ABSTRACT
When writing the objectives for a questionnaire (see Chapter 1), you should have a specific population in mind. A population is the group in which you are interested. It may be small, such as all social workers in a town, or large, such as all social workers in the nation. If you survey all members of a population (and they all participate in the study), you are conducting a census-that is, a count or survey of all members of a population. When populations are large, researchers often draw just a sample of its members, administer the questionnaire to them, and infer that what they learn about the sample is also true of the population. In this chapter, we will consider and evaluate basic methods of sampling. 1
While the population is the group a researcher is interested in, the accessible population is the group to which he or she has access.2 Consider, for example, a researcher who wishes to survey a sample of all accountants in the United States (her population) with a questionnaire, but has access to only the mailing lists of the Association of Women Accountants and the Association of Government Accountants. The accountants who are on the mailing lists of these organizations constitute her accessible population-those to whom she has access. Notice that even though she wants to study all accountants (a census), she will be able to study only a sample of them-the accessible population. Because sampling bias results when some members of the population have a greater chance of being selected (such as female and government accountants in our example) than other members of the population (such as male and nongovernment accountants), her sample will be biased. Because biased samples can produce misleading results, many statisticians warn against attempting to generalize from a biased sample to the population. At best, any generalizations from biased samples should be made with considerable caution.
