Rampant cheating has driven the establishment of testing centers at large universities like School where new assessment strategies are being explored. Since creating ours in 2023, use has grown from 8 to 27 courses and 5,550 to 8,015 exams. Yet, there remains a lack of research on these types of testing centers. To address this, we observed four different types of computer science courses, both undergraduate and graduate level. We found that scheduling patterns form a universal and distinctive pattern. Students tend to schedule exams as late as possible with the greatest amount of students scheduling late in the middle of the quarter. Yet, nearly every student who receives a grade of C or lower is a late test taker.

Initially, we hoped to gather data to help optimize seating capacity, and we have by adjusting hours according to peak usage patterns. However, these results provide new insights pedagogically as well. We can dispel widespread concerns about collusion or assumptions that late test takers have an upper hand, something faculty worry about with flexible testing. We can also identify at-risk students and provide early intervention. Ultimately, our goal is to share insight about testing centers at large universities and to support goals like mastery learning, flexible testing, and multiple chance exams given ever expanding enrollments.