This post provides a quick summary of some important things to consider when taking and studying a sample from a population. In order to provide reliable estimates (inferences) relating to the population, the sample has to fulfil a strict set of criteria.
Probability sampling is important in the social sciences because it allows us to take information from a relatively small number of sampling units (e.g. people) and represent the population of interest with a known precision. By taking a sample, we can save time and money, and still produce analyses which are sufficiently accurate to meet our needs. However, for these analyses to be unbiased and sufficiently precise, the sampling must fulfil strict criteria. Four important considerations are discussed in the sections below (click to expand the numbered items):
1. Random selection from the population of interest
2. Independence of sampling units
3. Sample size considerations
4. Missing data
Hopefully, the sections above provide a concise intro to some important aspects of probability sampling. Of course, each of these is a topic of study in its own right, with its own literature and research community. Anyone seeking to use probability samples in their research would do well to read up on these topics.
1. Adapted from: Bartlett et al., “Organizational Research: Determining Appropriate Sample Size in Survey Research”, Information Technology, Learning, and Performance Journal, Vol. 19, No. 1, Spring 2001