Construction dissertation guide part 4 - Questionnaires and interviews

This is part of a series of Construction Dissertation Guides for students. Click here for the whole series.

Typically, the method of data accumulation used for dissertations involves questionnaires or interviews, or a combination of both. However, other methodologies can be used:

• Experimental.
• Historical studies.
• Case studies.
• Mixed methodologies.

This article deals with questionnaires or interviews.

[edit] Questionnaires

Questionnaires are best used to accumulate data for descriptive or analytical research and are useful in establishing the opinions of a wide range of people. However, there may be a lack of depth in the answers due to ‘closed’ questions being asked such as ‘yes or no’, the use of rating factors and so on. This is inflexible in terms of gaining more of an insight into the individual’s personal opinions on the subject and in terms of potential inaccuracy or ambiguity in answers received.

Another problem can be a lack of response. The construction industry is a high-pressure, busy working environment, particularly in a tough economic climate, and completing a student questionnaire may be low on the list of priorities unless there is an existing relationship between the student and the individual.

It is important therefore that the questionnaire is as focussed as possible. When deciding on the sorts of question to include, the following considerations should be borne in mind:

• What is the objective of the question?
• Is the question relevant to the aim?
• Is the question relevant to the hypothesis?
• Could the question be answered from other sources?

Before honing the questions to a final format, compile some ‘first thought’ questions that arise whilst undertaking the literature review. The more specific the topic, the easier it will be to compose questions that are suitable.

Once the initial questions have been identified, the next stage is to pick out main topic areas that can serve as sections/categories. The list of questions should then be organised as best as possible into one or other of these to give the questionnaire some structure.

See example questionnaire.

[edit] Wording the questions

Once there is a list of rough questions, the wording needs to be constructed:

• Don’t include long-winded questions.
• Don’t make the questions too leading, they should be as neutral as possible.
• Avoid asking double questions. Respondents are less likely to answer them in enough detail or at all.
• Avoid questions that presume a position.
• Don’t use hypothetical or ambiguous questions.
• Try to make the sequence of questions flow.
• Try to ensure the questionnaire is presented professionally and proof-read for errors.

[edit] Types of questions

[edit] Open / closed questions

The following issues should be considered when deciding whether to use open or closed questions:

• The questionnaire’s objective.
• The level to which the respondent may be an expert on the topic.
• The ease with which respondents can realistically and accurately respond to the question.

[edit] Open/unrestricted questions

These seek the respondent’s opinions or free responses. For example, ‘what do you think about….?’

Open questions have the advantage of allowing the expression of views which can elicit more informative and valuable answers. They are useful when the possible answers may not be known and so it is difficult to pose a closed question with pre-determined options.

However, open questions inevitably demand more of the respondent and may be off-putting if they think it will take more time and effort to answer. Longer, personal answers are also more difficult to analyse compared with closed options that are more easily measurable.

[edit] Closed/restricted questions

These offer selected answer/s from a list of pre-determined options, such as; ‘yes’ or ‘no’, ‘agree’ or ‘disagree’. These are both easy to construct and quick and easy to answer, so the response rate is likely to be higher. The drawback is that they can bias the respondent with the available options, and the depth of information obtained is limited.

• Checklists: A list of items for the respondent to check if appropriate. This is straight-forward and can be easily analysed.
• Grids: Similar to a checklist but can allow answers to two or more questions at the same time.
• Rating scale: The respondent is able to rate the choice of answers on a numerical scale of relevance or importance depending on their level of agreement.
• Numerical rating scale: Respondents are asked to score answers based on their opinion of them.
• Ranking: Respondents are asked to place answers in ranked order.
• Scale: Respondents are asked to indicate their position on a scale of several points. This can be good for illustrating variants in perception.

Where scales are used, it can be advisable to select an odd number for the scale. This means the respondent has to make an active decision about whether to choose a higher or lower part of the scale as there is no option to just select the middle option for every answer.

[edit] Factual questions

Other questions may also be asked relating to:

• The individual/organisation’s background.
• Projects and case studies (such as, the type of construction, materials, procurement method, number of operatives, construction cost, etc.).
• Information accumulated from secondary data (such as, accident rate statistics).

[edit] Pilot study

A pilot study is a test run of the questionnaire, giving the opportunity to see how it works in practice in terms of the quality of response. This is a valuable means of assessing whether the wording, structure or layout of the questionnaire is appropriate or could be improved.

It might even be advisable to ask test respondents a handful of feedback questions to gauge their thoughts on the questionnaire. Things that might be improved by this feedback include:

• Length of time to complete.
• Thoughts on the layout.
• Thoughts on the subject matter and whether the key issues were addressed by the line of questioning.
• Clarity of the questioning and appropriateness of the format.

[edit] Covering email / letter

This should be sent out with the invitation to complete the questionnaire. It should concisely but accurately explain the purpose of the questionnaire, university course details, and why they would be a suitable respondent (this can be very general, i.e. ‘as a construction professional working on major infrastructure projects…’). The better the covering email or letter, the better the chance of a high response rate.

[edit] Interviews

The purpose of interviews is to gauge the views and opinions of professionals or others involved in the industry. They should be able to draw from a range of different viewpoints that can then inform the conclusions and recommendations of the dissertation.

Interviews are used where it is necessary to ask complex questions, that may be more subjective in nature, and that will elicit more detailed responses. A higher rate of response may be achieved compared to the largely unpredictable response-rate of questionnaires sent out impersonally via post or email. However, the results can be more difficult to analyse, particularly if the interview is not well thought out.

[edit] Interview design stage

Interviews can take an un-structured or structure form.

• Unstructured interview involve posing ‘open’ questions in a more explorative style, often pitched at a very general level. This is prevalent in the preliminary stages of research; for example, to determine the appropriateness of a proposed hypothesis in relation to the wider industry.
• Structured interviews are more formal and aim to discover as much as is possible about the topic. Interviews may be formulaic in that the questions presented are in the same sequence and with the same phrasing to all interviewees, although depending on the individual responses the interviewer may press for further details, allowing the exploration of tangents to the main subject.

Other aspects of the interview that need to be considered include, the optimum time to interview, the number of interviewees, and the sampling method.

[edit] When to interview during the research process

The best time to begin interviews is when the interviewee and their opinions are known well enough by the researcher in terms of their influence on the subject that only the most relevant questions for the research need to be asked. Clearly it is wise to pursue structured interviews only once the bulk of the research has been undertaken so that the researcher has sufficient knowledge to be able to engage with the interviewee and extract valuable information from them.

[edit] Number of interviewees

It can be beneficial to minimise the number of interviewees to a just a few that the researcher can focus on and perhaps tailor the questions to a form most appropriate for the individual. However, there is the potential disadvantage of this that the information yielded may not give a full view of the subject.

Interviews should ideally follow a loosely-structured format, allowing the interviewer’s own thoughts/ideas developed as the questions are answered by the interviewee to be incorporated in the same way as in an informal discussion.

Interviews should ideally be conducted after the Literature Review has been completed and perhaps data from questionnaires has been received and reviewed. The researcher should then be able to pinpoint the main areas of concern and so focus on these in the interviews.

Interviews should be recorded (with the permission of the interviewee) allowing attention to be focused on the interview itself, and for later transcription.

Continue to the next stage: Analysis and preparation of the results.