The validity of clerkship performance evaluations

The validity of narrative evaluations of clinical performance in medical school was investigated for three successive classes of medical school graduates. The criteria were ratings obtained from residency supervisors one year after entering the programme. Bivariate correlational analysis, multiple regression analysis and canonical redundancy analysis were performed to examine the associations between the evaluations by teachers of competence in medical school clerkships and the subsequent ratings by residency supervisors. The results, best summarized in the canonical analysis, showed two relationships between the two sets of measures (lambda 1 = 0.54 and lambda 2 = 0.47). The first relationship revealed an overall-competence-in-medicine dimension; the second suggested an interpersonal-skills/professional-maturity dimension. Overall, knowledge of clerkship performance accounted for 15% of the variance in residency ratings. The findings suggest the need to improve the assessment of professional maturity and behaviour during medical school.