection and representation from unions is not guaranteed in most parts of the world
seeing that construction industry union representation varies and many sites are nonunion (Walters, 2009:3; Bomel 2001:2.5). At professional and industry level, the
industry is dominated by many bodies that claim to represent interests of various
players in the industry and sometimes demarcations between various bodies are
unclear, and their roles are confusing. The industry has many issues and these areas
need to be considered if improvement in the industry has to be realised (Bomel,
2001:2.5).
Clearly, the construction industry is complex, unique and dogged with many
difficulties. However this unique nature of the industry should not hinder performance
improvement because the industry is critical to national economies. Egan (1998:6)
with reference to the UK’s construction industry rightly points out that this industry is
one of the pillars of the economy. The construction industry worldwide is a USD
Three trillion giant, accounting for around ten Percent of the world’s gross domestic
product and employs 180 million people, or seven Percent of the global employment.
Apart from economic contribution, the industry improves the quality of life for
product end users. The state of the art residential and commercial facilities,
recreational and transportation facilities all testify to the contribution that the
construction industry has made to civilisation (Egan, 1998:6). One area where
performance improvement is needed is on H&S.
2.3.
H&S IN CONSTRUCTION INDUSTRY
H&S is a combination of two terms, namely, health and safety. Health refers to
occupational health and is defined as the wellness of workers relative to their
occupation or place of work (ILO, 1983:1491). Safety, on the other hand, is the
prevention of risk to the health of workers, and visitors, to a work place and to
property (ILO, 1983:1491). Workers’ H&S is constantly at risk either due to
conditions within and outside of their work place or their inability to recognise and
handle the risks. The construction industry is not an exception when it comes to this
problem.
24
It has, in fact been reported that construction workers are more at risk of an injury or
even a fatality at their workplace than other manufacturing based industries (Suraji,
Duff & Peckitt, 2001:337). Many researchers such as Teo et al. (2005:329) and Kines,
Spangenberg & Dyreborg (2007:53) argue that generally construction sites are still
one of the most dangerous workplaces because of the high incidence of accidents. In
the construction industry, the risk of a fatality is said to be at least five times more
likely than in other manufacturing based industries (Sawacha, Naum, & Fong,
1999:309; Loughborough University & UMIST, 2003:vii). The construction industry
is considered to be hazardous and is among the most hazardous industries
(Abdelhamid & Everett, 2000:52) considering its work related mortality rate, injury
rates, and worker’s compensation payments (McDonald, Lipscomb, Bondy &
Glazner, 2009:53). In some studies such as one by Bomel (2001) the construction
industry is described as being a hazardous environment where workers have direct
exposure to height, forces, and power. According to the Irish Government’s
assessment, statistics maintained by the H&S Authority, illustrate that the
construction industry has a consistently poor record on worker H&S relative to other
industries (DETE, 2006:1). The construction industry, despite some gains in recent
years in terms of accidents, continues to lag behind many industries in H&S
performance (Gambatese, Hinze & Haas, 1997:32).
Research findings trace the hazardous conditions and unsafe behaviour in the
construction industry to the following sources:

Organisational factors and arrangements (Sawacha, Naoum & Fong,
1999:309; Suraji, Duff & Peckitt, 2001:337; Mullen, 2004:278; Manu,
Ankrah, Proverbs & Suresh, 2010:4);

Competitive tendering that is characteristic of the industry (Kartam, Flood &
Koushki, 2000:163; Svedung et al., 2002:398);

Inadequate consideration of H&S during design of structures (Mohamed,
2002:375; Loughborough & UMIST, 2003:vii; Manu, et al., 2010:4);

Short term and transitory nature of the industry (Sawacha et al., 1999:309);

The H&S culture of the industry (Bomel, 2001:5.3; IOSH, 2004:6);
25

Inadequate or lack of training and education (Sawacha, et al., 1999:309;
Bomel, 2001:8.3; Loughborough & UMIST, 2003:viii; Spangenberg, Baarts,
Dyreborg, Jensen, Kines & Mikkelsen, 2003:525; Gholipour, 2004:63);

Lack of incentives (Smallman, 2001:398; Tang et al., 2008:465);

Government policies and regulations (Kartam et al., 2000:173);

Extensive use of subcontractors (Lingard & Rowlinson, 1997:243; Kartam et
al., 2000:173; Manu, et al., 2010:4);

Confrontational cultures (Gholipour, 2004:63);

Small size of most construction firms (Kartam, et al. 2000:173; Ringen &
Englund, 2006:388);

High labour turnover (Kartam, et al., 2000:174);

Work pressure (Mitropoulos et al., 2005:817; Mullen, 2004:278);

The macho culture characteristic of the industry (Bomel 2001:0.5; Mullen,
2004:278);

Unsafe proximal conditions (Abdelhamid & Everett, 2000:59; Suraji et al.,
2001:337); and

Unsafe behaviour of workers (IOSH, 2004:6)
In addition to these reasons, Bomel (2001), Abdelhamid & Everett, (2000) and Suraji
et al.( 2001) identified the proximal fundamental sources of H&S risk as being factors
that will directly cause harm and that may include weights of materials, equipment,
height, electricity, pressure, collision, energy, human factors, heat, noise, and
numerous hazardous substances. However, beyond these proximal factors, there are
other distal factors such as commitment and culture of clients which have an influence
on H&S performance (Bomel, 2001).
There has also been awareness that organisational, managerial and human factors
rather than purely technical failures were prime causes of accidents (Flin, Mearns,
O’Connor & Bryden, 2000:178). In fact, most theories not only attribute the cause of
accidents to factors that relate to human actions but also to remote factors generated at
project inception (Suraji et al., 2001:337). To this effect Behm (2005:590) argue that
the H&S of any operation is determined long before the people, procedures and
equipment come together at the work site. In the construction industry this scenario
relates to factors that need attention at the inception stage.
26
Suraji et al., (2001:337) argued that management and organisational aspects were the
causes of accidents. They contended that from the introduction of the domino theory
by Heinrich (1969), accidents have not only been viewed as a consequence of
operative unsafe actions and unsafe site conditions, but also that the lack of
management control and organisational failures were often precursors of accidents
(Mansingh & Haupt, 2008:466).
Further causes of accidents have been identified as management system failure and
human error (Petersen, 2000:38; Chua & Goh, 2004:542). This theory highlights the
fact that trying to find the unsafe act or condition deals only at the symptomatic level,
because the act or condition may be the proximate cause but invariably not the root
cause (Behm, 2005:590). This theory holds that the root cause of accidents is in the
organisational and management processes (Abdelhamid & Everett, 2000:59; Behm,
2005:590; Bellamy & Geyer, 2007:15). Therefore, effort aimed at addressing H&S,
should be directed more at addressing organisational and project management factors.
These factors should include management at industry, project and company or
organisation level.
In brief, accidents are prevalent in the construction industry. Various causes of
accidents and ill health have been identified. Workers in construction faced a real risk
to their H&S compared to their counterparts in other industries. The source of hazards
and unsafe behaviour had its origins not only from proximal conditions but also from
distal conditions including organisational and management conditions.
2.4.
H&S PERFORMANCE OF THE CONSTRUCTION INDUSTRY
2.4.1
H&S Performance at International level
The construction industry employs about 180 million people, or seven Percent of
global employment (ILO, 2005:2; Murie, 2007:5), which is nearly the same as the
percentage quoted for the construction industry in the European Union (EU). In 2001
and 2002 the EU was said to have employed about eight Percent of the total
workforce (Dias, 2004:1; Karjalainen, 2004:3; Ringen & Englund 2006:388). Despite
its size in terms of the workforce, the construction industry in comparison with other
economic sectors of the EU causes about 20 to 25% of all fatal accidents (Dias,
2004:1; Karjalainen, 2004:3; Ringen & Englund, 2006:388). Clearly, in comparison
27
with other industries, the construction sector does not have an enviable record or
reputation and its H&S performance can best be described as poor (Haslam et al.,
2005:401). In fact during 2001, preliminary data released for the European Union,
showed that there were about 822,000 accidents at work resulting in more than three
days absence from work and about 1,200 fatal accidents (Karjalainen, 2004:3). The
European Agency for safety and health at work reported that many more accidents
occurred in construction per 100,000 workers than the overall EU national workforce
(Karjalainen, 2004:3). The incidence rate of non-fatal accidents at work was nearly
two-fold in construction compared to other sectors and even greater for fatal
accidents. Table 1 show that there were 10.4 fatalities per 100,000 workers in
construction and 4.2 per 100,000 overall for other industries (Karjalainen, 2004:3).
Table 1: Incidence rate of accidents at work in EU-15
Year
Incidence rate (per 100000 workers)
Non-fatal
Fatal
Construction 9 NACE
Construction 9 NACE
1994
9014
4539
14.7
6.1
1995
9080
4266
14.8
5.9
1996
8023
4229
13.3
5.3
1997
7963
4106
13.1
5.2
1998
8008
4089
12.8
5.0
1999
7809
4088
11.7
4.8
2000
7548
4016
11.4
4.6
2001
7213
3830
10.4
4.2
9NACE= agriculture, manufacturing, electricity, gas and water supply, wholesale and
retail sale, repairs, hotels and restaurants, transport and communications, financial,
real estate and business activities.
(Source: Karjalainen, 2004:3)
The most recent statistics for the United Kingdom (UK) show that there were 152
workers fatally injured in 2009/10 equivalent to a rate of 0.5 fatalities per 100,000
workers (HSE, 2010b:online). The construction sector in the UK had the highest rate
of 42 fatalities (National Statistics UK, 2010:8). In terms of self-reported work-related
illness and reportable non-fatal injury, the average rate per 100,000, showed that the
construction industry was the third worst performer after transport, communication
and agriculture, considering the average for the period 2007/08-2009/10 (National
Statistics UK, 2010:16). However reference should be made to the number of people
employed in the UK construction industry.
28
In 2009/10 the UK construction industry accounted for only 4% of the total workforce
but had 27% of fatalities and 7% of the reported injuries to employees (HSE, 2010b:
online). The construction industry had the highest number of fatal injuries among the
main industry groups accounting for 42 fatal injuries giving a rate of 2.2 per 100,000
workers (HSE, 2010b: online). The construction industry also accounted for 35% of
all reported injuries involving high falls and 24 % involving electricity. Although
there has been a marked improvement over the years as Figure 2 shows, the
construction industry has consistently lagged behind other industries. Figure 1
illustrates a comparison of the rate of fatal injury for all industries of 0.5 in 2009/10 to
that of 2.2 for the construction industry in Figure 2 for the year 2009/10.
350
Number of fatal injuries
300
250
200
150
100
50
0
96/97 97/98 98/99 99/00 00/01 01/02 02/03 03/04 04/05 05/06 06/07 07/08 08/09 09/10
Years
Figure 1: Fatal injuries in all sectors in UK
(Source: HSE, 2010:8)
29
Figure 2: Fatal injuries in UK’s construction industry
(Source: HSE, 2010b: online)
The H&S performance of the construction industry in the United States of America
(USA) is not much different from the UK and the rest of the EU. The performance is
poor compared to other industries. Statistics released for the year 2010 show that most
fatal falls occur in the construction industry (Figure 3). Workers in the construction
industry incurred the most fatalities of any industry in the private sector in 2009. Of
the total 4,340 fatal injuries, the construction sector accounted for 816 fatalities.
Although in 2009 the accident record “improved” compared to the year 2008, this
improvement has however been ascribed to economic factors. Among those factors is
the total number of hours worked in 2009 which fell by 6%, with percentages in
construction estimated to be even higher (BLS, 2010:1).
30
Utilities
Information
Financial activities
Mining
Educational and health services
Other services (exc. public admin.)
Wholesale trade
Leisure & hospitality
Retail trade
Manufacturing
Professional and business services
Government
Agriculture, forestry,fishing, and…
Transportation and warehousing
Construction
0
Total fatal injuries = 4,340
All worker fatal injury rate = 3.3
200
400
600
800
1000
Figure 3: Fatal accidents in USA
(Source: BLS, 2010:3)
The construction industry in Singapore appears to have a similar record. Although the
construction sector in 2005 contributed less than 10% to the gross domestic product it
accounted for more than 37% of all industrial accidents (Teo et al., 2005:329). H&S
performance, even at the global level is not impressive. As a matter of fact, the
International Labour Organisation (ILO) identified the construction industry as one of
the high risk sectors and estimated that at least 60,000 people were killed on building
sites every year and ranked the construction industry third highest risk sector (ILO,
2003:11). Clearly there was an urgent need to improve construction H&S
performance.
2.5
H&S IN DEVELOPING COUNTRIES & SOUTH AFRICA
H&S performance in South Africa, Botswana and other developing countries is poor.
Occupational H&S remains neglected in developing countries (Nuwayhid,
2004:1916). In comparison to developed countries, performance in H&S is poorer
(ILO, 2003:11; Anvuur, Male & Kumaraswamy, 2006:2). According to Kheni, Gibb
& Dainty, (2007:160), it is partly because of lack of resources in developing countries
that has constrained H&S performance improvement. In addition, the inadequate legal
and institutional arrangements for the management of H&S have compounded the
31
Utilities
Information
Financial activities
Mining
Educational and health services
Other services (exc. public admin.)
Wholesale trade
Leisure & hospitality
Retail trade
Manufacturing
Professional and business services
Government
Agriculture, forestry,fishing, and…
Transportation and warehousing
Construction
0
Total fatal injuries = 4,340
All worker fatal injury rate = 3.3
200
400
600
800
1000
Figure 3: Fatal accidents in USA
(Source: BLS, 2010:3)
The construction industry in Singapore appears to have a similar record. Although the
construction sector in 2005 contributed less than 10% to the gross domestic product it
accounted for more than 37% of all industrial accidents (Teo et al., 2005:329). H&S
performance, even at the global level is not impressive. As a matter of fact, the
International Labour Organisation (ILO) identified the construction industry as one of
the high risk sectors and estimated that at least 60,000 people were killed on building
sites every year and ranked the construction industry third highest risk sector (ILO,
2003:11). Clearly there was an urgent need to improve construction H&S
performance.
2.5
H&S IN DEVELOPING COUNTRIES & SOUTH AFRICA
H&S performance in South Africa, Botswana and other developing countries is poor.
Occupational H&S remains neglected in developing countries (Nuwayhid,
2004:1916). In comparison to developed countries, performance in H&S is poorer
(ILO, 2003:11; Anvuur, Male & Kumaraswamy, 2006:2). According to Kheni, Gibb
& Dainty, (2007:160), it is partly because of lack of resources in developing countries
that has constrained H&S performance improvement. In addition, the inadequate legal
and institutional arrangements for the management of H&S have compounded the
31
problem of H&S performance (Kheni, Gibb & Dainty, 2007:160). Other contributors
to poor H&S performance include a lack of effective mechanisms to implement laws
(Cotton, Sohail, & Scott, 2005:21), lack of infrastructure, equipment and rampant
corruption and a lack of concerted effort by policy makers to address H&S (Kheni,
Dainty & Gibb, 2007:159).
H&S performance in the construction industry of the developing countries is even
worse. The industry in developing countries is dominated by small & medium
contractors and most of these do not have effective systems to manage H&S (Kheni,
et al., 2007:159). In addition, the industry is also labour intensive and utilises old
technology which has an impact on H&S performance. The construction industry is
also characterised by a poor knowledge and awareness of H&S (Musonda &
Smallwood, 2008:85, Kheni, et al., 2007:160). The result has been a poor
performance as revealed in the number of both fatal and non-fatal injuries.
A research conducted in Botswana in 2006, revealed that most accidents occurred in
the construction industry (Van Ooteghem, 2006:43). Similarly the South African
Construction Industry Development Board (CIDB) report of 2008 showed that the
construction sector had the third highest number of fatalities per 100,000 workers
after fishing and transport (CIDB, 2008:9). The construction industry has also seen an
increase in the number of accidents that have been reported from 2004 to 2008 (Table
2). In comparison to other manufacturing based industries, the construction industry
rank first as can be seen in Table 3.
Table 2: Accidents in South Africa's construction industry
Year
Accidents
Non-Fatal
Fatal
Total
2004/05
170
54
224
2005/06
257
81
338
2006/07
255
79
334
2007/08
416
162
578
(Source: CIDB, 2008:5)
32
Table 3: Fatality rate in South Africa
Industry
Fatality rate
(Per 100000 of workers)
Construction Industry
25.5
Mining
23.5
Agriculture and forestry
12.5
All industries
11.4
(Source: CIDB, 2008:5)
According to the CIDB there was gross underreporting of accidents and therefore the
actual statistics could be even much higher (CIDB, 2008:2). The construction industry
in South Africa and Botswana has lagged behind other industries and the performance
record for the South African industry has not been improving when compared to its
past record. The construction industry in South Africa is known to be one of the most
hazardous with an unacceptably high level of injuries and fatalities (Geminiani,
Smallwood & Van Wyk, 2008).
2.6
CHAPTER SUMMARY
Although statistics are not reliable to inform on H&S performance, H&S performance
in the construction industry is almost the same in most parts of the world. The
construction sector in comparison to other industries has not improved much. It has
consistently lagged behind other industries and its position is among the worst H&S
performers. The level of H&S performance in the construction industry established
from even the unreliable statistics of accidents is therefore unacceptable, specifically
with the South African industry which has seen an increase in accidents in recent
years.
The status quo cannot be allowed to continue both for moral and economic reasons.
The ILO (2003:15) has estimated that four Percent of gross domestic product is lost
due to accidents and work-related injuries. This estimate for developing countries
could even be much higher than 10% of the gross domestic product. The impact of
accidents is severe (Hermanus, 2007:531). An occupational injury to one worker can
seriously reduce the income of a household. In the USA, for example, workers who
receive a partial disability due to work place injury lose about 40% of their income
33
over five years. Workplace injuries and illness typically reduce the overall earnings of
households (ILO, 2003:17).
It is clear that in comparison to other industries, the construction industry has (a) a
higher accident rate; (b) one of the highest fatality rates in comparison to other
sectors; (c) a poor H&S record and therefore a higher risk on H&S compared to other
sectors. The construction industry’s H&S performance is unenviable. Further, H&S
performance in developing countries is poor and the situation is even dire in the
construction industry of the developing countries. Therefore strategies should be
devised to improve H&S and should be considerate of political, economic and socialcultural factors (Kheni, et al., 2007:161). As Nuwayhid (2004:1916) argue, the
traditional workplace focused H&S performance improvement, has been insufficient
in developing countries and therefore real progress in H&S can only be achieved by
linking H&S to the broader context of social and national development.
34
CHAPTER THREE
H&S PERFORMANCE IMPROVEMENT AND CULTURE
3.1
INTRODUCTION
The Oxford dictionary defines ‘improve’ as ‘make or become better’. Improvement is
defined as the action of improving or the state of being improved (Oxford, 2001:451).
Therefore it may be that one intends to improve on a previous record or indeed
produce something better than in the past. It may also entail adding something, a
component, to enhance performance. Relative to H&S improvement, it has to do with
a better than previous H&S record expressed in terms of accidents and incidents. It
may be expressed in terms of statistics on injuries and fatalities, a better performance
in comparison to other industries or indeed in the way things are done relative to H&S
namely, the H&S culture. For H&S culture, it means that improvement should also be
reflected in the perception of H&S performance by stakeholders in an organisation or
indeed at the industry level. Therefore, improvement is said to be attained or
experienced if the new record or way of doing things has become better when
compared to a previous state, record or indeed a way of doing things. This
improvement would also be reflected in the positive perception of H&S performance
by the concerned stakeholders.
3.1.1
Performance improvement
From the definition, it appears that improvement could be dramatic, or a gradual
gravitation towards a set goal or vision which does not necessarily mean attaining the
goal, whether clearly defined or not. In this way, improvement is described in terms
of the extent to which the initial state has been exceeded or what progress has been
made cf National Statistics UK (2010). An example of this measure of improvement
could be the improvement in the perception of employees concerning the state of
H&S. From the perception surveys conducted among employees in year 2000 and
2004 by Petersen (2005), there was an improvement of 9% in the H&S system
between the stated periods. The reason for his conclusion was that there had been a
shift in workers’ perceptions about the H&S status (Petersen, 2005).
35
Improvement could also be slight, moderate, or significant depending on the scale that
has been used. However, the level or extent of improvement depends on the defined
or set goal. For H&S, the goal could be zero accidents or simply a reduction in the
number of accidents from an existing level to a new one. For example, Egan
(1998:16) proposed improvement to H&S in UK’s construction industry of 20%
reduction in the number of reportable accidents per year. Another example is the
initiative by the H&S commission in the UK cited by Smallman (2001:409), being a
target to reduce the incident rate of fatal and major injury accidents by 10% by 2010.
In 2000 the incident rate of fatal injuries in the UK was 5.9 per 100,000. The goal was
therefore to achieve an incident rate of fatal accidents of 5.4 per 100000 by year 2010.
Statistics released for the 2010 show a fatality rate of 2.2 in the construction industry
(HSE, 2010b: online). Compared to the goal of 5.4, a fatality rate of 2.2 could
therefore be described as a significant improvement. However the rate of 2.2 is almost
4½ times higher than the national average of 0.5. Nonetheless, compared to the goal
that had been set, there has been a significant progress. Therefore, improvement
relates to the targets that were initially set.
Furthermore, the speed with which an industry, entity or individual moves towards
achieving the desired goal may not be a fundamental issue in determining
improvement, but what may matter is that there is a movement towards the desired
goal. In the previous example, the fatality rate was 2.2 per 100,000 workers. This
fatality rate was calculated from 42 people who died in 2010 in the UK as a result of
workplace accidents, yet it can be said that there has been progress or rather an
improvement because in 2000, 100 people actually died (HSE, 2010b: online). The
desirable state is that no one should die as a result of his or her work. However
insensitive the measure might be, it is fairly acceptable to measure and determine
progress or improvement in this manner (National Statistics UK, 2010 and BLS,
2010).
On the other hand, improvement also has to do with benchmarking. Benchmarking
compares for example an organisation’s performance against the ‘best in class’,
wherever that is found (Johnson & Scholes, 2002:174). Benchmarking provides the
context for assessing performance. It is an approach to process improvement and is
defined as a systematic process of measuring and comparing an organisation’s
36
performance against that of other similar organisations in key activities (Rankin et al.,
2008:896). The reason benchmarking is popular in for example business circles, is
because it can change the mind-set of managers so that improvements in performance
will be gradual as a result of incremental changes. Benchmarking breaks the frame
within organisations and industries regarding performance standards to be achieved
(Johnson & Scholes, 2002:174). If benchmarking had been used to evaluate progress
made on H&S performance in the UK construction industry the verdict could
probably have been ‘no or minimal improvement’ considering the fatal rate of 2.2 in
construction against the rate of 0.5 in all industries. Therefore, statistics on accidents
alone may not be useful to determine and plan for improvement. Jafri et al.,
(2005:707) contend that incident and injury rates are important, but they are not
always useful for occupational H&S performance improvement.
3.1.2
Measuring H&S performance improvement
In order to measure H&S performance improvement it is not adequate to use only
statistics of accidents. Suitable and appropriate measures are therefore important in
addressing the issue of H&S performance improvement. Cameron & Duff (2007:870),
contend that current measures of accidents and incidents are today recognised as
invalid measures of H&S at the workplace except in very large organisations or over a
long period because they are insensitive to real changes in H&S behaviour or
conditions. Besides, these statistics are often under reported. As a result, an alternative
measurement that is based on evaluating management actions and worker perceptions
has been proposed. Evaluating management actions and worker perceptions can
differentiate health and safe sites based on the observed level of H&S management
commitment and perceptions of workers (Petersen, 2005:48; Cameron & Duff,
2007:870).
Worker perception surveys were used by Petersen (2005) to measure improvement.
He argues that perception is one of the better measures available to truly define
reality. This is in agreement with many researchers who argue that using statistics
alone is not sufficient to measure performance improvement (Carder & Ragan,
2003:158; Jafri et al., 2005:707). They argue that incident rates which involve
statistics are important, but they are not always useful for process improvement. They
contend that without proper investigation of causes, incident rates indicate that there
37
is a problem but they do not inform what the problem is. There are better methods to
reflect the reality than injury statistics alone (Petersen, 2005:48). A method such as
climate survey is used to measure H&S performance. Consequently, it is possible to
implement a safety cultural audit through a climate survey as a valid and reliable
measure of H&S performance in the construction industry (IOSH, 2004:7).
Determining cultural improvement has become important because it has been found to
be at the heart of most accidents.
3.2
TECHNIQUES TO IMPROVE H&S PERFORMANCE
3.2.1
Influence network technique
Several techniques have been proposed to improve H&S performance in the
construction industry such as the Influence Network Technique (Bomel, 2001). The
use of the Influence Network Technique is operationalised by first assessing factors
systematically at the environmental, policy, organisational, and direct levels that may
have influenced an outcome (top event) such as accidents or poor performance on
H&S. In order to improve performance, the identified factors with inadequacies are
then targeted to ensure that they meet the required standards.
A top outcome event, such as, H&S performance level or a specific event such as fatal
falls from height is first identified. Secondly, influences at the direct causal level of
the top event relative to the predefined categories of hardware, human and external
events are defined (Fig 4). This process is repeated for the subsequent levels of
organisation, policy and environment. The weighting and rating of current status or
practice of each influence are then made by a panel of experts either through a focus
group setting or a Delphi procedure. The weighting is based on an influence scale
totalling unity for influences at one level of influences. On the other hand, the rating
value is based on a scale of 1 to 10 to describe worst to best performance status. A
calculated risk level is obtained from the sum of all products of rating and weighting
agreed upon by the panel. Critical influences are then identified for the top event and
the necessary risk controls instituted. After implementation of the risk controls, the
procedure
is
repeated
for
influences
38
at
various
levels
(Bomel,
2001)
TOP EVENT
DIRECT LEVEL
COMPETENCE
MOTIV
/MORALE
TEAM
WORK
SITUATION
AWARENES/
RISK
PERCEPTION
FATIGUE
/ALERTNESS
HEALTH
COMMS
AVAILABILITY
OF
INFO/ADVICE
COMPLIANCE
AVAILABILITY
OF SUITABLE
RESOURCES
INSPECT
& MAINT
EQUIPMENT
OPERABILITY
WORKPLACE
ENVIRONMENT
EXT
CONDITION
ORGANISATIONAL LEVEL
RECRUITMENT
& SELECTION
TRAINING
PROCEDURES
PLANNING
INCIDENT
MANAGEMENT
& FEEDBACK
COMMS
MANAGEMENT
/ SUPERVISION
SAFETY
CULTURE
EQUIPMENT
PURCHASING
INSPECT
& MAINT
DECISION
PAY /
CONDITION
POLICY LEVEL
CONTRACTING
STRATEGY
OWNERSHIP &
CONTROL
COMPANY CULTURE
SAFETY MANAGEMENT
ORGANISATIONAL
STRUCTURE
LABOUR RELATIONS
ENVIRONMENTAL LEVEL
POLITICAL INFLUENCE
REGULATORY INFLUENCE
MARKET INFLUENCE
Figure 4 : Influence Network Technique
(Bomel, 2001: 0.9)
39
SOCIETAL INFLUENCE
COMPANY PROFITABILITY
Analysis of the Influence Network Technique suggests that the technique is very
useful as an analysis tool of causes of either accidents or poor performance relating to
one aspect of H&S. The Health and Safety Executive’s incident investigation reports
for example, provide some evidence on say an accident and in addition to this
knowledge, the experience that those working in the industry provide further
complementary insight on the accident (Bomel, 2001:6.2).
However, according to Bomel, evaluation of this data only provides evidence of what
has occurred but it is far more important to be able to implement appropriate courses
of action to prevent the accident happening again. Therefore in order to determine
what these actions might be, they propose that it is important to be able to structure
the disparate sources of evidence to provide a comprehensive ‘model’ of accidents
and therefore the necessity of the Influence Network Technique. Influence network
technique makes use of past events such as accidents and or evidence of poor
performance to formulate improvement strategies which after implementation could
be analysed again. The technique appears to be useful for evaluating the possible
effect or level of improvement that may be achieved on the top risk event if
improvement is made on either direct, organisational, policy or environmental level
influences.
Nonetheless, applicability of the technique in the construction industry to improve
H&S Performance may be limited because it is not clear as to where this tool may be
used; whether it is at the organisation industry or at national level. If at all the tool has
to be used at the organisation level, then its reliance on experts and complexity makes
it impractical to use because the industry is dominated by small contracting firms
(Kartam et al., 2000:173; Ringen & Englund, 2006:388) who may have limited
capacity to fully utilise its benefits. On the other hand, if it is meant to be used at the
policy government, industry or client level, then it may not be very useful as
involvement in the evaluation of practice is best done by key stakeholders that are
affected such as workers (Petersen, 2005:45). It is also worth noting that the Influence
Network Technique relies on past records, namely, accidents and or H&S
performance. However, recent trends have been a movement away from methods that
rely only on past records, such as accidents towards more proactive methods that rely
40
on leading indicators (Carder & Ragan, 2003:158; Jafri et al., 2005:703; Petersen,
2005:45; Cameron & Duff, 2007:870).
3.2.2
Climate / Perception surveys
Other techniques to improve H&S performance have been identified. One such
technique is the use of perception or climate surveys. Petersen (2005:45) proposed
that perception surveys can be used as an indicator of a H&S system at a point in
time. In a perception survey, members of an organisation or industry are asked to give
their opinion on H&S performance in their organisation or industry. A perception
survey is a better predictor of the future H&S record than any other indicator and it
helps to clearly target what needs to be done to improve H&S systems in an
organisation. In a study of 160 organisations in which questionnaires were
administered to workers in year 2000 and 2004, concluded after comparing results of
worker perceptions from the two data points, that there had been an improvement of
about 9% in the H&S system. A perception survey established a baseline and
diagnosed what needed to be fixed at a point in time in an organisation or indeed at
industry or organisation level (Petersen, 2005:48).
A similar technique was proposed by Carder & Ragan (2003) based on a survey
system for measurement and improvement. Their study involved a H&S perception
survey of over 6000 employees in 50 chemical plant sites. The difference between
their study and that of Petersen (2005) was the feedback system with employees that
took part in the survey. A feedback system was included as well as convening focus
groups within an organisation to better understand the results before developing
focused action plans. Carder & Ragan (2003:161) argued that organisations that
undertook a perception survey, provided feedback to employees, convened focus
groups to understand the results, developed action plans that were reviewed by senior
management before implementation would experience a reduction in accidents of 25
to 50%.
However the improvement level of 25% to 50% may not necessarily be generalised to
include for example the construction industry, as the survey in Carder & Ragan
(2003) study was conducted in chemical processing plants and organisations. It also
assumes that an organisation would have systems in place that are assessed from time
to time. As in the study of Petersen (2005), the technique of Carder & Ragan (2003)
41
was based on the philosophy that there had to be a measuring tool to determine
current status in order to develop improvement actions on various elements that are
found to be weak or poor.
3.2.3
Behavioural audits
Another method that has been proposed for H&S performance improvement is the use
of a H&S management behaviour audit with goal setting and feedback. Cameron &
Duff (2007:870) argued that an intervention of management audit, goals and feedback
could be used to improve H&S management performance. They argued that it was
possible to implement a H&S management behavioural audit as a valid and reliable
measure of management safety performance in the construction industry and that goal
setting can improve management H&S performance. However, Lingard & Rowlinson
(1997) found that there was a limitation of all behaviour based H&S management
programmes in that worker behaviour, although important, is only one factor affecting
industrial H&S performance and H&S behaviour can only be achieved where a basic
H&S infrastructure is already in place. Nevertheless, the Lingard & Rowlinson (1997)
study was directed at workers while that of Cameron & Duff (2007), was directed at
behaviour modification of managers after realising that behaviour based method may
not work without management commitment.
3.2.4
Conformance to H&S systems
H&S performance can also be improved by emphasising conformance to H&S
systems (Manzella, 1999:33). Therefore, by measuring the level of conformance to
system requirements, risks can be identified before an injury occurs. Consequently,
improving systems based on identified deficiencies can help reduce risks and
produces improved statistical performance over time. According to Manzella
(1999:33), identifying root causes for system deficiencies and taking corrective
actions will produce long-term, improvements by continuing to lower risks.
The assumption is that organisations will have H&S systems in place and that the
systems are effective However, even though a systems approach may work, as argued
by Manzella (1999), It is only organisations that have intrinsic ‘H&S’ that are able to
withstand its operational dangers and still achieve its objectives. A systems approach
is not adequate in itself to prevent accidents and improve H&S performance (Hudson,
42
1999:8-10; Leveson, 2004:238). In high reliability organisations which are prime
examples of systems approach, organisational culture is in fact of profound
importance and therefore suggesting that best results are achieved with an
improvement in H&S culture (Reason, 2000a:770).
3.2.5 Developing a H&S culture
Reason (2000b:3) argue that only a H&S culture is able to provide any degree of
lasting protection. Consequently, many authors see cultural change as the way to
improve H&S performance (Fitzgerald, 2005:324; Wamuziri, 2006:173; Choudhry et
al., 2007:207; Fernandez-Muniz et al., 2007:636). As in other improvement
techniques discussed in this section, there has to be an assessment of the current
situation in order to provide a basis for an improvement plan. If H&S performance
can only be improved through culture change, then there has to be a way of measuring
the baseline H&S culture as well as the H&S culture after interventions to the baseline
(Fitzgerald, 2005:325).
The assessment of the H&S culture could be achieved by using climate assessment
tools to gauge the prevailing attitudes to H&S within organisations. A climate survey
highlights the areas in which there are issues and provides a basis for defining the
required improvement plan. However climate analysis is only an initial diagnosis step.
To operationalise the culture change approach, a methodology is proposed that first
measures the climate through a perception survey, followed by a planning workshop
with senior managers to develop action plans. The process is then repeated after
implementation of the action plans to determine the improvement (Fitzgerald,
2005:326).
Improving H&S culture was central to the Australian Government in developing the
model client framework for H&S improvement. The aim of the model client
framework was to facilitate the development of a positive H&S culture (Lingard et al.,
2009:134). According to Lingard et al., (2009:134), through the model client
framework, the Australian Government for whom the framework was developed
would strive to ensure that the major stakeholders involved in the planning, design
and execution of construction work, work collaboratively to allocate responsibility for
H&S and integrate H&S considerations into all project-decision making.
43
3.2.6
Improvement through design
Apart from the H&S performance improvement methods described earlier most of the
studies are directed at improving specific disparate parts of H&S management. Such
proposals include having laws or a set of regulations enacted that address for instance
the requirement for H&S to be taught in higher education. The argument is that this is
the best prospect for long term improvement across the whole industry (Anderson,
2010:97).
Other H&S performance improvement techniques include prevention through design
(Gambatese et al., 1997:32, Kinnersley & Roelen, 2007:35). Prevention through
design and therefore performance improvement is based on the reasoning that
designers set the scene for a project by defining the work to be done. Designers are in
a unique position to eliminate hazards from the project and to alter the design so that
risks can be tackled at source (Hetherington, 1995:5). This approach does not
however, enjoy much support from both sides of sciences. For example, Smallman
(2001:412) argue that employing “designing-in” H&S is certainly feasible through
employing orthodox high reliability theory. However, according to him, history has
revealed spectacular failures of systems designed around these principles and suggest
that it is probably time to re-think this philosophy. He further contend that from the
point of view of normal accident theory “designing-in” H&S is somewhat of an
oxymoron, since the act of developing high reliability in fact exacerbates a system’s
propensity to fail (Smallman, 2001:412).
Therefore, regardless of the proposed method and in order to achieve a holistic
approach to the problem of H&S, the input or involvement of key stakeholders in the
process of H&S performance improvement is critical. Lingard et al (2009:132) argue
that success in H&S management can only be achieved through teamwork between all
the project stakeholders, namely: the client, the designer, and the contractor. In fact,
Barrett (2007:279) on the issue of general improvement of the construction industry,
argue that institutions, governments, clients and procurers of constructed facilities
could all create the momentum needed to improve the industry.
It appears therefore that in order to address H&S performance improvement
holistically, measurement has to be conducted, and that success may only be achieved
if several approaches are utilised. However what is also apparent is that the H&S
44
culture is increasingly being recognised to be at the centre and critical to H&S
performance improvement (Wamuziri, 2006:169; Chinda & Mohamed 2008:114). In
addition, the involvement of stakeholders, such as clients, designers and contractors,
has also been identified as critical to H&S performance. A proposal for a method to
improve H&S performance should therefore consider the following:

applicability to the construction industry;

ability to be operationalised at both industrial and organisational level;

holistic in its approach to improvement;

based on H&S cultural change;

simplicity and ease of interpretation, and

Possibility of performance measurement to be done.
Performance improvement on H&S in the construction industry has been illusive
partly because of improvement methods which do not exhibit the above. Current
disagreements on how to operationalise the concept of H&S culture is one example of
one concept that is not being utilised even though it has the potential to contribute to
improving H&S performance in the industry.
3.3
CULTURE
A review of H&S performance improvement techniques or methods points to the fact
that there is no single approach that may be successful. However culture has been
identified to be all encompassing and transcends all the proposed techniques or
methods. It appears that H&S performance improvement in the construction industry
may not be possible without a major change in culture within the organisations or
industry involved (Riley & Clare-Brown, 2001:149; Baram & Schoebel, 2007:632;
Chinda & Mohamed, 2008:114). Culture has been identified as key to industry or
indeed organisational improvement (Fitzgerald, 2005:324). Culture has been
recognised as the main driver for change in other industries (Schneider, 1997:81)
including on issues such as H&S (Wamuziri, 2006:169; Chinda & Mohamed
2008:114). Therefore this section presents a review of literature on the concept of
culture and in particular, the H&S culture.
45
3.3.1
Why H&S culture?
The H&S performance record has been attributed to an improvement in the H&S
culture of organisations in many studies, including that of CRC (2006); Dingsdag et
al., (2006); Baram & Schoebel (2007); Fernandez-Muniz et al., (2007) and Chinda &
Mohammed, (2008). A better H&S outcome has been associated with the prevalence
of a positive H&S culture within an organisation and indeed the industry (Molenaar et
al., 2002; CRC, 2006; Chinda & Mohamed, 2008). According to Glendon & Stanton
(2000:193), organisational culture, however defined, is widely acknowledged to be
critical to an organisation’s success or failure. The prevailing culture is therefore very
important in so far as the improvement of construction worker’s H&S is concerned. It
is because of this observation that Chinda & Mohamed (2008:114) observed that it
seems that attempts to improve H&S performance will not be fully realised until the
H&S culture is improved. In fact, the Royal Academy of Engineering as quoted by
Riley & Clare-Brown, states that:
“The construction industry would benefit significantly from the study and
adoption of best practices from manufacturing and other industries. Elements
are identified which could lead to early advantage: they include better supply
chain management and considerable improvements in culture and
organisation.” (Riley & Clare-Brown, 2001: 150)
The achievement of an effective H&S culture is recognised to be a vital element of
achieving and maintaining satisfactory standards of H&S performance. In his
foreword to ‘A construction safety competency framework: improving H&S
performance by creating and maintaining a safety culture’, Fisher claims that in his
view, sustained improvement would not happen without cultural change (Dingsdag et
al., 2006:iii). The IOSH (2004:6) contends that it is insufficient, for example, to
provide safe equipment, systems and procedures if the culture is not conducive to a
healthy and safe working environment. To illustrate the inadequacy of systems and
procedures, a “poor H&S culture” was identified as a factor contributing to the
Chernobyl disaster by the international Atomic Energy Agency (INSAG, 1992:18)
and recently, the H&S culture was questioned by the National Commission on the BP
deep-water Horizon Oil Spill and offshore drilling (2011:vii). The report on the BP
deep-water Horizon spill and offshore drilling revealed that the immediate causes of
the Macondo well blowout could be traced to a series of identifiable mistakes made
by BP, Halliburton and Transocean that reveal such systematic failures in risk
46
management, that they place in doubt the safety culture of the entire industry
(National Commission on the BP deep-water Horizon Oil Spill and offshore drilling,
2011:vii).
To illustrate how other industries regard H&S culture, the so called ‘high reliability’
industries such as the Air traffic, Petro Chemical and Nuclear installations were
observed, where significant hazards were always present, operating organisations and
their regulators paid considerable attention to H&S assessment. They noted that these
assessments were mainly focused on leading indicators regarding H&S climate - a
measure or determinant of the prevailing culture, because, as stated earlier, most of
these industries had realised the importance of organisational culture (Flin et al.,
2000:178). Figure 5 shows that there has been a movement away from a systems
approach to culture based solutions and the results have shown an accompanying
reduction in the number of incidents (Hudson, 2007:700).
Figure 5: Relationship between interventions and number of incidents
(Source: Hudson, 2007:700)
Furthermore, the reason why the H&S culture of an organisation was considered to
transcend all techniques and methods of H&S performance improvement was that it
influenced the deployment and effectiveness of the H&S management resources,
policies, practices and procedures (Gadd & Collins, 2002:1). H&S culture did not
47
operate in a vacuum or in isolation. Rather it affected and in turn was affected by
other operational processes or organisational systems (Cooper, 2000:113). Dingsdag
et al. (2006:2) argued that it was the H&S culture that guided actions and reactions to
situations as they arose.
Culture is a powerful, latent, and often unconscious set of forces that determine both
our individual and collective behaviour, ways of perceiving, thought patterns and
values (Schein, 1999). According to Reason (2000a:770), it is the imbedded culture
that provides workers in an organisation both the reminders and tools to help them
remember the requisite reaction or action when faced with a situation. Even in terms
of innovation in response to situations in an organisation, culture lies at the heart of it
(Martins & Terblanche, 2003:64). It is argued that a positive culture leads to both
improved H&S and as well as organisational performance (Dingsdag, et al., 2006:2).
Culture creates a homogeneous set of assumptions and decision premises in which
compliance occurs without surveillance (Grote, 2007:642).
On the other hand, the presence of an inappropriate H&S culture of an organisation
can lead to catastrophic consequences once a hazard strikes (IET, 2009:1). A culture
of H&S is something that was lacking in the organisations that were affected by
disasters. Reason (2000a:770) argues that organisational culture takes up a profound
significance in responding to situations.
However, there is evidence of mismatch between problems and solutions to H&S as
has been revealed in the major accident investigations (Glendon & Stanton, 2000:201;
Mearns, Whittaker & Flin, 2003:642) such as the Chernobyl and Piper Alpha
accidents. Accidents are not always caused by human or technical failures only.
According to Wiegmann et al. (2002:2) accident causation has moved from technical
causes, human error causes, technical causes, social technical causes to a stage that is
referred to as organisational culture. Gadd & Collins (2002:3) and Flin et al.
(2000:178) argue that an organisation’s susceptibility to accidents does not arise from
just ‘human errors’, chance, or technical failures alone. On the contrary, it is the
entrenched organisational policies, procedures and standards that have been shown to
be at the heart of the accidents. This conclusion is especially reinforced by the fact
that progress has been made towards ‘engineering out’ the physical causes of
accidents in high technology plants.
48
Therefore it is now generally acknowledged that individual frailties and pervasive
organisation defects lie behind the majority of the remaining accidents (Lee &
Harrison, 2000:61). Recent accident causation theories, such as the constraintresponse theory by Suraji et al. (2001), also refer to the influence of culture. In this
theory, the basic assumption is that inappropriate human behaviour – response, is the
central feature in accident causation. However, it has been shown that the action or
reaction by an individual, at whatever level in response to whatever constraint, is
guided by culture ( Reason, 2000b:9). Since it has been established that the cause of
most accidents has its roots in the organisational H&S culture, the solution should
therefore lie in addressing the H&S culture and that is the reason why culture is so
important. Therefore, the solution has to match the problem.
Relying on a technical system approach to the problem of H&S would be a mismatch
in as much as addressing this aspect is also important. In a technical system, many
accidents are anticipated through H&S rules, prescriptive procedures and management
processes. However people did not always do what they were supposed to do and
according to Lee & Harrison (2000:62), undermined the system of multiple defences
that an organisation constructed and maintained to guard against injury to its workers
and damage to property. Pidgeon & O’Leary (2000:16), on the theory of man-made
disasters, explained that, despite the best intentions of all involved, the objective of
safely operating technological systems could be subverted by some very familiar and
‘normal’ processes of organisational life. Lee & Harrison (2000:62) argued that the
H&S management system was essentially a social system wholly reliant on the
employees who operated it and therefore the concept of H&S culture had been
identified as the way of formulating and addressing the problem of H&S.
Reason (2000a:770) noted that high reliability organisations are prime examples of
the system approach. These organisations anticipate the worst and equip themselves to
deal with it at all levels of the organisation. However, Reason (2000a:770) argues that
it is hard, even unnatural, for individuals to remain chronically uneasy, and so their
organisational culture takes on a profound significance because when individuals are
faced with constraining events, they may forget to be afraid, but the organisational
culture provides them with both the reminders and the tools to help them remember.
For these organisations, the pursuit of H&S is not so much about preventing isolated
49
failures, either human or technical, as about making the system as robust as is
practicable in the face of its human and operational hazards. High reliability
organisations are not immune to adverse events, but they have developed a culture, a
H&S culture to be able to deal with these setbacks (Reason, 2000a:770).
Therefore culture has become popular among the various methods to improve H&S
performance. H&S culture offers a more powerful support for changing and achieving
the desired improvement in the industry (Chinda & Mohamed, 2008:114). In fact,
IOSH (2004:11) contend that occupational H&S professionals need to recognise the
importance of a good H&S culture in influencing the H&S performance of the
organisations they advise, and appreciate the characteristics and benefits of a sound
H&S culture.
Furthermore, it has been shown that major accidents have been attributed to a poor
H&S culture. In addition, the problem of H&S has its roots in the H&S culture of a
nation, industry or organisation and therefore the solutions should match this problem,
which is to address the H&S culture. It has therefore become necessary and important,
to understand what culture (or more specifically), H&S culture is, in order to
appreciate the concept better and for it to become useful because:
“When we know what culture is, we know what needs to be changed for
culture to change. Only once we appreciate its nature can we understand how
it might be changed. When we know its role, we can comprehend its
importance” (Williams et al. 1993:11)”
3.3.2
What is culture?
There has been an increasing interest in the concept of organisational culture
particularly the H&S culture (Cooper, 2000:111; Dingsdag et al., 2006:1), mainly
because of its impact on H&S outcomes (Cooper, 2000:111; Wiegmann et al., 2002:3;
Choudhry, Fang & Mohamed, 2007:207;, Fernandez-Muniz et al., 2007:636). It is
therefore not surprising that there are also many different definitions of the concept.
According to Fernandez-Muniz et al. (2007:627), it is the recognition of the
importance of the H&S culture in preventing occupational accidents that has led in
part to the numerous attempts to define and assess the H&S culture in many
organisations.
50
As for the concept of culture itself, its conceptualisation and definitions have been
derived from the more general notion of organisational culture. Complications and
differences in definitions of H&S culture have also been due to the interdisciplinary
nature of culture, and as a result, the definitions and methods for studying
organisational culture tend to vary according to the academic discipline from which
they originate (Wiegmann et al., 2002:4). Although it is generally accepted that there
is no unanimity on the definition of H&S culture (Choudhry et al., 2007:207;
Fernandez-Muniz et al., 2007:627), or culture in general, many authors are however in
agreement that there is no unanimity on the definition of H&S culture (Cooper,
2000:111; Hopkins, 2006:876; Fernandez-Muniz et al., 2007:627) and as such there is
no common definition of culture (Baram & Schoebel, 2007:633; Mengolini &
Debarberis, 2007:520; Molenaar, Park & Washington, 2009:488). Fernandez-Muniz
et al. (2007:627) argue that despite several attempts to define the concept, there are
still gaps which remain in the literature and that there is no universal agreement about
the definition or content of the concept.
According to ACRCCI (2001:4) culture can be defined as being a set of mores,
values, attitudes, beliefs and meanings that are shared by the members of a group or
organisation. A similar definition is given by Dingsdag et al. (2006:2), who contends
that H&S culture is characterised by H&S beliefs, values and attitudes that are held by
an organisation.
Culture has also been defined in terms of it being a product (HSE 1993; Choudhry et
al. 2007; Grote, 2007). For instance, Choudhry et al. (2007:211) defined culture as
being a product of individual and group behaviours, attitudes, norms and values,
perceptions and thoughts that determine the commitment to, and style and proficiency
of, an organisation’s system and how its people act and react in terms of the
company’s on-going H&S performance in construction site environments. However
contrary to Choudhry et al. (2007) of culture being a product of behaviour, Reason
(2000b:13) postulate that behaviour is a product of culture just as much as accidents
are a product of the prevailing culture (IET, 2009:1; Glendon & Stanton, 2000:201;
Wiegmann et al., 2002:5). Culture has been defined to be a product of individual or
group behaviour, that some researchers argue is difficult to measure because
behaviour is difficult to measure with, say, a questionnaire except through
51
ethnographic studies or observation (Hopkins, 2006:877). Therefore, the aspect of
culture being a product in the definition by Choudhry et al., (2007:211) seems to be
problematic and makes the measurement of H&S culture difficult.
A definition by Wiegmann et al., (2002:8) is probably clearer because it uses the
concept of “product” appropriately. He defines H&S culture as being the enduring
value and priority placed on worker and public H&S by everyone in every group at
every level of an organisation. H&S culture refers to the extent to which individuals
and groups will commit to personal responsibility for H&S, act to preserve, enhance
and communicate H&S concerns, strive to actively learn, adapt and modify (both
individual and organisational) behaviour based on lessons learned from mistakes and
be rewarded in a manner consistent with these values. This definition suggests that an
action or behaviour is an outcome or a product of a culture. Several authors agree that
behaviour is an outcome or a product of culture (IET, 2009:1; Glendon & Stanton,
2000:195; Wiegmann et al., 2002:8). In much the same way, Dingsdag et al., (2006:2)
define culture as a description of values, norms, attitudes and beliefs that are held
collectively towards H&S within an organisation and it is these values, attitudes,
norms and beliefs that guide behaviour by indicating to employees what will be
rewarded or punished by the organisation.
Table 4 presents a number of definitions of the concept of H&S culture that were
reviewed in this study. It is clear from the definitions, understanding and perspectives
of H&S culture are wide but what is central to the H&S culture concept is the notion
of sharing - the sharing of experiences, beliefs, values, and assumptions (Molenaar et
al., 2002; Fitzgerald, 2005; Dingsdag et al., 2006; Fernandez-Muniz et al., 2007;
Mengolini & Debarberis 2007) which guide actions and reactions. The key terms
common to most definitions of culture include sharing, values and beliefs. These in
turn influence attitudes and produce perception and behaviour.
Therefore from a synthesis of all definitions, culture can be defined as a characteristic
set of assumptions, beliefs, values, knowledge, attitudes and symbols shared and held
by all members of a group which influences behavioural patterns and perceptions.
These can be surfaced through observation and or description of what goes on by
those that are part of the organisation. This definition is also summed up by the
statement ‘the way we do things here’ (Cooper, 2000:115).
52
Table 4: Definitions of culture
Author
INSAG (1991:1)
Hudson (1999:8-2)
Cooper (2000:115)
Guldenmund (2000: 251)
Cullen (2001: 60)
ACRCCI (2001:4)
Gadd & Collins (2002:1)
Molenaar et al., (2002:8)
Wiegmann et al. (2002:8)
Martins & Terblanche
(2003:65)
IOSH (2004:6)
Hartog & Verburg
(2004:58)
Definition of culture
Safety culture is that assembly of characteristics and attitudes in organisations and individuals which establishes that, as
an overriding priority, safety (Nuclear) issues receive the attention warranted by their significance
Who and what we are, what we find important, and how we go about doing things round here
Is that observable degree of effort with which all organisational members direct their attention and actions towards
improving safety on a daily basis
Safety culture is those aspects of the organisational culture which will impact on attitudes and behaviour related to
increasing or decreasing risk
The product of individual and group values, of attitudes and patterns of behaviour that lead to a commitment to an
organisation’s health and safety management. Organisations with a positive safety culture are characterised by
communication founded on mutual trust, by shared perception of the importance of safety and by confidence in the
efficiency of preventative measures
Culture as being a set of mores, values, attitudes, beliefs and meanings that are shared by the members of a group or
organisation
Concept that describes the shared corporate values within an organisation which influences the attitudes and behaviours
of its members. Safety culture is a part of the overall culture of the organisation and is seen as affecting the attitudes and
beliefs of members in terms of H&S Performance
Culture is defined as the beliefs, values and behaviours that are consistent among all members of the corporation
The enduring value and priority placed on worker and public safety by everyone in every group at every level of an
organisation. It refers to the extent to which individuals and groups will commit to personal responsibility for safety, act to
preserve, enhance and communicate safety concerns, strive to actively learn, adapt and modify (both individual and
organisational) behaviour based on lessons learned from mistakes and be rewarded in a manner consistent with these
values
Organisation culture is the deeply seated (often subconscious) values and beliefs shared by personnel in an organisation
Defined safety culture as consisting of shared values (what is important) and beliefs (how things work) that interact within
an organisation’s structure and control systems to produce behavioural norms (the way we do things around here)
Organisational culture is a set of core values, behavioural norms, artefacts and behavioural patterns which govern the way
people in an organisation interact with each other and invest energy in their jobs and the organisation at large
53
Table 4 (continued)
Author
Fitzgerald (2005:325)
Definition of culture
Consists shared values (what is important) and beliefs (how things work) that interact with an organisation’s structure and
control systems to produce behavioural norms (the way we do things around here)
Dingsdag et al. (2006:2)
safety culture is characterised by safety beliefs, values and attitudes that are held by an organisation
Choudhry et al. (2007:211)
the product of individual and group behaviours, attitudes, norms and values, perceptions and thoughts that determine the
commitment to, and style and proficiency of, an organisation’s system and how its people act and react in terms of the
company’s on-going safety performance in construction site environments
Mengolini& Debarberis
A pattern of shared, taken for granted basic assumptions held by the members of an organisation and developed through
(2007:520)
a process of external adaptation and integration
Fernandez-Muniz et al.
A set of values, perceptions, attitudes and patterns of behaviour with regard to safety shared by members of the
(2007: 628)
organisation; as well as a set of policies, practices and procedures relating to the reduction employees’ exposure to
occupational risks, implemented at every level of the organisation, and reflecting a high level of concern and commitment
to the prevention of accidents and illnesses
Misnan et al. (2007:405)
Culture means doing something automatically, spontaneously without having second thought about it
IET (2009:1)
The product of the individual or group values, attitudes, competencies and patterns of behaviour that determine the
commitment to, and the style and proficiency of an organisation’s H&S management
54
Using the H&S climate as an indicator for H&S culture as in Cox and Cheyne’s study
(2000), it is possible to determine the prevailing H&S culture in an organisation or a
project. The synthesised definition adopted for this study makes it possible to
operationalise the concept of H&S culture as it makes perceptions of those that are
part of an organisation to be a product of the prevailing culture. The observable
degree of effort with which all organisational members direct their attention and
actions towards improving H&S on a daily basis, also makes the product to be
measurable (Cooper, 2000:115).
Members of an organisation are able to report on what they perceive to be each
member’s effort, the observable degree of effort for ensuring H&S in an organisation.
The behaviour of members within the group could also be determined and
consequently be able to determine the H&S culture from the units of “that observable
degree of effort with which all organisational members direct their attention and
actions towards improving H&S on a daily basis” (Cooper, 2000:115; Wiegmann et
al., 2002:8). The units of ‘effort’ could differ and could be the degree to which
members give priority to H&S over production. However, this measurement process
may become complex when it comes to determining whether the culture is positive or
not. The other outcome of a H&S culture is injury rates although a reduction in injury
rates albeit very important, are not sufficient in themselves to indicate the presence or
quality of a H&S culture, whereas “that observable degree of effort” and the
perceptions or report of those that are part of the organisation or project, is something
that can always be measured and assessed (Cooper, 2000:115).
The assessment of H&S is also possible, because for all industries, organisations or
projects, there is always some form of culture that is present (Hudson, 1999:8-11;
Hudson, 2007:702). However, the differentiating thing is that some cultures could be
described as being positive as in Lee & Harrison’s study (2000), negative, a reporting
culture, or perhaps a tolerant culture depending on how the organisations wish to
describe the prevailing culture. Furthermore, organisational culture and indeed H&S
culture exists on a continuum and organisations can either be, for instance at the
pathological level or at the generative level as shown in Figure 6 (Hudson, 2007:704).
H&S culture may also be described as either good or poor, positive or negative (Biggs
et al., 2005:2; Misnan & Mohammed, 2007:404). At the pathological level which for
55
example may be described as poor H&S culture, organisations do not really care
about H&S. However, at the generative level, which is the good H&S culture, H&S is
described as being the way business is done and H&S is a way of life.
Culture can also be improved from one level to another. It is interesting to note that
the word ‘culture’ stems from the word cultivate or to till the soil (Høivik et al.,
2009:993; Merriam-Webster, 2011: online). The behaviour, attitudes and values of
members are dependent upon the sets of both conscious and unconscious beliefs that
individual members possess and that these beliefs are seen as a key element of
organisational culture (ACRCCI, 2001:4). It is also clear that in order to produce
better behaviour, the underlying beliefs should be appropriate. Similarly, good fruit
production needs good soil, so the support structure (stem), which are values and
attitudes, need to be appropriate in order to hold the upper fruit bearing branches.
Culture therefore, could be cultivated and improved, in order to have better outcomes.
In addition, it is also important to note that H&S culture does not operate in a vacuum
(Cooper 2000:113). Rather it affects and is affected by the external environment.
Wamuziri (2006:169) argue that not only can culture be influenced and changed, it
can also be learned. H&S culture can also be said to be a subset, sub-culture or part of
the overall organisational culture (Cooper, 2000:113; Wiegmann et al., 2002:8;
Wamuziri, 2006:169; Hudson, 2007:702).
3.3.3
Climate and culture
Related to culture is the concept of climate. Culture is sometimes confused with
climate and these concepts are sometimes used interchangeably (Hopkins, 2006:877;
Yang, Wang, Chang, Gou & Huang, 2009:1149). Culture embodies values, beliefs
and underlying assumptions, whereas climate is a descriptive measure reflecting a
group’s perceptions of the organisational atmosphere and the way things are done
(Flin et al. 2000:178). The H&S climate in a work place reflects employees’
perceptions about the organisation’s H&S management system including policies,
practices and procedures that show how H&S is implemented within the working
environment. It is exactly the snapshot that describes “the way we do things around
here now” (Choudhry et al., 2007:208) as well as the way things are understood,
judged and valued (Davies, Nutley & Mannion, 2000:112).
56
Figure 6: H&S culture ladder
(Source: Hudson, 2007:704)
A H&S climate can be regarded as the surface features of the H&S culture discerned
from the workforce’s behaviours, attitudes and perceptions at a given point in time
(Cox & Cheyne, 2000:114; Flin et al., 2000:178). It is a snapshot of the state of H&S
providing an indicator of the underlying H&S culture of a work group or organisation
(Flin et al., 2000:178).
Culture is not climate. However, culture can be described from the behavioural
patterns and from what goes on, the perceptions, by those that are part of an
organisation or industry.
To identify the prevailing culture one needs to complete an assessment of the existing
organisational climate to act as an indicator of the culture (IOSH, 2004:7; Wamuziri,
2006:173). The reason climate is used as an indicator for culture is that culture which
is generally intangible, leads to a tangible manifestation through climate (INSAG,
1991:1). The main task therefore as INSAG (1991:1) proposed, is to develop means to
57
use the tangible manifestations to test what is underlying, the intangible prevailing
culture. Observing and measuring the tangible elements is the only practical way of
determining the underlying culture because, as Wiegmann et al. (2002:4) argue, while
an organisation’s culture is revealed in its general patterns of attitudes and actions, the
deeper structure of its culture is often not immediately interpretable by outsiders.
Therefore a climate or a perception survey can be conducted to establish the
underlying culture. Figure 7 illustrates and summarises the synthesised framework of
what culture is and where climate fits in.
Culture influencing
factors
Factors of Health
& safety culture






Leadership
Involvement
Procedures
Commitment
Communication
Competence
Underlying culture
Shared & understood but
tacitly
Culture indicator
Climate
Culture



Beliefs
Values
Assumptions
Behaviour
Culture outcomes
Tangible/ observable
Outcomes / results

