Reliability of peer and self-assessment scores compared with trainers' scores following third molar surgery

CONTEXT It is sometimes claimed that self-assessment is inaccurate and that clinicians over-rate their performance. There is a need to find out why this should be. Is poor self-assessment caused by some clinicians' inability to accurately judge performance? Or does over-scoring result from a desire to convey a more favourable impression? Peer assessment is widely advocated and is said to be of benefit to both assessor and assessee. METHODS In this study, we wanted to see if postgraduates were able to peer-assess and if this form of assessment was more reliable than self-assessment when compared with assessment by a trainer. We used checklist and global rating scales to evaluate surgical skills in removing a mandibular third molar tooth. RESULTS There was no statistically significant difference between peer-assessed and trainer-assessed scores. We found that, on average, peer assessment (especially global rating scales) reflected trainer scores more accurately than self-assessment of surgical skills. Self-assessment scores were significantly higher on average than those given in peer assessment. DISCUSSION Although peers and trainee surgeons came from the same group, the surgeons were more likely to over-score when measuring their own performances. The greatest variability (and over-scoring) between assessor and trainee surgeon appeared to occur in those with lower mean scores. Formative peer assessment may be a useful and less stressful mechanism for encouraging reflection.