Performance rating scale for peer and self assessment.

Abstract

A performance rating scale was developed and tested on a class of junior medical students who rated themselves and four to ten of their peers. When 928 ratings were factor analysed, two strong factors, knowledge and relationship, emerged. Test-retest reliabilities were good. Validity was measured by correlation of ratings with grades, and though both sources of ratings correlated significantly with grades given by faculty, peer ratings were more highly related to grades than were self ratings. Students tended to rate themselves lower than they were rated by their peers. Grades are probably not the best estimate of performance, but are currently one of the most reliable. Use of the scale to judge performance of physicians in practice has not been tested. The question of how such evaluation of peer and self would relate to other measures of quality of care is raised.