AQUAMONEY:
UK CASE STUDY REPORT
Ian J. Bateman, Silvia Ferrini and
Stephanie Hime
CSERGE, University of East Anglia,
Norwich, NR4 7TJ, UK.
7th November 2008
CSERGE Team
Page 1
TABLE OF CONTENT
1.
2.
Introduction.............................................................................................................. 3
Description of the case study ................................................................................. 10
2.1
Location of the case study area ...................................................................... 10
2.2
Water system characteristics .......................................................................... 12
2.3
Short characterization of water use and water users ...................................... 15
2.4
Main water management and policy issues in the context of the WFD ........ 16
3. Study Design .......................................................................................................... 19
3.1
Questionnaire design...................................................................................... 19
3.2
Sampling procedure and response rate .......................................................... 23
4. Valuation results .................................................................................................... 29
4.1
Respondent characteristics and sample representative ness .......................... 29
4.1.1.
Demographic characteristics ...................................................................... 29
4.1.2.
Socio-economic characteristics.................................................................. 30
4.1.3.
Water use characteristics ........................................................................... 32
4.2
Public perception of water management problems ........................................ 34
4.3
Estimated economic values for water resource management ........................ 35
4.4
Factors explaining economic values for water resource management .......... 38
4.5
Total Economic Value ................................................................................... 40
5. Conclusions ............................................................................................................ 41
6. Best practice recommendations ............................................................................. 42
7. References .............................................................................................................. 43
Appendix: Survey questionnaire .................................................................................... 62
CSERGE Team
Page 2
1.
INTRODUCTION
The European Union Water Framework Directive (WFD) requires all EU member states
to achieve “good ecological status” in their water bodies by 2015. This ambitious target is
likely to involve major policy changes, aimed at reducing pollution from both diffuse and
point sources. These improvements will have effects on both environmental quality and
human wellbeing. However, the cost of such changes is expected to be substantial at
between £450-630 million per year just in the UK (ENDS, 2008). Since the WFD allows
exemptions from implementation of improvements in cases of “disproportionate costs”
(RPA, 2004), estimating the benefits arising from water quality improvements assumes
crucial policy relevance and has been subject of some researches (Bateman et al. 2006,
Hanley et al. 2006, Birol et al. 2006).
The changes to the quality of open access waters (such as rivers and lakes), envisaged by
the WFD, typically generate public goods benefits. As such, analysts wishing to value
such changes are forced to rely upon non-market valuation techniques. These can be
broadly subdivided into two groups. Revealed preference methods rely upon assumptions
of weak complementarity to infer values from observed behaviour (Champ et al, 2003).
Stated preference (SP) methods, such as choice experiments or the contingent valuation
method, attempt to directly elicit values by asking a direct valuation questions posed to
respondents via survey interviews (Bateman et al., 2002). SP approaches ask the
individual to make hypothetical choices that can be formulated to identify total values
given by the sum of use and non-use values. Therefore, while both revealed and stated
preference methods have a long history of applications, recent years have seen an increase
CSERGE Team
Page 3
in the use of the latter (Carson, 2007). Given the fact that improvements in open water
quality may generate both use values and non-use values, the Aquamoney study adopts a
stated preference approach.
The UK Aquamoney project aims to value improvement in the water quality using the
Contingent Valuation (CV) method (Bateman et al., 2002). This allows estimation of the
total values attached to the WFD environmental improvements. However, having in mind
some common anomalies reported for the CV studies (Bateman et al. 2004), the study has
been designed to take into account: (a) the scope effect and (b) the ordering effect.
The sensitivity to scope, documented in the economic literature (Bateman et al. 2004,
Bateman et al 2001), implies that respondent values equally goods of different sizes. This
result contrasts with the neo-classical assumption of consumer behaviour rationality
(transitivity, complete and reflexive). The extent to which these assumptions hold true is
important for the interpretation of the results and the accuracy of the elicited willingnessto-pay (WTP) values, and consequently the degree to which the results provide valuable
information for policy makers. The questionnaire has been designed to test scope effects
in the non-market valuation of water quality improvements, focusing on quantity changes.
This test also allows assessing the internal validity of the study. The scope test is being
conducted using a split sample experiment in which one or two river stretches is valued
from respondent. In designing this test we take into account another well discussed bias:
the ordering effect.
The ordering effect (also termed procedure invariance effect) may be observed as a
consequence of the order in which scenarios appear in the questionnaire. Therefore in the
case of one of two stretches we want test if the individual willingness to pay (WTP)
CSERGE Team
Page 4
changes with the order in which they are presented to the respondents. Ordering effects
can be seen as a symptom of context dependent preferences and heuristics being applied to
overcome the cognitive burden of answering complex questions. Therefore, if an ordering
effect is observed this signifies the use of heuristics in the decision making process
suggesting that people may not have existing preference sets for the valued good.
Consequently, the forming of preferences may be influenced by the order in which the CV
scenarios are presented. The issue of ordering effects in CVs has been discussed in the
environmental economics literature (Clark and Friesen, 2006, Powe and Bateman 2003,
Carson et al. 1998) and it can undermine the reliability of the valuation studies. Therefore,
in the our case study we take account of both order and scope effects to test the internal
validity of the results.
The benefits of WFD-induced water quality changes will vary across space in a complex
manner, depending not just on the distribution and physical response of catchments, rivers
and estuaries, but also upon the distribution of present and potential future beneficiaries.
Another aim for research is therefore to capture the complex interplay and spatial
distribution of the key benefits of differences in river qualities.
The spatial element is central in the AquaMoney project since we believe that it is the
type, level and geographic characteristics of water improvement that are crucial to
assessing the individual willingness to pay. Consequently, throughout the project we
have made extensive use of spatially sensitive modelling routines and made routine
application of geographical information systems (GIS). The WFD benefits include use
values such as improved opportunities for, and qualities of, informal recreation 1 and
(perhaps more contentiously) non-use benefits, such as the values individuals may hold
1
Furthermore, where direct contact with water is likely, such as in bathing areas, then some health risk
reduction benefits may arise.
CSERGE Team
Page 5
for improvements in wildlife habitat which are not incorporated within recreation and
amenity values. Therefore to test the distance decay effect we explicitly include the
geographical location of the improved goods respect to the respondents’ home. As
documented in Bateman et al. (2006), we expect that the WTP values are correlated
with distance and they approach to zero after a certain distance. This effect is well
documented from the first law of geography which states "all things are related, but
near things are more related than far things." (Tobler). This law becomes fundamental
when we aggregate benefits across individuals.
Individual characteristics also vary spatially (for example, those with higher WTP
values and/or higher incomes may live closer to a given set of sites, etc.) and a
consideration of space is therefore vital to the accurate aggregation of benefit values for
improvements to amenity waterways (and indeed most spatially confined
environmental resources).
The spatial analytic capabilities of a GIS provides an ideal medium for harmonising the
diverse data necessary to undertaking such an aggregation exercise (Bateman et al,
2000). In particular a GIS readily allows the researcher to specify a valuation function
which varies across space according to a variety of factors including: (i) the distribution
of rivers, lakes, estuaries, etc., (ii) the change in quality to those resources, with
improvements tending to convert former non-users into users in a spatially non-random
manner; (iii) the accessibility of complementary and substitute assets; (iv) the
distribution and socio-economic/demographic characteristics of the population. The
inclusion of such factors allows the analyst to observe any ‘distance decay’ in values as
we progressively consider households which are more remote from a given
improvement. Furthermore, once such a valuation function is estimated, by applying it
CSERGE Team
Page 6
within a GIS to data detailing all explanatory variables for all locations we can define
the appropriate ‘economic jurisdiction’ (that area within which values are non-zero) for
calculating total benefit values (Bateman et al., 2006b). This avoids common
aggregation problems associated with artificial ‘political jurisdictions’ typically defined
by convenient administrative areas rather than the benefits generated by a scheme 2 .
The estimation of spatially sensitive valuation functions also allows us to investigate
the potential for ‘benefit transfers’ (Brouwer, 2000; Barton, 2002; Ready et al., 2004).
Here value functions, estimated using data such as that described above, are applied to
generate values for policy relevant sites which may not of themselves have been part of
the valuation survey exercise. In theory, once a robust valuation function has been
estimated the researcher need only know the attribute levels and improvement scheme
which characterises an unsurveyed site in order to estimate (via the valuation function)
the benefit generated by improving that site. However, in practice the track record of
benefits transfer work is strewn with failure as the majority of analyses fail recognised
tests of transfer validity.
Despite the empirical problems of benefit transfer, its
potential to obviate the need for conducting individual site surveys each time an
environmental improvement is to be assessed makes this an attractive proposition.