Perceptions

Artefacts

Actions/
behaviour

Accidents /
incidents
Environment
Figure 7: H&S culture framework
3.3.4
What are the factors of H&S culture?
Measuring H&S culture is one area where confusion has reigned partly because of the
many terms that have been used to describe what constitutes and influences a H&S
culture. Some studies have referred to the parts which form, shape or make up H&S
culture as characteristics (Hudson, 1999:8-1; Molenaar et al., 2002:19; Fitzgerald,
2005:325; Molenaar et al., 2009:488), indicators (Flin, Means, O’Connor & Bryden,
2000:178; Fernandez-Muniz et al., 2007:628), factors (Misnan, Mohammed,
Mahmood, Mahmud, & Abdullah, 2008:1911), determinants (Chinda & Mohamed,
2008:114), elements(Fernandez-Muniz et al., 2007:628), enablers (Chinda &
Mohamed, 2008:114), and attributes (Chinda & Mohamed, 2008:116) of H&S
58
culture. It is therefore important to know what these terms mean to partly reduce the
confusion. The Oxford dictionary (2002) defines the identified terms as follows:
1. Attributes: (v) or attribute is a characteristic quality. An object traditionally
associated with a person or thing. It can also be referred to as characteristic or
feature;
2. Characteristic: (adj) a typical feature or quality of something or somebody. A
distinguishing feature;
3. Determinant: (n) A factor which determines the nature or outcome of
something;
4. Element/content: (n) A basic part of something. It can also be referred to as
component, constituent;
5. Enabler: (v) This is from the word enable which is to provide with the ability
or means to do something. To make possible;
6. Factor: (n) A circumstance, fact, or influence that contributes to a result;
7. Indicator: (n) a thing that indicates a state or level.
A scrutiny of the above terms reveals that the terms ‘attribute’ and characteristic refer
to the description of quality of something. Therefore, with reference to culture this
would refer to the quality or an identifying feature such as a reporting culture.
The term ‘determinant’ and ‘factor’ refer to a circumstance or aspect that will
contribute to a result. Therefore, referring to culture, this term would describe an
influence that has a bearing on the quality or type of the culture. In other words,
without the factor or determinant it is impossible to have the type or quality of culture
being sought. Factors therefore may be leading indicators of a H&S culture.
The term ‘enabler’ on the other hand is more of a catalyst to achieving a result. A type
or quality of the culture does not necessarily depend on the enabler but it would be
beneficial if it is present.
An ‘indicator’ is a thing usually described as being an object such as a meter, a clock,
etc. that indicates the level of a result. Therefore, with reference to H&S culture, this
could be an instrument or certain exhibits from the culture that could be observed or
measured to tell the type or quality of the prevailing culture.
59
From the definition of culture, it is said to be composed or consisted of beliefs, values,
assumptions, knowledge and attitudes (IOSH, 2004:6; Biggs et al., 2005:2; IET,
2009:1). Therefore beliefs, values, assumptions, knowledge and attitude are
“elements” of culture generally and in particular H&S culture. Hudson (1999:8-3)
correctly refers to beliefs and values as components of a H&S culture. An element is a
part of something or somebody. Because the elements, namely beliefs, values,
knowledge, assumptions and attitude constitute H&S culture, they collectively dictate
behaviour in all members of a group because behaviour is an outcome of culture
(Cooper, 2000:116; Reason, 2000b:13).
On the other hand, the external or internal forces that influence or determine H&S
culture are referred to as factors of H&S culture. Both elements and factors can
inform of the prevailing culture. An attempt to measure these can therefore inform of
the level of H&S culture that is prevalent in a project, organisation or industry.
Assessing or measuring H&S is however not easy and is a subject that is without
controversy. Consequently there is no agreement on the method that should be used to
measure culture (Cooper, 2000:111; Fernandez-Muniz et al., 2007:628). Therefore the
continued debate on how culture should be measured and what methods should be
used has resulted in problems to operationalise the concept.
Nevertheless, it has been established that culture can be determined by either
measuring the elements that make up H&S culture or the factors that determine H&S
culture. Both these aspects are leading indicators. However, the elements are
intangible and therefore measurement may be problematic. On the other hand, it is
relatively easier to measure factors. Factors of H&S culture can be thought of in terms
of factors that contribute to the health of an individual. The health of an individual is
dependent on the diet that that individual is following. Following a good healthy diet
could result in a better health. In this case, the diet is a contributor or a ‘factor’ but not
the only ‘determinant’ to the health of that individual. Diet is therefore in this case a
leading indicator of the individual’s health. It is certain that a good diet will result in
good health. On the other hand, a bad diet may result in bad health immediately, or
after some years. However, for positive results, emphasis should be placed on a good
diet. Molenaar et al., (2002:27) illustrate this point by referring to a cholesterol test.
They contend that just as a poor cholesterol test does not absolutely predict a heart
60
attack, a poor H&S culture test does not indicate an impending accident. However,
both are good indicators that a catastrophe is more likely and therefore change is
required.
3.3.4.1 Factors of H&S culture
In order to be able to assess or measure the factors which are essentially leading
indicators of H&S culture, the factors should be known first. According to Chinda &
Mohamed, (2008:128), the aspect of leadership, policy and strategy of an
organisation, people, partnerships and resources, processes and goals are key
contributors to and shape the outcome of a H&S culture. In order to change culture,
that is to influence culture positively, IOSH (2004:7) contend that there is need for
commitment to change and there has to be leadership at the highest management
level. Wiegmann et al., (2002:11) identified organisational commitment and
involvement, employee empowerment, a reward and reporting system as being
contributors to the H&S culture of an organisation. Specific aspects such as education
and training have also been identified as vital contributors to a good H&S culture
(Fitzgerald, 2005:326; Pellicer & Molenaar, 2009:44). In a study by Choudhry et al.
(2009:46), 11 factors were identified namely commitment and involvement,
procedure, psychological and economical features, self-esteem, workers’ experience,
performance pressure, working environment, job security and education as being
contributors to the H&S culture of an organisation.
Other factors such as communication (Gadd & Collins 2002:14; Mohamed, 2002:376;
IOSH, 2004:7; Dingsdag et al., 2006:2; Havold, 2007:175, IET, 2009:1); competence
(Gadd & Collins,2002:12; Mohamed, 2002:376; IOSH, 2004:6; IET, 2009:1), and
leadership (Fitzgerald, 2005:325; Dingsdag et al., 2006:2; IET, 2009:1) have also
been identified to contribute to H&S culture. Risk perception of workers (Entec
1999:16; Flin et al. 2000:186; Gadd & Collins, 2002:14) and more generally policies,
procedures and rules equally contribute to H&S culture (Flin eta al, 2000:180;
Mohamed, 2003:81; Fernandez-Muniz et al. 2007:628). In addition, the aspect of
performance measurement and feedback of results, determines the prevailing H&S
culture (Pidgeon & O’Leary, 2000:18; Gadd & Collins, 2002:17; IOSH, 2004:11;
Fitzgerald, 2005:326). Other authors add the factors of rewards, incentives and
61
disincentives to factors that determine H&S culture (Wiegmann et al., 2002:12;
Molenaar et al., 2009:493).
Table 5 is a summary of the factors of H&S culture that have been identified. The
factors common to most studies include leadership; involvement; procedures;
commitment; communication and competence. These have been found to be common
to most studies and some have been combined in to one factor.
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Procedures/rules
/policy/plans
X
X
X
X
X
X
X
X
X
Competence
/Training/talks
X
Involvement
Commitment
Booth(1995)
Glendon & McKenna (1995)
Entec (1999
Flin et al. (2000)
Pidgeon & O’Leary (2000)
Gadd & Collins (2002)
Wiegmann et al. (2002)
IOSH (2004)
Fitzgerald (2005)
Human Engineering (2005)
Dingsdag et al.(2006)
Parker et al. (2006)
Wamuziri (2006)
Havold (2007)
IET (2009)
Molenaar et al. (2009)
Leadership
Cultural element
Communication
Table 5: Common factors of H&S culture
X
X
X
X
X
X
X
X
X
X
X
x
X
X
H&S leadership
The Leadership factor seemed to be common to a number of studies on culture
(Booth, 1995, IOSH, 2004, Fitzgerald, 2005, Dingsdag et al., 2006, IET, 2009).
Leadership is a critical factor of H&S culture and has been recognised as such in
many studies (Harvey, 2002; Hopkins, 2006; Baram & Schoebel, 2007; Choudhry et
62
al., 2007; Mengolini & Debarberis, 2007). Since leadership is a critical factor of H&S
culture, there is need therefore to start looking at how clients for example are leading.
Krause & Weekley (2005:34) argue that H&S improvement to higher levels require a
new approach to leadership in H&S that accounts for the leader’s role in reducing
exposure and in creating a climate and culture which is favourable for H&S.
Consequently the type of leadership they are talking about is a H&S leadership that
can be demonstrated through inter alia, incorporating H&S considerations at every
level of decision making, having a policy on H&S, active monitoring of H&S
programmes, monitor and control all stakeholders and coordination of all stakeholders
involved (Lingard et al., 2009:134).By so doing, clients will set the tone for H&S and
impact the overall project culture. Furthermore, clients also need to be involved in
H&S programs apart from providing leadership.
Involvement
Involvement, whether of clients, contractor’s top management, designers or
employees has been mentioned in many studies to be a factor of H&S culture (Flin et
al., 2000; Gadd & Collins, 2002; Wiegmann et al., 2002; Fitzgerald, 2005; Teo et al.,
2006; IET, 2009; Molenaar et al., 2009). The involvement factor is critical to H&S
culture (ACRCCI, 2001:24; Harvey et al., 2002:31; Mohamed, 2003:82; Grote,
2008:452). All stakeholders including employees are considered to be involved in
H&S management by inter alia the, extent to which they get personally involved in
critical H&S activities; presence and contribution to H&S meetings, planning
sessions; management’s contribution to training; active management oversight of
H&S critical operations; ability of management to “stay-in touch” and extent to which
there is good communication about H&S issues from and between all stakeholders.
According to Teo et al. (2006:1586), it is difficult to inculcate a H&S culture in an
organisation or indeed a project if management is not involved. Involvement however
also entails having procedures in place for H&S management.
Procedures
The term procedures, was taken to be a combination of various terms or descriptions
such as the established processes, rules, programs and protocols. Procedures therefore
may include inter alia monitoring and analysis of H&S implementation (evaluation)
(Cheung et al., 2004:160; Makino, 2006:2); formal inspections and audits (Entec,
63
1999:40; Toellner, 2001:44); H&S planning and definition of goals (Saurin et al.,
2005:1182); schedule H&S in prequalification and contracts for all parties;
performance measurement (Pidgeon & O’Leary, 2000:21; Gadd & Collins, 2002:15;
Fitzgerald, 2005:327); established rules, policies, programs and protocol (Flin et al.,
2000:188; Mohamed, 2003:81; Fernández-Muñiz, 2007:628); having a H&S structure
(Entec, 1999:43); hazard identification and risk assessment and H&S design, plan and
specification (Entec, 1999:39; Teo et al., 2006:1587). A review and thematic analysis
of H&S culture indicated that having formal procedures has an influence on H&S
culture (Glendon & Stanton, 2000:206; Flin et al., 2000:180; Wiegmann et al.,
2002:11; Dingsdag et al., 2006:3; Havold, 2007:191; Molenaar et al., 2009:495).
However, having procedures alone is not enough because there has to be commitment
from all involved and especially the client.
Commitment
H&S commitment especially from management such as clients, contractor top
management or indeed designers has also been recognised to be critical to H&S
performance (Entec, 1999:39; Flin et al., 2000:179; Toellner, 2001:47; Mohamed,
2003:81). Commitment to H&S can be demonstrated by inter alia the client for
example having a positive attitude toward H&S (Misnan & Mohammed, 2007:403);
actively promoting H&S in a consistent manner across all levels; providing adequate
finance and other resources for the implementation of H&S and by supporting the
development and implementation of various H&S activities (Misnan & Mohammed,
2007:403). Commitment could also be seen through for example the client
establishing a reward system for a safe behaviour; conducting regular H&S tours
(Toellner, 2001:47); getting involved in incident, accidents and ill-health
investigations and essentially through that observable effort to ensure every aspect of
operations and work schedules are routinely evaluated and modified if necessary
(Cooper, 2009:37). Probably the best descriptor of commitment for the client could be
elevating the status of H&S above production and profits in everything that the client
does and this can be seen in their attitude to say workers stopping work because of a
H&S issue.
Management commitment is key in influencing the H&S culture of an organisation
(Gadd & Collins, 2002:4; Wiegmann, 2002:11, IOSH, 2004:7; Havold, 2007:175;
64
Fernandez-Muniz et al., 2007:628; Choudhry et al., 2009:211). Management attitude,
and therefore commitment, can influence many aspects of the organisation and
project, including success of H&S initiatives, reporting of near-misses, incidents and
accidents; employees taking work related risks; influencing production pressures;
health interventions; effectiveness and credibility of H&S officers and committees
(Gadd & Collins, 2002:22).
Issues such as the status of H&S officers and the effectiveness of H&S committees
can also be a reflection of management commitment. Further, management
commitment could be seen in worker perceptions which inform things such as
employees’ willingness to report near misses, incidents, and or accidents (Gadd &
Collins, 2002:22). Clearly management commitment is an important factor to H&S
culture. However commitment is also a function of H&S communication (Hofmann &
Morgeson 1999:288). Therefore characterisation of H&S culture may not be adequate
without the H&S communication factor.
Communication
H&S communication factor is widely recognised in many studies as being critical to
H&S management (Entec, 1999; Dingsdag et al., 2006; Gadd & Collins, 2002;
Mohamed, 2003; IOSH, 2004; Havold, 2007; IET, 2009). Effective H&S
communication may entail inter alia having, formal reporting systems (Mearns et al.,
2003:644); feedback system; worker involvement in planning and review of H&S,
effective dissemination of risk findings to all stakeholders; clear H&S policy
statements on expected performance standards regarding H&S; risk control
information available to all stakeholders (Mearns et al., 2003:670); H&S briefings,
posters and or bulletins (Human Engineering, 2005:10). The communication system
may also be said to be working properly if all parties are comfortable to use the
reporting system. It is also important to observe that it is not so much the channels of
communication that matter but also the quality of communication (Entec, 1999:19).
Competence
The competence factor includes the aspect of education and training (Entec, 1999:4;
INSAG, 2002:5; Bomel, 2001:8.3; Fitzgerald, 2005:326; Fernandez-Muniz et al.,
2007;631). ; In addition competence also refers to the capacity of the organisation to
manage considering its skills and experience (Gadd & Collins, 2002:12). It has been
65
argued that increasing the skill and competence base of H&S critical role holders
would lead to behaviours that would build and maintain a H&S culture that would
work according to external industry demands and influences (Cipolla et al., 2006:4).
An organisation can therefore be considered competent if: it has permanent H&S
staff, conducts H&S training at all levels; H&S staff have knowledge and the skills
that are accompanied by formal qualifications in H&S management. Training, which
is just one aspect of the competence factor, has been identified to be critical to H&S
performance in many studies (Entec, 1999; Biggs et al., 2005; Fernandez-Muniz et al.,
2007; Molenaar et al., 2009). Therefore a level of training activities obtaining in an
organisation can be indicative of the level of that organisation’s competence.
Conversely this factor is essential to H&S culture because it impacts behaviour
(Molenaar et al., 2009:489). Table 6 summarises the indicators of factors of H&S
culture as discussed in this section.
However, the internal culture of an organisation is also influenced by other external
factors. Cooper (2000:113) rightly argues that culture does not operate in a vacuum it
affects and is affected by external factors. Therefore, the next section describes the
external factors namely: legislative, political, economic, technology and social
environment influence on a project culture. The influence of the construction client as
an external factor to H&S culture of construction project stakeholders is also
discussed.
3.4
EXTERNAL ENVIRONMENT’S INFLUENCE ON H&S CULTURE
3.4.1 Legislative
Legislation has a certain influence on the H&S culture of a nation, industry,
organisation and or a construction project. The manner in which people act is
conditioned by requirements set at a high level such as legislative (INSAG, 1991:5).
According to INSAG (1991:5), the highest level affecting H&S is the legislative level,
at which the national basis for H&S culture is set. Consequently the European Union
used the legislative framework as one of the tools to promote a culture of risk
prevention in their strategy to improve quality and productivity (Lethbridge, 2008:14).
Similarly, the health and safety authority (HSA) in Ireland has also advocated for the
introduction of a regulatory framework of duties for clients, designers and supervisors
to ensure that due consideration is given to the health, safety and welfare of persons at
66
work on construction sites, noting that this was the most appropriate method to
improve performance (DoETE, 2006:3).
Table 6: Indicators of factors of H&S Culture
Factors
Indicators
Leadership
 Incorporating H&S considerations at every level of decision
making;
 Having a policy on H&S;
 Active monitoring of H&S programmes;
 Monitor and control all stakeholders;
 Coordination of all stakeholders involved
Involvement
 Extent to which stakeholders get personally involved in critical
H&S activities;
 Presence and contribution to H&S meetings, planning sessions etc.;
 Management’s contribution to training;
 Active management oversight of H&S critical operations;
 Ability of management to stay ‘in-touch’;
 Extent to which there is good communication about H&S issues
from and between all stakeholders;
Procedures
 Monitoring and analysis of H&S implementation;
 Formal inspections and audits;
 H&S planning and defined goals;
 Schedule H&S in prequalification and contracts for all parties;
 Performance measurement;
 Established rules, policies and protocol;
 Having a H&S structure;
 Hazard identification and risk assessment;
 H&S design, plan and specification.
Commitment
 Demonstrating, a positive attitude toward H&S;
 Actively promoting H&S in a consistent manner across all levels
 Providing adequate finance and other resources for the
implementation of H&S;
 Supporting the development and implementation of various H&S
activities;
 Demonstrating that effort has been put forth to ensure every aspect
of operations, and work schedules are routinely evaluated and
modified if necessary;
 Establishing a reward system for safe behaviour;
 Conducting regular H&S tours;
 Getting involved in incident, accidents and ill-health investigations;
 Deliberately setting H&S as an important agenda item in meetings;
 Elevating the status of H&S above production and profits.
67
Table 6 (continued)
Competence





communication








An organisation having permanent H&S staff;
H&S training at all levels;
Having H&S knowledge and skills;
Conducting induction programmes and refresher courses
H&S formal qualifications for H&S staff
Formal reporting systems;
Formal structured feedback system;
Worker involvement in planning and review of H&S;
All parties comfortably use the reporting system;
Timely and valuable feedback to all parties;
Risk findings being disseminated to all concerned;
Clear H&S policy statements made by management;
Clear statements on expected performance standards regarding
H&S;
 Risk control information being made available to all available;
 H&S briefings, and or bulletins.
Research has shown that legislation or targeted regulations can influence H&S
performance of either a project, industry or a stakeholder. A study conducted in South
Africa revealed that construction regulations were perceived to have had an impact on
H&S performance (CIDB, 2008:11). It was observed from this study that the
manifestations of the impact of the construction regulations in South Africa were wide
spread and in particular increased H&S awareness and consideration for, or reference
to, H&S by project managers and general contractors. A more specific relationship
between legislation and H&S performance can be seen from the new ergonomics rule
in the USA. An improvement in H&S performance was observed when the rule was
introduced in 2000 and it dropped when the rule was repealed to remove it in 2003
(Foley, Silverstein, Polissar & Neradilek, 2009:1).
Regulation that is appropriate works. A study conducted in the United Kingdom, by
the Chartered Institute of Building (CIOB) informs that about 61% of the respondents
felt that the construction industry had become healthier and safer after the
implementation of the Construction Design and Management (CDM) regulations
(CIOB, 2009:20).
Despite these notable benefits, other studies have advocated for self-regulation. Under
the philosophy of self-regulation, which is a principle that those who create the risks
should be responsible for their control (Fairman & Yapp, 2005:498), the central
68
responsibility is placed on each organisation’s management for developing,
implementing and monitoring its own H&S management system. Self-regulation may
be described as the control of a collective behaviour by an industry or the private
parties concerned without the direct involvement of external authorities (Lenox &
Nash, 2003:).
However, there is contention about whether organisations should be self-regulated or
strictly follow government policies. Findings by Lazarevic et al., (2004) for example,
were that although the Occupational H&S Act in Australia allowed for businesses to
become self-regulated, small companies were not able to do that. In this study
respondents argued that the nature of the industry being a conglomeration of small
businesses made self-regulation difficult. Further, the lack of a level playing field
between small and large companies meant that some would not worry about selfregulation and do whatever was necessary to get the job done cheaply. Lazarevic et al
(2004) found that self-regulation was open to too much interpretation and as such
there needed to be government regulations to adhere to.
Self-regulation as an alternative in the construction industry is likely to be ineffective
and therefore not appropriate in developing a suitable H&S culture considering the
industry’s uniqueness and complexity. According to a review by Hasle & Limborg
(2006), there are high accident risks in smaller enterprises and this is especially the
case for fatal and other serious injuries. Hasle & Limborg (2006) observed that the
risk is higher in smaller enterprises and the ability to control risks is lower. Therefore
the consequences of non-compliance are too severe to exclude the involvement of the
courts in enforcement actions. In addition, without explicit sanctions for nonconformance, self-regulation is likely to attract more polluting organisations (Lenox
& Nash, 2003:).
In most parts of the world the problem is not necessarily with whether the legislative
regime is that of forced or self-regulation. The problem appears to be that of
enforcement. In the United Kingdom, the House of Commons noted that although,
breaches of H&S regulations are serious criminal offences, and legislation provides
for penalties, courts have, however, tended not to impose maximum penalties
available (House of Commons, 2004:4). The House of Commons suggested that the
HSE needed to increase the deterrent effect of prosecution (House of Commons,
69
2004:14). In South Africa, the problem was identified to be that of enforcement
describing it to be inadequate (CIDB, 2008:ii).
The legislative environment can influence H&S culture. Therefore it is easy for
example to allocate resources if there is pressure from regulations. For instance, by
demanding that resource allocation for H&S be specified in the proposal, the
customer can ensure that money is actually put aside for this purpose, since at this
stage the potential contractor is not inclined to question customer demands if the
customer is responding to legislative requirements (Torner & Pousette, 2009:402).
However, legislative measures alone are not adequate without economic incentives.
According to the European Agency for Safety and Health at Work (2010:24),
regulation and economic incentives are complementary.
3.4.2
Economic
Critical to the success of a project is the use of incentives as a method for promoting a
culture within which technical and process innovation can flourish (Tang et al.,
2008:465). Similarly, better results in H&S may be achieved with economic
incentives. The European Agency for Safety and Health at Work (2010:9) observe
that economic incentives can be effective in promoting H&S. It is in view of this that
Elsler & Nikov (2003:913) contend that there is a need for economic incentives to
proactively promote H&S. Some of the reasons why economic incentives should be
considered include the failure of strict regulation approaches, costs involved in
bringing organisations to courts for non-compliance and the low level of fines which
have failed to encourage companies to comply (Elsler & Nikov, 2003:911). However,
in order for the economic incentives to be effective, they should be directed at
organisation or national level (European Agency for Safety and Health at Work,
2010:14). These economic incentives may entail linking fiscal incentives such as
lower accident insurance premiums or tax rates to a good H&S performance. Other
methods to incentivise employers to implement H&S include matchup funds where a
grant is given to an employer equal in amount to the amount to be spent on H&S or
linking an incentive amount to a voluntary audit or inspection (European Agency for
Safety and Health at Work, 2010:26).
According to Tang et al (2008:457), incentives give contractors an opportunity to
work in partnership with clients to achieve good results by giving contractors a direct
70
financial stake in the efficient execution of the project and may create a more
proactive, cooperative relationship between contracting parties. They argue that
incentives enable all parties to make reasonable return and to bear appropriate risks
and help to build a flexible system that apportions the risks depending on the
circumstances of each party and the ability to bear the risk.
The design and use of economic incentives should take into account the constraints
and risks of a project, organisation or indeed the Nation (Elsler & Nikov, 2003:913;
Tang et al., 2008:457). Incentives should make risk allocation fairer, because
incentives can be seen as the sharing of rewards from good performance and this may
motivate the participants to perform better (Tang et al., 2008:459).
The reason why economic incentives work or are seen as one of the solutions to
proactively improve H&S, is partly because of the cost of providing H&S by the
project implementers such as contractors. Contractors work at reducing the cost in
order for them to remain competitive. Bishop et al (2009:243) rightly argue that the
unacceptability of occupational H&S performance of the building and construction
industry is attributed to the powerful competitive forces in the industry which
ultimately work against H&S. He observed that the industry strives to complete
projects on time in order to reduce costs and too often H&S is neglected. The solution
may be a cultural and behavioural change and this may only come about by
harnessing the competitive forces in the industry to work for occupational H&S.
Economic impact on H&S performance can also be seen in the lack of resources. A
lack of resources or underfunding for H&S programs limits any action. For example,
in Tanzania, less than 1% of the Labour Department’s budget was allocated to
occupational H&S (Kamuzora, 2006:65). This kind of allocation can result in a low
capacity to enforce legislation and failure to conduct inspection and surveillance.
Investment in research and training may be impacted on and one can argue that
ultimately the H&S culture may be influenced in a negative way (Kamuzora,
2006:65). According to Cotton et al. (2005:23), contractors or indeed other
stakeholders are unlikely to see the need of implementing H&S without the
application of incentives or sanctions in the developing countries.
71
The benefits of incentives are clear. The European Agency for Safety and Health and
Work (2010:10) demonstrated from a case study of six organisations in Europe that
improvements of 25 to 70% were possible with economic incentives. However for the
incentives to be effective, they should be provided by national and or international
organisations. Consequently political will is necessary for the national or international
organisations to be involved.
3.4.3
Political influence
Political will, has influenced the H&S culture, especially at national level. Political
will was identified as one of the many barriers to the implementation of existing H&S
policies at international, national and sector levels (Lethbridge, 2008:14). Lack of
political will is a barrier and according to Lethbridge, lack of political will can
manifest itself in things such as a low level of ratifications of ILO conventions that
address H&S, low level of resources allocated to the management of H&S, and a lack
of investment in ensuring compliance as well as capacity building (Lethbridge,
2008:19). Consequently, as Kamuzora (2006:65) rightly contends, a low funding for
occupational H&S programs effectively limits any action and results in a poor H&S
culture.
Political will is even more critical for developing countries. The majority of
developing countries do not have adequate political mechanisms to translate for
example scientific findings into meaningful occupational H&S policies (Nuwayhid,
2004:1916). Consequently, standards are not enforced by most Government agencies
partly because of the lack of an enabling environment to promote H&S (Kheni, et al.,
2007:160).
Lack of political will is probably the most damaging for occupational H&S policy
implementation. Most national governments in developing countries have a
centralised type of decision making. A centralised decision making makes it more
difficult to implement H&S agreements if there is no political will. Most governments
in developing countries are faced with the challenge of for example economic
development growth. As a result, H&S standards remain neglected because of these
competing social, economic and political challenges (Nuwayhid, 2004:1916;
Kamuzora, 2006:65).
72
Political will can also be seen in the allocation of resources or even in the structuring
of the government departments. Lethbridge (2008:19) argues that this can be a
fundamental problem. The author observes that this problem can be seen in the
separation of responsibilities for H&S between different departments. An example
would be the management of H&S by two different departments such as the
department of Health and that of Labour. Both of these departments handle issues to
do with H&S therefore would sometimes result in confusion of roles and inaction.
3.5
CHAPTER SUMMARY
This chapter has discussed H&S performance improvement, the concept of culture
and the influence of the external environment on the organisation’s or industry’s
culture. It appears from literature that in order to address H&S performance
improvement holistically, measurement has to be conducted, and that success may
only be achieved if several approaches are utilised. However what is also apparent is
that the H&S culture is increasingly being recognised to be at the centre and critical to
H&S performance improvement (Wamuziri, 2006:173; Chinda & Mohamed
2008:114). In addition, the involvement of stakeholders, such as clients, designers and
contractors, has also been identified as critical to H&S performance. The use of
economic incentives coupled with the deterrent effect of legislation is considered to
be effective to influence the prevailing H&S culture
73
CHAPTER FOUR
INFLUENCE OF CONSTRUCTION CLIENTS ON PROJECT
H&S PERFORMANCE
4.1
CLIENTS’ ROLE IN H&S PERFORMANCE
H&S during the construction process is conventionally considered to be the
contractor’s responsibility. When construction accidents happen, perceived factors of
causation are always associated with the contractor’s management failures or site
operative failures to control unsafe site conditions or unsafe actions (Abdelhamid &
Everett, 2000:55; Suraji et al., 2006:52). The general perception is that construction
H&S is a matter of construction management rather than client’s management or
indeed other participants’ management (Suraji et al., 2006:52). Studies concentrating
on factors that relate to the contractor create the impression that the main problem is
with contractors and therefore H&S performance improvement can only be achieved
by addressing contractor issues.
It should be appreciated however, that the H&S of any operation is determined long
before people, procedures and equipment come together at the work site. In other
words before the contractor commences work. As Behm (2005:590) puts it, many if
not all hazards are designed into construction projects. Design in this case is a process
that happens long before the contractor commences work on site. However, this is not
to say contractors have no role in accident causation. To the contrary, contractors’
especially top management has influence over H&S performance.
The contractor’s influence on H&S can be seen in the studies that have been
undertaken before. It has been found that for a number of accidents, the inappropriate
construction planning, construction control and operation are some of the frequent
contributing factors (Suraji, Duff & Peckitt, 2001:339). Research also indicates that
the attitude of contractor’s top management impacts on H&S. According to Gould &
Joyce (2002:367), top management’s attitude can be reflected on the job site in many
74
ways, such as training, housekeeping, toolbox talks, meetings, and adherence to H&S
measures, maintenance of equipment and tools, and intolerance of violations. It is a
natural tendency for those in authority to exert control and show exemplary behaviour
to those under them. Managerial commitment is reflected in the H&S culture that is
prevalent and the number of accidents that result. Usually, these occur because
management allows

employees to work in unsafe environments;

unsafe conditions to remain in a work environment;

at-risk behaviours to continue without intervening; and

unsafe machines and equipment to be used.
The commitment of top management to H&S determines the successfulness of any
H&S programme. A study conducted by Atkinson (1999), revealed that simple errors
at management level often surface as operative errors. A similar study, conducted by
Mackenzie et al. (1999), revealed that positive action by management within the UK
construction industry could have prevented about 70% of the fatalities. Similarly,
Yule, Flin & Murdy (2007:139) found that management commitment had a positive
effect on worker H&S performance. The commitment of top management is a
prerequisite for the successful implementation of H&S in the construction industry as
it addresses factors at the construction stage (Human Engineering, 2005:8).
It is however unlikely that H&S performance improvement can be achieved
throughout the industry by only focusing on addressing issues at the construction
stage and the contractor specifically. This is due partly to the difficulty conditions that
contractors operate in. Suraji et al. (2006:59) argue that contractors operate under a
number of constraints including the actions of designers as well as the action of
clients, and therefore may fail to provide safe working conditions, at least in part as a
result of these constraints.
Despite what Suraji et al. (2006:59) refer to as constraints under which contractors
operate, contractors in most parts of the world are also forced by the legislative
regime and the power of money to consider H&S as a fundamental part of project
management (Shash & Ahcom, 2006:68). One may argue therefore that most
75
contractors are only complying because they are forced to do it, and questions of
sustainability and culture change may therefore arise.
The constraints continue to create a negative attitude in some contractors as they are
seemingly forced to bear the H&S responsibility alone. Therefore, in most cases
contractors opt for non-compliance as implementing H&S increases operating costs.
Loosemore & Andonakis (2007:587), in a research conducted in Australia among
subcontractors, found that cost is by far the greatest barrier to compliance by
contractors. They found that it is possible when contractors perceive compliance to
exceed the benefits in terms of cost; to opt for non-compliance or simply turn a blind
eye to issues of H&S. The first element to be eliminated when it comes to cost
reduction would most probably be H&S.
With such contractor problems for instance, there is a need to move upstream for
interventions. Bomel (2001:5.3) suggest that the culture of client organisations
presents considerable opportunities for H&S improvement in the construction
industry. Therefore, there should be an increasing interest to find interventions that
address client issues and procedures as well.
Striving for better H&S performance will remain elusive if the client is not seen to be
actively involved in H&S implementation. Huang & Hinze (2006a:164) rightly argue
that the involvement of clients is an essential requirement for the zero injuries
objective. The importance of the client to H&S management is well documented.
Construction H&S can be successfully influenced by clients (Smallwood, 1998:182;
Bomel, 2001:5.3; Lingard et al., 2009:132).
Clients are usually at the centre of most decisions that are made on construction
projects; many of these decisions have an impact on the H&S outcome. It is because
of this that Suraji et al. (2001:340) for instance, contend that construction accidents
may be caused by client’s inappropriate responses to certain constraints and the
environment. Examples of such constraints could be client responses, which could be
actions or omissions in response to constraints that emerge during the development of
a project scope such as reducing the project budget, adding new project criteria,
changing project objectives and accelerating the design or construction efforts of the
project. Suraji et al. (2001:340) argue that all of these examples are factors that may
76
impact negatively on H&S depending on the decisions that clients may make. There is
therefore no argument that clients have a positive role to play in lowering injury rates
(Smallwood, 1998:181; Huang & Hinze, 2006a:164) and by extension, the possibility
of a change in the construction project H&S culture.
Successful implementation of H&S depends on the extent to which clients participate
in the process (Loosemore et al., 1999:884). Leadership in H&S must come from
clients. Without this, the construction industry has a long way to go in changing
attitudes and ultimately the H&S culture (Loosemore et al., 1999:884). There are
several activities which could for example show client participation. Levitt &
Samuelson (1993:215) observed that one of the activities through which the client can
make a difference is monitoring.
It is argued that clients must take responsibility to prevent accidents for example by
carefully considering H&S control in ordering works, exercising supervision, and
issuing instructions (Watanabe & Hanayasu, 1999:60). By so doing, many affirm that
clients can influence H&S on site (Smallwood, 1998:182; Suraji et al., 2001:339;
Huang & Hinze, 2006a:172). Consequently H&S culture within client organisations is
important because H&S culture has been identified to have an impact on H&S goals
(Dingsdag et al., 2006:2).
However, from evidence in several earlier studies it seems that most clients have not
shown serious commitment to H&S. A study conducted by Smallwood (1998) in
South Africa found that most clients give priority equally to cost and quality in
comparison to H&S being largely overlooked. This situation is a challenge to H&S
performance improvement.
A further challenge is the perception that H&S management is primarily the
responsibility of contractors, despite the emergence in recent times of legislative and
regulatory frameworks that redistribute responsibility for construction H&S to all
parties involved in the construction process. Interventions, such as H&S audits, are
usually designed only to find risks or hazards at the technical or operational level but
less concern at managerial level of the project organisation as a whole. Few strategies
are directed at improving upstream elements including those involving clients. H&S
77
campaigns are only made for operatives rather than for those who are involved during
the concept or design phases of a construction project (Suraji et al., 2006:51).
A study conducted by Alinaitwe (2008) revealed that the majority of clients do not
regularly attend the scheduled project meetings because, on average, they attended
about 40% of those scheduled. Lack of regular attendance at meetings implies that the
clients do not always keep track of the project developments. This might lead to
variations for which the client might not have a clear background. Public clients in
most parts of the world are even worse off in terms of performance. According to
Alinaitwe (2008:76), public clients are seldom good clients. There is usually poor
financial planning and management involved.
Architects are not sure whether clients provide adequate support to the contractors in
ensuring the H&S of workers (Alinaitwe, 2008:76). The author observes that this
situation could be the reason for many accidents in the building industry in Uganda.
Not surprisingly from this study, the majority of architects believed that the
productivity of workers was influenced by the way clients play their roles (Alinaitwe,
2008:76).
A review of literature confirms that examination of the role and culture of clients are
almost absent from most studies. Most studies focus on the construction phase of
projects and the related operational processes of contractors (Sawacha et al., 1999;
Hudson, 2001; Carder & Ragan, 2003; Saurin et al., 2003; Teo et al., 2005). The
exception however, is the study by Huang & Hinze (2006a) which focused on clients.
Clients, if involved, can influence worker health & safety (Smallwood, 1998:182;
Huang & Hinze, 2006a:172). Using total recordable injury rate (TRIR), to determine
the relationship between H&S Performance and owner involvement, Huang & Hinze
(2006a) demonstrated that clients can influence H&S outcomes. Smallwood
(1998:189) found that most general contractors believed that the client could
potentially influence their H&S performance.
According to Suraji et al. (2006:62), improving H&S means making clients, client
representatives, designers and contractors as well as employees be aware of their roles
in the improvement process. Huang & Hinze (2006a:164) further argue that
participation of clients is an essential requirement for the zero injuries objective. In
78
fact according to Gambatese (2000:668), owners should participate with contractors
in all project’s H&S activities. Clients could be involved in setting up work
procedures, raising awareness, requiring attendance at meetings and appointments
because these have been found to be effective in H&S performance improvement
(Smallwood 1998:189; Said, Shafiel & Omran, 2009:129). Clients could further show
commitment and involvement by conducting regular audits and inspections
(Smallwood 1998:189; Human Engineering, 2005:9). Huang & Hinze (2006a:172)
further elucidate that clients could show commitment, by providing adequate
resources for H&S instead of relying on contractors. This could include inter alia,
providing financial support, the inclusion of H&S as a prequalification criterion for
contractors, scheduling H&S requirements prior to bidding process structuring
contract documentation to allow for H&S, and selecting suitable contractors
(Smallwood 1998:182; Huang & Hinze, 2006a:165). Successful implementation of
H&S depends on the extent to which construction clients participate and assign
resources to the process.
The opportunity to improve H&S performance exists with the clients because they
bear the responsibility for example, in selecting design consultants, who are able to
offer designs that are not only safe to use, but also are capable of being built safely
(Bomel, 2001:8.18; Alinaitwe 2008:74). It also follows that since clients are the
initiators of projects, and without their decision to develop, projects would not be
implemented. Therefore it is only reasonable to deduce that risks should be equitably
apportioned between client and contractor (Tang et al., 2008:457). In addition, focus
on H&S performance should also be directed at construction clients.
H&S performance improvement depends on the extent to which clients provide
leadership on H&S matters. Loosemore, Lingard, Walker, & Mackenzie (1999:884)
argue that the lead must come from clients themselves on H&S. They maintain that
without client’s leadership, the construction industry has a long way to go in changing
attitudes and the H&S culture. According to Levitt & Samuelson (1993:215) simple
monitoring exercises can show leadership, as it makes a difference, and that excellent
H&S performance can be obtained with the active participation of clients, even from
average contractors. It is argued that clients have the moral if not the legal duty to
take reasonable care to ensure H&S to all workers on construction sites (Suraji et al.,
79
2006:55). Client leadership can be achieved by the client considering carefully H&S
control in ordering works, exercising supervision, and providing instructions.
According to Huang & Hinze (2006b:181), clients set the H&S culture tone for a
project.
4.2
ROLE OF CLIENTS IN DESIGNING FOR H&S
Designing for H&S can help improve H&S performance in the construction industry
(Gambatese, Behm & Hinze, 2005:1035; Hecker, Gambatese and Weinstein,
2005:35). Designing for H&S entails consideration of H&S in the design of structures
or appraising designs in terms of H&S and then action plans are developed to ensure
that risks are engineered out of the system before they are able to cause injury,
disease, damage, or even loss of life on site. Behm (2005:590) defines design for
construction H&S as being the consideration of site H&S in the design of a project.
Therefore according to Behm (2005:590) designing for H&S includes:

modifications to the permanent features of the construction project in such a
way that construction site H&S is considered;

attention during preparation of plans and specifications for construction in
such a way that construction site H&S is considered;

the utilisation of specific H&S suggestions in the design for construction; and

the communication of risks regarding the design in relation to the site and the
work to be performed (Behm, 2005:590).
Similarly, Hecker et al. (2005:32), consider design for H&S as interventions to
eliminate hazards before they appear on the jobsite. The foci of designing for
construction H&S efforts are typically the incorporation of construction knowledge in
the design effort and consideration of H&S early on and throughout the project. Hinze
et al. (1999:400), advocate for a holistic approach of designing for the entire life cycle
of a project, including the construction process. Effectively addressing construction
H&S issues means that the designer must consciously assess the implications of each
construction phase on H&S as the facility is being built. It means therefore that a
thorough risk assessment of each design component should be done (Hinze et al.,
1999:420).
80
Design should include the design specification and requirements at one end and the
instructions and procedures for use at the other. Safe design therefore means a design
that allows and conditions, as far as feasible, healthy and safe use across the whole
life cycle including demolition and disposal. However, this conclusion also has
implications for the definition of “design errors’. There should not be therefore talk of
“design errors” but rather of errors in a specified step in the design process. Design
for H&S therefore calls for an extensive knowledge on H&S as opposed to a mere
general awareness of the subject (Hale et al., 2007:315).
The probable reason design is touted as one of the solutions to the problem of H&S is
that design is driven largely by a logical conclusion that systems development begin
with design, and therefore design offers the earliest, and hopefully, the cheapest place
to intervene and ensure it is done correctly (Hecker et al., 2005:32; Hale et al.,
2007:308; Frijters & Swuste, 2008:280). This statement makes sense because in order
for a problem to be solved, the root cause should be addressed first.
However designers have to be compelled or motivated to design for H&S. According
to Hale et al. (2007:309), the factors that compel designers to consider H&S in their
designs include:

ethical considerations and concern for the organisation’s reputation;

liability claims resulting from damage and injury; and
at a legal level an increasing emphasis is placed on the liability of the designer for
incorrect design decisions. However, Hale et al. (2007:310) contend that this
liability is limited in most cases, whether under strict liability or tort law systems,
to what the designer has control of and can reasonably be expected to do.
Apart from these factors, some designers lack motivation to consider H&S in their
designs. According to Gambatese et al. (2005:1030), some of the factors that have
contributed to designers’ lack of motivation to design for H&S and therefore a barrier
to H&S improvement include inter alia:

weak or absent regulatory requirements for designers to design for the H&S of
construction workers;

placement of responsibility on the contractor;

liability concerns among architects and engineers;

narrow specialisation of construction and design;
81

limited availability of safety-in-design tools, guidelines and procedures;

limited pre-construction collaboration between the designer and the contractor
due to the traditional contracting structure of the construction industry;

the limited education architects and engineers receive on issues of construction
worker H&S and how to design for H&S;

the structure of the construction contracting process;

a lack of knowledge and acceptance of the concept;

competing project objectives; and

motivation to implement the concept (Gambatese et al., 2005:1030).
Limited education on H&S is a challenge in designing for H&S. There is little
knowledge by designers concerning problems such as on how the operation or
construction will be undertaken (Kirwan, 2007:161). Furthermore, Kirwan (2007:161)
argue that there is often little detail, if any, on the procedures to be followed when
implementing the design. This, according to him, amounts to a lack of a mature
operational concept, one that is sufficiently detailed to allow H&S hypotheses (e.g.
what would happen if….?) to be answered (other than - well it depends how we
operate or implement it’). Kirwan (2007:161) argue that this problem, coupled with
the requirement of H&S assessment of new concepts, requires incorporating expert
judgements where data is not available or not representative. There is need therefore
for designers to be adequately equipped in H&S.
Apart from a lack of motivation and qualification on the part of designers, a further
problem or barrier to achieving a total design for H&S is the nature of design being a
distributed process especially in the construction industry. The nature of design as a
distributed process raises the same sort of concerns as the division of labour that
characterised the Taylorian approach to production and assembly line manufacture (cf
Peaucelle, 2000). The Taylorian approach led to problems because no individual
participant in the process has the overview of, or the sense of ownership for, the
product being made (Suarez-Barazza, Ramis-Pujol & Heras, 2010:77). Such
Taylorian production lines only work when there is a strong central planning and
control function, which ensures this overview and the necessary communication and
optimisation. The same lack of ownership of the total design and the problems of
interfaces between the different actors can be seen in the design process in the
82
construction industry and therefore pose a problem or barrier to H&S Improvement
(Hale et al., 2007:312).
Notwithstanding the identified barriers, designers can have an impact and
significantly reduce the number of injuries and fatalities by considering construction
H&S in their designs (Weinstein et al., 2005:1126). One of the benefits of an early
involvement of designers is that it will lead designers to think about H&S from the
start rather than thinking that it is something that comes later and does not form part
of their job or concern (Kirwan, 2007:162). Other residual benefits which are very
important and have a lasting impact and influence on H&S are the new cultures that
are created. A H&S culture can be enhanced by early consideration of H&S in the
design process. Not only do the designers become more exposed to H&S and its
mission and practices, but other stakeholders from the project managers to contractors
taking part in early simulations, realise that H&S is being addressed in a useful way
and therefore reinforcing its importance for all concerned and its continual presence
throughout the entire project life cycle (Kirwan, 2007:162).
The concept of designing for H&S has benefits, and these are well documented, and
probably the reason it is also advocated for. A number of studies have indicated that
design can influence H&S on construction sites. They include

an investigation across aviation and nuclear industries by Kinnersley & Roelen
(2007). They found that 50% of all accidents had their root in design;

Gibb et al. (2004) who observe that in almost 50% of accident cases, design is
involved;

a study by Behm (2006) who found that design was linked to accidents in
approximately 34% of the 450 accidents that he analysed.;

Hecker et al. (2004) study which identified elements in design, planning,
scheduling, and material specifications as probable contributors to working
conditions that pose risks to musculoskeletal injuries during the actual
construction process; and

an examination by Loughborough & UMIST (2003) of 100 accidents. They
found that up to half of the accidents could have been mitigated through a
design change.
83
Therefore addressing H&S at the design stage may offer an opportunity to improve
H&S performance in the construction industry. It cannot be denied that consideration
of H&S in design of facilities is a very significant step in working towards H&S
performance improvement. Kirwan (2007:162) argue that since accidents often have
their roots in design, the sooner H&S is introduced the better. In particular, hazard
causes identified early on may become more difficult to find or correct later, with the
risk that they become latent errors in the system design. As a result of this realisation,
building designers in Europe, now have a legal obligation to take working conditions
throughout the project into account in their designs. The obligation contained in
Directive 92/57/EEC is now incorporated in most European Union Countries’ laws.
However, in order for H&S design to be effectively used as an intervention, albeit not
on its own, an additional need exists to address procurement systems and this can only
be successfully done with the client’s involvement (Hecker, Gambatese & Weinstein,
2005:43). This means that different types of projects require different approaches in
design. The type of project delivery method can impact the extent to which H&S is
addressed in the design. The forms of project delivery essentially alter the roles
played by the different parties, and most importantly, the allocation of responsibility
(therefore liability) is also redistributed. For example, the traditional design-bid-build
approach and others of a similar nature keep the parties apart and there is presumably
no payback for the designer to address construction worker H&S. In this way, the
designer is a stand-alone entity, and as an isolated entity, designers often revert to
their traditional role of not getting involved in addressing H&S (Gambatese et al.,
2005:1031).
Where there is no existing organisation with a powerful central role in managing the
parallel design processes, there is a task for government or other such clients in
bringing together the players in the design process to define and coordinate their roles.
This view is endorsed by Hale et al. (2007:313), who argue that for designs within the
diverse systems with many uncoordinated players, the issues of responsibility for
predicting risks and making choices to control them is very important and is only
sometimes identified. However, according to Hale et al., (2007:313), the allocation of
responsibilities and above all the possibility of checking and enforcing that those
responsibilities are carried out, is almost non-existent. The best practice for coping
84
with the fragmented design process by different professionals is bound to differ across
different systems, but there should be more explicit attention to this question in the
sectors with less developed design processes. The solution may however lie with the
client taking a lead role. Gambatese et al. (2005:1031) argue that the client is the key
to getting the designer involved in the H&S process because the client can influence
the way the project is procured, and address specific issues regarding H&S in the bid
documents, including organisation for the project. It is also clients who can ensure
that opportunities are created for designer-contractor interaction in the course of
specific projects as this has been identified as critical to designing for H&S (Cosman,
2004:59).
Furthermore, it is the client that can convince designers of the role that they have to
play. The client is vital to getting the designer involved in the H&S process because
the client can alter the way in which the project is procured and address specific
issues regarding H&S in the contract as well as the coordination (Gambatese et al.,
2005:1031). According to Huang & Hinze (2006b:181), clients set the H&S culture
tone for a project even by emphasising designing for H&S.
4.3
CHAPTER SUMMARY
Chapter four discussed the role of the client on H&S performance and particularly on
their influence on design. H&S performance improvement depends on the extent to
which clients provide leadership on H&S matters. Without client’s leadership, the
construction industry has a long way to go in changing attitudes and the H&S culture.
Simple monitoring exercises can show leadership and excellent H&S performance can
be obtained even from average contractors with the active participation of clients.
Furthermore, literature informs that it is the client that can convince designers of the
role that they have to play. The client is vital to getting the designer involved in the
H&S process because the client can alter the way in which the project is procured and
address specific issues regarding H&S in the contract as well as the coordination.
Clients set the H&S culture tone for a project
85
CHAPTER FIVE
RESEARCH METHODS
5.1
INTRODUCTION
A number of studies exist on H&S performance improvement in the construction
industry at an international level but little has been done at the local level in the
Southern African region. Studies have been conducted on the awareness and
implementation of H&S in Botswana (Musonda & Smallwood, 2008), impact of
construction regulations in South Africa (Smallwood & Haupt, 2007), designers’
perceptions on H&S implementation in South Africa (Smallwood, 1998) and a review
of the status of H&S in South Africa (CIDB, 2008). There is little if any that has been
written on the improvement of H&S in the construction industry in Southern Africa
especially looking at the role of construction clients and how they can contribute to
H&S improvement.
The above scenario, coupled with the fact that the construction industry does not have
an enviable record on H&S performance (Dias, 2004:1; Karjalainen, 2004:3; Ringen
& Englund, 2006:388), motivated the commencement of this particular study. This
study therefore set out to achieve the listed general objectives below. These were to:
GO1. establish from literature, the status of H&S in the construction industry, the role
of clients, designers and the top management of contractors in H&S
performance;
GO2. examine current trends in H&S performance improvement;
GO3. establish the role and impact of construction clients and their potential
contribution to and influence on project H&S performance;
GO4. develop a client-centred model for H&S performance improvement in the
construction industry; and
GO5. validate the conceptualised client-centred model for H&S performance
improvement by comparing the Delphi and literature review outcome with the
field questionnaire survey outcome.
86
5.2
RESEARCH DESIGN
In order to meet the general research objectives, the following strategies were
adopted:
General objective G01
For the first general objective, which was to establish the status of H&S in the
construction industry, a review of literature was conducted on the performance of the
construction industry in terms of H&S. Published articles, accidents and incidents
reports and status reports were reviewed. Both international and local Southern
African status reports and literature were reviewed. The expected outcome from this
objective was information as well as a depiction of H&S performance in the
construction industry. The information and a general picture were useful for the reader
to have an understanding of how the construction industry is performing in terms of
H&S and the extent of the problem.
General objective G02
The second general objective of the research was to establish the current trends in
H&S performance improvement in the construction and manufacturing based
industries. This objective entailed that a rigorous and exhaustive review of literature
had to be conducted. This review was conducted from a wide source of publications
including journals, conference proceedings, books and monographs obtained from the
library and the internet.
Current topical research is the best source for current ideas on solutions that are being
advanced for problems that face mankind and therefore it was decided that in order to
achieve the objective, a review of literature on the subject of H&S performance
improvement should be done. The expected output from this second general objective
was information on the current trends and theories on H&S improvement, especially
to determine which areas most studies focus on, what common themes are apparent
and the type of methodologies that were used in the research. This information was
necessary because it formed the core literature on which the current research project
was based.
87
General objective G03
Literature was reviewed to achieve the third general objective, which was to establish
how construction clients and other stakeholders could influence H&S performance.
Literature was reviewed to identify key responsibilities of various stakeholders but
particularly those of construction clients and an attempt was made to establish their
role in H&S improvement. The output from this objective was an outline of factors
associated with clients that may be critical to H&S performance.
General objective G04
The Delphi method was used to achieve the fourth objective which was to determine
the impact of client influence on the overall project H&S performance and the extent
to which H&S could be improved as a result of client and other stakeholders’
influence. The Delphi method was the best method to use in this instance as the
objective entailed soliciting expert opinions on what would happen to H&S
performance if there was an influence from various stakeholders particularly from the
construction clients. This type of question can only be addressed by methods such as
the Delphi and or focus groups apart from experimental procedure which was not
feasible for this study.
Apart from the Delphi approach, focus groups could have been used except that there
was a challenge of bringing international experts to one destination and making them
deliberate for a minimum duration of eight hours per day for at least two days. Not
only was the focus group method not feasible, it was also considered costly beyond
what was budgeted for and it would have defeated the purpose of conducting a
rigorous process to achieve the objective. In a Delphi method, bias is eliminated by
members of the expert panel remaining completely anonymous to each other and
therefore there is no undue influence from other peers. This is not the case when using
the focus group method.
The expected output was an estimation of the extent to which project H&S
performance could be improved if influenced by construction clients as well as an
identification of factors that are critical to H&S performance. From these factors and
interrelationships, a conceptual model, which was centred on the client, was
developed for H&S improvement.
88
General objective G05
A questionnaire survey was conducted and the collected data was analysed using
structural equation modelling in order to achieve the fifth general objective of the
research which was to test and validate the conceptual model. Data obtained from the
questionnaire sought to establish the influence of clients on contractor, designer and
project H&S performance. In addition, the influence of the external environment
factors on client H&S performance was investigated.
The method was considered to be suitable for the type of information that was being
collected as the basic aim was to establish H&S practice in the construction industry.
A detailed explanation of the survey concerning population, sampling procedure and
analysis of results is presented in the methods section. The expected output from the
fifth objective was information on the goodness-of-fit of the conceptual model to the
sample data. This information was then used to validate the postulated model and to
finalise the best fit model for H&S performance improvement.
5.3
METHODS
Table 7 summarises the research methods that were employed to achieve the general
objectives of the research. Figure 8 is an outline of how the entire study was
conducted. The first stage involved establishing the need for the study, conducting a
review of literature and theorising on client influence on H&S performance. The
second stage involved conducting a Delphi study, developing a conceptual model and
evolving a questionnaire for the survey at the third stage. The third and final stage
involved conducting a survey, analysing and modelling the results, determining the
best fit model and comparing the conceptual model to the sample data from the
questionnaire survey.
89
Table 7: Research procedure followed in current study
Stage
General
Objective
Data collection
Method
Data analysis method
Establish status of H&S in

construction industry
Establish current theories and 
literature on H&S
improvement
Establish clients’ role and

influence on H&S performance
Literature review

Literature synthesis
Literature review

Literature synthesis
Literature review

Literature synthesis
2.0
Delphi method
Determine impact of client

influence on H&S performance
Delphi technique


Desk study

3.0
Questionnaire
survey
Develop a Client-centred
conceptual model for H&S
performance improvement
Test and validate conceptual
model for H&S performance
improvement

Questionnaire
survey

1.0
Literature review
90
Output
Status of H&S in
construction industry
Theories and literature on
H&S improvement
Client Factors associated
with good H&S
performance
Descriptive statistics Consensus on Impact level
of client influence on H&S
performance
Client-centred model
Theory
Structural equation
modelling (SEM)
o EQS
o MPlus
Information to validate
conceptual modelGoodness-of-fit test results
Stage 1
Stage 2
Stage 3
Questionnaire
Develop conceptual
model based on
clients’ influence
Evaluate Clients’
influence on H&S
performance
Delphi
Analyse & model
results to validate
conceptual model
Literature review
Best fit model &
recommendations
Need for study
Figure 8: Research design outline
(Adapted from Manu, Ankrah, Proverbs & Suresh, 2010:29)
5.3.1
Literature review
Literature is the foundation of research (Boote & Beile, 2005:3). Review of literature
is one of the most important aspects of developing a study and also as a way to know
what has already been written on the subject, methodologies that have been used to
investigate similar concepts or phenomena and to establish the trends on the solutions
that are being advanced to solve the many problems that face mankind (Heppner &
Heppner, 2004:52).
It was therefore necessary to conduct a literature review in order to establish the

general H&S status in the construction industry from research work of others;

theories and literature on H&S performance improvement ; and

client factors associated with good H&S performance in the industry.
In order to guarantee integrity and sophistication of the study, effort was made to
ensure that the review was thorough and exhaustive. Studies reviewed were
synthesised and they considered methods adopted or used in other studies. A critical
review of methods adopted in other studies also ensured that the study did not only
report the claims made in the existing literature because this is one of the remedies
against the trap of simply reporting other people’s claims (Boote & Beile, 2005:3).
91
The resources used for literature review included books, reviews of articles on the
subject both published and unpublished, such as dissertations, people and search
engines. Further, names of key leading contributors on the subject where drawn and
searched to establish their publication history and then conducted focused searches
within these searches. Furthermore, Harzing’s publish or perish software was used to
identify articles on the subject that were frequently cited.
Articles from these sources were read, re-read and analysed in order to establish the
progression of research in the area, specifically on the topic under study. The process
of conducting literature review as recommended by Boote & Beile (2005:13)
specifically involved, inter alia

finding a broad range of high-quality, specific articles, books, dissertations
and reviews directly related to the study;

reading and re-reading to establish progressions and trends;

summarising of the studies read;

identifying methodologies adopted in the studies;

relating the current study to those reviewed; and

writing the literature.
The output from the literature review was a clear perspective of the topic and an
indication of where the study fits in relation to other studies on the subject as well as
provided a framework for comparing the results of the study with others.
5.3.2
Delphi method
Suitability of the method
The Delphi method was used for the second stage of the study to explore the impact
of client H&S culture on contractor, designer and project H&S performance. In
addition, the influence of the external environment on client H&S performance was
investigated. The Delphi method was first developed by the Rand Corporation (Holey
et al., 2007:1). The Delphi method is a structured process requiring experts to respond
to non-leading unambiguous statements with the aim of achieving consensus (Holey
et al., 2007:1). The Delphi method was preferred to common survey methods because
the current study was addressing the ‘what can-if’ kind of questions as opposed to the
92
‘what is’ kind of questions. Delphi is more suited for these kinds of questions in order
to explore concepts that are difficult to measure except through experimental
methods. Unfortunately, an experimental survey was not feasible and appropriate for
the current study.
The Delphi method was also considered to be a robust method for a rigorous query of
experts. Unlike ordinary survey research, the Delphi’s strength also lies in the rounds
used which provide an opportunity for initial feedback, collation of feedback, and
distribution of collated feedback to participations for further review. This unique
process requiring group communication is central to the strength of the Delphi
(Gohdes & Crews, 2004:56).
Selection of Delphi panel members
In this study, the Delphi study involved and retained 11 active panel members. This
number of panellists was considered adequate based on literature recommendations
and experience from what other Delphi studies have used. Skulmoski, Krahn &
Hartman (2007:10) suggest that 10 to 15 panellists could be sufficient if the
background of the panellists was homogenous. A review by Rowe & Wright (1999)
indicates that the size of a Delphi panel has ranged from three to 80 in peer reviewed
studies. According to Okoli & Pawlowski (2004:18) and Skulmoski et al., (2007:10) a
panel size of about 10 to 18 members is adequate. Hallowell & Gambatese (2010:103)
suggest that since most studies incorporate between eight and 16 panellists, a
minimum of eight should suffice. The size of the panel should be dictated by
characteristics of the study such as the number of available experts, the desired
geographical representation and the capacity of the facilitator. Based on this
discussion, and the fact that the Delphi method does not depend on the statistical
power, but rather on group dynamics for arriving at consensus among experts, the
panel of 11 members was considered to be adequate.
The selection of panellists was based on criterion sampling. Panellists were
purposefully selected to apply their knowledge to a concept raised in the study based
on the criteria that was developed from the research questions under investigation. A
Delphi study does not depend on a statistical sample that attempts to be representative
of any population. It is a group-decision mechanism requiring qualified experts who
93
have a deep understanding of the issues (Okoli et al., 2004:20). Therefore, one of the
most critical requirements is the selection of qualified experts because it is the most
important step in the entire Delphi process. It directly relates to the quality of the
results generated. Therefore, careful selection of the panel of experts is the keystone
to a successful Delphi study (Gohdes & Crews, 2004:60; Hsu & Sandford, 2007:3).
In view of the above, successful panel members had to meet the following criteria
adopted from Skulmoski et al. (2007:10):

knowledge and experience in construction H&S;

knowledge and experience in construction project management;

appropriate academic qualifications;

professional registration with a recognised built environment or H&S
registration body;

published articles in peer reviewed journals, books and or conferences;

industry experience of at least five years;

capacity and willingness to participate;

sufficient time to participate; and

effective communication skills.
Delphi panel
Panel members were identified from three sources. The first source was the CIB
W099 register of members located on the CIB W099 website (CIB, 2010: Online).
The CIB W099 is a working commission that was set up on royal appointment to
enable researchers on construction H&S in the world to collaborate as well as to
protect H&S. The second source was the conference proceedings of the CIB W099
from year 2005 to 2009. Individuals who had frequently appeared as authors or
keynote speakers in these proceedings were identified as potential experts on the
study. CIB W099 conferences are one of the avenues (if not the only avenue) where
leading researchers in construction H&S present their research findings and also
present as keynote speakers. The third and final source was the references of
individuals who had committed their lives working in the area of H&S in the
construction industry locally in Southern Africa.
94
From these sources, 48 invitations were sent to the identified potential panel
members. Out of 48 invitations, 33 potential panel members responded to the
invitation, 15 completed the first round and 11 panel members were retained
throughout the study. Panellists were judged to be qualified as experts and included in
the study based on their curriculum vitae that they were requested to submit in
response to the initial invitation. A Delphi method is a very rigorous and time
consuming process and this could have probably led to most of the potential members
falling out at the introductory stage when they learnt of their obligations.
The panel consisted of two members from South Africa, three from the United States
of America (USA), three from the United Kingdom (UK), one each from Singapore,
Hong Kong, and Sweden. The continents that were not represented on the panel are
South America, Australia and the Middle East. See Table 8.
Table 8: Delphi panel composition by Regions
Region
Number of experts
Africa
2
Asia
2
Europe
4
North America
3
Total
11
The panellists’ highest qualifications are as tabulated in Table 9. One panel member
had a Doctor of Science (DSc) degree, six had a Doctor of Philosophy (PhD) degrees,
two panellists had Master of Science (MSc) degrees, one had a bachelor of Science
(BSc.) degree and the last panellist had a Diploma in H&S management. All panellists
specialized in construction H&S.
In terms of the panellists’ current occupation, three of the panellists were employed
by contracting organisations, two by consulting organisations, and six by Universities.
All panellists held very senior positions in their organisations and were involved in
community service. The panel consisted of six professors, a dean at the University, a
contracts manager in a multi-national contracting organisation, a private consultant to
major international organisations including the International Labour Organisation
95
(ILO) on human resources and safety, a director in a contracting organisation and a
H&S coordinator and inspector in a construction companies’ registration body.
The panel had a cumulative 243 years of experience. The minimum being seven and
the maximum was 45 years. The mode for number of years of experience was 15, the
mean was 22.1 years and the median was 15 years. Experience was an important
factor in determining who an expert was and therefore a minimum number of years
was set to be five years. As can be seen from Table 10, two panellists had 6-10 years
of experience, four had 11-20 years of experience, two had 21-30 years of experience
and finally three panellists had above 31 years of experience.
All panellists were professionally registered at the highest level with profession
regulating bodies such as the South African Council for Construction Project
Managers, Institute of Civil Engineers, Chartered Institute of Building, American
society of Civil Engineers and the Royal Institute of Quantity Surveyors. Two panel
members were registered as fellows, four were chartered or professional engineers,
two were chartered builders, one was a registered construction project manager and
the last one was a registered H&S practitioner.
Table 9: Delphi panel composition by highest qualifications
Highest qualification
Number of experts
Doctor of Science ( DSc)
1
Doctor of Philosophy (PhD)
6
Master of Science degree (MSc)
2
Bachelor of Science degree (BSc)
1
Diploma (Dip.)
1
Total
11
96
Table 10: Delphi panel composition by number of years of experience
Years of experience
Number of experts
1- 5
0
6-10
2
11-20
4
21-30
2
Over 31 years
3
Mean
22.1
Mode
15
Cumulative years of experience
243
In terms of publications, 10 of the panellists had published in peer reviewed journals,
conferences and books. Between them, they had published 57 books and monographs,
19 chapters in books, 187 peer reviewed academic journals, 345 recent conference
papers and 341 other publications comprising of articles in professional journals,
technical reports, policy papers, expert witness documentation and key note
addresses. Between them, they had led and managed 108 funded research projects.
Three panellists served on editorial boards of 43 peer reviewed journals and
conference proceedings (Table 11). Figure 9 shows the contribution of panellists to
the above mentioned publications. Five of the panellists had published books and
monographs, three had published chapters in books, seven had published articles in
peer reviewed journals, ten panellists had published articles in peer reviewed
conference proceedings, seven of them had led and managed funded research projects
and all of the panellists had published an article in other publications, such as
professional journals and technical reports. Three of the panellists served on editorial
boards for journals, six had served as a referee for conference proceedings and three
had been appointed as referee or reviewer for journal publications.
Specific objectives of the Delphi
The external environment and the client H&S culture have a certain level of influence
on project H&S performance. Literature informs that improving H&S culture would
lead to improvement in H&S performance. What was however not clear in the
literature was specifically the level or extent of impact of client H&S culture on
project H&S performance. An attempt was therefore made to infer client’s influence
97
by relating total recordable injury rates (TRIR) to client involvement (Huang &
Hinze, 2006a). However, it was noted that accidents, incidents, injury or even fatality
records are not a reliable and certainly not the only measure of H&S performance.
These are lagging indicators and the TRIR is also a lagging indicator. The H&S
culture of an organisation and other leading indicators are now the accepted measure
of H&S performance.
Table 11: Number of publications by Delphi panel members
Panel publications
No. of publications
Books and monographs
57
Chapters in books
19
Peer reviewed Journals
187
Peer reviewed Conference proceedings
345
Funded research
108
Other publications
341
Editorial board membership
43
Referee for journals
22
Referee for Conference proceedings
30
Other publications
Editorial board member
Publications
Referee for Conference proceedings
Referee for journals
Funded research
Contributors
Conference proceedings
Journals
Chapters in books
Books/monographs
0
2
4
6
8
Number of contributors
Figure 9: Delphi panel members’ contribution to publications
98
10
12
An alternative measure of the impact of client influence was therefore necessary in
order to establish not only whether clients do have influence on H&S performance but
also the extent or level of this influence. Such a study would ordinarily call for an
experimental kind of research. However, this kind of research was not feasible and
practical considering the time frame, ethical issues and the willingness of would be
participants.
The Delphi method was therefore considered the most suitable method to achieve the
general objective of determining the impact of client’s influence on project H&S
performance. Therefore, the specific objectives of the Delphi study entailed
establishing the impact of the:
D1. external environment factors on client H&S performance;
D2. client H&S culture on H&S consideration (hazard identification, risk analysis,
assessment and mitigation) throughout the project life cycle;
D3. client H&S culture on contractor H&S performance (top management);
D4. client H&S culture on designer H&S performance; and
D5. client H&S culture on the overall project H&S performance.
Achieving the above objectives, resulted in the following outcomes:
1. factors of client H&S culture, contractor and designer H&S performance, and
external environment that were of critical significance to H&S performance
improvement; and
2. a conceptual client-centred model on H&S performance improvement.
Data collection through Delphi
The Delphi method involved three rounds of an iterative process with the view of
achieving consensus
between the
panel
members
regarding the external
environment’s impact on client H&S performance/culture and what the client H&S
culture’s impact was on contractor, designer and project H&S performance.
A Delphi questionnaire, attached as Appendix C, was sent out electronically to all
panel members who were then asked to take time and respond to the questions
according to their ability and expertise. The Delphi questionnaire was developed
99
based on the findings from the literature review and was specifically designed to
address the five Delphi objectives D1 to D5.
Regarding the questionnaires, panel experts were specifically requested to rate the
probability that the contractor or designer would implement a range of H&S elements
as a result of client H&S culture’s influence. The probability scale ranged from 1 to
10, representing 0 to 100% on an ordinal scale. Furthermore, experts were requested
to rate how severe the impact on H&S performance for instance would be if a
particular factor of client H&S culture was absent or not apparent. The impact scale
was based on a 10 point rating scale, ranging from 0 (negligible or low impact) to 10
(very high impact). This aspect indicated the severity of the factor of client H&S
culture.
The critical function of the Delphi method was for panel members to reach consensus
on the probability rating that an aspect of H&S would be implemented as well as the
impact level of a factor of client H&S culture on contractor, designer and project
H&S performance. The ultimate goal therefore was for panellists to reach consensus
on all issues.
Responses were received for each round of the Delphi. In order to establish consensus
on each and every question, group medians were calculated for each response for each
element as shown on the questionnaire attached as Appendix D. The group mean,
median and mode can all be used as measures of central tendency in Delphi analysis
(Hsu & Sandford, 2007:4). In the current study the group median was the chosen
measure of central tendency because it indicated a general inclination of the group on
a particular question. The median was more appropriate and suitable for the type of
information that was being collected. The median eliminates biasness and takes into
consideration outlier responses. It makes the consensus notion more reasonable. Other
measures of central tendency, such as the mean for example, may not reflect a
reasonable central tendency because it does not consider frequencies.
Upon receipt of responses from the first round, group medians were computed for
each question. In the second round, the same questionnaire was sent back to panellists
individually with their own responses from the first round with the group median
responses included so that responses in the second round could be made taking into
100
account the group median. In the second round expert panel members were asked to
either maintain their original responses made in the first round, or they could change
their initial response to either be in agreement with the group median or make a new
rating altogether. The panellists, who had ratings of two units either above or below
the group median on any one particular question, were requested to state the reasons
for their dissenting opinion if at all they opted to stick to their rating. The stated
reasons were sent to all panellists so that together with the calculated group medians
in the second round, panellists could take cognisance of those comments in making
their new ratings in the third round. Panellists were specifically requested to consider
reasons from the outliers made in the second round in making their decisions in the
third round.
After the second round, group medians and comments made by those with dissenting
views were sent to panellists together with each panellist rating made in the second
round for a third round of responses. Calculated group medians and comments were
sent to all panellists and individual ratings were sent only to those that had made them
for anonymity.
After the third round, group medians and the absolute deviations were again
computed for the third round. Calculations for the third round of the Delphi process
indicated that there was no need to proceed to the fourth round as there was no further
value that could be added to the degree of consensus that was attained at that level.
Throughout the Delphi process, anonymity of panel members was maintained to avoid
undue influence on other members. The aspect of anonymity was crucial to the
credibility of the Delphi process. Figure 10 outlines the Delphi process.
Analysis of data from Delphi
A two-stage analysis of data from the Delphi study was conducted using Microsoft
Office Excel, a spread sheet software program. The first stage involved an analysis to
establish or confirm consensus on responses to the predetermined criteria. This
involved determining the group median responses for each question. After the third
round of the Delphi, absolute deviations ( Di ) about the group medians ( mX  ) of
each rating for every question were calculated using Equation 1. In addition to this,
average (mean) absolute deviations (MAD) were calculated for every question. The
101
mean absolute deviation is a calculated mean of all absolute deviations for all
panellists about the median on each question.
Appendix G shows a table of calculated MAD’s for each of the Delphi rounds.
Further analysis involved determining the statistical range in ratings by panellists on
each question and the percentage of panellists with a similar opinion inclination on
each and every question. Consensus was determined to have been achieved when the
MAD was less than one unit below or above the group median, the range in ratings on
each question between all panellists was below 4.0 and the percentage of panellists
that were of a similar inclination in opinion was 60% and above on a particular
question (Appendix G).
Researcher actions
Expert panel actions
Round 1
Circulates questionnaire to
panellists
Rates likely improvement and impact on
H&S due to a factor influence
Round 2
Calculates group medians and recirculate questionnaires
Reviews individual rating in view of the
group’s median. Gives reason if required
Re-Calculates group medians,
standard deviations & compile
comments
Opportunity to reconsider rating
Round 3
Determine consensus and
terminate process
Figure 10: Delphi process
(Adapted from Thangaratinam & Redman (2005:124)
102
The second stage of Delphi data analysis, involved determining the impact
significance of environmental factors on client H&S performance, the impact
significance of factors of client H&S culture on contractor and designer H&S
performance, H&S consideration throughout a construction project’s lifecycle and the
overall project H&S performance.
Equation 1
Where:
Di
 Absolute deviation
xi
 Panellist rating
m X   Measure of central tendency
The impact significance of designer and contractor H&S performance on project H&S
performance was also investigated from the Delphi data. The significance of the
impact of various factors associated with the external environment, client, designer
and contractor was categorised as being critical, major, moderate, minor or low. The
categorisation was helpful in determining which stakeholder factor and relationship
was key and relevant for H&S performance improvement.
This method of analysis has been used in many studies. The impact significance of a
factor of client H&S culture for instance, was obtained as a product of the rated
probability (likelihood) that this factor of client H&S culture would influence
contractor, or designer to implement H&S elements. The rated negative impact
(severity) of a factor for instance of client H&S culture, was the rating of the negative
impact that would result on contractor and or designer H&S performance if this factor
was absent or was not apparent. This relationship is illustrated in Equation 2.
As an example, Table 12 will be used to illustrate how the impact significance was
determined. Column B lists a number of H&S elements that need to be implemented
by the client. Column C to H lists experts’ probability rating that had been agreed
upon to say that elements numbered 1 to 11 would be implemented if the client was
pressured by the external environment factors namely: political, social, economic,
103
technology, legislative and professional bodies (labelled as: PE, SE, TECH, EE, LE,
and PB). Column I list the total score rating for each element numbered 1 to 11that it
would be implemented if the client was pressured by external environment factors in
column C to H. Each element in this example had a maximum of 60 points (6 factors
rated on a 10 point ordinal scale). Column J, list the calculated overall probability that
each client element numbered 1 to 11 would be implemented if the client was
pressured by all the factors. Each factor’s likelihood to influence clients to implement
all the H&S elements for example, was calculated by obtaining the total ratings
achieved by the factor and calculated it as a percentage of the maximum available
ratings which in this case was 110. The impact significance of a factor for instance
PE, was obtained as a product of the likelihood that client would implement H&S
elements which in the example was calculated as 63%, and its severity determined to
be 8. The severity rating was taken to be the rated negative impact on client
implementing the elements 1 to 11 if the factor for example PE was absent.
Equation 2
A prediction of the likelihood that H&S elements would be implemented if pressured
by the identified factors, was also made concerning clients’, designers’ and
contractors’ top management. Predictions were made about the influence of
environmental factors on client H&S performance, and also about client H&S
culture’s influence on contractor and designer H&S performance, H&S consideration
throughout the project lifecycle as well as on project H&S performance.
The impact significance of external environment factors, and factors of client H&S
culture, contractor and designer H&S performance were also investigated. These
values were entered as shown in row 15 of Table 12.
The findings from the Delphi analysis were presented as numbers and percentages in
tables, column and bar charts. These were expert predictions of the likelihood that
H&S elements would be implemented by clients, contractors and designers.
Predictions were also made about the severity and therefore the impact significance of
the evaluated factors.
104
The impact significance was categorised as ‘critical’, ‘major’, ‘moderate’, ‘minor’
and ‘low or negligible’ depending on the calculated value as shown in Table 13. In the
same way, the likelihood of an element being implemented, was described as being
‘very likely to occur’, ‘likely to occur’, ‘may occur half the time’, ‘unlikely’ or ‘very
unlikely’, depending on the rating by the expert. See Table 14.
Establishing consensus from the Delphi process
It was important that consensus was reached on all questions. Measuring or
determining consensus is a highly contended subject in literature. It should in fact be
mentioned that there is no agreement in literature on how consensus can be said to
have been attained regarding a set of opinions (Hsu & Sandford, 2007:4). Holey et al.
(2007:2) suggest that consensus is the same as agreement and that agreement can be
determined by the following:

