Assessing implicit gender bias in Medical Student Performance Evaluations

Eval Health Prof. 2010 Sep;33(3):365-85. doi: 10.1177/0163278710375097.


For medical schools, the increasing presence of women makes it especially important that potential sources of gender bias be identified and removed from student evaluation methods. Our study looked for patterns of gender bias in adjective data used to inform our Medical Student Performance Evaluations (MSPEs). Multigroup Confirmatory Factor Analysis (CFA) was used to model the latent structure of the adjectives attributed to students (n = 657) and to test for systematic scoring errors by gender. Gender bias was evident in two areas: (a) women were more likely than comparable men to be described as ''compassionate,'' ''sensitive,'' and ''enthusiastic'' and (b) men were more likely than comparable women to be seen as ''quick learners.'' The gender gap in ''quick learner'' attribution grows with increasing student proficiency; men's rate of increase is over twice that of women's. Technical and nontechnical approaches for ameliorating the impact of gender bias on student recommendations are suggested.

MeSH terms

  • Confidence Intervals
  • Data Interpretation, Statistical
  • Educational Measurement / methods*
  • Educational Status
  • Female
  • Gender Identity*
  • Humans
  • Likelihood Functions
  • Male
  • Multivariate Analysis
  • Prejudice*
  • Sex Factors
  • Students, Medical / statistics & numerical data*
  • Task Performance and Analysis