Transferring well specified functions (if they can be estimated) allows the analyst to
obtain a value which is adjusted for the characteristics and environment of the site in
question. The key issue therefore is to determine the robustness or otherwise of the
valuation function to be transferred. This is typically undertaken by taking a function
estimated from one subset of observations and using it to estimate values for an
2
The typical problem here arises when survey sampling is predefined to occur in a set area which is not
representative of the economic jurisdiction. In such cases aggregation approaches which do not rely
upon valuation functions but instead use survey sample means may (although this is not inevitable)
result in substantial bias. For example, if sampling is confined to an area close to a site but the sample
mean is applied to a larger aggregation area then total value assessments may be upwardly biased.
CSERGE Team
Page 7
alternative set of sites for which independent value estimates are already held. The
proposed research will investigate such robustness analysis. In so doing we will be
guided in part by recent findings which suggest that the transfer of statistical ‘best-fit’
functions may inflate value estimation errors (Brouwer and Bateman, 2005). This is
because such functions may include site specific contextual factors which are not
relevant to those sites the function is transferred to. Conversely the specification of
functions on the basis of those general factors identified in core economic theory (e.g.
income levels, usage parameters, etc.) can produce valid transfers which outperform
simple mean value approaches in terms of the errors generated.
A final advantage offered by our spatially sensitive, GIS-based, methodology is an
ability to examine the distributional implications of the water quality improvements
offered by the WFD. While valuation studies often note an association between WTP
and household income, the implications of this association are rarely explored.
Furthermore, given that the distribution of benefits is likely to be both spatially and
socially uneven, the potential exists that some groups will fare better than others in
capturing the non-market benefits of the Directive. The estimation of a value function
which varies across space and socio-economic dimensions allows us to use the GIS to
link between census measures of deprivation and corresponding WFD benefit values. It
seems likely that many environmental benefits, such as those generated by National
Parks, are captured by the richer sections of society. This may be the case with water
quality benefits; however the case is less clear cut. Informal waterway recreation is
open to all and, unlike National Parks, is much more widespread and accessible.
Furthermore, if the WFD is indeed implemented as it stands then it seems likely that the
benefits and beneficiaries may be clustered within urban and surrounding areas.
However, while this Directive seems potentially redistributive in terms of its benefits,
CSERGE Team
Page 8
their urban concentration contrasts markedly with the predominantly rural incidence of
WFD costs.
CSERGE Team
Page 9
2.
DESCRIPTION OF THE CASE STUDY
2.1 . Location of the case study area
The area selected for our case-study is located in and around the city of Bradford in the
northern part of the Humber basin in the United Kingdom (Fig. 1).
Figure 1 The Humber Basin
The Humber basin cover an area of 25000 km2, more than 20% of the land area of
England, has a population of over 10 million people. The basin drains 28% of England,
mainly via its two principle river catchments, the Ouse and the Trent. The rivers Trent
and Ouse, which provide the main freshwater flow into the Humber, drain large
industrial and urban areas to the south and west (River Trent), and less densely
populated agricultural areas to the north and west (River Ouse). These two drainage
basins are almost identical in area.
The Humber is an ideal case study area as it captures a full range of emission levels of
key pollutants such as phosphates, nitrogen and pesticides. In addition to its overall size
CSERGE Team
Page 10
and significance, the contrasting characteristics of sub-catchments across the Humber
basin provide useful variation for subsequent extrapolation.
In this area there are several major urban centres: Nottingham, Leicester and the West
Midlands/Birmingham conurbation which are drained by the Trent, the Leeds–Bradford
area in West Yorkshire is drained by the Aire/Calder and the Sheffield/Doncaster area
in South Yorkshire is drained by the Don. The area selected for the case study is the
Ouse catchment with the largest urban area being the Leeds– Bradford conurbation,
totalling 1, 200,000 inhabitants. Three main rivers flow in the area: Calder, Wharfe and
Aire. They are surrounded by high-grade agricultural land and high population/industry
area therefore they present a great variety of water quality levels. The area selected for
the study is reported in Fig. 2 and the map correspondents to the one used in the
questionnaire.
CSERGE Team
Page 11
Figure 2 The case study area
2.2 . Water system characteristics
The Humber water system has been analysed in Oguchi et al. (2000) with the aim to
assess the river water quality across the Humber catchment. The Environment Agency
data including water quality data for rivers, sewage and trade effluents were coupled
with the Water Information System from the Institute of Hydrology and they verify the
chemical quality of the Humber water system. The results show a high variability in
water quality for many determinants across the Humber catchment mainly due to
sewage discharges and industrial effluents. The study demonstrates the importance of
anthropogenic influences on the large-scale regional water quality of the Humber
catchment. However, these results do not describe completely the river quality. In fact,
river quality encompasses other attributes or properties such as ecological/biological,
aesthetic, geological and flow characteristics (Holmes et al., 1999).
CSERGE Team
Page 12
Current river water quality within England and Wales is assessed annually under
Environment Agency general quality assessments (GQAs) through reference to
standards concerning recommend levels of dissolved oxygen, biological oxygen
demand, ammonia (total/un-ionised), pH, water hardness (CaCO3), dissolved copper,
and total zinc (Environment Agency, 2007a). The WFD requires a substantial shift in
assessment practice with the focus moving toward outcomes, in the form of ecological
status, rather than chemical composition. However, this poses practical and
methodological problems. The UK, has long time series data concerning the chemical
composition of open waters although, assessments of macro-invertebrates and aesthetic
river features have been a part of river water quality assessment in the UK since 1988.
To date there are few if any systematic assessments of ecological status in its entirety,
where ecological status includes all features of the river environment e.g. aquatic
plants, macro-invertebrates, bank-side vegetation and algae. Indeed even the meaning
of ‘ecological status’ is the subject of a pan-European debate to be concluded late in
2008. However, as an initial focus UKTAG has sought to determine the biological
elements associated with high ecological quality in terms of macro-invertebrates and
their links to measured levels of biological oxygen demand (BOD) and Ammonia
(UKTAG, 2008).
Given the impossibility of generating a measure of ‘good ecological quality’ prior to its
final definition, data collected as part of the GQAs of river water in the UK has been
used to assess the water quality system in the survey area. Consistently with the quality
levels used in the description of the new water quality ladder (cfr. Section 3.1), we
define four quality levels: Blue as pristine rivers, Green as a good quality, Yellow as
fair quality and Red for poor quality. Then joining these quality levels with the EA data
CSERGE Team
Page 13
we describe the current water system quality as it is in Figure 3. The BOD measure and
Ammonia level reported in Fig. 3 show good river quality level almost in all the study
area. Some variations in term of water quality can be found in the urban area around
the two main cities: Bradford and Leeds. These cities are highly populated and an
improvement in water quality can be a desirable policy for residents but less relevant
for residents whose live far from these areas. Therefore using this description of quality
and a good spatial distribution of sample the survey will determine the aggregate
benefit value of the WFD.
(a)
(b)
Figure 3 The mean generic water quality level from
(a) - biological oxygen demand (BOD) measures and (b) – Ammonia concentrations taken at
Environmental Agency sampling points from 1986-1997, on three rivers; the Wharfe, Aire and
Calder in the Humber region.
We point out that at the time that the our quality level was being developed within the
CSERGE team, UKTAG were defining their own standards with regard to the high,
good, moderate and poor ecological status of river water in the UK.
Therefore
nowadays we can compare our quality level with the UKTAG quality level finding that
they are consistently similar (Hime and Bateman 2008).
CSERGE Team
Page 14
2.3 . Short characterization of water use and water users
The waterways in the Humber are manly used for agricultural activities, industries
activities and domestic water. The agricultural land is generally arable and horticultural
land, growing mainly cereal crops, oilseed rape, and root crops, and grazing land. The
industry sector was traditional represented by the production of textiles, iron and steel
while nowadays the chemical and petrochemical industries are flourishing as is the
power industry. A considerable activity in the Humber catchment is the commercial
fisheries which takes place in the upper reaches of the estuary, and in some of the rivers
leading to it (Gray, 1995). On the other hand, considering the type of users of the
waterways is a fundamental task to understand the ability of individuals to value water
quality improvement. In fact, the benefits provided by rivers may provide examples of
both well-formed and poorly-formed preference. For example, high-intensity,
recreational users of waterways, such as anglers, seem likely to have well formed
preferences for river resources and have robust values for the any changes in river
quality. Following Plott (1996), it is the high degree of consumption (use) experience
which has led such anglers to ‘discover’ robust, theoretically consistent, economic
preferences. Following this logic it seems likely that the absence of such experience is
likely to mean that preferences for the non-use benefits of rivers are less well formed
and therefore derived values are more liable to exhibit anomalies relative to standard
theoretical expectations. Using data from previous unpublished studies and official
statistics we characterized the main uses of the area (Fig. 4).
CSERGE Team
Page 15
Figure 4 The distribution of users in the case study area
The mainly uses of the area are for walking and fishing.