the aggregate of judgements;

a move to a subjective level of central tendency; or

alternatively by confirming stability in responses with the consistency of
answers between successive rounds of the study.
To determine agreement or consensus, some researchers have used frequency
distribution and the criterion of 60% responding to any given response category
(Gohdes & Crews, 2004:63). Other studies such as one conducted by Rayens & Hahn
(2000), used means and standard deviations with a decrease in standard deviations
between rounds indicating an increase in agreement. Coupled with the means and
standard deviations, measures such as the inter-quartile deviation (IQD) have also
been used to determine consensus (Rayens & Hahn, 2000:308) with smaller values of
the inter-quartile range indicating higher degrees of consensus. In his study, Rayens &
Hahn (2000) included another criterion to determine consensus (in addition to the
IQD) in order to achieve what he referred to as stability. The criterion to achieve
consensus was that the IQD should equal to one unit for which more than 60% of
respondents should have answered either generally positive or generally negative.
Items which had an IQD =1, for which the percentage of generally positive or
generally negative responses was between 40 and 60%, were determined to indicate a
lack of consensus or agreement.
105
Table 12: Calculation of Likelihood and Impact significance
Probability that client would implement element if pressured by:
(A)
ID
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
(B)
Client element
Finance H&S Management
Appoint H&S agent
Employ permanent H&S staff
Be involved in design & planning
H&S prequalification criteria for
contractors
Select procurement method suitable for
H&S
Schedule H&S requirements prior to
bidding process
Schedule H&S in contracts
Conduct H&S inspections & audits
Have effective H&S policy, procedures &
goals
Assume H&S leadership role
Total contribution of factor maximum
point = 110
Percent probability of influence
severity rating (values from Q1.9)
Impact significance
(C)
PE
6
7
8
5
7
(D)
SE
4
5
5
4
6
(E)
TECH
5.5
5
6.5
5
5
(F)
EE
8
9
9
8
9
(G)
LE
8
9
9
8
9
(H)
PB
6
5
5
6
5
(I)
Total
score
(Max =60)
37.5
40
42.5
36
41
6
4
5.5
8
8
6
37.5
63
3
4
5
9
9
6.5
36.5
61
4
8
8
6
6
6
5
7.5
7.5
9
9
9
9
9
9
6
6
6
39
45.5
45.5
65
76
76
7
69
4
54
7.5
65
8
95
8
95
5
62.5
39.5
66
Ave-67%
63%
8
5.02
49%
6
2.95
59%
6
3.55
86%
9
7.77
86%
9
7.77
57%
6
3.41
106
(J)
Probability
(%)
63
67
71
60
68
Table 13: Impact significance and severity rating scale
0>1
1>3
3>5
5>7
Low/negligible
Minor
Moderate
Major
Table 14: Likelihood rating scale
0>20%
20>40%
40>60%
Very
Unlikely
May occur 1/2
unlikely
7>10
Critical
60>80%
80%>100%
Likely to occur
Very likely
the time
to occur
Holey et al. (2007:3) used the following criteria to determine consensus:

percentage response;

percentages for each level of agreement for each question to compensate for
varying response rates;

computation of median, standard deviation and their associated group
rankings;

computation of the means, standard deviation and their associated group
rankings using the importance ratings; and

computation of the Weighted Kappa (k) values to compare the chance
eliminated agreement between rounds.
Based on the above estimations, consensus is reached when the following is present:

an increase in percentage agreements;

convergence of importance rankings;

increase in kappa values;

a decrease in comments as rounds progress;

a smaller range of responses; and

smaller values of standard deviations
107
Although there is little consensus on how to determine consensus, it is however
apparent that there has to be some measure of central tendency about which a measure
of dispersion of individual responses can be done. It is also apparent that agreement is
said to have been reached when responses congregate around the central tendency
measure and if variability in responses is small. It is in fact agreeable from the
practice that for consensus to be said to have been attained, there has to be a
convergence of ideas and reasoning towards a subjective central tendency measure.
It was therefore decided for the current study that instead of using one measure as a
criteria for consensus, several other measures in line with what has been agreed upon
were necessary and formed the criteria to determine consensus. These included:
1. the median as the measure of central tendency;
2. percentage of panellists with a generally positive or negative rating on a
question not to be less than 60%;
3. the mean absolute deviation to be used as the measure of dispersion about the
central tendency measure;
4. the calculated average (mean) of the absolute deviation (MAD) not to be more
than one unit; and
5. the range in responses not to be more than four units.
After three rounds of the Delphi process, the mean absolute deviation was found to be
-0.09 which was less than one unit. The value was negative because it was below the
measure of central tendency - the median.
The percentage of panellists leaning either generally positive or negative was 91.6%
up from 78% and 91.0% from the first and second rounds respectively. This was an
indication of a movement towards a convergence of ideas and reasoning. The range in
responses between the highest rating and the lowest rating was 3.50. This value was
lower than the upper limit of 4.0.
These results are summarised in Table 15 and Figure 11 and 12. With those values
after the third round, it was determined that consensus had been reached and therefore
there was no need to proceed to round four.
108
Table 15: MAD, range and percentage agreeing values
Round
MAD
Range in ratings Generally leaning positive or
negative
1
-1.34
5.80
78.0%
2
-0.09
3.90
91.0%
3
-0.09
3.50
91.6%
5.00
5.80
3.90
4.00
3.50
3.00
Range in ratings
2.00
1.00
-0.09
-0.09
2
3
0.00
1
Mean absolute
deviation
-1.00
-2.00
-1.34
Delphi rounds 1 to 3
Figure 11: Plot of MAD and range to determine consensus
95.0%
91.0%
91.6%
90.0%
85.0%
Ratings leaning
generally negative or
positive
80.0%
78.0%
75.0%
70.0%
1
2
3
Delphi rounds 1 to 3
Figure 12: Percent of panellists with similar inclination on an issue
109
Reliability and validity of the Delphi method
Reliability has to do with the extent to which a procedure produces similar results
under constant conditions at all times (Els & Delarey, 2006:52). However in a Delphi
study, this kind of statistical reliability is not possible because another panel may
reach a different conclusion depending on their knowledge of the subject area and
interest. To ensure reliability in the current study therefore, care was taken that
credibility was shown in truthfulness and response consistency from all participants.
In addition personal communication between panel members and the researcher to
clarify questions ensured that the correct phenomenon was being measured.
Validity has to do with how dependable the results are (Els & Delarey, 2006:52). The
panel credibility was one criterion to ensure validity. Credibility was assured during
the selection of the panel. All panellists had distinguished themselves in the work that
they did and the management of 108 funded research projects spoke for itself.
Internal validity was enhanced by the removal of bias or influence from other
members by keeping all members completely anonymous from each other and
therefore completely eliminated the ‘bandwagon’ effect. Results from the Delphi
study were therefore reliable because there was no peer pressure.
In addition, the number of rounds that were adopted in the Delphi process also
enhanced the internal validity. Participants were given chance to change their opinion
or maintain it with a written explanation or argument for dissenting views. This gave
panel members a chance to think through their responses having read what other panel
members had said.
Feedback with the researcher and constant email communication between the
researcher and the panellists individually was another way of ensuring internal
validity of the study. The communication clarified questions and also ensured that
participants remained interested in the discussions. Constant communication also
ensured that almost all the participants continued with the discussions throughout the
three rounds.
The external validity of a study has to do with whether the results can be generalised
to a larger population. This is usually determined by how participants are selected to
110
be part of the study. External validity test was however not a significant factor to the
Delphi study because the result from this study was going to be validated through a
different method- the field questionnaire survey method.
However, the selection of the panel members offered an acceptable level of external
validity. The panel comprised of members from all sectors of the construction
industry, from major geographical locations of the world, and all members were
highly experienced, with 91% being highly published. The Delphi study therefore met
acceptable levels of external validity.
5.3.3
Questionnaire survey
Stage three of the research involved collecting data from the field via questionnaires
in order to meet the general objective GO5 and specific objectives Q1 to Q 5 of the
overall research, and stage three respectively. Stage three formed the pinnacle of the
research. The Delphi study findings were that:
1. the external environment had influence on construction client H&S
performance;
2. clients had influence on designer H&S performance;
3. clients had influence on contractor H&S performance; and
4. project H&S performance was likely to be better with client influence.
Therefore in order to validate findings from the Delphi study, the specific objectives
of the questionnaire survey were to:
Q1. identify external environment factors that have a higher influence on client
H&S culture;
Q2. establish clients’ influence on designer H&S performance;
Q3. establish clients’ influence on contractor H&S performance;
Q4. determine clients’ influence on the overall construction project H&S
performance; and
Q5. determine the goodness–of–fit of the hypothesized client-centred-model to the
sample data.
111
The conceptual model was made up of the following general interrelationships which
in essence constituted the hypotheses that were tested:
H1. the external environment has a direct positive influence on client H&S culture;
H2. client H&S culture has a direct positive influence on contractor H&S
performance;
H3. client H&S culture has a direct positive influence on designer H&S
performance;
H4. client H&S culture has a direct positive influence on project H&S
performance;
H5. contractor H&S performance has a direct positive influence on project H&S
performance;
H6. designer H&S performance has a direct positive influence on project H&S
performance;
H7. client H&S culture, has an indirect positive influence on project H&S
performance mediated by contractor H&S performance; and
H8. client H&S culture, has an indirect positive influence on project H&S
performance mediated by designer H&S performance.
Given the objectives that were to be met, it was decided that in order to collect
information relating to the actual practice in the industry, a field survey was the most
suitable method of collecting the required data. Data was required that related clients
participation or lack of it in H&S management and the practice of H&S by designers,
contractors (top management) and the overall construction project. This type of
information could not be obtained through other means except a field survey. Apart
from the realisation that other means could not have given an acceptable picture of the
various relationships on H&S performance in the industry, the following reasons,
inter alia, motivated the adoption of the method

the philosophy underpinning the research was based on the positivist theory
which uses quantitative methods and from this philosophical view, collection
of data is via tools such as questionnaires;

validation of the conceptual model developed at stage two entailed using an
alternative method to the one initially used in the second stage (Delphi). This
therefore eliminated the use of methods similar to Delphi and its derivatives
112
and called for collection of data on the practice of H&S in the industry through
a field survey;

the field survey was considered to be more representative in that there were
more participants from the construction industry; and

interpretation and presentation of the data could easily be done and understood
by various readers when adopting a positivist philosophy of research because
it follows a logical explanation of the method.
Questionnaire survey instrument
A questionnaire was used to collect data during the field survey. The questionnaire
was based on the literature review conducted at the first stage of the research, as well
as the findings from the Delphi study at the second stage of the research. The
questionnaire was made up of two sections.
The first section was meant for collecting general construction project information
that the respondents were reporting on. The information collected included particulars
such as contract value, type of client, type of project (whether civil engineering,
building construction or other), procurement method used for the project and the type
of contract that was in use on the project. Other information in this first section
pertained to particulars of the respondents. This information included position and
profession of the respondent as well as which party on the project employed the
respondents.
Section two of the questionnaire included questions on H&S practice on the particular
construction projects that the respondents had been involved or familiar with. The
questions specifically sought to establish the extent to which construction clients were
involved, committed, competent, provided leadership on H&S, communicated and
had H&S procedures in place on construction projects.
In addition, information on designers’ and contractors’ H&S performance,
particularly on the extent to which they had procedures in place, how effective their
communication was and how competent they were in terms of H&S was solicited
from respondents. This section of the questionnaire comprised fourteen questions
which required respondents to rate 105 items on the extent to which they agreed or
disagreed on H&S practice of clients, designers and contractors on the project.
113
Furthermore information on what motivated clients’ involvement in H&S
performance was also collected using the second section of the questionnaire. Each
question was rated on the five point Likert scale.
The length of the questionnaire was nine pages including the cover letter. It took
approximately 15 minutes to complete although respondents were informed in the
cover letter that it would take 20 minutes to complete.
The questionnaires were administered via email and by hand and were self-completed.
This type of questionnaire administration was preferred to telephone or face to face
interviews because of, inter alia:

collection of adequate information is possible with minimal effort;

protection of respondent’s anonymity is
guaranteed (Wilkinson &
Birmingham, 2003:39);

the lower cost;

consideration of responses by respondents in their own time;

less bias from the interviewer; and

possibility of covering a wider audience within a shorter time period (Burfold
et al., 2009:207).
However, it was also recognised that this type of questionnaire administration had a
few weaknesses in that there was a:

need to keep questions relatively simple;

need to limit the number of questions to only the critical information to the
research. There are too many questionnaires in circulation competing for
respondent’s time. This situation may result in the return of superficial data;

absence of probing beyond the answer given;

lack of control over who answers the questionnaire; and

possibility of a low response rate due to the amount of questionnaires in
circulation (Wilkinson & Birmingham, 2003:10).
The above weaknesses were addressed by refining the questions and keeping them
relatively simple. However, care was taken not to deviate from the objectives of the
instrument, keeping the overall questionnaire within the recommended limits and
114
ensuring that only the right person completed the questionnaire by constant
communication with respondents. The absence of further probing is characteristic of
this type of data collection and is not a major concern because the data to be collected
was meant for validating the conceptual model initially developed at the second stage.
Questions in the final questionnaire eventually used to collect information were
reviewed and refined by consulting with the research supervisor, the statistics
department and through a pilot survey. Minor changes were deemed necessary after a
pilot survey among 10 construction professionals and discussions with the statistics
department. A copy of the final questionnaire that was used for the study is attached
as Appendix F.
Variables
The research instrument was designed to measure the latent variables namely:
external environment (EE); client H&S culture (CL); designer H&S performance
(DE); contractor H&S performance (CO) and project H&S performance (PP). The
exogenous variable EE (external environment) was hypothesised to be characterised
by indicator variables namely: legislative (LGN), economic (ECO), social (SOC),
professional bodies (PR) and technology (TEC). The indicator variables for the
external environment construct collectively constituted the questionnaire items.
Figure 13 illustrates this relationship between the factors of the external environment
and the latent variable EE.
115
Figure 13: External environment construct
The exogenous variable, client H&S culture, was theorised to be a construct of six
latent endogenous variables namely: leadership (CLL); involvement (CLI);
procedures (CLP); commitment (CLT); communication (CLN) and competence
(CLC). Figure 14 shows the conceptualised client H&S culture measurement model.
The questionnaire items used to measure the leadership factor were marked CLLP1-3,
client involvement was measured by the items labelled CLIP1-3 while that of client
procedures was measured by the items labelled CLPP1-3. Client commitment was
measured by questionnaire items labelled CLTP1-3 and communication was
measured by questionnaire items CLNP1-3. The variable, competence was measured
by items labelled CLC1-4. The numbers of indicator variables measuring one factor
were shown by for instance 1-3, which means there were three indicator variables.
This type of presentation was adopted for clarity and neatness of presentation only.
All indicator variables collectively constituted questionnaire survey items or
questions. During the analysis stage, these variables were parcelled.
116
Figure 14: Client H&S culture measurement model
(CLL= Leadership, CLI = Involvement, CLP = Procedures, CLT = Commitment,
CLN = Communication and CLC = Competence.)
Contractor H&S performance measurement model was defined by indicator variables
namely: procedures (COPP1-3), commitment (COTP1-3) and communication
(CONP1-3). The designer H&S performance was measured by indicator variables of
procedures (DEPP1-3), commitment (DETP1-3) and communication (DENP1-3). The
subscales for contractor and designer H&S performance were theorised to be first
order models with indicator variables grouped in parcels (Figure 15 and Figure 16).
The outcome endogenous variable from the priori structural model was the project
H&S performance factor. The indicator variables theorised to measure project H&S
performance are labelled PP1 to PP17 in Figure 17. The indicator variables also
constituted questionnaire items or questions.
117
Figure 15: Contractor H&S performance measurement model
(COPP= Procedures, COTP= Commitment and CONP= Communication)
Figure 16: Designers H&S Performance measurement model
(DEPP= Procedures, DETP= Commitment and DENP= Communication)
118
Figure 17: Project H&S performance measurement model
Population
The study was conducted in Gaborone and Johannesburg in Botswana and South
Africa respectively. The two cities where chosen for the research because the
researcher was familiar with the areas and it was the intention of the researcher to
delimit the study to these two population areas. It was also viewed that there were
more construction projects in those two areas. Furthermore, accessibility to
respondents was much easier compared to other places in the two countries. It was
also necessary to have two different environments with different laws and practice
governing the common denominator, which was the construction project and more
specifically, the practice of H&S on construction projects. In other words, because of
the external environment factor in the conceptual model, at least two different
environments were necessary which had a different political, legislative and economic
landscape.
119
The unit of measurement was construction professionals who were currently working
or involved in construction projects or had just completed working on a construction
project. The construction professionals represented all organisations that operated in
the construction industry. Those organisations included client, consulting/designer,
contractor and subcontractor entities. The mix was necessary because questions on the
practice of H&S related to all organisations working in the industry.
In addition, cognisance was taken of the likelihood that respondents would respond to
the questionnaire survey, as a function of their:
1. authority to respond where they might not have the formal or informal
authority to respond on behalf of the organisation;
2. capacity to respond where they might not have capacity or the relevant
knowledge to complete adequately the questionnaire; and
3. motive to respond where they might not be personally or organisationally
motivated to disclose information about the organisation (Tomaskovic-Devey
et al., 1994:440).
Information on the practise of H&S by clients, contractors (top management) and the
overall project was needed. Therefore, professionals, who were not working on or had
not been involved in a project within the two years of the study, were excluded from
the study. At least four professionals per project were generated from the lists of
projects that second year Building students at the University of Johannesburg were
attached to for industrial training within Johannesburg. In addition, a referral system
was used for Johannesburg and Gaborone. From these two systems, a sample of 281
responses was realised. The responses obtained from Johannesburg were 217 and the
remaining 64 were obtained from Gaborone in Botswana.
The city of Gaborone is much smaller and the number of both construction projects
and organisations involved in construction business is much smaller compared to that
of Johannesburg and therefore obtaining a smaller number of responses was not
surprising.
Sampling
The department of construction management and quantity surveying at the University
of Johannesburg was contacted for information on the National Diploma students who
120
were currently on industrial training as a requirement for their qualification. From this
list of students, a list of construction projects within Johannesburg and the
surrounding areas was generated. This was one of the most practical and feasible
ways to identify on-going construction projects within the region because there was
no readily accessible database that could be consulted to identify on-going
construction projects. This generated list of projects was used as a target for potential
respondents to the questionnaire.
Coupled with the above method, the researcher used contacts within the industry to
further generate other links and potential respondents where generated by a referral
system. Contacts included professional bodies such as the South African Council for
Project and Construction Managers (SACPCMP). This approach was necessitated by
the links and networks that the researcher had forged within the industry. A list of ongoing projects in Gaborone, Botswana was generated from enquiries within client
organisations, professional contacts and from physical visits to construction sites. This
approach was possible in Gaborone because the area is small compared to
Johannesburg.
Other options to generate lists of potential respondents had been identified and in fact
tried at the pilot stage. One such method was a compilation of a list of contracting
organisations from databases set up by contractor registration bodies such as MasterBuilders Gauteng (MBA-GP) and the Construction Industry Development Board
(CIDB) for organisations registered in Johannesburg and the Gauteng Province.
Enquires were made about on-going projects that the organisations were involved in,
particularly requesting for contact information about construction professionals that
were involved in the construction projects. However, obtaining this information
proved to be very difficult because organisations were not willing to issue information
telephonically or via email to an individual they had never met before.
Similarly, a list of consultants/designers was generated for Johannesburg and
Gaborone. It was again difficult to obtain information on projects and individuals that
could be contacted to complete questionnaires on the projects that they were working
on. As was the experience with contracting organisations, consultants could also not
give the information that was needed to generate a list of potential respondents.
121
The lack of a centralised database, where information on current construction projects
could be accessed, coupled with the difficulties to obtain on-going project
information, made it clear that the response would be very low if the seemingly
logical method of contacting contractors and designers first, was used. Therefore, it
would not have been possible to generate the amount of response that was required for
the study and the type of modelling that was required to validate the conceptual
model. The sample size in this study was an important factor.
Data collection
After generating a list of potential respondents, together with their contact
information, questionnaires were distributed via emails as well as by hand.
Respondents were asked to complete the questionnaires either electronically or
manually, using a pen or pencil. Upon completion respondents could then fax, email
or hand-deliver their response back to the researcher. Other completed questionnaires
were collected physically from the respondents by the researcher. The most common
method used by respondents was returning the questionnaires back to the researcher
electronically via email. This was the case even for those that had been completed
manually because they were scanned and sent back to the researcher as an attachment.
A telephonic interview option was tested at the pilot stage. However, the telephone
interview method proved to be time consuming, costly and required well trained
research assistants to conduct the interviews. Whereas most respondents reported that
they took on average 15 minutes to complete the questionnaire, the telephone
interview was found to require at least 25 minutes. Therefore it was concluded that
the telephone interview option was not suitable. The researcher opted for the selfcompletion method.
The process of data collection, starting from questionnaire circulation to receiving
them back, took about three months. All questionnaires were completed by
respondents in their own time and without undue pressure, because respondents took
time to think through their responses.
After it had become clear that there would not be any more receiving of
questionnaires, data collection was called off especially that the minimum target
number of 200 responses had been exceeded. Each questionnaire was then marked for
122
identity and sent to the University of Johannesburg’s statistical department for data
capturing. Figure 18 is a summary of how the empirical survey was conducted.
Data analysis from questionnaire survey
Raw data from the questionnaires was entered into the Statistical Package for Social
Sciences (SPSS) software and was later exported to the structural equation modelling
(SEM) software EQS version 6.1 and MPlus version 6.0 for analysis. The motivation
for the choice of structural equation modelling and particularly the use of the software
EQS is explained in the next sections.
SEM is the most inclusive statistical procedure in social and scientific research
catering for all operations of the general linear modelling (GLM) group of statistics
such as ANOVA, MANOVA and multiple regression (Kline, 2005:14). Additionally
as Dion (2008:365) explains, SEM simultaneously estimates all coefficients in the
model and therefore it is able to assess the significance and strength of a particular
relationship in the context of the entire postulated model. In addition, considering that
the postulated model in this study consisted of unobservable (latent) variables which
had to be estimated from observable variables, methods of analysis such as ANOVA
could not be used as they lack a direct way to distinguish between observed measures
and the underlying constructs (Kline, 2005:14). In addition, in SEM, a distinction is
made between true variance and error variance, which implies that model parameters
are estimated by taking measurement error into consideration.
The choice of the software EQS for analysis was dictated by the benefit of utilising
the Satorra-Bentler scaled statistic (
), which provides an adjusted, more
robust measure of fit for non-normal data. This approach is more accurate than the
normal chi-square test statistic (
) (Byrne, 2006:22). According to Kline (2005:83),
EQS offers several different estimation methods for non-normal data as well,
including the robust maximum likelihood (RML).
123
Generate list of construction projects
Compiled from list of UJ 2nd year National Diploma Building students on industrial
attachment
Identify potential respondents
obtained from list of projects generated in
first stage above
Obtained from researcher's industry network
and referrals
Circulate questionnaire
Email
hand delivery
Respondents complete questionnaire
Electronically
Manually with pen or pencil
Respondents return back completed questionnaires
Email
Collected by
researcher
Fax
hand delivered
Data compilation and analysis
Statistics data capturers
SPSS
SEM/EQS
Figure 18: Questionnaire survey procedure
Data screening and preparation
Before a detailed analysis of the postulated model was conducted to determine fit,
screening of the data was essential. Pre-analysis data screening focused on
establishing whether there were any missing data, outliers, the distribution
characteristics of the data, and the identification of the model.
According to Raykov, Tomer & Nesselroade (1991:500), missing values and outliers
can adversely affect SEM results by their presence in the raw data. It was therefore
necessary to identify, at the pre-analysis stage, any missing values and outliers in
order to determine the best way to handle them. According to Kline (2005:52),
missing values may pose a very serious challenge to analysis. In addition to
124
identifying missing values and outliers, assessment of multivariate normality was
equally important because the choice of the estimation method depends on it
(Schreiber et al., 2006:327). For example, the common estimation method in SEM is
maximum likelihood which carries with it the assumption of multivariate normality
(Kline, 2005:48; Jackson & Gillaspy, 2009:9). In addition, other problems have been
identified with non-normal data. Jackson & Gillaspy (2009:9) and Kline (2005:137)
reported that failure to meet the assumption of multivariate normality could lead to an
overestimate of the chi-square statistic and therefore to an inflated type I error
(Rejecting a model which should not be rejected). As a result, examination of
statistics of skew, kurtosis, Mardia’s coefficient and the multivariate kurtosis was
conducted to establish normality of the data.
In order to proceed to model analysis, it was critical to determine whether the
postulated model could be analysed or not. Model complexity is determined through
establishing whether a model is just-identified, under-identified or over-identified. A
just-identified model is one in which there is a one-to-one correspondence between
the data and the structural parameters namely, the number of data variances and
covariances being equal to the number of parameters to be estimated (Kline,
2005:108; Byrne, 2006:31). The number of variances and covariances are referred to
as the observations (Kline, 2005:100). According to Byrne (2006:31), a just-identified
model is not scientifically interesting because it has no degrees of freedom and
therefore can never be rejected. Accordingly, an over-identified model is desirable as
it results in positive degrees of freedom that allow for rejection of the model therefore
rendering it to be of scientific use (Byrne, 2006:31). An over-identified model is one
in which the number of estimable parameters is less than the number of observations
(Byrne, 2006:31; Kline, 2005:109). On the other hand, an under-identified model is
one in which the number of parameters to be estimated exceed the number of
variances and covariances (observations) (Byrne, 2006:31). An under-identified
model contains insufficient information to attain a determinate solution of parameter
estimation. As a result, there can be an infinite number of solutions and therefore
defeat the purpose of the analysis (Kline, 2005:109; Byrne, 2006:106).
125
In summary, for a model to be analysed, there has to be at least as many observations
as parameters to be estimated which is that the degrees of freedom (df) should be
greater than zero (
) (Kline, 2005:100).
Parameter estimation and input matrix method
Examination of the degrees of freedom of the postulated model in this study revealed
that the model was over-identified. The least value for the degrees of freedom was
found to be nine. This degree of freedom related to the external environment
construct. All values of degrees of freedom for the model constructs were positive and
therefore indicative of an over-identification of the measurement models (Kline,
2005:100).
After screening the data, it was established that the data was non-normal with the
lowest Mardia’s coefficient of 41.0290 (contractor H&S performance) and the highest
Mardia’s coefficient of 443.7814 (Client H&S culture). See Table 23.
The non-normality of the data influenced the choice and use of the robust maximum
likelihood (RML) estimation method. The RML gives several robust fit indices
(Bartholomew, Loukas, Jowers & Allua, 2006:72). According to Byrne (2006:22),
one of the outputs from the RML estimation method is the robust chi-square statistic (
) referred to as the Satorra-Bentler scaled statistic (
) and robust standard
errors which are corrected for non-normality in large samples. The structural equation
modelling software, EQS version 6.1 was used in part, due to the ability of the
programme to adjust standard errors for the non-normality of the data (Byrne,
2006:22).
The covariance matrix was the chosen in-put matrix for the analysis in the current
study. The analysis strategy adopted to analyse the postulated model was first to
estimate the measurement part of the model and thereafter analyse the measurement
and structural parts of the model simultaneously. Similarly the results from the
analysis were reported in the same manner namely, results from the measurement
model analysis referred to as the confirmatory factor analysis (CFA) were presented
first and thereafter the results from the analysis of the entire structural model referred
to as the full latent variable model (FV) were presented.
126
Model analysis
Having satisfied the pre-analysis conditions, selected the input matrix of the data and
selected the method of estimation; analysis of the postulated model was the next step
in the analysis process. The following fit indices identified from Raykov et al.
(1991:501); Hu & Bentler (1999:5); Boomsma (2000:473), Kline (2005:134), and
Schreiber et al. (2006:327) were examined to determine model fit. These statistics
were:

Chi-square values
;

Bentler comparative fit index (CFI);

Satorra-Bentler scaled Chi-square statistic (

Standardised root mean square residual (SRMR);

Root mean square error of approximation (RMSEA);

Root mean square error of approximation with its 90% confidence interval
);
(RMSEA @ 90% CI); and

Tucker Lewis index (TLI).
The fit indexes
, CFI, TLI and (
) belong to the incremental or comparative
fit indexes while the SRMR and RMSEA belong to the absolute fit indexes (Byrne,
2006:99). The adopted cut-off values for the above fit indices are as tabulated in
Table 16.
The decision on model fit was based on the proposal by Hu & Bentler (1999:28) to
use a two index presentation of incremental and absolute fit indexes because they
seemed to perform superior to a single index presentation strategy. Hu & Bentler
(1999:27) suggested therefore, that the maximum-likelihood based SRMR and a
supplemental fit index such as a CFI or RMSEA, would result in minimum Type I
(namely, the probability of rejecting the null hypothesis when it is true) and a Type II
error (namely, the probability of accepting the null hypothesis when it is false).
The statistical significance of parameter estimates was established by examining the
ratio output of the parameter estimate divided by its standard error (therefore
analogous of z-values) and tests that the estimate is statistically different from zero
(Byrne, 2006:103). This is in agreement with a proposal by Schreiber et al.
(2006:327) to determine statistical significance by examining the z-value test statistic.
127
Based on the
level of 0.05, the test statistic had to be greater than 1.96
before the hypothesis could be rejected (Byrne, 2006:103).
Table 16: Cut-off criteria of fit statistics
Statistic
Cut-off criteria
Chi-square  ratio
to
2 0r 3
Bentler comparative fit
index (CFI)

Standardised root mean
square residual (SRMR)


Root mean square error of
approximation (RMSEA)



Root mean square error of
approximation with its 90%
confidence interval (RMSEA
at 90% CI)
Tucker Lewis index (TLI)

Reference
(Kline, 2005;
Schreiber et al. 2006)
value should be 0.95 for (Bartholomew et al.,
2006; Dion, 2008;
good fit
Schreiber et al. 2006);
the value should be 0.08 (Kline, 2005;
a value of 0.1 is also Schreiber et al., 2006)
acceptable
value should be < 0.05 for (Hu & Bentler; 1999;
Kline,2005;
good fit
values < and 0.08 indicate a Bartholomew et al.,
reasonable
error
of 2006; Dion, 2008)
approximation
values of > 0.10 suggests a
poor fit
(Schreiber et al. 2006)
value to be 0.06 to 0.08
should be > 0.9
(Hu & Bentler; 1999)
Reliability and validity
In addition to the pre-analysis and post-analysis tests described above, tests on score
reliability and construct validity were conducted.
In order to determine the score reliability, the internal consistency reliability measure
statistic of Rho coefficient and Cronbach’s alpha ( ) were examined. According to
Kline (2005:59), the Cronbach’s alpha measures the degree to which responses are
consistent across all items within a single measure and if this statistic is low, the
content of the items may be so heterogeneous that the total score is not the best
possible unit of analysis for the measure.
However, acceptability of the Cronbach’s alpha to measure internal homogeneity is
limited. Byrne (2006:133) argue that the use of Cronbach’s alpha coefficient to latent
128
variable models especially models with multi- dimensional structure is questionable
because it is based on a very restrictive model that requires all factor loadings and
error variances to be equal. Therefore, in determining score reliability, the Rho
coefficient was relied upon more than the Cronbach’s alpha coefficient because it
provides a good estimate of internal consistency especially because the model that
was analysed in the current study was a full latent variable model (Byrne, 2006:133).
Apart from establishing the score reliability, construct validity was established
through evaluation of the convergent validity and magnitude of the parameter
estimates of all relations in the model using the confirmatory factor analysis (CFA) of
the SEM (Kline, 2005:60). In addition, due to the absence of another external criterion
against which comparison could be made of measures, discriminant validity was also
used to examine construct validity. According to Kline (2005:60), a set of variables
postulated to measure different aspects shows discriminant validity if their intercorrelations are not too high.
Questionnaire response
Table 17 is a breakdown of the profile of respondents. A sample of 281 responses was
realised when the survey closed. Of the total 281 responses, 84 were quantity
surveyors, 42 were engineers (civil, mechanical, electrical and services), 36 were
consulting designers, 42 were project managers and 20 were construction H&S
professionals. The number of respondents employed by a consulting or designer firm
was 107 and 145 were employed by contracting firms. The number of respondents
employed by client organisations was 24.
The profile of the projects that construction professionals reported on is as shown in
Table 18. The majority of respondents reported on building construction projects
contributing 209 responses while civil engineering projects contributed 50 responses.
The other 18 respondents indicated that they were working on other projects such as
renovation works.
In terms of size of projects reported on, 12.8% had a contract value of between 4 and
6.5 million Rands, 13.2% between 6.5 and 13 million Rands, 23.8% between 13 and
40 million Rands, 21.4% between 40 and 130 million Rands while 26% reported on
construction projects with contract values of over 130 million Rands (See Table 18).
129
Most respondents were working on projects where the client was a public or
government institution (Table 20). The percentage of respondents that indicated this
was 53.6% followed by those working for private property developers with 21.2%.
Respondents that worked on projects where the client was a state owned enterprise
(SOE) was 11.7%. The mining organisations only had 4.4% respondents (See Table
20.)
Most projects reported on by respondents were procured through the common method
of open tender with 58.5 % respondents indicating thus. The selective tender method
had 30.7% respondents and only 8.9% indicated that the projects they reported on
were procured through negotiation. The rest were procured by other methods. This
finding was not surprising since most of the respondents indicated that the clients on
the projects they were reporting on were public or government institutions. These
institutions are still practicing old conversional methods.
In terms of the project delivery system used for the projects reported on, 22.1%
respondents indicated that they were reporting on projects that were delivered through
the design and build method, 35.8% through the traditional design bid build, and the
remaining 42.1% indicated that the projects were procured through other methods
including construction management. Table 21 is a summary of the profile of project
types that respondents reported on regarding the practice of H&S.
The sample of respondents spanned across all major project types, procurement
methods and worked with all the major clients in the industry. Respondents
themselves were of different critical backgrounds ranging from site agents to safety
officers. The diversity shown in the respondent characteristics was considered to be a
factor enhancing the research’s prudence.
130
Table 17: Respondents profile
Respondent
Clerk of works
Consulting designer
Contracts manager
Engineer
Health and safety consultant
Project manager
Quantity surveyor
Safety officer
Site agent
Not stated
Total
Frequency
Percent %
1
36
16
42
1
42
84
20
32
7
281
0.4
12.8
5.7
14.9
0.4
14.9
29.9
7.1
11.4
2.5
100
Table 18: Contract value of projects reported on in this study
Contract value
Frequency
(ZAR Million)
4-6.5
36
6.5-13
37
13-40
67
40-130
60
Over 130
73
Total
273
Valid Percent
13.2
13.6
24.5
22.0
26.7
100
Table 19: Project type reported on in this study
Project Type
Percent
Civil engineering
Building Construction
Other
18.4
74.7
6.9
Total
100
131
Table 20: Client organisations represented in the survey
Client
Public / state
Private property developers
State owned enterprise (SOE)
Mining organisations
Other clients
Total
Table 21: Contractor selection method reported by respondents
Method
Frequency
Open tender
Selective tender
Negotiated
Other
Total valid
5.4
158
83
24
5
270
Percent responses %
53.6
21.2
11.7
4.4
9.1
100
Percent
58.5
30.7
8.9
1.9
100
CHAPTER SUMMARY
This chapter provided a framework of how the current study was conducted. The first
stage involved establishing the need for the study, conducting a review of literature
and theorising on client influence on H&S performance. The second stage involved
conducting a Delphi study, developing a conceptual model and evolving a
questionnaire for the survey for use at the third stage. The third and final stage
involved conducting a survey, analysing and modelling the results using structural
equation modelling, determining the best fit model and comparing the conceptual
model to the sample data from the questionnaire survey. Conclusions and
recommendations were made based on findings from both the Delphi and the
questionnaire survey. Chapters six, seven and eight will therefore now present
findings from the Delphi and questionnaire survey.
132
CHAPTER SIX
RESULTS FROM THE DELPHI STUDY
6.1
INTRODUCTION
A Delphi study was principally conducted to estimate the level of impact of client
influence on project H&S performance in the construction industry. Other aspects
were also investigated and they included the estimation of the level of impact of the
external environment on client H&S performance. In addition, the level of impact of
client H&S culture’s influence on designer and contractor H&S performance was
investigated.
The latent variable, client H&S performance, was defined by factors of client H&S
culture namely: leadership, involvement, commitment and competence. H&S
performance was interpreted in terms of H&S culture. Therefore in this study, a better
performance means a better H&S culture. Consequently, client H&S culture is used
throughout this study.
This section will present results from the Delphi study. The study comprised three
rounds. Computations for each and every question were made regarding the estimated
probability improvement in H&S performance by the client, contractor, designer or
the overall project as a result of an external influence.
The objectives of conducting the Delphi survey were, inter alia, to estimate the impact
of the:
D1. external environment factors on client H&S performance;
D2. client H&S culture on H&S consideration (hazard identification, risk analysis,
assessment and mitigation) throughout the project life cycle;
D3. client H&S culture on contractor H&S performance (top management);
D4. client H&S culture on designer H&S performance; and
D5. client H&S culture on the overall project H&S performance.
133
The associated outcomes from achieving the above objectives were to:
1. identify factors of client H&S culture, designer and contractor H&S
performance, and environmental factors that have a significant influence on
H&S performance; and
2. develop a client-centred model that was based on the identified factors with
significant influence on H&S performance.
Results of the Delphi study are therefore presented in relation to the specific Delphi
objectives.
6.2
FINDINGS FROM THE DELPHI STUDY
6.2.1
Impact of the external environment on client H&S culture – D1
The external environment was defined by six factors, namely: political, social,
economic, legislative, professional bodies and technology. The impact significance of
these factors’ influence on client H&S culture was obtained as a product of client’s
likelihood to implement H&S elements and the severity rating or negative impact on
client’s H&S performance if the factors were absent.
The level of influence was determined by assessing the extent to which client would
implement various H&S elements if pressured by the external environment. Severity
of an environmental factor was the rated negative effect on client H&S performance
that would result from an environmental factor’s absence. The severity rating was
based on an ordinal scale of 0 to 10 with 0 being negligible and 10 critical. The
Impact significance was obtained as a product of the severity rating of an
environmental factor and the likelihood of client implementing a particular H&S
element (Refer to equation 2).
Of the six environmental factors, three of them namely, political, economic and
legislative, were determined to have an impact significance of over 5.0. The economic
and legislative factors had an impact significance of 7.77 each. According to the
classification scale used in this study (Table 13), a rating of 7.77 was considered to be
‘critical’. The rating suggested that the two factors namely, economic and legislative
were critical to client implementing the required H&S elements or programmes (See
Figure 19).
134
The influence of the political factor had a lower impact significance rating of 5.02.
This estimate of 5.02 was considered to be of ‘major’ impact significance. This
impact significance was lower than that of the economic and legislative factors, by
about 35%. From this finding, it seemed that the economic and legislative factors
were considered to have a higher impact than political influence on client H&S
performance.
The likelihood of clients implementing H&S elements, as a result of external
environment’s influence was 67% on average (see Figure 20). The standard deviation
in the likelihood ratings was 0.06. The small standard deviation suggested that the
likelihood of the client implementing the H&S elements was almost the same.
However, clients were least likely to be involved in design and planning of H&S
activities. The likelihood for this element was determined to be 60% (see Figure 20).
On the other hand, clients were most likely to conduct H&S audits and inspections
and have H&S policies, procedures and goals. The likelihood for these H&S elements
was determined to be 76% each.
Eight out of eleven client H&S elements were considered ‘likely’ to be implemented
as a result of external environment influence. This translated to 73% of all H&S
elements that were ‘likely’ to be implemented and 27% were ‘very likely’ to be
implemented. The H&S elements that had likelihood above 70% were, for the client
to:

have effective H&S policy, procedures, and goals (76% likelihood);

conduct H&S inspections and audits (76% likelihood); and

employ permanent H&S staff (71% likelihood).
Therefore these H&S elements were considered, ‘very likely to occur’ (see Table 14)
as a result of external environment’s influence. Three factors of the external
environment namely: political, economic and legislative, had a higher impact on
client H&S performance compared to the influence of professional bodies, social and
technology factors. The political, economic and legislative influence contributed 59%
to the total likelihood that clients would implement H&S elements. On the other hand,
professional bodies, social environment and technology contributed 41% to the total
135
likelihood that clients would implement the H&S elements. The standard deviations
for these two groups of factors were determined to be 0.02.
Figure 19: Impact significance of external environment factors to client culture
Figure 20: Client likelihood to implement H&S elements
136
Figure 21: Client likelihood to implement H&S elements
6.2.2
Client influence on H&S consideration in the project lifecycle - D2
Client H&S culture was defined by four factors, namely: commitment, competence,
involvement and leadership. The influence of these factors was estimated. As
explained in the previous section, the level of influence of a factor was inferred from
the calculated value of the impact significance.
The H&S performance on a construction project depends in part on the extent to
which it is considered throughout the project life cycle. H&S consideration entails
H&S risk identification, analysis, assessment, mitigating, implementing and assessing
the implementation of H&S measures. In order to evaluate H&S performance at the
project level, it was necessary to evaluate H&S consideration throughout the project
life cycle.
137
The impact significance of all factors of client H&S culture was determined to be
‘critical’. The findings seemed to suggest that client H&S culture was critical to
achieving the required level of H&S consideration. The average impact significance
of all factors of client H&S culture was determined to be 7.36. This rating was
considered to be of ‘critical’ impact significance.
The impact significance of the commitment, involvement, competence and leadership
factors was determined to be 7.59, 8.10, 7.07 and 6.68 respectively (see Figure 22).
The standard deviation of these values was determined to be 0.62. The small standard
deviation seemed to suggest that all factors of client H&S culture were almost of
equal importance. This finding also seemed to suggest that a good total client H&S
culture was desirable as opposed to individual isolated client efforts to improve
project H&S performance.
9.00
8.00
7.59
AVE=7.36
SD = 0.62
8.10
7.07
Impact significance
7.00
6.68
6.00
5.00
4.00
3.00
2.00
1.00
0.00
Committment
Involvement
Competence
Leadership
Client H&S culture aspects
Figure 22: Impact significance of factors of client H&S culture
The influence of the involvement factor was determined to be more ‘critical’ when
compared to other three factors of client H&S culture. The impact significance of the
involvement factor was determined to be 8.10. This rating was considered to be
‘critical’ impact significance (see Table 13). The factor with the least impact
138
significance on H&S consideration was leadership. The impact significance of the
leadership factor was determined to be 6.68. This rating was considered to be of
‘major’ impact significance (see Table 13). However although the rating of 6.68 was
low, when compared to other factors of client H&S culture, its impact significance
and therefore its influence was considered to be ‘major’. Therefore the contribution of
the leadership factor to H&S performance may not be overlooked.
The likelihood of H&S being considered as a result of client H&S culture’s influence
was determined to be 83% (Figure 13). This rating of 83% suggested that H&S
consideration throughout the project life cycle was ‘very likely to occur’ (Table 14).
Figure 23: Likelihood of H&S consideration due to client’s influence
It was further determined that H&S consideration was ‘very likely to occur’ at all the
project stages except at the procurement stage. With the exception of the procurement
stage, the likelihood of H&S consideration at all other project stages was determined
to be above 80%. On the other hand, the likelihood of H&S consideration at the
procurement stage was determined to be 73%. The likelihood of 73% meant that H&S
consideration was ‘likely to occur’.
H&S consideration was most likely at the construction stage. The likelihood was
determined to be 95%. It seemed from this finding that H&S consideration was almost
certain at the construction stage with client H&S culture’s influence.
139
The likelihood of H&S consideration at all project stages was high. This finding
seemed to suggest that client H&S culture was critical to project H&S performance.
All factors of client H&S culture contributed almost equally to the total influence on
H&S consideration.
6.2.3
Impact of client H&S culture on contractor H&S performance - D3
The influence of client H&S culture on contractor H&S performance was estimated.
Client H&S culture was defined by the factors namely: commitment, competence,
involvement and leadership.
The impact significance of client H&S culture on contractor H&S performance was
determined to be 6.60 (see Figure 24). This rating was considered to be ‘major’
impact significance. The rating suggested that client H&S culture had a major impact
on contractor H&S performance.
7.20
Impact significance
7.00
6.80
7.00
Ave = 6.60
SD = 0.34
6.70
6.50
6.60
6.40
6.20
6.20
6.00
5.80
Committment
Involvement Competence
Client H&S culture aspect
Leadership
Figure 24: Impact significance of client factors on contractor H&S performance
All factors of client H&S culture had an impact significance rating of between 5.0 and
7.0. The impact for this scale range was considered to be ‘major’ impact significance’.
Figure 24 shows that the leadership factor had the most impact significance when
compared to all other factors of client H&S culture. The impact significance of this
factor on contractor H&S performance was determined to be 7.0. This rating was
140
considered to be ‘critical’ impact significance. The finding seemed to suggest that the
leadership factor was critical to contractor H&S performance.
The competence factor had the least impact significance on contractor H&S
performance. This factor had a rating of 6.20. Although the impact significance was
lower when compared to all other factors of client H&S culture, its impact was
considered to be ‘major’ (Table 13). Apart from this, the difference in the impact
rating between all factors was minimal. The standard deviation in impact significance
values was found to be 0.34 suggesting a small variability between all impact
significance estimates.
The likelihood that contractors would implement H&S elements as a result of
influence from client H&S culture was determined to be 83% on average. This rating
suggested that the implementation of H&S elements by contractors due to client H&S
culture’s influence was ‘very likely to occur’ (Table 14). The standard deviation of all
likelihood estimates was found to be 0.03. This standard deviation value was very
small and therefore indicated a small variability in the estimates.
Figure 25: Contractor likelihood to implement H&S elements
141
Contractors were ‘very likely’ to implement the following six H&S elements as a
result of client H&S culture’s influence:
1. employ permanent H&S staff;
2. conduct H&S audits and inspections;
3. carry out hazard identification and risk assessment;
4. improve the level of top management commitment to H&S;
5. consult and communicate H&S information to all stakeholders; and
6. establish and implement H&S policies, procedures and goals.
The likelihood of implementing all these six H&S elements was determined to be
above 80%.
When compared to all other H&S elements, contractors were least likely to involve
workers in H&S management and also to develop staff H&S competence. These two
H&S elements had the likelihood of less than 80% but crucially not less than 70%.
The likelihood of implementing each one of these H&S elements was determined to
be 78%. According to Table 14, this rating meant that contractors were ‘likely’ to
implement H&S elements even though they had a low rating.
6.2.4
Impact of client H&S culture on designer H&S performance - D4
The role of designers has been recognised as being important to H&S performance on
construction projects. However, some studies have suggested that designers are
unwilling to participate fully through their designs or take up a leadership role, for
example, in managing H&S. This reluctance has in a way inhibited better
performance in H&S. It seems as though, designers can only fully participate in H&S
performance with the help of an external influence.
One such external influence could be the client who in fact is their employer. It was
therefore the objective of this study to partly establish the impact significance of
client H&S culture on designer H&S performance. The impact significance of factors
of client H&S culture namely: commitment, involvement, competence and leadership
was evaluated. In addition, designers’ likelihood to implement H&S elements as a
result of client’s influence was determined.
142
The impact significance of client H&S culture on designer H&S performance was
determined to be 6.45 on average. The estimate of 6.45 was slightly lower than the
impact significance of client H&S culture on contractor H&S performance. The
impact significance on contractor H&S performance was determined to be 6.60.
However, there was more variability in the impact significance values relative to
designer H&S performance when compared to the impact significance on contractor
H&S. The standard deviation in the impact significance values relative to designer
H&S performance was determined to be 0.8. On the other hand, the standard
deviation of the impact significance values relative to contractor H&S performance
was found to be 0.34.
The involvement factor was considered to be more critical to designer H&S
performance and its impact significance was determined to be 7.31. The factor of
client H&S culture with the least impact significance on designer H&S performance
was competence. This factor had an impact significance of 5.42. The leadership factor
was the second most influential factor followed by commitment. These factors were
determined to have an impact significance of 6.75 and 6.30 respectively (Figure 26).
Apart from the involvement factor, all other factors of client H&S culture were
considered to have ‘major’ impact significance on designer H&S performance.
When the impact significance of factors of client H&S culture are compared between
those obtained for designer H&S performance and contractor H&S performance, the
competence factor was found to be the least rated. The competence factor had an
impact significance of 6.20 on contractor H&S performance and 5.42 on designer
H&S performance (Figure 27). The difference in estimates between that of impact on
designers and on contractors was suggestive of client’s competence being slightly
more critical to contractors’ than to designers.
143
8.00
Impact significance
7.00
7.31
6.75
6.30
6.00
5.42
Ave = 6.45
SD = 0.80
5.00
4.00
3.00
2.00
1.00
0.00
Committment
Involvement Competence
Client H&S culture aspect
Leadership
Figure 26: Impact significance of client factors on designer H&S performance
In the case of H&S consideration throughout the project life cycle, clients’
competence was rated to be 7.07. This was suggestive of the fact that client
competence was extremely necessary and ‘critical’ to ensure that H&S consideration
was undertaken throughout the project life cycle.
The involvement factor had the most impact significance on both designer H&S
performance and H&S consideration throughout the project lifecycle. The impact
significance of this factor was determined to be 7.31 on designer H&S performance
and 8.10 on H&S consideration. The high impact significance of the involvement
factor on both designer H&S performance and H&S consideration seemed to suggest
that this factor was critical to get designers to participate in H&S management. This
finding appears to be logical because it is expected that H&S consideration
throughout the project lifecycle would be possible if designer H&S performance is
favourable.
144
9.00
8.10
8.00
Impact significance
7.00
7.59
6.30
6.70
7.31
7.07
6.50
6.00
6.75
7.00
6.68
6.20
5.42
5.00
4.00
3.00
2.00
1.00
0.00
Commitment
Involvement
Client H&S culture aspect
Competence
Leadership
Designers
Contractor top management
H&S consideration
Figure 27: Overall impact significance of client H&S culture factors
All factors of client H&S culture were found to be important to designer H&S
performance. The likelihood that designers would implement H&S elements as a
result of client H&S culture’s influence was determined to be 78% on average (Figure
28). This likelihood value suggested that designers were ‘likely’ to implement H&S
elements when influenced by client H&S culture (Table 14).
Designers were most likely to implement the following four H&S elements:
1. involve contractors in design reviews;
2. conduct hazard identification and risk analysis;
3. design for H&S and
4. establish effective H&S policies, procedures and goals.
145
The likelihood of designers implementing these factors was determined to be above
80%. This rating suggested that designers were ‘very likely’ to implement H&S
elements when influenced by client H&S culture. Therefore, it can be inferred from
this finding that client H&S culture could be a leading indicator of designer H&S
performance. Specifically, the findings suggest that designers were very likely to
involve contractors, conduct hazard identification and risk assessment, design for
H&S and also establish effective H&S policies, procedures and goals when influenced
by a positive client H&S culture.
The standard deviation in the likelihood values was found to be 0.04 (Figure 28). This
value was very small and therefore seemed to indicate that all H&S elements had an
almost equal possibility of being implemented.
Design for H&S
83%
Hazard identification & risk assessment
83%
Involve Contractors in design reviews
83%
Establish & implement H&S Policies,…
Conduct H&S audits and inspections
H&S element
Consult & communicate H&S information to…
80%
73%
75%
Employ permanent H&S staff
75%
Develop staff competency on H&S
75%
Ave = 78%
SD = 0.04
60%
70%
80%
Likelihood
90%
Figure 28: Designer likelihood to implement H&S elements due to client influence
146
The other H&S elements which had above 75% likelihood of being implemented
included the designer to:

