Sample Design

A sample design is the framework, or road map, that serves as the basis for the selection of a survey sample and affects many other important aspects of a survey as well. In a broad context, survey researchers are interested in obtaining some type of information through a survey for some population, or universe, of interest. One must define a sampling frame that represents the population of interest, from which a sample is to be drawn. The sampling frame may be identical to the population, or it may be only part of it and is therefore subject to some undercoverage, or it may have an indirect relationship to the population (e. g. the population is preschool children and the frame is a listing of preschools). The sample design provides the basic plan and methodology for selecting the sample.

A sample design can be simple or complex. For example, if the sampling frame consists of a list of every unit, together with its address, in the population of interest, and if a mail survey is to be conducted, then a simple list sampling would be appropriate; for example, the sample design is to have a sampling interval of 10 (select every 10th unit) from the list. The sample design must vary according to the nature of the frame and the type of survey to be conducted (the survey design). For example, a researcher may want to interview males through a telephone survey. In this case, the sample design might be a relatively simple one-stage sample of telephone numbers using [Page 777]random-digit dialing. One aspect of the sample design in this case is to determine whether all males in a household are to be interviewed and, if not, how to select a second-stage sample of males from each sampled telephone number that reaches a household.

The simplest type of sample design is purposive sampling, or convenience sampling. Usually, however, the survey researcher wants every unit in the frame to have a known probability of selection, so a more complex sample design is needed. In many situations, for purposes of efficiency and reducing costs, a multi-stage sample is desirable. For example, suppose a researcher wants to do face-to-face interviewing for a population consisting of African American women. In this case, the survey researcher might select an area probability sample, with the first stage of selection being a sample of counties, or primary sampling units. The next stage of selection might be a sample of blocks within the sampled counties, followed by a sample of housing units from the sampled blocks. The appropriate sampling frame at the first stage of selection for such a survey would be all counties in the United States. However, the researcher might decide to restrict the frame to only those counties with more than a particular percentage or number of African Americans. In this case, the survey researcher is introducing undercoverage into the sample design, which will likely result in some degree of bias in survey estimates. In deciding on the preferred sample design, the researcher must weigh the bias concerns against cost issues. If the researcher has a fixed budget, then sampling from all counties will result in higher variances or standard errors, but in lower bias, than will sampling from a restricted frame of counties.

...