2.4 . Main water management and policy issues in the context of
the WFD
The Water Framework Directive (WFD) (European Parliament, 2000) represents a
fundamental change in the management of water quality in Europe. The Directive imposes
outcome based targets, requiring a shift away from standards framed in terms of the
chemical composition of water in favour of an approach which assesses the ecological
quality of water bodies. These standards will be water body specific and hence require
differentiated action to improve all European waters to “good ecological status” by 2015.
Although the definition of such status depends upon reference conditions, it is generally
agreed that implementation of the WFD will require substantial reductions in pollutant
inputs to rivers both from point and diffuse sources (Environment Agency, 2004).
However, there is uncertainty with the definition of the water quality level that satisfy the
“good ecological status”. This represents an issue in determining the WFD benefits.
CSERGE Team
Page 16
A number of approaches have been used to convey information regarding water quality.
Perhaps the most well known of these is the Resources For the Future (RFF) water quality
ladder (Vaughan, 1986; Mitchell and Carson, 1989; Carson and Mitchell, 1993). This is a
use-based measure, describing open water quality in an ascending scale from having no
uses, to being suitable for boating, fishing and then swimming. Variants of this approach
have been used in a number of stated preference studies right up to the present day
(Desvousges et al., 1987; Bateman et al., 2006a). While providing excellent service
through the years, the categories used in the RFF are somewhat limited regarding the
extent to which they convey the ecological changes (and associated use and, importantly,
non-use benefits) implicit in movement up or down the ladder. Furthermore, the ladder
focuses upon use categories which do not readily relate to national data on water quality
(which to date typically tend to focus upon water chemistry measures)3. This limits the
transferability of results and the RFF water quality ladder have been thrown into sharper
relief by the introduction of the EU Water Framework Directive (WFD; European
Parliament, 2000). Therefore in section 3.1 we are introducing new water quality ladder
useful to represent the WFD improvements.
The EU WFD Directive requires the catchment based management of water, entailing a
substantial expansion in the spatial scale of water management. Therefore, the UK has
replied to the EU request creating 11 river basin liaison panels in England and Wales. By
the end of 2008 each river basin will have draft management plans with associated
programmes of measures. These measures will be aimed at reducing the pressures sectors
3
As part of the implementation process for the Water Framework Directive, work is ongoing to agree an
EU-wide measure of ecological status. This should be available toward the end of 2008. The water
quality ladder described in the here should be readily adaptable to such a measure.
CSERGE Team
Page 17
exert upon the water environment. The measures are decided upon by the stakeholders in
these liaison panel meetings, with public consultation upon the reports it produces.
The Humber river basin liaison panel has been created including a range of interested
partners who represent the activities within the river basin with different impacts on the
water environment. The sectors involved are:
ƒ
agriculture
ƒ
diffuse pollution
ƒ
fisheries
ƒ
industry
ƒ
spatial planning
ƒ
catchment abstraction management strategies
The management plans of the Humber has not been published yet.
.
CSERGE Team
Page 18
3.
STUDY DESIGN
3.1 . Questionnaire design
In designing the questionnaire we took into account the useful line of research provided
by work in the psychological sciences (Lipkus and Hollands, 1999; Hibbard et al.,
2002; Hibbard and Peters, 2003). We avoided using numeric and textual approaches to
represent river quality improvements and we preferred the presentation of information
in visual form. This representation choice aimed to increase the individual
understanding and eliciting values for land use change. Therefore, as part of the
questionnaire design process, a water quality index has been constructed to enable each
respondent to visualise the overall ecological qualities of each particular site (Hime and
Bateman 2008).
Given the importance to use visual information we used a computerized questionnaire that
has been tailored for our case study. The advantage of computerised presentation is that
river maps are tailored to respondents’ locations, in some extent, each respondent
received as much information as possible to asses the water quality improvements.
The questionnaire was divided in three main sections:
ƒ
Introduction
ƒ
Valuation task
ƒ
Socio-demographic data.
A simplified version of the computerized questionnaire is reported in the Appendix.
CSERGE Team
Page 19
In the first part of the questionnaire, we collect information regarding the respondent’s
home, recreation activities over the last 12 months and recreation activities on the river
sites on the study are map (Fig. 2). This section was fundamental to introduce the generic
measure of water quality level. To define our measure of water quality we have taken into
account the following considerations.
We introduce a scale depicting changes in the level of that overall quality. In constructing
that scale we extend the well tested ‘water quality ladder’ approach of Vaughan (1986)
and Carson and Mitchell (1993). This depicts different levels of water quality in terms of
the characteristics of that level. The original work in this area focussed upon use-based
characteristics such as a rivers suitability for boating, fishing, swimming, etc. Given the
need to capture non-use as well as use value (and the complexity of natural environment
factors which may determine non-use value) we have rejected the conventional water
ladder and we designed our own version. Here a graphic artist has been hired to create
images depicting four different levels of water quality, ranging from high to low. The
process of defining these levels is as follows:
ƒ
Ecological data informs the choice of river profile (bank and bed form, breadth,
etc)
ƒ
Ecological research also determines the flora and fauna which characterise each
quality level.
ƒ
The definition of each level is chosen to span the breadth of quality levels
charactering rivers and be linked to existing information concerning the quality
of rivers. It is vital that levels can be related to well defined, numeric or
categorical measures of river quality so that resulting values can be
meaningfully interpreted and related to changes induced by the WFD.
ƒ
The definition of differences between levels must be such that they reflect the
likely effects which implementation of the WFD may have on open water
quality.
CSERGE Team
Page 20
ƒ
Direct use of the river by humans (e.g. fishing, walking, boating) is not depicted
in these images as it is felt that respondents may react in diverse ways to such
usage and that this might distract from focus upon those factors under policy
control. Furthermore, it is felt that respondents can readily visualise use of
rivers.
Each of the water quality levels has been assigned a distinctive colour. This colour has
been used to indicate the levels of the water quality attribute in a river when depicted on
maps presented to survey respondents (see Appendix).
Taking these facets together, our water quality ladder both conveys levels of quality to
survey respondents and is defined such that responses and values can be directly related to
those changes in water quality which are likely to arise from implementation of the WFD.
Therefore, these water quality ladders provide us with levels for our generic water quality
attribute. In essence this provides the basis for answering question concerning the value of
improving water quality.
The second section of the questionnaire introduces the valuation exercises.
Alternatively the respondent could receive first a contingent valuation (CV) task or a
choice experiment. For sack of brevity we focus here only on the CV exercise with
payment card. Firstly, we showed respondents a status quo map with the current water
quality level and then respondent entered in the CV exercise.
This exercise has been design to take into account anomalies generate by ordering and
scope effects. The order test controls for the procedural invariance of answers, as to say
answers should be independent from the sequence of questions. The scope effect verify
CSERGE Team
Page 21
that the WTP is equal or higher for a larger quantity of a good than for a small
quantity.
Therefore we define two different quantities of river improvement:
A1= Small quantity improvement – one river stretch
A2= Large quantity improvement – one river stretches
A2 is defined so that it contains all of A1 plus an additional river stretch to ensure that,
in quantity (and quality) terms, A2 > A1.
In order to test for ordering effect, we ask some respondents to value improvement A1
and then A2 while other respondents value A2 and then A1. By varying the order of
presentation and denoting the 1st and 2nd question by subscripts, we obtain four
different versions of CV questions as reported in Table 1.
Table 1 The CV questions
Order of question
Size of the good
Small
Large
1st
2nd
The computer system automatically randomized the versions of questionnaire of Tab. 1
displaying the correct sequence for each respondent. Therefore each respondent
received two different sequences of questions with variation in river quantity.
Respondents should value the water quality improvement identifying the maximum
willingness to pay (WTP) to achieve it. To help them to define the WTP value we used
CSERGE Team
Page 22
a list of money values (see payment card on Appendix). After the first valuation
exercise, we showed the appropriate follow up questions to verify the reason of their
answers. The aim of these questions were to identify protesters and true zero bid and
further to separate the first valuation exercise to the second.
The last section of the questionnaire was dedicated to the collection of the socioeconomic information such as age, income, etc.
The computerized system automatically validated the data entries minimizing the number
of mistakes and missing data. The questionnaire was administered face to face with the
support of a laptop. Interviewers have been carefully trained to use the system and to
follow the survey guidelines defined for our case study.
3.2 . Sampling procedure and response rate
One of the novel approach of the UK Aquamoney project was to explicitly consider the
treatment of the natural resources location. To date the issue of location has virtually
been ignored within stated preference studies 4 . Yet a large body of research (notably
travel cost studies) shows that location is a highly important driver of values for natural
resource change. This shows that the value of improvements to the on-site attributes of
a water body may decline with increasing distance to that site from a survey
respondents’ home location. In short, there is evidence to support the common-sense
presumption that individuals care more about improvements to local, as opposed to
distant, resources. Failure to account for such a proximity effect may result in the value
of improvements being incorrectly estimated. In contrast, capturing the nature and form
4
Exceptions to this are reviewed in Bateman et al., (2006) which also presents a case study illustrating
the major impact which location can have upon natural resource benefit values.