develop staff competence,

consult and communicate H&S information and

employ permanent H&S staff.
The rating of 75% likelihood meant that designers were ‘likely’ to implement H&S
elements (Table 14).
Designers were least likely to conduct H&S audits and inspections. The likelihood of
designers conducting H&S audits and inspections as a result of client H&S culture’s
influence was determined to be 73%. Although this likelihood was low when
compared to the likelihood of other H&S elements being implemented, it was actually
a favourable result considering that conducting H&S audits and inspections was
‘likely to occur’.
Figure 29: Designer & contractor likelihood to implement H&S elements
147
Contractors were more likely to implement H&S elements than designers when
influenced by client H&S culture. The likelihood of contractors implementing H&S
was 83% while that of designers was 78% (Figure 29). The largest disparity was
observed in the likelihood that the following H&S elements would be implemented,
namely:
1. employ permanent H&S staff;
2. consult and communicate H&S information to all stakeholders; and
3. conduct H&S audits and inspections.
Contractors were more likely to implement these H&S elements. According to the
likelihood scale (Table 14), contractors were ‘very likely’ to implement H&S
elements while designers were ‘likely’ to implement H&S elements. Similarly, the
impact significance of client H&S culture was greater on contractor H&S
performance than it was on designer H&S performance (Figure 30). The average
impact significance of client H&S culture on contractor H&S performance was 6.60
while that on designer H&S performance was 6.45.
Figure 30: Impact of client factors on designer and contractor H&S performance
148
6.2.2.2 To estimate the impact significance of designer H&S performance on H&S
consideration throughout a project life cycle
The influence of designer H&S performance on H&S consideration was also
evaluated. Designer H&S performance was defined by four factors namely:
leadership, involvement, commitment and competence. The impact of each of these
factors of designer H&S performance was therefore evaluated.
The impact significance of the leadership, competence and commitment factors was
determined to be 7.80, 7.35 and 7.05 respectively (Figure 31). The impact
significance of the involvement factor was also determined to be 7.05. All factors of
designer H&S performance had impact significance values above 7.0. All factors of
designer H&S performance were therefore considered to be ‘critical’ to H&S
consideration throughout the project lifecycle.
The leadership factor was found to have the largest impact significance when
compared to all other factors of designer H&S performance. This factor had an impact
significance of 7.80 on H&S consideration. The competence factor had the second
largest impact significance determined to be 7.35.
The standard deviation of the impact significance for the factors of designer H&S
performance was found to be 0.35 (Figure 31).
9.00
Impact significance
8.00
7.00
7.05
7.05
7.35
7.80
6.00
5.00
4.00
3.00
2.00
1.00
0.00
Committment Involvement Competence Leadership
Designers H&S culture aspects
Figure 31: Impact significance of factors of designer H&S culture
149
Ave = 7.31
SD = 0.35
The leadership factor, of designer H&S performance, contributed about 33% to the
total impact on H&S consideration. This value suggested that leadership from
designers was critical for H&S consideration to be implemented.
When compared to clients’ influence on H&S consideration, designers’ influence on
H&S consideration was found to be equally critical. The average impact significance
of client H&S culture on H&S consideration was found to be 7.36 while that of
designer H&S performance was 7.31 (Figure 32). The impact significance of client
H&S culture was just marginally higher than that of designer H&S performance. Both
these ratings were considered to be ‘critical’ to H&S consideration.
9.00
8.1
7.80
7.59
8.00
7.05
7.35
7.05
7.07
Impact significance
7.00
6.68
Client
Ave = 7.36
SD = 0.62
Designer
Ave = 7.31
SD = 0.35
6.00
5.00
4.00
3.00
Designer
2.00
Client
1.00
0.00
Committment
Involvement
Competence
Leadership
H&S culture aspects
Figure 32: Impact of client and designer H&S culture on H&S consideration
The average likelihood was 81% of H&S being considered at all the project phases as
a result of designer H&S performance. Figure 33 shows that the likelihood of H&S
consideration was above 80% with the exception of H&S consideration at the
initiation or concept stage. The average likelihood of 81% was indicative of the
possibility that H&S consideration throughout the project life cycle was ‘very likely
to occur’ with designers’ influence. Therefore, the finding seemed to suggest that
150
designer H&S performance was a leading indicator of a better H&S performance
generally and specifically for H&S consideration throughout the project lifecycle.
H&S consideration was found to be least likely at the initiation stage. The likelihood
of H&S consideration at the initiation stage was found to be 73%. Although the 73%
likelihood of H&S consideration at the initiation/concept stage was lower when
compared to the likelihood at other construction stages, it was nevertheless ‘likely to
occur’ (Table 14).
The project phases at which H&S consideration was ‘very likely to occur’ due to
designers’ influence included the:
1. design;
2. procurement;
3. commission and testing, and
4. closeout stage.
The likelihood of H&S consideration was above 80% at these construction stages.
The likelihood of H&S consideration at both the construction and design stage was
the highest with 85% likelihood each.
Overall, H&S consideration throughout the project life cycle was considered to ‘very
likely to occur’. In addition, the variance in the likelihood values was found to be very
small. The standard deviation between all likelihood estimates was found to be 0.05
(Figure 33). It was also observed that the likelihood of H&S consideration was
highest at the phases where designers are typically most active. This is during the
design and construction stages.
When compared with the clients’ influence on H&S consideration, designers had a
lower influence on H&S consideration. The likelihood was 83% that H&S would be
considered throughout the project lifecycle as a result of clients’ influence. On the
other hand, the likelihood was 81% that H&S would be considered as a result of
designers’ influence. The likelihood of H&S consideration was highest at the
construction stage in both cases. The likelihood of H&S consideration at the
construction stage was 95% with clients’ influence and 85% with designers’
influence.
151
Figure 33: Likelihood of H&S consideration due to designers' influence
Figure 34: Likelihood of H&S consideration due to client and designer influence
152
6.2.5
Direct client’s influence on project H&S performance - D5
The direct impact of client H&S culture on project H&S performance was evaluated.
In addition, the direct impact of both designer and contractor H&S performance were
evaluated. The likelihood that clients would implement H&S elements as a result of
external environment’s influence was determined to be 67% (Table 22)
The severity of client H&S culture to project H&S performance was determined to be
8.88. From Equation 2, the direct impact significance of client H&S culture on project
H&S performance was found to be 5.92. This rating was considered to be ‘major’
impact significance.
On the other hand, the likelihood that designers would implement H&S elements as a
result of client H&S culture’s influence was determined to be 78%. The severity of
designer H&S performance on project H&S performance was rated to be 8.75. This
value was slightly lower than that of client H&S culture. The impact significance of
designer H&S performance was therefore determined to be 6.84. This impact
significance rating was considered to be ‘major’.
Contractor H&S performance had the most impact significance. The likelihood that
contractors would implement H&S elements as a result of client H&S culture’s
influence was determined to be 83%. The severity of contractor H&S performance
was determined to be 9.38. From Equation 2, the impact significance of contractor
H&S performance on project H&S performance was determined to be 7.73. This
impact significance rating was considered to be ‘critical’. However, an observation
was made that contractors had a high likelihood of implementing H&S elements as a
result of client H&S culture’s influence. This high likelihood of implementing H&S
elements enhanced contractors’ impact significance on project H&S performance.
Other findings from the study were that client H&S culture could possibly be a six
factor construct. The aspect of clients having standard procedures and effective
communication was highlighted. The elements mentioned under procedures included
the client having H&S programmes, goals, conducting H&S inspections, carrying out
hazard identification and risk assessments and specifically specifying H&S
requirements in construction contracts.
153
Table 22: Influence level of various project stakeholders
Stakeholder
Likelihood Stakeholder
severity to
project H&S
Client’s likelihood of
67%
8.88
implementing H&S elements
due to environmental influence
Designer’s likelihood of
78%
8.75
implementing H&S elements
due to clients influence
Contractor’s top management
83%
9.38
likelihood of implementing H&S
elements due to clients’
influence
6.3
DISCUSSION OF DELPHI RESULTS
6.3.1
Objective D1
Impact
significance
Remark
5.92
Major
6.84
Major
7.73
Critical
The first objective of the Delphi study was to estimate the influence of the external
environment factors on client H&S performance. This was achieved by determining
the impact significance of the factors of external environment on client H&S culture.
The level of influence of the external environment was inferred from the impact
significance on client H&S performance of the external environment factors. The
external environment was defined by six factors namely:
1. political;
2. economic;
3. social;
4. materials and methods (Technology);
5. professional bodies; and
6. legislative.
154
In addition, the likelihood of clients implementing H&S elements as a result of being
influenced by factors of the external environment was also assessed. The H&S
elements were for the client to:
1. provide finance for H&S;
2. appoint H&S agent;
3. employ permanent H&S staff;
4. be involved in design and planning;
5. select procurement methods suitable for H&S promotion;
6. schedule H&S requirements prior to bidding;
7. conduct H&S inspections and audits;
8. have effective H&S policy, procedures and goals; and
9. assume H&S leadership role.
The likelihood of clients implementing all of these H&S elements as a result of the
external environment’s influence was determined to be 67%. According to the five
point likelihood scale with ratings ranging from ‘not likely’ (0%) to ‘very likely to
occur’ (100%) the value of 67% meant that the implementation of H&S elements by
clients as a result of external environment’s influence was ‘likely to occur’ (Table 14).
This finding seemed to indicate that the external environment’s influence was
necessary in order to assure that clients implemented these H&S elements or
programmes. This finding not only confirms that the external environment was
necessary for client H&S performance; it also provided an estimate of the extent to
which client H&S performance could be influenced.
The factors of the external environment which had major influence on client H&S
performance were also identified. The political, economic and legislative factors were
found to have major influence on client H&S performance. The impact significance of
these factors was determined to be critical. The significance of this finding was that
these factors may be considered to be leading indicators of client H&S performance.
With the influence of these factors, clients were ‘likely’ to implement H&S elements.
Therefore a continued influence from external environment may entail a continued
implementation of H&S elements and therefore improve client H&S performance.
155
The social, technology and professional bodies’ factors, were found to have less
influence on client H&S performance than the political, economic and legislative
factors. The impact significance of these factors was considered to be ‘moderate’. The
significance of this finding was that the social, technology and professional bodies
factors were also necessary albeit their influence on client H&S performance being
moderate.
6.3.2
Objective D2
The second objective of the Delphi study was to estimate the influence of the factors
of client H&S culture on H&S consideration throughout a project lifecycle. This was
achieved by determining the impact significance of the factors of client H&S culture
on H&S consideration.
The level of influence of client H&S culture was inferred from the impact
significance on H&S consideration. Client H&S consideration was defined by four
factors namely:
1. commitment;
2. involvement;
3. competence; and
4. leadership.
In addition, the likelihood of H&S consideration as a result of client’s influence was
assessed. The likelihood of H&S consideration was assessed at the following
construction stages, namely the:
1. initiation;
2. design;
3. procurement;
4. construction;
5. commissioning (and testing);
6. close out; and
7. operations and maintenance stage.
156
The likelihood of H&S consideration at all construction stages as a result of client’s
influence was determined to be 83%. According to the five point likelihood scale with
ratings ranging from ‘not likely’ (0%) to ‘very likely to occur’ (100%) the value of
83% meant that H&S consideration at all construction stages was ‘very likely to
occur’ (Table 14). This finding seemed to indicate that the client’s influence was
necessary in order to assure H&S consideration. This finding confirmed that client
influence was necessary for H&S consideration to occur. Furthermore, the finding
also provided an estimate of the extent to which H&S consideration could be
influenced.
The factors of client H&S culture which had major influence on H&S consideration
were also identified. The commitment, involvement and competence factors were
found to have major influence on H&S consideration. The impact significance of
these factors was determined to be critical. The significance of this finding was that
these factors may be considered to be leading indicators of H&S consideration. With
the influence of these factors, H&S consideration was ‘very likely to occur’ at all
construction stages. Therefore a continued influence from a positive client H&S
culture may entail a continued consideration of H&S throughout all construction
stages and thereby lead to an improvement in project H&S performance.
According to Hodgson & Milford, (2005:5), H&S consideration is not done in many
construction projects. Therefore the finding suggested that getting clients involved,
committed and acquire the requisite knowledge may ensure that H&S is considered
throughout the project life cycle.
Although the leadership factor was found to be less influential than the commitment,
involvement and competence factors on H&S consideration, its impact was found to
be ‘major’. The finding seemed to suggest that leadership on H&S consideration was
not as critical as the commitment, involvement and competence factors were.
6.3.3
Objective D3
The third objective was to estimate the influence of the factors of client H&S culture
on contractor H&S performance. This was achieved by determining the impact
significance of the factors of client H&S culture on contractor H&S performance.
157
The level of influence of client H&S culture was inferred from the impact
significance on contractor H&S performance. Client H&S culture was defined by four
factors namely: commitment, involvement competence and leadership.
In addition, the likelihood of contractors implementing H&S elements as a result of
being influenced by client H&S culture was assessed. The H&S elements were for the
contractor to:
1. develop staff competence in H&S;
2. employ permanent H&S personnel;
3. consult and communicate information on H&S;
4. have effective H&S policies, procedures and goals;
5. involve workers in H&S management;
6. conduct H&S hazard identification and risk assessment; and
7. achieve top management commitment and involvement.
The likelihood of contractors implementing all of these H&S elements as a result of
the influence from client H&S culture was determined to be 83%. The likelihood
value of 83% meant that the implementation of H&S elements by contractors as a
result of client’s influence was ‘very likely to occur’ (Table 14). This finding seemed
to indicate that client H&S culture’s influence was necessary in order to assure that
contractors implemented these H&S elements or programmes. This finding not only
confirmed that client H&S culture was necessary for contractor H&S performance; it
also provided an estimate of the extent to which contractor H&S performance could
be influenced. An estimate of 83% likelihood was almost an assurance that
contractors would implement H&S elements.
The factors of client H&S culture which had major influence on contractor H&S
performance were also identified. The commitment, involvement, competence and
leadership factors were all found to have major influence on contractor H&S
performance. The impact significance of these factors on contractor H&S
performance was determined to be major. The impact significance ranged from 6.20
to 7.00.
158
The significance of this finding was that these factors may be considered to be leading
indicators of contractor H&S performance. Therefore, with the influence of these
factors, contractors were ‘very likely’ to implement H&S elements. Consequently, a
continued influence from a positive client H&S culture may entail a continued
implementation of H&S elements by contractors and therefore lead to an
improvement in their H&S performance.
Top management of contractors has been identified as being of critical importance to
H&S performance in the construction industry (Cooper, 2006:6; Zhou et al., 2011:94).
Therefore it is important that contractors (top management) are influenced to become
committed.
6.3.4
Objective D4
The fourth objective was to estimate the influence of client H&S culture on designer
H&S performance. This was achieved by determining the impact significance of the
factors of client H&S culture on designer H&S performance.
The level of influence of client H&S culture was inferred from the impact
significance on designer H&S performance. Client H&S culture was defined by four
factors namely: commitment, involvement competence and leadership.
In addition, the likelihood of designers implementing H&S elements as a result of
being influenced by client H&S culture was assessed. The H&S elements were for the
designer to:
1. develop staff competence in H&S;
2. employ permanent H&S staff;
3. conduct H&S inspections and audits;
4. have effective H&S policies, procedures and goals;
5. involve contractors in design reviews;
6. conduct H&S hazard identification and risk assessment; and
7. design for H&S.
Designers were ‘likely’ to implement the above elements as a result of client
influence. The likelihood was determined to be 78%. This finding seemed to indicate
that client H&S culture’s influence was necessary in order to assure that designers
159
implemented these H&S elements or programs. This finding not only confirmed that
client H&S culture was necessary for designer H&S performance; it also provided an
estimate of the extent to which designer H&S performance could be influenced. A
78% likelihood was almost an assurance that designers would implement the H&S
elements.
The involvement factor of client H&S culture had major influence on designer H&S
performance. The impact significance of this factor on designer H&S performance
was determined to be critical (7.31 impact significance). The commitment,
competence and leadership factors were considered to have major impact significance
on designer H&S performance. However, the leadership factor was found to be more
influential on designer H&S performance when compared to other client H&S factors.
The significance of this finding was that these factors may be considered to be leading
indicators of designer H&S performance. Therefore, with the influence of these
factors, designers were ‘very likely’ to implement the H&S elements. Consequently, a
continued influence from a positive client H&S culture may entail a continued
implementation of H&S elements by designers and therefore lead to an improvement
in their H&S performance.
The role of designers and their contribution to H&S improvement in the industry has
been highlighted in many studies (Gambatese et al., 1997; Smallwood, 2004;
Gambatese et al., 2005; Hecker et al., 2005; Kinnersley & Roelen, 2007). Designers
play a critical role in H&S performance improvement. Therefore, getting designers to
be involved and remain so, may require influence from client H&S culture.
6.3.5
Objective D5
The fifth objective of the Delphi study was to estimate the direct and indirect
influence of client H&S culture on the overall project H&S performance. This was
achieved by determining the impact significance of the factors of client H&S culture
on project H&S performance.
The indirect influence of client H&S culture on project H&S performance was found
to be greater than the direct one. The indirect influence of client H&S culture on
project H&S performance, mediated by contractor and designer H&S performance,
160
was considered to be critical. On the other hand, the direct influence of client H&S
culture was found to be less than the indirect influence, contractor and designer H&S
performance’s influence. However, the direct influence of client H&S culture on
project H&S performance was considered to be major.
This finding was in agreement with Yule & Mearns’ (2004) study, who concluded
that clients had influence on project H&S performance. However, the results also
show that clients’ influence on project H&S performance had greater impact when
mediated by either contractor or designer H&S performance. The finding seemed to
suggest that clients need all parties in order to achieve greater influence on H&S
performance.
In addition to the four factors of client H&S culture assessed during the Delphi study,
two factors were added. These were the fact that clients should have standard
procedures and effective communication. This finding was in agreement with
literature. Procedures and communication have been identified as factors of a H&S
culture (Gadd & Collins, 2002; Dingsdag et al., 2006).
6.4
CHAPTER SUMMARY
Findings from the Delphi study have been discussed in this chapter. The key findings
were that the external environment had a major influence on client H&S
performance/culture. Furthermore, client H&S culture had a major influence on
contractor, designer and project H&S performance.
These findings were in agreement and lend support to findings by other authors on the
influence of the external environment on client H&S culture as well as client H&S
culture’s influence on contractor, designer and project H&S performance.
The findings in this Delphi study and together with findings from literature formed
the basis of the conceptual model presented in chapter seven. The model postulates
that client H&S culture was critical to influence contractor, designer and the overall
project H&S performance. It was also further postulated that client H&S culture was
influenced by the prevailing environmental factors. Therefore the postulated model
was one that was centred on the client with the view to improve H&S performance.
161
CHAPTER SEVEN
THE CONCEPTUAL CLIENT-CENTRED CONSTRUCTION H&S
PERFORMANCE IMPROVEMENT MODEL
7.1
INTRODUCTION
Various improvement methods have been suggested to improve H&S performance in
the construction industry. However, it seems the most feasible way to improve H&S
performance in the industry is through culture change (Riley & Clare-Brown,
2001:150; Fitzgerald, 2005:324; Chinda & Mohamed, 2008:114). Despite general
agreement that H&S improvement may only be realised with an improvement in H&S
culture, the concept of culture, and in particular H&S culture, is still confusing. This
confusion has resulted in a myriad of definitions and measurement methods of the
concept. Despite these differences on what culture is and how it should be measured,
there is a general agreement on the efficacy of the concept to improve H&S
performance (Dingsdag et al., 2006; Molenaar et al., 2006; Chinda & Mohamed,
2008). It is in fact suggested that clients’ culture could offer an opportunity for
addressing the problem of H&S performance (Bomel, 2001:5.3). It has been
suggested that the impetus for change lies with the clients of construction projects
because clients can influence H&S performance.
Consequently, given that H&S performance may only be realised with an
improvement in H&S culture and that clients can influence both contractors and
designers on a construction project, and the fact that the impetus required for H&S
performance improvement can be provided by clients, the holistic model of H&S
performance improvement should be one that is based on cultural change driven by
construction clients. Therefore a client-centred model is proposed in this study. Based
both on the relevant literature and findings from the Delphi study, the model theorises
that contractor’s and designer’s H&S performance can be influenced by client’s H&S
culture and that this would lead to an improved project H&S culture and therefore
performance. Few models of H&S performance improvement have addressed the
162
unique and complex nature of the construction industry and therefore may not
effectively address the problem of H&S performance.
The model proposed by Lund & Aaro (2004), identifies factors such as economic,
legislative, aspects of the cultural, organisation and physical surroundings as being
critical to H&S improvement. This model was not developed to be used for the
construction industry. However, the factors mentioned in the model are applicable at
industry level. In the client-centred model, recognition is made that external
environment factors such as political, economic, social, technology and legislative,
have an impact on the H&S culture of all project stakeholders.
On the other hand, Zohar & Luria (2003) proposed a model that uses supervisory
practices to leverage the improvement of H&S behaviour of workers and ultimately
reduce accidents on sites. The model is based on the principle that modification
should be done to the value function for H&S behaviour by introducing short term
rewards that outweigh immediate costs. This is based on the ABC (antecedents –
behaviour- consequences) theory of behaviour based H&S. According to the authors,
behaviour modification principles are considered to be the most effective intervention
framework. However the authors recognise the shortcomings of behaviour based
safety and note that it does not take full account of unique attributes of the
organisational context as the onus for change lies with individual workers and the
reinforcement provided by the feedback and incentive is also directed at the workers
(Zohar & Luria 2003:568). However, focusing only on the workers for H&S
improvement is problematic because the efforts do not include immediate supervisors,
despite the primary role that supervisors play to influence subordinate behaviour. Line
supervisors continually provide the antecedents and consequences employed in the
behavioural safety interventions. Accordingly, Zohar & Luria (2003) proposed an
intervention framework that seeks to improve supervisory H&S practices. The model
is an organisational level model and is based on behaviour modification principles.
However even with the inclusion of supervisors being targeted for behaviour in order
to alter the behaviour of those at the front line, the model does not address the entire
context that affect behaviour which is the overall organisational culture. Behaviour is
a product of the prevailing culture in an organisation or industry (Guldenmund,
2000:251; Reason, 2000b:13). Therefore targeting behaviour of workers, or
163
supervisors is not adequate because behaviour (practices) is an output of culture and
therefore the target should be the organisational or industry culture.
The model client framework of Australia (Lingard et al., 2009), is a more practical
model for H&S performance improvement in the construction industry. The model
was specifically developed for the construction industry in Australia. The framework
could be adopted for use by project managers anywhere in the world. The aim of the
framework was to improve the H&S culture of the construction industry especially on
public funded construction projects. The framework is a set of booklets which contain
a number of instructions, checklists and pro-forms which can be used by government
agents to ensure that specific duties are performed at various levels of the project to
ensure that H&S is implemented. In undertaking these duties the client agents are also
charged with the responsibility to ensure that they demonstrate commitment to
developing a H&S culture, lead, develop relationships, promote H&S in planning and
design, consult and communicate, manage H&S risk hazards, measure and monitor,
and evaluate H&S performance.
However, because of the confusion surrounding the concept of H&S culture, the
model client framework risks being just another mechanical process and result in a
pile of paperwork without necessarily developing the required H&S culture as
intended. The framework does not outline how commitment to culture development
may be achieved. The model framework seems more of a technical approach rather
than a recognised culture change. Improvement of H&S culture lies in the assessment
of and addressing specific factors that specifically influence H&S culture such as
leadership, commitment, communication, involvement and competence (Human
Engineering, 2005:38). The concept of culture itself is intangible as it exists in a state
of opacity; therefore it needs a clear identification if it is to be applicable as an audit
issue (Karlsen & Valen, 2011:2). The model client framework does not indicate how
culture would be audited.
Culture change is the starting point and offers the best way to improve H&S in the
construction industry. Improving culture is more effective than increased supervision
or rigorous procedures to enhance H&S performance (Parker, Lawrie & Hudson,
2006:552). It was because of this and the lack of a model that seeks to improve H&S
in a holistic way that the client-centred model was proposed in this study. The model
164
postulates that cultural change in construction clients will lead to a better H&S
performance in construction projects.
Specifically, the model theorised that client H&S culture had influence on contractor,
designer and the overall project H&S performance. This relationship is illustrated in
Figure 35. It was further theorised that client H&S culture was a function of the
factors of the external environment namely: political, economic, social, technology
and legislative. From the Delphi study and the literature review, the factors of client
H&S culture were found to be, leadership, involvement, procedures, commitment,
communication and competence. These six factors of client H&S were collectively
referred to as LIP+3C.
It was also theorised that using a perception or climatic survey to establish the
practice of the factors of client H&S culture, contractor, designer and the overall
project H&S performance, a H&S culture of an organisation or a project could be
established. The perception or climate survey could expose the “way things are done”
or that “degree of effort” concerning the identified factors (Cooper, 2000:115; IOSH,
2004:7). In addition to these climate surveys, follow-up procedures such as focus
groups and interviews to supplement the climate surveys as a way of obtaining
feedback from the respondents and be able to understand the behaviour are
recommended (Wiegmann et al., 2002:13; IOSH, 2004:7).
Consequently, continuous improvement of H&S culture could typically follow
Deming’s plan, do, check and act cycle approach. Alternatively, the H&S maturity
model and principles of total quality management combined could also be used to
build a H&S culture by: assessing the current level, developing plans to improve,
implementing the plans, monitoring the implemented plans and reassessing the
implementation. The entire cycle is repeated for continuous improvement (IOSH,
2004:9)
165
7.2
CLIENT-CENTRED H&S IMPROVEMENT MODEL
Figure 35 is the conceptual client-centred construction H&S performance
improvement model. The postulated relationships in the conceptualised model were
that:
H1. the external environment has a direct positive influence on client H&S culture;
H2. client H&S culture has a direct positive influence on contractor H&S
performance;
H3. client H&S culture has a direct positive influence on designer H&S
performance;
H4. client H&S culture has a direct positive influence on project H&S
performance;
H5. contractor H&S performance has a direct positive influence on project H&S
performance;
H6. designer H&S performance has a direct positive influence on project H&S
performance
H7. client H&S culture, has an indirect positive influence on project H&S
performance mediated by contractor H&S performance; and
H8. client H&S culture, has an indirect positive influence on project H&S
performance mediated by designer H&S performance.
The conceptual model was validated in an alternative questionnaire survey. Results of
this questionnaire survey are presented in the next section.
From the theorised model in this study, it seemed easier and practical to view the
identified factors of client H&S culture namely; leadership, involvement, procedures,
commitment, communication and competence (LIP+3C), as action items. Continuous
assessment and control of these items could be used to improve the client H&S
culture and therefore H&S performance in a project. The LIP+3C could therefore be
used as a tool to check on the practice of these factors.
166
Consequently, the process of H&S performance improvement could simply follow the
following cycle:
1. Assess LIP +3C;
2. Develop strategies in consultation with stakeholders to enhance LIP + 3C;
3. Implement strategies;
4. Assess LIP +3C;
5. Act to improve LIP+3C;
6. Repeat process 1-5.
The above is based on Deming’s PDCA cycle and the same approach is suggested by
IOSH (2004:9).
Contractor
performanc
e
Client
culture
Environment
Project
performanc
e
Designers’
performanc
e
Figure 35: Theorised client centred H&S performance improvement
7.3
CHAPTER SUMMARY
In chapter seven, a conceptual model was theorised that client H&S culture had
influence on contractor, designer and the overall project H&S performance. It was
further theorised that client H&S culture was a function of the factors of the external
environment namely: political, economic, social, technology and legislative. From the
Delphi study and the literature review, the factors of client H&S culture were found to
be,
leadership,
involvement,
procedures,
commitment,
communication
and
competence. These six factors of client H&S were collectively referred to as LIP+3C.
Findings from the validation of the conceptual model developed in chapter seven will
now be presented in chapter eight.
167
CHAPTER EIGHT
QUESTIONNAIRE SURVEY RESULTS
8.1
INTRODUCTION
The postulated sub-models are diagrammatically shown in Figure 36 and 37. Indicator
variables for the exogenous variables of external environment and client H&S culture,
and the endogenous variables of designer, contractor and project H&S performance
are shown in each measurement model. The theory behind the hypothesised clientcentred H&S performance improvement model was based on literature and on
experts’ opinion acquired during the Delphi study described and presented in chapter
six of this study.
Raw data from questionnaires were entered into the Statistical Package for Social
Sciences (SPSS) software and was later exported to the structural equation modelling
(SEM) software EQS version 6.1 for analysis (Bentler, 1999). The sample used for the
analysis of the model was 281 cases. According to Kline (2005:15), a sample size of
281 is classified as large. A small sample of less than 100 cases tends to be
problematic when it comes to structural equation model analysis (Kline, 2005:15).
Figure 36: Hypothesised model-Client influence on project H&S performance
168
Leadership
Procedures
External
Environment
Commitment
Communicatio
n
Competence
Figure 37: Environmental influence on client H&S culture
169
Client H&S culture
Involvement
8.2
STATISTICS ON SEM ASSUMPTIONS
8.2.1
Outliers and Missing data
Inspection of data sets revealed that some data sets had missing values. A detailed
examination of the patterns of missing data led to the conclusion that the missing data
was missing at random (MAR) as opposed to data missing completely at random
(MCAR). According to McDonald & Ho (2002:70), the condition that data was
missing completely at random is a situation where the presence or absence of the
observation is independent of other observed variables and the variable itself.
Arguably, the condition MCAR is a very strict assumption that may be difficult to
justify in practice (McDonald & Ho, 2002:70). Therefore the assumption of the
condition MAR was adopted. Consequently, the robust maximum likelihood
estimation solution in EQS was used to address the problem. The method of
maximum likelihood has produced better results compared to other methods
(Boomsma, 2000:469; Kline, 2005:56). The assumption in this method was that the
means, variances and covariances were sufficient statistics. Therefore, cases with
missing variables were skipped and not included in the analysis.
Similarly, examination of the analysis output revealed that there were a few outliers in
the data. EQS result output includes case numbers with the largest contribution to
Mardia’s normalised multivariate Kurtosis. Examination of these case numbers
revealed the case numbers that included outliers and it was upon these examinations
that the conclusion was reached that there were a few outliers in the data. The chosen
method of estimation namely, robust maximum likelihood was adequate to address
the problem of outliers. The method is reliable because it replaces ordinary sample
covariances with the robust estimates of the covariances (Boomsma, 2000:469).
8.2.2
Distribution characteristics of the data
The estimation method of maximum likelihood assumes multivariate normality.
Therefore it was very important that the distribution characteristics of the data were
established before model analysis could be undertaken.
The EQS result output included univariate statistics such as mean, Skewness (G1),
Kurtosis (G2) and the standard deviations. In addition, the multivariate kurtosis
formed part of the result output. Examination of the univariate statistics and Mardia
170
based multivariate Kurtosis suggested non-normality in the sample. All Mardia
estimates of multivariate Kurtosis were greater than the upper limit value of 3.0
(DeCarlo, 1997:292). Therefore the distributions were described as highly Kurtotic.
Mardia’s coefficient and other univariate statistics are presented in Table 23.
The non-normality of the data led to the adoption of the robust maximum likelihood
estimation method of the postulated model.
8.2.3
Identifiability of the model
A further requirement for SEM analysis is the Identifiability of the model. In order for
a model to be analysed, it has to fulfil the conditions of model identification.
Boomsma (2000:466) argue that it is the duty of a researcher to examine whether a
model is theoretically identified or not. Kline (2005:105) explains that a model is said
to be identified if it is theoretically possible to derive a unique estimate of each
parameter. Although the identification test result is presented in the result section,
identification is a property of the model and not the data. However, checking for
model identification is a requirement before model analysis could be done. Kline
(2005:105) contends that a model that is not identified remains so no matter the
sample size and any attempts to analyse it may prove fruitless.
A model is said to be identified if there are at least as many observations as free
model parameters (namely, the degree of freedom
0) and that every unobserved
variable must be assigned a scale (Kline, 2005:105). According to Byrne (2006:31), a
model could be just-identified, over-identified or under-identified. It is desirable to
have an over-identified model. Byrne (2006:31) explains that an over-identified
model is one in which the number of parameters to be estimated is less than the
number of data variances and covariances of the observed variables and therefore
results in a positive degree of freedom.
171
Table 23: Univariate and Mardia’s normalised Multivariate estimates
Factor
Indicator Mean Skewness Kurtosis
SD
Mardia’s
variable
(G1)
(G2)
coefficient
CLLP1
3.9683 -0.9772
0.7028 0.8303
CLLP2
3.9792 -1.0170
0.9037 0.8693
CLLP3
3.8956 -0.8988
0.5920 0.8989
CLIP1
3.5446 -0.6839
-0.0601 0.9804
CLIP2
3.6209
7.3232
92.5874 1.6082
CLIP3
3.5330 -0.6424
-0.1771 1.0329
CLPP1
3.7220 -0.5064
-0.5933 0.9272
CLPP2
3.7030 -0.6105
-0.0474 0.8911
CLPP3
3.6790 -0.4736
-0.3164 0.9547
Client H&S
CLTP1
3.9573
4.5314
54.9245 1.2022 443.7814
culture
CLTP2
3.8889 -1.1463
1.0599 0.9067
CLTP3
3.7393 -0.9150
0.2673 0.9629
CLNP1 3.7279 -0.6784
-0.3974 0.9696
CLNP2 3.7040 -0.6742
-0.3112 0.9574
CLNP3 3.7322 -0.6171
-0.3804 0.9387
CLCP1
3.9188 -1.1509
1.2552 0.9588
CLCP2
3.6875 -0.5896
-0.4836 1.1009
CLCP3
3.7279 -0.7941
-0.1483 1.0864
CLCP4
3.7610 -0.7443
-0.4131 1.1517
COPP1 3.8944 -0.9701
0.4788 0.9310
COPP2 3.8694 -0.8651
0.1425 0.9726
COPP3 3.9158 -1.0083
0.7937 0.9064
COTP1 3.9866 -1.1079
1.3724 0.8148
Contractor H&S
COTP2 3.8913 -1.0482
1.0960 0.8466 41.0290
Performance
COTP3 3.9377 -1.1069
1.1619 0.8790
CONP1 3.9081 -1.0392
0.8474 0.8957
CONP2 3.8407 -0.8506
0.4371 0.8859
CONP3 3.8413 -0.8920
0.7397 0.8767
DEPP1 3.7466
4.5422
52.5736 1.2351
DEPP2 3.7011 -0.6495
0.1887 0.8969
DEPP3 3.6734 -0.6189
0.0158 0.9400
DETP1
3.8702 -0.7208
0.4373 0.8401
Designer H&S
DETP2
3.8007 -0.5168
-0.1803 0.8479 225.6381
Performance
DETP3
3.8438 -0.6010
0.0438 0.8425
DENP1 3.8063 -0.8180
0.2194 0.9427
DENP2 3.7620 -0.7117
0.0073 0.9952
DENP3 3.6943 -0.7156
0.0501 0.9749
LGN
4.0219 -1.1863
1.0521 0.9942
ECON
3.5600 -0.5801
-0.4881 1.1136
External
SOC
3.4618 -0.5603
-0.6029 1.1689
47.9571
environment
PRI1
3.4854 -0.4408
-0.7585 1.1743
PRI2
3.6703 -0.7003
-0.2381 1.0612
TECH
3.6473 -0.6705
-0.1495 1.0718
172
The significance of model over-identification is that it allows for a model to be
rejected and therefore rendering it of scientific value (Byrne, 2006:31). A justidentified model cannot be rejected and it is impossible to obtain a solution for an
under-identified model. Examination of the EQS results revealed that the lowest value
for the degree of freedom in the current study was 2 and the highest was 35. These
values indicated a positive value of degree of freedom and therefore were suggestive
of an over-identified model.
There were no convergence problems because all parameter estimates for client H&S
culture, designer H&S performance, contractor H&S performance and project H&S
performance constructs stabilised at fewer than 10 iterations each. The desired
circumstance is the situation whereby only a few iterations are needed to reach
convergence and these should not exceed the default value of 30 (Byrne, 2006:102).
According to Byrne (2006:102) the number of iterations exceeding 30, results in nonconvergence and the output may not be trusted.
8.3
FIT STATISTICS ON MEASUREMENT MODELS (CFA)
8.3.1
Client H&S culture
The number of cases that were analysed for the client H&S culture subscale was 273
cases from a sample of 281. Eight cases were skipped because of missing variables.
The client H&S culture subscale had 19 dependent variables, 25 independent
variables and 53 free parameters. The number of fixed non-zero parameters was 25.
The following hypotheses for the client H&S culture subscale were tested. These
hypotheses, diagrammatically presented in Figure 38 described the relationships
between the indicator variables and the client culture construct. The model postulates
that client H&S:
1. culture may be explained by six factors namely: leadership, involvement,
procedures, commitment, competence and communication ( LIP+3C model);
2. leadership factor is explained by indicator variables CLLP1 to 3;
3. commitment factor is explained by indicator variables CLTP1 to 3;
4. involvement factor is explained by indicator variables CLIP1 to 3;
173
5. communication factor is explained by indicator variables CLNP1 to 3;
6. competence factor is explained by indicator variables CLCP1 to 3; and
7. procedures factor is explained by indicator variables CLPP1 to 3.
These indicator variables are presented in Table 24. In order to establish how well the
model fit the sample data and the strength of the hypothesised relations between
variables, results presented on residual covariance matrix, distribution of standardised
residuals, goodness-of-fit statistics and parameter estimates’ statistical significance at
probability level of 5% were examined.
In addition the Cronbach’s alpha and the Rho coefficient of internal consistency were
examined for score reliability. The construct validity of the measurement model was
determined from the convergent validity and the magnitude of parameter coefficients.
Results of the above analysis are presented in this section.
Figure 38: Theorised six factor client H&S culture model
174
Reliability results are presented for each factor of the client H&S culture construct.
Indicator variables were analysed in parcels in order to mitigate the possibility of a
type I error. In certain instances models do not fit simply because there are too many
indicator variables and one of the solutions is to analyse the indicator variables in
parcels. When indicator variables are in parcels, the number of variables to be
analysed is reduced by grouping uni-dimensional indicator variables together. Table
24 presents all indicator variables for each factor, un-parcelled, as they were presented
in the questionnaires.
8.3.1.1 Diagnostic fit analysis –Residuals
Average absolute residual values of the client H&S culture are presented in Table 25.
Both unstandardised and standardised average absolute residuals are presented.
Results revealed that all the absolute residuals and the average off-diagonal absolute
residual values both unstandardised and standardised were very close to zero. The
unstandardised average off-diagonal residual was 0.0200 while the standardised
average off diagonal residual was 0.0197. Byrne (2006:94) suggests that a value can
be said to be large if it is greater than 2.58. Therefore, since the values in Table 25
were all less than 2.58, results were suggestive of an acceptable fit to the sample data.
In addition, 100% of standardised residuals fell between -0.1 and +0.1
The significance of this distribution is that for a model to be described as a wellfitting model, the distribution of standardised residuals should be symmetrical and
centred around the zero (Byrne, 2006:94). From this information, results suggested a
model that was well fitting albeit minimal discrepancy in fit between the hypothesised
model and the sample data. Since the initial examination indicated a good fit, further
tests of goodness-of-fit were possible to conclusively make a decision on the fit and
appropriateness of the model.
175
Table 24: Postulated model variables for client culture
Latent variable
Indicator variables
(factor)
(The project client…
Leadership
considers H&S implications before making decisions on
the project
has an effective H&S policy
monitors H&S on the project throughout all stages
monitors designers’ H&S implementation
monitors contractor’s H&S implementation
mandated designers to manage project H&S
requires that the contractor manages project H&S
coordinates designers & contractor to ensure good H&S
Commitment
demonstrated positive attitude toward H&S
actively promoted H&S in a consistent manner across all
levels
provided finance for H&S
supported implementation of H&S activities
put in efforts to ensure every aspect of work & operations
are routinely evaluated for H&S
conducted regular H&S tours on the project
has been involved in investigations of accidents, incidents
& ill-health on the project
set H&S as an important agenda item in every project
progress meeting
set H&S as a No.1 priority on the project
Involvement
is personally active in critical project H&S activities
is always present in project H&S meetings
contributes to H&S training
is active in overseeing of H&S on critical operations
has constantly stayed “in-touch” on H&S issues
always communicates information on H&S to all parties
conducts regular audits & inspections
Communication has set up a formal reporting system of incidents &
accidents on the project
involved all parties in planning for H&S on the project
involves all parties in H&S review
has provided timely feedback on reported accidents &
incidents on the project
communicates risk findings to all parties on the project
clearly made H&S policy statements for the project
has clearly outlined H&S roles & responsibilities for all
parties on the project
has clearly communicated the expected performance on
H&S to all stakeholders
has provided information on H&S risk control to all parties
176
Label
CLLP1
CLLP2
CLLP3
CLLP1
CLLP2
CLLP3
CLLP1
CLLP2
CLTP1
CLTP2
CLTP3
CLTP1
CLTP2
CLTP3
CLTP1
CLTP2
CLTP3
CLIP1
CLIP2
CLIP3
CLIP1
CLIP2
CLIP3
CLIP1
CLNP1
CLNP2
CLNP3
CLNP1
CLNP2
CLNP3
CLNP1
CLNP2
CLNP3
Table 24: (Continued)
Competence
representatives have demonstrated knowledge of H&S
conducts H&S training for its own staff
deployed staff on the project that are qualified to manage
H&S
ensured that H&S induction to client staff was done on the
project
Procedures
has programmes to monitor and analyse H&S
implementation
has clear project H&S goals
scheduled H&S as a key contract prequalification criteria
for all parties involved in the project
scheduled H&S in all contracts for the parties involved in
the project
conducts regular H&S Performance measurement
has its own H&S committee
conducts hazard identification & risk assessment
required that designers adequately address H&S in their
designs
Table 25: Average absolute residuals for measurement models
Variable
Unstandardised
Standardised
Ave
Ave offAve
Ave offabsolute
diagonal
absolute
diagonal
residual
absolute
residual
absolute
residual
residual
Client
H&S
0.0180
0.0200
0.0177
0.0197
culture
Contractor
0.0049
0.0062
0.0062
0.0078
H&S
Performance
Designer H&S
0.0059
0.0079
0.0066
0.0083
Performance
Project H&S
0.0377
0.0424
0.0345
0.0388
Performance
External
0.0289
0.0405
0.0241
0.0337
environment
177
CLCP1
CLCP2
CLCP3
CLCP4
CLPP1
CLPP2
CLPP3
CLPP1
CLPP2
CLPP3
CLPP1
CLPP2
% falling
between
100
100
100
94.77
100
8.3.1.2 Goodness-of-fit statistics - Robust maximum likelihood method (RML)
A two statistic strategy of fit indexes was adopted and is reported on in this study. The
robust comparative/incremental index of Comparative Fit Index (CFI) and the robust
absolute fit index of root mean square error of approximation (RMSEA) at 90%
confidence interval were evaluated in order to establish the fit of the CFA models and
are reported on. In addition, the Satorra-Bentler scaled chi-square (
) and the
standard root mean squared residual (SRMR) were evaluated in order to compliment
the conclusion on model fit.
The sample data on the six factor client H&S culture model yield the
of
219.323 with 137 degrees of freedom (N=273; P=0.00001). The chi-square was
significant. This chi-square value indicated that the postulated model significantly
differed from the sample data. However the chi-square test is very sensitive to sample
size and therefore not very reliable. Kline (2005:136) observed that the chi-square test
tends to be affected by the sample size with a propensity to reject models if the
samples are large. Therefore a normed Chi-square value is usually adopted by most
researchers (Kline, 2005:137). Normed chi-square is the procedure of dividing the
chi-square by the degrees of freedom. The normed values of up to 3.0 and even 5.0
are recommended. From the above chi-square and degrees of freedom values the ratio
was found to be 1.60. This ratio was much lower than the upper limit of 3.0 or 5.0
advocated for by some authors (Kline, 2005:137) and therefore the model fit may be
described as acceptable.
In addition, the CFI was found to be 0.979 and the RMSEA with 90% confidence
interval (lower bound value = 0.035 and upper bound value = 0.058) was found to be
0.047. The CFI value was higher than the lower cut-off limit of 0.95 for a model to be
described as good fit. Equally, the RMSEA value of 0.047 was less than the upper
cut-off value of 0.05 for a good fit model. The absolute fit index SRMR was found to
be 0.025. The SRMR index also met the cut-off criteria of not exceeding 0.05 for a
good fitting model. These fit indexes for the client H&S culture measurement model
were therefore suggestive of a good fit overall (Table 26).
Additionally, parameter estimates were scrutinised to determine whether the model
worked properly and was reasonable. This involved assessing the magnitude, signs
178
and statistical significance of the parameter estimates. These statistics are presented in
Table 27.
Table 26: Robust fit indexes for client H&S culture construct
Fit Index
Cut-off value
Estimate
Df
0
CFI
0.9 acceptable
Comment
219.323
137
Acceptable
0.979
Good fit
0.025
Good fit
0.047
Good fit
0.035:0.058
Acceptable range
0.95 Good fit
SRMR
0.08 acceptable
0.05 Good fit
RMSEA
0.08 acceptable
0.05 Good fit
RMSEA 90% CI
8.3.1.3 Statistical significance of parameter estimates
Raykov et al. (1991:501) recommended that further examination of factor loadings
(parameter coefficients), standard errors and the test statistics should be conducted in
addition to the analysis of fit statistics before conclusions could be made about the
appropriateness of the postulated models. Therefore these estimates were examined
and are presented in this section.
According to Byrne (2006:94) estimates are said to be unreasonable if they have
correlation values greater than 1.00, have negative variances and the correlation or
covariances are not definite positive. Furthermore, the test statistic need to be greater
than 1.96 based on the probability level of 5% before the hypothesis can be rejected
(Byrne, 2006:103). The test statistic reported in this study is the parameter estimate
(factor loading) divided by its standard error and therefore it functions as a Z-statistic
to test that the estimate is statistically different from zero. The coefficient was
therefore referred to as the Z-statistic.
Inspection of the correlation values, standard errors and the test statistic presented in
Table 27 reveal that all correlation values were not greater than 1.00; all test statistics
179
were greater than 1.96 and the signs were appropriate (positive). The estimates were
therefore reasonable as well as statistically significant.
The parameter with the highest standardised coefficient was found to be the parcel
indicator variable CLPP1. The parameter coefficient was 0.961. Individually, the
indicator variable, induction, explained the factor of client competence better with a
factor loading of 0.997. All parameter estimates had high correlation values close to
1.00 suggesting a high degree of linear association between the indicator variables
and the latent variables. In addition, the
values were close to the desired value of
1.00 indicating that the exogenous variables explained more of the variance in the
indicator variables. However the parcel CLIP2 was the only exception with an
of
0.332 (Table 27).
Equally, Correlation values between the six factors of the client H&S culture
construct indicated that the factors were related to each other. The correlation values
ranged from 0.653 (Involvement– Leadership) to 0.942 (Commitment - Involvement).
Table 29 shows exogenous variable correlations. However, inspection of the
individual indicator variables postulated to measure the exogenous variable,
involvement, in Table 28, revealed that all the variables were close to 1.00. The lower
value for the parcel CLIP2 may probably have resulted from the combination of
the indicator variables making up the parcel CLIP2. The test statistic and the standard
errors for the overall client H&S culture also revealed that the estimates were
reasonable and statistically significant. The covariances among independent variables
at 5% level were also found to be statistically significant.
8.3.1.4 Internal reliability and validity of scores
In order to determine the internal consistency of the composite of the measurement
models the Rho coefficient and the Cronbach’s alpha coefficient were examined to
establish reliability (Byrne, 2006:133). According to Kline (2005:59), the reliability
coefficient should fall between zero and 1.00. Values close to 1.00 are desired. The
Rho coefficient of internal consistency was found to be 0.978. This value was above
the minimum value of 0.70. In addition, the Cronbach’s alpha was also found to be
above the minimum 0.70 at 0.963. Both those values showed a high level of internal
consistency and therefore reliability.
180
Table 27: Parameter estimates of client H&S culture measurement model
Indicator
Variable
(Parcelled)
Un-standardised
Coefficient (λ)
Standardised
Coefficient (λ)
Z- Statistic
CLLP1
0.739
0.892
16.788
0.795
Yes
CLLP2
0.816
0.940
16.817
0.884
Yes
CLLP3
0.765
0.853
14.758
0.727
Yes
CLIP1
0.913
0.933
19.736
0.871
Yes
CLIP2
0.928
0.578
18.874
0.334
Yes
CLIP3
0.949
0.921
20.386
0.848
Yes
CLPP1
0.888
0.961
25.153
0.924
Yes
CLPP2
0.837
0.943
20.561
0.889
Yes
CLPP3
0.848
0.891
19.029
0.794
Yes
CLTP1
0.836
0.697
15.097
0.486
Yes
CLTP2
0.841
0.929
16.426
0.863
Yes
CLTP3
0.900
0.937
18.794
0.878
Yes
CLNP1
0.904
0.935
21.869
0.874
Yes
CLNP2
0.917
0.960
22.777
0.922
Yes
CLNP3
0.895
0.956
22.866
0.914
Yes
CLCE1
0.780
0.816
12.765
0.666
Yes
CLCE2
0.966
0.880
20.101
0.774
Yes
CLCE3
0.971
0.897
18.495
0.804
Yes
CLCE4
1.000
0.870
18.857
0.758
Yes
(Robust statistical significance at 5% level (parcelled)
181
Significant at
5% level?
Table 28: Factor loading and Z-statistics of client model
Latent variable
Indicator variable
Parameter
Coefficient
(λ)
Z- Statistic
Significant?
.741
.796
.816
.841
.792
.717
.618
.798
.839
.806
.758
.775
.942
11.011
13.209
14.657
17.700
13.641
11.047
8.759
14.022
10.628
13.316
11.307
12.254
17.663
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
.956
.887
17.884
15.811
Yes
Yes
.799
.957
.829
.919
.934
.934
.987
.962
13.116
17.347
12.956
12.653
17.308
18.362
19.825
18.066
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
(The client...)
Leadership
Commitment
Involvement
considers H&S implications before making decisions on the project
has an effective H&S policy
monitors H&S on the project throughout all stages
monitors designers’ H&S implementation
monitors contractor’s H&S implementation
mandated designers to manage project H&S
requires that the contractor manages project H&S
coordinates designers & contractor to ensure good H&S
demonstrated positive attitude toward H&S
actively promoted H&S in a consistent manner across all levels
provided finance for H&S
supported implementation of H&S activities
put in efforts to ensure every aspect of work & operations are routinely
evaluated for H&S
conducted regular H&S tours on the project
been involved in investigations of accidents, incidents & ill-health on the
project
set H&S as an important agenda item in every project progress meeting
set H&S as a No.1 priority on the project
is personally active in critical project H&S activities
is always present in project H&S meetings
contributes to H&S training
is active in overseeing of H&S on critical operations
has constantly stayed “in-touch” on H&S issues
always communicates information on H&S to all parties
182
Table 28: (Continued)
Communication
Competence
Procedures
conducts regular audits & inspections
has set up a formal reporting system of incidents & accidents on the project
involved all parties in planning for H&S on the project
involves all parties in H&S review
has provided timely feedback on reported accidents & incidents on the
project
communicates risk findings to all parties on the project
clearly made H&S policy statements for the project
has clearly outlined H&S roles & responsibilities for all parties on the project
has clearly communicated expected performance on H&S to all
has provided Information on H&S risk control to all parties
representatives have demonstrated knowledge of H&S
conducts H&S training for its own staff
deployed staff on the project that are qualified to manage H&S
ensured that H&S induction to client staff was done on the project
.954
.902
.865
.820
.919
17.972
16.644
16.419
17.241
18.530
Yes
Yes
Yes
Yes
Yes
.977
.894
.895
.890
.965
.785
.976
.963
.997
21.252
17.517
18.733
16.611
18.807
13.082
20.689
18.160
18.883
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
has programmes to monitor and analyse H&S implementation
has clear project H&S goals
scheduled H&S as a key contract prequalification criteria for all parties
involved in the project
scheduled H&S in all contracts for the parties involved in the project
conducts regular H&S Performance measurement
has its own H&S committee
conducts Hazard identification & risk assessment
required that designers adequately address H&S in their designs
.908
.827
.839
19.753
15.959
16.215
Yes
Yes
Yes
.770
.945
.839
.963
.698
16.205
20.996
13.471
19.335
11.397
Yes
Yes
Yes
Yes
Yes
(Robust statistical significance at 5% level (un-parcelled)
183
The construct validity was determined by examining the convergent validity of the
factors of client H&S culture construct and from the magnitude of the parameter
coefficients (factor loadings). High correlation values indicate convergence and
parameter coefficients greater than 0.5 indicate close relation between the factor and
an indicator variable. A parameter coefficient of 0.5 is interpreted as 25% of the total
variance in the indicator variable being explained by the latent variable (factor).
Accordingly a parameter coefficient has to be greater than 0.7 to explain about 50%
of the variance in an indicator variable (Hair et al., 1998:111). The relationship
between exogenous variables should show high correlations in order to indicate
convergence. Inspection of unstandardised parameter coefficients presented in Table
27, revealed that they were sufficiently high with the minimum of 0.739 for the
relationship between CLLP1 and the factor leadership. The parameter estimate of
0.739 meant that the leadership factor accounted for about 55% of the variance in the
indicator variable, CLLP1 and therefore indicative of a good fit between the two
variables.
Additionally, the correlation values between the six factors postulated to reflect client
H&S culture were high and therefore confirmed convergence. Convergence is said to
have been reached when measures of constructs that theoretically should be related to
each other are, in fact, observed to be related to each other indicated by higher
correlation values. Table 29, shows correlation values between the six factors of client
H&S culture. All correlation values were above 0.4 with the minimum correlation
value being 0.653.
Therefore, the client H&S culture construct satisfied both the internal reliability
criteria and the construct validity criteria. The Rho value was above the minimum
value of 0.70 (Table 30) and the convergent validity characterised by high correlation
values, magnitude, sign and statistical significance of parameter coefficients were
found to be satisfactory.
A further check for model mis-specification was conducted. The model was checked
against mis-specification by examining results from the Lagrange Multiplier Test (LM
test). In EQS, a model can be said to be mis-specified if there are any mis-fitting
parameters using a LM test (Byrne, 2006:113). The criterion is to identify any
significant drop in the
values of parameters. Additionally, in the univariate and
184
multivariate analysis, the probability that a parameter estimate is equal to zero should
be less than 0.05 in order to be rejected. This is also an indication of mis-specification
(Byrne, 2006:113). Inspection of the LM test output revealed that there were no
significant mis-fitting variables that would have warranted model re-specification.
Table 29: Correlations between factors of client H&S culture
Client H&S culture
factors
Leadership (CLLP)
Involvement (CLIP)
Procedures (CLPP)
Commitment (CLT)
Communication
(CLNP)
Competence (CLCP)
CLLP
CLIP
CLPP
CLTP
CLNP
1.000
0.653
0.813
0.691
0.821
1.000
0.709
0.942
0.749
1.000
0.682
0.841
1.000
0.721
1.000
0.719
0.734
0.819
0.746
0.780
CLCP
1.000
Table 30: Reliability and construct validity of client H&S culture model
Factor
Client H&S
culture
No. of indicator
variables/parcels
Indicator
variable
Factor
loading
Cronbach’s
Alpha
Rho
coefficient
19
CLLP1
CLLP2
CLLP3
CLIP1
CLIP2
CLIP3
CLPP1
CLPP2
CLPP3
CLTP1
CLTP2
CLTP3
CLNP1
CLNP2
CLNP3
CLCE1
CLCE2
CLCE3
CLCE4
0.892
0.940
0.853
0.933
0.578
0.921
0.961
0.943
0.891
0.697
0.929
0.937
0.935
0.960
0.956
0.816
0.880
0.897
0.870
0.963
0.978
* Parameter estimates are based on standardised solutions
185
8.3.1.5 Solution evaluation and summary on client H&S culture
The residual covariance estimates fell within the acceptable range, the robust fit
indexes met the cut-off index criteria and all parameter estimates were found to be
statistically significant and feasible. Having these criteria, the measurement model for
the client H&S culture subscale fit the sample data well. As a result, there was no
need to modify and re-specify the model before including it in the full latent variable
model. Furthermore, there was no significant evidence that indicated model misspecification and the LM test supported the conclusion that the measurement model
for client H&S culture subscale fit the sample data well. Apart from this, as Byrne
(2006:113) explains, it is unwise to modify a model that fits well because
modifications may only be fitting small characteristic features of the sample.
Therefore, the hypotheses could not be rejected that client H&S:
1. culture was explained by six factors namely: leadership, involvement,
procedures, commitment, competence and communication (LIP+3C model);
2. leadership factor is explained by indicator variables CLLP1 to 3
3. commitment factor is explained by indicator variables CLTP1 to 3
4. involvement factor is explained by indicator variables CLIP1 to 3
5. communication factor is explained by indicator variables CLNP1 to 3
6. competence factor is explained by indicator variables CLCP1 to 3
7. procedures factor is explained by indicator variables CLPP1 to 3
Therefore, the six factor client H&S culture construct, comprising: leadership,
involvement, procedures commitment, communication and competence, adequately
described the client H&S culture framework and the indicator variables adequately
measured these exogenous factors. Therefore the measurement model could be used
in the analysis of the full latent variable model.
8.3.2
Project H&S Performance
The number of cases that were analysed for the project H&S performance subscale
was 272 cases. The number of cases that were skipped was nine because they had
missing variables. The project H&S subscale had 17 dependent variables, 18
independent variables and 34 free parameters. The number of fixed non-zero
parameters was 18.
186
The hypothesis that the project H&S performance construct is explained by indicator
variables PP1 to PP17 shown in Table 31, was evaluated. This hypothesis was also
presented diagrammatically in Figure 17. The project H&S performance construct,
was a measurement model which was part of the overall structural equation model.
In order to establish how well the model fit the sample data and the strength of the
hypothesised relations between variables, results presented on residual covariance
matrix, distribution of standardised residuals, fit statistics and statistical significance
at probability level of 5% were examined. In addition, the Cronbach’s alpha and the
Rho coefficient of internal consistency were examined for score reliability. The
construct validity of the measurement model was determined from the convergent
validity and the magnitude of parameter coefficients.
Table 31: Indicator variables for the project H&S Performance construct
Latent
Variable
(factor)
Project H&S
Performance
Indicator variable
(On the project......)
label
there is an effective H&S policy
H&S programmes are regularly monitored
H&S implementation is well coordinated between all parties
all key project participants attend H&S meetings
regular H&S inspections / audits are conducted
H&S is well addressed in all project contract documents
H&S committee is effective
there is a permanent H&S officer
H&S requirements are clearly stated on drawings /
specifications
there is a H&S risk register
there is an incentive programme for H&S performance
H&S is a major agenda item in all project meetings
all accidents/ incidents are reported by workers
all parties know what the H&S risks are on the project
H&S induction is conducted for all workers on the project
formal H&S training is conducted for all workers on the project
H&S skill & knowledge is a key consideration for all staff
recruitment
PP1
PP2
PP3
PP4
PP5
PP6
PP7
PP8
PP9
187
PP10
PP11
PP12
PP13
PP14
PP15
PP16
PP17
8.3.2.1 Diagnostic fit analysis –Residuals
Average absolute residual values of the Project H&S performance are presented in
Table 25. The average absolute residuals, both unstandardised and standardised, are
presented. Results revealed that all the absolute residual values and the average offdiagonal absolute residual were close to zero. The unstandardised average off
diagonal residual was 0.0424 while the standardised average off diagonal residual was
found to be 0.0388. These values were considered to be very small and therefore
acceptable. An absolute residual value is considered to be large if it is more than 2.58
(Byrne, 2006:94). The results obtained for project H&S performance measurement
model were therefore suggestive of an acceptable fit to the sample data since all
residual values were below the 2.58. In addition, 94.77% of standardised residuals fell
between -0.1 and +0.1 which is the acceptable range.
In order for a model to be described as well-fitting, the distribution of standardised
residuals should be symmetrical and centred around the zero (Byrne, 2006:94).
Results suggested a model that was well fitting albeit minimal discrepancy in fit
between the hypothesised model and the sample data. About 5.23% of the residuals
fell outside the -0.1 to 0.1 range.
Therefore, since the above initial examination of residuals indicated a fairly good fit,
further tests of goodness-of-fit were possible to conclusively make a decision on the
fit and appropriateness of the model. The goodness-of-fit test results will now be
presented in the next sections.
8.3.2.3 Goodness-of-fit statistics - RML
A two statistic strategy of fit indexes is reported.
The sample data on project H&S performance measurement model yield the
of 294.515 with 119 degrees of freedom (N=272; P=0.0000). The chi-square was
significant. This chi-square value indicated that the postulated model significantly
differed from the sample data. However, the chi-square test is very sensitive to sample
size and therefore not very reliable. The chi-square test tends to be affected by the
sample size with a propensity to reject models if the samples are large. Therefore a
normed Chi-square value is usually adopted by most researchers (Kline, 2005:137).
Normed chi-square is the procedure of dividing the chi-square by the degrees of
188
freedom. The normed values of up to 3.0 or even 5.0 are recommended. From the
above chi-square and degrees of freedom values the ratio was found to be 2.475. This
ratio was lower than the upper limit of 3.0 or 5.0 advocated for by some authors
(Kline, 2005:137) and therefore the model fit may be described as acceptable.
In addition, the CFI was found to be 0.923 and the RMSEA with 90% confidence
interval (lower bound value = 0.064 and upper bound value = 0.086) was found to be
0.075. The CFI value was lower than the cut-off limit of 0.95 for a model to be
described as having a good fit. However the value was within the acceptable range to
accept the model. The lower limit is 0.9. Likewise, the RMSEA value of 0.075 was
less than the upper cut-off value of 0.08 for an acceptable fit. The model could be
accepted although the fit was not good. The absolute fit index SRMR was found to be
0.048. The SRMR index met the cut-off criteria for good fit. The SRMR cut-off value
is 0.5. These fit indexes (Table 32) for the project H&S performance measurement
model suggested that the measurement model had an adequate fit to the sample data.
Further, the fit statistics indicated that the model was working properly and could be
included in the full latent variable model analysis.
Additionally, parameter estimates were scrutinised to determine whether the model
worked properly and was reasonable. This involved assessing the magnitude, signs
and statistical significance of the parameter estimates. These statistics are presented in
Table 33.
Table 32: Robust fit indexes for project H&S Performance construct
Fit Index
Df
CFI
SRMR
RMSEA
RMSEA 90% CI
Cut-off value
0
0.9
0.95
0.08
0.05
0.08
0.05
acceptable
Good fit
acceptable
Good fit
acceptable
Good fit
Estimate
Comment
294.515
119
0.923
Acceptable
Acceptable
0.048
Good fit
0.075
Acceptable
0.064:0.086
Acceptable range
189
8.3.2.4 Statistical significance of parameter estimates
Inspection of the correlation values, standard errors and the test statistic in Table 33,
revealed that all correlation values were not greater than 1.00, all Z-statistics were
greater than 1.96 (P<0.05) and the signs were appropriate (positive). The estimates
were therefore found to be reasonable as well as statistically significant. The
parameter with the highest standardised coefficient was the indicator variable PP11.
The parameter coefficient was found to be 0.947. The indicator variable PP11 (there
is an incentive programme for H&S performance on this project) was found to be
more closely associated with the construct of project H&S performance than all
indicator variables. However, all parameter estimates showed high correlation values
close to 1.00 suggesting a high degree of linear association between the indicator
variables and the factor, project H&S performance. In addition,
values were
found to be close to the desired value of 1.00. The only exceptions were the indicator
variables PP8, PP9 and PP13. The
values for these variables were below 0.5.
Table 33: Factor loading and Z-statistic of Project H&S performance model
Indicator
Variable
Unstandardised
Coefficient (λ)
Standardised
Coefficient
(λ)
Z- Statistic
PP1
PP2
PP3
PP4
PP5
PP6
PP7
PP8
PP9
PP10
PP11
PP12
PP13
PP14
PP15
PP16
PP17
0.824
0.836
0.838
0.802
0.808
0.704
0.877
0.688
0.730
0.806
0.947
0.726
0.685
0.772
0.885
0.907
0.881
0.855
0.862
0.859
0.805
0.808
0.721
0.857
0.698
0.634
0.779
0.805
0.743
0.694
0.746
0.790
0.800
0.775
15.205
14.829
15.748
14.239
14.161
11.413
18.523
11.169
11.945
13.806
22.201
12.093
11.386
13.861
15.539
16.185
18.191
(Robust statistical significance at 5% level)
190
Significant at
5% level?
0.732
0.742
0.738
0.648
0.652
0.520
0.735
0.487
0.402
0.607
0.649
0.553
0.482
0.556
0.625
0.640
0.600
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
8.3.2.5 Internal reliability and validity of scores
In order to determine the internal consistency of the composite of the project H&S
performance measurement model the Rho coefficient and the Cronbach’s alpha
coefficient were examined to establish reliability (Byrne, 2006:133). According to
Kline (2005:59), the reliability coefficient should fall between zero and 1.00. Values
close to 1.00 are desired. The Rho coefficient of internal consistency was found to be
0.963. This value was above the minimum value of 0.70. Similarly, the Cronbach’s
alpha was also found to be above the minimum value of 0.70 at 0.963. Both these
values showed a high level of internal consistency and therefore reliability.
The construct validity was determined by examining the magnitude of the parameter
coefficients (factor loadings). High parameter coefficients of greater than 0.5 indicate
a close relation between the factor and an indicator variable. A parameter coefficient
of 0.5 is interpreted as 25% of the total variance in the indicator variable being
explained by the latent variable (factor). Accordingly a parameter coefficient has to be
greater than 0.7 to explain about 50% of the variance in an indicator variable (Hair et
al., 1998:111). Inspection of unstandardised parameter coefficients presented in Table
33, revealed that they were sufficiently high with the minimum of 0.685 for the
relationship between PP13 and the factor, project H&S performance. The parameter
estimate of 0.685 suggested that the factor accounted for about 47% of the variance in
PP13. This value was however above the acceptable level. On the other hand, all other
parameter estimates were above 50% and therefore indicative of an adequate fit
between the indicator variables and the factor.
Further, the LM test did not reveal any significant mis-specification of the project
H&S performance construct. In addition to the LM test, statistical significance of
parameter estimates at P<0.05, also indicates a level of mis-specification (Byrne,
2006:113). Inspection of the LM test output and the parameter estimates significance
revealed that there were no significant mis-fitting variables that would have justified
re-specification of the measurement model.
Therefore the factor, project H&S performance, satisfied both the internal reliability
and the construct validity criteria. The Rho value was above the minimum of 0.70
191
(Table 34) and the construct validity criteria was justified by the magnitude, signs and
statistical significance of all parameter coefficients.
Table 34: Reliability and construct validity of Project H&S performance model
Factor
Project H&S
performance
No.
of Indicator
indicator variable
variables
17
PP1
PP2
PP3
PP4
PP5
PP6
PP7
PP8
PP9
PP10
PP11
PP12
PP13
PP14
PP15
PP16
PP17
Factor
loading
Cronbach’s
Alpha
0.855
0.862
0.859
0.805
0.808
0.721
0.857
0.698
0.634
0.779
0.805
0.743
0.694
0.746
0.790
0.800
0.775
0.963
Rho
coefficient
0.963
* Parameter estimates are based on standardised solutions
8.3.2.6 Solution evaluation and summary on project H&S performance
The residual covariance estimates fell within the acceptable range, the robust fit
indexes met the cut-off indexes and the parameter estimates were found to be
statistically significant and feasible. Based on these criteria the measurement model
for the project H&S performance subscale was found to adequately fit the sample
data. Consequently, there was no need to modify the model before including it in the
full latent variable model analysis. Furthermore, there was no significant evidence of
model mis-specification and the LM test supported the conclusion that the
measurement model for project H&S performance subscale adequately fit the sample
data. Apart from this, it is unwise to modify a model that fits well because
modifications may only be fitting small characteristic features of the sample (Byrne,
2006:103).
192
Therefore, the factor, project H&S performance, appeared to be explained by the
indicator variables PP1 to PP17 and hence adequately measured the project H&S
performance construct. The measurement model on project H&S performance could
therefore be used in the analysis of the full latent variable model.
8.3.3
Contractor H&S performance
The number of cases that were analysed for the contractor H&S performance subscale
was 275 cases. The number of cases that were skipped was six because of missing
variables. The contractor H&S performance subscale had 9 parcelled dependent
variables, 12 independent variables and 21 free parameters. The number of fixed nonzero parameters was 12.
A three factor contractor H&S performance construct comprising commitment,
communication and procedures was analysed before inclusion in the full latent model.
The indicator variables were organised in parcels as shown in Table 35 and Figure 39.
In order to establish how well the model fit the sample data and the strength of the
hypothesised relations between variables, results presented on residual covariance
matrix, distribution of standardised residuals, fit statistics and statistical significance
at probability level 5% were examined. In addition, the Cronbach’s alpha and the Rho
coefficient of internal consistency were examined for score reliability. The construct
validity of the measurement model was determined from the convergent validity and
the magnitude of parameter coefficients.
8.3.3.1 Diagnostic fit analysis – Analysis of residual covariance estimates
The average absolute residual values of the contractor H&S performance are
presented in Table 25. The average absolute residuals, both unstandardised and
standardised, are presented. Results revealed that all the absolute residual values and
the average off-diagonal absolute residual values were close to zero. The
unstandardised average off-diagonal residual was 0.0062 while the standardised
average off diagonal residual was found to be 0.0078. A residual value greater than
2.58 is described as large (Byrne, 2006:94). The results obtained for contractor H&S
performance measurement model suggested a fairly acceptable fit to the sample data
because the absolute residuals were all below 2.58. In addition, 100% of standardised
residuals fell between -0.1 and +0.1 which is the acceptable range.
193
In order for a model to be described as well-fitting, the distribution of standardised
residuals should be symmetrical and centred around zero (Byrne, 2006:94). From the
above information, results suggested a measurement model that had an adequate fit.
Therefore, since this diagnostic fit analysis indicated a good fit, further tests of
goodness-of-fit were possible to conclusively make a decision on the fit and
appropriateness of the measurement model.
Table 35: Postulated contractor H&S performance measurement model
Latent variable
(factor)
Commitment
Communication
Procedures
Indicator variable
(Contractor top management…
has demonstrated a positive H&S attitude
has actively promoted H&S
has provided adequate resources for H&S implementation
has put in effort to ensure routine H&S evaluation in all work
schedules
has set up incentives for good H&S behaviour
has set H&S as a major agenda item in all project meetings
actively monitors H&S programmes on the project
always attends H&S meetings on the construction site
conducts H&S inspections/audits regularly
is always involved in accident /incident investigations
has established formal H&S reporting system
has a formal H&S feedback system
involves workers in H&S planning
provides timely feedback on incidents/accidents
always communicates risk findings to all workers
has made a clear H&S policy statement on the project
has made clear the expected H&S performance standard on the
project
has provided information on H&S risk control
regularly makes H&S briefs
has “stayed –in- touch” on H&S issues
has an effective H&S policy
has an effective H&S plan
scheduled H&S in all subcontractor contracts or prequalification
documents
conducts regular H&S performance measurement
conducted a hazard identification and risk assessment for the
project
actively oversees H&S on critical operations
has set up a formal H&S training programme
considers H&S knowledge to be a requirement for employment
has permanent H&S staff
ensures that H&S induction is conducted for all employees
194
Parcel
label
COTP1
COTP2
COTP3
COTP1
COTP2
COTP3
COTP1
COTP2
COTP3
COTP1
CONP1
CONP2
CONP3
CONP1
CONP2
CONP3
CONP1
CONP2
CONP3
CONP1
COPP1
COPP2
COPP3
COPP1
COPP2
COPP3
COPP1
COPP2
COPP3
COPP1
COPP1
COPP2
COPP
COTP1
COTP2
COTP
COTP3
CONP1
Contractor H&S
Performance
COPP3
CONP
CONP2
CONP3
Figure 39: Contractor H&S performance construct
8.3.3.2 Goodness-of-fit statistics - RML
The analysis strategy of goodness-of-fit for the contractor H&S performance construct
followed a two statistic strategy of fit indexes and is reported on in this section.
The sample data on contractor H&S performance measurement model yield the
of 25.0664 with 24 degrees of freedom and (N=275; P=0.40216,). The chisquare was insignificant. This chi-square value indicated that the departure of the
sample data from the postulated measurement model was not significant and hence
indicative of an acceptable fit. However the chi-square test is very sensitive to sample
size and is used more as a descriptive index of fit rather than as a statistical test
(Kline, 2005:136). Therefore a normed Chi-square value is usually adopted by most
researchers. Normed chi-square is the procedure of dividing the chi-square by the
degrees of freedom. The normed values of up to 3.0 or even 5.0 are recommended
(Kline, 2005:137). From the above chi-square and degrees of freedom values the ratio
was found to be 1.044. This ratio was lower than the upper limit of 3.0 or 5.0
advocated for by some authors (Kline, 2005:137).
195
In addition, the CFI was found to be 0.999 and the RMSEA with 90% confidence
interval (lower bound value = 0.000 and upper bound value = 0.051) was found to be
0.013. The CFI value was higher than the cut-off limit of 0.95 for a model to be
described as having a good fit. Similarly, the RMSEA value of 0.013 was less than the
upper cut-off value of 0.05 for a good fitting model. The absolute fit index SRMR
was found to be 0.009. The SRMR index also met the cut-off criteria for a good fit
model. A good fit model has an SRMR value of not more than 0.05. These fit indexes
for the contractor H&S performance measurement model suggested that the
postulated model adequately fit the sample data and could therefore be included in the
full latent variable model analysis (Table 36).
In addition to fitting the measurement model, it was also necessary to determine
whether the postulated model was reasonable. To establish this, analysis of each
parameter estimate was done to determine statistical significance. Statistics of these
parameter estimates are presented in Table 37.
Table 36: Robust fit indexes for Contractor H&S Performance construct
Fit Index
Df
CFI
SRMR
RMSEA
RMSEA 90% CI
Cut-off value
0
0.9
0.95
0.08
0.05
0.08
0.05
acceptable
Good fit
acceptable
Good fit
acceptable
Good fit
Estimate
Comment
25.0664
24
0.999
Acceptable
Good fit
0.009
Good fit
0.013
Good fit
0.000:0.051
Acceptable range
8.3.3.3 Statistical significance of parameter estimates
Inspection of the correlation values, standard errors and the test statistics in Table 37
revealed that all correlation values were not greater than 1.00, z-statistics were greater
than 1.96 and the signs were appropriate. The estimates were therefore deemed
reasonable as well as statistically significant. The parameter with the highest
standardised coefficient was the indicator variable CONP1. The parameter coefficient
was found to be 0.982. The parcel CONP1, containing the indicator variables namely:
196
(1) contractor top management has established formal H&S reporting system; (2)
provides timely feedback on incidents/accidents and (3) has made clear the expected
H&S performance standard on the project, was found to associate more with the
communication factor than indicator variables in CONP2 and CONP3. However, all
parameter estimates had high correlation values close to 1.00. The high correlation
values suggested a high degree of linear association between the indicator variables
and the factors of contractor H&S performance construct namely, procedures,
communication and commitment. In addition, the
values were also close to the
desired value of 1.00 indicating that the factors explained more of the variance in the
indicator variables.
Table 37: Factor loading and Z-statistics of contractor H&S performance model
Indicator
Unstandardised
Standardised
Z-
Significant at
Variable
Coefficient (λ)
Coefficient (λ)
Statistic
5% level?
COPP1
0.904
0.971
19.674
0.942
Yes
COPP2
0.927
0.954
20.535
0.909
Yes
COPP3
0.847
0.934
17.363
0.872
Yes
COTP1
0.768
0.943
15.912
0.890
Yes
COTP2
0.793
0.937
16.240
0.878
Yes
COTP3
0.849
0.966
17.387
0.934
Yes
CONP1
0.880
0.982
18.900
0.964
Yes
CONP2
0.857
0.968
19.744
0.937
Yes
CONP3
0.822
0.938
17.487
0.880
Yes
(Robust statistical significance at 5% level)
In addition, correlation values between the three factors of contractor H&S
performance construct indicated that the factors were related to each other and
belonged to one construct. The correlation values ranged from 0.896 (procedures and
commitment) to 0.932 (Commitment - Communication). Table 38 is a tabulation of
correlations between the factors of contractor H&S performance.
197
Table 38: Correlations of factors of contractor H&S performance model
Factors
of
contractor
H&S COPP
COTP
CONP
performance
Procedures (COPP)
1.000
Commitment (COTP)
0.896
1.000
Communication (CONP)
0.921
0.932
1.000
8.3.3.4 Internal reliability and validity of scores
The internal consistency and reliability of scores for the contractor H&S performance
construct was determined from the Rho and the Cronbach’s alpha coefficient.
According to Kline (2005:59), the reliability coefficient should fall between zero and
1.00. Values close to 1.00 are desired. The Rho coefficient of internal consistency was
found to be 0.989. This value was above the minimum required value of 0.70.
Similarly, the Cronbach’s alpha was above the minimum acceptable value of 0.70.
The Cronbach’s alpha was found to be 0.981. Both these values revealed a high level
of internal consistency and therefore reliability.
Construct validity was determined from the magnitude and reasonableness of the
parameter coefficients (factor loadings). High parameter coefficients of greater than
0.5 indicate a close relation between the factor and an indicator variable. A parameter
coefficient of 0.5 is interpreted as 25% of the total variance in the indicator variable
being explained by the latent variable (factor). Accordingly a parameter coefficient
has to be greater than 0.7 to explain about 50% of the variance in an indicator variable
(Hair et al., 1998:111). Inspection of unstandardised parameter coefficients presented
in Table 37, revealed that all coefficients were sufficiently high with the minimum of
0.768 for the relationship between COTP1 and the commitment factor. This parameter
estimate suggested that the commitment factor accounted for about 59% of the
variance in COTP1. The magnitude of the parameter estimate was above the 50%
minimum. Likewise, all other parameter estimates were above 59% and therefore
indicated an adequate fit. This in addition indicated a strong relationship between the
indicator variables and the factors of contractor H&S performance construct.
198
The contractor H&S performance construct satisfied both the internal reliability and
the construct validity criteria. The Rho value was above the minimum value of 0.70,
the magnitude, signs and statistical significance of the parameter estimates were
appropriate.
Table 39: Reliability and construct validity of contractor H&S performance model
Factor
No. of parcels
Contractor
H&S
Performance
9
Indicator
Factor
Cronbach’s
Rho
variable
loading
Alpha
coefficient
COPP1
0.971
COPP2
0.954
COPP3
0.934
COTP1
0.943
COTP2
0.937
COTP3
0.966
CONP1
0.982
CONP2
0.968
CONP3
0.938
0.981
0.989
* Parameter estimates are based on standardised solutions
8.3.3.5 Solution evaluation and summary on contractor H&S performance
The residual covariance estimates fell within the acceptable range, the robust fit
indexes met the cut-off index criteria and all the parameter estimates were statistically
significant and feasible. Considering these criteria, the measurement model for the
contractor H&S performance subscale was found to adequately fit the sample data.
Consequently, there was no need to improve the measurement model before it could
be included in the full latent variable model. Furthermore, the LM test did not indicate
any significant evidence of model mis-specification. Therefore, the contractor H&S
performance construct was adequately measured by the indicator variables and hence
could be used in the analysis of the full latent variable model.
8.3.4
Designers H&S Performance
The number of cases that were analysed for the designer H&S performance subscale
was 271 cases. The number of cases that were skipped was 10 because of missing
variables. The designer H&S performance subscale had 9 parcels of dependent
199
variables, 12 independent variables and 21 free parameters. The number of fixed nonzero parameters was 12.
A three factor designer H&S performance construct comprising commitment,
communication and procedures, was analysed before it could be included in the full
latent variable model. The indicator variables were organised in parcels as shown in
Table 40 and Figure 40.
In order to establish how well the measurement model fit the sample data and how
well the indicator variables measured the factors of designer H&S performance
construct, results presented on residual covariance matrix, distribution of standardised
residuals, fit statistics and statistical significance at probability level 5% were
examined. In addition the Cronbach’s alpha and the Rho coefficient of internal
consistency were examined for score reliability. Construct validity of the
measurement model was determined from model convergence and the magnitude of
parameter coefficients.
DEPP1
DEPP2
DEPP3
DEPP
DETP2
Designer H&S
Performance
DETP1
DETP
DETP3
DENP1
DENP
DENP2
DENP3
Figure 40: Designer H&S performance measurement model
(DEPP = procedures, DETP = Commitment and DENP= Communication)
200
8.3.4.1 Diagnostic fit analysis –Residuals
The average absolute residual values of the designer H&S performance construct are
presented in Table 25. Both unstandardised and standardised average absolute
residuals are presented. Results revealed that all the absolute residual values and the
average off-diagonal absolute residuals, both unstandardised and standardised, were
close to zero. The unstandardised average off diagonal residual was 0.0074 while the
standardised average off-diagonal residual was found to be 0.0083. These values were
smaller than the 2.58 upper limit and therefore suggested that the model could have an
adequate fit to the sample data. In addition, 100% of standardised residuals fell
between the acceptable range of -0.1 and +0.1 (Byrne, 2006:94).
The favourable diagnostic fit analysis tests justified further tests of goodness-of-fit on
the designer H&S performance construct.
8.3.4.2 Goodness-of-fit statistics- RML
The sample data on designer H&S performance measurement model yield the
of 35.6033 with 24 degrees of freedom (N=271; P=0.05994). The chisquare was insignificant. This chi-square value indicated that the departure of the
sample data from the postulated measurement model was not significant and hence
indicative of an acceptable fit. The ratio of the chi-square to the degrees of freedom
was found to be 1.4835. This ratio was lower than the upper limit of 3.0 or 5.0
advocated for by some authors (Kline, 2005:137). The measurement model was
therefore considered to be of an acceptable fit.
In addition to the Chi-square test, the CFI was found to be 0.994. The CFI value was
higher than the minimum value of 0.95 set for a good fit criteria. On the other hand,
the RMSEA (90% CI) was found to be 0.042 (lower bound value = 0.000; upper
bound value = 0.070). Equally, the RMSEA value of 0.042 was less than the upper
cut-off value of 0.05 for a good fitting model. The absolute fit index SRMR was
found to be 0.010. The SRMR index also met the cut-off criteria of not exceeding
0.05 for a good fitting model. These fit indexes for the designer H&S performance
measurement model suggested that the model adequately fit the sample data and
therefore could be included in the full latent variable model (Table 41).
201
Additionally, parameter estimates were scrutinised to determine whether the model
worked properly and was reasonable. This involved assessing the magnitude, signs
and statistical significance of the parameter estimates. These statistics are presented in
Table 42.
Table 40: Postulated designer H&S Performance model
Latent variable
(factor)
Commitment
Communication
Procedures
Indicator variable
(Designers …
have demonstrated positive H&S attitude
actively promote H&S on the project
provide information on H&S
put in effort to ensure routine H&S evaluation in all work
schedules
conduct regular H&S inspections/audits
ensure that H&S is always on the agenda of project meetings
regularly monitor H&S programmes
regularly attend H&S meetings for the project
are always involved in accident/incident investigations
have set up a formal H&S reporting system
involve the contractor to plan for H&S
provide timely feedback on H&S issues such as incidents /
accidents
always communicate risk findings
made clear the expected H&S standard performance
provided information on H&S risk control
make regular H&S briefings / reports
have stayed ‘in-touch’ on H&S issues
have a H&S policy for the project
have a H&S plan for the project
regularly measure H&S performance
or other designers oversee H&S on critical operations
or other designers have clearly stated H&S requirements on
contract drawings / specifications
conducted a Hazard identification and risk assessment for the
project
have clearly defined H&S roles and responsibilities for the
project
have demonstrated knowledgeable of H&S
ensure that H&S induction is conducted for all on the project
202
Parcel
label
DETP1
DETP2
DETP3
DETP1
DETP2
DETP3
DETP1
DETP2
DETP3
DENP1
DENP2
DENP3
DENP1
DENP2
DENP3
DENP1
DENP2
DEPP1
DEPP2
DEPP3
DEPP1
DEPP2
DEPP3
DEPP1
DEPP2
DEPP3
Table 41: Robust fit indexes for designer H&S performance construct
Fit Index
Df
CFI
SRMR
RMSEA
Cut-off value
0
0.9
0.95
0.08
0.05
0.08
0.05
acceptable
Good fit
acceptable
Good fit
acceptable
Good fit
RMSEA 90% CI
Estimate
Comment
35.6033
24
0.994
Acceptable
Good fit
0.010
Good fit
0.042
Good fit
0.000:0.070
Acceptable range
8.3.4.3 Statistical significance of parameter estimates
Examination of the correlation values, standard errors and the test statistics in Table
42 revealed that all correlation values were not greater than 1.00, Z-values were
greater than 1.96 and the signs were appropriate and reasonable. All parameter
estimates were therefore considered to be reasonable as well as statistically
significant. The parameter with the highest standardised coefficient was the indicator
variable DENP3. The parameter coefficient was found to be 0.984. The parcel
DENP3, containing the indicator variables namely: (1) designer provide timely
feedback on H&S issues such as incidents/accidents and (2) designer has provided
information on H&S risk control, was found to be more associated with the
communication factor than the indicator variables in parcels DENP1 and DENP2.
However, all parameter estimates had high correlation values close to 1.00 suggesting
that all indicator variables measured the communication factor.
The high correlation values suggested a high degree of linear association between the
indicator variables and the factors of designer H&S performance construct. In
addition, the
values were also found to be close to the desired value of 1.00 and
hence indicating that the factors of designer H&S performance explained more of the
variance in the indicator variables.
Examination of correlation values in Table 43 between the three factors of the
designer H&S performance construct indicated that the factors were related to each
other and to one construct. The correlation values ranged from 0.818 (procedures and
203
commitment) to 0.924 (Commitment - Communication). Table 41 shows exogenous
variable correlations.
Table 42: Factor loading and Z-statistics of designer H&S performance model
Indicator
Variable
Unstandardised
Coefficient (λ)
Standardised
Coefficient (λ)
Z- Statistic
DEPP1
DEPP2
DEPP3
DETP1
DETP2
DETP3
DENP1
DENP2
DENP3
0.868
0.867
0.886
0.793
0.788
0.812
0.888
0.944
0.959
0.703
0.967
0.942
0.944
0.930
0.964
0.942
0.949
0.984
18.346
20.411
20.069
18.925
20.778
21.192
19.696
20.801
22.096
Significant at
5% level?
0.494
0.935
0.888
0.890
0.864
0.929
0.887
0.900
0.968
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
(Robust statistical significance at 5% level)
Table 43: Correlations of factors of designer H&S performance model
Client H&S culture factors
DEPP
DETP
Procedures (DEPP)
1.000
Commitment (DETP)
0.818
1.000
Communication (DENP)
0.825
0.924
DENP
1.000
8.3.4.4 Internal reliability and validity of scores
The Rho coefficient and the Cronbach’s alpha coefficient were examined in order to
establish score reliability (Byrne, 2006:133). According to Kline (2005:59), the
reliability coefficient should fall between zero and 1.00. Values close to 1.00 are
desired. The Rho coefficient of internal consistency was found to be 0.976. This value
was above the minimum required of 0.70. Similarly, the Cronbach’s alpha was found
to be above the minimum 0.70 at 0.962. Both these values indicated a high degree of
internal consistency and homogeneity.
Construct validity was determined by examining the magnitude and signs of the
parameter coefficients. High parameter coefficients of greater than 0.5 indicate a close
relation between the factor and an indicator variable. A parameter coefficient of 0.5 is
204
interpreted as 25% of the total variance in the indicator variable being explained by
the latent variable (factor). Accordingly a parameter coefficient has to be greater than
0.7 to explain about 50% of the variance in an indicator variable (Hair et al.,
1998:111). Inspection of standardised parameter coefficients in Table 42 revealed that
all coefficients were sufficiently high with the minimum factor loading being 0.703
for the relationship between DEPP1 and the procedures factor. The parameter
estimate of 0.703 suggested that the factor accounted for about 49% of the variance in
DEPP1. All other parameter estimates were above 80% and therefore indicated that
the measurement model had a relatively acceptable fit and the solution was
reasonable.
As a result, the designer H&S performance construct satisfied both internal reliability
and construct validity criteria because the Rho value was above the minimum value of
0.70 and the magnitude, signs and statistical significance of parameter estimates were
appropriate.
Table 44: Reliability and construct validity of designer H&S performance model
Factor
Designer H&S
performance
No. of parcels
9
Indicator
variable
Parameter
coefficient
Cronbach’s
Alpha
Rho
coefficient
DEPP1
DEPP2
DEPP3
DETP1
DETP2
DETP3
DENP1
DENP2
DENP3
0.703
0.967
0.942
0.944
0.930
0.964
0.942
0.949
0.984
0.962
0.976
* Parameter estimates are based on standardised solutions
8.3.4.5 Solution evaluation and summary on designer H&S performance
The measurement model for the designer H&S performance subscale revealed an
adequate fit to the sample data. The residual covariance estimates fell within the
acceptable range of -0.1 to 0.1; the robust fit indexes (CFI= 0.994; RMSEA = 0.042 &
SRMR = 0.010) met the cut-off index criteria and the parameter estimates were found
to be statistically significant at 5% level and feasible. In addition the LM test did not
205
reveal significant need for model re-specification. Consequently, there was no need to
improve the measurement model before including it in the full latent variable model
8.3.5
External environment
Out of 281 cases that were available for analysis, 275 cases were analysed for the
external environment subscale. Six cases were skipped because of missing variables.
The external environment subscale had six dependent variables, seven independent
variables and 12 free parameters. The number of fixed non-zero parameters was
seven. The six dependent indicator variables for the external environment were:
legislative, economic, social, professional bodies, technology and materials. These
indicator variables are presented in Table 45. The external environment measurement
model shown in Figure 41 was analysed before it could be included in the full latent
variable model.
In order to establish how well the model fit the sample data and the strength of the
hypothesised relations between the variables, results on residual covariance matrix,
distribution of standardised residuals, fit statistics and statistical significance at
probability level 5% were examined. In addition, the Cronbach’s alpha and the Rho
coefficient of internal consistency were examined for score reliability. Results on
these statistics are presented in this section for the external environment construct.
Table 45: Postulated external environment influence model
Latent variable
Indicator variable
(factor)
(Clients …
External
ensure H&S implementation due to Legislative requirements,
environment
actively participate in H&S implementation due to the
economic reasons
actively participate in H&S implementation due to
construction workers’ union’s influence
consider H&S in making decisions on cost, time & quality as a
result of designers’ persuasion
ensure H&S implementation due to Professional bodies’
influence
consider H&S due to Materials & technology adopted for the
project
206
label
LGN
ECON
SOC
PRI1
PRI2
TECH
Figure 41: External environment construct
(LGN= Legislative, ECON= Economic, SOC = Social, PRI= Professional and TECH
= Technology & materials)
8.3.5.1 Diagnostic fit analysis –Residuals
The average absolute residual values of the external environment construct are
presented in Table 46. Results revealed that all the absolute residual values and the
average off-diagonal absolute residuals, both unstandardised and standardised, were
close to zero. The smallest unstandardised average off diagonal residual was 0.0289
while the largest was 0.0405. Similarly, the smallest standardised average off diagonal
residual was 0.0241 while the largest was 0.0337.These residual values were
considered small as they were all less than 2.58 (Byrne, 2006:94) In addition, 100%
of standardised residuals fell within the acceptable range of -0.1 and +0.1. The
significance of this distribution is that for a model to be described as well-fitting, the
distribution of standardised residuals should be symmetrical and centred around the
zero (Byrne, 2006:94). From the above information, the results seemed to suggest that
the model had a good-fit to the sample data. Therefore, since this initial examination
of residuals indicated a fairly good fit, further tests of goodness-of-fit were justifiable.
207
Table 46: Residual values for external environment influence model
Residual
Index
Unstandardised
Average absolute residual
Average off-diagonal absolute residual
Standardised
Average absolute residual
Average off-diagonal absolute residual
% falling between
0.0289
0.0405
0.0241
0.0337
100%
8.3.5.2 Goodness-of-fit statistics - RML method
The sample data on the external environment factor yield an
of 15.490 with
9 degrees of freedom. The associated p-value was determined to be 0.07832 for the
analysed sample of 275 cases. The chi-square value suggested that the difference
between the sample data and the postulated external environment measurement model
was insignificant. In addition the ratio of
to the degrees of freedom was
determined to be 1.72 which was lower than the upper limit value of 3.0 (Kline,
2005:137).
Similarly, other fit indexes indicated a good fit of the model to the sample. The robust
CFI index of 0.989 was greater than the cut-off value for a good fitting model. A
model is said to be a good fit if the CFI is above the cut-off value of 0.950 (Hu &
Bentler, 1999:27). The robust RMSEA with 90% confidence interval (lower bound
value = 0.000 and upper bound value = 0.094) was found to be 0.052. This value was
marginally above the maximum value of 0.05 for a good fit model. On the other hand,
the absolute fit index SRMR was found to be 0.033. This value indicated a very good
fit because a good fitting model is expected to have an SRMR index lower or equal to
0.05 while an index of 0.08 is sufficient to accept the postulated model. The absolute
fit index SRMR accounts for the average discrepancy between the sample and the
postulated correlation matrices and therefore it represents the average value across all
standardised residuals and ranges between zero and 1.00 (Byrne, 2006:94). Evaluation
of the SRMR, RMSEA (90% CI) and the CFI fit indexes indicated a very good fit of
the measurement model for the external environment factor because those indexes
met the condition for a good fit (See Table 47).
208
Additionally, parameter estimates were scrutinised to determine whether the model
worked properly and was reasonable. This involved assessing the magnitude, signs
and statistical significance of the parameter estimates. These statistics are presented in
Table 48.
Table 47: Robust fit indexes for external environment influence construct
Fit Index
Df
CFI
SRMR
RMSEA
RMSEA 90% CI
Cut-off value
0
0.9
0.95
0.08
0.05
0.08
0.05
acceptable
Good fit
acceptable
Good fit
acceptable
Good fit
Estimate
Comment
15.490
9
0.989
Acceptable
Good fit
0.033
Good fit
0.052
Good fit
0.000:0.094
Acceptable range
8.3.5.3 Statistical significance of parameter estimates
Apart from assessing the goodness-of-fit or the lack of it, feasibility of a model can be
judged by a further inspection of the obtained solution and this involves inspection of
parameter estimates, standard errors and the test statistics (Raykov, 1991:501).
Estimates are said to be unreasonable if in the output there are estimates that have
correlation values that are greater than 1.00, have negative variances and the
correlation or covariances are not definite positive (Byrne, 2006:103). Additionally
the test statistic needs to be greater than 1.96 based on the probability level of 5%
before the hypothesis can be rejected (Byrne, 2006:103).The test statistic reported in
this study was the parameter estimate divided by its standard error and therefore it
functions as a Z-statistic to test that the estimate is statistically different from zero.
Inspection of the correlation values, standard errors and the test statistic in Table 48
revealed that all correlation values were not greater than 1.00, test statistics were
greater than 1.96 and the signs were appropriate. The estimates were therefore found
to be reasonable as well as statistically significant. The parameter with the highest
209
standardised coefficient was the indicator variable PRI2. The parameter coefficient
was found to be 0.833. The indicator variable PRI2 namely: clients ensure H&S
implementation due to professional bodies’ influence, was found to be more
associated with the construct of external environment than all indicator variables.
However, all parameter estimates had high correlation values that were above 0.5
suggesting a high degree of linear association between the indicator variables and the
external environment factor. In addition, although the
values were not very close
to the value of 1.00, they were all crucially above 0.5 with the exception of LGN. The
external environment factor accounted for 42% of the variance in the indicator
variable, LGN.
Table 48: Factor loading and Z-statistics of the external environment model
Indicator
Variable
Unstandardised
Coefficient (λ)
Standardised
Coefficient (λ)
Z- Statistic
LGN
ECON
SOC
PRI1
PRI2
TECH
0.644
0.879
0.935
0.955
0.884
0.834
0.650
0.791
0.801
0.816
0.833
0.783
8.990
18.026
17.519
20.021
16.044
14.183
Significant at
5% level?
0.422
0.625
0.642
0.666
0.695
0.613
Yes
Yes
Yes
Yes
Yes
Yes
(Robust statistical significance at 5% level
8.3.5.4 Internal reliability and validity of scores
The Rho and Cronbach’s alpha coefficient were examined in order to establish score
reliability (Byrne, 2006:133). According to Kline (2005:59), the reliability coefficient
should fall between zero and 1.00. Values close to 1.00 are desired. The Rho
coefficient of internal consistency was found to be 0.905. This value was above the
minimum required value of 0.70. Similarly, the Cronbach’s alpha was found to be
above the minimum 0.70. The Cronbach’s alpha was found to be 0.903. Both these
values indicated a high degree of internal consistency and therefore reliability.
Construct validity was determined by examining the magnitude and signs of the
parameter coefficients. High parameter coefficients of greater than 0.5 indicate a close
relation between the factor and an indicator variable. A parameter coefficient of 0.5 is
210
interpreted as 25% of the total variance in the indicator variable being explained by
the factor. Accordingly a parameter coefficient has to be greater than 0.7 to explain
about 50% of the variance in an indicator variable (Hair et al., 1998:111). Inspection
of both unstandardised and standardised parameter coefficients presented in Table 48,
revealed that all coefficients were sufficiently high with the minimum of 0.650 for the
relationship between LGN and the external environment factor. The parameter
estimate of 0.650 suggested that the factor accounted for about 42% of the variance in
the indicator variable, LGN. However all other parameter estimates were above 50%
and therefore indicated a reasonable measurement model.
Subsequently, the external environment construct satisfied both internal reliability and
construct validity criteria because the Rho value was above the minimum value of
0.70 and the magnitude, signs and statistical significance of parameter estimates were
appropriate.
Table 49: Reliability and construct validity of the environment factor
Factor
External
environment
No. of indicator
variables
Indicator
variable
Factor
loading
Cronbach’s
Alpha
Rho
coefficient
6
LGN
ECON
SOC
PRI1
PRI2
TECH
0.650
0.791
0.801
0.816
0.833
0.783
0.903
0.905
* Parameter estimates are based on standardised solutions
8.3.5.5 Solution evaluation and summary on the external environment subscale
The residual covariance estimates fell within the acceptable range, the robust fit
indexes met the cut-off index criteria and all the parameter estimates were found to be
statistically significant and feasible. It was concluded therefore, that the measurement
model for the external environment subscale, had an adequate fit to the sample data.
Consequently, there was no need to improve the measurement model before it could
be included in the full latent variable model. Furthermore, the LM test did not reveal a
significant evidence of model mis-specification and therefore the measurement model
was considered to work well.
211
8.4
CONSTRUCT VALIDITY OF THE SEM MODEL
Table 50 present values for the construct validity of the postulated model and internal
reliability of scores. The magnitude of the factor loadings informed the extent to
which the measurement models’ constructs could be described as valid. High values
of more than 0.45 are desirable. The factor loadings or coefficients in Table 50 were
all found to be higher than 0.45 and therefore indicated a significant relationship
between the factors. Similarly, the Rho coefficients were found to be sufficiently high
with values close to 1.00 and therefore the internal consistency criteria was met.
8.5
FIT STATISTICS ON THE STRUCTURAL MODEL
Results from the measurement models indicated that the models worked well and it
was therefore feasible to test the full latent variable model. The full structural model
was analysed in two parts and the results are presented as such due to its complexity.
The first part, Model 1.0, theorised that the external environment had an influence on
client H&S culture. The second part, model 2.0, theorised that client H&S culture had
a positive influence on designer, contractor and project H&S performance.
In the full structural model, the contractor H&S performance construct was analysed
as a one factor model with indicator variables in parcels. Similarly, the designer H&S
performance construct was analysed in the full structural model as a one factor model
with indicator variables in parcels. This approach was adopted for ease of analysis of
the full structural model and enabled a much clearer presentation of the full structural
model. All indicator variables defining the contractor and designer H&S performance
constructs were maintained. The only effect that could result from analysing the
contractor and designer H&S performance constructs as one factor models was a
compromise in the overall fit of the full latent variable model. This meant that there
was a high probability of rejecting a model that ideally should not be rejected.
Moreover, the question of whether measurement models should be checked before
analysing the full SEM model, (which was the strategy adopted in this study) or not,
is simply a strategy a researcher adopts (Bollen, 2000; Hayduck & Glaser, 2000 and
Boomsma, 2000). Furthermore the question of how many factors a construct should
have is also debatable (Bollen, 2000). However, assessing the measurement models
first has an advantage. The advantage of first analysing the measurement models
212
separately before analysing the full SEM model, is that the researcher is assured of a
proper working measurement model before analysing the full SEM model. By so
doing, the researcher avoids the frustration of re-specifying the full model if a solution
cannot be obtained.
Besides these observations, the analysis presented in the current study was a pure
confirmatory analysis and therefore recommendations were based on whether the
postulated priori model fit the sample data or not. Therefore for these reasons, the
contractor and designer H&S performance constructs were included in the full latent
variable model analysis as one factor models (See Figure 43).
8.5.1 Model 1.0 – External influence on client H&S culture
A confirmatory factor analysis of model 1.0 was conducted. Model 1.0 postulated that
the external environment had a direct positive influence on client H&S culture (Figure
42). The indicator variables of the external environment factors were: legislative,
economic, social, professional bodies and materials and technology. Model 1.0 is
presented in Figure 42. The theory and basis of the model was presented in section
7.2. Model 1.0 was part of the full structural model for the client-centred H&S
performance improvement model. The model was analysed and presented separately
due to complexity of the full structural model and also to improve model eloquence.
The number of cases that were analysed for the full latent variable model 1.0 was 238
out of the available 281 cases. Out of the available 281 cases, 43 cases were skipped
because of missing variables. The model had 57 dependent and 58 independent
variables. It also had 108 free parameters. The number of fixed non-zero parameters
was 64. The covariance matrix of the model was analysed using the robust maximum
likelihood estimation method. Raw data was used for the analysis.
213
Table 50: Reliability and construct validity of the client H&S culture model
Factor
No. of indicator Indicator
variables/parcels
variable
Client H&S
culture
19
Contractor H&S
Performance
9
Designer H&S
Performance
9
CLLP1
CLLP2
CLLP3
CLIP1
CLIP2
CLIP3
CLPP1
CLPP2
CLPP3
CLTP1
CLTP2
CLTP3
CLNP1
CLNP2
CLNP3
CLCE1
CLCE2
CLCE3
CLCE4
COPP1
COPP2
COPP3
COTP1
COTP2
COTP3
CONP1
CONP2
CONP3
DEPP1
DEPP2
DEPP3
DETP1
DETP2
DETP3
DENP1
DENP2
DENP3
214
Parameter
coefficient
0.892
0.940
0.853
0.933
0.578
0.921
0.961
0.943
0.891
0.697
0.929
0.937
0.935
0.960
0.956
0.816
0.880
0.897
0.870
0.971
0.954
0.934
0.943
0.937
0.966
0.982
0.968
0.938
0.703
0.967
0.942
0.944
0.930
0.964
0.942
0.949
0.984
Cronbach’s
Alpha
Rho
coefficient
0.963
0.978
0.981
0.989
0.962
0.976
Table 50: (Continued)
Project H&S
performance
17
External
environment
6
PP1
PP2
PP3
PP4
PP5
PP6
PP7
PP8
PP9
PP10
PP11
PP12
PP13
PP14
PP15
PP16
PP17
LGN
ECON
SOC
PRI1
PRI2
TECH
0.855
0.862
0.859
0.805
0.808
0.721
0.857
0.698
0.634
0.779
0.805
0.743
0.694
0.746
0.790
0.800
0.775
0.650
0.791
0.801
0.816
0.833
0.783
* Parameter estimates are based on standardised solutions
215
0.963
0.963
0.903
0.905
P1
CLNP
P2
Client H&S culture
P3
P4
H1a
CLTP
P6
H1b
P7
LGN
CLPP
ECON
SOC
PRI1
H1c
External
Environment
H1d
PRI2
TEC
P5
P8
P9
P 10
CLIP
P 11
P 12
H1e
P 13
CLLP
P 14
P 15
H1f
P 16
CLCP
P 17
P 18
P 19
Figure 42: Model 1.0 - External environment influence on client H&S culture
216
8.5.1.1 Hypothesised relationships for model 1.0
Model 1.0 generally postulated that the external environment had a direct positive
influence on client H&S culture (H1).
Specifically, model 1.0 hypothesized that the external environment had a direct
positive influence on the following factors of client H&S culture, namely:
1. Leadership (H1a);
2. Involvement (H1b);
3. Procedures (H1c);
4. Commitment (H1d);
5. Communication (H1e); and
6. Competence (H1f)
8.5.1.2 Model 1.0 goodness-of-fit statistics - RML method
The sample data on the model yield an
of 2523.9043 with 1218 degrees of
freedom. The associated p-value was determined to be 0.000 with a sample of 238
cases. The chi-square index suggested that the difference between the hypothesised
model and the sample data matrix was significant. However, the chi-square test of fit
is very sensitive and therefore could not be relied upon to determine fit. However the
normed chi-square index, which is the ratio of the scaled chi-square
to the
degrees of freedom yield a value of 2.072. This value was lower than the upper limit
value of 3.0 and therefore indicative of a reasonable fit of the model.
The robust CFI index was found to be 0.844. The CFI index was less than 0.900
which is the lower limit value for model acceptance. However, a two statistic strategy
is considered satisfactory to accept or reject a model (Hu & Bentler, 1999:28).
Therefore RMSEA and SRMR statistics were used to decide on the acceptability of
the model.
The robust RMSEA with 90% confidence interval was found to be 0.067. (lower
bound value = 0.063 and upper bound value = 0.071) The RMSEA index was just
above the upper limit of 0.050 for the model to be described as good. However the
value of 0.67 indicated that the model was acceptable. In addition, the absolute fit
index, SRMR, was found to be 0.074. The SRMR fit index indicated an adequate fit
of the full structural model 1.0 to the sample data.
217
All the indexes with the exception of the CFI met the condition for model acceptance.
Although the fit index CFI did not matter so much in the two fit statistic strategy, its
value of 0.844 was not far from the lower limit value of 0.900 for model acceptance.
See Table 51.
Table 51: Robust fit indexes for SEM Model 1.0
Fit Index
Df
CFI
SRMR
RMSEA
Cut-off value
0
0.9
0.95
0.08
0.05
0.08
0.05
acceptable
Good fit
acceptable
Good fit
acceptable
Good fit
RMSEA 90% CI
Model 1.0
Comment
2523.9043
1218
0.844
Acceptable
Inadequate
0.074
Acceptable
0.067
Acceptable
0.063:0.071
Acceptable range
8.5.1.3 Model 1.0 Hypotheses testing
Rejection of the hypotheses depended on how reasonable the parameter estimates
were in terms of their magnitude, signs and statistical significance. In addition, if the
output showed estimates that had correlation values greater than 1.00, had negative
variances and the correlation or covariances were not definite positive then they were
said to be exhibiting unreasonable estimates (Byrne, 2006:103). Additionally the test
statistic had to be greater than 1.96 based on the probability level 5% before the
hypothesis could be rejected (Byrne, 2006:103).The test statistic reported was the
parameter estimate divided by its standard error and therefore it functioned as a Zstatistic to test that the estimate was statistically different from zero. The significance
test was used to evaluate the hypotheses H1a to H1f.
Testing the influence of the external environment on client H&S culture
(H1)
It was generally hypothesised that the external environment had a direct positive
influence on client H&S culture (H1). Specifically the hypotheses were that the
external environment had a direct positive influence on the six factors of client H&S
218
culture, namely: leadership, involvement, procedures, commitment, communication
and competence.
Results from the SEM analysis yield support for all the hypothesised relations, H1a to
H1f. The hypothesised relationships between the external environment and all
endogenous factors of client H&S culture were found to be significant and they all
had definite positive directions. The relationship between the procedures factor and
the external environment was found to be the most significant. The parameter
coefficient for this relationship was 0.933 and the Z-statistic was 10.610. Similarly,
the relationship between the communication factor and the external environment was
found to be statistically significant. This relationship had a parameter coefficient of
0.932 and a Z – statistic of 9.781. These values are presented in Table 52.
Therefore the general hypothesis that the external environment had a direct positive
influence on client H&S culture could not be rejected (Table 52). In terms of the
magnitude of the parameter coefficients, a comparison of these revealed that the
influence of external environment on the commitment factor was found to be the
lowest at 0.784.
Table 52: Model 1.0 factor loadings and Z-statistic
Label
Hypotheses
Factor
External environment has a positive loading
direct influence on the factors of
(λ)
client H&S culture namely :
H1a
Leadership
0.880
9.569
Yes
H1b
Involvement
0.791
8.938
Yes
H1c
Procedures
0.933
10.610
Yes
H1d
Commitment
0.784
9.455
Yes
H1e
Communication
0.932
9.781
Yes
H1f
Competence
0.867
9.534
Yes
(Robust statistical significance at 5% level)
219
Z- Statistic
Significant?
8.5.1.4 Solution evaluation of the model 1.0
The robust fit indexes of SRMR and the RMSEA met the cut-off index criteria and
the parameter estimates were found to be statistically significant and reasonable. The
postulated structural model was therefore acceptable and considered to adequately fit
the sample data. Since the analysis was confirmatory, there was no need to further
improve the fit of the structural model at this stage. Furthermore, the LM test did not
indicate a significant evidence of model mis-specification. Byrne (2006:113) points
out, that for most models, model improvement is merely an exercise that tries to fit
small characteristic features of the sample and does not necessarily add value to that
already fitted. Therefore the hypothesised model 1.0 was accepted with its level of fit.
8.5.2
Model 2.0 – Client’s influence on project H&S performance
A confirmatory factor analysis was conducted on the postulated full latent variable
model 2.0. The hypothesis was that client H&S culture has a direct positive influence
on contractor, designer and project H&S performance. This model is presented in
Figure 43. The structural equation modelling software, MPlus version 6.0, was used to
analyse model 2.0. The use of MPlus was influenced by the complexity of the model
vis-à-vis the number of cases, and the fact that the model was a mixed one. The model
was mixed in the sense that some of the indicator variables were parcelled and others
were not. MPlus has more capabilities to handle these kinds of complexities. In
addition, MPlus has the capacity to handle up to 1000 iterations during analysis while
the default for EQS version 6.1 is only 30. In EQS therefore, type I error is more
likely when analysing complex models with samples that are not very large
The indicator variables for client H&S culture, contractor and designer H&S
performance were analysed in parcels while those of project H&S performance were
analysed as individual indicator variables. The number of cases that were analysed for
Model 2.0 was 259 cases. The number of cases that were skipped was 22 due to
missing variables. Only complete cases were analysed for the model. The model had
54 dependent and 60 independent variables. It also had 197 free parameters.
The model was analysed using the robust maximum likelihood method. As in the case
of measurement models, the covariance matrix was analysed as opposed to the
correlation matrix for reasons explained in the methods chapter.
220
Figure 43: Model 2 - Client influence on project H&S performance
221
8.5.2.1 Hypothesised relationships of model 2.0
Model 2.0 generally postulated that:
1. client H&S culture had a direct positive influence on contractor H&S
performance (H2);
2. client H&S culture had a direct positive influence on designer H&S
performance (H3);
3. client H&S culture had a direct positive influence on the project H&S
performance (H4);
4. contractor H&S performance had a direct positive influence on project H&S
performance (H5);
5. designer H&S performance had a direct positive influence on project H&S
performance (H6);
6. client H&S culture had a positive indirect influence on project H&S
performance that was mediated by contractor H&S performance (H7);
7. client H&S culture had a positive indirect influence on project H&S
performance that was mediated by designer H&S performance (H8).
8.5.2.2 Model 2.0 goodness-of-fit statistics- RML method
As shown in Table 51, the sample data on model 2.0 yield a chi-square statistic (
)
of 2966.661 with 1342 degrees of freedom. The associated p-value was determined to
be 0.000 indicating a significant difference between the postulated model and the
sample. However, the problems of the chi-square are well known in its use as a fit
statistic. Kline (2005:136) observed that the chi-square test tends to be affected by the
sample size with a propensity to reject models if the samples are not very large.
Therefore a normed Chi-square value is usually adopted by most researchers (Kline,
2005:137). Normed chi-square is the procedure of dividing the chi-square by the
degrees of freedom. The normed values of up to 3.0 and even 5.0 are recommended
(Kline, 2005:137). The ratio of
obtained for model 2.0 to the degrees of freedom
yield a value of 2.211. This value was lower than the upper limit value of 3.0 and
suggested an acceptable fit of the model.
222
Table 53 presents the fit indexes for model 2.0. The robust RMSEA with 90%
confidence interval was found to be 0.068 (lower bound value = 0.065 and upper
bound value = 0.072). The RMSEA index was just above the upper limit of 0.050 for
the model to be described as a good fit. The 0.065 however indicated that the fit was
adequate. In fact, some authors have argued that a good fitting model could have
RMSEA values of up to 0.06 (Hu & Bentler, 1999:27).
Apart from assessing the RMSEA, the absolute fit index SRMR was also evaluated.
The SRMR was found to be 0.045. The SRMR of 0.045 was much lower than the cutoff value of 0.05 for a good fitting model. This value therefore suggested that there
was a good fit between the model and the sample data.
Additionally, the CFI was evaluated and was found to be close to the lower limit of
0.90 at 0.88. The CFI index was not greater than 0.90 which is the lower limit value
for model acceptance if the CFI is considered in the combination rules. In the current
study, a two statistic model fit evaluation strategy proposed by Hu & Bentler
(1999:28) was followed. Therefore the decision on model fit was based on the SRMR
and the RMSEA fit indexes.
An evaluation of the SRMR, RMSEA and the CFI fit indexes and considering a
combination rule for RMSEA > 0.07 and SRMR > 0.05 which was found by Hu &
Bentler (1999:27) to be very sensitive to mis-specified models, indicated a reasonable
fit of the overall SEM model 2.0 to the sample data. The Lagrange multiplier (LM)
test conducted on the sample data did not reveal any significant indications of model
mis-specification. Because the model fit the sample data adequately, it was now
possible to evaluate the statistical significance of the hypothesised parameters.
223
Table 53: Robust fit indexes for SEM Model 2.0
Fit Index
Cut-off value
Model 1.0
Comment
2966.661
Df
0
1342
Acceptable
CFI
0.9 acceptable
0.88
Barely acceptable
0.045
Good fit
0.068
Acceptable
0.065:0.072
Acceptable range
0.95 Good fit
SRMR
0.08 acceptable
0.05 Good fit
RMSEA
0.08 acceptable
0.05 Good fit
RMSEA 90% CI
8.5.2.3 Model 2.0 hypotheses testingThe rejection of the hypotheses depended on how reasonable the parameter estimates
were in terms of their magnitude, signs and statistical significance. In addition, if the
output revealed estimates that had correlation values greater than 1.00, contained
negative variances and the correlation or covariances were not definite positive then
they were said to be exhibiting unreasonable estimates (Byrne, 2006:103).
Additionally the test statistic had to be greater than 1.96 based on the probability of
5% before the hypothesis could be rejected (Byrne, 2006:103).The test statistic
reported in this study was the parameter estimate divided by its standard error and
therefore it functioned as a Z-statistic to test that the estimate was statistically
different from zero. The significance test was used to evaluate the general hypotheses
H2 to H6. Results of this analysis and an evaluation of the hypotheses will now be
presented in the next section.
Testing the direct influence of client H&S culture on contractor H&S
performance (H2)
It was generally hypothesised that client H&S culture had a direct positive influence
on contractor H&S performance. Specifically the hypotheses which collectively
formed hypothesis H2 were that the factor of client H&S culture, namely:
H2a. leadership, had a direct positive influence on contractor H&S performance;
224
H2b. involvement, had a direct positive influence on contractor H&S
performance;
H2c. procedures, had a direct positive influence on contractor H&S
performance;
H2d. commitment, had a direct positive influence on contractor H&S
performance;
H2e. communication, had a direct positive influence on contractor H&S
performance;
H2f. competence, had a direct positive influence on contractor H&S
performance.
Results from the confirmatory factor analysis of the full structural model, presented in
Table 54, yield support for hypothesis H2c (procedures) and H2d (commitment) but
did not support the hypothesis H2a (leadership), H2b, (involvement), H2e
(communication) and H2f. (competence). The relationship between the factor,
procedures, and contractor H&S performance, was found to be significant at the
probability level of 5% (λ = 0.494, Z= 4.407 and P= 0.000). Similarly, the
hypothesised relationship between the factor, commitment, and contractor H&S
performance, was found to be statistically significant. That relationship yield
significant parameter estimates at 5% probability level with λ = 0.616, Z=2.393 and
P= 0.017.
On the other hand, although the hypothesised relationship between the factor,
involvement, and contractors H&S performance, was significant at 5% probability
level (λ= -0.663, Z= -2.402, P=0.016) the direction was not positive definite. The
result indicated that with the increase in client H&S involvement, contractor H&S
performance decreased by 0.663 units. This result was interesting because it was
expected that with an increase in client involvement there would be an increase in
contractor performance. However , the measurement model on client H&S culture,
revealed high collinearity between commitment and involvement factors of client
H&S culture (Table 27). The high collinearity may probably explain the unreasonable
parameter estimate exhibited for hypothesis H2b, (involvement).
The insignificant relationships were found to be between the competence factor and
contractor H&S performance, (λ = 0.081, Z=0.675 and P=0.500), leadership factor
225
and contractor H&S performance, (λ = 0.204, Z=1.720, P= 0.086) and between the
communication factor and contractor H&S performance (λ = 0.026, Z= 0.212, P=
0.832). Therefore the postulated specific hypotheses for these relationships were not
supported.
However, the general hypothesis H2 which postulated that client H&S culture had a
direct positive influence on contractor H&S performance could not be rejected
because two of the six specific hypotheses were found to be statistically significant
and were definite positive.
Testing the direct influence of client H&S culture on designer H&S
performance (H3)
It was generally hypothesised that client H&S culture had a direct positive influence
on designer H&S performance. Specifically, the hypotheses were that the factors of
client H&S culture, namely:
H3a. leadership, had a positive direct influence on designer H&S performance;
H3b. involvement, had a direct positive influence on designer H&S performance
H3c. procedures, had a direct positive influence on designer H&S performance;
H3d. commitment, had a direct positive influence on designer H&S
performance;
H3e. communication, had a direct positive influence on designer H&S
performance;
H3f. competence, had a direct positive influence on designer H&S performance.
Results from the confirmatory factor analysis of the full structural model presented in
Table 52, yield support for H3c (procedures) and H3e (communication) but did not
support the hypothesis H3a (leadership), H3b, (Involvement), H3d (commitment) and
H3f. (competence). The relationship between the procedures factor and designer H&S
performance, was found to be significant at 5% probability level with λ = 0.439, Z=
3.009, and P= 0.003. Similarly, the hypothesised relationship between the
communication factor and designer H&S performance was found to be statistically
significant. This relationship yield significant estimates at the 5% probability level of
λ = 0.348, Z=3.346 and P= 0.001. The parameter estimates for the two factors of
client H&S culture namely, procedures and communication, indicated that with an
226
increase of one unit in procedures, designer H&S performance increased by about
0.439. Similarly, an improvement of one unit in client communication caused an
improvement of 0.348 in designer H&S performance.
The insignificant relationships were found to be those between the competence factor
and designer H&S performance (λ = -0.196, Z=-1.747, P=0.081), leadership factor
and designer H&S performance (λ = 0.182, Z=1.618, P= 0.106) and between the
factor, commitment, and designer H&S performance (λ = 0.188, Z= 0.681, P= 0.496).
The hypotheses for these relationships were not statistically significant. In addition,
the factors of competence and involvement were found to have a negative relationship
with designer H&S performance. This result was surprising because it was expected
that an increase in client competence and involvement would result in an increase in
designer H&S performance.
However, the general hypothesis H3, which postulated that client H&S culture had a
positive direct influence on designer H&S performance, could not be rejected because
two of the six hypotheses were found to be statistically significant. Further, two more
hypotheses were found to be statistically significant although the signs were not
definite positive.
Testing the direct influence of client H&S culture on project H&S
performance (H4)
The general hypothesis was that client H&S culture had a direct positive influence on
construction project H&S performance. Specifically, the hypotheses which
collectively formed the hypothesis H4 were that the factors of client H&S culture,
namely:
H4a. leadership, had a direct positive influence on project H&S performance;
H4b. involvement, had a direct positive influence on project H&S performance;
H4c. procedures, had a direct positive influence on project H&S performance;
H4d. commitment, had a direct positive influence on project H&S performance;
H4e. communication, had a direct positive influence on project H&S
performance; and
H4f. competence, had a direct positive influence on project H&S performance.
227
Table 54: Parameter estimates and test statistic for model 2.0
Hypothesis
H2a
H2b
H2c
H2d
H2e
H2f
H3a
H3b
H3c
H3d
H3e
H3f
H4a
H4b
H4c
H4d
H4e
H4f
H5
H6
Parameter
CLLP→ CONT H&S PERFORMANCE
CLIP→ CONT H&S PERFORMANCE
CLPP→ CONT H&S PERFORMANCE
CLTP→ CONT H&S PERFORMANCE
CLNP→ CONT H&S PERFORMANCE
CLCP→ CONT H&S PERFORMANCE
CLLP→ DESG H&S PERFORMANCE
CLIP→ DESG H&S PERFORMANCE
CLPP→ DESG H&S PERFORMANCE
CLTP→ DESG H&S PERFORMANCE
CLNP→ DESG H&S PERFORMANCE
CLCP→ DESG H&S PERFORMANCE
CLLP→ PROJ H&S PERFORMANCE
CLIP→ PROJ H&S PERFORMANCE
CLPP→ PROJ H&S PERFORMANCE
CLTP→ PROJ H&S PERFORMANCE
CLNP→ PROJ H&S PERFORMANCE
CLCP→ PROJ H&S PERFORMANCE
CONT → PROJ H&S PERFORMANCE
DESGN → PROJ H&S PERFORMANCE
Unstandardised
Estimate (λ)
Z- statistic
0.228
-0.608
0.474
0.618
0.024
0.090
0.188
-0.132
0.388
0.175
0.298
-0.202
0.018
-0.137
0.127
0.219
0.033
0.095
0.518
0.163
(Robust statistical significance at 5% level)
228
1.738
-2.348
4.282
2.324
0.212
0.675
1.665
-0.538
2.744
0.674
3.299
-1.750
0.270
-0.968
1.359
1.481
0.442
1.122
7.124
2.636
Standardised estimates
Estimate (λ)
Z-Statistic
P-Value
0.204
-0.663
0.494
0.616
0.026
0.081
0.182
-0.155
0.439
0.188
0.348
-0.196
0.016
-0.158
0.139
0.231
0.038
0.090
0.546
0.159
1.720
-2.402
4.407
2.393
0.212
0.675
1.618
-0.540
3.009
0.681
3.346
-1.747
0.270
-0.965
1.371
1.502
0.445
1.141
8.021
2.582
0.086
0.016
0.000
0.017
0.832
0.500
0.106
0.589
0.003
0.496
0.001
0.081
0.787
0.334
0.171
0.133
0.656
0.254
0.000
0.010
The results for these hypotheses presented in Table 52, did not yield support for all
hypothesised direct relationships between the factors of client H&S culture and
project H&S performance. The test statistics revealed that the direct relationships
between the factor, leadership, and project H&S performance had a parameter
coefficient λ= 0.016 and the test statistic Z = 0.270. The probability P was found to be
0.787 for this relationship. The relationship between the factor, involvement, and
project H&S performance yield λ = - 0.158, Z= -0.965 and P=0.334. On the other
hand, parameter estimates for the relationship between the factor, procedures, and
project H&S performance were, λ = 0.139, Z= 1.371 and P = 0.171. The relationships
between the factor, commitment, and project H&S performance (λ = 0.231, Z= 1.502,
P= 0.133) and between communication and project H&S performance (λ = 0.038, Z=
0.445, P = 0.656) were also not significantly different from zero or the null
hypothesis. Therefore, the general hypothesis (H4) which was that client H&S culture
had a direct positive influence on project H&S performance was rejected.
Testing the direct influence of contractor H&S performance on project
H&S performance (H5)
Results of the SEM analysis yield support for the hypothesis that contractor H&S
performance had a direct positive influence on project H&S performance. The test
statistics were significantly different from zero (λ=0.546, Z = 8.02, P=0.000). Given
these results, the hypothesis H5 could not be rejected. The parameter estimate
between contractor H&S performance and project H&S performance indicated that
for every unit improvement in contractor H&S performance, project H&S
performance would improve by 0.546 units. The contractor in this case is referring to
the top management of contractors.
Testing the direct influence of designer H&S performance on project
H&S Performance (H6)
The results from the SEM analysis yield support for the hypothesis that designer H&S
performance had a direct positive influence on project H&S performance. The test
statistics were significantly different from zero (λ=0.159, Z = 2.582, P=0.010).
Therefore, the hypothesis H6 could not be rejected given these parameter estimates.
The parameter estimate between designer H&S performance and project H&S
performance indicated that for every unit improvement in designer H&S performance,
229
project H&S performance would improve by 0.159. This coefficient was however
lower than the desired 0.400. Nonetheless, the relationship was found to be
statistically significantly different from zero indicating that designers H&S
performance had an influence on project H&S performance.
Testing indirect influence of client H&S culture on project H&S
performance mediated by contractor and designer H&S performance (H7
and H8)
An indirect relationship is said to exist between two variables if the direct relationship
between the two is completely insignificant or tends to diminish in the face of an
increased indirect significance. The indirect effect results of client H&S culture on
project H&S performance are presented in Table 55 and Table 56 respectively. The
direct relationship between client H&S culture and project H&S performance was
found to be insignificant (Table 54). However, the direct relationship between client
H&S culture and contractor H&S performance was found to be significant. Similarly,
the relationship between client H&S culture and designer H&S performance was also
significant. In addition, the direct influences of contractor and designer H&S
performance on project H&S performance were found to be significant (Table 54).
The indirect effects on project H&S performance by three factors of client H&S
culture, namely: involvement, procedures, and commitment, mediated by contractor
H&S performance were found to be significant at 5% probability level. The
standardised indirect effects of the involvement factor, yield parameter estimates λ= 0.362, Z= -2.335 and P= 0.020. As for the factor, procedures, the estimates were λ =
0.270, Z= 3.877 and P= 0.000 indicating that the effect was significant. The specific
standardised indirect effects of the commitment factor, on project H&S performance
mediated by contractor H&S performance yield parameter estimates λ = 0.337, Z=
2.303 and P=0.021. These estimates too, indicated a significant effect. The effects of
three factors of client H&S culture namely: communication, leadership and
competence, were found to be not significant (Table 55).
Examination of the indirect influence of client H&S culture on project H&S
performance mediated by designers revealed that two relationships were significant.
The indirect effect of the communication factor, on project H&S performance
mediated by designer H&S performance was found to be significant. The standardised
230
parameter estimates of the indirect relationship were: λ = 0.055, Z = 1.977 and P=
0.048. The indirect effect of the procedures factor, had unstandardised parameter
estimates of λ= 0.063, Z= 1.968 and P= 0.049. However, the standardised estimates
for the factor, procedures, were found to be insignificant.
The sum of indirect effects of client H&S culture on project H&S performance
revealed that two factors namely: procedures and commitment, had a significant total
indirect effect on project H&S performance (Table 56). The standardised estimates for
the total indirect effect were found to be λ = 0.340, Z = 4.619 and P= 0.000 for the
procedures factor. The standardised estimates of the total indirect effect of the
commitment factor, were found to be λ= 0.366, Z = 2.052 and P=0.040.
Therefore, client H&S procedures and commitment were found to have a significant
effect and hence influence on project H&S performance mediated by contractor and
designer H&S performance. This finding confirmed the mediatory role that contractor
and designer H&S performance played in model 2.0. In addition, the findings also
confirmed that although the client H&S culture did not exhibit a direct positive
influence on project H&S performance, the findings revealed a significant indirect
influence on project H&S performance.
Therefore the hypothesis that client H&S culture generally had an indirect positive
influence on project H&S performance, mediated by contractor and designer H&S
performance could not be rejected. Specifically, procedures and commitment factors
were found to be statistically significant at 5% probability level.
8.5.2.4 Solution evaluation for the structural model 2.0
The robust fit indexes, SRMR and the RMSEA, met the cut-off index criteria and the
parameter estimates were found to be statistically significant and reasonable. The
postulated structural model 2.0, which hypothesised that client H&S culture had
influence on contractor, designer and project H&S performance, adequately fit the
sample data. In view of the fact that the analysis was confirmatory of the priori model,
there was no need to further improve the structural model at this stage. Further
investigation of alternative models could be a matter for further studies as the current
study was a confirmatory analysis of the priori.
231
Notwithstanding, the LM test did not reveal significant evidence of model misspecification warranting a re-specification. Byrne (2006:113) points out, that for most
models, model improvement is purely a process that attempts to adjust small features
of the sample and does not necessarily add value to that already fitted model. Model
2.0 as presented in Figure 44, was therefore accepted with its level of fit. The line
indicating a direct influence of client H&S culture on project H&S performance is
dotted because this relationship was found to be not statistically significant. Client
H&S culture’s influence on project H&S performance was only statistically
significant when mediated by contractor and designer H&S performance.
Table 55: Specific indirect effects of client H&S culture
Parameter
CLLP→ CONT →PROJ. H&S
CLIP→ CONT →PROJ. H&S
CLPP→ CONT →PROJ. H&S
CLTP→ CONT →PROJ. H&S
CLNP→ CONT →PROJ. H&S
CLCE→ CONT →PROJ. H&S
CLLP→ DESG →PROJ. H&S
CLIP→ DESG →PROJ. H&S
CLPP→ DESG →PROJ. H&S
CLTP→ DESG →PROJ. H&S
CLNP→ DESG →PROJ. H&S
CLCE→ DESG →PROJ. H&S
Unstandardised
Indirect
ZP-Value
effect
statistic
0.118
1.738
0.082
-0.315
-2.258
0.024
0.245
3.748
0.000
0.320
2.216
0.027
0.012
0.211
0.833
0.047
0.667
0.505
0.031
1.412
0.158
-0.021
-0.533
0.594
0.063
1.968
0.049
0.028
0.662
0.508
0.049
1.999
0.046
-0.033
-1.488
0.137
Standardised estimates
Indirect
ZPeffect
Statistic
Value
0.111
1.738
0.082
-0.362
-2.335
0.020
0.270
3.877
0.000
0.337
2.303
0.021
0.014
0.211
0.833
0.044
0.667
0.505
0.029
1.413
0.158
-0.025
-0.534
0.593
0.070
1.948
0.051
0.030
0.665
0.506
0.055
1.977
0.048
-0.031
-1.427
0.153
Table 56: Total indirect effects of client H&S culture
Parameter
CLLP→ PROJ. H&S
CLIP→ PROJ. H&S
CLPP→ PROJ. H&S
CLTP→ PROJ. H&S
CLNP→ PROJ. H&S
CLCE→ PROJ. H&S
Unstandardised
Indirect
Zeffect
statistic
0.149
1.809
-0.336
-1.991
0.308
4.496
0.348
1.987
0.061
0.922
0.014
0.173
232
Standardised estimates
Indirect
ZPeffect
Statistic
Value
0.140
1.810
0.070
-0.387
-2.045
0.041
0.340
4.619
0.000
0.366
2.052
0.040
0.069
0.928
0.353
0.013
0.173
0.862
Significant
at P> 0.05
level
No
Yes
Yes
Yes
No
No
Contractor
H&S
performance
Project
H&S
performance
Client H&S
culture
Designer
H&S
performance
Figure 44: Finalised model 2.0 for client H&S culture influence
Statistically significant relationship
Statistically insignificant relationship
8.5
CHAPTER SUMMARY
The postulation for the overall model was that the external environment had influence
on client H&S culture and which in turn had influence on contractor, designer and
project H&S performance.
SEM results of measurement models were presented in this section. These results
were obtained from an analysis to determine whether the indicator variables
(questionnaire items) actually measured the constructs that they were supposed to
measure. In addition, results were also presented from the evaluation of the submodels to establish whether the theorised number of factors for the sub-models was
working properly before inclusion in the analysis of the full latent variable model.
Further, results on reliability and construct validity were also presented. It was
important to ensure that the measurement model was in a good working order before
the full latent model could be analysed.
233
The analysis of the full latent variable model was conducted by separating the
postulated model in to two parts namely, model 1.0 and model 2.0. The division of the
model into two parts was merely for the purpose of simplicity. Apart from this, the
focus of the study was on the influence of client H&S culture on project H&S
performance. However, it was also necessary to establish the influence of the external
environment on the client H&S culture.
The findings were that the external environment had influence on client H&S culture.
Further, client H&S culture was found to have a direct positive influence on
contractor and designer H&S performance. However, client H&S culture was found
to have an indirect positive influence on project H&S performance that was mediated
by contractor and designer H&S performance. Therefore the finalised overall clientcentred H&S performance improvement model (combining model 1 and 2) was found
to be as presented in Figure 45.
Model 1.0
Model 2.0
Contractor
H&S
performance
External
Environment
Client H&S
culture
Project H&S
performance
Designer H&S
performance
Figure 45: Finalised Client-centred H&S performance improvement model
234
CHAPTER NINE
DISCUSSION OF RESULTS
The purpose of this study was to model construction client H&S culture’s influence
on construction project H&S performance. The research’s primary objectives were to
establish the (a) status of H&S in the construction industry and the role of clients,
designers and contractors top management (contractors) in H&S performance; (b)
current trends on theories of H&S performance improvement in manufacturing based
industries such as the construction industry; (c) role of construction clients and their
influence on project H&S performance (d) client-centred model for H&S performance
improvement (e) validity of the developed client-centred H&S performance model
and (f) identify factors of client H&S culture with significant influence on contractor,
designer and project H&S performance.
The finalised client-centred H&S performance improvement model showed that the
factors of client H&S culture namely: commitment, procedures, and communication
had significant influence on contractor, designer and project H&S performance.
Likewise, the factors of the external environment namely: legislative, economic,
social, professional bodies and materials and technology were found to have a
significant influence on client H&S performance.
9.1
QUESTIONNAIRE SURVEY RESULTS
Results of the hypotheses testing revealed that of the eight general hypotheses, seven
could not be rejected and one was rejected.
9.1.1
External environment’s influence on client H&S culture
The general hypothesis was that the external environment had a direct positive
influence on client H&S culture (H1) which could not be rejected. All six specific
hypotheses which collectively formed the H1 hypothesis could not be rejected. The
specific hypotheses stated that the external environment had a direct positive
influence on the factors of client H&S culture namely: leadership, involvement,
235
procedures, commitment, communication and competence. The results indicated that
at least 62.6% of variance in client H&S culture was explained by the external
environment. The external environment’s influence was weakest on client H&S
involvement compared to the influence on other factors of client H&S culture. The
effect was found to be strongest on the factor, procedures. Generally the findings
suggested that clients were more likely to lead, be involved, set up procedures, be
committed, communicate on H&S issues and develop competence in H&S as a result
of the external environment influence. Specifically, the results suggested that it was
possible for client H&S culture to be modified as a result of external environment’s
influence. It was this change in client H&S culture that was needed for H&S
performance to be realised in the construction industry. Bomel (2001:5.3) observed
that the culture of client organisations presented considerable opportunities for H&S
improvement in the construction industry.
The implications of this finding are that clients may effectively participate in H&S
management and if they do, they would influence project H&S performance
continuously and therefore lead to a general H&S improvement on construction
projects. Research in Southern Africa has shown that despite the acknowledged
significance of clients to H&S performance, clients have not participated significantly
in H&S management (Kikwasi, 2008:58; Musonda & Smallwood, 2008:87; Musonda,
Haupt & Smallwood, 2009:71). Similarly, a study conducted by Loughborough &
UMIST (2003), established that clients give insufficient consideration to H&S despite
their obligations under the CDM regulations. Bomel (2001:5.3) observed that the
culture of client organisations presents considerable opportunities for H&S
improvement in the construction industry. The findings in the current study were
therefore significant in the sense that with an increased incentive to clients to
participate in H&S management through their culture change resulting from all
factors of the external environment, the much desired participation of clients in H&S
management may be realised. Further, the findings make it possible for policy makers
to address factors of the external environment namely, legislative, economic, social,
professional bodies and materials and technology in such a way that the external
environment enables clients’ participation in H&S management.
236
The legislative factor was found to have a significant influence on client H&S culture.
The findings lend support to the comments by INSAG (1991:5), stating that the
manner in which people act is conditioned by requirements set at a high level such as
legislative. The current study also validates the findings by CIDB (2008), that there
was a general perception in the construction industry that the construction regulations
promulgated in 2003 in South Africa seemed to have had a positive impact. This was
also found to be the case in the UK concerning the CDM regulations (CIOB, 2009).
Apart from the legislative influence, economic conditions were also found to
influence clients’ actions. The European Agency for Safety and Health at Work
(2010:186) observe that economic incentives can be effective in promoting H&S. Pan,
Soetanto & Sidwell (2010:16) observed that the economic situation in the UK
influenced the homebuilders (clients) to slow down on the use of cross walls precast
technology despite the benefits that came with the new technology when the housing
markets were promising. The clients’ decision on the type of technology to use in this
case was highly influenced by the economics despite the benefits including the H&S
benefits that would have arisen. In this case, due to a lack of economic incentive in
the method, H&S was the casualty. The situation described by Pan et al. (2010) lend
support to the findings in the current study that the economic situation as is the case
with the legislative framework had a significant influence on client decisions and
hence client H&S performance.
In the same study by Pan et al. (2010), clients, influenced by the new technology were
forced to assume new roles such as producing outline designs, detailed design
coordination, procurement and construction. In other words, they were influenced to
change the way they did things as a result of technology. This again supports the
findings in the current study that the technology and materials factor of the external
environment, had influence on client H&S culture. In fact in projects where the new
technology was deployed, an observation was made that they had experienced a
reduction in the H&S risk and also enhanced the building quality (Pan et al.,
2010:19).
Worker unions (social), as a factor of the external environment, have also been found
to influence H&S in the construction industry and therefore ties in with the findings of
the current study. According to Fraser (2007:15), unions in Australia, influenced a
237
significant improvement in regulations concerning workers’ H&S. It would appear
therefore that client H&S culture could be enhanced with an increased participation
from the social economic environment such as the workers union.
Therefore the finding that the external environment exhibited a direct positive
influence on client H&S culture not only validated what other authors have stated
before but it also offered a platform and a set of minimum factors that may be
required to be addressed in order to change or influence client H&S culture. It would
seem that a single approach may not be so successful. The European Agency for
Safety and Health at Work (2010:24), observed that incentives and legislation were
complementary. For example, clients in the UK were slow to take up their
responsibilities on H&S even though the CDM regulations required them to do so
(Baxendale & Jones, 2000:33). It could be argued that what they probably needed was
an incentive from the external environment as established in the current study.
Economic, social, technology and the legislative environment all have to be
supportive of client H&S culture.
9.1.2
Influence of client on contractor H&S performance (H2)
The findings suggested that client H&S culture had an influence on contractor H&S
performance. Two of the six specific hypotheses which collectively formed the
hypothesis that client H&S culture had a direct positive influence on contractor H&S
performance were found to be statistically significant. The two hypotheses related to
the influence of the factors of client H&S culture namely: procedures and
commitment. The indicator variables for the procedures factor included the client to:

