I am collecting data in university courses over 2-3 cohorts. So good to see that learning behaviour is mostly similar across cohorts with a similar effect on exam performance (stay tuned for specific results). But what a pain to explain differences in taking the first or second exam, etc.