ABSTRACT
For many research projects, studying the entire population of interest is not possible. Rather, the researcher must engage in sampling, deciding which members or units of the population of interest will provide the data for the study at hand. In this chapter, random sampling techniques are described, and the logic (and basic math) that shows why random samples are the best choice is also discussed. In order to draw a random sample, in which every unit has an equal chance of being selected, the researcher must have access to a comprehensive list of the units in the population. Simple random sampling, stratified sampling, systematic sampling, and cluster or multistage sampling are examined as types of probability sampling. When obtaining such a list is not feasible, a non-random sample is drawn, and this chapter introduces particular approaches to drawing non-random samples, as well. These include convenience samples, purposive samples, quota samples, and snowball samples. The goal of choosing a sample that will represent the wider population is advanced in the chapter, as are means of determining sample size. Sampling error, standard error, the central limit theorem, normal distribution, and confidence intervals are also defined. Importantly, the chapter raises particular considerations for (and from) a social justice perspective, including the possibility of oversampling or using a weighted sample in order to be sure minoritized voices are heard.