CSERGE Team
Page 23
of such effects may assist in decision making if authorities wish to target improvements
so as to maximise the utility they generate.
The sampling procedure is therefore a crucial decision for many studies given that the
sampling biases can produce misleading conclusions (Champ et al. 2003, Bateman et
al. 2002, Edwards and Anderson 1987). Therefore, in our study to fully capture the
distance-decay effect and to minimize the sampling biases we follow a well targeted
sampling strategy, called efficient sampling approach.
The sampling strategy has been designed to maximize the statistical information of the
model to be estimated. In the following we briefly describe our approach.
We assume that the individual preference function for the water improvement is
governed by the following function:
y i = f (θ , X i )
(1)
where yi represents the individual WTP to the site characteristics (X) and θ is a vector
of unknown parameters. The site characteristics can be represented by the quantity or
quality level of the site improved and its distance to the respondents’ home. Specifying
eq. 1 in the simplest way we can rewrite yi as:
f (θ , X i ) = I jα j + β d j + ε i
(2)
where the error terms are iid N(0,σ2), I is a dummy variable which describes the
characteristics of the selected site, dj is the individual distance to the improved site j
and θ=(α,β) is the vector of unknown parameters.
CSERGE Team
Page 24
The aim of the CV studies is in general to recover the parameter of the model in order
to identify the relationship between the dependent variable and the independent
variables.
Determination of the optimal sampling strategy is a key initial task within any CV
study. Following standard sampling theory we might choose the random sample or the
stratified sample. The first strategy randomly selects a sample of respondents to be
interviewed. The second divides the target population according to the distance and then
randomly samples inside each stratum. Alternatively, we can design an efficient sample
taking into account that the selection of respondents directly influence the reliability of the
results, as to say the efficiency of the estimates defined by the covariance matrix of the
model specified. Therefore given the covariance matrix of a generic CV model, defined
as:
⎛ ∂ log l ( y; d , θ ) ∂ log l ( y; d ,θ ) ' ⎞
Ωθ ( d ) = E ⎜
⎟
∂θ
∂θ
⎝
⎠
we can define a criteria of efficiency as discussed in Federov (1972). The best known
measure of efficiency is the D-error index which defines the distribution of sample, i.e. the
combination of (dj), for which the determinant of the Fisher information matrix is the
maximum.
Assuming a linear specification for the CV model and given a specific sample size, we can
demonstrate that the D-error index suggests to select only two sampling zones: the nearest
location and the farthest location. We acknowledge that this sampling strategy relies on a
strong assumption regarding the specification of the CV model to estimate the survey data,
because rarely the researcher knows perfectly the model that he is going to specify with
CSERGE Team
Page 25
the data that he should still collect. Therefore to overcome this limitation, we test many
different sampling strategies performing a Monte Carlo simulation where we compared
the sampling efficiency according to the D-error index. The best sampling distribution, in
term of estimation efficiency has been saved and used as the Aquamoney sampling
strategy.
Technically, we identified about 30 possible “sampling sites” (each one corresponding to a
part of a town, a small village, etc.) within the sampling area, each of them with enough
population to allow one day or more of interviews. The sampling sites (SS) have been
identified in order to be scattered across the entire sampling area, and therefore to
maximize the spatial variation, and are represented in Figure 5.
The optimal number of observations from each SS will be derived using a Monte Carlo
simulation based on some prior parameters values. We have randomly simulated different
sampling strategies among the SS and seen which one gave the best parameters estimates
in terms of the determinant of the covariance matrix (the D-error index). Fig. 5 and Tab. 2
report the sampling allocation of the 1,000 respondents of the survey.
CSERGE Team
Page 26
Table 2 The sample distribution
Site No.
Site name
Respondent No.
1
Skipton
80
2
Addingham
80
3
Ilkley
4
Harrogate
5
Knaresborough
0
6
Earby
0
7
Keighley
80
8
Baildon
60
9
Yeadon
0
10
Bramhope
40
11
Collingham
80
12
Trawden
0
13
Haworth
0
14
Shipley
60
15
Pudsey
40
16
Leeds (Manston)
60
17
Hebden Bridge
20
18
Halifax
20
19
Brighouse
20
20
Dewsbury
60
21
Wakefield
40
22
Rochdale
0
23
Marsden
60
24
Huddersfield
60
25
Clayton West & Skelmanthorpe
26
Royston
0
60
0
80
CSERGE Team
Page 27
Figure 5: Survey area showing sample sites (orange background gives site
numbers from Table 2) and corresponding respondent numbers (yellow
background)
The data collection took place from the 18th August 2008 until the 20th September 2008.
Interviewers approached respondents knowing on the door of the selected streets and the
respondents could be interviewed in their home or in some city public places.
The face-to-face approach gave us a good response rate but precise information of it will
be available as soon as the interviewers’ note, recorded outside the computerised system,
will be analyzed.
CSERGE Team
Page 28
4.
VALUATION RESULTS
4.1 . Respondent characteristics & sample representative ness
In total the data collected for the Aquamoney project are 1,000 but only 439 have been
collected using the payment card format. Therefore we are focusing our analysis only on
this sub sample.
Interview durations ranged from 15 to 59 minutes with 24 minutes the mean interview
time and 6 the standard deviation.
4.1.1. Demographic characteristics
Of the 439 respondents interviewed 239 (54%) were female and 200 (46%) were male.
The mean age of respondents was 50.1 with a standard deviation (SD) of 18.7. The age
distribution is reported in table 3 and we can see that, except the class between 25 and 30
years old, all the other categories are well represented in the sample.
Table 3 The age distribution
Age Categories
%
18-24
10%
25-30
8%
31-40
19%
41-50
15%
51-60
16%
61-70
16%
Over 70
17%
CSERGE Team
Page 29
The household composition is reported in Table 4 and the most represent category is given
by family with two members.
Table 4 The household composition
Household
members
%
1
16%
2
38%
3
17%
4
18%
5
6%
More than 5
5%
4.1.2. Socio-economic characteristics
The high proportion of retired respondents (31%) is coherent with the technique used to
recruit respondents. In fact, the door-to-door technique increases the chance to recruit
people with spend the majority of their time at home, such as retired. However, to balance
the proportion of retired the survey took place including some weekends to ensure
inclusion of working people. The second well represented category is the employed fulltime (24%), follow by the employed part-time (13%) and the self-employed (9%).
Around 8% of respondents were responsible to look after the home/children and 5% were
students. The description of the job distribution is reported in Tab. 5.
CSERGE Team
Page 30
Table 5 The job distribution
Employment status
%
Self employed
9%
Employed full-time (30+ hrs)
24%
Employed part-time (<30 hrs)
13%
Student
5%
Unemployed - seeking employment
4%
Unemployed – other
1%
Look after the home / children
8%
Retired
31%
Unable to work due to sickness or
disability
3%
Other
3%
Household gross income distribution is reported in Tab. 6 and the most represented
category is between £12,000 and £18,000. Considering the mean value per each category 5
rescaled by the annual taxes, we get that the average household net income is around
£19.500 with a standard deviation of £11,000. The minimum level of income in the
sample is £4,500 and the maximum is £50,000.
5
For the income last category we assume a maximum value of £78000 gross annual income.
CSERGE Team
Page 31
Table 6 The gross household annual income distribution
Category
%
< £6000
8%
£6001 - £12,000
11%
£12,001 - £18,000
14%
£18.001 - £24,000
12%
£24,001 - £30,000
4%
£30,001 - £36,000
6%
£36,001 - £42,000
7%
£42,001 - £48,000
4%
£48,001 - £54,000
2%
£54,001 - £60,000
3%
£60,001 - £66,000
1%
£66,001 - £72,000
0.5%
> £72,001
3%
Don't know
7%
Refused
18%
4.1.3. Water use characteristics
Analysing the recreation behaviour to water bodies, we report in Tab.7 river sites visits
distribution and in Tab. 8 lake sites visits distribution. The majority of respondents made
from 1 to 5 visits to a river or lake site in the last 12 months. The proportion of non-users
is higher for lakes than rivers.
CSERGE Team
Page 32
Table 7 The river sites visits distribution over the last 12 months
River frequency
%
0
27%
1-5
34%
5-10
7%
11-30
18%
31-50
2%
51-100
5%
more than 100
7%
Table 8 The river sites visits distribution
Lakes frequency
%
0
37%
1-5
45%
5-10
7%
11-30
8%
31-50
0%
51-100
3%
more than 100
1%
Comparing the water bodies visits with the total recreation occasions in one year of time
we obtain that for the majority of times respondents selected a non water bodies visit.
However, the proportion of water bodies visits over the total number of recreation visits is
0.45 with a standard deviation of 0.33.
CSERGE Team
Page 33
Most respondents stated that they visited the water bodies for walking/rambling.
