Writing process
5 Questionnaires and interviews
General
Questionnaires and interviews are two methods of empirical research.
Questionnaires are documents that ask a group of people (called ‘respondents’) to complete a number of questions or statements about a topic. They are usually a quantitative method of research, provided you have enough data (for more details see below). Otherwise, they are like interviews a way of getting primary qualitative data.
Interviews are meetings between researchers and an expert about a topic during which the researcher questions the expert. They are used for qualitative research. Results can differ considerably from interviewee to interviewee and are not necessarily representative. Data obtained is often difficult to consolidate. Table 2 below illustrates the main differences.
Table 2: Main differences between questionnaire and interview
| questionnaire | interview | |
|---|---|---|
| Purpose | to gather information from a typical audience about the topic | to gather information about the topic from an authority |
| Respondents | (usually) people chosen at random | authority about the topic |
| Design | written statements or questions designed with possible answers on the document given to many respondents respondents answer the survey alone and return it to the designer, usually by mail |
written (open) questions designed to be answered by the interviewee usually one person per interview interviewee answers directly (face to face, on the phone, by email etc) |
Which method is adequate depends on your research question(s) or hypotheses, your aims and objectives. You need to consider what you want to achieve, what kind of data you require, which results you expect from a certain method, whether you want more in depth information or need representative answers. You also have to consider which method is feasible within the scope of your research.
What is a representative sample?
In quantitative research the group of people surveyed are called “sample”. This can be representative or not. A representative sample does not significantly deviate from the sampling universe. A rough guideline is that for a survey you need about 1’100 answers per investigated group. So, if you investigate the attitude towards costs for public transport of the general public within the Canton of Lucerne, you need at least 1’100 responses from people who feature similar characteristics as the whole population in terms of gender, age, income, origin and other factors possibly influencing your conclusions. However, in case your sample consists mainly out of rich people, it is unlikely to be representative for the whole population in terms of their attitude towards costs for public transport even if there are more than 1’100 answers.
If you use questionnaires as a means to conduct qualitative research, the groups of people asked are the “respondents”.
See next two chapters on more details on how to design, conduct and write up results of questionnaires and interviews.