Correlating peer and tutor assessment on a low-stakes engineering assignment.

Abstract

Peer assessment has been a subject of great debate in recent years. The way students perceive assessment and what motivates them when assessing may differ significantly from the tutor. This paper discusses a study designed to correlate students marking with the marks awarded by their tutors when peer assessing one another from in-class oral presentations. A new and alternative approach to correlate results is presented, which is based on the normalisation of the quantitative judgements based on determined criteria. The methodology was blind and holistic, as described in previous works: some guidelines were provided to the students on what is considered acceptable without getting into detail (holistic marking), and peer-assessment marks were made confidential (blind approach). It was observed that students have a tendency to overrate fellow students - especially where lower marks might be awarded. There is, however, direct agreement with the tutors marking in terms of qualitative judgements, which is highlighted by the presented correlation method used to adjust students marks. The presented methodology to correlate marks between the students and tutor showed to be a promising one. After processing the data with this simple and straightforward algorithm, peer and tutor assessment practically showed a perfect match.