Ethnic and social disparities in performance on medical school selection criteria.

CONTEXT Medical schools in Western societies seek measures to increase the diversity of their student bodies with respect to ethnicity and social background. Currently, little is known about the effects of different selection procedures on student diversity. OBJECTIVES This prospective cohort study aimed to determine performance differences between traditional and non-traditional (i.e. ethnic minority and first-generation university candidates) medical school applicants in academic and non-academic selection criteria. METHODS Applicants in 2013 (n = 703) were assessed on academic and non-academic selection criteria. They also completed a questionnaire on ethnicity and social background. Main outcome measures were 'not selected' (i.e. failure on any criteria), 'failure on academic criteria' and 'failure on non-academic criteria'. Odds ratios (ORs) with 95% confidence intervals (CIs) were estimated by logistic regression analysis for ethnic subgroups (Surinamese/Antillean, Turkish/Moroccan/African, Asian, Western) compared with Dutch applicants, adjusted for age, gender, additional socio-demographic variables (first-generation immigrant, first-generation university applicant, first language, medical doctor as parent) and pre-university grade point average (pu-GPA). Similar analyses were performed for first-generation university applicants. RESULTS Compared with Dutch applicants, Surinamese/Antillean applicants underperformed in the selection procedure (failure rate: 78% versus 57%; adjusted OR 2.52, 95% CI 1.07-5.94), in particular on academic criteria (failure rate: 66% versus 34%; adjusted OR 3.00, 95% CI 1.41-6.41). The higher failure rate of first-generation university applicants on academic criteria (50% versus 37%; unadjusted OR 1.66, 95% CI 1.18-2.33) was partly explained by additional socio-demographic variables and pu-GPA. The outcome measure 'failure on non-academic criteria' showed no significant differences among the ethnic or social subgroups. CONCLUSIONS The absence of differences on non-academic criteria was promising with reference to increasing social and ethnic diversity; however, the possibility that self-selection instigated by the selection procedure is stronger in applicants from non-traditional backgrounds cannot be ruled out. Further research should also focus on why cognitive tests might favour traditional applicants.