Effect of unproctored versus proctored examinations on student performance and long-term retention of knowledge.

Abstract

To compare unproctored and proctored online exams among chiropractic students. Pre-existing data of 234 students across 4 consecutive endocrinology classes were analyzed for this study. The course was comprised of 3 lectures (50 minutes per lecture) each week. Student performance was evaluated by midterm exam and summative exam (S1). The students from 3 classes were asked to take a voluntary second summative exam (S2) approximately 7 months after the S1. Since this study was partially conducted during the COVID pandemic, some classes took the midterm and the S1 proctored in the classroom while others took them unproctored from a remote location. The mean midterm exam (p < .001) and S1 scores (p = .01) for the unproctored group (93.6 ± 7.0 and 88.8 ± 8.2) were significantly higher than the proctored group (88.1 ± 8.2 and 83.9 ± 11.2). The mean time taken by students was much greater for the unproctored exams than for the proctored exams (midterm: 40.7 ± 10.2 versus 16.7 ± 7.0, p < .001; S1: 47.0 ± 8.7 versus 21.5 ± 9.0, p < .001). By contrast, the mean unproctored S2 scores were lower than the proctored group (60.2 ± 14.7 versus 88.1 ± 8.2, p < .001). A linear regression test showed that the final exam was a statistically significant predictor of the recall exam (p < .01, R2 = 28.3%). The findings suggest that student performance is significantly altered by test format.