have programmes to monitor and analyse H&S implementation;

have clear project H&S goals,

schedule H&S as a key contract prequalification criteria for all parties to be
involved in a project;

schedule H&S in all contracts;

conduct regular H&S performance measurement,

have their own H&S committee,

conduct hazard identification and risk assessment; and

require designers to adequately address H&S in the designs.
238
The finding was that the influence of the procedures factor of client H&S culture, on
contractor H&S performance was statistically significant. This finding in the current
study, supported the findings of Huang & Hinze, (2006a) who in their study observed
that projects where owners (clients), tracked the individual H&S performances of
each contractor on site had significantly better H&S performances.
Although Huang & Hinze, (2006a) only referred to one indicator variable namely;
performance
measurement
(programmes
to
monitor
and
analyse
H&S
implementation), their study lends support to the findings in the current study that
clients need to have clear procedures if they were to influence contractor H&S
performance.
The indicator variables for the commitment factor were inter alia, for the client to:

demonstrate a positive H&S attitude;

actively promote H&S;

provide adequate resources for H&S implementation;

put in effort to routinely evaluate H&S in all work schedules;

set up incentives for good H&S behaviour;

set H&S as a major agenda item in project meetings;

actively monitor H&S programmes;

always attend H&S meetings on the construction site;

conduct H&S inspections and audits; and

always be involved in accident or incident investigations.
The finding was that the commitment factor of client H&S culture had a statistically
significant influence on contractor H&S performance. The findings in the current
study was supported by observations made by Toellner (2001), Wiegmann (2002)
Mohamed (2003), Ng et al. (2005), Cameron & Duff (2007), and Choudhry et al.
(2009) who found that management commitment was critical to H&S performance.
Although reference was made to commitment of top management of contractors by
these authors, the principle remained that commitment was crucial. There is little
evidence of studies that have been conducted to evaluate the effect of client
commitment on contractor H&S performance.
239
The influence of the factors of client H&S culture, namely: leadership, involvement,
communication and competence, was found to be statistically insignificant. This
finding was surprising because those factors were expected to have influence on
contractor H&S performance. However, statistical significance can be greatly affected
by the sample size and the type of population that is sampled. It would however be
interesting to discover what the results could be in another study with a different
sample type and size.
However, the involvement factor was found to have a negative relationship with
contractor H&S performance. The involvement factor was defined by indicator
variables which inter alia required the client to:

personally be active in critical project H&S activities:

always be present in project H&S meetings;

contribute to H&S training;

actively oversee H&S on critical operations;

constantly stay in touch on H&S issues;

always communicate information on H&S to all parties and

conduct regular audits and inspections.
The findings revealed that when clients increased their level of undertaking activities
that defined the involvement factor, contractor’ H&S performance reduced. This
finding was surprising as contractor’s performance was expected to improve with
client involvement. However, it was realised that the result could have been
influenced by a thin differentiating line between the involvement and commitment
factor. Results in the current study revealed a high collinearity between these two
factors. The involvement factor had a correlation value higher than 0.9 with the
commitment factor. It was therefore speculated that the unreasonable result where
client involvement caused a reduction in contractor H&S performance may probably
have been a result of the high collinearity between the two factors. This may in fact be
the reason why some authors use and refer to the two factors as being one factor,
namely: commitment and involvement (Harvey et al., 2002:31; Mohamed, 2003:82)
and many other authors only identify the commitment factor (INSAG, 1991; Cooper,
2000; Guldenmund, 2000; Biggs et al., 2005; Huang & Hinze, 2006a; FernándezMuñiz et al., 2007; Misnan & Mohammed, 2007). These studies seem to suggest that
240
management (client) involvement may be a subset of commitment. In fact getting
involved or to actively participate in H&S programs may entail commitment
(Molenaar et al., 2009:494).
In situations such as the one described above where two factors have high correlation
values between them, Tabachnick & Fidell (2007) advises that one factor should be
adopted. Therefore since there was a high correlation between commitment and
involvement, it made sense that the commitment factor be adopted.
The findings on the hypothesis that client H&S culture had a direct positive influence
on contractor H&S performance entails therefore that the minimum that the client
could do in order to significantly influence contractor H&S performance was to have
procedures in place and to be committed to H&S performance.
The findings offered a minimum requirement that could be used by clients seeking to
influence contractor H&S performance. A checklist of items defining the factors of
procedures and commitment could ensure that clients meet the basic required criteria
to influence contractor H&S performance.
9.1.3
Influence of client on designer H&S performance (H3)
The finding on H3 was that client H&S culture had influence on designer H&S
performance. Two of the six specific hypotheses which collectively formed the
hypothesis that client H&S culture had a direct positive influence on designer H&S
performance, were found to be statistically significant. The two hypotheses were that
the procedures and communication factor had direct positive influence on designer
H&S performance.
The procedures factor, was defined by indicator variables being, the client to

have programmes to monitor and analyse H&S implementation;

have clear project H&S goals;

schedule H&S as a key contract prequalification criteria for all parties to be
involved in a project;

schedule H&S in all contracts;

conduct regular H&S performance measurement; and
241

have own H&S committee, conducting hazard identification and risk
assessment and require designers to adequately address H&S in the designs.
The finding was that the influence of the procedures factor, on contractor H&S
performance was statistically significant. This finding in the current study validated a
proposal by Bomel (2004:xi) that designers could be able to design for H&S with the
mobilisation of client influence. Designing for H&S was an aspect of H&S
performance by designers. It would appear that if clients had clear programmes on
H&S, it would be easier for designers to perform on their H&S obligations. The
clients having procedures in place also had a significant effect on designer H&S
performance.
The communication factor, was defined by the variables being inter alia, the client to

have formal reporting system of incidents and accidents;

involve all parties in planning for H&S on the project;

involve all parties to review H&S;

provide timely feedback on reported accidents and incidents;

communicate risk findings to all parties on the project;

have clearly outlined H&S roles and responsibilities;

have clearly communicated expected performance on H&S to all; and

provide information on H&S risk control to all parties.
The effect on designer H&S performance by the client performing these activities
(Communication), was found to be statistically significant. The direct positive
influence of the factors of client H&S culture namely: leadership, involvement,
commitment and competence were found to be statistically insignificant. This finding
was unexpected considering that these factors were identified in the Delphi study to
have influence on the general H&S performance.
Surprisingly, the involvement factor had a negative effect on designer H&S
performance. The indicator variables for the involvement factor were that the client:

is personally active in critical project H&S activities;

is always present in project H&S meetings;

contributes to H&S training;
242

is active in overseeing of H&S on critical operations;

has constantly stayed in touch on H&S issues;

always communicates information on H&S to all parties and

conducts regular audits and inspections.
The findings revealed that when clients increased their level of undertaking these
activities, designer H&S performance decreased. Possible explanation for this result
has been discussed in the previous sections. This relates to the commitment and
involvement factors. The findings regarding the hypothesis that client H&S culture
had a direct positive influence on designer H&S performance entailed therefore that
the minimum that the client could do in order to significantly influence designer H&S
performance was to have procedures in place and to provide effective communication
on H&S. However, although the influence of leadership, involvement, commitment
and competence factors was not statistically significant, they remain important to
influence designer H&S performance and part of the factors of client H&S culture.
The findings offer a minimum requirement that could be used by clients seeking to
influence designers’ H&S performance. A checklist of items defining the factors,
procedures and communication, could ensure that clients meet the basic required
criteria to influence designers’ H&S performance. Designer H&S performance
especially designing for H&S is critical to the overall project H&S performance and
therefore one way to achieve that is to mobilise clients’ influence (Bomel, 2004:xi).
9.1.4
Influence of client on project H&S performance (H4, H7 and H8)
The finding was that client H&S culture had an indirect positive influence on project
H&S performance. This influence was mediated by contractor and designer H&S
performance. The effects on project H&S performance of two factors of client H&S
culture, namely procedures and commitment were found to be statistically significant.
However the direct positive influence of client H&S culture on project H&S
performance was found to be statistically insignificant. None of the factors of client
H&S
culture
namely:
leadership,
involvement,
procedures,
commitment,
communication and competence had a statistically significant direct positive influence
on project H&S performance.
243
Nonetheless, findings indicated that client H&S culture was important to project H&S
performance particularly the factors with statistical significance, namely procedures
and commitment. The procedures factor was made up of, the client:

having programmes to monitor and analyse H&S implementation;

having clear project H&S goals;

scheduling H&S as a key contract prequalification criteria for all parties to be
involved in a project;

scheduling H&S in all contracts;

conducting regular H&S performance measurement;

having own H&S committee;

conducting hazard identification and risk assessment; and

requiring designers to adequately address H&S in the designs.
On the other hand, client commitment can be seen through clients;

demonstrating a positive H&S attitude;

actively promoting H&S;

providing adequate resources for H&S implementation;

putting in effort to routinely evaluate H&S in all work schedules;

setting up incentives for good H&S behaviour;

setting H&S as a major agenda item of each and every project meeting;

actively monitor H&S programmes;

always attending H&S meetings at the construction site;