However, dog walking, picnicking, feeding birds, cycling and motorised water sports are
also mentioned. In particular, 28% of respondents own a canoe, surfboard etc. and 5% of
respondents have got a fishing license
4.2 Public perception of water management problems
The survey was designed to avoid any specific reference to the WFD in order to avoid
protest. Therefore the perception of water management can be obtained analysing some
side questions included in the questionnaire. Firstly, the current water quality level
depicted in the local map (screen shot 7 in the Appendix) has been compared with the
perceive quality by each respondent. Fig. 6 reports the proportion of answers for the
following question: “which is your reaction to the information concerning the current
quality of rivers in the area shown on the map?”. The majority of respondents agreed
with the quality represented where the water quality of one of three rivers is very good,
one if fair and one is good.
60%
50%
40%
30%
20%
10%
0%
Much better
Little better
Same
Little worse
Much worse
Figure 6 Answers proportion for the question “which is your reaction to the information concerning
the current quality of rivers in the area shown on the map?”
CSERGE Team
Page 34
Subsequently, talking about the improvement envisaged in the CV task, we asked
respondents “how likely do you think it is that the improvements in water quality
described would actually happen?”. Fig. 7 shows that the majority of respondents believe
in the water quality improvement, although a significant proportion of respondents
consider the improvement unlike.
45%
40%
35%
30%
25%
20%
15%
10%
5%
0%
Very likely
Somewhat
likely
Neutral
Somewhat
unlikely
Very unlikely Don’t know
Figure 7 Answers proportion for the question “how likely do you think it is that the improvements in
water quality described would actually happen?”
Finally, as a measure of consideration of the water management problems, we report the
proportion of protesters in the CV task that is roughly 3%.
4.3 .
Estimated
management
economic
values
for
water
resource
The individual willingness to pay for water quality improvement has been calculated
taking into account the analysed CV anomalies. Therefore in Tab. 9 we report the mean
and median WTP in different cases. In the first rows, we report the WTP values for the
CSERGE Team
Page 35
small and large river improvement. We obtain that the WTP for the small good is smaller
than the value for the large good. However, considering the order and scope effect
together we get different valuations driven by anomalies.
Table 9 WTP (€ PPP) descriptive statistics per country
€ (PPP)
(st.dev)
Average WTP- Small WTP (A1)
20 (29)
10
Median WTP-Small
Average WTP- Large WTP (A2)
26 (37)
12
Median WTP- Large
Average WTP- Question 1-Small WTP (
)
23 (33)
10
Median WTP- Question 1-Small
Average WTP- Question 2-Small WTP (
)
16 (25)
10
Median WTP- Question 2-Small
Average WTP- Question 1-Large WTP (
)
12
Median WTP- Question 1-Large
Average WTP- Question 2-Large WTP (
25 (32)
)
Median WTP2- Question 2-Large
27 (41)
12
Notes: WTP values recalculated based on Purchasing Power Parity indicators
(World Bank, 2007). Protest bids are excluded in the estimation of WTP statistics.
The results presented in table 9 suggest the influence of scope and order effect on the
WTP values. We expect respondents to give a higher or equal value to the large than/as
to the small improvement. Considering the behaviour of respondents respect to order
and scope effects we report in table 10 the proportion of sample respect to the
relationship between small water improvement in the first question (
), large water
improvement in the first question ( ), small water improvement in the second question
( ) and large water improvement in the second question (
). The most interesting
insight from this table is that the majority of the sample gives an equal WTP response
for the small and large improvement. These respondents are either not sensitive to the
CSERGE Team
Page 36
scope of the change, or they perceive the two water bodies changing in the large
improvement scenario as perfect substitutes.
Table 10 Response behaviour in valuing water quality improvement with order and scope effects
UK
WTP (
) < WTP (
)
14%
WTP (
) > WTP (
)
21%
= WTP ( )
37%
= WTP ( ) =0
25%
WTP (
WTP (
Finally, we compare WTP values addressing ordering and scope effects simultaneously.
More specifically, we compare the Bottom Up (1st small, 2nd large) with the Top Down
(1st large, 2nd small) approach. We reject the hypothesis that these approaches result in
the same size of scope effect. In the BU approach differences are smaller than in the top
down approach. Figure 8 graphically displays the two approaches. In the BU approach,
respondents increase their WTP little, compared to the decrease in WTP in the TD
approach. These results are coherent with the findings in Bateman et al. (2004).
30
BU
25
20
TD
15
10
5
0
Small
Large
Figure 8 Sensitivity to scope of bottom up and top down approach
CSERGE Team
Page 37
4.4 Factors explaining economic values for water resource
management
To analyze the CV data and the impact of socio-economic factors and the scope and order
effect we specify a multivariate regression analysis. Excluding the protesters and
modelling the possibility to have a zero-bid, we specify a Tobit model. Since each
respondent gave two WTP answers, we model the data with a panel structure estimating a
random effects Tobit panel data model. The WTP function is defined as:
WTPijt = f ( ΔQijt , Dijt , Dikt , Dict , O, Yi ) + ε ijt + αi
where WTPijt is individual i’s willingness-to-pay for a water quality change in water body
j in time period t, ΔQijt is the quantity change of the improvement, Dijt is the distance from
the respondent’s residence to water body j, Dikt is the distance from the respondent to the
substitute water body, Dict is the distance from the respondent to the nearest site on the
coast, O is a vector of dummies which take into account the scope and order effects and Yi
is a vector of socio-economic and demographic characteristics of individual i. Finally αi
and εijt are the two error terms, the former is a time invariant individual effect error and the
other is a time-varying error term. The unobserved effect αi is assumed to be uncorrelated
with each explanatory variable, formally written as:
Cov(X,ai)= 0 for each time period t and variable 1…k
When the unobserved effects (αi) are large, the random effects estimates will be similar to
the fixed effects model. When the unobserved effects are small, relative to the variance of
εijt, the random effects estimates will be closer to a pooled model. The results of the
model are presented in table 10.
CSERGE Team
Page 38
Table 11 Results of the Random effects Tobit panel data model
Variables
Explanation
Coefficient
Constant
Respondent characteristics
Income
Net household income in ppp € per
year
Gender
Gender of respondent (1 if female)
Age
Age of respondent in years
Distance
to Distance to the site that changes in
site
quality
Distance
to Distance to the nearest water
substitute
substitute
Distance
to Distance to the nearest coast site
coast
Scope
and
order effects
Large1
= 1 if respondent values the large
improvement in the first WTP
question
Small2
= 1 if the respondent values the
small improvement in the second
WTP question
Large2
= 1 if the respondent values the
large improvement in the second
WTP question
Std.
pErr value
5.28
1.21
6.40
10.07
0.001 0.00001
1.97
1.24
-0.076
0.034
1.59
-2.25
-0.274
0.184
0.071
0.151
-3.82
1.22
0.122
0.060
2.03
6.251
1.548
4.04
-3.83
1.536
-2.50
6.643
1.359
4.89
Sigma α
Sigma ε
rho 6
34.70
10.87
0.91
0.762
0.372
0.007
45.57
29.19
LogLikelihood
Wald chi2(9):
-2361
236
157 observations censored; 457 uncensored
The total model and almost all parameters are significant. The signs of the sociodemographic characteristics of the respondents are conform to theoretical expectations.
Income has a positive effect on the WTP. The WTP decreases with age and gender is
not significant. We find a significant distance-decay effect. The further away a
respondent lives from the site, the lower the WTP. The closer to a substitute site (river
site or coast) the respondent lives, the less he is willing to pay for the site under
valuation. Although, the distance to river site substitute is not significant.
6
Rho= var_ α /(var_ α +var_ ε)
CSERGE Team
Page 39
The sigmas represent the variances of the two error terms α and ε .Their relationship is
described by the variable rho. If this variable is zero, the panel-level variance
component is irrelevant, and the estimator is not different from the pooled estimator.
The results show that in this case, the panel data structure of the WTP answers has to be
taken into account.
The dummies for order and size are highly significant confirming the CV biases as in
Bateman et al. (2004). The dummy Large1 is significant, which implies that the first
WTP value stated by respondents is sensitive to the size of the improvement. When the
small improvement is presented in the second question (Small2), WTP for the small
improvement decreases significantly compared to the baseline. When the large
improvement is valued secondly, respondents are willing to pay significantly more than
for the small improvement in the first question.
4.5 . Total Economic Value
The estimate of the total economic value relies on the aggregation of individual WTP.
This procedure is relatively straightforward is some conditions are satisfied (Bateman
et al. 2002). Some of these conditions rely on sample selection and population and
point out the importance to take into account the response rate and the sample
distribution. Therefore to proceed with a correct aggregation procedure we should wait
for the response rate per sampling area.
CSERGE Team
Page 40
5.
CONCLUSIONS
This project is highly interdisciplinary and policy orientated but also addresses current
frontier issues in the literature on non-market benefits analysis, and its spatial aspects.
The valuation results point out the impact of scope and order effects on the WTP values
underlying the distance-decay effect.
