Determining your sample
Most of us are experts in our disciplines and not in statistics. Following are some general suggestions for data collection related to instructional assessment. You are invited to inquire further with our colleagues who are experts in statistics, Donna Raymond or Monte Cheney (and thanks to them for the enthusiastic invitation).
When you assess students, you are targeting a specific population for whom you will generalize your results. For example, you may want to target all graduating students in a CTE program, or all students who have completed a specific general education course in a given year. Your intended population shapes your assessment methods.
If you are not assessing the entire population, you are attempting to find a representative sample. The sample size and methods will help you ensure you have a good representation for your population.
Determining an effective sample size
We often worry about the sample size when collecting data. If you want assistance determining an appropriate sample size, you can use a sample size calculator (general recommendations are to select a margin of error no greater than 10% and a confidence interval of at least 90%). Or, consult our trusty COCC's math faculty who teach statistics.
Method of data collection
If you choose to evaluate a sample of student work to represent the student population, your goal is to get a random sample. From a statistical perspective, the method of data collection is of a much greater concern for inference than sample size. It is frustrating to have a great quantity of data that is not useful due to errors in data collection. To get a random sample you might select every third, or fourth (or nth) student from your roster, where the actually increment of students is driven by the number of samples you need. Beware of voluntary response
Protecting student identity
When you report your data, please do not share any individual student identifiers such as part or all of their names or student ID's.