4. lecture_Collecting primary data

Collecting primary (qualitative?)
-.. techniques of data collection in which each
person is asked to respond to the same set of
questions in a predetermined
+ large samples
+ quantitative analysis
- hard to produce a good questionnaire
- only one possibility to ask the sample you are
interested in
Questionnaire design
…affects reliability, validity, response rates
• Careful design of individual questions;
• Clear layout of the questionnaire form;
• Lucid explanation of the purpose of the
• Pilot testing;
• Carefully planned and executed
When to use
• …for descriptive research – to identify /
describe the variability of different
phenomena (e.g., attitudes, opinions,
• …explanatory/analytical research - examine /
explain relationships between variables;
cause-and-effect relationships
Suitable in a multiple-methods research design
Types of questionnaires
• Self-administered questionnaires (internet, email, post, delivered by hand)
• Interviewer-administered questionnaires
(telephone; face-to face or structured
The choice of questionnaire
…depends on various factors:
• Characeristics of the respondents from whom you wish to
collect data
• Importance of reaching a particular person as respondent
• Possible contamination of respondent’s answers
- insufficient knowledge/experience - uninformed response
- socially desirable answers
- discussing answers with others
- the poor recording
• Size of sample required for analysis (likely response rate)
• Types and number of question you need to ask
Questionnaire design requirements
You should think of…
- representativeness/ accurateness of sample –>
generalising; comparing/relating results to
earlier research
- are your measurement tools compatible
- good overview of the relevant literature define the theory you wish to test
- understanding of research context (e.g.,
Designing the questionnaire
• Validity and reliability - does the questions
and answers of the questionnaire make sense
The question must be understood by the
respondent in the way intended by the
researcher and the anser given by the
respondent must be understood by the
researcher in the way intended by the
Assessing validity
• Internal validity - the ability of your questionnaire to measure what you
intend to measure
• Content validity – refers to the extent to which the measurement tool
provides adequate coverage of the questions you study. How to check:
1) definition of the research through the literature
2) discussion with some experts-supervisor
3) panel of individuals to assess whether each measurement question is
essential or not essential
• Criterion-related validity/ predictive validity – concerned with the ability
of the measurement tool to make accurate predictions. How to check:
- compare the data you gathered with that specified in the criterion in some
way (e.g., correlation)
• Construct validity – refers to the extent to which your measurement
questions actually measure the presence of those constructs you intended
them to measure – (attitude scales, ability/ personality tests). How to
- how well can you generalise from your measurement questions to your
Testing for reliability
…whether or not your tool will produce consistent findings at
different times and under different conditions (different
samples; different interviewers etc)
1) Test re-test – correlating data with those from the same
questionnaire collected under as near equivalent
conditions as possible
2) Internal consistency – correlating the responses to each
question with those to other questions in the
questionnaire (e.g., Cronbach’s alpha – we, psychologists,
all love it)
3) Alternative form – comparing responses to alternative
forms of the same question / groups of questions - check
questions (However, respondents may get mad)
Designing individual questions
• Adopt, adapt, or develop
• Open questions (open-ended questions) - allow
respondents to give answers in their own way
• Closed questions (forced-choice questions) provide a number of alternative answers from
which the respondent is instructed to choose
- quicker/ easier to answer
- easier to compare the responses
- easier to analyse
- must be clearly interpretable
Types of closed questions
• List – list of items, any of which may be selected
Useful: when you need to be sure that the respondent has
considered all possible responses.
For example: „What kind of music styles to you listen to?“
indie/ rock/ folk/ indie/ jazz/ easy listening/ indie/ world /
classical/ indie
• Category – only one response can be selected from a given
set of categories
Useful: if you need to collect data about behaviour or
attributes; categories should not overlap and should cover all
possible responses!
For example: „How often do you go to the cinema?“
2 or more times a week/ once a week/ couple times a month/
• Ranking – the respondent is asked to place something in
Useful: when you want to discover the relative importance of
the things to the respondent. Seven / eight items maximum ->
respondents rank accurately when they can remember all
For example: „Rank the factors listed in order of importance to
you in your choice of field of study“
other people’s advice/ personal interest/ potential future
income/ career possibilities / possibility to travel etc.
• Rating – a rating device to record responses
Useful: when you need to collect opinion data; Likert-type
rating scales; 4, 5, 6, 7-point rating scale.