Subsequent analysis will compare these findings with those obtained from common
design studies carried out by Aquamoney partners in other EU countries.
CSERGE Team
Page 41
6.
BEST PRACTICE RECOMMENDATIONS
The study of water quality improvements suggests different aspect to be taken into
account to produce reliable benefits value. Firstly, we verify that the valuation of
improvements differs from users and non users and including vaguely the location of
improvement we end up with overvalued benefits. Therefore, a clear spatial definition
of the good is fundamental in order to capture the distance decay effect. This effect
suggests to carefully define the population of interest and the sampling strategy. A
simply, random selection of the sample can produce under representation of spatial
distribution of respondents and bias the estimate. Therefore, further research can be
useful to understand the best sampling strategy for spatially distributed goods.
The second aspect to be consider in valuing water improvement is objective of
valuation results. The finding of scope and order effects suggests that more effort needs
to me made to encourage respondents to attend to the quantitative information.
CSERGE Team
Page 42
7.
REFERENCES
Barton, D. (2002) The transferability of benefit transfer: contingent valuation of water
quality improvements in Costa Rica, Ecological Economics, 42(1-2): 147-164
Bateman, I.J., Langford, I.H. and Nishikawa, N. and Lake, I. (2000) The Axford debate
revisited: A case study illustrating different approaches to the aggregation of
benefits data, Journal of Environmental Planning and Management, 43(2): 291302.
Bateman, I.J., Cooper, P., Georgiou, S. and Poe, G.L. (2001) Visible choice sets and
scope sensitivity: an experiment and field test of the study design effects upon
nested contingent valuation, CSERGE Working Paper EDM-2001-01, Centre for
Social and Economic Research on the Global Environment, University of East
Anglia.
Bateman I.J., Cole, M., Cooper P., Georgiou S., Hadley D., Poe G.L. (2004) On visible
choice sets and scope sensitivity, Journal of Environmental Economics and
Management, 47: 71-93.
Bateman, I. J., Cole, M. A., Georgiou, S., and Hadley, D. J. (2006a) Comparing
contingent valuation and contingent ranking: A case study considering the
benefits of urban river water quality improvements, Journal of Environmental
Management, 79: 221-231.
Bateman, I.J., Brouwer, R., Davies, H., Day, B.H., Deflandre, A., Di Falco, S.,
Georgiou, S., Hadley, D., Hutchins, M., Jones, A.P., Kay, D., Leeks, G., Lewis,
M., Lovett, A.A., Neal, C., Posen, P., Rigby, D. and Turner, R.K. (2006c)
Analysing the Agricultural Costs and Non-market Benefits of Implementing the
Water Framework Directive, Journal of Agricultural Economics, vol. 57: 221–237
Bateman, I.J., Carson, R.T., Day, B., Hanemann, W.M., Hanley, N., Hett, T., JonesLee, M., Loomes, G., Mourato, S., Özdemiroğlu, E., Pearce, D.W., Sugden, R.
and Swanson, J. (2002) Economic Valuation with Stated Preference Techniques:
A Manual, Edward Elgar Publishing, Cheltenham.
Bateman, I.J., Day, B.H., Georgiou, S. and Lake, I. (2006b) The aggregation of
environmental benefit values: Welfare measures, distance decay and total WTP,
Ecological Economics, 60(2): 450-460.
Birol, E., Karousakis, K., Koundouri, P., (2006) Using economic valuation techniques
to inform water resources management: A survey and critical appraisal of
available techniques and an application, Science of the Total Environment, 365:
105-122.
Brouwer, R. (2000) Environmental value transfer: state of the art and future prospects.
Ecological Economics, 32: 137-152.
CSERGE Team
Page 43
Brouwer, R. and Bateman, I.J. (2005): The temporal stability and transferability of
models of willingness to pay for flood control and wetland conservation, Water
Resource Research, in press.
Carson, R. T. and Mitchell, R.C. (1993) The Value of Clean Water: The Public's
Willingness to Pay for Boatable, Fishable and Swimmable Quality Water, Water
Resources Research, 29(7): 2445-2454.
Carson, R.T., Flores, N.E., Hanemann, W.M., (1998) Sequencing and valuing public
goods. Journal of Environmental Economics and Management, 3: 314-323.
Carson, R.T., Flores, N.E., Hanemann, W.M., (1998) Sequencing and valuing public
goods. Journal of Environmental Economics and Management, 3: 314-323.
Carson, Richard T., (2007) Contingent Valuation: A Comprehensive Bibliography and
History. Northampton, MA: Edward Elgar.
Champ, P.A., Boyle, K. and Brown, T.C. (eds.) (2003) A Primer on Non-market
Valuation, The Economics of Non-Market Goods and Services: Volume 3,
Kluwer Academic Press, Dordrecht.
Clark J., Friesen L., (2006) The Causes of Order Effects in Contingent Valuation
Surveys: An Experimental Investigation," Working Papers in Economics 06/06,
University of Canterbury, Department of Economics.
Desvousges, W.H., Smith, V.K. and Fisher, A. (1987) Option price estimates for water
quality improvements: a contingent valuation study of the Monongahela River,
Journal of Environmental Economics and Management, 14: 248-267.
Edwards A. and Anderson G., (1987) Overlooked Biases in Contingent Valuation
Surveys: Some Considerations Land Economics, 63(2): 168-178 .
ENDS (1998) Water abstraction decision deals savage blow to cost-befit analysis, 278:
16-18.
ENDS (2008) Getting to grips with the Water Framework Directive, Services, t
http://www.endsdirectory.com/index.cfm?action=articles.view&articleID=200402
,
Environment Agency (2004) Water Framework Directive Characterisation Atlas –
Pressures and Impacts Assessment Review, Environment Agency, available from
http://www.environment-agency.gov.uk/business/
Environment Agency (2007) River water quality standards, available at
http://www.environment-agency.gov.uk/subjects/fish/569882/?version=1&lang=_e
accessed, November 2007
Federov V.V. (1972) Theory of optimal experiments. Academic Press, New York and
London.
Hanley, N., Wright, R.E. and Alvarez-Farizo, B. (2006) Estimating the economic value
of the improvements in river ecology using choice experiments: an application to
CSERGE Team
Page 44
the water framework directive, Journal of Environmental Management, 78: 183193.
Hibbard J.H., and Peters E., (2003) Supporting informed consumer health care
decisions: Data Presentation Approaches that Facilitate the Use of Information in
Choice. Annual Review of Public Health, 24: 413-433.
Hibbard JH, Slovic P, Peters E, Finucane ML. (2002) Strategies for reporting health
plan performance information to consumers: evidence from controlled studies.
Health Serv. Res. 37: 291-313.
Hime S. and Bateman I.J, (2008) A transferable water quality ladder for conveying use
and ecological information within public surveys. Forthcoming as CSERGE
Working Paper, Centre for Social and Economic Research on the Global
Environment, University of East Anglia.
Holmes, N., Boon, P. and Rowell, T. (1999) Vegetation communities of British rivers.
Joint Nature Conservation Committee
Lipkus I.M., and Hollands, J.G. (1999) The visual communication of risk. Journal of
the National Cancer Institute Monographs, 25: 149-163
Mitchell, R. C., and Richard T. C. (1989) Using Surveys to Value Public Goods: The
Contingent Valuation Method. Washington: Resources for the Future, 1989
Oguchi, T., Jarvie, H.P. and Neal, C. (2000) River water quality in the Humber
Catchment: An introduction using GIS-based mapping and analysis. The Science
of the Total Environment, 251/252: 9-26
Plott, C.R., (1996) Rational individual behaviour in market and social choice processes:
the discovered preference hypothesis. In: Arrow, H.J., Colombatto, E., Perlman,
M., Schmidt, C. (Eds.). The Rational Foundations of Economic Behaviour,
International Economic Association, Macmillan, London, St. Martin’s, New
York, 225-250
Powe, N. A. & Bateman, I. J., (2003) "Ordering effects in nested 'top-down' and
'bottom-up' contingent valuation designs," Ecological Economics, 45(2): 255-270
Ready, R.C. S. Navrud, B. Day, R. Dubourg, F. Machado, S. Mourato, F. Spanninks
and M.X.V. Rodriquez (2004) Benefit Transfer in Europe. How Reliable Are
Transfers Between Countries? Environmental and Resource Economics, 29: 6782
RPA (2004) Water Framework Directive – Indicative Costs of Agricultural Measures,
RPA Consultants, Loddon, Norfolk.
UKTAG, 2008. UK Environmental Standards and Conditions (Phase 1) Final Report.
Vaughan, W. J. (1986). The RFF Water Quality Ladder, Appendix B in Robert
Cameron Mitchell and Richard T. Carson, The Use of Contingent Valuation Data
CSERGE Team
Page 45
for Benefit/Cost Analysis in Water Pollution Control, Final Report. Washington,
D.C.: Resources For the Future.
