Comparison of Student Performance by Computer Tests to Paper Tests on Histology and Embryology Lecture Material

Abstract

In 2002 Indiana University School of Medicine started a course server ANGEL for student access to course material. The content made available depended on faculty interest in each course. At IUSM-Fort Wayne lecture PowerPoint presentations were made available after each lecture in the Histology and Embryology course. Another feature that was of interest was use of the ‘quiz’ mode for regular unit exams. The quiz mode was designed with minimal security for low stakes exams but was not of a major concern since Fort Wayne had a small class size (16) and tests could be administered in computer labs with proctors present. We now are going on four years of experience with using computer given lecture exams and although security has increased with newer versions of ANGEL, tests are still given in the same format. Forty one students have completed the course from 2002 and for this study a comparison was made to 46 students who had completed the course in the preceding three years with similar paper tests. Class averages (mean ± St.dev., n) for years 1999–2001 were 86.35% ± 4.24% (15), 83.47% ± 5.16% (15), and 84.77% ± 4.05% (16), respectively. Class averages for 2002–2004 were 87.12% ± 4.78% (12), 86.45% ± 3.49% (15), and 85.72% ± 6.32% (14), respectively. Student performance data was pooled into paper and computer test groups and analyzed for homogeneity, then by ANOVA for statistical differences. Data was not heterogeneous with mean performance for paper exams at 84.86 ± 0.67 (46) and for computer exams slightly higher with 86.40 ± 0.76 (41). Students may benefit from computer exams in that they will be taking NBME exams on computers at the end of their sophomore year and may even perform slightly better on computer exams in course material, however, this increase in average performance was not statistically different (p=0.133).