conducting H&S inspections and audits, and

always involved in accident or incident investigations.
Therefore, the finding that client H&S culture had an indirect positive influence on
project H&S performance validates an observation made by Bomel (2001:5.3) that
clients’ culture offered an opportunity upon which H&S performance could be
improved on construction projects.
The findings also suggest that project H&S performance and improvement may not be
achieved by focusing on one party such as the contractor, or designer or indeed the
client. However, all parties are critical to achieving the desired H&S performance.
244
The influence of client H&S culture on project H&S performance was found to be an
indirect one and was mediated by both designer and contractor H&S performance.
Furthermore, although the direct influence of designers and contractors was found to
be significant, the results suggested that these two factors also needed influence from
client H&S culture. In order to continuously achieve or improve project H&S
performance, client H&S culture was found to be necessary.
The importance of having conducted a structural equation modelling analysis to
determine the influence of client H&S culture on the contractor, designer and project
H&S performance, was that it was possible to identify specifically which factors of
client H&S culture had a causal effect and direction of that effect as opposed to the
general blanket statement that client H&S culture had influence on project H&S
performance.
9.1.5
Influence of contractor on project H&S performance (H5)
The finding was that contractor H&S performance (top management) had a direct
positive influence on project H&S performance. Contractor H&S performance was
defined by indicator variables that defined three factors, namely: procedures,
commitment and communication.
The finding suggested that contractor H&S performance had a direct positive
influence on the overall project H&S performance. The finding was consistent with
the Lin & Mills (2001) study, which found that when contractors scored highly in the
management responsibility and H&S system elements, their total H&S standards
tended to be higher.
Once again the study result highlighted the role of contractors to influence the overall
project H&S performance. The results also highlighted the fact that it was beneficial
to start looking at H&S performance at the top management level as opposed to
concentrating at factors found at the construction stage only. This significance was
also highlighted by Jaselskis et al. (1996:69), who argued that management
characteristics, H&S meetings and budget allocations improved H&S performance
and activities are top management functions.
245
9.1.6
Influence of designer on project H&S performance (H6)
The finding was that designer H&S performance had a direct positive influence on
project H&S performance. Designer H&S performance was defined by indicator
variables that described three factors, namely: procedures, commitment and
communication.
The finding suggested that designer H&S performance had a direct positive influence
on the overall project H&S performance. The finding was consistent with that of
Behm (2005) and Gambatese et al. (2006) who found a link between the design and
construction site injury and fatality incidents. However, in those studies, reference
was only made to the design aspect whereas in the current study; design was only one
aspect through which project H&S performance could be influenced.
Once again the current study highlighted the role of designers to influence the overall
project H&S performance. This finding was against the backdrop of a lack of
understanding by some designers as to the extent of their influence on H&S
performance (Gambatese, 1997:654; Toole, 2005:204). Therefore, the finding was
significant in that it provides designers with knowledge of the fact that designers have
influence on H&S performance not only in the negative sense but rather in a positive
way.
Similarly, the finding was significant because for clients, who are employers of
designers, designer H&S performance framework, could constitute a checklist of
activities that could be used to evaluate designers on, in order to ensure acceptable
project H&S performance. Similarly, by using the designer H&S performance
framework, designers would have knowledge of activities that have to be
implemented in order to ensure an acceptable level of project H&S performance.
These activities could also constitute leading indicators for all stakeholders involved
in a project.
9.2
QUESTIONNAIRE AND DELPHI SURVEY RESULTS
9.2.1
Direct external environment’s influence on client H&S culture
Findings from the Delphi study were that the external environment had influence on
client H&S performance. Clients were ‘very likely’ to implement H&S elements as a
246
result of the external environment influence. In addition, the impact of external
environment factors namely: legislative, economic, social, materials and technology
and professional bodies were found to be significant. This finding from the Delphi
study was validated by the questionnaire survey which was analysed and modelled
using the SEM method.
The questionnaire survey results indicated that the external environment had a direct
positive influence on client H&S culture. This finding was consistent with the Delphi
study finding and also with other studies that have identified aspects of the external
environment to be of influence on client H&S performance. These include legislative
(CIDB, 2008; CIOB, 2009), economic (European Agency for Safety and Health at
Work, 2010, Pan et al., 2010), social (union influence) (Fraser, 2007) and materials
and technology (Baxendale, 2000).
9.2.2
Direct client’s influence on contractor H&S performance
The findings from the Delphi study were that client H&S culture had influence on
contractor H&S performance. Results were that contractors were ‘very likely’ to
implement H&S elements with the influence of client H&S culture. The factors of
client H&S culture namely: leadership, involvement, commitment, and competence
were found to have ‘major’ impact significance on contractor H&S performance.
The Delphi study was validated by a field questionnaire survey. The results suggested
that, client H&S culture had influence on contractor H&S performance. In the
questionnaire survey, the hypothesis that client H&S culture had a direct and positive
influence on contractor H&S performance, could not be rejected because two out of
the six specific hypotheses could not be rejected. The factors of client H&S culture
namely: commitment and procedures were found to have a statistically significant
influence on contractor H&S performance. Findings from both the Delphi and the
questionnaire survey therefore suggested that client H&S culture had influence on
contractor H&S performance.
The advantage of using SEM modelling, to validate the Delphi findings, was that it
was possible to specifically identify which factors of the client H&S culture construct
had significant influence on contractor H&S performance. Therefore instead of
making a general blanket statement that client H&S culture had influence on
247
contractor H&S performance, it was possible to specifically state that the factors
namely: commitment and procedures, had a statistically significant influence on
contractor H&S performance.
9.2.3
Direct client’s influence on designer H&S performance
As for the Delphi findings on the influence of client H&S culture on designer H&S
performance, results revealed that clients had influence on designer H&S
performance. Designers were ‘likely’ to implement H&S elements as a result of client
H&S culture’s influence. Furthermore, the impact of the factors of client H&S culture
was found to be of ‘major’ impact significance on designer H&S performance. The
factors included: commitment and involvement, competence and leadership.
The findings from the Delphi study were validated by the questionnaire survey. The
questionnaire survey results revealed that there was a certain level of influence from
client H&S culture on designer H&S performance. In the questionnaire survey, the
hypotheses that the factors of client H&S culture namely: communication and
procedures had a direct and positive influence on designer H&S performance, could
not be rejected. These relationships were statistically significant.
Two factors, namely procedures and communication, were not analysed at the Delphi
stage. These factors were included at the questionnaire survey stage on
recommendation from the Delphi panel. The finding however from the current study
was that client H&S culture had influence on designer H&S performance. This
finding validates an argument that for designers to effectively contribute to H&S
implementation, clients’ influence and commitment was necessary (Bomel, 2004:xi;
Hecker, Gambatese & Weinstein, 2005:43). The argument in this study is that the
recognition of the need to mobilise client influence can only come about from the
belief that clients have influence. The current study has therefore consolidated that
view and shown that clients have influence on designer H&S performance especially
if clients have laid out procedures and communicate effectively on H&S.
9.2.4
Direct client’s influence on project H&S performance
The findings from the Delphi study revealed that client H&S culture had influence on
project H&S performance. Furthermore, the indirect influence on project H&S
performance of client H&S culture through influence on contractors and designers
248
was found to be greater than the direct client H&S culture influence. Similarly results
from the questionnaire survey, also revealed that clients had an indirect statistically
significant influence on project H&S performance.
The factors of client H&S culture that were found to have a statistically significant
indirect influence on project H&S performance were procedures and commitment. As
stated earlier, the advantages of using SEM modelling to analyse complex
relationships lies in being able to specifically determine the factors that have a
statistically significant influence as well as the direction of the influence. In the
current study it has been shown that client H&S culture has influence on project H&S
performance. However this influence was found to be an indirect one mediated by
contractor and designer H&S performance.
9.3
CHAPTER SUMMARY
In summary therefore, the findings from the questionnaire survey generally supported
the predictions that were made by the experts from the Delphi study. The validated
predictions were that the external environment has influence on client H&S
performance and that client H&S culture has influence on designer, contractor and
project H&S performance.
In addition, existing literature lends support to the findings of the current study. The
supported findings were that the factors, commitment and procedures, are critical to
project H&S performance. Further, communication was necessary for effective H&S
management.
249
CHAPTER TEN
CONCLUSIONS
10.1
CONCLUSION ON THE OVERALL STUDY
The broad overall objective of the current study was to develop a client-centred H&S
performance improvement model and specifically identify critical factors of client
H&S culture with significant effect on project H&S performance improvement on
construction projects.
In order to achieve that general objective, the study adopted a mixed methodology of
conducting an extensive literature review, a Delphi study and a field questionnaire
survey which was analysed using structural equation modelling. The field
questionnaire survey was conducted in order to validate findings from the Delphi
study. Conclusions regarding the study are presented relative to the objectives of the
study.
10.1.1 Objective GO1
The first objective of the study was to establish the status of H&S in the construction
industry based on literature and also the role of clients, designers and the top
management of contractors in H&S performance. In order to achieve this objective, a
review of literature was conducted. Findings were that the construction industry
continues to lag behind other industrial sectors on H&S performance. H&S
performance in the construction industry remained a problem and this had resulted in
a negative impact on the image of the industry.
The literature also informed that clients offered potential for H&S performance
improvement. Similarly, designers and top management of contractors were identified
as being critical to H&S performance in the construction industry. Findings from
literature were that more research and effort was required to try and address the
problem of H&S in the industry.
250
10.1.2 Objective GO2
The second objective was to investigate the current trends in H&S performance
improvement. A review of literature was carried out in order to achieve this objective.
The findings revealed that culture offered more prospects to improve H&S
performance than any other approach. It was clear from literature that change in
culture would achieve the goal of accident and incident reduction as opposed to
isolated strategies. There was however no consensus on the definitions and
measurement of H&S culture in literature. This current study therefore offered a
synthesised definition of what culture was and presented a framework of H&S
culture. From the synthesised literature on culture, this current argued that it was
possible to characterise the prevailing H&S culture in an organisation or in a project.
10.1.3 Objective GO3
The third objective was to establish the impact of construction clients and their
potential contribution to and influence on project H&S performance.
A Delphi study was conducted in order to achieve this objective. Findings were that
clients had a significant impact on contractor, designer and project H&S performance.
Experts predicted that contractors and designers were likely to implement H&S
elements with client influence. This finding indicated that the potential influence and
contribution of clients to H&S performance, was significant. In addition, experts
predicted that clients were very likely to implement H&S elements with the external
environment’s influence.
10.1.4 Objective GO4
The fourth objective of the study was to develop a conceptual client-centred H&S
performance improvement model based on both literature and the Delphi study.
A synthesis of the reviewed literature together with the findings from the Delphi study
was used to achieve this objective. The conceptual model theorised that the external
environment had influence on client H&S culture. In addition, client H&S culture had
influence on contractor, designer and project H&S performance. This hypothesis was
validated through a structural equation modelling of data from the field questionnaire
survey.
251
10.1.5 Objective GO5
The fifth and final objective of the study was to test and validate the conceptual
client-centred H&S performance improvement model by conducting a questionnaire
survey and analysing it using SEM. A questionnaire survey and analysis of the results
using SEM software, EQS version 6.1 and MPlus version 6.0 was conducted in order
to achieve this objective.
Findings were that client H&S culture had influence on contractor, designer and
project H&S performance. In addition, the external environment had significant
influence on client H&S performance. These findings validated the conceptual clientcentred H&S performance improvement model developed from literature and the
Delphi study.
10.2
CONTRIBUTION AND VALUE OF THE RESEARCH
The value and contribution of the research is described at three levels. These are the
methodological, theoretical and practical levels of the research findings.
10.2.1 Methodological

Most studies have used univariate statistical analysis methods such as
ANOVA or MANOVA to model clients’ influence on project H&S
performance. However, the current study used SEM which is more robust and
superior to the methods mentioned to determine causality of factors in a model
and their direction of influence (Kline, 2005:14).

With SEM analysis, it was possible to identify the factors of client H&S
culture which had significant effect and hence influence on project H&S
performance as opposed to a general blanket statement that clients had
influence on project H&S performance.

In order to measure H&S performance, a shift was made away from the use of
accidents data and incidents (TRIR) to more acceptable leading indicators.
These leading indicators were validated during the checking of measurement
models of the SEM.

The questionnaire survey instrument had high internal reliability values and
therefore could be used in similar studies to validate the current study or for
similar purposes.
252

Findings from the Delphi study and a conceptual model developed from both
literature review and the Delphi study was validated by conducting a
questionnaire survey. Data from the questionnaire survey was analysed using
SEM software, EQS version 6.1 and MPlus version 6.0. As a result of this
mixed method, a parsimonious model was developed.
Apart from the study contributing and adding value to the body of knowledge in terms
of the methodological approach, a contribution to theory was also achieved.
10.2.2 Theoretical

The results of the SEM analysis indicated that the external environment had
influence on client H&S culture generally and specifically on all the factors of
client H&S culture. The factors of client H&S culture that the external
environment had significant influence over were: leadership, involvement,
procedures, commitment, communication and competence. The researcher
could not find evidence of a similar study that has been conducted in the
construction industry.

The SEM results also indicated that client H&S culture had influence on
contractor, designer and project H&S performance. Specifically, the factors of
client of H&S culture, namely: procedures and commitment had significant
influence on contractor H&S performance. On the other hand, the factors,
procedures and communication, had significant influence on designer H&S
performance. The factors, procedures and commitment, had a significant
influence on project H&S performance although the influence was an indirect
one through contractor and designer H&S performance. This finding enforced
the theory that all stakeholders should be involved in H&S management.

Indicator variables could be used to characterise client H&S culture,
contractor and designer H&S performance. They could also be used for H&S
performance measurement.

A comprehensive review of literature on H&S in the construction industry,
H&S performance improvement, H&S culture and the role of construction
clients on H&S management was undertaken. This review provides a
253
synthesised compilation of the status of H&S in the industry and trends in
H&S performance improvement.

Literature review did not reveal evidence of a similar study to the current one
and therefore suggested that this type of research has not yet been conducted
in the construction industry especially in Southern Africa.

Furthermore, there was no evidence that suggested that a mixed method of
using a Delphi and SEM had been used in the construction industry in
Southern Africa. Therefore this study may offer a base for other researchers to
use for other follow up studies.

The current study modelled the influence of client H&S culture on contractor,
designer and project H&S performance. Previous studies have tried to model
client involvement’s influence on project H&S performance. However,
involvement is only one aspect of client culture. This study has shown that
there is more than one factor that can influence project H&S performance.
10.2.3 Practical

Clients, especially in Southern Africa, have not realised the significance of
their participation in H&S management. However Delphi results have
indicated that contractors and designers were likely to implement H&S
elements on a project with clients’ influence. Further, SEM results indicated
that client H&S culture had influence on contractor, designer and on project
H&S performance.

Knowledge of the influence of client H&S culture particularly the influence of
the factors namely: procedures, commitment and communication on
contractor, designer and project H&S performance could help clients to plan,
organise, coordinate and control all aspects relating to H&S implementation
on construction projects. Clients could use this knowledge to help with
decisions on how to allocate financial and human resources on H&S
performance.

Similarly, designers have not realised the significance of their participation in
H&S over and above the task of simply designing for H&S. Both the Delphi
and the SEM results suggested that designers had a significant influence on
project H&S performance. Therefore this knowledge is essential for those that
254
train designers and those that award professional status to designers. In
addition, designers themselves may use this knowledge when it comes to
determining their role, allocation of resources and level of participation in
project H&S implementation.

The practical significance of the findings for contractors is that the knowledge
of contractor’s influence namely, top management, means that addressing
H&S cannot be a project level issue only but that it should be a strategic
management issue as well.

As for policy makers; professional bodies; researchers into new materials and
technology and the organised labour or civil societies, the knowledge that
clients have influence on project H&S performance, offers an opportunity for
them to get involved in H&S management by insisting that clients should be
more committed. As a result of this knowledge, appropriate policies and
regulations could be formulated that seek to influence clients’ H&S
performance. In addition, research into newer materials and technology that
would result in clients making decisions that are favourable to H&S might be
buoyed by the findings of this study. In addition, campaigns for better worker
H&S can be successfully directed at the appropriate stakeholders as well as to
check on the requirements for an acceptable level of H&S performance.

The knowledge of the fact that designers have a significant influence on
project H&S performance may enable clients to insist on designers’ full
participation. Therefore the client may use a list of indicator variables as part
of a design or commission brief to designers.

The use of client H&S culture, contractor and designer H&S performance
indicator variables as an evaluation tool is possible in order to determine
whether the required H&S elements were in place to ensure an acceptable
H&S performance standard on a project.

The study offers an opportunity for further research to improve the model
developed in this study and probably refine indicator variables to suit specific
environments.
Therefore the implications for practice of all these areas in which the current study
may add value and contribute were considered to be many.
255
10.3
IMPLICATIONS FOR PRACTICE
As a result of the identified contributions that the current study may make, the
following implications for practice have been identified

The policy makers, researchers and the social environment especially the
organised labour (unions) should not continue to be on the fringe when it
comes to H&S performance improvement. These publics have a significant
effect on client H&S culture and hence performance.

Clients or owners of construction projects may not continue to leave H&S
performance to contracting organisations alone. Findings indicated that clients
have influence on project H&S especially when they have procedures in place
and are committed.
Contractors are very likely to implement H&S
programmes with client’s influence.

Designers should not continue to doubt their contribution to H&S performance
and will also realise that their involvement may go beyond the minimum
requirement of simply designing for H&S. The effect of designer H&S
performance on project H&S performance was found to be significant.

Contractors’ top management have a reason to take H&S to strategic
management level as opposed to relegating H&S to be a project level
responsibility issue.

Professional bodies and those institutions that regulate the construction
industry have a basis upon which to revise standard contract documents, clear
definition of duties of all stakeholders in the industry and a clear identification
of project deliverables that are inclusive of H&S.
Recommendations are made from the methodological, theoretical and practical points
of view.
10.4
RECOMMENDATIONS
10.4.1. Methodological

It is recommended that a mixed method of using a Delphi and questionnaire
survey be encouraged in studies such as the current one where a test – retest
methodology may not be feasible to validate a study. This situation is common
256
in engineering and construction management studies and most studies end at
Delphi or questionnaire surveys and as such renders generalisation of
conclusions especially on causality to be questionable. The recommended
method could commence with a Delphi study followed by a questionnaire
survey or vice-versa in order to validate a study and therefore improve its
generalisability.

There are many research studies in engineering and construction management
that try to establish cause - effect relationships between different latent
constructs. However most of these studies use inadequate analytical methods
such as ANOVA and multiple regressions. ANOVA and even multiple
regressions which are basically standard statistical procedures do not offer a
convenient and a straightforward way to test a hypothesis at a higher level of
abstraction (Kline, 2005:14). Therefore, for similar studies such as the current
study, SEM is recommended to be used as the analysis method for better
results.

It is also recommended that a similar study should be conducted with a
different population and sample size to improve its application in the
construction industry.

More research should be conducted on the indicator variables to establish any
improvement in model fit as the current study was purely a confirmatory
factor analysis. There is a possibility that client H&S culture could be defined
by more indicator variables just as contractor, designer and project H&S
performance could be. Recognition should be made however that there is no
such a thing as a perfect model fit. Nonetheless, there should be a move to try
and improve on the current model rather than invent a new model.
10.4.2 Theoretical

It was observed from literature that there were still different definitions and an
understanding of the concept of H&S culture. This has led in the past to a
limited utilisation of the concept to improve H&S in the construction industry.
Further, there has not been consensus on how H&S culture should be
measured. However, in the current study, literature was reviewed and
synthesised on the concept of H&S culture. Coupled with expert knowledge
257
obtained through the Delphi study, a six factor H&S culture model was arrived
at especially for the clients. Those factors were identified as leadership,
involvement, procedures, commitment, communication and competence
(LIP+3C). This client H&S culture model was tested and was found to fit the
sample data. The same factors could also be used to characterise contractor,
designer and or project H&S culture. It is therefore recommended that the
developed model and theory of H&S culture, with particular emphasis on
operationalising it, form a basis for further refinement of the concept and
thereby make it beneficial.

It is recommended that the influence of client H&S culture be integrated in
H&S performance improvement models that have been already proposed in
other studies but do not include client H&S culture’s influence.
10.4.3 Practical

It is recommended that clients, designers and contractors realise which
indicator variables constitute their own H&S performance and hence influence
project H&S performance.

It is further recommended that clients should know what elements define
contractor and designer H&S performance so that they may be able to include
them in design briefs as deliverables as well as in prequalifying and contract
documents. Most contract documents that have been revealed do not
breakdown H&S deliverables and therefore providing for H&S remains vague.
In addition, knowing what elements to include could help in preparing
estimates for H&S implementation on particular projects.

All stakeholders should know which indicator variables define project H&S
performance so that H&S performance planning, organising, monitoring,
measurement and control could be possible.

As opposed to measuring lagging indicators of H&S performance namely:
accidents and incidents, use of leading indicators such as the indicator
variables used in the current study should be encouraged. Leading indicators
are a pro-active way to ensure improvement in H&S performance and as a
result they do not necessarily depend on project type for example.
258
10.5
LIMITATIONS
The following limitations regarding the current study should be considered, namely:

Research was only conducted in Johannesburg and Gaborone. Given enough
resources, it would be preferable to conduct a similar research study with a
wider population.

Although by SEM requirements the sample size of 281 cases could be
described as large, a larger sample size exceeding 500 cases would have
benefited the study considering the complexity of the model analysed in the
current study.

Several nested models especially for the measurement models, could have
been evaluated to check out the suitability of other alternative models. The
current study was purely confirmatory in nature.

Although the internal reliability tests indicated high internal consistency and
therefore a well-constructed research tool, some constructs revealed high
correlation values such as the relationship between commitment and
involvement. This may be due to the fact that only one questionnaire was used
to collect information relative to clients, designers and contractors. A review
of the research tool may benefit findings in this study.
10.6
SUGGESTIONS FOR FURTHER RESEARCH
The following suggestions for further research have been identified:

Research on the relationship between contractor and designer H&S
performance and this relationship’s influence on project H&S performance
would have benefited the current study.

An SEM evaluation of a direct influence of the external environment factors
namely: legislative, economic, social, professional, technology and materials’
influence on client H&S culture would have given insight on the level of
influence of each one of these factors. Although this was done during the
Delphi study, the same evaluation at an abstract level would have been an
additional contribution to knowledge in this aspect.

The direct influence of the external environment on project H&S performance
with all stakeholders playing a mediating role was not analysed in the current
259
study for the purpose of model parsimony. An SEM evaluation of direct or
indirect influence on project H&S performance could have provided insight on
the magnitude of external environment influence specifically on project H&S
performance.

