Last update:
19-Nov-2012

Epidemiological studies rely on data from samples rather than populations. In epidemiology, "population" refers to all the possible observations that could theoretically be made in a study, while "sample" is a part of these observations. By using samples the precision of measurements is decreased, but in practice studies are rarely conducted with the participation of entire populations under study. Researchers conducting an epidemiological study have to choose a representative sample from a population. When the sample is representative of the population that it is derived from it is possible to generalize the results of the study of the sample to the whole population. In this way, external validity of the study is achieved. Selection of a representative sample of the population for epidemiological study is achieved by random sampling. The possible methods of random sampling are simple, systematic, stratified and cluster random sampling. In order to select a random sample, the first step is to define with precision the study population that gives rise to the sample. This population is known as the sampling frame and can be thought of as a list of all the observations from which the sample is to be derived. In random sampling, the probability of an observation being selected is equal to the probability of any observation of the sampling frame being selected. Participation of certain categories or strata of observations in a sample study is achieved by stratified random sampling, which increases the representativeness of the sample and decreases random error. Stratified random sampling can be separated into proportional and non proportional. In proportional stratified random sampling, the proportions of the strata in the sample are identical to the proportions of the strata in the population from which that sample is derived. Cluster sampling is applied in cases where the sampling frame of the population does not need to be determined with precision. In cluster sampling, the first step is to divide the population in a random way into separate clusters, following which the observations of the sample arise from the separate clusters.

Key words: Random sampling, Sample, Sampling, Sampling frame, Stratified sampling.