The survey

When the researcher questions people, either by interview or by having people fill in a questionnaire themselves, and the response format is structured so that responses can be quantified, then a survey is taking place. The whole of a population may be surveyed (a census) but more often a relatively small number of individuals will be selected from the population. Something about the whole population will be inferred from what the representative if the people in the sample are. Hence the importance of sampling. (You may like to note that the survey in market research came out of the older practice of opinion polling.)

A survey will be initiated by identifying the target group in the context of the research problem. This group becomes 'the population'. Following this, a sample method will be chosen and the sample size estimated. The next stage in survey work is the development, testing and administration of the questionnaire. There is often a trade-off in questionnaire design between the information wanted and the ease and cost of collecting the data. Computer software is used in the analysis of the survey data, so questionnaires are designed with this in mind. There now exists computer aided telephone interviewing (CATI) and, a lesser extent, computer aided personal interviewing (CAPI) but so far these lend themselves only to fairly simple, straightforward questionnaires.

We are now going to concentrate on sampling. This will be in the context of the survey, although you should note that sampling may, in certain circumstances, take place with the observation methods and also in experimentation (in the selection of retail outlets).