Student background, admission routes, and academic success: a structural mediation analysis

Medical school admissions increasingly seek to balance cognitive merit with broader notions of potential and diversity. Yet, the extent to which different selection routes predict academic success, and how social background shapes these pathways, remains debated. Using linked administrative and survey data from 644 students at a large German medical school, we examined how sociodemographic background, prior educational trajectories, and parental characteristics influence both admission routes and early academic performance. We estimated a structural equation model (SEM) with observed variables to represent temporally ordered educational pathways linking background characteristics, school performance, admission routes, and academic success. Direct and indirect effects were quantified using mediation analysis, including the proportion mediated to assess the relative importance of indirect pathways. Indicators of cognitive achievement in secondary education emerged as the strongest and most consistent predictors of medical school performance. Parental education and school achievement facilitated direct access to medical school via more favourable educational trajectories. Gender had no meaningful indirect or direct influence on admission routes or performance. Admission routes that placed greater weight on non-cognitive criteria were independently associated with lower exam performance, even after accounting for prior achievement. Our findings indicate that within the highly selected and academically rather homogeneous population of medical students, effects of demographic and socioeconomic background on study success and student selection operate predominantly through earlier educational pathways. Inequalities in opportunity arise primarily before university admission, not within it. For medical schools, this implies that fairness is best promoted not by compensating for background through increasingly complex non-cognitive selection tools, but by maintaining transparent, evidence-based admission systems anchored in criteria with demonstrated predictive validity for study success.