CSERGE Team
Page 46
APPENDIX
THE UK COMPUTERISED AQUAMONEY QUESTIONNAIRE
[ This questionnaire and all elements therein is copyright protected.
For reproduction and usage permission contact:
Ian Bateman, CSERGE, UEA. Norwich, UK.
Email: i.bateman@uea.ac.uk ]
Good morning/afternoon/evening. My name is [INTERVIEWER NAME] from the
[THE UNIVERSITY OF EAST ANGLIA] and I am carrying out research on a local
issue regarding this area. The questionnaire is anonymous and all answers you
give will be treated in confidence. The interview will take approximately 20
minutes.
[Click the WQSystem shortcut on the desktop, this will take you to the WQ
system’s starting page. ]
SECTION 1 : INTERVIEW & INTERVIEWEE DETAILS
1.1
1.2
1.3
1.4
1.5
1.6
1.7
INTERVIEWER ID:
LOCATION OF INTERVIEW:
DATE: collected automatically
START TIME: collected automatically
HOME POSTCODE (if no postcode enter XXX XXX)
TOWN (you must enter a town)
ROAD NAME (only if post code is not supplied)
CSERGE Team
Page 47
Screen shot 1:Initial Screen
1.8
Can you tell me how long you have lived either at this [YEARS] [MONTHS]
address or an address in the surrounding mile or so?
1.9
If less than 1 years can you tell me your previous address
[TOWN] {COUNTY]
(AT LEAST THIS)
The main focus of this survey is outdoor recreation. We want to get a balanced
picture and are just as interested in talking to people who don’t like outdoor
recreation as those that do. Looking at the following categories [IN 1ST DROP
DOWN ON SCREEN 2], which best describes how often, over the past 12 months,
you have been on trips to any type of outdoor recreation site, including the
countryside, parks, forests, rivers, lakes, the seaside, etc.
[RECORD RESPONSE] on the total outdoor trips drop down menu on screen 2.
That would imply that you made about to outdoor recreation sites in total over
the past year.
[RECORD RESPONSE] in total outdoor trips text box on screen 2
B3. Looking again at these categories [IN 2nd DROP DOWN ON SCREEN 2],
which best describes how often, over the past 12 months; you have been on trips to
rivers or riverside sites. [RECORD RESPONSE] on the total river trips drop down
menu on screen 2
CSERGE Team
Page 48
That would imply that you made about [READ LABEL FROM SCREEN], to rivers
in total over the past year.
[RECORD RESPONSE] in the text box with the label Total River Trips on screen 2.
Looking again at these categories [IN 2nd DROP DOWN ON SCREEN 2], which
best describes how often, over the past 12 months; you have been on trips to canals
or canal side sites. [RECORD RESPONSE] on the total river trips drop down menu
on screen 2
That would imply that you made about [READ LABEL FROM SCREEN], to rivers
in total over the past year.
[RECORD RESPONSE] in the text box with the label Total canal Trips on screen 2.
Looking again at these categories [IN 3rd DROP DOWN ON SCREEN 2], which
best describes how often, over the past 12 months; you have been on trips to lakes
or lakeside sites. [RECORD RESPONSE] on the total lake trips drop down menu on
screen 2
That would imply that you made about [READ LABEL FROM SCREEN], to lakes
in total over the past year.
Looking at the following categories [IN 4th DROP DOWN ON SCREEN 2], which
best describes how often, over the past 12 months, you have been on trips to any
other type of outdoor recreation site, including the countryside, parks, forests, the
seaside, etc.
[RECORD RESPONSE] on the total other outdoor trips drop down menu on screen 2.
That would imply that you made about [READ LABEL FROM SCREEN] to
outdoor recreation sites in total over the past year.
You said that you took about [Number from screen 2 –River TRIPS text box] trips
to rivers each year. How many were to sites on this map?
CSERGE Team
Page 49
Screen shot 2: Previous Address and trip information
In this survey we will be talking about possible changes to the quality of river
water. These changes would not affect any other aspect of water services such as
drinking water quality. The only things affected by changes in river water quality
are the plants and animals that live there and the types and quality of recreation
that visitors can enjoy.
In these next questions we are going to use these pictures [SHOW WATER
QUALITY LADDER] to indicate different river quality. As you can see the
pictures are arranged from higher to lower water quality.
CSERGE Team
Page 50
CSERGE Team
Page 51
[INDICATE THE HIGHEST QUALITY PICTURE - BLUE]
This picture marked with the blue circle, shows the highest river quality. Please
take a moment to consider this.
[PAUSE FOR 10 SECONDS]
The symbols here (indicate all symbols) show that a river of this quality is suitable
for game fish [POINT TO
] such as salmon and trout, which are sensitive to
pollution; coarse fish such as carp and chub [POINT TO
2ND SYMBOL], which
are less sensitive to pollution. It is also suitable for swimming [POINT TO
],
boating [
], bird watching and enjoying nature. The picture also shows the
plant life in and around the river you would expect to see and shows quite clear
water.
In the second picture marked with a green circle [INDICATE THE 2ND PICTURE GREEN] you can see that there has been a change with the pollution sensitive fish
rarely found; different plants in and around the river and water of slightly
different clarity. However, there are still coarse fish, the opportunity for
swimming boating, bird watching and enjoying nature [POINT TO SYMBOLS].
In the third picture marked with a yellow circle [INDICATE THE 3RD PICTURE YELLOW], once again there are very rarely pollution sensitive fish species and the
number of coarse fish also becomes lower [POINT TO SYMBOLS]. In addition to
this you can no longer swim in the water; the water plants have started to be
replaced with algae and the water is of lower clarity. However, you can still go
boating and see birds.
This final picture shows the river water of the lowest quality and is marked red
[INDICATE THE 4TH PICTURE - RED]. Here there are virtually no fish, no
opportunities for swimming or boating and the number of common birds has also
decreased [POINT TO SYMBOLS]. In addition to this the plants both in and out
of the water have changed and the water is of lower clarity [INDICATE THE 4TH
PICTURE - RED].
GIVE THE WATER LADDER TO THE RESPONDENT
We will be using these pictures throughout the interview so please take your time
to familiarise yourself with them, [PAUSE], then continue.
These next questions concern the quality of rivers in your area.
Please take a look at the following map of the area. Can you point to where your
home is? – All respondents must identify their home – even if they do not visit rivers!
CSERGE Team
Page 52
Screen shot 3: Marking home
This is the home square
B4a). Can you point to the site (on a river) that you visited most often?
Screen shot 4: Site one show the site visited in the past 12 months
CSERGE Team
Page 53
B4b). Looking at the water quality ladder which colour best describes the quality
of the site.
B4c). How often did you visit this site in the last 12 months?
B4d). What was the purpose of most of your visits [SHOWCARD 1]?
SHOWCARD1
A
Water sports
B
Dog walking
C
Angling
D
Rambling
E
Running
F
Cycling
G
Climbing
H
Feeding birds
I
Picnic
J
Motorised recreation
K
Wildlife watching
L
Other
Screen shot 5: Details of site visits
B5). We are going to mark all of the sites that you visited in this area in the last 12
months, in exactly the same way as before. Did you visit any other sites in this
area over the last 12 months?
IF YES
LOOP questions B4 - UP TO A MAXIMUM OF TEN SITES
CSERGE Team
Page 54
IF NO GO TO the NEXT SECTION
CONTINGENT VALUATION
I now want to ask a different type of question. Please look at this map you
live here [POINT TO MAP].
[SHOW MAP OF CURRENT SITUATION AREA – on screen]
We have used the colours from the pictures to show the current quality of
rivers in this area. [This is based on information from the Environment
Agency, which is the official body that monitors river quality in the UK].
As you can see, at present the river closest to you [INDICATE RIVER
CLOSEST TO RESPONDENTS TOWN], the [GIVE NAME] which is
[STATE COLOUR OF THAT RIVER]. That means that on average its
quality is like this [POINT TO PICTURE CORRESPONDING TO
COLOUR]. This is the river [INDICATE AND GIVE NAME OF THE
RIVER 2ND CLOSEST TO RESPONDENTS TOWN] which is [STATE
COLOUR OF THAT RIVER] and on average its quality is like this [POINT
TO PICTURE CORRESPONDING TO COLOUR].
Finally, at present the River [INDICATE AND GIVE NAME OF THE
RIVER 3RD CLOSEST TO RESPONDENTS TOWN] is coloured [STATE
COLOUR OF THAT RIVER]. That means that, on average, its quality is
like this [POINT TO PICTURE CORRESPONDING TO COLOUR].
E1. Looking at the categories on screen, which phrase best, describes your
reaction to the information concerning the general current quality of
rivers in the area[SHOWCARD 2]?
