A Statistical Analysis of Medical School Exam Score Distributions

Mean, median, and standard deviation are integral components in basic statistical analysis. In the context of medical education, few have endeavored to statistically analyze large exam score data sets for these measures of central tendency and deviation. This study determines basic statistical measures of distribution among exams scores from first and second year medical students. Raw exam score data from the 2019-2020 academic year was collected from all exams given to first and second year medical students at Oklahoma State University College of Osteopathic Medicine. This data was deidentified and grouped according to Course Title and Course Exam in Microsoft Excel. This data was imported into the software suite R (version 4.0.2) for preliminary numerical and graphical analysis, with a distinction drawn between system and non-system courses. From these analyses it was concluded that Systems courses appeared to have more tightly clustered exam score means, medians, and standard deviations. This result lends itself to further exploration of a possible correlation between clustering of the calculated values and normality of exam score distribution curves. Additionally, Non-Systems exams showed a large degree of variability in measures of central tendency and dispersion. This observation, once assessed for statistical significance and skewness, may indicate a need to restructure non-systems courses in order to decrease the variance in exam performance. This study highlights the utility of simple statistical methodology in analyzing large scale data sets and will allow for future evidence-based curriculum decisions to be made.