Guidance appreciated: I have approx 40k records from about 50 colleges. Data is FERPA-compliant one student per record, with age (integer), gender, ethnicity, college, course, and section. Course is selected from a set of size three. I need some guidance on how to characterize this population (or these samples from a much larger population). My intent is to run simulations using section sizes and student compositions slightly (randomly) different from (centered about) the real-life data.