Student Peer Assessment in Higher Education: A Meta-Analysis Comparing Peer and Teacher Marks

TLDR

Forty-eight quantitative peer assessment studies comparing peer and teacher marks were meta‑analysed. The meta‑analysis found that peer marks more closely match teacher marks when global, well‑understood criteria or academic products/processes are used, high design quality improves validity, while hypotheses about advanced courses, STEM disciplines, and multiple ratings were unsupported; differences between self and peer assessments were noted and future research directions suggested.

Abstract

Forty-eight quantitative peer assessment studies comparing peer and teacher marks were subjected to meta-analysis. Peer assessments were found to resemble more closely teacher assessments when global judgements based on well understood criteria are used rather than when marking involves assessing several individual dimensions. Similarly, peer assessments better resemble faculty assessments when academic products and processes, rather than professional practice, are being rated. Studies with high design quality appear to be associated with more valid peer assessments than those which have poor experimental design. Hypotheses concerning the greater validity of peer assessments in advanced rather than beginner courses and in science and engineering rather than in other discipline areas were not supported. In addition, multiple ratings were not found to be better than ratings by singletons. The study pointed to differences between self and peer assessments, which are explored briefly. Results are discussed and fruitful areas for further research in peer assessment are suggested.