For example: „I feel that my opinion is important in decision
processes in the organisation“
agree/tend to agree/ tend to disagree/ disagree
• …purposeful discussion between two or more people 
• Structured interviews – questionnaires based on a predetermined/
standardised / identical set of questions -> intervieweradministered questionnaires (already discussed). The social
interaction should not mediate the responses, the questions should
be read exactly as written and in the same tone of voice – be as
computer-like as possible. Also known as quantitative research
• Semi-structured interviews – non-standardized; list of
themes/questions to be covered, partly varying from interview to
interview; additional questions may be added; the order of
questions may vary. Also known as qualititive research interview
• In-depth interviews – informal; to explore in depth a general area
of interest; no predetermined list of questions; only clear idea
about the aspect(s) you want to explore. The interviewee is given
the opportunity to talk freely. Also known as non-directive interview
Links to the purpose and research strategy
• Semi-structured and in-depth interviews:
- „what“, „how“ and „why“ questions
- exploratory studies - to find out what is happening and
to seek new insights
- explanatory studies - semi-structured interviews -> to
understand the relationships between variables
• Useful:
- to help identify the questions for your questionnaire
- explore / explain themes that have emerged from your
- to validate findings from your questionnaire
Situations favouring non-standardised
• The purpose of the research – to understand the reasons for the decisions
of participants; to understand the reasons for the decisions for their
attitudes/ opinions. Understanding if interviewees use words/ ideas in a
particular way; to lead discussion into areas you hadn’t previously
considered; to formulate some new question
• The significance of establishing personal contact – managers likely to
agree to be interviewed rather than complete a questionnaire. Receiving
the questionnaire through the post ->problematic to provide sensitive
information; trust the way in which the information is used; understanding
the meaning of the questions
• The nature of questions – large number of questions; complex or openended questions; the order and logic of questioning may need to be
• Length of time required and completeness of the process – issues that
are complex, unclear, or large in number; empty answers in questionnaire
give no hint, why they are unanswered; by personal interview, you will
form some indication of why a response could not provided -> incentive to
modify the question
Data quality issues
• Forms of bias
– the interviewer bias –
comments, tone or
non-verbal behaviour of the
interviewer creates bias in the
way that interviewees
respond to the questions
being asked. Imposing
interviewer’s own beliefs/
frame of reference 
bias in interpreting
Data quality issues
• Response/ interviewee bias – respondent’s
sensitiveness to the unstructured exploration to certain
themes –> not reveal and discuss a topic in question;
providing a partial „picture“ of the situation casting
him/herself in a socially desirable role
• Validity and generalisability
– getting the right sample - some parts of the sample
remove themselves from taking part because of the lack
of time -> biased sample
- no generalisations can be made about the population small and unrepresentative samples
Successful interview preparation
• Level of knowledge about the organisational / situational context - your
prior knowledges raise your credibility
• Level of information supplied to the interviewee – credibility through the
supply of relevant information to participants before the interview;
possibility for them to consider the information being requested and
assemble supporting organisational documentation if needed
• Appropriate location – comfortable; unlikely to be disturbed; quiet
• The beginning of interview
– shaping the credibility
- explaining your research
- gaining the consent/ confidence of the interviewee
- supplying the information how the data is used/ how the anonymity is
Successful interview preparation 2
• Approach to questioning –
- questions be phrased clearly, understandably, with neutral tone of
- avoid too long questions;
- avoid questions made up of two or more questions;
- avoid too many theoretical concepts;
- ground your questions in the real-life experiences of your
participant rather than abstract concepts;
- use critical incident technique - respondent is asked to describe in
detail a critical incident that is key to the research question;
- leave sensitive questions until near the end - greater time for the
building up the trust; if the question upset the respondent, the
irritation will not affect further questions :P
Successful interview preparation 3
• Interviewer’s behaviour during the interview
– comments / nonverbal behaviour indicating any bias in your
thinking should be avoided;
- neutral, but not uninterested response;
- do not look bored!
- listen attentively – hold back your own thoughts, give
enough time to develop the response, avoid projecting
your own views. Careful listening allows yo to identify
comments that may be significant to the research topic.
• Test your understanding – e.g., summarize an explanation
provided by the respondent – so he/she can evaluate the
adequacy of your interpretation
• If your research question/objective is
concerned with what people do, an obvious
way in which to discover this is to watch them
do it.
• Definition - the systematic observation,
recording, description, analysis and
interpretation of people’s behaviour
Participant observation
- participate fully in the lives and activities of
subjects and thus becomes a member of their
group, organisation or community
- not only observes, but also feels, what is
The meanings and context of the phenomena
are discovered (hopefully).
Researcher roles
• Complete participant – you as a researcher attempt to
become a member of a group in which you conduct
Example: studying luchtime drinking in particular work setting.
The question of ethics. Threat: in gaining the trust of your
„colleagues“, you might value that trust so much that you lose
the sight of your research purpose
• Complete observer – you would not reveal the purpose of
your activity to those you are observing. You do not take
part in the activities of the group.
Example: studying the consumer behaviour in the
supermarkets - how people tend to react to the „rush hours“
Structured observation
has high level of predetermined structure
quantify behaviour
how often things happen rather than why they
• …moves more and more into the Internet –
„indirect observation“, traces of behaviour
• personal digital assistants
Threats to validity and reliability
• Subject error – observing a situation that is not representative – but you
do not know it –> choose subjects who in as many respects as possible are
„normal“ examples of the population under study
• Time error – at different times a day and a week there may be very
different processes occuring
• Observer effect – the process of the observer’s observation of behaviour
changes the nature of that behaviour owing to the fact that the subject is
conscious of being observed. It is experimentally demonstrated that also
cockroacher run faster when other cockroaches are watching.. To
overcome this effect:
1) to observe in secret
2) minimal interaction – observer tries to melt into the background, has no
or little interaction, avoids eye contact, sits in an unobtrusive position
3) habituation – subjects become familiar with the process of observation
so they take it for granted. Several observation sessions are necessary in
the same setting with the same subjects
Homework 3
• You have for now chosen your research
strategy. Now, specify your data collection
method (write just a short passage to extend and make
more clear, how your measurement tool should look like. For
example – if you wanted to use questionnaire, design some
questions and explain, why you choose to ask these. If you
want to interview someone, then design some topic/question
and specify, would it be structured or semi/structured
interview, etc)
• Email it.. Before our next lecture 