10.7
Validate the Musonda H&S culture framework presented in Figure 7.
CHAPTER SUMMARY AND CONCLUSION
A client-centred H&S performance improvement model was developed using H&S
performance improvement theories grounded in H&S culture. It was theorised that
client H&S culture had influence on construction project H&S performance as well as
on contractor and designer H&S performance. It was also postulated that client H&S
culture was a six factor construct made up of leadership, involvement, procedures,
commitment, communication and competence (LIP+3C) factors. Further it was
postulated that the external environment had influence on client H&S culture.
The postulated models were analysed with the SEM software EQS version 6.1 and
MPlus version 6.0. The fit statistics for the measurement and structural models had an
adequate fit to the sample data. The finalised empirical model revealed that the
external environment factor had a statistically significant influence on client H&S
culture. In particular, the external environment explained more variance in the
procedures factor of client H&S culture than in any other factors. Further, the
finalised empirical model revealed that client H&S culture had influence on
contractor, designer and project H&S performance. Specifically, the procedures and
commitment factors had a statistically significant influence on contractor H&S
performance while the factors, procedures and communication, had a statistically
significant influence on designers. Client H&S culture had an indirect influence on
project H&S performance. This influence was mediated by contractor and designer
H&S performance. The factors of client H&S culture namely, procedures and
commitment had a statistically significant influence on project H&S performance.
The findings have theoretical value because respondents were drawn from client,
contractor, designer and subcontractor organisations. Further, respondents had
working knowledge of the projects that they were reporting on. In addition, the
questionnaire survey whose results were modelled using the SEM was a validating
260
study of a conceptual model developed from synthesised theories established from
literature and more importantly from the Delphi study.
The current study, lends support to other studies that have utilised alternative methods
to establish client’s influence on project H&S performance. These studies have
concluded that clients can influence H&S performance on construction projects. The
current study utilised a more robust modelling method, of SEM. By adopting that
methodology, the current study was able to specifically model client influence on
project H&S performance and also identify client H&S factors that were statistically
significant. The practical implication was that, it was possible to assure H&S
performance in a construction project by ensuring that client factors with a significant
influence were implemented.
261
REFERENCES
Abdelhamid, T.S. and Everett, J.G. (2000). Identifying root causes of construction
accidents. Journal of Construction Engineering and Management, 126(1):52-60.
ACRCCI (2001). Industry culture: A need for change. Australian Cooperative Research
Centre for Construction Innovation.
Agarwal, P. and Everett, J.G. (1997). Strategies for construction contractors to reduce
workers' compensation costs. Journal of Management in Engineering, 13(5):70-75.
Ahasan, R. (2001). Legacy of implementing industrial health and safety in developing
countries. Journal of physiological anthropology and applied human science,
20(6):311-319.
Alinaitwe, H.M. (2008). An assessment of clients' performance in having an efficient
building process in Uganda. Journal of civil engineering and management,
14(2):73-78.
Alves Dias, L.M. (2004). Occupation safety and health coordination in the construction
industry in European Union countries. Lisbon: International Social Security
Association- Construction Section.
Anderson, J. (2010). Briefing: Regulations to improve construction health and safety.
Proceedings of the civil engineers, 163(August):97-100.
Anvuur, A., Male, S. and Kumaraswamy, M. (2006) Taking forward public
procurement reforms in Ghana. In:, CIB W107 Construction in Developing
Economies International Symposium, "Construction in Developing Economies:
New issues and Challenges", 18-20 January, Santiago, Chile
Arezes, P.M. and Miguel, A.S. (2003). The role of safety culture in safety performance
measurement. Measuring business excellence, 7(4): 20-28.
Atkinson, A.R. (1999). Implications for construction safety studies of the management
of human error. In Implementation of Safety and Health on construction sites.
Edited by Singh, A., Hinze, J. H. and Coble, R. J., Rotterdam: A. A. Balkema.
Baram, M. and Schoebel, M. (2007). Safety culture and behavioural change at the
workplace. Safety Science, 45, 631–636.
Barrett, P. (2007). Revaluing Construction: A holistic model. Building research and
Information, 35(3): 268-286.
Bartholomew, J.B., Loukas, A., Jowers, M.E. and Allua, S. (2006). Validation of the
physical activity self-efficacy scale: Testing measurement invariance between
Hispanic and Caucasian children. Journal of physical activity and health, 3, 70-78.
Baxendale, T. and Jones, O. (2000). Construction design and management safety
regulations in practice - progress on implementation. International Journal of
project management, 18, 33-40.
Behm, M., Veltri, A. and Kleinsorge, K.I. (2004). The cost of safety- cost analysis
model helps build business case for safety. Professional Safety, April, 22-29
Behm, M. (2005). Linking construction fatalities to the design for construction safety
concept. Safety Science, 43(8):589-611.
Behm, M. (2006). An analysis of construction accidents from a design perspective.
Silver Spring: CPWR
Bellamy, L.J. and Geyer, T.A.W. (2007). Development of a working model of how
human factors, safety management systems and wider organisational issues fit
together Norwich: HSE Books.
Bellamy, L.J., Geyer, T.A.W. and Wilkinson, J. (2008). Development of a functional
model which integrates human factors, safety management systems and wider
organisational issues. Safety Science, 46, 461-492.
Biggs, H.C., Dingsdag, D.P., Sheahan, V.L. and Stenson, N.J. (2005). The role of
collaboration in defining and maintaining a safety culture: Australian perspectives
in the construction sector. Conference proceedings of the 21st annual conference of
the association of researchers in construction management, held in London.
Bishop, D., Felstead, A., Fuller, A., Jewson, N., Unwin, L. and Kakavelakis, K. (2009).
Constructing learning: Adversarial and collaborative working in the British
construction industry. Journal of education and work, 22(4):243-260.
BLS (2010). National census of fatal occupational injuries in 2009 (preliminary
results). Washington, Bureau of labor statistics, U.S. Department of Labor.
BLS (2008). Workplace injuries and illnesses in 2007. Washington: Bureau of labor
statistics, U.S. Department of Labor.
Bollen, K.A. (2000). Modelling strategies: In search of the Holy Grail. Structural
equation modelling, 7(1):74-81.
Bomel (2004). Improving health and safety in construction phase 2 – depth and
breadth. Volume 5- falls from height. Understanding causes and risk control in the
construction industry. Norwich: HSE Books.
Bomel (2001). Improving health and safety in construction. Phase 1: Data collection,
review and structuring. Norwich: HSE Books.
Boomsma, A. (2000). Reporting analyses of covariance structures. Structural equation
modelling, 7(3):461-483.
263
Boote, D.N. and Beile, P. (2005). Scholars before researchers: On the centrality of the
dissertation literature review in research preparation. Educational researcher,
34(6):3-15.
Booth, R.T. & Lee, T.R. (1995). The role of human factors and safety culture in safety
management. Proceedings of the institution of mechanical engineers, part B:
Journal of engineering manufacture, 209(5):393-400.
Botswana Federation of Trade Unions (2007). Policy on health and occupational safe
environment in Botswana. Gaborone: BFTU.
Burfold, B., Hesketh, A., Wakeling, J., Bagnall, G., Colthart, I., Illing, J., Kergon, C.,
Morrow, G., Spencer, J. and Van Zwanenberg, T.T. (2009). Asking the right
questions and getting meaningful responses: 12 tips on developing and
administering a questionnaire survey for healthcare professionals. Medical teacher,
31, 207-211.
Byrne, B.M. (2006). Structural equation modelling with EQS- Basic concepts,
Applications and programming 2nd ed. Mahwah: Lawrence Erlbaum Associates,
Inc.
Cameron, I. and Duff, R. (2007). Use of performance measurement and goal setting to
improve construction managers’ focus on health and. Construction Management
and Economics, 25, 869–881.
Carder, B. and Ragan, P.W. (2003). A survey based system for safety measurement and
improvement. Journal of Safety Research, 34, 157-165.
Central Statistics Office (2009). June 2008 formal employment statistics. Gaborone:
CSO.
Chan, A.P.C. and Chan, A.P.L. (2004). Key performance indicators for measuring
construction success. Benchmarking: An International Journal, 11(2):203-221.
Chen, T.W., Lu, S.C. and Huang, Y.H. (2011). Investigating the safety cognition of
Taiwan’s construction personnel. Journal of Marine Science and Technology,
19(4):398-408.
Cheung, S.O., Cheung, K.K.W. and Suen, H.C.H. (2004). CSHM: Web-based safety
and health monitoring system for construction management. Journal of Safety
Research, 35, 159 – 170.
Chinda, T. and Mohamed, S. (2008). Structural equation model of construction safety
culture. Engineering, Construction and Architectural Management, 15(2):114-131;
114.
Chinda, T. and Mohamed, S. (2007). Causal relationships between enablers of
construction safety culture. Conference proceedings of the Fourth International
264
conference on construction in the 21st century (CITC-IV) conference held in
Goldcoast from July 11-17. CITC
Choudhry, R.M., Fang, D. and Mohamed, S. (2007). Developing a model of
construction safety culture. Journal of Management in Engineering, 23(4):207-212.
Choudhry, R.M., Fang. D. and Mohamed, S. (2009). Closure to “Developing a model of
construction safety culture” by Rafiq M. Choudhry, Dongping fang and Sherif
Mohamed. Journal of Management in Engineering, 25(1):45-47.
Choudhry, R.M., Fang, D. and Mohamed, S. (2007). The nature of safety culture: A
survey of the state-of-the-art. Safety Science, 45(10):993-1012.
Chua, D.K.H. and Goh, Y.M. (2004). Incident causation model for improving feedback
of safety knowledge. Journal of Construction Engineering and Management,
130(4):542-551.
CIB General Secretariat. (2010), Available from:
http://cibworld.xs4all.n1/pages/ftp/cmb_dir/com_list/w099/address.pdf. (Accessed
15 March 2010).
CIDB (2008). Construction health and safety in South Africa. Pretoria: Construction
Industry Development Board.
CIDB (2004). South Africa construction industry status report - 2004. Pretoria: CIDB.
CIOB (2009). Health and safety in the construction industry 2009. Berkshire: Chartered
Institute of Building.
Cipolla, D., Dingsdag, D.P., Sheahan, V. & Briggs, H., (2006). Using Safety Culture to
Overcome Market Force Influence on Construction Site Safety. In Clients Driving
Innovation, Moving Ideas into Practice. Conference proceedings of the second
International Conference of the CRC for Construction Innovation. Available from:
http://eprints.qut.edu.au/3801/.
Cooper, D. (2009). Behaviour safety interventions: A review of process design factors.
Professional Safety, February, 36 - 45.
Cooper, D. (2006). The impact of management’s commitment on employee behaviour.
In: American society of safety Engineers. Conference proceedings of the 7th
Professional development conference and exhibition held in Bahrain from March
18-22. ASSE.
Cooper, D. (2001). Treating safety as a value. Professional Safety, February, 17-21.
Cooper, M.D. (2000). Towards a model of safety culture. Safety Science, 36, 111-136.
265
Cosman, M. (2004). Roles, culture, outcomes. What does the UK experience mean? In
Designing for safety and health in construction. Edited by Hecker, S., Gambatese,
J. and Weinstein, M., September, 15-16, 59.
Cotton, A.P, Sohail, M. and Scott, R.E (2005). Towards improved labour standards for
construction of minor works in low-income countries. Engineering, Construction
and Architectural Management, 12(6), 617-632
Cox, S.J. and Cheyne, A.J.T. (2000). Assessing safety culture in offshore environments.
Safety Science 34 (1-3): 111-129.
Cox, S. and Flin, R. (1998). Safety culture: Philosopher’s stone or man of straw? Work
and Stress, 13(12): 189-201.
Creaser, W. (2008). Prevention through design (PtD) safe design from an Australian
perspective. Journal of Safety Research, 39(2):131.
Cullen, P.C. (2001). The Ladbroke grove rail inquiry - part 2 report. Norwich: HSC.
Cullen, W.D. (1990). The public inquiry into the piper alpha disaster. London:
Department of Energy, HMSO.
Dainty, A.R.J., Briscoe, G.H. and Millet, S.J. (2001). New perspectives on construction
supply chain integration. Supply chain management: An International Journal,
6(4):163-173.
Davies, H.T.O., Nutley, S.M. and Mannion, R. (2000). Organisational culture and
quality of health care. Quality in Health Care 9,111–119
De Leur, P. and Sayed, T. (2003). A framework to proactively consider road safety
within the road planning process. Canadian Journal of Civil Engineering, 30, 711719.
De Carlo, T.L. (1997). On the meaning and use of kurtosis. Psychological methods,
2(3):292-307.
Delbecq, A., Van de Ven, A. and Gustafson, D.H. (1975). Group techniques for
program planning: a guide to nominal group and Delphi processes. Glenview:
Scott, Foresman and company.
Dingsdag, D.P., Biggs, H.C., Sheahan, V.L. and Cipolla, C.J. (2006). A construction
safety competency framework: Improving OH&S performance by creating and
maintaining a safety culture. Brisbane: Cooperative Research Centre for
Construction Innovation.
Dion, P.A. (2008). Interpreting structural equation modelling results: A reply to Martin
and Cullen. Journal of Business Ethics, 83, 365-368.
266
DOETE (2006). Regulatory impact assessment safety, health and welfare at work
(construction) regulations 2006. Dublin: DOETE.
Dubois, A. and Gadde, L.E. (2001). The Construction Industry as a Loosely Coupled
System- Implications for productivity and innovativity. Conference proceedings of
the 17th IMP Conference, held in Oslo from 9-11th September. IMP
Duff, A.R., Robertson, I.T., Phillips, R.A. and Cooper, M.D. (1994). Improving safety
by the modification of behaviour. Construction Management and Economics, 12,
67-78.
Egan, J. (1998). Rethinking construction. London: Department of Trade and Industry.
Els, D.A. and Delarey, R.P. (2006). Developing a holistic wellness model. SA Journal
of Human Resource Management, 4(2):46-56.
Elsler, D. and Nikov, A. (2003) European approaches for economic incentives in
occupational safety and health. In Quality of work and products. Edited by Strasser,
H., Kluth, K., Rausch, H. and Bubb, H., Stuttgart: Ergonomia Verlag, 911-913.
Entec (1999). Development of a business excellence model of safety culture. London,
Entec.
Esposito, P.A. (2009). Safety audits comparing three types of assessments. Professional
Safety, 54(12):42-43.
European Agency for Safety and Health at Work (2010) Economic incentives to
improve occupational safety and health: A review from the European perspective.
Luxemburg, Publications Office of the European Union.
Eyiah, A.K. and Cook, P. (2003) Financing small and medium-scale contractors in
developing countries: A Ghana case study. Construction Management and
Economics, 21(4), 357-367.
Fairman, R. and Yapp, C. (2005). Enforced self-regulation, prescription and conceptions
of compliance within small businesses: The impact of enforcement. Law & Policy,
27(4):491-519.
Fellows, R and Liu, A. (2010). Culture as a component of complexity in construction.
Conference proceedings of the 2010 World Congress of CIB on Building a Better
world,
held
in
Salford
from
10-14
May.
Available
from:
http://hdl.handle.net/10722/136494
Fennell, D. (1988). Investigation into king's cross underground fire. London:
Department of Transport, HMSO.
267
Fernández-Muñiz, B., Montes-Peón, J.M. and Vázquez-Ordás, C.J. (2007). Safety
culture: Analysis of the causal relationships between its key dimensions. Journal of
Safety Research, 38, 627–641.
Findley, M., Smith, S., Kress, T., Petty, G. and Enoch, K. (2004). Safety programs
elements in construction- which ones best prevent injuries and control related
workers' compensation costs? Professional Safety, February, 14-21.
Fitzgerald, I. and Howarth, T. (2009). A study of migrant worker health and safety
issues in the UK construction industry. In Working together: Planning, designing
and building a healthy and safe construction industry. Conference Proceedings of
the CIB W099 Conference held in Melbourne from 21-23 October. CIB
Fitzgerald, M.K. (2005). Safety performance improvement through culture change.
Trans IChemE, part B, process safety environmental protection, 83 (B4):324-330.
Flin, R., Mearns, K., O'Connor, P. and Bryden, R. (2000). Measuring safety climate:
Identifying the common features. Safety Science, 34(1-3):177-192.
Foley, M, Silverstein, B., Polissar, N., Neradilek, B., (2009). Impact of Implementing
the Washington State Ergonomics Rule on Employer Reported Risk Factors and
Hazard Reduction Activity’, American Journal of Industrial Medicine, 52, 1-16
Fraser, L. (2007). Significant developments in occupational health and safety in
Australia’s construction industry. International Journal occupational
environmental health, 13, 12-20.
Frijters, A.C.P. and Swuste, P.H.J.J. (2008). Safety assessment in design and
preparation phase. Safety Science, 46, 272-281.
Gadd, S. and Collins, A.M. (2002). Safety culture: A review of the literature. Sheffield:
Health & Safety Laboratory.
Galvin, J.M. (2005). Occupational health and safety acts – performance and prosecution
in the Australian minerals industry. Mining technology: Transactions of the
Institute of mining & metallurgy, section A, 114, 251-256.
Gambatese, J.A., Hinze, J. and Haas, T. (1997). Tool to design for construction worker
safety. Journal of Architectural Engineering, 3(1):32-41.
Gambatese, J.A. and Hinze, J. (1999). Addressing construction worker safety in the
design phase. Automation in construction, 8(6):643-649.
Gambatese, J.A., Behm, M. and Hinze, J. W. (2005). Viability of designing for
construction worker safety. Journal of Construction Engineering and Management,
131(9):1029-1036.
268
Gambatese, J.A. (2009). MMT Online magazine, mmtmagazine. Available from:
http://www.mmtmagazine.org/page/index8f02.html?id. (Accessed 15 September
2011).
Gambatese, J.A. (2000). Owner involvement in construction site safety, In: Proceedings
of the sixth Construction Congress, 661–669.
Gholipour, Y. (2004). Injury minimisation in construction projects. Injury control and
safety promotion, 11(1):63-65.
Gibb, A. and Bust, P. (2006). Construction health and safety in developing countries.
Great Britain: European Construction Institute
Gibb A.G.F, Pavitt, T.C. and McKay, L.M., (2004) Designing or health and safety in
cladding installation – implications from pre-assembly. In Proceedings of
International Conference on Building Envelope Systems and Technologies
(ICBEST 2004), Sydney: Australia, 1-7.
Glendenning, P.M. (2001). Employee safety incentives: A best practices survey of
human resource practitioners. Professional Safety, 46(2):22.
Glendon, A.I. and Litherland, D.K. (2001). Safety climate factors, group differences and
safety behaviour in road construction. Safety Science (39): 157-188.
Glendon, A.I. and Stanton, N.A. (2000). Perspectives on safety culture. Safety Science
34, 193-214.
Glendon, A.I. and McKenna, E.F. (1995). Human Safety and Risk Management London:
Chapman & Hall.
Gohdes, S.L.W. and Crews, B.T. (2004). The Delphi technique: A research strategy for
career and technical education. Journal of career and technical education,
20(2):55-67.
Goodrum, P.M. and Gangwar, M. (2004). Safety incentives, a study of their
effectiveness in construction. Professional Safety, July, 24-34.
Gould, F.E. and Joyce, N.E. (2002). Construction management. Professional ed.
London: Prentice Hall.
Government of the Republic of South Africa (2003). The national occupational health
and safety policy.3rd Ed. Pretoria: Government of the Republic of South Africa.
Griffin, M.A. and Neal, A. (2000). Perceptions of safety at work: A framework for
linking safety climate to safety performance, knowledge and motivation. Journal of
occupational health psychology, 5(3):347-358.
269
Grote, G. (2008). Diagnosis of safety culture: A replication and extension towards
assessing "safe" organisational change processes. Safety Science, 46, 450-460.
Grote, G. (2007). Understanding and assessing safety culture through the lens of
organizational management of uncertainty. Safety Science, 45, 637-652.
Guldenmund, F.W. (2000). The nature of safety culture: A review of theory and
research. Safety Science, 34, 215-257.
Gyeke, S.A. and Salminen, S. (2009). Educational status and organizational safety
climate: Does educational attainment influence workers’ perceptions of workplace
safety? Safety Science, 47, 20-28.
Hair, J.F. Jr., Anderson, R.E., Tatham, R.L., & Black, W.C. (1998). Multivariate Data
Analysis, (5th Edition). Upper Saddle River, NJ: Prentice Hall
Hale, A., Kirwan, B. and Kjellen, U. (2007). Safe by design: Where are we now? Safety
Science, 45, 305-327.
Hale, A.R. (2000). Culture's confusions. Safety Science, 34, 1-14.
Hallowell, R.M. and Gambatese, J.A. (2010). Qualitative research: Application of the
Delphi method to CEM research. Journal of Construction Engineering and
Management, 136(1):99-107.
Hallowell, M. (2008). A formal model for construction safety and health risk
management. Oregon State University.
Hämäläinen, P., Saarela, K.L. and Takala, J. (2009). Global trend according to
estimated number of occupational accidents and fatal work-related diseases at
region and country level. Journal of Safety Research, 40, 125–139.
Hamid, A.R.A., Singh, B., Yusuf, W.Z.W. and Yang, A.K.T. (2004). Integration of
safety, health, environment and quality (SHEQ) management system in
construction: A review. Jurnal kejuruteraan awam, 16(1):24-37.
Hansen, L.L. (2000). The architecture of safety excellence. Professional Safety. High
Beam Research. Available from: http://www.highbeam.com. (Accessed 16
February 2012).
Harvey, J., Erdos, G., Bolam, H., Cox, M.A.A., Kennedy, J.N.P. and Gregory, D.T.
(2002). An analysis of safety culture attitudes in a highly regulated environment.
Work & Stress, 16(1):18-36.
Haslam, R.A., Hide, S.A., Gibb, A.G.F., Gyi, D.E., Pavitt, T., Atkinson, S. and Duff,
A.R. (2005). Contributing factors in construction accidents. Applied ergonomics,
36, 401-415.
270
Hasle, P. and Limborg, H.J. (2006). A review of the literature on preventive
occupational health and safety activities in small enterprises. Industrial Health, 44,
6-12.
Haupt, T.C. (2001). The performance approach to construction worker safety and
health. Unpublished thesis. Florida: University of Florida.
Havold, J.I. (2007). National cultures and safety orientation: A study of seafarers
working for Norwegian shipping companies. Work & Stress, 21(2):173-195.
Hayduck, L.A. and Glaser, D.N. (2000). Doing the four-step, right-2-3, wrong-2-3: A
brief reply to Mulaik and Milsap; Bollen; Bentler; and Herting and Costner.
Structural equation modelling, 7(1):111-123.
Hecker, S., Gambatese, J. and Weinstein, M. (2005). Designing for safety- moving the
construction safety process upstream. Professional Safety, September, 32-44.
Hecker, S., Gambatese, J. and Weinstein, M. (2004). Outcomes of a design-for-safety
process In Designing for safety and health in construction. Edited by. Hecker, S.
Gambatese, J. A. & Weinstein, M., Eugene: University of Oregon Press, 242-263.
Heppner, P.P. and Heppner, M.J. (2004). Writing and publishing your Thesis,
Dissertation & Research- A guide for students in the Helping Professions Belmont:
Brooks/Cole- Thomson Learning.
Hetherington, T. (1995). Why involve design professionals in construction safety?
Structural survey, 13(1):5-6.
Hermanus, M.A. (2007). Occupational health and safety in mining - Status, new
developments and concerns. The Journal of the Southern African Institute of
Mining and metallurgy, 107, 531-538
Hinze, J. and Gambatese, J.A. (1996). Addressing construction worker safety in project
design, Bureau of Engineering Research, University of Texas at Austin.
Hinze, J., Coble, R.J. & Elliot, B.R. (1999). Integrating construction worker protection
into project design. In Implementation of safety and health on construction sites.
Edited by Singh, A., Hinze, J. H. and Coble, R., Rotterdam: A.A. Balkema, 395402.
Hinze, J. (2005). A paradigm shift: leading to safety. In Rethinking and revitalising
construction safety, health, environment and quality. Edited by Haupt, T.C. and
Smallwood, J., Port Elizabeth: Construction Research Education and Training
Enterprises.
Hodgson, S. and Milford, R. (2005). Construction Safety, Health, Environment and
Quality: A Framework for Performance Improvement in South Africa. In
Rethinking and revitalising construction safety, health, environment and quality.
271
Edited by Haupt, T.C. and Smallwood, J., Port Elizabeth: Construction Research
Education and Training Enterprises.
Hofmann, D.A. and Morgeson, F.P. (1999). Safety-related behaviour as a social
exchange: The role of perceived organisational support and leader-member
exchange. Journal of applied psychology, 84(2):286-296.
Høivik, D., Moen, B.E., Mearns, K. and Haukelid, K. (2009). An explorative study of
health, safety and environment culture in a Norwegian petroleum company. Safety
Science, 47, 992-1001.
Høivik, D., Tharaldsen, J.E., Baste, V. and Moen, B.E. (2009). What is most important
for safety climate: The Company belonging or the local working environment? – A
study from the Norwegian offshore industry. Safety Science, 47:1324-1331.
Hola, B. (2007). General model of accident rate growth in the construction industry.
Journal of Civil Engineering and Management, 13(4):255-264.
Holey, E.A., Feeley, J.L., Dixon, J. and Whittaker V.J. (2007). An exploration of the
use of simple statistics to measure consensus and stability in Delphi studies. BMC
medical research methodology, 7(52):1-10.
Hoonakker, P., Loushine, T., Carayon, P., Kallman, J., Kapp, A. and Smith, M.J.
(2005). The effect of safety initiatives on safety performance: A longitudinal study.
Applied Ergonomics, 36, 461-469.
Hopkins, A. (2006). Studying organisational cultures and their effects on safety. Safety
Science, 44, 875-889.
House of Commons (2004). Health and safety executive: Improving health and safety in
the construction industry. London: The stationary office Limited.
HSE (2010a). Self-reported work-related illness and workplace injuries in 2008/09:
Results from the labour force survey. Caerphilly: HSE Books.
HSE (2010b). Work-related injuries and ill health in construction. Suffolk: HSE Books.
HSE (2001). A guide to measuring health and safety performance Suffolk: HSE Books.
LSC (2006). Standards for health and safety. Coventry, Learning Skills Council.
Available from: http://readingroom.lsc.gov.uk/lsc/2006/quality/goodpractice/natstandardsforhealthandsafety-re-feb2006.pdf. (Accessed 04 March 2011).
Hsu, C. and Sandford, B.A. (2007). The Delphi technique: Making sense of consensus.
Practical assessment, research and evaluation, 12(10):1-8.
272
Hu, L. and Bentler, P.M. (1999). Cut-off criteria for fit indexes in covariance structure
analysis: Conventional criteria versus new alternatives. Structural equation
modelling, 6(1):1-55.
Huang, X. and Hinze, J. (2006a). Owner's role in construction safety. Journal of
Construction Engineering and Management, 132(2):164-173.
Huang, X. and Hinze, J. (2006b). Owner's role in construction safety: Guidance model.
Journal of Construction Engineering and Management, 132(2):174-181.
Hudson, P. (2007). Implementing a safety culture in a multi-national. Safety Science 45,
697-722.
Hudson, P. (1999). Safety culture -Theory and practice. In: The Human Factor in
System Reliability – Is Human Performance Predictable? Proceedings of the RTO
HFM Workshop held on 1-2 December, Siena.
Human Engineering (2001). A review of safety culture and safety climate literature for
the development of the safety culture inspection toolkit. Bristol: HSE Books.
IET (2009). Safety culture. Stevenage: IET.
ILO (2005). Facts on safety at work. Geneva: ILO.
ILO (2003). Safety in numbers: Pointers for a global safety culture at work. Geneva:
ILO
ILO (1983). Encyclopaedia of occupational health and safety. Geneva: ILO.
INSAG (2002). Key practical issues in strengthening safety culture. Vienna: IAEA.
INSAG (1992). The Chernobyl accident: Updating of INSAG-1. Vienna: IAEA.
INSAG (1991). Safety culture. Vienna, IAEA.
IOSH (2010). Global best practices in contractor safety. IOSH/ASSE good practice
guidelines, 10.2. Leicestershire: IOSH.
IOSH (2004). Promoting a positive culture - a guide to health and safety culture. 4.2.
Leicestershire: IOSH.
Jackson, D.L., Gillaspy Jr., J.A. and Stephenson, R.P. (2009). Reporting practices in
confirmatory factor analysis: An overview and some recommendations.
Psychological Methods, 14(1):6-23.
Jackson, D.L. (2007). The effect of the number of observations per parameter in misspecified confirmatory factor analytic models. Structural equation modelling,
14(1):48-76.
273
Jafri, H., Ali, M.W., Ahmad, A. and Kamsah, M.Z. (2005). Effective occupational
health and safety performance measurements. ICCBPE / SOMChE, 702-708.
Jaselskis, J.E., Anderson, D.S. and Russell, S.J. (1996). Strategies for achieving
excellence in construction safety performance. Journal of Construction
Engineering and Management, 122(1):61-70.
Johnson, G. and Scholes, K. (2002). Exploring corporate strategy, text and cases. 6th
ed. Harlow: Prentice Hall.
Kamuzora, P. (2006). Non-decision making in occupational health policies in
developing countries. International Journal of occupational and environmental
health, 12, 65-71.
Karjalainen, A. (2004). A statistical portrait of health and safety at work in the
construction industry: Actions to improve safety and health in construction.
Luxembourg: European Agency for safety and health at work.
Karlsen, J.E. and Valen, H. (2011). The social construction of HSE culture. Safety
Science monitor, 15(1):1-9.
Kartam, N.A., Flood, I. and Koushki, P. (2000). Construction safety in Kuwait: Issues,
procedures, problems and recommendations. Safety Science 36, 163-184.
Kennedy, R. and Kirwan, B. (1998). Development of a hazard and operability-based
method for identifying safety management vulnerabilities in high risk systems.
Safety Science 30, 249-274.
Kheni, N.A, Dainty, A.R.J. and Gibb, A.G.F. (2007) Influence of political and sociocultural environments on health and safety management within SMEs: a Ghana
case study. In: Boyd, D (Ed) Proceedings of the 23rd Annual ARCOM Conference,
3-5 September 2007, Belfast, UK, Association of Researchers in Construction
Management, 159-168
Kheni, N.A., Gibb, A.G.F. and Dainty, A.D.F. (2006) The management of construction
site health and safety by small and medium-sized construction businesses in
developing countries: a Ghana case study. In: Boyd, D (Ed) Proceedings of the
22nd Annual ARCOM Conference, 4-6 September 2006, Birmingham, UK,
Association of Researchers in Construction Management, 273-282
Kibert, C.J. and Coble, R.J. (1995). Integrating safety and environmental regulation of
construction industry. Journal of Construction Engineering & Management,
121(1):95-99.
Kikwasi, G.J. (2008). Client involvement in construction safety and health. In Evolution
of and directions in construction safety and health. Edited by Hinze, J., Boener, S.
and Lew, J., Rotterdam: In-House publishing, 55-69.
274
Kines, P., Spangenberg, S. and Dyreborg, J. (2007). Prioritizing occupational injury
prevention in the construction industry: Injury severity or absence? Journal of
Safety Research 38, 53-58.
Kinnersley, S. and Roelen, A. (2007). The contribution of design to accidents. Safety
Science, 45, 31-60.
Kirwan, B. (2007). Safety informing design. Safety Science, 45, 155-197.
Kline, B.R. (2005). Principles and practice of structural equation modelling. 2nd Ed.
New York: The Guilford Press.
Koehn, E., Kothari, R.K. and Pan, C. (1995). Safety in developing countries:
Professional and bureaucratic problems. Journal of Construction Engineering &
Management, 121(3):261-265.
Krause, T.R. and Weekley, T. (2005). Safety leadership- A four factor model for
establishing a high functioning Organisation. Professional Safety, November, 3440.
Kulchartchai, O. and Hadikusumo, B.H.W. (2010). Exploratory study of obstacles in
safety culture development in the construction industry: A grounded theory
approach. Journal of construction in developing countries, 15(1):45-66.
Langford, D., Rowlinson, S. and Sawacha, E. (2000). Safety behaviour and safety
management: Its influence on the attitudes of workers in the UK construction
industry. Engineering Construction and Architectural Management, 7(2):133-140.
Laurence, D. (2005). Safety rules and regulations on mine sites-the problem and a
solution. Journal of Safety Research, 36(1):39-50.
Lazarevic, S.P. and Perry, M. (2004). Occupational health and safety shortcomings in
the Australian construction industry. Monash University, Business and Economics.
Lee, S., Halpin, D.W. and Chang, H. (2006). Quantifying effects of accidents by fuzzylogic and simulation- based analysis. Canadian Journal of Civil Engineering 33,
219-226.
Lee, T. and Harrison, K. (2000). Assessing safety culture in nuclear power stations.
Safety Science 34, 61-97.
Lenox, M.J. and Nash, J. (2003). Industry self-regulation and adverse selection: A
comparison across four trade association programs. Business strategy and the
environment, 12(6):343-356.
Lethbridge, J. (2008). Occupational health regulations and health workers: Protection
or vulnerability? London: Public Services International.
275
Leveson, N. (2004). A new accident model for engineering safer systems. Safety
Science, 42(4):237-270.
Levitt, R.E. and Samuelson, N.M. (1993). Construction safety management, 2nd Ed.
New York: John Wiley & Sons.
Lin, J. and Mills, A. (2001). Measuring the occupational health and safety performance.
Facilities, 19(3/4):131-138.
Lingard, H. and Rowlinson, S. (1997). Behaviour-based safety management in Hong
Kong’s construction industry. Journal of Safety Research, 28, 243-256.
Lingard, H., Blismas, N., Cooke, T. and Cooper, H. (2009). The model client
framework - resources to help Australian government agencies to promote safe
construction. International Journal of managing projects in business, 2(1):131-140.
Lingard, H., Stranieri, A. and Blismas, N. (2006). Supporting the design OHS process:
A knowledge-based system for risk management. Cooperative Research Centre.
Loosemore, M. and Andonakis, N. (2007). Barriers to implementing OHS reforms - the
experiences of small subcontractors in the Australian construction industry.
International Journal of project management, 25, 579-588.
Loosemore, M., Lingard, H., Walker, D.H.T. and Mackenzie, J. (1999). Benchmarking
safety management systems in contracting organisations against best practice in
other Industries. In Implementation of Safety and Health on construction sites.
Edited by Singh, A., Hinze, J. H. and Coble, R.J., Rotterdam: A.A. Balkema.
Loughborough University and UMIST (2003). Causal factors in construction accidents.
Norwich: HSE Books.
Lund, J. and Aaro, L.E. (2004). Accident prevention. Presentation of a model placing
emphasis on human, structural and cultural factors. Safety Science, 42, 271-324.
Mackenzie, J., Gibb, A.G.F. and Bouchlaghem, N.M. (1999). Communication of health
and safety in the design phase. In Implementation of Safety and Health on
construction sites. Edited by Singh, A., Hinze, J. H and Coble, R. J., Rotterdam:
A.A. Balkema, 419-426.
Makino, M. (2006). Monitoring toward cultivation of safety culture. International
symposium on safety culture. Tokyo: Nuclear safety Commission.
Mansingh, K.S and Haupt, T.C. (2008). Construction Accident Causation: An
Exploratory Analysis. In Evolution of and Directions in Construction Safety and
Health. Edited by Hinze, J., Boehner, S and Lew, J, CIB W099, Rotterdam, 465482.
276
Manu, P., Ankrah, N., Proverbs, D. and Suresh, S. (2010). Exploring the influence of
construction project features in Accident causation. In Proceedings of CIB 2010
World Congress. Edited by Barrett, P., Amaratunga, D., Haigh, R., Keraminiyage,
K., Pathirage, C., Salford: CIB.
Manzella, J.C. (1999). Measuring safety performance to achieve long term
improvement. Professional Safety, September, 33-36
Marsh, H.W., Hau, K.T. and Wen, Z. (2004). In search of golden rules: Comment on
hypothesis- testing approaches to setting cut-off values for fit indexes and dangers
in over generalising Hu and Bentler's (1999) findings. Structural equation
modelling, 11(3):320-341; 320.
Martins, E.C. and Terblanche, F. (2003). Building organisational culture that stimulates
creativity and innovation. European Journal of Innovation Management, 6 (1): 6474.
McDonald, M.A., Lipscomb, J.H., Bondy, J. and Glazner, J. (2009). Safety is everyone's
job: The key to safety on a large university construction site. Journal of Safety
Research, 40, 53-61.
McDonald, R.P. and Ho, R.M.R. (2002). Principles and practice in reporting structural
equation analyses. Psychological methods, 7(1):64-82.
Mearns, K., Whitaker, S.M. and Flin, R. (2003). Safety climate, safety management
practice and safety performance in offshore environments. Safety Science 41, 641680.
Mengolini, A. and Debarberis, L. (2007). Safety culture enhancement through the
implementation of IAEA guidelines. Reliability engineering and system safety, 92,
520-529.
Misnan, M.S. and Mohammed, A.H. (2007). Development of safety culture in the
construction industry. In proceedings of the conference on sustainable building
South East Asia. Kuala Lumpur: CIB.
Misnan, M.S., Mohammed, A.H.B., Mahmood, W.Y.W., Mahmud, H.S. and Abdullah,
N. M. (2008). Development of safety culture in the construction industry: the
leadership and training roles. Conference proceedings of the 2nd International
Conference on Built Environment in Developing Countries 2008 (ICBEDC) held in
Penang. USM.
Mitropoulos, P., Abdelhamid, S.T and Howell, G.A. (2005). Systems model of
construction accident causation. Journal of Construction Engineering and
Management, 131(7):816-825.
Mohamed, S. (1999). Empirical investigation of construction safety management
activities and performance in Australia. Safety Science, 33(3):129-142.
277
Mohamed, S. (2003). Scorecard approach to benchmarking organizational safety culture
in construction. Journal of Construction Engineering and Management, 129(1):8088.
Mohamed, S. (2002). Safety climate in construction site environments. Journal of
Construction Engineering and Management, 128(5):375-384.
Molenaar, K.R., Park, J and Washington, S. (2009). Framework for measuring corporate
safety culture and its impact on construction safety performance. Journal of
Construction Engineering and Management, 135(6):488-496.
Molenaar, K., Brown, H., Caile, S and Smith, R. (2002). Corporate culture: A study of
firms with an outstanding construction safety. Professional Safety, July, 18-27.
Mullen, J. (2004). Investigating factors that influence individual safety behaviour at
work. Journal of Safety Research, 35, 275-285.
Murie, F. (2007). Building Safety-An international perspective. International Journal
occupational environmental health, 13, 5-11.
Musonda, I and Smallwood, J.J. (2008). Health and safety awareness and
implementation in Botswana’s construction industry. Journal of Engineering,
Design and Technology, 6(1):81-90.
Musonda, I., Haupt, T.C. and Smallwood, J.J. (2009). Client attitude to health and
safety - A report on contractors' perceptions. Acta Structilia, 16(2):69-85.
Nanfosso, T.A.R. and Mbassi, A.N.S. (2011). Organisational changes and productivity
of manufacturing enterprises in Cameroon. Business and Management Review,
1(4):49-58.
National Statistics UK (2010). The health and safety executive statistics 2009/10.
Caerphilly: HSE Books.
Ng, S.T., Cheng, K.P. and Skitmore, M. (2005). A framework for evaluating the safety
performance of construction contractors. Building and environment, 40(10):13471355.
Ng, S.T., Palaneeswaran, E. and Kumaraswamy, M.M. (2002). A dynamic e-reporting
system for contractor's performance appraisal. Advances in Engineering Software,
33, 339-349.
Nuwayhid, I.A., (2004). Occupational health research in developing countries: A
partner for social justice. American Journal of public health, 94(11):1916-1921.
Okoli, C. and Pawlowski, S.D. (2004). The Delphi method as a research tool: An
example, design considerations and applications. Information and management, 42,
15-29.
278
Oil Spill Commission (2011). Deep water the gulf oil disaster and the future of offshore
drilling. Washington: OSC
Oxford (2001). Compact Oxford dictionary thesaurus and word power guide, Oxford:
Oxford University Press.
Pan, W., Soetanto, R. and Sidwell, R. (2010). How Environments Shape Innovation:
The Case of Precast Concrete Cross-wall for Multi-Storey Residential Building
Construction. In Proceedings of CIB 2010 World Congress. Edited by Barrett, P.,
Amaratunga, D., Haigh, R., Keraminiyage, K., Pathirage, C., Salford: CIB, 14-25.
Parker, D., Lawrie, M. and Hudson, P. (2006). A framework for understanding the
development of organisational safety culture. Safety Science, 44, 551–562.
Paul, J. (2009). London 2012: A new approach to CDM coordination. Civil
Engineering, 162(2):76-86.
Peaucelle, J.L. (2000). From taylorism to post-taylorism: Simultaneously pursuing
several management objectives. Journal of organizational change management,
13(5):452-467.
Pellicer, E. and Molenaar, K.R. (2009). Discussion of “Developing a model of
construction safety culture” by Rafiq M. Choudhry, Dongping, Fang and Sherif
Mohamed. Journal of Management in Engineering, 25(1):44-47.
Petersen, D. (2005). Safety improvement - perception surveys can reveal strengths and
weaknesses. Professional Safety, January, 45-48.
Petersen, D. (2000). The behavioural approach to safety management. Professional
Safety, (March):37-39.
Pidgeon, N. and O'Leary, M. (2000). Man-made disasters: Why technology and
organizations (sometimes) fail. Safety Science, 34, 15-30.
Pivo, G. (2008). Responsible property investment criteria developed using the Delphi
method. Building Research & Information, 36(1):20-36; 20.
Podgórski, D. (2006). Factors influencing implementation of occupational safety and
health management systems by enterprises in Poland. Human factors &
Ergonomics in manufacturing, 16(3):255-267.
Pungvongsanuraks, P. and Chinda, T. (2010). Investigation of safety perceptions of
management and workers in Thai construction industry. Suranaree Journal of
science and technology, 17(2):177-191.
Rahman, M.M., Kumaraswamy, M.M. and Yng Ling, F.Y. (2007). Building a relational
contracting culture and integrated teams. Canadian Journal of Civil Engineering,
34(1):75-88.
279
Rajendran, S. and Gambatese, J.A. (2009). Development and initial validation of
sustainable construction safety and health rating system. Journal of Construction
Engineering & Management, 135(10):1067-1075.
Rankin, J., Fayek, A.R., Meade, G., Haas, C. and Manseau, A. (2008). Initial metrics
and pilot program results for measuring the performance of the Canadian
construction industry. Canadian Journal of Civil Engineering, 35, 894-907.
Rayens, M.K. and Hahn, E.J. (2000). Building consensus using policy Delphi method.
Policy, politics and nursing practice, 1(4): 308-315.
Raykov, T., Tomer, A. and Nesselroade, J.R. (1991). Reporting structural equation
modelling results in psychology and aging: Some proposed guidelines. Psychology
and aging, 6(4):499-503.
Reason, J. (2000a). Human error: Models and management. BMJ, 320, 768-770.
Reason, J. (2000b). Safety paradoxes and safety culture. Injury control & safety
promotion, 7(1):3-14.
Resnick, M. (2009). Safety incentive programs. Avoiding the pitfalls. Professional
Safety, July, 46-48.
Riley, M.J. and Clare-Brown, D. (2001). Comparison of cultures in construction and
manufacturing industries. Journal of Management in Engineering, 17(3):149-158.
Ringen, K. and Englund, A. (2006). The construction industry. Annals of the New York
academy of sciences, 1076(1):388-393.
Rowe, G. and Wright, G. (1999). The Delphi technique as a forecasting tool: Issues and
analysis. International Journal of forecasting, 15(4):353-375.
Rundmo, T. and Hale, A.R. (2003). Managers’ attitudes towards safety and accident
prevention. Safety Science, 41(7):557.
Saji, G. (2003). Safety goals in ‘risk-informed, performance-based’ regulation.
Reliability Engineering & System Safety, 80(2):163.
Said, I., Shafiel, M.W.M. and Omran, A. (2009). The roles of clients in enhancing
construction safety. Journal of engineering, VII (2), 127-134.
Salem, O., Lothlikar, H., Genaidy, A. and Abdelhamid, T. (2007). A behaviour-based
safety approach for construction projects. In, Proceedings of the conference of the
International Group for Lean Construction, Michigan: IGLC.
Saurin, A.T., Formoso, T.C. and Guimaraes, L.B.M. (2004). Safety and production: An
integrated planning and control model. Construction Management and Economics,
22, 159-169.
280
Saurin, A.T., Formoso, T.C. and Cambraia, F.B. (2005). Analysis of a safety planning
and control model from the human error perspective. Engineering, construction
and architectural management, 12(3):283-298.
Sawacha, E., Naoum, S. and Fong, D. (1999). Factors affecting safety performance on
construction sites. International Journal of project management, 17(5):309-315;
309.
Schein, E.H. (1999). The corporate culture survival guide. San Francisco: Jossey-Bass
Inc.
Schneider, J. (1997). Organisational behaviour modification and occupational health
and safety performance improvement. Conference proceedings of the Queensland
Mining Industry Health and Safety Conference.
Schreiber, J.B., Stage, K.F. and King, J. (2006). Reporting structural equation modeling
and confirmatory factor analysis results: A review. The Journal of educational
research, 99(6):323-337.
Seo, D.C., Torabi, M.R., Blair, E.H. and Ellis, N.T. (2004). A cross validation of safety
climate scale using confirmatory factor analytic approach. Journal of Safety
Research, 35, 427-445.
Shash, A.A. and Ahcom, J. (2006). Organisational aspects of safety management
subsystem. Journal of construction research, 7(1-2):65-80.
Sheehy, E.J. (2004). Effective safety incentives. A pilot program based on risk
reduction Professional Safety, February, 40-48.
Simon, I.S. and Frazee, R.P. (2005). Building a better safety vehicle. Professional
Safety, January, 36-44.
Skulmoski, G., Hartman, F.T. and Krahn, J. (2007). The Delphi method for graduate
research. Journal of Information Technology Education, 6, 1-21.
Smallman, C. and John, G. (2001). British directors’ perspectives on the impact of
health and safety on corporate performance. Safety Science, 38, 227-239.
Smallman, C. (2001). The reality of revitalizing health and safety. Journal of Safety
Research, 32, 391-439.
Smallwood, J. (1998). Client influence on contractor health and safety in South Africa.
Building research & information, 26(3):181-189.
Smallwood, J.J. and Haupt, T.C. (2007). Impact of the South African construction
regulations on construction health and safety: Architects' perceptions. Journal of
engineering, design and technology, 5(1):23-34.
281
Smallwood, J.J. (2004). The influence of engineering designers on health and safety
during construction. Journal of the South African Institution of Civil Engineering,
46(1):2-8.
Spangenberg, S., Baarts, C., Dyreborg, J., Jensen, L., Kines, P. and Mikkelsen, K.L.
(2003). Factors contributing to the differences in work related injury rates between
Danish and Swedish construction workers. Safety Science, 41, 517-530.
Statistics South Africa (2011). Quarterly labour force survey. Pretoria: Statistics South
Africa.
Suárez-Barraza, M.F., Ramis-Pujol, J. and Heras, M. A. (2010). Reflecting upon
management systems: Content analysis and synthesis. International Journal of
business research and management, 1(2):64-86.
Suraji, A., Duff, A.R. and Peckitt, S.J. (2001). Development of causal model of
construction accident causation. Journal of Construction Engineering and
Management, 127(4):337-344.
Suraji, A., Sulaiman, K., Mahyuddin, N. and Mohamed, O. (2006). Rethinking
construction safety: An introduction to total safety management. Journal of
construction research, 1(1&2):49-63.
Svedung, I. and Rasmussen, J. (2002). Graphic representation of accident scenarios:
Mapping system structure and the causation of accidents. Safety Science, 40, 397417.
Tabachnick, B.G. and Fidell, L.S. (2007). Using multivariate statistics 5th ed. Boston:
Allyn and Bacon.
Tam, C.M., Zeng, S.X. and Deng, Z.M. (2004). Identifying elements of poor
construction safety management in china. Safety Science, 42, 569-586.
Tang, W., Qiang, M., Duffield, C. F., Young, D.M. and Lu, Y. (2008). Incentives in the
Chinese construction industry. Journal of Construction Engineering and
Management, 134(7):457-467.
Teo, A.L.E. and Ling, Y.Y.F. (2006). Developing a model to measure the effectiveness
of safety management systems of construction sites. Building and environment, 41,
1584-1592.
Teo, E.A.L., Ling, F.Y.Y. and Chong, A.F.W. (2005). Framework for project managers
to manage construction safety. International Journal of project management,
23(4):329-341.
Thangaratinam, S. and Redman, W.E.C. (2005). The Delphi technique. The obstetrician
& gynaecologist, 7, 120-125.
282
The Keil Centre. (2003). Behaviour modification to improve safety: Literature review.
Edinburg: HSE Books.
Toellner, J. (2001). Improving safety and health performance: Identifying and
measuring leading indicators. Professional Safety, September, 42-47.
Tomaskovic-Devey, D., Leiter, J. and Thompson, S. (1994). Organisational survey
nonresponse. Administrative science quarterly, 39(4):439-457.
Toole, T.M. and Gambatese, J. (2008). Trajectories of prevention through design in
construction. Journal of Safety Research, 39, 225-230.
Toole, T.M. (2005). Increasing engineers’ role in construction safety: Opportunities and
barriers. Journal of professional issues in engineering education & practice,
131(3):199-207.
Toole, T.M. (2002). Construction site safety roles. Journal of Construction Engineering
& Management, 128(3):203.
Törner, M. and Pousette, A. (2009). Safety in construction – a comprehensive
description of the characteristics of high safety standards in construction work,
from the combined perspective of supervisors and experienced workers. Journal of
Safety Research, 40, 399-409.
Tuchman, J.L. (2003). Owners play a major role in site safety. ENR: Engineering newsrecord, 251(6):14.
Tzortzopoulos, P., Sexton, M. and Cooper, R. (2005). Process models implementation
in the construction industry: A literature synthesis. Engineering, construction and
architectural management, 12(5):470-486.
Ulang, N.M., Gibb, A. and Anumba, C.J. (2009). The Communication of Health and
Safety Information in Construction. In Proceedings of the W099 - Special Track
18th CIB World Building Congress. Edited by Barrett P., Amaratunga D., R. Haigh
& Keraminiyage K. and Pathirage C., Salford: CIB, 264-277.
Van Ooteghem, P. (2006). Work-related injuries and illnesses in Botswana.
International Journal occupational environmental health, 12, 42-51.
Viscusi, W.K. (1989). Safety through markets. Society, 27(1):9-10.
Waara, F. and Bröchner, J. (2006). Price and non-price criteria for contractor selection.
Journal of Construction Engineering & Management, 132(8):797-804.
Walls, J., Pidgeon, N., Weyman, A. and Horlick-Jones, T. (2004). Critical trust:
Understanding lay perceptions of health and safety risk regulation. Health, risk &
society, 6(2):133-150.
283
Walters, D. (2009). The role of worker representation and consultation in managing
health and safety in the construction industry. Cardiff: ILO.
Wamuziri, S. (2006). Safety culture in the construction industry. Municipal Engineer,
159(ME3):167-174.
Watanabe, T. and Hanayasu, S. (1999). Philosophy of construction safety management
in Japan. In Implementation of Safety and Health on construction sites. Edited by
Singh, A., Hinze, J. H. and Coble, R. J., Rotterdam: A.A. Balkema, 55-62.
Weinstein, M., Gambatese, J. and Hecker, S. (2005). Can design improve construction
safety? Assessing the impact of a collaborative safety-in-design process. Journal of
Construction Engineering and Management, 131(10):1125–1134.
Wiegmann, D.A., Zhang, H., Thaden, T.V., Sharma, G. and Mitchell, A. (2002). Safety
culture: A review. Savoy: Federal Aviation Administration Atlantic City
International Airport.
Wilkinson, D. and Birmingham, P. (2003). Using Research Instruments: A Guide for
Researchers London: Taylor & Francis.
Williams, A., Dobson, P. and Walters (1993). Changing culture: New organisational
approaches. 2nd ed. London: Institute of personal management.
Wilson, E.S. (2008). Toward a model of the impact organisation, human and technology
factors have on the effectiveness of safety management systems. Journal of
achievements in materials and manufacturing engineering, 31(2):827.
Winch, G.M. (2000). Institutional reform in British construction: Partnering and private
finance. Building Research & Information, 28(1):141-155.
World Bank, (2008). World development Indicators, Botswana - Country Brief
http://web.worldbank.org/WEBSITE/EXTERNAL/COUNTRIES/AFRICAEXT/B
OTS (Accessed 04 March 2009).
Wright, F. (2007). Health and Safety Commission – A case for reform? Law, science
and policy, 3, 157-175.
Yang , C.C., Wang, Y.S., Chang, S.T., Guo, S.E. and Huang, M.F. (2009). A study on
the leadership behaviour, safety culture and safety performance of the healthcare
industry. World academy of science, engineering and technology, 53, 1148-1155.
Yean, Y.L., Rahman, M.M. and Tiong, L.N. (2006). Incorporating contractual
incentives to facilitate relational contracting. Journal of professional issues in
engineering education & practice, 132(1):57-66.
Yule, S. and Mearns, K. (2004). The client has most influence on the contractor's HS&E
culture. In: proceedings of the SPE international conference on Health, Safety and
284
Environment in Oil and Gas exploration and production. Society of Petroleum
Engineers.
Yule, S., Flin, R. and Murdy, A. (2007). The role of management and safety climate in
preventing risk-taking at work. International Journal of risk assessment and
management, 7(2):137-151.
Zhou, Q., Fang, D. and Mohamed, S. (2011). Safety climate improvement: Case study
in a Chinese construction company. Journal of Construction Engineering and
Management, 137(1):86-95.
Zhou, Q., Fang, D. and Wang, X. (2008). A method to identify strategies for the
improvement of human safety behaviour by considering safety climate and
personal experience. Safety Science, 46, 1406-1419.
Zohar, D. and Luria, G. (2003). The use of supervisory practices as leverage to improve
safety behaviour. Journal of Safety Research, 34, 567-577.
285
APPENDIX A
INVITATION LETTERTO PARTICIPATE IN A DELPHI STUDY
17 October 2009
Dear Sir,
Innocent Musonda is registered for a PhD in the Faculty of Built Environment at the
University of Johannesburg under the supervision of Prof Theo C Haupt ( Director:
Building Construction Science (BCS), College of Architecture, Art and Design,
Mississippi State University, and Visiting Professor, University of Johannesburg, South
Africa, Department of Construction Management ).
The area of his research is on developing a client centred model on health and safety
Performance improvement. He will be using a Delphi approach and needs to compile a
panel of experts in the field to participate in this process. It would be appreciated if you
would consent to participating in the study in this capacity.
Kind regards
Theo C. Haupt, Ph.D., M.Phil., Pr. CM.
Director: Building Construction Science (BCS), College of Architecture, Art and
Design, Mississippi State University
Research Associate: Southern African Built Environment Research Centre (SABERC),
Cape Peninsula University of Technology, South Africa ,Visiting Professor: University
of Johannesburg, South Africa, Department of Construction Management ,President:
Association of Schools of Construction of Southern Africa (ASOCSA) ,Joint
International coordinator: CIB TG59 - People in Construction , Editor-in-chief: Journal
of Engineering, Design and Construction (JEDT) , Co-editor: Journal of Construction
(JOC)
899 College view, 128 Giles Hall, P.O. Box AQ, Mississippi State, MS 39762
Voice: 662.325.2547 (office) , Fax: 622.325.8872, Mobile: 662.312.5328
Email: thaupt@caad.msstate.edu
APPENDIX B
REQUEST FOR EXPERTS’ CURRICULLUM VITAE
13 November 2009
Dear Sir/Madam
I would like to thank you for accepting the invitation to participate as an expert in our
project to develop a client-centred model to improve health and safety performance in
the construction industry.
The process of collecting input from the expert panel will probably involve no less than
three rounds. The first round will be at the end of January 2010, the second round will
be held at the end of February 2010 and the third round is anticipated to be held at the
end of March 2010.
To start with, I would like to request for your curriculum vitae for our records and to
confirm your area of expertise. We would appreciate your response by the end of
November 2009.
Kind regards
I. Musonda PhD Candidate
University of Johannesburg
Faculty of Engineering and the Built Environment
Tel: +2711 559 6655
Mobile: +27730769652
Fax: +2711 559 6630
Email: innocentmusonda@gmail.com
APPENDIX C
DELPHI INSTRUCTIONS FOR ROUND 1 AND QUESTIONNAIRE
DELPHI SURVEY – ROUND 1
Thank you once again for serving on the Delphi panel for this research. Your participation
is greatly appreciated.
This first Round survey is intended to be completed in approximately 25-30 minutes.
Subsequent surveys will require significantly less time to complete. When you have
finished answering all of the questions, please email your response, in Word format, to
innocentmusonda@gmail.com.
You will be given the opportunity to change your response later on after all Delphi
participants have completed the first Round survey and results have been analysed.
Results will be in simple statistics e.g. median response, average and range.
INSTRUCTIONS
Please answer all of the following questions to the best of your ability. Please indicate
your response by placing an ‘X’ in the appropriate boxes. The survey requests that you
rate the likelihood of the element happening due to external influence or pressure and the
impact level of various Health and safety (H&S) elements on stakeholder culture and on
H&S outcome. The probability scale is presented below and only a number should be
used for a probability range. If say you consider the probability range to be between 41 &
50% of the element happening then you should mark ‘X’ under the box ‘5’. If the impact
is considered to be medium, then the ‘X’ should be marked under the ‘5’ or ‘6’ box. (See
below).
Please use your experience and judgement to rate what you believe the average negative
impact of the various elements on stakeholders would be if the described elements were
not implemented or absent.
PROBABILITY SCALE (likelihood in percentage)
0112131415110%
20%
30%
40%
50%
60%
1
2
3
4
5
6
X
IMPACT SCALE
No impact
Low impact
1
2
3
4
Medium
impact
5
6
X
288
6170%
7
7180%
8
High impact
7
8
8190%
9
91100%
10
Very high
impact
9
10
1.1
Client element
Client financing H&S
Appointment of H&S agent
H&S staffing
Choice of procurement method
Client involvement in design
H&S inspections & audits
H&S policy, procedures & goals
H&S leadership
Probability = of Client Implementing listed elements if
pressured by political environment. (1=low probability,
10=high probability)
1
2
3
4
5
6
7
8
9
10
X
X
X
X
X
X
X
X
1.2 (Social pressure = Refers to requirements for community approval of projects on
satisfying H&S standards)
Probability = of Client Implementing listed elements if
Client element
pressured by social environment. (1=low probability, 10=high
probability)
1
2
3
4
5
6
7
8
9
10
Client financing H&S
X
Appointment of H&S agent
X
H&S staffing
X
Choice of procurement method
X
Client involvement in design
X
H&S inspections & audits
X
H&S policy, procedures & goals
X
H&S leadership
X
1.3
Client element
Client providing finance for H&S
Appointment of H&S agent
H&S staffing
Choice of procurement method
Client involvement in design
H&S inspections & audits
H&S policy, procedures & goals
H&S leadership
Probability = of Client Implementing listed elements if there
is an improvement in materials & methods. (1=low
probability, 10=high probability)
1
2
3
4
5
6
7
8
9
10
X
X
X
X
X
X
X
X
289
1.4 (Economic pressure = Refers to economic incentive for implementing H&S and
tangible loss for not implementing H&S)
Probability = of Client Implementing listed elements if
Client element
pressured by economics. (1=low probability, 10=high
probability)
Client providing finance for H&S
Appointment of H&S agent
H&S staffing
Choice of procurement method
Client involvement in design
H&S inspections & audits
H&S policy, procedures & goals
H&S leadership
1.5
Client element
Client providing finance for H&S
Appointment of H&S agent
H&S staffing
Choice of procurement method
Client involvement in design
H&S inspections & audits
H&S policy, procedures & goals
H&S leadership
1.6
Client element
Client providing finance for H&S
Appointment of H&S agent
H&S staffing
Choice of procurement method
Client involvement in design
H&S inspections & audits
H&S policy, procedures & goals
H&S leadership
1
2
3
4
5
6
7
8
X
9
X
X
X
10
X
X
X
X
Probability = of Client Implementing listed elements if
pressured by regulations & legislative framework. (1=low
probability, 10=high probability)
1
2
3
4
5
6
7
8
X
X
X
X
X
X
X
X
9
10
Probability = of Client Implementing listed elements if
pressured by professional bodies. (1=low probability, 10=high
probability)
1
2
3
4
5
6
7
8
9
10
X
X
X
X
X
X
X
X
290
1.7
Client element
Client providing finance for H&S
Appointment of H&S agent
H&S staffing
Choice of procurement method
Client involvement in design
H&S inspections & audits
H&S policy, procedures & goals
H&S leadership
1.8
Client element
Client providing finance for H&S
Appointment of H&S agent
H&S staffing
Choice of procurement method
Client involvement in design
H&S inspections & audits
H&S policy, procedures & goals
H&S leadership
1.9
Environment element
Political pressure
Social pressure
Economic pressure
Improvement on materials &
methods
Contractor commitment &
involvement
Regulations & legislative frame
work
Professional bodies influence
Impact = on contractors’ H&S culture if listed client element
is lacking. (1=no impact, 10=very high negative impact)
1
2
3
4
5
6
7
8
9
X
X
X
X
10
X
X
X
X
Impact = on designers’ H&S culture if listed client element is
lacking. (1=no impact, 10=very high negative impact)
1
2
3
4
5
6
7
8
X
X
X
9
X
10
X
X
X
X
Impact = on clients’ H&S culture if listed element is lacking.
(1=no impact, 10=very high negative impact)
1
2
3
4
5
6
X
X
7
8
X
9
10
X
X
X
X
291
2.1
Project phase
Project initiation
Project design
Procurement
Construction
Project commissioning
Project closeout
Project operations & maintenance
Probability = of H&S consideration at listed project phases if
client is committed & involved in H&S management. (1=low
probability, 10=high probability)
1
2
3
4
5
6
7
8
9
10
X
X
X
X
X
X
X
2.2 (Site H&S outcome refers to accidents, incidents, safe worker behaviour,
awareness, safe work place, safe tools & equipment and safe work methods)
Impact = on site H&S outcome if H&S is not considered at
Project phase
listed project phases. (1=no impact, 10=very high negative
impact)
1
2
3
4
5
6
7
8
9
10
Project initiation
X
Project design
X
Procurement
X
Construction
X
Project commissioning
X
Project closeout
X
Project operations & maintenance
X
3.1
Contractor element
Development of staff competency
H&S staffing
Consult & communicate H&S
information
H&S audits and inspections
Policy, procedures & goals
Worker involvement
Hazard identification & risk
assessment
Top management commitment
Probability = of Contractors implementing listed elements
if clients provide finance for H&S (1=low probability,
10=high probability)
1
2
3
4
5
6
X
X
7
X
X
X
X
X
292
8
X
9
10
3.2
Contractor element
Development of staff competency
H&S staffing
Consult & communicate H&S
information
H&S audits and inspections
Policy, procedures & goals
Worker involvement
Hazard identification & risk
assessment
Top management commitment
3.3
Contractor element
Development of staff competency
H&S staffing
Consult & communicate H&S
information
H&S audits and inspections
Policy, procedures & goals
Worker involvement
Hazard identification & risk
assessment
Top management commitment
Probability = of Contractors implementing listed elements
if clients are adequately staffed for H&S (1=low probability,
10=high probability)
1
2
3
X
4
X
X
X
5
6
7
8
9
10
X
X
X
X
Probability = of Contractors implementing listed elements
if clients select procurement methods suitable for H&S
(1=low probability, 10=high probability)
1
2
3
X
4
5
X
X
X
X
X
X
X
293
6
7
8
9
10
3.4
Contractor element
Development of staff competency
H&S staffing
Consult & communicate H&S
information
H&S audits and inspections
Policy, procedures & goals
Worker involvement
Hazard identification & risk
assessment
Top management commitment
3.5
Contractor element
Development of staff competency
H&S staffing
Consult & communicate H&S
information
H&S audits and inspections
Policy, procedures & goals
Worker involvement
Hazard identification & risk
assessment
Top management commitment
3.6
Contractor element
Development of staff competency
H&S staffing
Consult & communicate H&S
information
H&S audits and inspections
Policy, procedures & goals
Worker involvement
Hazard identification & risk
assessment
Top management commitment
Probability = of Contractors implementing listed elements
if clients are involved in design & planning for H&S (1=low
probability, 10=high probability)
1
2
3
4
5
6
X
X
7
8
10
X
X
X
9
X
X
X
Probability = of Contractors implementing listed elements
if clients conduct H&S audits & inspections (1=low
probability, 10=high probability)
1
2
3
4
5
6
7
X
X
8
9
10
X
X
X
X
X
X
Probability = of Contractors implementing listed elements
if clients have effective H&S policies, procedures & goals
(1=low probability, 10=high probability)
1
2
3
4
5
X
6
7
8
X
X
X
9
X
X
X
X
294
10
3.7
Contractor element
Development of staff competency
H&S staffing
Consult & communicate H&S
information
H&S audits and inspections
Policy, procedures & goals
Worker involvement
Hazard identification & risk
assessment
Top management commitment
3.8
Contractor element
Development of staff competency
H&S staffing
Consult & communicate H&S
information
H&S audits and inspections
Policy, procedures & goals
Worker involvement
Hazard identification & risk
assessment
Top management commitment
4.1
Designer element
Development of staff competency
H&S staffing
Consult & communicate H&S
information
H&S audits and inspections
Policy, procedures & goals
Contractor involvement in design
review
Hazard identification & risk
assessment
Design for H&S
Probability = of Contractors implementing listed elements
if clients lead on H&S (1=low probability, 10=high
probability)
1
2
3
4
5
6
7
8
X
X
9
10
X
X
X
X
X
X
Impact = on project H&S culture if listed contractor’s
element is lacking. (1=no impact, 10=very high negative
impact)
1
2
X
X
X
X
3
X
4
5
6
7
8
9
10
X
X
X
Probability = of Designers implementing listed elements if
clients provide finance for H&S (1=low probability,
10=high probability)
1 2 3
4
5
6
7
8
9
10
X
X
X
X
X
X
X
X
295
4.2
Designer element
Development of staff competency
H&S staffing
Consult & communicate H&S
information
H&S audits and inspections
Policy, procedures & goals
Contractor involvement in design
review
Hazard identification & risk
assessment
Design for H&S
4.3
Designer element
Development of staff competency
H&S staffing
Consult & communicate H&S
information
H&S audits and inspections
Policy, procedures & goals
Contractor involvement in design
review
Hazard identification & risk
assessment
Design for H&S
Probability = of Designers implementing listed elements if
clients are adequately staffed for H&S (1=low probability,
10=high probability)
1 2 3
4
5
6
7
8
9
10
X
X
X
X
X
X
X
X
Probability = of Designers implementing listed elements if
clients select procurement methods suitable for H&S
(1=low probability, 10=high probability)
1 2 3
4
5
6
7
8
9
10
X
X
X
X
X
X
X
X
296
4.4
Designer element
Development of staff competency
H&S staffing
Consult & communicate H&S
information
H&S audits and inspections
Policy, procedures & goals
Contractor involvement in design
review
Hazard identification & risk
assessment
Design for H&S
4.5
Designer element
Development of staff competency
H&S staffing
Consult & communicate H&S
information
H&S audits and inspections
Policy, procedures & goals
Contractor involvement in design
review
Hazard identification & risk
assessment
Design for H&S
Probability = of Designers implementing listed elements if
clients are involved in design & planning for H&S (1=low
probability, 10=high probability)
1
2
3
4
5
6
7
8
X
X
X
9
10
X
X
X
X
X
Probability = of Designers implementing listed elements if
clients conduct H&S audits & inspections (1=low
probability, 10=high probability)
1
2
3
4
5
6
7
8
9
X
X
X
X
10
X
X
X
X
297
4.6
Designer element
Development of staff competency
H&S staffing
Consult & communicate H&S
information
H&S audits and inspections
Policy, procedures & goals
Contractor involvement in design
review
Hazard identification & risk
assessment
Design for H&S
4.7
Designer element
Development of staff competency
H&S staffing
Consult & communicate H&S
information
H&S audits and inspections
Policy, procedures & goals
Contractor involvement in design
review
Hazard identification & risk
assessment
Design for H&S
Probability = of Designers implementing listed elements
if clients have effective H&S policies, procedures & goals
(1=low probability, 10=high probability)
1
2
3
4
5
6
7
8
9
X
X
X
X
10
X
X
X
X
Impact = on project H&S culture if listed designer’s
element is lacking. (1=no impact, 10=very high negative
impact)
1
2
3
4
5
6
7
8
X
X
X
X
X
9
10
X
X
X
298
5.1
Project phase
Project initiation
Project design
Project procurement
Project construction
Project commissioning
Project closeout
5.2
Project phase
Project initiation
Project design
Project procurement
Project construction
Project commissioning
Project closeout
5.3
Project phase
Project initiation
Project design
Project procurement
Project construction
Project commissioning
Project closeout
5.4
Project phase
Project initiation
Project design
Project procurement
Project construction
Project commissioning
Project closeout
Probability = of H&S consideration at listed project phases if
designers are adequately staffed for H&S (1=low probability,
10=high probability)
1
2
3
4
5
6
7
8
9
10
X
X
X
X
X
X
Probability = of H&S consideration at listed project phases if
designers have H&S competent staff (1=low probability,
10=high probability)
1
2
3
4
5
6
7
8
9
10
X
X
X
X
X
X
Probability = of H&S consideration at listed project phases if
designers consult & communicate H&S information to all
stakeholders (1=low probability, 10=high probability)
1
2
3
4
5
6
7
8
9
10
X
X
X
X
X
X
Probability = of H&S consideration at listed project phases if
designers conduct H&S audits & inspections (1=low
probability, 10=high probability)
1
2
3
4
5
6
7
8
9
10
X
X
X
X
X
X
299
5.5
Project phase
Project initiation
Project design
Project procurement
Project construction
Project commissioning
Project closeout
5.6
Project phase
Project initiation
Project design
Project procurement
Project construction
Project commissioning
Project closeout
5.7
Project phase
Project initiation
Project design
Project procurement
Project construction
Project commissioning
Project closeout
Probability = of H&S consideration at listed project phases if
designers have effective H&S policy, procedures & goals
(1=low probability, 10=high probability)
1
2
3
4
5
6
7
8
9
10
X
X
X
X
X
X
Probability = of H&S consideration at listed project phases if
designers involve contractors in design reviews (1=low
probability, 10=high probability)
1
2
3
4
5
6
7
8
9
10
X
X
X
X
X
X
Probability = of H&S consideration at listed project phases if
designers conduct hazard identification & risk assessment
(1=low probability, 10=high probability)
1
2
3
4
5
6
7
8
9
10
X
X
X
X
X
X
300
5.8 (Site H&S outcome refers to accidents, incidents, safe worker behaviour,
awareness, safe work place, safe tools & equipment and safe work methods)
Impact= on site H&S outcome if H&S is not considered at
Project phase
the listed project phases (1=no impact, 10=very high negative
impact)
1
2
3
4
5
6
7
8
9
10
Project initiation
X
Project design
X
Project procurement
X
Project construction
X
Project commissioning
X
Project closeout
X
6.1
H&S project elements
Consultation & communication of
H&S information
Frequent H&S audits & inspections
Client commitment & involvement
Designer commitment &
involvement
Contractor commitment &
involvement
Safety program
H&S staffing
Staff competency on H&S
Prequalifying contractors &
subcontractors on H&S
Designing & planning for H&S
H&S risk assessment &
management
H&S meetings
Site worker involvement in H&S
Effective policy, procedures &
clear goals
Impact= on project H&S culture if listed elements for all
parties are lacking (1=no impact, 10=very high negative
impact)
1
2
3
4
5
6
7
8
9
10
X
X
X
X
X
X
X
X
X
X
X
X
X
X
301
7.1 PERSONAL INFORMATION OF EXPERT PANEL MEMBERS
Title (Mr, Mrs, Ms, Dr, Prof)
Dr & Prof
Highest qualification
DSc
Field of specialisation
Construction management
Professional registration (Pr E,
Chartered Civil Engineer, FICE & Chartered Builder,
PE etc.)
FCIOB
Years of experience (construction 40
H&S)
Current employer
Retired, now independent consultant
Position
Professor Emeritus
Country
Wales, UK
Thank you for taking your time to fill out this first round survey. The second round of the
Delphi process will begin on March 15, 2010.
Please do not hesitate to contact me or my promoter Prof Theo Haupt if you have any
questions about this survey or about the research project in general. Kindly see contact
details below.
Contact details:
Innocent Musonda
Ph.D. Candidate
Dept. of Construction Management and Quantity Surveying, University of Johannesburg
Doornfontein Campus 2028,
Johannesburg, South Africa.
Tel.: +27115596655
Mobile: +27730769652
Email: innocentmusonda@gmail.com
Prof Theo Haupt, PhD, MPhil, Pr.CM
Director: Building Construction Science, College of Architecture, Art and Design,
Mississippi state University,
Visiting Professor: Dept. of construction Management & Quantity Surveying, University
of Johannesburg, South Africa.
899 College view, 128 Giles Hall, P.O. AQ, Mississippi State, MS 39762,
Tel: 662 325 2547
Mobile: 662 312 5328
Email: thaupt@caad.msstate.edu
APPENDIX D
302
DELPHI INSTRUCTIONS FOR ROUND 2 AND AN EXAMPLE OF
COMPLETED QUESTIONNAIRE WITH GROUP MEDIAN
DELPHI SURVEY – ROUND 2
Thank you for completing Round 1 Delphi survey. We recognize that the survey required
a significant time investment to complete thoughtfully. We appreciate your time and
effort. This Round 2 survey continues the Delphi process for this study. The purpose of
Round 2 is to provide you with the opportunity to change your response, if desired, given
the median group response for each question and element.
The second round survey is intended to take approximately 20 minutes as you are only
being asked to review your previous responses given the collective group median. When
you have finished answering all of the questions, please email your response to
innocentmusonda@gmail.com by Friday, April 30, 2010.
INSTRUCTIONS
For each element you will see 2 values: your response from the Round 1 survey (indicated
with a yellow highlighted box), and the group median from the Round 1 survey indicated
in the column to the far right hand of each table. Please take one of the following three
actions for each category:
1.
2.
3.
4.
Accept the group median response by leaving the field completely unchanged.
Maintain your original response by placing an ‘X’ in the highlighted field*.
Indicate a new response by placing an ‘X’ in the appropriate field*.
Provide fresh answer in blank boxes as these were not provided in first round.
* *If your response is more than ten Percent above or below the group median please
provide a reason for your outlying response in the field provided.
303
1.1 POLITICAL INFLUENCE: The profile and practices within government related to the
construction industry. This is reflected in Government’s position to give much higher
weighting to H&S in tender evaluation, contract award, bonuses or alternative rewards such as
tax and training incentives.
Client element
What is the Probability of a Client Implementing each of
the listed elements if pressured by political environment?
(1=low probability, 10=high probability)
1 2 3 4 5 6 7 8 9 10 MEDIAN
Finance H&S Management
Appoint H&S agent
Employ permanent H&S staff
Be involved in design &
planning
H&S prequalification criteria for
contractors
Select procurement method
suitable for H&S
Schedule H&S requirements
prior to bidding process
6
7
8
5
X
X
X
X
7
X
6
X
3
X
Schedule H&S in contracts
Conduct H&S inspections &
audits
Have effective H&S policy,
procedures & goals
Assume H&S leadership role
4
8
X
X
X
X
1.2 SOCIAL INFLUENCE: Reflected in for example, requirements for community
approval of projects upon satisfying H&S standards
Client element
What is the Probability of a Client Implementing each of
the listed elements if pressured by the social
environment? (1=low probability, 10=high probability)
1 2 3 4 5 6 7 8 9 10 MEDIA
N
Finance H&S Management
X
4
Appoint H&S agent
X
5
Employ permanent H&S staff
X
5
Be involved in design &
X
4
planning
H&S prequalification criteria for
X
6
contractors
Select procurement method
X
4
suitable for H&S
Schedule H&S requirements
X
4
prior to bidding process
Schedule H&S in contracts
X
6
Conduct H&S inspections &
X
6
audits
Have effective H&S policy,
X
6
procedures & goals
Assume H&S leadership role
X
4
304
8
7
1.3 MATERIALS & METHODS: Refers to new materials and methods of construction
adopted or advocated for in the industry that enhance H&S, such as use of prefabricated units.
Client element
What is the Probability of a Client Implementing each of
the listed elements if H&S compliant materials &
methods of construction are adopted in the industry?
(1=low probability, 10=high probability)
1 2 3 4 5 6 7 8 9 10 MEDIAN
Finance H&S Management
Appoint H&S agent
Employ permanent H&S staff
Be involved in design &
planning
H&S prequalification criteria for
contractors
Select procurement method
suitable for H&S
Schedule H&S requirements
prior to bidding process
Schedule H&S in contracts
Conduct H&S inspections &
audits
Have effective H&S policy,
procedures & goals
Assume H&S leadership role
X
5.5
5
6.5
5
X
X
X
X
5
X
5.5
X
5
X
5
7.5
X
X
7.5
X
7.5
1.4 ECONOMIC PRESSURE: Refers to economic incentive for implementing H&S as well
as a tangible loss for not implementing H&S
Client element
What is the Probability of a Client Implementing each of
the listed elements if pressured by economic
environment? (1=low probability, 10=high probability)
1 2 3 4 5 6 7 8 9 1 MEDIAN
0
Finance H&S Management
X
8
Appoint H&S agent
X
9
Employ permanent H&S staff
X
8
Be involved in design & planning
X
8
H&S prequalification criteria for
X
8
contractors
Select procurement method
X
8
suitable for H&S
Schedule H&S requirements prior
X
8
to bidding process
Schedule H&S in contracts
Conduct H&S inspections &
audits
Have effective H&S policy,
procedures & goals
Assume H&S leadership role
X
8
9
X
X
X
305
9
8
1.5 REGULATIONS AND LEGISLATIVE FRAMEWORK: Refers to regulations &
legislature appropriate for H&S improvement and an effective enforcement structure.
Client element
What is the Probability of a Client Implementing each of
the listed elements if pressured by regulations &
legislative framework? (1=low probability, 10=high
probability)
1
2 3 4 5 6 7 8 9 1 MEDIA
0 N
Finance H&S Management
X
8
Appoint H&S agent
X
9
Employ permanent H&S staff
X
9
Be involved in design & planning
X
8
H&S prequalification criteria for
X
9
contractors
Select procurement method
X
8
suitable for H&S
Schedule H&S requirements prior
X
9
to bidding process
Schedule H&S in contracts
X
9
Conduct H&S inspections &
X
9
audits
Have effective H&S policy,
X
9
procedures & goals
Assume H&S leadership role
X
8
1.6 CONSTRUCTION PROFESSIONAL BODIES: Refers to bodies empowered to register
designers, contractors and clients in order that registered entities may conduct their business in
the construction industry
Client element
What is the Probability of a Client Implementing each of
the listed elements if pressured by construction
professional bodies? (1=low probability, 10=high
probability)
1 2 3 4 5 6
7 8 9 10 MEDIA
N
Finance H&S Management
Appoint H&S agent
Employ permanent H&S staff
Be involved in design & planning
H&S prequalification criteria for
contractors
Select procurement method
suitable for H&S
Schedule H&S requirements prior
to bidding process
Schedule H&S in contracts
Conduct H&S inspections &
audits
Have effective H&S policy,
procedures & goals
Assume H&S leadership role
X
X
6
5
5
6
5
X
6
X
6.5
X
X
6
6
X
X
X
X
X
306
6
5
1.7 H&S CULTURE: Refers to that observable degree of effort with which all organisational
members direct their attention and actions towards improving H&S on a daily basis. The way
things are done in an organisation. The way things are done in an organisation.
Client element
What is the Impact on contractors’ H&S culture if each of
the listed client elements is lacking? (1=no impact, 10=very
high negative impact)
No
Low
Medium High
Very
impact impact impact
impact High
impact
1
2 3 4 5
6 7 8 9 10 MEDIAN
Commitment to H&S
Involvement in H&S management
H&S competence
H&S leadership
1.8 H&S CULTURE: (see 1.7)
Client element
X
X
8
8
8
8
X
X
What is the Impact on designers’ H&S culture if each of
the listed client elements is lacking? (1=no impact, 10=very
high negative impact)
No
Low
Medium High
Very
impact impact impact
impact High
impact
1
2 3 4 5
6 7 8 9 10 MEDIAN
Commitment to H&S
Involvement in H&S management
H&S competence
H&S leadership
X
8
9
7
9
X
X
X
1.9 H&S CULTURE: (see 1.7)
Environment element
Political pressure
Social pressure
Economic pressure
Improvement on materials &
methods
Contractor commitment &
involvement
Regulations & legislative frame
work
impacting
directly
on client
Professional
bodies
influence
What is the Impact on clients’ H&S culture if each of
the listed elements is lacking? (1=no impact, 10=very
high negative impact)
No
Low
Mediu High
Very
impact impact m
impact High
impact
impact
1
2
3
4
X
5
6
7
8
9
X
X
X
X
X
X
307
1 MEDIAN
0
8
6
9
6
6
9
6
2.1 CLIENT COMMITMENT: Reflected in allocation of resource, procurement methods
adopted, establishment of policies, procedures and goals.
What is the Probability that H&S will be considered at
Project phase
each of the following listed project phases if a client is
committed to H&S management? (1=low probability,
10=high probability)
1 2 3 4 5 6 7 8 9 1 MEDIAN
0
Project initiation
X
8
Project design
X
8
Procurement
X
8
Construction
X
9
Project commissioning
X
9
Project closeout
X
9
Project operations & maintenance
X
8
2.2 CLIENT INVOLVEMENT: Refers to client active participation as reflected in
conducting of H&S audits & inspections, involved in design & planning, monitor & review
performance and consult and communicate H&S information
Project phase
What is the Probability that H&S will be considered at
each of the following listed project phases if a client is
involved in H&S management? (1=low probability,
10=high probability)
1 2 3 4 5 6 7 8 9 10 MEDIAN
Project initiation
X
9
Project design
X
9
Procurement
X
8
Construction
X
10
Project commissioning
X
9
Project closeout
X
9
Project operations & maintenance
X
9
2.3 H&S COMPETENCE: Refers to an organisation having attained capability to manage &
improve H&S through employing permanent H&S competent personnel and by improving
staff competence through education & training.
Project phase
What is the Probability that H&S will be considered at
each of the following listed project phases if a client is
H&S competent? (1=low probability, 10=high
probability)
1 2 3 4 5 6 7 8 9 10 MEDIAN
Project initiation
Project design
Procurement
Construction
Project commissioning
Project closeout
Project operations & maintenance
X
X
X
X
X
X
X
308
8
8
6
9
8
8
8
2.4 H&S LEADERSHIP: Refers to the responsibility of assuming a lead role and take full
responsibility for H&S planning, organisation, control, coordination, feedback and eventual
outcome
Project phase
What is the Probability that H&S will be considered at
each of the following listed project phases if a client
assumes H&S leadership on projects? (1=low
probability, 10=high probability)
1 2 3 4 5 6 7 8 9 1 MEDIAN
0
Project initiation
X
8
Project design
X
7
Procurement
X
7
Construction
X
10
Project commissioning
X
8
Project closeout
X
8
Project operations & maintenance
X
7
2.5 SITE H&S OUTCOME: Refers to accidents, incidents, safe worker behaviour, awareness,
safe work place, safe tools & equipment and safe work methods
Project phase
What is the Impact on site H&S outcome if H&S is not
considered at each of the following listed project phases?
(1=no impact, 10=very high negative impact)
1 2 3 4 5 6 7 8 9 1 MEDIAN
0
Project initiation
X
7
Project design
X
7
Procurement
X
8
Construction
X
10
Project commissioning
X
8
Project closeout
X
8
Project operations & maintenance
X
9
309
3.1 CLIENT COMMITMENT: see 2.1
Contractor element
What is the Probability of Contractors implementing
each of the listed elements if clients are committed to
H&S management? (1=low probability, 10=high
probability)
1 2 3 4 5 6 7 8 9 1 MEDIAN
0
Develop staff competency on
X
8
H&S
Employ permanent H&S staff
X
9
Consult & communicate H&S
X
8
information to stakeholders
Conduct H&S audits and
inspections
Establish & implement H&S
Policies, procedures & goals
X
X
9
8
Involve workers in H&S
management
Hazard identification & risk
assessment
Top management commitment
X
X
X
8
9
8
3.2 CLIENT INVOLVEMENT: (see 2.2)
Contractor element
What is the Probability of Contractors implementing
each of the listed elements if clients are involved in
H&S management? (1=low probability, 10=high
probability)
1 2 3 4 5 6 7 8 9 1 MEDIAN
0
Develop staff competency on
X
8
H&S
Employ permanent H&S staff
X
8
Consult & communicate H&S
X
8
information to stakeholders
Conduct H&S audits and
inspections
Establish & implement H&S
Policies, procedures & goals
X
X
8
9
Involve workers in H&S
management
Hazard identification & risk
assessment
Top
management commitment
X
X
7
8
9
X
310
3.3 H&S COMPETENCE: (see 2.3)
Contractor element
What is the Probability of Contractors implementing
each of the listed elements if clients are H&S
competent? (1=low probability, 10=high probability)
1 2 3 4 5 6 7 8 9 1 MEDIAN
0
Develop staff competency on H&S
X
7
Employ permanent H&S staff
X
8
Consult & communicate H&S
X
8
information to stakeholders
Conduct H&S audits and
inspections
Establish & implement H&S
Policies, procedures & goals
X
X
8
7
Involve workers in H&S
management
Hazard identification & risk
assessment
Top management commitment
X
X
X
8
8
8
3.4 H&S LEADERSHIP: (see 2.4)
Contractor element
What is the Probability of Contractors implementing
each of the listed elements if clients assume H&S
leadership? (1=low probability, 10=high probability)
1 2 3 4 5 6 7 8 9 1 MEDIAN
0
Develop staff competency on H&S
X
8
Employ permanent H&S staff
X
9
Consult & communicate H&S
X
9
information to stakeholders
Conduct H&S audits and
inspections
Establish & implement H&S
Policies, procedures & goals
X
X
Involve workers in H&S
management
Hazard
identification & risk
assessment
Top management commitment
X
X
X
311
9
9
8
9
9
3.5 H&S CULTURE: (see 1.7)
Contractor element
What is the Impact on project H&S culture if each of
the following listed contractor’s elements is lacking?
(1=no impact, 10=very high negative impact)
No
impact
1
2
Develop staff competency on
H&S
Employ permanent H&S staff
Consult & communicate H&S
information to stakeholders
Low
impact
3
4
Mediu
m
impact
5
6
X
High
impact
7
8
Very
High
impact
9
1 MEDIAN
0
8
X
8
8
X
Conduct H&S audits and
inspections
Establish & implement H&S
Policies, procedures & goals
X
X
Involve workers in H&S
management
Hazard identification & risk
assessment
Top management commitment
X
X
X
9
8
9
9
9
4.1 CLIENT COMMITMENT: (see 2.1)
Designer element
What is the Probability of Designers implementing each
of the listed elements if clients are committed to H&S
management? (1=low probability, 10=high probability)
1 2 3 4 5 6 7 8 9 1 MEDIA
0 N
Develop staff competency on
X
7
H&S
Employ permanent H&S staff
X
8
Consult & communicate H&S
X
8
information to stakeholders
Conduct H&S audits and
inspections
Establish & implement H&S
Policies, procedures & goals
X
X
Involve Contractors in design
reviews
Hazard identification & risk
assessment
Design for H&S
7
X
X
X
312
7
9
8
8
4.2 CLIENT INVOLVEMENT: (see 2.2)
Designer element
What is the Probability of Designers implementing each
of the listed elements if clients are involved in H&S
management? (1=low probability, 10=high probability)
1 2 3 4 5 6 7 8 9 1 MEDIAN
0
Develop staff competency on
X
8
H&S
Employ permanent H&S staff
X
8
Consult & communicate H&S
X
8
information to stakeholders
Conduct H&S audits and
inspections
Establish & implement H&S
Policies, procedures & goals
X
8
X
8
Involve Contractors in design
reviews
Hazard identification & risk
assessment
Design for H&S
X
8
8
9
X
X
4.3 H&S COMPETENCE: (see 2.3)
Designer element
What is the Probability of designers implementing each
of the listed elements if clients are H&S competent?
(1=low probability, 10=high probability)
1 2 3 4 5 6 7 8 9 1 MEDIAN
0
Develop staff competency on
X
8
H&S
Employ permanent H&S staff
X
8
Consult & communicate H&S
X
7
information to stakeholders
Conduct H&S audits and
inspections
Establish & implement H&S
Policies, procedures & goals
X
7
X
8
Involve Contractors in design
reviews
Hazard identification & risk
assessment
Design for H&S
X
X
X
8
8
8
313
4.4 H&S LEADERSHIP: (see 2.4)
Designer element
What is the Probability of designers implementing each
of the listed elements if clients assume H&S leadership?
(1=low probability, 10=high probability)
1 2 3 4 5 6 7 8 9 1 MEDIAN
0
Develop staff competency on
X
7
H&S
Employ permanent H&S staff
X
6
Consult & communicate H&S
X
7
information to stakeholders
Conduct H&S audits and
inspections
Establish & implement H&S
Policies, procedures & goals
X
7
X
8
Involve Contractors in design
reviews
Hazard identification & risk
assessment
Design for H&S
X
8
X
X
314
9
8
4.5 H&S CULTURE: (see 1.7)
Designer element
What is the Impact on project H&S culture if each of
the following listed designer’s element is lacking? (1=no
impact, 10=very high negative impact)
No
impact
1
2
Develop staff competency on
H&S
Employ permanent H&S staff
Consult & communicate H&S
information to stakeholders
Low
impact
3
4
Mediu
m
impact
5
6
High
impact
7
8
X
X
Conduct H&S audits and
inspections
Establish & implement H&S
Policies, procedures & goals
1 MEDIAN
0
6.5
X
X
7.5
7
X
315
9
6.5
8
X
Involve Contractors in design
reviews
Hazard identification & risk
assessment
Design for H&S
Very
High
impact
8
X
8
X
9
5.1 DESIGNERS H&S COMMITMENT: As for client commitment defined in 2.1
What is the Probability that H&S will be considered at
Project phase
each
MEDIAN
1 of
2 the3following
4 5 listed
6 project
7 8phases
9 if1designers
were
committed
to
H&S
management?(1=low
0
Project initiation
X
7
probability, 10=high probability)
Project design
X
8
Project procurement
X
8
Project construction
X
8
Project commissioning
X
8
Project closeout
X
8
5.2 DESIGNER INVOLVEMENT: As for client involvement defined in 2.2
What is the Probability that H&S will be considered at
Project phase
each
MEDIAN
1 of
2 the3following
4 5 listed
6 project
7 8phases
9 if1designers
were
involved
in
H&S
management?
(1=low
0
Project initiation
X
7
probability, 10=high probability)
Project design
X
8
Project procurement
X
8
Project construction
X
8
Project commissioning
X
8
Project closeout
X
8
316
5.3 H&S COMPETENCE: (see 2.3)
What is the Probability that H&S will be considered at
Project phase
each
MEDIAN
1 of
2 the3following
4 5 listed
6 project
7 8phases
9 if1designers
developed
H&S
competence?
(1=low
probability,
0
Project initiation
X
7
10=high probability)
Project design
X
9
Project procurement
X
8
Project construction
X
9
Project commissioning
X
8
Project closeout
X
8
5.4 H&S LEADERSHIP: (see 2.4)
What is the Probability that H&S will be considered at
Project phase
each
MEDIAN
1 of
2 the3following
4 5 listed
6 project
7 8phases
9 if1designers
assumed
leadership
role
on
projects?
(1=low
0
Project initiation
X
8
probability, 10=high probability)
Project design
X
9
Project procurement
X
8
Project construction
X
9
Project commissioning
X
9
Project closeout
X
9
317
5.5 SITE H&S OUTCOME: (see 2.5)
What is the Impact on site H&S outcome if each of the
Project phase
listed elements for all parties is lacking? (1=no impact,
10=very high negative impact)
No
Low
Medium High
Very
impact impact impact
impact High
impact
1 2 3 4 5
6 7 8 9 1 MEDIAN
0
Client commitment
X
7
Client involvement
X
7
Client leadership
X
7
Client H&S competence
X
8
Designer commitment
X
9
Designer involvement
X
9
Designer leadership
X
9
Designer H&S competence
X
9
Contractor top management
X
9
commitment
Contractor top management
X
8
involvement
Contractor top management
X
8
leadership
Contractor H&S competence
X
9
318
6.1 H&S CULTURE: (see 1.7)
H&S project elements
Client commitment
Client involvement
Client leadership
Client H&S competence
Designer commitment
Designer involvement
Designer leadership
Designer H&S competence
Contractor top management
commitment
Contractor top management
involvementtop management
Contractor
leadership H&S competence
Contractor
What is the Impact on the project H&S culture if each of
the listed elements for all parties is lacking? (1=no impact,
10=very high negative impact)
No
Low
Medium High
Very
impact impact impact
impact High
impact
1 2 3 4
5
6 7 8 9 1 MEDIAN
0
X
9
X
9
X
8.5
X
9
X
9
X
9
X
8
X
9
X
10
X
10
X
9
X
8.5
Thank you for taking your time to respond to this second round survey. Your input is
incredibly valuable and the time you have spent on this panel is greatly appreciated. You
will be informed of this round’s results by Friday 14 May, 2010.
Please do not hesitate to contact me or my promoter Prof Theo Haupt if you have any
questions or contributions on this survey or about the research project in general. Kindly
see contact details below.
Contact details:
Innocent Musonda
Ph.D. Candidate
Dept. of Construction Management and Quantity Surveying, University of Johannesburg
Doornfontein Campus 2028,
Johannesburg, South Africa.
Tel.: +27115596655,
Mobile: +27730769652
Email: innocentmusonda@gmail.com
319
Prof Theo Haupt, PhD, MPhil, Pr.CM
Director: Building Construction Science, College of Architecture, Art and Design,
Mississippi state University,
Visiting Professor: Dept. of construction Management & Quantity Surveying, University
of Johannesburg, South Africa.
899 College view, 128 Giles Hall, P.O. AQ, Mississippi State, MS 39762,
Tel: 662 325 2547
Mobile: 662 312 5328
Email: thaupt@caad.msstate.edu
320
APPENDIX E
INSTRUCTIONS TO EXPERTS ON DELPHI STUDY ROUND 3
DELPHI SURVEY – ROUND 3
Thank you for completing Round 2 Delphi survey. We recognize that the survey required
a significant time investment to complete thoughtfully. We appreciate your time and
effort. This Round 3 survey concludes the Delphi process for this study. The purpose of
Round 3 is to provide you with a final opportunity to change your response, if desired,
given the median group response and reasons for outlying responses for each element.
The round 3 survey is intended to take significantly less time than the first two rounds as
you are only being asked to review your previous responses given the collective group
median and reasons for the outlying responses. When you have finished answering all of
the questions, please email your response to innocentmusonda@gmail.com by Friday,
July 9, 2010.
INSTRUCTIONS
For each element you will see 2 values: your response from Round 2 survey (indicated
with a yellow highlighted box), and the group median from the Round 2 survey indicated
in the column to the far right hand of each table. Please take one of the following three
actions for each category:
1. Accept the group median response by leaving the field completely unchanged.
2. Maintain your original response by placing an ‘X’ in the highlighted field*.
3. Indicate a new response by placing an ‘X’ in the appropriate field*.
* *If your response is more than ten Percent (one unit) above or below the group median
please provide a reason for your outlying response in the field provided if you have not
done so already.
We URGE you to review and consider the median and the responses provided by the
other expert panellists when considering your final responses for each element.
321
APPENDIX F
RESEARCH INTRODUCTION LETTER AND QUESTIONNAIRE
UNIVERSITY OF JOHANNESBURG,
FACULTY OF ENGINEERING AND BUILT ENVIRONMENT
September 20, 2010
Dear Respondent,
We are inviting you to participate in a research project to study Health and Safety (H&S)
practice in the Construction industry. Along with this letter is a short questionnaire that
asks a variety of questions about H&S. We ask you to look over the questionnaire and, if
you choose to do so, complete it and send or give it back to me. It should take you about
20 minutes to complete.
We do not know of any risks to you if you decide to participate in this survey and we
guarantee that your responses will not be identified with you personally. You are also
encouraged not to put your name on the questionnaire.
I hope you will take the time to complete this questionnaire and return it. Your
participation is voluntary. Regardless of whether you choose to participate, please let us
know if you would like a summary of our findings. Contact details are at the end of the
questionnaire.
If you have any questions or concerns about completing the questionnaire or about being
in this study, you may contact Mr I. Musonda at +27730769652. The University of
Johannesburg has approved this study. If you have any concerns about your rights as a
participant in this study you may also contact the promoter of this study, Prof Theo Haupt
on Mobile: 662 312 5328, Email: thaupt@caad.msstate.edu.
Sincerely
Innocent Musonda
University of Johannesburg
Tel: +27 11 559 6655, Mobile: +27 73076 9652, email: innocentmusonda@gmail.com
322
HEALTH AND SAFETY (H&S) PRACTICE QUESTIONNAIRE
(To be filled for a project that you are participating or have recently participated in)
Please tick (√) in the appropriate box provided.
SECTION 1.0 (General project particulars)
1.1
Kindly state your position on the project:
___________________________________
1.2
What is your profession?
___________________________________
1.3
What type of organisation are you currently employed with:
Consulting/Designer
1.4
Contractor
Client
What type of construction project are you currently working on (to which this
questionnaire will refer to)?
Civil engineering
project
1.5
Building Construction
project
other
Please indicate below the range in which the contract value for this project falls
4 - 6.5
1.6
1.7
project
What type of client is it for the project?
Property
developer
Parastatal
organisation
Public/
Government
Mining
organisation
Other
Please indicate below the method that was used to select the contractor for this
Open tender
1.8
Contract value in Millions of South African Rands
6.5 - 13
13 - 40
40 - 130
Over 130
Million
Selective
tender
Negotiated
Other
Kindly indicate below the project delivery system used for this project
323
Designtender-build
1.9
Design &
build
Construction
management
Other
Please indicate below the type of contract in use on this project
Lump
Sum
SECTION 2.0
Bill Of
Quantities
based (Fixed)
Bill Of
Quantities based
(Fluctuating)
Cost
reimbursement
(Health and safety (H&S) practice)
2.1
Strongly
agree
Agree
Neither
Disagree
Strongly
disagree
Please rate how strongly you agree or disagree with each of the following
statements regarding H&S consideration at various project stages by placing a
check mark in the appropriate box
Health and safety:
Extent to which you agree
PR1 Was adequately considered at project initiation
PR2 Was adequately considered in detail at design
stage
PR3 Was a major consideration in appointing the
contractor
PR4 Is emphasised during construction stage
PR5 Is a key consideration at project closeout
2.2
L1
L2
L3
L4
L5
L6
L7
Considers H&S implications before making
decisions on the project
Has an effective H&S policy
Monitors H&S on the project throughout all
stages
Monitors designers’ H&S implementation
Monitors contractor’s H&S implementation
Mandated designers to manage project H&S
Requires that the contractor manages project
324
Strongly
agree
Agree
Neither
Disagree
Strongly
disagree
Please rate how strongly you agree or disagree with the following statements
regarding the client
The project client:
Extent to which you agree
L8
H&S
Coordinates designers & contractor to ensure
good H&S
2.3
CT1
CT2
CT3
CT4
CT5
CT6
CT7
CT8
CT9
Strongly
agree
Agree
Neither
Disagree
Strongly
disagree
Please rate how strongly you agree or disagree with the following statements
regarding the client
The project client has:
Extent to which you agree
Demonstrated positive attitude toward H&S
Actively promoted H&S in a consistent
manner across all levels
Provided finance for H&S
Supported implementation of H&S activities
Put in efforts to ensure every aspect of work
& operations are routinely evaluated for H&S
Conducted regular H&S tours on the project
Been involved in investigations of accidents,
incidents & ill-health on the project
Set H&S as an important agenda item in
every project progress meeting
Set H&S as a No.1 priority on the project
2.4
I1
I2
I3
I4
I5
I6
I7
Is personally active in critical project H&S
activities
Is always present in project H&S meetings
Contributes to H&S training
Is active in overseeing of H&S on critical
operations
Has constantly stayed “in-touch” on H&S
issues
Always communicates information on H&S to
all parties
Conducts regular audits & inspections
325
Strongly
agree
Agree
Neither
Disagree
Strongly
disagree
Please rate how strongly you agree or disagree with the following statements
regarding the client
The project client:
Extent to which you agree
2.5
Strongly
agree
Agree
Neither
Disagree
Strongly
disagree
Please rate how strongly you agree or disagree with the following statements
regarding the client
The project client:
Extent to which you agree
Has set up a formal reporting system of
incidents & accidents on the project
CN3 Involved all parties in planning for H&S on
the project
CN4 Involves all parties in H&S review
CN5 Has provided timely feedback on reported
accidents & incidents on the project
CN6 Communicates risk findings to all parties on
the project
CN7 Clearly made H&S policy statements for the
project
H&S roles
&
CN8 Has clearly outlined
responsibilities for all parties on the project
expected
CN9 Has clearly communicated
performance on H&S to all
CN10 Has provided Information on H&S risk
control to all parties
CN1
2.6
CE1 Representatives have demonstrated knowledge
of H&S
CE2 Conducts H&S training for its own staff
CE3 Deployed staff on the project that are qualified
to manage H&S
CE4 Ensured that H&S induction to client staff was
done on the project
326
Strongly
agree
Agree
Neither
Disagree
Strongly
disagree
Please rate how strongly you agree or disagree with the following statements
regarding the client
The client:
Extent to which you agree
2.7
P6
P7
P8
2.8
CT4
CT5
CT6
CT7
CT8
Strongly
agree
Agree
Extent to which you agree
strongly
disagree
The contractor’s top management:
CT1
CT2
CT3
Neither
Please rate how strongly you agree or disagree with the following statements
regarding the Contractor’s top management (Contractor’s top management
refers to contractors staff with a position of project manager and higher)
Has demonstrated a positive H&S attitude
Has actively promoted H&S
Has provided adequate resources for H&S
implementation
Has put in effort to ensure routine H&S
evaluation in all work schedules
Has set up incentives for good H&S
behaviour
Has set H&S as a major agenda item in all
project meetings
Actively monitors H&S programs on the
project
Always attends H&S meetings on the
construction site
327
Strongly
agree
P5
Agree
P4
Neither
P2
P3
Has programs to monitor and analyse H&S
implementation
Has clear project H&S goals
Scheduled H&S as a key contract
prequalification criteria for all parties involved
in the project
Scheduled H&S in all contracts for the parties
involved in the project
Conducts regular H&S performance
measurement
Has its own H&S committee
Conducts Hazard identification & risk
assessment
Required that designers adequately address
H&S in their designs
Disagree
P1
Disagree
Strongly
disagree
Please rate how strongly you agree or disagree with the following statements
regarding the client
The client:
Extent to which you agree
CT9
CT10
CN1
CN2
Conducts H&S inspections/audits regularly
Is always involved in accident /incident
investigations
Has established formal H&S reporting system
Has a formal H&S feedback system
Involves workers in H&S planning
Provides timely feedback on
incidents/accidents
CN5 Always communicates risk findings to all
workers
CN6 Has made a clear H&S policy statement on
the project
CN7 Has made clear the expected H&S
performance standard on the project
CN8 Has provided information on H&S risk
control
CN9 Regularly makes H&S briefs
CN10 Has “stayed –in- touch” on H&S issues
CN3
CN4
2.9
P1
P2
P3
P4
P5
P6
P7
P8
P9
P10
Has an effective H&S policy
Has an effective H&S plan
Scheduled H&S in all subcontractor contracts
or prequalification documents
Conducts regular H&S performance
measurement
Conducted a Hazard identification and risk
assessment for the project
Actively oversees H&S on critical operations
Has set up a formal H&S training program
Considers H&S knowledge to be a
requirement for employment
Has permanent H&S staff
Ensures that H&S induction is conducted for
all employees
328
Strongly
agree
Agree
Neither
Disagree
Strongly
disagree
Please rate how strongly you agree or disagree with the following statements
regarding the Contractor’s top management (Contractor’s top management
refers to contractors staff with a position of project manager and higher)
The contractor’s top management:
Extent to which you agree
2.10
CT1
CT2
CT3
CT4
CT5
CT6
CT7
CT8
CT9
CN1
CN2
CN3
CN4
CN5
CN6
CN7
CN8
Has demonstrated positive H&S attitude
Actively promotes H&S on the project
Provides information on H&S
Puts in effort to ensure routine H&S
evaluation in all work schedules
Conduct regular H&S inspections/audits
Ensures that H&S is always on the agenda of
project meetings
Regularly monitor H&S programs
Regularly attend H&S meetings for the
project
Are always involved in accident/incident
investigations
Has set up a formal H&S reporting system
Involved the contractor to plan for H&S
provides timely feedback on H&S issues such
as incidents / accidents
Always communicates risk findings
Made clear the expected H&S standard
performance
Provided information on H&S risk control
Makes regular H&S briefings / reports
Has stayed ‘in-touch’ on H&S issues
329
Strongly
agree
Agree
Neither
Disagree
Strongly
disagree
Please rate how strongly you agree or disagree with the following statements
regarding the lead designer or designers (the term designer refers to all
consultants on the construction project. i.e. QS, Architect, Engineers and Project
Manager)
The lead designer/consultant:
Extent to which you agree
2.11
P1
P2
P3
P4
P5
P6
P7
P8
P9
Has a H&S policy for the project
Has a H&S plan for the project
Regularly measure H&S performance
Or other designers oversee H&S on critical
operations
Or other designers have clearly stated H&S
requirements on contract drawings /
specifications
Conducted a Hazard identification and risk
assessment for the project
Has clearly defined H&S roles and
responsibilities for the project
Has demonstrated knowledgeable of H&S
Ensures that H&S induction is conducted for
all on the project
330
Strongly
agree
Agree
Neither
Disagree
Strongly
disagree
Please rate how strongly you agree or disagree with the following statements
regarding the lead designer or designers (the term designer refers to all
consultants on the construction project. i.e. QS, Architect, Engineers and Project
Manager)
The lead designer/consultant:
Extent to which you agree
2.12
L1
L2
L3
L4
L5
I1
I2
I3
I4
P1
P2
P3
P4
P5
CT1
CT2
CT3
CT4
CT5
CT6
CT7
There is visible H&S leadership
There is an effective H&S policy
H&S programs are effective
H&S programs are regularly monitored
H&S implementation is well coordinated
between all parties
All project participants are actively involved In
H&S activities
All key project participants attend H&S
meetings
Adequate H&S training is provided on the
project
Regular H&S inspections / audits are
conducted on the project
H&S is well addressed in all project Contract
documents
H&S committee is effective
There is a permanent H&S officer
H&S requirements are clearly stated on
drawings / specifications
There is a H&S risk register
There is a general positive attitude towards
H&S
H&S is actively promoted
Every aspect of operations is evaluated in terms
of H&S
There is an incentive program for H&S
performance
Incidents & accident are always investigated
H&S is a major agenda item in all project
meetings
H&S is a number one priority just as cost, time
& quality are
331
Strongly
agree
Agree
Neither
Disagree
Strongly
disagree
Please rate how strongly you agree or disagree with the following statements
regarding the project
On this project,:
Extent to which you agree
2.13
CN1 Workers on site know where to report an
incident or accident
CN2 Workers are able to report accidents comfortably
CN3 Feedback on H&S issues is always provided to
workers
CN4 All accidents / incidents are reported by workers
CN5 All parties know what the H&S risks are on the
project
CN6 All parties know what the H&S policy for the
project is
CN7 All parties know what is expected of them on
H&S performance
CN8 All parties on the project have information on
H&S risk control
CN9 H&S induction is conducted for all project
participants
CE1 Formal H&S training is conducted for all
workers on the project
CE2 Formal H&S training is conducted for key
project staff
CE3 H&S skill and knowledge is a key consideration
for all staff recruitment
CE4 H&S skill and knowledge is a key consideration
for all professional staff recruitment
P.T.O.
332
Strongly
agree
Agree
Neither
Disagree
Strongly
disagree
Please rate how strongly you agree or disagree with the following statements
regarding the project
On this project,:
Extent to which you agree
2.14
E1
E2
E3
E4
E5
E6
2
Disagree
3
Neither
4
Agree
5
Strongly
agree
1
Strongly
disagree
Please rate how strongly you agree or disagree with the following statements
regarding the Environmental influence on construction clients
On this project, the client was compelled to:
Extent to which you agree
Ensure H&S implementation due to Legislative
requirements,
Actively participate in H&S implementation due
to the Economic reasons
Actively participate in H&S implementation due
to construction workers’ union’s influence
Consider H&S in making decisions on cost,
time & quality as a result of designers’
persuasion
Ensure H&S implementation due to Professional
bodies’ influence
Consider H&S due to Materials & technology
adopted for the project
Thank you for your contribution. We really value your contribution and time spent on
completing this questionnaire.
If you have any queries, please do not hesitate to contact the undersigned.
Innocent Musonda
University of Johannesburg
Tel: +27 11 559 6655, Mobile: +27 73076 9652, email: innocentmusonda@gmail.com
333
APPENDIX G
DELPHI MEAN ABSOLUTE DEVIATIONS
CONSENSUS ANALYSIS
MAD
QUESTION RND 1
RND 2
RND 3 RND 1
1.1
-1.119
-0.207 -0.033
1.2 -0.0797
-0.116 -0.165
1.3
-1.455
-0.508 -0.355
1.4
-1.985
0.136
0.173
1.5
-0.734
0.066 -0.041
1.6
-1.255
-0.417 -0.417
1.7
-0.596
-0.159 -0.227
1.8
-0.596
-0.205 -0.159
1.9
-0.022
0.091
0.091
2.1
-0.308
0.052
0.169
2.2
-0.549
-0.039 -0.013
2.3
-4.846
-0.13 -0.247
2.4
-4.923
0.312
0.247
2.5
-0.121
0.039
0.026
3.1
-0.279
-0.011
0.034
3.2
-0.385
-0.114 -0.091
3.3
-0.087
-0.148 -0.295
3.4
-0.971
0.17
0.182
3.5
-0.923
-0.841 -0.591
4.1
-0.144
0.25
0.182
4.2
-0.317
0.125
0.068
4.3
0.029
-0.034 -0.182
4.4
-4.837
0.295
0.125
4.5
-0.889
-0.301 -0.455
5.1
-0.333
-0.03
0.015
5.2
0.103
-0.136
-0.03
5.3
-0.256
-0.197 -0.152
5.4
-5.372
-0.485 -0.515
5.5
-5.122
0.144
0.038
6.1
-1.814
-0.163 -0.076
ALL Average:
STDEV
-1.340
1.76
-0.085
0.25
-0.090
0.22
RANGE
RND 2 RND 3
7
5
6
4
8
7
8
3
5
2
7
6
8
5
7
3
7
4
5
3
5
2
3
6
3
3
5
2
7
4
6
3
6
6
6
2
8
7
6
4
6
4
6
6
4
3
6
4
6
2
5
3
5
5
4
4
4
3
5
2
5.8
3.9
334
RND 1
5
51%
3
64%
6
60%
3
70%
1
90%
6
57%
4
85%
2
83%
4
75%
2
91%
1
95%
5
65%
2
64%
2
88%
4
90%
3
88%
5
84%
2
83%
6
78%
4
87%
4
89%
6
92%
2
70%
4
68%
2
85%
4
92%
5
91%
4
64%
2
65%
2
77%
3.5
22
78.0%
Ave % response
LEANING RND 2 LEANING RND3
>5
64% >5
67%
<5
66% <5
71%
<5
51% <5
52%
>5
99% >5
100%
>5
100% >5
100%
<5
58% >5
58%
>5
91% >5
91%
>5
93% >5
95%
>5
92% >5
94%
>5
100% >5
100%
>5
100% >5
100%
<5
90% >5
90%
<5
99% >5
100%
>5
100% >5
100%
>5
99% >5
98%
>5
98% >5
99%
>5
91% >5
91%
>5
99% >5
99%
>5
84% >5
88%
>5
99% >5
99%
>5
99% >5
99%
>5
91% >5
91%
<5
99% >5
99%
>5
90% >5
86%
>5
98% >5
100%
>5
95% >5
97%
>5
94% >5
97%
<5
91% >5
89%
<5
99% >5
99%
>5
100% >5
100%
91.0%
91.6%
LEANING
>5
<5
>5
>5
>5
>5
>5
>5
>5
>5
>5
>5
>5
>5
>5
>5
>5
>5
>5
>5
>5
>5
>5
>5
>5
>5
>5
>5
>5
>5
APPENDIX H
MODEL 2.0 COVARIANCE MATRIXES
Table 57: Covariances - project performance variables on horizontal
CLCP1
CLCP2
CLCP3
CLCP4
PP1
PP2
PP3
PP4
PP5
PP6
PP7
PP8
PP9
PP10
PP11
PP12
PP13
PP14
PP15
PP16
PP17
CCLP_1
CCLP_2
CCLP_3
CLIP_1
CLIP_2
CLIP_3
CLPP_1
CLPP_2
CLPP_3
CLNP_1
CLNP_2
CLNP_3
CLTP_1
CLTP_2
CLTP_3
COPP_1
COPP_2
COPP_3
COTP_1
COTP_2
COTP_3
CONP_1
CONP_2
CONP_3
DEPP_1
DEPP_2
DEPP_3
DETP_1
DETP_2
DETP_3
DENP_1
DENP_2
DENP_3
CLCP1
CLCP2
CLCP3
CLCP4
PP1
0.875
0.781
0.726
0.736
0.425
0.457
0.510
0.488
0.479
0.407
0.425
0.400
0.404
0.474
0.437
0.466
0.338
0.404
0.400
0.434
0.402
0.457
0.474
0.483
0.502
0.476
0.512
0.529
0.531
0.492
0.534
0.529
0.528
0.500
0.435
0.505
0.433
0.441
0.417
0.354
0.370
0.385
0.428
0.402
0.395
0.348
0.366
0.307
0.382
0.313
0.367
0.374
0.364
0.416
1.213
0.917
0.973
0.540
0.581
0.630
0.540
0.577
0.478
0.583
0.491
0.537
0.626
0.565
0.566
0.390
0.502
0.518
0.544
0.554
0.507
0.546
0.553
0.667
0.633
0.700
0.700
0.701
0.650
0.675
0.688
0.668
0.574
0.559
0.654
0.530
0.545
0.530
0.401
0.443
0.471
0.501
0.479
0.489
0.412
0.457
0.474
0.486
0.433
0.469
0.510
0.453
0.490
1.157
0.960
0.555
0.608
0.629
0.578
0.589
0.523
0.586
0.573
0.510
0.568
0.603
0.582
0.420
0.469
0.537
0.569
0.586
0.545
0.567
0.532
0.679
0.640
0.711
0.677
0.695
0.625
0.679
0.715
0.685
0.614
0.598
0.701
0.505
0.544
0.503
0.383
0.421
0.442
0.469
0.449
0.443
0.494
0.470
0.463
0.500
0.430
0.487
0.547
0.489
0.531
1.304
0.636
0.623
0.680
0.590
0.589
0.572
0.687
0.575
0.602
0.686
0.766
0.641
0.373
0.522
0.617
0.659
0.630
0.564
0.567
0.557
0.681
0.630
0.707
0.712
0.739
0.663
0.712
0.742
0.708
0.612
0.611
0.683
0.599
0.584
0.573
0.461
0.508
0.500
0.523
0.526
0.510
0.473
0.472
0.509
0.504
0.423
0.486
0.531
0.482
0.541
0.915
0.722
0.713
0.679
0.706
0.567
0.710
0.602
0.590
0.640
0.709
0.560
0.563
0.590
0.689
0.665
0.675
0.465
0.508
0.477
0.424
0.463
0.496
0.569
0.554
0.502
0.550
0.563
0.564
0.496
0.451
0.506
0.633
0.653
0.564
0.521
0.548
0.539
0.563
0.565
0.563
0.502
0.489
0.487
0.480
0.462
0.473
0.537
0.487
0.509
335
Table 57: (Continued)
PP2
PP3
PP4
PP5
PP6
PP7
PP8
PP9
PP10
PP11
PP12
PP13
PP14
PP15
PP16
PP17
CCLP_1
CCLP_2
CCLP_3
CLIP_1
CLIP_2
CLIP_3
CLPP_1
CLPP_2
CLPP_3
CLNP_1
CLNP_2
CLNP_3
CLTP_1
CLTP_2
CLTP_3
COPP_1
COPP_2
COPP_3
COTP_1
COTP_2
COTP_3
CONP_1
CONP_2
CONP_3
DEPP_1
DEPP_2
DEPP_3
DETP_1
DETP_2
DETP_3
DENP_1
DENP_2
DENP_3
PP2
PP3
PP4
PP5
PP6
0.965
0.738
0.666
0.728
0.563
0.738
0.600
0.556
0.698
0.736
0.632
0.559
0.598
0.730
0.763
0.700
0.498
0.537
0.548
0.493
0.492
0.506
0.556
0.553
0.501
0.598
0.552
0.566
0.530
0.509
0.520
0.670
0.694
0.618
0.554
0.555
0.572
0.581
0.571
0.556
0.465
0.527
0.522
0.481
0.437
0.480
0.534
0.516
0.496
0.942
0.684
0.702
0.565
0.735
0.589
0.576
0.674
0.741
0.650
0.550
0.634
0.648
0.665
0.694
0.464
0.508
0.505
0.522
0.519
0.556
0.575
0.559
0.544
0.592
0.603
0.591
0.482
0.460
0.530
0.652
0.715
0.621
0.535
0.565
0.573
0.614
0.626
0.611
0.532
0.530
0.516
0.492
0.433
0.507
0.555
0.528
0.542
0.990
0.680
0.613
0.642
0.492
0.708
0.583
0.762
0.646
0.569
0.582
0.662
0.694
0.717
0.464
0.516
0.458
0.458
0.528
0.534
0.566
0.582
0.533
0.596
0.591
0.599
0.451
0.469
0.523
0.608
0.619
0.547
0.473
0.545
0.533
0.557
0.551
0.520
0.473
0.531
0.532
0.487
0.465
0.474
0.582
0.537
0.606
1.028
0.651
0.642
0.543
0.570
0.695
0.780
0.662
0.583
0.589
0.667
0.683
0.675
0.481
0.552
0.527
0.501
0.488
0.497
0.599
0.536
0.541
0.591
0.559
0.572
0.478
0.480
0.547
0.640
0.715
0.576
0.536
0.579
0.588
0.601
0.575
0.578
0.494
0.519
0.510
0.500
0.484
0.516
0.597
0.551
0.535
0.928
0.609
0.554
0.711
0.572
0.656
0.521
0.428
0.487
0.519
0.602
0.553
0.454
0.489
0.442
0.428
0.419
0.424
0.529
0.497
0.520
0.527
0.557
0.552
0.438
0.432
0.470
0.499
0.530
0.492
0.402
0.448
0.460
0.479
0.480
0.478
0.453
0.484
0.477
0.444
0.385
0.447
0.556
0.513
0.557
336
Table 57: (Continued)
PP7
PP8
PP9
PP10
PP11
PP12
PP13
PP14
PP15
PP16
PP17
CCLP_1
CCLP_2
CCLP_3
CLIP_1
CLIP_2
CLIP_3
CLPP_1
CLPP_2
CLPP_3
CLNP_1
CLNP_2
CLNP_3
CLTP_1
CLTP_2
CLTP_3
COPP_1
COPP_2
COPP_3
COTP_1
COTP_2
COTP_3
CONP_1
CONP_2
CONP_3
DEPP_1
DEPP_2
DEPP_3
DETP_1
DETP_2
DETP_3
DENP_1
DENP_2
DENP_3
PP7
PP8
PP9
1.042
0.713
0.699
0.691
0.860
0.638
0.516
0.635
0.730
0.788
0.771
0.464
0.503
0.480
0.453
0.482
0.500
0.572
0.573
0.547
0.576
0.582
0.581
0.452
0.450
0.481
0.679
0.717
0.663
0.546
0.593
0.629
0.637
0.644
0.627
0.491
0.526
0.544
0.483
0.424
0.475
0.547
0.509
0.540
0.974
0.562
0.583
0.604
0.530
0.448
0.489
0.546
0.573
0.558
0.421
0.444
0.402
0.384
0.378
0.396
0.470
0.449
0.452
0.449
0.448
0.474
0.432
0.403
0.447
0.527
0.544
0.575
0.413
0.448
0.475
0.468
0.466
0.458
0.376
0.406
0.436
0.458
0.399
0.459
0.502
0.482
0.484
1.347
0.611
0.791
0.567
0.388
0.534
0.504
0.655
0.720
0.499
0.501
0.506
0.427
0.520
0.455
0.626
0.647
0.608
0.561
0.560
0.593
0.433
0.360
0.476
0.516
0.535
0.538
0.410
0.458
0.453
0.516
0.526
0.520
0.620
0.666
0.604
0.541
0.487
0.536
0.626
0.599
0.619
337
PP10
PP11
1.075
0.852
0.615
0.542
0.609
0.740
0.665
0.701
0.426
0.457
0.458
0.416
0.366
0.458
0.568
0.569
0.524
0.553
0.530
0.570
0.406
0.436
0.484
0.635
0.633
0.589
0.510
0.536
0.528
0.559
0.559
0.565
0.442
0.513
0.524
0.484
0.438
0.465
0.551
0.519
0.536
1.388
0.744
0.537
0.687
0.833
0.956
0.993
0.584
0.595
0.578
0.560
0.581
0.603
0.719
0.670
0.684
0.718
0.715
0.696
0.513
0.534
0.608
0.759
0.795
0.730
0.594
0.697
0.665
0.681
0.682
0.676
0.616
0.598
0.637
0.552
0.507
0.536
0.653
0.590
0.649
Table 57: (Continued)
PP12
PP13
PP14
PP15
PP16
PP17
CCLP_1
CCLP_2
CCLP_3
CLIP_1
CLIP_2
CLIP_3
CLPP_1
CLPP_2
CLPP_3
CLNP_1
CLNP_2
CLNP_3
CLTP_1
CLTP_2
CLTP_3
COPP_1
COPP_2
COPP_3
COTP_1
COTP_2
COTP_3
CONP_1
CONP_2
CONP_3
DEPP_1
DEPP_2
DEPP_3
DETP_1
DETP_2
DETP_3
DENP_1
DENP_2
DENP_3
PP12
PP13
PP14
PP15
0.981
0.470
0.574
0.618
0.687
0.652
0.456
0.497
0.470
0.447
0.455
0.510
0.540
0.539
0.508
0.530
0.496
0.503
0.445
0.448
0.502
0.571
0.595
0.556
0.463
0.506
0.520
0.528
0.515
0.519
0.515
0.488
0.523
0.480
0.428
0.507
0.516
0.502
0.472
0.949
0.682
0.699
0.616
0.525
0.340
0.349
0.312
0.295
0.324
0.331
0.348
0.347
0.318
0.326
0.337
0.338
0.307
0.326
0.314
0.524
0.567
0.495
0.424
0.429
0.442
0.460
0.448
0.429
0.326
0.357
0.353
0.342
0.331
0.342
0.380
0.377
0.372
0.993
0.800
0.799
0.666
0.393
0.424
0.391
0.409
0.420
0.483
0.482
0.470
0.441
0.439
0.448
0.447
0.400
0.375
0.414
0.602
0.632
0.560
0.460
0.466
0.491
0.540
0.537
0.522
0.401
0.458
0.456
0.409
0.383
0.427
0.478
0.462
0.484
1.204
0.930
0.797
0.460
0.456
0.400
0.431
0.440
0.486
0.555
0.534
0.533
0.481
0.463
0.483
0.429
0.450
0.494
0.752
0.769
0.684
0.567
0.586
0.610
0.611
0.608
0.554
0.457
0.481
0.507
0.415
0.380
0.424
0.481
0.427
0.444
338
PP16
1.273
0.911
0.519
0.522
0.508
0.506
0.500
0.561
0.614
0.566
0.587
0.588
0.550
0.552
0.491
0.470
0.524
0.795
0.820
0.740
0.581
0.602
0.633
0.662
0.670
0.633
0.480
0.515
0.541
0.491
0.402
0.477
0.550
0.499
0.552
Table 58: Covariances – Client H&S culture variables on horizontal
PP17
PP17
CCLP_1
CCLP_2
CCLP_3
CLIP_1
CLIP_2
CLIP_3
CLPP_1
CLPP_2
CLPP_3
CLNP_1
CLNP_2
CLNP_3
CLTP_1
CLTP_2
CLTP_3
COPP_1
COPP_2
COPP_3
COTP_1
COTP_2
COTP_3
CONP_1
CONP_2
CONP_3
DEPP_1
DEPP_2
DEPP_3
DETP_1
DETP_2
DETP_3
DENP_1
DENP_2
DENP_3
1.279
0.533
0.556
0.544
0.454
0.463
0.544
0.625
0.608
0.608
0.590
0.583
0.570
0.461
0.438
0.516
0.763
0.818
0.697
0.576
0.642
0.639
0.680
0.695
0.668
0.502
0.556
0.566
0.513
0.451
0.516
0.592
0.523
0.627
CCLP_1
CCLP_2
CCLP_3
CLIP_1
0.707
0.631
0.575
0.449
0.406
0.455
0.551
0.540
0.516
0.584
0.566
0.556
0.422
0.408
0.467
0.473
0.500
0.460
0.381
0.399
0.404
0.424
0.425
0.411
0.445
0.427
0.405
0.448
0.374
0.414
0.475
0.433
0.465
0.780
0.644
0.514
0.533
0.505
0.573
0.558
0.556
0.609
0.618
0.614
0.480
0.469
0.521
0.471
0.531
0.468
0.384
0.415
0.432
0.440
0.430
0.441
0.465
0.470
0.461
0.508
0.460
0.475
0.546
0.507
0.539
0.801
0.509
0.496
0.458
0.569
0.555
0.545
0.575
0.559
0.542
0.454
0.426
0.507
0.457
0.499
0.444
0.355
0.396
0.417
0.427
0.434
0.438
0.459
0.460
0.458
0.468
0.439
0.440
0.476
0.436
0.489
0.978
0.925
0.882
0.570
0.605
0.564
0.633
0.636
0.603
0.721
0.731
0.781
0.400
0.439
0.390
0.288
0.340
0.349
0.349
0.349
0.354
0.464
0.455
0.479
0.422
0.373
0.414
0.470
0.407
0.467
339
Table 58: (Continued)
CLIP_2
CLIP_3
CLPP_1
CLPP_2
CLPP_3
CLNP_1
CLNP_2
CLNP_3
CLTP_1
CLTP_2
CLTP_3
COPP_1
COPP_2
COPP_3
COTP_1
COTP_2
COTP_3
CONP_1
CONP_2
CONP_3
DEPP_1
DEPP_2
DEPP_3
DETP_1
DETP_2
DETP_3
DENP_1
DENP_2
DENP_3
CLIP_2
CLIP_3
CLPP_1
CLPP_2
CLPP_3
2.687
0.902
0.563
0.585
0.552
0.596
0.582
0.592
0.680
0.740
0.751
0.377
0.421
0.391
0.310
0.369
0.362
0.357
0.344
0.357
0.448
0.482
0.492
0.427
0.420
0.434
0.477
0.408
0.470
1.086
0.586
0.645
0.541
0.637
0.662
0.648
0.752
0.763
0.816
0.422
0.471
0.428
0.352
0.394
0.390
0.389
0.399
0.387
0.534
0.502
0.496
0.471
0.429
0.457
0.541
0.481
0.545
0.853
0.747
0.745
0.686
0.668
0.689
0.507
0.454
0.545
0.528
0.557
0.501
0.415
0.452
0.453
0.496
0.507
0.512
0.621
0.586
0.600
0.543
0.466
0.531
0.580
0.527
0.572
0.804
0.673
0.665
0.663
0.672
0.561
0.519
0.598
0.510
0.521
0.493
0.430
0.455
0.455
0.495
0.506
0.499
0.606
0.569
0.582
0.519
0.457
0.510
0.575
0.528
0.575
0.863
0.647
0.634
0.631
0.454
0.422
0.518
0.515
0.529
0.503
0.360
0.417
0.413
0.458
0.461
0.475
0.590
0.524
0.547
0.473
0.410
0.461
0.508
0.456
0.506
CLNP_1
CLNP_2
CLNP_3
CLTP_1
CLTP_2
0.953
0.852
0.823
0.529
0.527
0.616
0.516
0.538
0.479
0.402
0.443
0.447
0.500
0.492
0.487
0.610
0.549
0.580
0.540
0.488
0.516
0.614
0.555
0.603
0.934
0.837
0.534
0.538
0.645
0.480
0.519
0.470
0.389
0.419
0.430
0.478
0.477
0.467
0.607
0.548
0.567
0.544
0.504
0.523
0.635
0.579
0.624
0.897
0.538
0.512
0.587
0.487
0.510
0.487
0.407
0.438
0.435
0.473
0.475
0.456
0.586
0.549
0.561
0.533
0.490
0.537
0.622
0.581
0.626
1.490
0.737
0.756
0.403
0.431
0.385
0.323
0.359
0.365
0.356
0.367
0.363
0.439
0.364
0.368
0.389
0.345
0.377
0.430
0.404
0.437
0.830
0.759
0.406
0.433
0.389
0.342
0.370
0.386
0.357
0.347
0.334
0.375
0.366
0.370
0.371
0.349
0.351
0.404
0.367
0.416
Table 58: (Continued)
CLNP_1
CLNP_2
CLNP_3
CLTP_1
CLTP_2
CLTP_3
COPP_1
COPP_2
COPP_3
COTP_1
COTP_2
COTP_3
CONP_1
CONP_2
CONP_3
DEPP_1
DEPP_2
DEPP_3
DETP_1
DETP_2
DETP_3
DENP_1
DENP_2
DENP_3
340
Table 59: Covariances – Contractor performance variables on horizontal
CLTP_3
COPP_1
COPP_2
COPP_3
COTP_1
COTP_2
COTP_3
CONP_1
CONP_2
CONP_3
DEPP_1
DEPP_2
DEPP_3
DETP_1
DETP_2
DETP_3
DENP_1
DENP_2
DENP_3
CLTP_3
COPP_1
COPP_2
COPP_3
COTP_1
0.937
0.441
0.497
0.444
0.348
0.388
0.415
0.413
0.397
0.405
0.449
0.447
0.466
0.462
0.423
0.414
0.506
0.427
0.481
0.848
0.830
0.756
0.609
0.627
0.673
0.703
0.685
0.654
0.439
0.459
0.461
0.425
0.363
0.404
0.474
0.433
0.451
0.954
0.788
0.651
0.671
0.717
0.751
0.733
0.702
0.459
0.498
0.489
0.472
0.418
0.451
0.535
0.492
0.491
0.828
0.593
0.607
0.647
0.674
0.658
0.626
0.407
0.454
0.446
0.425
0.373
0.411
0.476
0.457
0.449
0.661
0.624
0.637
0.620
0.599
0.564
0.337
0.374
0.363
0.354
0.317
0.335
0.393
0.384
0.356
COTP_2
COTP_3
CONP_1
CONP_2
CONP_3
0.753
0.695
0.645
0.621
0.598
0.351
0.388
0.376
0.367
0.336
0.362
0.432
0.392
0.380
0.778
0.688
0.669
0.634
0.339
0.407
0.399
0.393
0.360
0.377
0.452
0.399
0.410
0.785
0.735
0.709
0.397
0.443
0.438
0.435
0.384
0.414
0.479
0.444
0.457
0.785
0.707
0.418
0.456
0.464
0.447
0.389
0.437
0.491
0.449
0.468
0.772
0.421
0.462
0.464
0.456
0.391
0.444
0.496
0.465
0.467
Table 59: (Continued)
COTP_2
COTP_3
CONP_1
CONP_2
CONP_3
DEPP_1
DEPP_2
DEPP_3
DETP_1
DETP_2
DETP_3
DENP_1
DENP_2
DENP_3
341
Table 60: Covariances – Designer performance variables on horizontal
DEPP_1
DEPP_2
DEPP_3
DETP_1
DETP_2
DETP_3
DENP_1
DENP_2
DENP_3
DEPP_1
DEPP_2
DEPP_3
DETP_1
DETP_2
1.619
0.777
0.795
0.619
0.550
0.586
0.714
0.661
0.643
0.821
0.778
0.599
0.541
0.583
0.688
0.643
0.636
0.903
0.620
0.578
0.596
0.708
0.643
0.668
0.784
0.660
0.668
0.715
0.671
0.670
0.743
0.661
0.724
0.662
0.633
DETP_3
DENP_1
DENP_2
DENP_3
0.782
0.754
0.714
0.667
0.981
0.883
0.874
0.938
0.814
1.001
Table 60: (Continued)
DETP_3
DENP_1
DENP_2
DENP_3
342
APPENDIX J
MODEL 1.0 COVARIANCE MATRIXES
Table 61: Covariances – Environmental influence on client H&S culture
2.2.L1
2.2.L2
V1
V2
2.2.L1 V1
1.147
2.2.L2 V2
0.715
1.042
2.2.L3 V3
0.603
0.708
2.2.L4 V4
0.628
0.682
2.2.L5 V5
0.596
0.627
2.2.L6 V6
0.556
0.572
2.2.L7 V7
0.466
0.502
2.2.L8 V8
0.599
0.673
2.3.CT1 V9
0.489
0.493
2.3.CT2 V10
0.495
0.568
2.3.CT3 V11
0.389
0.428
2.3.CT4 V12
0.382
0.497
2.3.CT5 V13
0.484
0.544
2.3.CT6 V14
0.540
0.635
2.3.CT7 V15
0.501
0.593
2.3.CT8 V16
0.458
0.473
2.3.CT9 V17
0.617
0.615
2.4.I1 V18
0.422
0.495
2.4.I2 V19
0.235
0.486
2.4.I3 V20
0.527
0.565
2.4.I4 V21
0.560
0.585
2.4.I5 V22
0.508
0.610
2.4.I6 V23
0.476
0.611
2.4.I7 V24
0.517
0.581
2.5.CN1 V25
0.648
0.726
2.5.CN2 V26
0.665
0.690
2.5.CN3 V27
0.548
0.673
2.5.CN4 V28
0.676
0.746
2.5.CN5 V29
0.650
0.692
2.5.CN6 V30
0.516
0.663
2.5.CN7 V31
0.621
0.696
2.5.CN8 V32
0.583
0.640
2.5.CN9 V33
0.666
0.722
2.5.CE1 V34
0.532
0.554
2.6.CE2 V35
0.621
0.630
2.6.CE3 V36
0.649
0.634
2.6.CE4 V37
0.653
0.595
2.7.P1 V38
0.627
0.707
2.7.P2 V39
0.627
0.699
2.7.P3 V40
0.630
0.618
2.7.P4 V41
0.554
0.598
2.7.P5 V42
0.647
0.716
2.7.P6 V43
0.452
0.618
2.7.P7 V44
0.492
0.651
2.7.P8 V45
0.474
0.539
2.14.E1 V46
0.594
0.522
2.14.E2 V47
0.406
0.435
2.14.E3 V48
0.395
0.386
2.14.E4 V49
0.502
0.466
2.14.E5 V50
0.448
0.394
2.14.E6 V51
0.484
0.462
Variables V1-V51= Questionnaire items
343
2.2.L3
V3
2.2.L4
V4
0.985
0.702
0.772
0.585
0.509
0.635
0.491
0.465
0.398
0.471
0.554
0.686
0.611
0.487
0.646
0.509
0.406
0.490
0.542
0.612
0.584
0.616
0.694
0.657
0.640
0.663
0.679
0.576
0.591
0.596
0.628
0.556
0.588
0.616
0.579
0.645
0.619
0.601
0.586
0.701
0.593
0.663
0.434
0.538
0.404
0.350
0.370
0.327
0.429
1.069
0.689
0.659
0.514
0.736
0.449
0.460
0.376
0.430
0.474
0.534
0.654
0.385
0.621
0.500
0.323
0.510
0.524
0.545
0.529
0.534
0.697
0.645
0.714
0.709
0.693
0.632
0.693
0.661
0.721
0.546
0.565
0.608
0.639
0.632
0.654
0.687
0.665
0.680
0.585
0.664
0.618
0.588
0.558
0.476
0.587
0.462
0.582
2.2.L5
V5
0.909
0.601
0.499
0.622
0.459
0.481
0.337
0.479
0.475
0.639
0.548
0.466
0.567
0.455
0.498
0.507
0.563
0.609
0.561
0.592
0.662
0.641
0.598
0.652
0.657
0.595
0.537
0.608
0.657
0.523
0.550
0.627
0.598
0.620
0.601
0.563
0.523
0.647
0.551
0.612
0.423
0.608
0.454
0.381
0.405
0.382
0.455
Table 61: (continued)
2.2.L6
2.2.L7
V6
V7
2.2.L6 V6
1.059
2.2.L7 V7
0.397
0.758
2.2.L8 V8
0.735
0.526
2.3.CT1 V9
0.388
0.268
2.3.CT2 V10
0.354
0.353
2.3.CT3 V11
0.354
0.265
2.3.CT4 V12
0.396
0.294
2.3.CT5 V13
0.430
0.346
2.3.CT6 V14
0.536
0.416
2.3.CT7 V15
0.478
0.441
2.3.CT8 V16
0.372
0.276
2.3.CT9 V17
0.517
0.409
2.4.I1 V18
0.492
0.341
2.4.I2 V19
0.539
0.309
2.4.I3 V20
0.454
0.295
2.4.I4 V21
0.481
0.307
2.4.I5 V22
0.513
0.392
2.4.I6 V23
0.433
0.368
2.4.I7 V24
0.447
0.407
2.5.CN1 V25
0.607
0.466
2.5.CN2 V26
0.558
0.484
2.5.CN3 V27
0.522
0.482
2.5.CN4 V28
0.555
0.494
2.5.CN5 V29
0.580
0.472
2.5.CN6 V30
0.487
0.439
2.5.CN7 V31
0.505
0.388
2.5.CN8 V32
0.491
0.406
2.5.CN9 V33
0.593
0.441
2.5.CE1 V34
0.478
0.416
2.6.CE2 V35
0.545
0.419
2.6.CE3 V36
0.538
0.482
2.6.CE4 V37
0.574
0.460
2.7.P1 V38
0.562
0.514
2.7.P2 V39
0.577
0.487
2.7.P3 V40
0.571
0.364
2.7.P4 V41
0.516
0.343
2.7.P5 V42
0.593
0.472
2.7.P6 V43
0.511
0.463
2.7.P7 V44
0.540
0.516
2.7.P8 V45
0.518
0.348
2.14.E1 V46
0.504
0.378
2.14.E2 V47
0.508
0.401
2.14.E3 V48
0.522
0.342
2.14.E4 V49
0.482
0.406
2.14.E5 V50
0.359
0.306
2.14.E6 V51
0.432
0.370
Variables V1-V51= Questionnaire items
344
2.2.L8
V8
1.179
0.432
0.464
0.405
0.476
0.474
0.585
0.502
0.363
0.570
0.618
0.593
0.455
0.469
0.495
0.466
0.437
0.638
0.679
0.663
0.633
0.643
0.552
0.566
0.636
0.612
0.462
0.548
0.557
0.584
0.615
0.578
0.582
0.534
0.630
0.661
0.685
0.555
0.504
0.668
0.628
0.596
0.450
0.544
2.3.CT1
V9
2.3.CT2
V10
7.735
0.892
0.748
0.761
0.786
0.833
0.728
0.775
0.810
0.750
0.695
0.846
0.823
0.830
0.867
0.799
0.640
0.589
0.536
0.523
0.597
0.449
0.434
0.498
0.519
0.569
0.620
0.648
0.621
0.720
0.633
0.472
0.393
0.548
0.419
0.476
0.626
0.474
0.428
0.279
0.402
0.314
0.348
1.104
0.652
0.699
0.774
0.865
0.736
0.735
0.801
0.695
0.788
0.805
0.755
0.856
0.836
0.836
0.612
0.626
0.543
0.611
0.622
0.510
0.489
0.486
0.549
0.454
0.592
0.600
0.626
0.533
0.514
0.440
0.377
0.527
0.418
0.522
0.574
0.457
0.444
0.296
0.408
0.329
0.306
Table 61: (continued)
2.3.CT3
2.3.CT4
V11
V12
2.3.CT3 V11
1.230
2.3.CT4 V12
0.733
0.957
2.3.CT5 V13
0.778
0.774
2.3.CT6 V14
0.705
0.768
2.3.CT7 V15
0.665
0.680
2.3.CT8 V16
0.647
0.672
2.3.CT9 V17
0.721
0.727
2.4.I1 V18
0.666
0.710
2.4.I2 V19
0.484
0.629
2.4.I3 V20
0.775
0.705
2.4.I4 V21
0.717
0.717
2.4.I5 V22
0.730
0.748
2.4.I6 V23
0.718
0.753
2.4.I7 V24
0.671
0.651
2.5.CN1 V25
0.579
0.538
2.5.CN2 V26
0.483
0.524
2.5.CN3 V27
0.404
0.494
2.5.CN4 V28
0.421
0.428
2.5.CN5 V29
0.609
0.522
2.5.CN6 V30
0.388
0.475
2.5.CN7 V31
0.367
0.414
2.5.CN8 V32
0.423
0.467
2.5.CN9 V33
0.456
0.505
2.5.CE1 V34
0.539
0.462
2.6.CE2 V35
0.672
0.561
2.6.CE3 V36
0.693
0.568
2.6.CE4 V37
0.639
0.558
2.7.P1 V38
0.483
0.465
2.7.P2 V39
0.396
0.472
2.7.P3 V40
0.355
0.426
2.7.P4 V41
0.313
0.405
2.7.P5 V42
0.548
0.495
2.7.P6 V43
0.473
0.338
2.7.P7 V44
0.493
0.497
2.7.P8 V45
0.549
0.533
2.14.E1 V46
0.423
0.416
2.14.E2 V47
0.373
0.383
2.14.E3 V48
0.250
0.192
2.14.E4 V49
0.294
0.347
2.14.E5 V50
0.224
0.182
2.14.E6 V51
0.260
0.275
Variables V1-V51= Questionnaire items
345
2.3.CT5
V13
1.119
0.973
0.893
0.788
0.900
0.812
0.761
0.868
0.877
0.880
0.849
0.915
0.674
0.665
0.600
0.605
0.694
0.532
0.512
0.510
0.569
0.479
0.658
0.700
0.692
0.614
0.556
0.519
0.451
0.660
0.522
0.603
0.620
0.516
0.487
0.398
0.414
0.332
0.357
2.3.CT6
V14
1.338
0.981
0.789
0.973
0.902
0.854
0.891
0.892
0.978
0.898
1.014
0.802
0.756
0.680
0.741
0.840
0.659
0.647
0.629
0.680
0.545
0.707
0.740
0.699
0.702
0.623
0.564
0.495
0.729
0.570
0.713
0.635
0.525
0.515
0.431
0.486
0.413
0.399
2.3.CT7
V15
1.338
0.755
0.974
0.783
0.596
0.823
0.842
0.903
0.856
0.936
0.722
0.661
0.661
0.759
0.735
0.693
0.640
0.568
0.683
0.582
0.642
0.740
0.741
0.708
0.642
0.617
0.580
0.743
0.677
0.741
0.716
0.535
0.473
0.365
0.434
0.386
0.481
Table 61: (continued)
2.3.CT8
2.3.CT9
2.4.I1
2.4.I2
2.4.I3
2.4.I4
2.4.I5
2.4.I6
2.4.I7
2.5.CN1
2.5.CN2
2.5.CN3
2.5.CN4
2.5.CN5
2.5.CN6
2.5.CN7
2.5.CN8
2.5.CN9
2.5.CE1
2.6.CE2
2.6.CE3
2.6.CE4
2.7.P1
2.7.P2
2.7.P3
2.7.P4
2.7.P5
2.7.P6
2.7.P7
2.7.P8
2.14.E1
2.14.E2
2.14.E3
2.14.E4
2.14.E5
2.14.E6
V16
V17
V18
V19
V20
V21
V22
V23
V24
V25
V26
V27
V28
V29
V30
V31
V32
V33
V34
V35
V36
V37
V38
V39
V40
V41
V42
V43
V44
V45
V46
V47
V48
V49
V50
V51
2.3.CT8
V16
1.050
0.855
0.611
0.671
0.759
0.760
0.851
0.804
0.809
0.593
0.514
0.451
0.505
0.524
0.441
0.383
0.435
0.465
0.488
0.561
0.626
0.602
0.518
0.448
0.426
0.405
0.577
0.380
0.482
0.556
0.432
0.279
0.264
0.222
0.299
0.261
2.3.CT9
V17
2.4.I1
V18
2.4.I2
V19
2.4.I3
V20
1.415
0.832
0.794
0.992
0.885
0.986
0.900
0.939
0.851
0.808
0.719
0.816
0.887
0.716
0.668
0.691
0.782
0.585
0.748
0.834
0.841
0.770
0.722
0.731
0.592
0.792
0.559
0.692
0.690
0.549
0.466
0.376
0.559
0.454
0.527
1.322
1.048
0.828
0.862
0.838
0.808
0.821
0.754
0.620
0.580
0.613
0.689
0.621
0.571
0.588
0.620
0.455
0.640
0.673
0.686
0.574
0.562
0.547
0.488
0.632
0.679
0.684
0.661
0.454
0.604
0.445
0.499
0.282
0.496
8.381
0.783
0.955
0.961
0.927
0.884
0.702
0.637
0.604
0.638
0.544
0.612
0.426
0.479
0.553
0.436
0.630
0.648
0.609
0.615
0.512
0.537
0.414
0.488
0.631
0.643
0.715
0.385
0.590
0.307
0.488
0.272
0.385
1.373
1.000
0.969
0.901
0.911
0.779
0.741
0.643
0.671
0.798
0.615
0.524
0.549
0.695
0.474
0.723
0.786
0.742
0.669
0.650
0.550
0.431
0.703
0.482
0.668
0.674
0.511
0.536
0.362
0.481
0.372
0.435
Variables V1-V51= Questionnaire items
346
Table 61: (continued)
2.4.I4
2.4.I5
V21
V22
2.4.I4 V21
1.219
2.4.I5 V22
1.001
1.245
2.4.I6 V23
0.943
1.058
2.4.I7 V24
0.900
1.037
2.5.CN1 V25
0.733
0.797
2.5.CN2 V26
0.692
0.716
2.5.CN3 V27
0.636
0.662
2.5.CN4 V28
0.661
0.725
2.5.CN5 V29
0.745
0.788
2.5.CN6 V30
0.583
0.607
2.5.CN7 V31
0.553
0.559
2.5.CN8 V32
0.592
0.628
2.5.CN9 V33
0.685
0.707
2.5.CE1 V34
0.523
0.602
2.6.CE2 V35
0.689
0.745
2.6.CE3 V36
0.717
0.750
2.6.CE4 V37
0.729
0.728
2.7.P1 V38
0.667
0.691
2.7.P2 V39
0.610
0.634
2.7.P3 V40
0.588
0.580
2.7.P4 V41
0.472
0.526
2.7.P5 V42
0.719
0.709
2.7.P6 V43
0.503
0.535
2.7.P7 V44
0.663
0.705
2.7.P8 V45
0.698
0.704
2.14.E1 V46
0.559
0.536
2.14.E2 V47
0.483
0.459
2.14.E3 V48
0.367
0.353
2.14.E4 V49
0.423
0.439
2.14.E5 V50
0.335
0.345
2.14.E6 V51
0.450
0.393
Variables V1-V51= Questionnaire items
347
2.4.I6
V23
2.4.I7
V24
2.5.CN1
V25
1.237
0.994
0.745
0.682
0.678
0.694
0.763
0.588
0.588
0.625
0.716
0.592
0.749
0.710
0.722
0.661
0.668
0.610
0.584
0.712
0.594
0.752
0.745
0.508
0.476
0.334
0.459
0.379
0.430
1.344
0.803
0.744
0.664
0.736
0.758
0.655
0.594
0.634
0.705
0.577
0.739
0.757
0.763
0.681
0.593
0.549
0.552
0.729
0.622
0.674
0.614
0.544
0.450
0.391
0.363
0.402
0.421
1.343
0.961
0.825
0.897
0.927
0.838
0.801
0.800
0.887
0.630
0.770
0.804
0.796
0.818
0.787
0.766
0.682
0.911
0.722
0.846
0.613
0.630
0.539
0.407
0.516
0.456
0.519
Table 61: (continued)
2.5.CN2
2.5.CN3
V26
V27
2.5.CN2 V26
1.238
2.5.CN3 V27
0.921
1.030
2.5.CN4 V28
0.892
0.854
2.5.CN5 V29
0.923
0.882
2.5.CN6 V30
0.763
0.724
2.5.CN7 V31
0.727
0.738
2.5.CN8 V32
0.740
0.695
2.5.CN9 V33
0.855
0.795
2.5.CE1 V34
0.569
0.540
2.6.CE2 V35
0.751
0.670
2.6.CE3 V36
0.789
0.678
2.6.CE4 V37
0.800
0.688
2.7.P1 V38
0.748
0.723
2.7.P2 V39
0.751
0.714
2.7.P3 V40
0.729
0.712
2.7.P4 V41
0.643
0.615
2.7.P5 V42
0.822
0.729
2.7.P6 V43
0.613
0.566
2.7.P7 V44
0.770
0.761
2.7.P8 V45
0.559
0.588
2.14.E1 V46
0.573
0.488
2.14.E2 V47
0.514
0.540
2.14.E3 V48
0.431
0.424
2.14.E4 V49
0.588
0.581
2.14.E5 V50
0.499
0.477
2.14.E6 V51
0.530
0.538
Variables V1-V51= Questionnaire items
2.5.CN4
V28
2.5.CN5
V29
2.5.CN6
V30
1.247
1.056
0.863
0.833
0.788
0.879
0.589
0.705
0.716
0.807
0.778
0.730
0.667
0.605
0.809
0.703
0.780
0.542
0.600
0.550
0.507
0.606
0.584
0.586
1.297
0.901
0.881
0.861
0.965
0.598
0.802
0.835
0.842
0.826
0.794
0.734
0.625
0.919
0.693
0.869
0.618
0.632
0.697
0.603
0.778
0.629
0.646
1.072
0.841
0.781
0.880
0.495
0.623
0.683
0.676
0.785
0.722
0.640
0.608
0.835
0.672
0.813
0.558
0.547
0.517
0.443
0.581
0.469
0.553
2.5.CN9
V33
2.5.CE1
V34
2.6.CE2
V35
1.223
0.632
0.851
0.811
0.851
0.822
0.866
0.775
0.747
0.871
0.722
0.860
0.633
0.666
0.619
0.492
0.711
0.520
0.637
0.921
0.820
0.746
0.774
0.637
0.558
0.504
0.527
0.657
0.558
0.589
0.501
0.480
0.396
0.273
0.396
0.326
0.388
1.224
0.924
1.009
0.746
0.738
0.676
0.678
0.829
0.682
0.796
0.646
0.512
0.519
0.410
0.511
0.395
0.444
Table 61: (continued)
2.5.CN7
2.5.CN8
V31
V32
2.5.CN7 V31
1.060
2.5.CN8 V32
0.854
1.029
2.5.CN9 V33
0.917
0.947
2.5.CE1 V34
0.514
0.563
2.6.CE2 V35
0.670
0.689
2.6.CE3 V36
0.668
0.694
2.6.CE4 V37
0.693
0.737
2.7.P1 V38
0.753
0.736
2.7.P2 V39
0.745
0.719
2.7.P3 V40
0.687
0.647
2.7.P4 V41
0.691
0.662
2.7.P5 V42
0.821
0.747
2.7.P6 V43
0.673
0.655
2.7.P7 V44
0.754
0.747
2.7.P8 V45
0.565
0.574
2.14.E1 V46
0.578
0.647
2.14.E2 V47
0.544
0.587
2.14.E3 V48
0.478
0.530
2.14.E4 V49
0.637
0.625
2.14.E5 V50
0.495
0.473
2.14.E6 V51
0.585
0.560
Variables V1-V51= Questionnaire items
348
Table 61: (continued)
2.6.CE3
2.6.CE4
V36
V37
2.6.CE3 V36
1.145
2.6.CE4 V37
0.975
1.338
2.7.P1 V38
0.806
0.783
2.7.P2 V39
0.761
0.818
2.7.P3 V40
0.658
0.650
2.7.P4 V41
0.596
0.666
2.7.P5 V42
0.852
0.842
2.7.P6 V43
0.664
0.754
2.7.P7 V44
0.773
0.819
2.7.P8 V45
0.584
0.675
2.14.E1 V46
0.599
0.588
2.14.E2 V47
0.565
0.618
2.14.E3 V48
0.464
0.516
2.14.E4 V49
0.569
0.629
2.14.E5 V50
0.500
0.471
2.14.E6 V51
0.541
0.589
Variables V1-V51= Questionnaire items
2.7.P1
V38
2.7.P2
V39
2.7.P3
V40
1.145
0.839
0.767
0.698
0.953
0.820
0.901
0.639
0.573
0.518
0.477
0.571
0.499
0.594
1.045
0.738
0.709
0.841
0.699
0.851
0.614
0.543
0.536
0.482
0.624
0.467
0.591
1.123
0.842
0.774
0.618
0.779
0.717
0.554
0.527
0.448
0.545
0.445
0.579
2.7.P5
V42
2.7.P6
V43
2.7.P7
V44
2.7.P8
V45
1.195
0.835
0.975
0.640
0.633
0.531
0.480
0.553
0.527
0.600
1.329
0.941
0.550
0.517
0.604
0.646
0.452
0.447
0.509
1.358
0.783
0.570
0.692
0.597
0.564
0.485
0.640
1.158
0.536
0.530
0.505
0.574
0.365
0.535
Table 61: (continued)
2.7.P4
2.7.P5
2.7.P6
2.7.P7
2.7.P8
2.14.E1
2.14.E2
2.14.E3
2.14.E4
2.14.E5
2.14.E6
V41
V42
V43
V44
V45
V46
V47
V48
V49
V50
V51
2.7.P4
V41
0.976
0.753
0.668
0.719
0.624
0.540
0.439
0.396
0.436
0.381
0.513
Variables V1-V51= Questionnaire items
Table 62: Covariances – Indicator variables for the Environment factor
2.14.E1
2.14.E2
2.14.E3
2.14.E4
2.14.E5
2.14.E6
V46
V47
V48
V49
V50
V51
2.14.E1
V46
1.002
0.615
0.569
0.602
0.551
0.602
2.14.E2
V47
2.14.E3
V48
2.14.E4
V49
1.280
0.944
0.879
0.693
0.731
1.390
0.892
0.886
0.716
1.398
0.887
0.848
Variables V1-V51= Questionnaire items
349
2.14.E5
V50
1.143
0.779
2.14.E6
V51
1.143
TOP EVENT
DIRECT LEVEL
COMPETENCE
MOTIV
/MORALE
TEAM
WORK
SITUATION
AWARENES/
RISK
PERCEPTION
FATIGUE
/ALERTNESS
HEALTH
COMMS
AVAILABILITY
OF
INFO/ADVICE
COMPLIANCE
AVAILABILITY
OF SUITABLE
RESOURCES
INSPECT
& MAINT
EQUIPMENT
OPERABILITY
WORKPLACE
ENVIRONMENT
EXT
CONDITION
ORGANISATIONAL LEVEL
RECRUITMENT
& SELECTION
TRAINING
PROCEDURES
PLANNING
INCIDENT
MANAGEMENT
& FEEDBACK
COMMS
MANAGEMENT
/ SUPERVISION
SAFETY
CULTURE
EQUIPMENT
PURCHASING
INSPECT
& MAINT
DECISION
PAY /
CONDITION
POLICY LEVEL
CONTRACTING
STRATEGY
OWNERSHIP &
CONTROL
COMPANY CULTURE
SAFETY MANAGEMENT
ORGANISATIONAL
STRUCTURE
LABOUR RELATIONS
ENVIRONMENTAL LEVEL
POLITICAL INFLUENCE
REGULATORY INFLUENCE
MARKET INFLUENCE
Figure 4 : Influence Network Technique
(Bomel, 2001: 0.9)
39
SOCIETAL INFLUENCE
COMPANY PROFITABILITY
use the tangible manifestations to test what is underlying, the intangible prevailing
culture. Observing and measuring the tangible elements is the only practical way of
determining the underlying culture because, as Wiegmann et al. (2002:4) argue, while
an organisation’s culture is revealed in its general patterns of attitudes and actions, the
deeper structure of its culture is often not immediately interpretable by outsiders.
Therefore a climate or a perception survey can be conducted to establish the
underlying culture. Figure 7 illustrates and summarises the synthesised framework of
what culture is and where climate fits in.
Culture influencing
factors
Factors of Health
& safety culture






