Probability Sample

Probability versus Nonprobability Sampling

Due to limits on resources and time, researchers are rarely able to observe every member of a target population, or the group for which a researcher wishes to generalize the results of a study. Instead, a subset must be chosen. Members may be selected for study using either nonprobability or probability sampling methods. In nonprobability sampling, selection occurs in a nonrandom fashion usually based on availability. Two examples of nonprobability sampling methods are convenience and snow-ball sampling. Because all members of a population chosen via nonprobability sampling methods do not have a nonzero chance of being selected for the study, it is not possible to determine how closely a nonprobability sample resembles the target population and, therefore, results from these studies are not generalizable to the entire target population. However, probability sampling, which involves selection of members from the target population using random selection techniques, produces results that, in [Page 841]the absence of other biases, are generalizable. In probability sampling, all members of the target population have a nonzero opportunity of being chosen to be in the sample and the probability that any given one will be chosen can be calculated.

Types of Probability Sampling

Simple random sampling, systematic sampling, and stratified or cluster sampling are types of probability sampling. In simple random sampling, members are randomly and independently chosen from a list of all target population members; every member of the target population has an equal chance of being chosen for participation in a study. In systematic sampling, a sample is chosen by selecting the first member from a list at random and then by taking every kth member from the population list thereafter. In stratified or cluster sampling, a population is first subdivided into groups that share at least one common characteristic, then a sample is chosen from each stratum using simple random or systematic selection.

Probability Sampling and Sampling Error

The degree to which a sample obtained through probability sampling varies from the target population is measured by estimating the sampling error. Sampling error arises when only part of a population is observed; for any given target population, many alternative sample realizations, or sets of members or individuals selected for a sample, can be obtained by employing a given sampling method, and each may lead to a different summary statistic, such as mean, for the parameter being studied. Theoretically, sampling error is derived from the amount of variation that exists between the summary statistics for all possible realizations. In practice, the standard error of the study sample is used to construct a confidence interval for which with a certain level of confidence the true parameter of interest lies.

...