Tuesday, September 2, 2014


We are just a few days away from the deadline for submission to ACC-RAC. And I have had an opportunity to assist a number of faculty in developing their submissions. Each year, one of the things I see are projects that involve survey methodology. This got me thinking about what it means to use a sample. What methods exist for selecting a sample from a population? Peter Jarvis, is his text “The Practitioner-Researcher: Developing Theory from Practice” (pages 122-123) provides a full list of samples. Among them are:

  • Random Sample: Here, all we do is select a sample for the population in a way that gives everybody in the population an equal chance of being selected.
  • Systematic Random Sample: Here, we find our population and then use some system to select our sample. If we did this in a class, it might mean that we use every 3rd student’s data or something similar.
  • Stratified Random Sample: We divide our population into groups somehow, and then we select an equal number of people per group. For example, we could stratify by age range.
  • Cluster Sample: Jarvis uses the example of doing research about teaching medicine and selecting a number of medical schools in a particular area and then study the teachers in just those schools.
  • Stage Sample: This builds on a cluster sample, but then uses a second random sample in the population drawn from the cluster group.
  • Convenience Sample: We select our sample from a convenient population. This is what we do with most of our clinical trials- the community is invited, some people respond, and then we get our sample from those that do.
  • Quota Sample: We select a specific number of individuals from each category we seek to study.
  • Purposive Sample: We hand pick the sample. Perhaps we invite our colleagues to participate in our research.
  • Dimensional Sample: This is a modified quota sample where we also ensure that we capture all the dimensions of the practice to be studied.
  • Snowball Sample: We ask, for example some experts to be in our study, and then ask them to suggest additional experts. The sample then “snowballs” in size.

Thus, there are many ways to find a sample, and as always, we just need to use the best one to answer out question.

No comments: