Medical Student Evaluations of Medical Faculty: Characteristics of Highly and Lower-Rated Teachers

Purpose: Our purpose was to evaluate the characteristics of highly and poorly rated teachers as well as to assess the validity and reliability of those evaluations. Methods: We downloaded 14 years of medicine faculty evaluations completed by 3rd and 4th year medical students. We dichotomized overall teaching effectiveness as outstanding (receiving "outstanding") or inferior (rated as a "unsatisfactory," "marginal," or "acceptable"). We analyzed these using logistic regression (STATA v 18.0). We assessed validity and reliability using factor analysis, Cronbach's alpha, and intraclass correlation coefficients. Results: Most (57%) of the 722 faculty members were rated as outstanding. Medical students valued faculty that took advantage of opportunities to teach (OR, 3.0; 95% CI, 2.7-3.3), who were enthusiastic (OR, 2.6; 95% CI, 2.3-2.9), and clear/organized (OR, 2.5; 2.3-2.7). Faculty rarely were rated as inferior (7.7%). Among lower-rated faculty, 91% had more than one lower evaluation. Lower-rated teachers had lower ratings on most domains of evaluation including taking advantages of opportunities to teach (4.6 vs. 2.7, p < 0.0005), being clear and organized (3.0 vs. 4.6, p < 0.0005), enthusiasm (4.5 vs. 2.7, p < 0.0005), being supportive (4.5 vs. 2.5, p < 0.0005), providing feedback (4.4 vs. 2.6, p < 0.005), or clearly answering questions (4.6 vs. 3.1, p < 0.0005). While evaluations were highly consistent (Cronbach's alpha, 0.94), there were low levels of agreement with intraclass correlation coefficients ranging from 0.09 to 0.36. Conclusion: Most attendings received high ratings, while lower ratings were uncommon. Most teachers receiving lower ratings received more than one, suggesting that lower ratings may be a better discriminator of teaching effectiveness than outstanding ones. Teaching ratings had low inter-rater reliability, suggesting either low validity or that learners value different characteristics in teachers.