Leadership
Involvement
Procedures
Commitment
Communication
Competence
Underlying culture
Culture indicator
Shared & understood but
tacitly
Tangible/ observable
Climate
Culture



Culture outcomes
Beliefs
Values
Assumptions
Outcomes / results

Perceptions

Artefacts

Actions/
behaviour

Accidents /
incidents
Behaviour
Environment
Figure 7: H&S culture framework
3.3.4
What are the factors of H&S culture?
Measuring H&S culture is one area where confusion has reigned partly because of the
many terms that have been used to describe what constitutes and influences a H&S
culture. Some studies have referred to the parts which form, shape or make up H&S
culture as characteristics (Hudson, 1999:8-1; Molenaar et al., 2002:19; Fitzgerald,
2005:325; Molenaar et al., 2009:488), indicators (Flin, Means, O’Connor & Bryden,
2000:178; Fernandez-Muniz et al., 2007:628), factors (Misnan, Mohammed,
Mahmood, Mahmud, & Abdullah, 2008:1911), determinants (Chinda & Mohamed,
2008:114), elements(Fernandez-Muniz et al., 2007:628), enablers (Chinda &
Mohamed, 2008:114), and attributes (Chinda & Mohamed, 2008:116) of H&S
58
Stage 1
Stage 2
Stage 3
Questionnaire
Develop conceptual
model based on
clients’ influence
Evaluate Clients’
influence on H&S
performance
Delphi
Analyse & model
results to validate
conceptual model
Literature review
Best fit model &
recommendations
Need for study
Figure 8: Research design outline
(Adapted from Manu, Ankrah, Proverbs & Suresh, 2010:29)
5.3.1
Literature review
Literature is the foundation of research (Boote & Beile, 2005:3). Review of literature
is one of the most important aspects of developing a study and also as a way to know
what has already been written on the subject, methodologies that have been used to
investigate similar concepts or phenomena and to establish the trends on the solutions
that are being advanced to solve the many problems that face mankind (Heppner &
Heppner, 2004:52).
It was therefore necessary to conduct a literature review in order to establish the

general H&S status in the construction industry from research work of others;

theories and literature on H&S performance improvement ; and

client factors associated with good H&S performance in the industry.
In order to guarantee integrity and sophistication of the study, effort was made to
ensure that the review was thorough and exhaustive. Studies reviewed were
synthesised and they considered methods adopted or used in other studies. A critical
review of methods adopted in other studies also ensured that the study did not only
report the claims made in the existing literature because this is one of the remedies
against the trap of simply reporting other people’s claims (Boote & Beile, 2005:3).
91
mean absolute deviation is a calculated mean of all absolute deviations for all
panellists about the median on each question.
Appendix G shows a table of calculated MAD’s for each of the Delphi rounds.
Further analysis involved determining the statistical range in ratings by panellists on
each question and the percentage of panellists with a similar opinion inclination on
each and every question. Consensus was determined to have been achieved when the
MAD was less than one unit below or above the group median, the range in ratings on
each question between all panellists was below 4.0 and the percentage of panellists
that were of a similar inclination in opinion was 60% and above on a particular
question (Appendix G).
Researcher actions
Expert panel actions
Round 1
Circulates questionnaire to
panellists
Rates likely improvement and impact on
H&S due to a factor influence
Round 2
Calculates group medians and recirculate questionnaires
Reviews individual rating in view of the
group’s median. Gives reason if required
Re-Calculates group medians,
standard deviations & compile
comments
Opportunity to reconsider rating
Round 3
Determine consensus and
terminate process
Figure 10: Delphi process
(Adapted from Thangaratinam & Redman (2005:124)
102
The second stage of Delphi data analysis, involved determining the impact
significance of environmental factors on client H&S performance, the impact
significance of factors of client H&S culture on contractor and designer H&S
performance, H&S consideration throughout a construction project’s lifecycle and the
overall project H&S performance.
Equation 1
Where:
Di
 Absolute deviation
xi
 Panellist rating
m X   Measure of central tendency
The impact significance of designer and contractor H&S performance on project H&S
performance was also investigated from the Delphi data. The significance of the
impact of various factors associated with the external environment, client, designer
and contractor was categorised as being critical, major, moderate, minor or low. The
categorisation was helpful in determining which stakeholder factor and relationship
was key and relevant for H&S performance improvement.
This method of analysis has been used in many studies. The impact significance of a
factor of client H&S culture for instance, was obtained as a product of the rated
probability (likelihood) that this factor of client H&S culture would influence
contractor, or designer to implement H&S elements. The rated negative impact
(severity) of a factor for instance of client H&S culture, was the rating of the negative
impact that would result on contractor and or designer H&S performance if this factor
was absent or was not apparent. This relationship is illustrated in Equation 2.
As an example, Table 12 will be used to illustrate how the impact significance was
determined. Column B lists a number of H&S elements that need to be implemented
by the client. Column C to H lists experts’ probability rating that had been agreed
upon to say that elements numbered 1 to 11 would be implemented if the client was
pressured by the external environment factors namely: political, social, economic,
103
technology, legislative and professional bodies (labelled as: PE, SE, TECH, EE, LE,
and PB). Column I list the total score rating for each element numbered 1 to 11that it
would be implemented if the client was pressured by external environment factors in
column C to H. Each element in this example had a maximum of 60 points (6 factors
rated on a 10 point ordinal scale). Column J, list the calculated overall probability that
each client element numbered 1 to 11 would be implemented if the client was
pressured by all the factors. Each factor’s likelihood to influence clients to implement
all the H&S elements for example, was calculated by obtaining the total ratings
achieved by the factor and calculated it as a percentage of the maximum available
ratings which in this case was 110. The impact significance of a factor for instance
PE, was obtained as a product of the likelihood that client would implement H&S
elements which in the example was calculated as 63%, and its severity determined to
be 8. The severity rating was taken to be the rated negative impact on client
implementing the elements 1 to 11 if the factor for example PE was absent.
Equation 2
A prediction of the likelihood that H&S elements would be implemented if pressured
by the identified factors, was also made concerning clients’, designers’ and
contractors’ top management. Predictions were made about the influence of
environmental factors on client H&S performance, and also about client H&S
culture’s influence on contractor and designer H&S performance, H&S consideration
throughout the project lifecycle as well as on project H&S performance.
The impact significance of external environment factors, and factors of client H&S
culture, contractor and designer H&S performance were also investigated. These
values were entered as shown in row 15 of Table 12.
The findings from the Delphi analysis were presented as numbers and percentages in
tables, column and bar charts. These were expert predictions of the likelihood that
H&S elements would be implemented by clients, contractors and designers.
Predictions were also made about the severity and therefore the impact significance of
the evaluated factors.
104
Figure 13: External environment construct
The exogenous variable, client H&S culture, was theorised to be a construct of six
latent endogenous variables namely: leadership (CLL); involvement (CLI);
procedures (CLP); commitment (CLT); communication (CLN) and competence
(CLC). Figure 14 shows the conceptualised client H&S culture measurement model.
The questionnaire items used to measure the leadership factor were marked CLLP1-3,
client involvement was measured by the items labelled CLIP1-3 while that of client
procedures was measured by the items labelled CLPP1-3. Client commitment was
measured by questionnaire items labelled CLTP1-3 and communication was
measured by questionnaire items CLNP1-3. The variable, competence was measured
by items labelled CLC1-4. The numbers of indicator variables measuring one factor
were shown by for instance 1-3, which means there were three indicator variables.
This type of presentation was adopted for clarity and neatness of presentation only.
All indicator variables collectively constituted questionnaire survey items or
questions. During the analysis stage, these variables were parcelled.
116
Figure 14: Client H&S culture measurement model
(CLL= Leadership, CLI = Involvement, CLP = Procedures, CLT = Commitment,
CLN = Communication and CLC = Competence.)
Contractor H&S performance measurement model was defined by indicator variables
namely: procedures (COPP1-3), commitment (COTP1-3) and communication
(CONP1-3). The designer H&S performance was measured by indicator variables of
procedures (DEPP1-3), commitment (DETP1-3) and communication (DENP1-3). The
subscales for contractor and designer H&S performance were theorised to be first
order models with indicator variables grouped in parcels (Figure 15 and Figure 16).
The outcome endogenous variable from the priori structural model was the project
H&S performance factor. The indicator variables theorised to measure project H&S
performance are labelled PP1 to PP17 in Figure 17. The indicator variables also
constituted questionnaire items or questions.
117
Figure 15: Contractor H&S performance measurement model
(COPP= Procedures, COTP= Commitment and CONP= Communication)
Figure 16: Designers H&S Performance measurement model
(DEPP= Procedures, DETP= Commitment and DENP= Communication)
118
Figure 17: Project H&S performance measurement model
Population
The study was conducted in Gaborone and Johannesburg in Botswana and South
Africa respectively. The two cities where chosen for the research because the
researcher was familiar with the areas and it was the intention of the researcher to
delimit the study to these two population areas. It was also viewed that there were
more construction projects in those two areas. Furthermore, accessibility to
respondents was much easier compared to other places in the two countries. It was
also necessary to have two different environments with different laws and practice
governing the common denominator, which was the construction project and more
specifically, the practice of H&S on construction projects. In other words, because of
the external environment factor in the conceptual model, at least two different
environments were necessary which had a different political, legislative and economic
landscape.
119
Generate list of construction projects
Compiled from list of UJ 2nd year National Diploma Building students on industrial
attachment
Identify potential respondents
obtained from list of projects generated in
first stage above
Obtained from researcher's industry network
and referrals
Circulate questionnaire
Email
hand delivery
Respondents complete questionnaire
Electronically
Manually with pen or pencil
Respondents return back completed questionnaires
Email
Collected by
researcher
Fax
hand delivered
Data compilation and analysis
Statistics data capturers
SPSS
SEM/EQS
Figure 18: Questionnaire survey procedure
Data screening and preparation
Before a detailed analysis of the postulated model was conducted to determine fit,
screening of the data was essential. Pre-analysis data screening focused on
establishing whether there were any missing data, outliers, the distribution
characteristics of the data, and the identification of the model.
According to Raykov, Tomer & Nesselroade (1991:500), missing values and outliers
can adversely affect SEM results by their presence in the raw data. It was therefore
necessary to identify, at the pre-analysis stage, any missing values and outliers in
order to determine the best way to handle them. According to Kline (2005:52),
missing values may pose a very serious challenge to analysis. In addition to
124
likelihood that clients would implement the H&S elements. The standard deviations
for these two groups of factors were determined to be 0.02.
Figure 19: Impact significance of external environment factors to client culture
Figure 20: Client likelihood to implement H&S elements
136
Figure 21: Client likelihood to implement H&S elements
6.2.2
Client influence on H&S consideration in the project lifecycle - D2
Client H&S culture was defined by four factors, namely: commitment, competence,
involvement and leadership. The influence of these factors was estimated. As
explained in the previous section, the level of influence of a factor was inferred from
the calculated value of the impact significance.
The H&S performance on a construction project depends in part on the extent to
which it is considered throughout the project life cycle. H&S consideration entails
H&S risk identification, analysis, assessment, mitigating, implementing and assessing
the implementation of H&S measures. In order to evaluate H&S performance at the
project level, it was necessary to evaluate H&S consideration throughout the project
life cycle.
137
significance on H&S consideration was leadership. The impact significance of the
leadership factor was determined to be 6.68. This rating was considered to be of
‘major’ impact significance (see Table 13). However although the rating of 6.68 was
low, when compared to other factors of client H&S culture, its impact significance
and therefore its influence was considered to be ‘major’. Therefore the contribution of
the leadership factor to H&S performance may not be overlooked.
The likelihood of H&S being considered as a result of client H&S culture’s influence
was determined to be 83% (Figure 13). This rating of 83% suggested that H&S
consideration throughout the project life cycle was ‘very likely to occur’ (Table 14).
Figure 23: Likelihood of H&S consideration due to client’s influence
It was further determined that H&S consideration was ‘very likely to occur’ at all the
project stages except at the procurement stage. With the exception of the procurement
stage, the likelihood of H&S consideration at all other project stages was determined
to be above 80%. On the other hand, the likelihood of H&S consideration at the
procurement stage was determined to be 73%. The likelihood of 73% meant that H&S
consideration was ‘likely to occur’.
H&S consideration was most likely at the construction stage. The likelihood was
determined to be 95%. It seemed from this finding that H&S consideration was almost
certain at the construction stage with client H&S culture’s influence.
139
considered to be ‘critical’ impact significance. The finding seemed to suggest that the
leadership factor was critical to contractor H&S performance.
The competence factor had the least impact significance on contractor H&S
performance. This factor had a rating of 6.20. Although the impact significance was
lower when compared to all other factors of client H&S culture, its impact was
considered to be ‘major’ (Table 13). Apart from this, the difference in the impact
rating between all factors was minimal. The standard deviation in impact significance
values was found to be 0.34 suggesting a small variability between all impact
significance estimates.
The likelihood that contractors would implement H&S elements as a result of
influence from client H&S culture was determined to be 83% on average. This rating
suggested that the implementation of H&S elements by contractors due to client H&S
culture’s influence was ‘very likely to occur’ (Table 14). The standard deviation of all
likelihood estimates was found to be 0.03. This standard deviation value was very
small and therefore indicated a small variability in the estimates.
Figure 25: Contractor likelihood to implement H&S elements
141
The other H&S elements which had above 75% likelihood of being implemented
included the designer to:

develop staff competence,

consult and communicate H&S information and

employ permanent H&S staff.
The rating of 75% likelihood meant that designers were ‘likely’ to implement H&S
elements (Table 14).
Designers were least likely to conduct H&S audits and inspections. The likelihood of
designers conducting H&S audits and inspections as a result of client H&S culture’s
influence was determined to be 73%. Although this likelihood was low when
compared to the likelihood of other H&S elements being implemented, it was actually
a favourable result considering that conducting H&S audits and inspections was
‘likely to occur’.
Figure 29: Designer & contractor likelihood to implement H&S elements
147
Contractors were more likely to implement H&S elements than designers when
influenced by client H&S culture. The likelihood of contractors implementing H&S
was 83% while that of designers was 78% (Figure 29). The largest disparity was
observed in the likelihood that the following H&S elements would be implemented,
namely:
1. employ permanent H&S staff;
2. consult and communicate H&S information to all stakeholders; and
3. conduct H&S audits and inspections.
Contractors were more likely to implement these H&S elements. According to the
likelihood scale (Table 14), contractors were ‘very likely’ to implement H&S
elements while designers were ‘likely’ to implement H&S elements. Similarly, the
impact significance of client H&S culture was greater on contractor H&S
performance than it was on designer H&S performance (Figure 30). The average
impact significance of client H&S culture on contractor H&S performance was 6.60
while that on designer H&S performance was 6.45.
Figure 30: Impact of client factors on designer and contractor H&S performance
148
Figure 33: Likelihood of H&S consideration due to designers' influence
Figure 34: Likelihood of H&S consideration due to client and designer influence
152
Consequently, the process of H&S performance improvement could simply follow the
following cycle:
1. Assess LIP +3C;
2. Develop strategies in consultation with stakeholders to enhance LIP + 3C;
3. Implement strategies;
4. Assess LIP +3C;
5. Act to improve LIP+3C;
6. Repeat process 1-5.
The above is based on Deming’s PDCA cycle and the same approach is suggested by
IOSH (2004:9).
Contractor
performanc
e
Client
culture
Environment
Project
performanc
e
Designers’
performanc
e
Figure 35: Theorised client centred H&S performance improvement
7.3
CHAPTER SUMMARY
In chapter seven, a conceptual model was theorised that client H&S culture had
influence on contractor, designer and the overall project H&S performance. It was
further theorised that client H&S culture was a function of the factors of the external
environment namely: political, economic, social, technology and legislative. From the
Delphi study and the literature review, the factors of client H&S culture were found to
be,
leadership,
involvement,
procedures,
commitment,
communication
and
competence. These six factors of client H&S were collectively referred to as LIP+3C.
Findings from the validation of the conceptual model developed in chapter seven will
now be presented in chapter eight.
167
CHAPTER EIGHT
QUESTIONNAIRE SURVEY RESULTS
8.1
INTRODUCTION
The postulated sub-models are diagrammatically shown in Figure 36 and 37. Indicator
variables for the exogenous variables of external environment and client H&S culture,
and the endogenous variables of designer, contractor and project H&S performance
are shown in each measurement model. The theory behind the hypothesised clientcentred H&S performance improvement model was based on literature and on
experts’ opinion acquired during the Delphi study described and presented in chapter
six of this study.
Raw data from questionnaires were entered into the Statistical Package for Social
Sciences (SPSS) software and was later exported to the structural equation modelling
(SEM) software EQS version 6.1 for analysis (Bentler, 1999). The sample used for the
analysis of the model was 281 cases. According to Kline (2005:15), a sample size of
281 is classified as large. A small sample of less than 100 cases tends to be
problematic when it comes to structural equation model analysis (Kline, 2005:15).
Figure 36: Hypothesised model-Client influence on project H&S performance
168
Leadership
Procedures
External
Environment
Commitment
Communicatio
n
Competence
Figure 37: Environmental influence on client H&S culture
169
Client H&S culture
Involvement
5. communication factor is explained by indicator variables CLNP1 to 3;
6. competence factor is explained by indicator variables CLCP1 to 3; and
7. procedures factor is explained by indicator variables CLPP1 to 3.
These indicator variables are presented in Table 24. In order to establish how well the
model fit the sample data and the strength of the hypothesised relations between
variables, results presented on residual covariance matrix, distribution of standardised
residuals, goodness-of-fit statistics and parameter estimates’ statistical significance at
probability level of 5% were examined.
In addition the Cronbach’s alpha and the Rho coefficient of internal consistency were
examined for score reliability. The construct validity of the measurement model was
determined from the convergent validity and the magnitude of parameter coefficients.
Results of the above analysis are presented in this section.
Figure 38: Theorised six factor client H&S culture model
174
COPP1
COPP2
COPP
COTP1
COTP2
COTP
COTP3
CONP1
Contractor H&S
Performance
COPP3
CONP
CONP2
CONP3
Figure 39: Contractor H&S performance construct
8.3.3.2 Goodness-of-fit statistics - RML
The analysis strategy of goodness-of-fit for the contractor H&S performance construct
followed a two statistic strategy of fit indexes and is reported on in this section.
The sample data on contractor H&S performance measurement model yield the
of 25.0664 with 24 degrees of freedom and (N=275; P=0.40216,). The chisquare was insignificant. This chi-square value indicated that the departure of the
sample data from the postulated measurement model was not significant and hence
indicative of an acceptable fit. However the chi-square test is very sensitive to sample
size and is used more as a descriptive index of fit rather than as a statistical test
(Kline, 2005:136). Therefore a normed Chi-square value is usually adopted by most
researchers. Normed chi-square is the procedure of dividing the chi-square by the
degrees of freedom. The normed values of up to 3.0 or even 5.0 are recommended
(Kline, 2005:137). From the above chi-square and degrees of freedom values the ratio
was found to be 1.044. This ratio was lower than the upper limit of 3.0 or 5.0
advocated for by some authors (Kline, 2005:137).
195
variables, 12 independent variables and 21 free parameters. The number of fixed nonzero parameters was 12.
A three factor designer H&S performance construct comprising commitment,
communication and procedures, was analysed before it could be included in the full
latent variable model. The indicator variables were organised in parcels as shown in
Table 40 and Figure 40.
In order to establish how well the measurement model fit the sample data and how
well the indicator variables measured the factors of designer H&S performance
construct, results presented on residual covariance matrix, distribution of standardised
residuals, fit statistics and statistical significance at probability level 5% were
examined. In addition the Cronbach’s alpha and the Rho coefficient of internal
consistency were examined for score reliability. Construct validity of the
measurement model was determined from model convergence and the magnitude of
parameter coefficients.
DEPP1
DEPP2
DEPP3
DEPP
DETP2
Designer H&S
Performance
DETP1
DETP
DETP3
DENP1
DENP
DENP2
DENP3
Figure 40: Designer H&S performance measurement model
(DEPP = procedures, DETP = Commitment and DENP= Communication)
200
Figure 41: External environment construct
(LGN= Legislative, ECON= Economic, SOC = Social, PRI= Professional and TECH
= Technology & materials)
8.3.5.1 Diagnostic fit analysis –Residuals
The average absolute residual values of the external environment construct are
presented in Table 46. Results revealed that all the absolute residual values and the
average off-diagonal absolute residuals, both unstandardised and standardised, were
close to zero. The smallest unstandardised average off diagonal residual was 0.0289
while the largest was 0.0405. Similarly, the smallest standardised average off diagonal
residual was 0.0241 while the largest was 0.0337.These residual values were
considered small as they were all less than 2.58 (Byrne, 2006:94) In addition, 100%
of standardised residuals fell within the acceptable range of -0.1 and +0.1. The
significance of this distribution is that for a model to be described as well-fitting, the
distribution of standardised residuals should be symmetrical and centred around the
zero (Byrne, 2006:94). From the above information, the results seemed to suggest that
the model had a good-fit to the sample data. Therefore, since this initial examination
of residuals indicated a fairly good fit, further tests of goodness-of-fit were justifiable.
207
P1
CLNP
P2
Client H&S culture
P3
P4
H1a
CLTP
P5
P6
H1b
P7
LGN
CLPP
ECON
SOC
PRI1
H1c
TEC
P9
External
Environment
P 10
H1d
CLIP
PRI2
P8
P 11
P 12
H1e
P 13
CLLP
P 14
P 15
H1f
P 16
P 17
CLCP
P 18
P 19
Figure 42: Model 1.0 - External environment influence on client H&S culture
216
Figure 43: Model 2 - Client influence on project H&S performance
221
Contractor
H&S
performance
Project
H&S
performance
Client H&S
culture
Designer
H&S
performance
Figure 44: Finalised model 2.0 for client H&S culture influence
Statistically significant relationship
Statistically insignificant relationship
8.5
CHAPTER SUMMARY
The postulation for the overall model was that the external environment had influence
on client H&S culture and which in turn had influence on contractor, designer and
project H&S performance.
SEM results of measurement models were presented in this section. These results
were obtained from an analysis to determine whether the indicator variables
(questionnaire items) actually measured the constructs that they were supposed to
measure. In addition, results were also presented from the evaluation of the submodels to establish whether the theorised number of factors for the sub-models was
working properly before inclusion in the analysis of the full latent variable model.
Further, results on reliability and construct validity were also presented. It was
important to ensure that the measurement model was in a good working order before
the full latent model could be analysed.
233
The analysis of the full latent variable model was conducted by separating the
postulated model in to two parts namely, model 1.0 and model 2.0. The division of the
model into two parts was merely for the purpose of simplicity. Apart from this, the
focus of the study was on the influence of client H&S culture on project H&S
performance. However, it was also necessary to establish the influence of the external
environment on the client H&S culture.
The findings were that the external environment had influence on client H&S culture.
Further, client H&S culture was found to have a direct positive influence on
contractor and designer H&S performance. However, client H&S culture was found
to have an indirect positive influence on project H&S performance that was mediated
by contractor and designer H&S performance. Therefore the finalised overall clientcentred H&S performance improvement model (combining model 1 and 2) was found
to be as presented in Figure 45.
Model 1.0
Model 2.0
Contractor
H&S
performance
External
Environment
Client H&S
culture
Project H&S
performance
Designer H&S
performance
Figure 45: Finalised Client-centred H&S performance improvement model
234