Sociological Methods

A. Swaby-Burton
Contributor

THE MOST common type of survey is the cross-sectional or snapshot survey. The researcher gets information from a cross-section of the group under study at one point in time. Thus, the findings are gathered fairly quickly and cheaply to provide useful information.

A less common and more expensive type of social survey is the longitudinal survey. In this, a selected group is studied over a period of time. This enables the researcher to gain not only current information but also insights into the way people are changing their attitudes or behaviour over a much longer time-span.

Almost always used as part of a social survey is the questionnaire. This is where the sociologist designs a set of questions about a particular subject to be answered by respondents (the people being surveyed). Questionnaires are difficult to write but are worth the effort because the sociologist can get a lot of information from respondents. However, some types of questions are easier to handle than others.

(Adapted from R. Power et al.
Discovering Sociology)

a) According to the passage, what advantages does the longitudinal survey have over the cross-sectional survey?

(b) Name two different types of question which may be used in a questionnaire, and give an example of each.

(c) "... the people being surveyed" are chosen by different types of sampling. Identify the advantages and disadvantages of different sampling techniques.

(d) Examine the advantages and limitation of using questionnaire.

ANSWERS

A. The longitudinal survey has one major advantage over the cross-sectional survey in that it offers deeper insights into the way in which people's attitudes and behaviour change over time.

B. Two different types of questions, which may be used in a questionnaire are open ended and closed questions. An example of an open-ended question would be: "In what ways are you influenced by television advertising"?

An example of a closed question would be: "Do you own your own television set? Please tick appropriate box. Yes No

C. Random or probability sampling is perhaps the most widely used form. The choice of the sample may, itself, be problematic, particularly where the sample group comprises of, say, households rather than individuals. Perhaps the most straightforward method is to pull names out of a hat, or, alternatively use a table of random numbers. By choosing every 10th, 20th names we begin to use quasi-random or systematic sampling. It is clearly possible here that the chosen random sample could be unrepresentative, example, all men or women, when the researchers seek an equal number or both.

Stratified sampling offers a more representative method than the random method, allowing a wide range of personal characteristics to be sampled, example age, sex, religion. This method tends to decrease the likelihood of
sampling error.

Quota sampling consist of similar elements to stratified sampling, whilst containing one important point of difference. It is likely that this method will be used by a researcher in the field who goes out and personally locates the appropriate quota. Care has to be taken, however; if the quota is obtained at the same place and time then it would be unrepresentative of the general population, for example, it could miss shift workers. The method is generally speaking both quick and cheap, whilst not being either very accurate or as reliable as, for example, stratified
sampling.

Snowball sampling operates when a sample is built through one respondent (example a group member) who then suggests others to be interviewed. This has the advantage of assisting with access into the group but contains the danger, once more, that the sample might not be representative, but reflect the friendship grouping of the original respondent; the sample is clearly not random.

A major advantage of questionnaire is their ability to be circulated (often through the post) to large numbers of people at a relatively low cost, when compared with interviews or observational methods. It is important that the questionnaire itself and the questions asked are to be valid; that is, that they actually measure what it is that they are intending to measure. There are difficulties with the use of a postal questionnaire, in as much as the researcher has no direct or immediate way of knowing whether the
responses give a true reflection of the person's view.

The questionnaire must also be reliable and this level of reliability will be assessed in terms of both the consistency and the reputability of the response. In other words, whether the respondent's answers fit into a consistent pattern throughout, or contain contradictions, and whether if asked the same questions again, the respondent would give the same answers.