Data Sampling for Assessment
Sample size is a term used in research to define the number of subjects included in a survey, study, or experiment. In surveys with large populations, sample size is incredibly important. The reason for this is because it's unrealistic to get answers or results from everyone—instead, you can take a random sample of individuals that represent the population as a whole. https://www.qualtrics.com/blog/calculating-sample-size/
Considerations
If the population size is small enough to collect information on the entire population, this is ideal.
Sampling is often used for populations where it is not feasible and/or practical to collect information on every member of the population.
Some best practices for determining whether to use sampling in assessment data at the College:
- When assessing a course overall, if the course typically runs fewer than twelve sections per semester, you can most likely collect and use assessment data from the whole population of students in that course.
- If it is necessary to assess or analyze the course on a section-by-section basis, then you will likely still need to use population data or see the information on stratified sampling below.
- If the course typically runs more than twenty sections per semester, you will most likely want to consider sampling to keep the data set to a manageable size.
- If a course falls in between these two areas, the determination will be based on what kinds of artifacts you are assessing and for what purposes; please contact the Office of Assessment and Evaluation to discuss your situation in more detail.
Stratified Sampling for Equity Assessment
Although random sampling is often seen as the “gold standard” in sampling, there are instances when we need to subsample based on specific characteristics (e.g., race/ethnicity, gender identity, Pell eligibility).
- When it comes to stratified samples, there may be instances where there is not enough of a population of a specific group to sample; in these instances, please contact OIE/OAE/OIR for consultation.
- These subpopulations are known as strata (each subpopulation is known as a stratum).
Once stratum are selected, they can then be sampled randomly
- Alternatively, strata can be further divided into clusters and then random sampled.
- This would be helpful in the instance of looking at specific sub-populations (e.g., Hispanic females).
Ideally, stratified sampling consists of proportionate stratification where the sample size of each stratum is proportionate to the population size of the stratum.
- To find a statistically valid sample size, one must know the population size as well as the proportion of the population; if this is not readily available, Institutional Research can help provide this information.
Calculating Sample Size
A plethora of sample size calculators exist online: