Are we seeing the full picture of peer-assessment in medical education?

Dear Sir

We read with interest the peer rating article by Basehore et al. (2014). We note that the findings of this study add to the growing body of evidence on the reliability of peer assessment. While acknowledging the many advantages of peer assessment, we suggest that it is important to consider some of the potential limitations of peer assessment:

1. The quality of feedback: the assessment of a clinical interaction needs to be accompanied by meaningful feedback that would enable the student to improve. Are peers at similar levels of development able to provide this level of meaningful feedback?

2. The authenticity of the assessment experience: the perceived authenticity and weighting attached to a peer-assessment activity is different to that attached to an assessment by a clinical teacher. This has an impact on the ‘backwash effect’ of assessment, which plays a significant role in driving learning. Peer-to-peer interactions are limited in their ability to perform this function.

3. Reduced opportunities for student–clinician interaction: standard clinical assessment usually involves a multi-dimensional clinical experience that is guided by a clinician. This usually involves analysis of clinical issues beyond the narrow aspects being assessed, and is a critical component of clinical learning. Peers with limited clinical experience are unlikely to be able to facilitate such an experience.

4. Perpetuation of misconceptions that exist among peers: misconceptions around various aspects of clinical presentations are common amongst medical students, and these are often based on popular media. Peer-assessment is unlikely to identify and address these misconceptions, and may play a role in re-iterating and perpetuating these misconceptions.

We therefore contend that the limitations of peer-assessment need to be recognised in any decision to increase its use in medical education. We are particularly concerned that the move towards peer-assessment may be driven more by workload considerations rather than by educational considerations. As the authors of the study recognise, this is a major driver of peer-assessment, and we need to be cognisant of the trade-off that may be associated with such a move. An acceptable level of reliability should not be the sole determinant of the quality of assessment.

Declaration of interest: The authors report no conflicts of interest.