SHOWCARD 2
1 Much better than expected
2 A little better than expected
3 About the same as expected
4 A little worse than expected
5 Much worse than expected
CSERGE Team
Page 55
Screen shot 6: Current water quality levels
[THERE ARE TWO VERSIONS OF THIS QUESTIONNAIRE – THEY DIFFER IN
THE ORDERING OF PRESENTATION OF QUESTION A AND B]
I now want to show you a second map [POINT TO “ALTERNATIVE”
MAP]. This shows an alternative situation, where river water treatment
works are undertaken to improve the stretch of river shown here
[INDICATE CHANGED STRETCH]. Comparing the two maps you can see
that in this stretch the river water quality has improved from YELLOW to
BLUE.
[INDICATE WATER QUALITY LADDER] We can see that’s a move from here
[INDICATE INITIAL QUALITY] to here [INDICATE FINAL QUALITY]. All
other parts of all the rivers stay as they currently are.
Before asking you whether your household would want this alternative
situation or not, please consider the following:
•
Any money you spend on improving river water quality obviously
would not be available for spending on any other purchases.
•
Please think about the location of the improvement, how close it is
to your home, and whether you would benefit from it.
DICTOMOUS CHOICE
E2. So which would you prefer [INDICATE] the current situation where
your water bill stays the same and there is no change in river water
CSERGE Team
Page 56
quality or [INDICATE] the alternative situation where your water bill
increase by [LABEL] and river water quality improves as shown.
CSERGE Team
Page 57
A
Screen shot 7: This example shows one stretch in the alternative with a
bid level
PAYMENT CARD
PAYMENT CARD – A TEXT BOX APEARS ON THE SCREEN UNDER
THE ALTERNATE MAP
To help you work out how much, if anything, this scheme is worth to your
household please consider this card.
[SHOW PAYMENT CARD]
E2. For each amount please ask yourself whether or not your household
would be prepared to pay this amount each year to get the improvement
shown. Then tell me the amount which is the most your household would be
prepared to pay on top of your normal water bill in order to get this
improvement.
E3. Looking at this list, please tell me the two most important reasons for
your answer, press NEXT
[SHOW PAYMENT CARD]
-----00000----- SHOWCARD: PAYMENT CARD -----00000-----
CSERGE Team
Page 58
Place a tick next to the amount which is the most your household would be prepared to
pay each year for this water quality improvement.
€0
€30
€65
€100
€135
€190
€350
€700
€1050
€3
€35
€70
€105
€140
€200
€400
€750
€1100
€5
€40
€75
€110
€145
€225
€450
€800
€1150
€10
€45
€80
€115
€150
€250
€500
€850
€1200
€15
€50
€85
€120
€160
€275
€550
€900
>
€1200,
namely
€…
€20
€55
€90
€125
€170
€300
€600
€950
Other:
€ …..
€25
€60
€95
€130
€180
€325
€650
€1000
Don’t
know
-----00000----- END OF SHOWCARD: PAYMENT CARD -----00000-----
CSERGE Team
Page 59
Screen shot 7: Payment card example
Looking at this list, please tell me the two most important reasons for your
answer
[FOLLOWING IS FOR PEOPLE WHO CHOSE >€0 the correct page is
automatically loaded]
Screen shot 8: Reasons for selection
[FOLLOWING IS FOR PEOPLE WHO CHOSE €0]
CSERGE Team
Page 60
[Tick one in each column] – see screen below and press NEXT
Screen shot 9: reasons for selection
Now I would like you to consider a second alternative.
Again this concerns an improvement from [INDICATE] YELLOW to
BLUE quality but now for this stretch of the river.
SECOND QUESTION FROM A OR B
DICOTOMOUS CHOICE
E4. If the alternative situation required an increase of [INSERT BID
LEVEL FROM SCREEN] in your annual water bill to give the increase in
water quality shown, then would you prefer that or the current situation
where there is no change in your annual water bill but water quality stays
at its current level? SKIP PAYMENT CARD
PAYMENT CARD
To help you work out how much, if anything, this scheme is worth to your
household please consider this card.
[SHOW PAYMENT CARD – see above]
E4. As before for each amount please ask yourself whether or not your
household would be prepared to pay this amount each year to get the
improvement shown. Then tell me the amount which is the most your
household would be prepared to pay on top of your normal water bill in
order to get this improvement.
CSERGE Team
Page 61
Screen shot 7a: Payment card example
GO TO CE IF NOT COMPLETED OTHERWISE GO TO SOCIOECONOMIC
SECTION
CHOICE EXPERIMENT
Plans are being considered for improving the quality of some rivers across the UK
including some in your area. Improvements in river water quality would require
investments which would increase water bills.
The official water watchdog, OFWAT, wants to know whether individuals feel it is
worthwhile improving water quality as the necessary investments would cost
money. This cost would have to be paid for by higher water bills. All water users
would have to pay, including industry and farmers but also households because
they also contribute to water pollution. The extra amount would have to be paid
every year. This survey is your opportunity to express your views about these
plans.
I am going to show you maps of the local area in which in which we use the same
colours as before to indicate river water quality. [SHOW WATER QUALITY
LADDER AND INDICATE COLOURS]. We use red [INDICATE] to indicate the
lowest quality rivers, then yellow [INDICATE], green [INDICATE] and finally blue
[INDICATE] to show increasing levels of river quality.
[SHOW FIRST CE DUAL MAP]
CSERGE Team
Page 62
Screen shot 10: example choice question
The next few questions all show two maps showing different water qualities along
with their associated increases in annual water bills. For each question choose the
situation you prefer
GO TO CV IF NOT COMPLETED OTHERWISE GO TO SOCIOECONOMIC
SECTION
CONTROL QUESTIONS
CONTROL QUESTIONS
G1
Overall, how easy or difficult did you find it to answer the questions
involving changes in water quality and water bills (SELECT FROM
DROP DOWN ON SCREEN)?
Very easy
Fairly easy
Neither easy nor difficult
Fairly difficult
Very difficult
Don’t know
G2
1
2
3
4
5
6
How likely do you feel it is that the river quality improvements
discussed would be provided as described?(SELECT FROM
DROP DOWN ON SCREEN)
CSERGE Team
Page 63
Very likely
Somewhat likely
Neutral
Somewhat unlikely
Very unlikely
Don’t know
1
2
3
4
5
6
G3
READ STATEMENTS OFF SCREEN – I am going to read out a list of
statements. Please tell me, the extent you agree or disagree with each.
SELECT RADION BUTTONS ON SCREEN
G4
READ STATEMENTS OFF SCREEN – How important was each of the
following issues in determining your answers to the questions where you chose
between two maps?
Please tell me, the extent to which they were important when making
decisions. SELECT RADION BUTTONS ON SCREEN
[READ FROM SCREEN – AS THE ORDER WILL CHANGE]
Screen shot 11: control questions
8.
SOCIO-ECONOMIC AND DEMOGRAPHIC QUESTIONS
To finish off, I just have a few more questions about you and your household.
These will only be used for statistical purposes to see if we have interviewed a fair
range of people and please remember that all of these answers are completely
confidential.
CSERGE Team
Page 64
I1. [RECORD RESPONDENT’S GENDER]
I2
What is your age?
……years
I3
How many people are in your household …………persons
including yourself?
I4
How many people younger than 18 are in …………persons
your household?
I5. Which of these statements [SHOWCARD 3] best describes your current
employment status?
SHOWCARD 3
Self-employed
Employed full-time (30+ hrs)
Employed part-time (up to 30 hrs)
Student
Unemployed – seeking work
Unemployed – other
Looking after the home/children full-time
Retired
Unable to work due to sickness or
disability
Other
I6. Looking at this card and tell me which, if any, of these organisations you or
any others in your household are members of.
I8. Do you own a boat or canoe, surfboard, motorboat, yacht, etc? SELECT FROM
DROP DOWN ON SCREEN
I9. Do you hold a fishing licence? SELECT FROM DROP DOWN ON SCREEN
IF YES
I9a. How much does your annual license cost? RECORD IN TEXTBOX ON SCREEN
I10. Do you belong to an angling club ?
IF YES
I10a. Can you give me the name(s) of the angling/fishing club(s) you belong to?
I10b. What is your annual membership fee?
I10c. Are non-members able to fish on your club waters?
IF NO Press “next”
CSERGE Team
Page 65
I16. Looking at the categories on screen could you tell me which best
approximates your total household income before tax (select one on screen),
[IF NECESSARY, REASSURE RESPONDENT THAT ALL INFORMATION IS
COMPLETELY CONFIDENTIAL AND THIS IS THE BEST INDICATOR OF WHETHER
WE HAVE INTERVIEWED A REPRESENTATIVE RANGE OF PEOPLE]
Screen shot 12: socioeconomic questions
CLOSING
That was the last of my questions. This survey will continue for several weeks. At
the end of that time there is a possibility that my supervisor might have some
follow up questions about which he would like to call you. Could you please give
me a telephone number where you can be contacted and your first name. This will
be kept strictly confidential and will not be given to anyone else.
That's the end of the interview!
Thank you very much for your time and help, it is very
much appreciated!
CSERGE Team
Page 66