Effects of Early Warning Emails on Student Performance

We use learning data of an e-assessment platform for an introductory mathematical statistics course to predict the probability of passing the final exam for each student. Based on these estimated probabilities we sent warning emails to students in the next cohort with a low predicted probability to pass. We analyze the effect of this treatment and propose statistical models to quantify the effect of the email notification. We detect a small but imprecisely estimated effect suggesting effectiveness of such interventions only when administered more intensively.