Can we improve the criteria for selection of medical students in Sri Lanka?

Abstract

OBJECTIVE: To examine the validity of selected entry level characteristics in relation to the GCE Advanced Levei (AL) examination as predictors of performance of medical students. METHODS: Data on two consecutive batches of students at the Medical Faculty, Kelaniya University, were analysed. These included students' sex. AL attempt, AL z-score, English placement test marks on registration at the faculty, and aggregate marks at summadve examinations held in the 2nd (Phase I), 3rd and 4l (Phase II) years. The effect of predictor variables on medical school marks was assessed by multivariate analysis with causal path diagrams. RESULTS: The sample consisted of 294 students; 58.8% were females. Most had entered after the 2nd attempt (32% from 1s1, 48.3% from 2nd, 19.7% from 3rd attempt). The mean z-score was 2.017 (SD0.142). Mean English placement test mark was 59.1 (SD14.2). Mean marks in Phase I and II examinations were 54.3 (SD7.3) and 56.5 .(SD7.6) respectively. Significantly more males had entered after the 3rd attempt than females (27.2% of males, vs 14.5% of females, Pearson x2=0.027). Females did significantly better in English and in Phase I and II examinations. Path analysis showed that z-score, attempt and English placement test marks were all significant predictors of outcome at the First Examination. There was high correlation between performance at First and Second Examinations. CONCLUSION: Requiring prospective medical students to pass General English and restricting the number of attempts at the AL to two instead of three, will result in selection of medical students that perform better.