1996 OPEN FORUM Abstracts
Population and Samples: Concepts and Examples
Joseph L. Rau, Jr., PhD, RRT Wednesday, November 6, 1996
The topic of populations and samples is introduced with two case examples of sampling which resulted in erroneous estimates of population values. The terms 'target population', 'accessible population' and 'sample' are defined, and used to analyze the flaws in the cases. The concept of a population is characterized, with an example from a published research study to identify criteria for the target and source of the accessible population. General considerations of samples are reviewed with an emphasis on the need for the sample to represent the target population. Sample size will affect the precision of sample estimates of population values, but not guarantee representativeness of the sample for population values. Sample values, or statistics, are estimates of population values (parameters), and sample statistics are used in inferential statistical tests. The difference of probability and nonprobability sampling is reviewed, with a definition of a probability, or random, sample. An example of the method used for obtaining a simple random sample is given, using a hypothetical sampling frame and table of random numbers. Methods of random sampling (simple, stratified, systematic, cluster) are identified. Random sampling is not possible for experimental medical studies in which voluntary informed consent must be obtained. Use of population criteria can safeguard the representativeness of a sample however. Random sampling is not the same as random assignment; the former is a method of obtaining a sample for a study; the latter randomly assigns the sample subjects to membership in the study groups.
References: Cochran WG. Sampling Techniques, 3rd ed. Wiley, 1977. Levy PS, Lemeshow S. Sampling for Health Professionals. Belmont, CA: Lifetime Learning Publications, 1991.