Understanding Academic Self-Concept and Performance in Medical Education

Problem: Medical students' academic self-concept (ASC) is an important factor in better understanding noncognitive mediators of performance in medical school. However, research is limited on ASC in medical students across multiple phases of undergraduate medical education curriculum. This pilot study explored the relationship between ASC and academic performance across different phases of a U.S. medical school curriculum, specifically at the end of the second (preclinical) and third (clinical) years.

Approach: Medical students across 2 cohorts at Virginia Commonwealth University School of Medicine, Richmond, Virginia, were surveyed using an ASC confidence subscale in 2019. Multiple linear regression analysis was conducted using medical student ASC scores in preclinical (n = 190) and clinical (n = 149) phases and performance data. Clinical performance was calculated through a weighted mean of clerkship grades based on the number of weeks for each clerkship.

Outcomes: Preclinical performance was related to ASC, gender, and performance after year 1. ASC scores varied significantly by gender in the preclinical cohort ( P < .01), with men reporting higher ASC than women (mean [SD], 2.94 [0.41] vs 2.78 [0.38]). Significant gender differences in performance were found at the end of year 3 ( P < .01), with women performing more favorably compared with men (mean [SD], 94.1 [59.04] vs 124.24 [64.54]). The relationship between ASC and performance at the end of year 2 suggested students with higher ASC perform better during their preclinical phase.

Next steps: This pilot study supports future scholarship in 2 areas: (1) identification and assessment of additional factors that influence the relationship between ASC and academic performance across the entire undergraduate medical education curriculum and (2) development and implementation of evidence-based interventions to support student ASC and performance and enhance the learning environment. Analyzing longitudinal trends across multiple cohorts will drive evidence-based interventions at learner and programmatic levels.