The mentor effect in student evaluation

To determine if faculty mentors rate their mentored students higher than do nonmentors, and to ascertain if gender is a factor.

All third-year students (n = 101) from academic years 1996-1998, who performed their obstetrics and gynecology clerkship at the Milton S. Hershey Medical Center in the Penn State Geisinger Health System and were evaluated by full-time faculty (n = 18), were included in the study (total observations = 545). Students were rated by faculty on an ordinal scale in five categories. Generalized estimating equation methodology was used to fit proportional odds models for ordinal data to assess whether there were statistically significant mentor or faculty/student gender effects.

Student evaluations from mentors were more likely to have better scores than student evaluations from nonmentoring-faculty for all five categories (all P <.01). The odds ratios (OR) for the mentor effect ranged from 2.1 (95% confidence interval [CI] 1.4, 3.2) for fund of knowledge to 3.2 (95% CI [2.1, 4. 8]) for attitude. For problem-solving and technical skills, male faculty were more likely than female faculty to give male students better scores (problem-solving skills: odds ratio [OR] = 1.7, 95% CI [1.0, 2.7]; technical skills: OR = 2.2, 95% CI [1.1, 4.6]). Mentoring-faculty evaluations were not strongly correlated with the students' objective examination scores.

Overall, mentors score their mentored students statistically higher than do nonmentors. Gender differences in evaluation, while present, are less consistent and smaller than the mentor effect.