Research project assessments and supervisor marking: maintaining academic rigour through robust reconciliation processes

Le résumé fourni par la source

Research project modules are a key part of UK undergraduate and postgraduate bioscience degree programmes. Report marking invariably uses two assessors, but marking models are mixed with some institutions using two independent markers and others using the project supervisor as one of the assessors. This latter model is controversial with critics suggesting that it is vulnerable to supervisor bias whilst proponents argue that it ensures subject expertise in the assessment process. Our undergraduate bioscience programmes utilise the supervisor as one of the report assessors, whilst postgraduate programmes do not. With the aim of exploring the impacts, if any, of using the supervisor as an assessor, the grades relating to undergraduate and postgraduate reports marked and reconciled during the period 2011/12 to 2016/17 were compared. Analyses of undergraduate reports (897), showed the grades awarded by the supervisor were on average 2.3% higher (p < 0.01) than those awarded by the independent assessor, although 22.8% of report grades differed by 10% and required reconciliation. After reconciliation the supervisor grades were an average of 1.3% higher (p < 0.01). Evaluation of postgraduate reports (894) showed grades awarded by ‘home’ assessors were on average 1.0% higher (p < 0.05) than the grades awarded by assessors from other universities, but 33.1% of the report grades differed by 10%. Following their reconciliation the average difference was 0.4%. Modification of the undergraduate reconciliation process so the two assessors agreed a single grade resulted in an average 0.5% difference (p < 0.05; Cohen’s d value 0.1), which supported minimal supervisor bias and defended supervisor involvement as a project report marker.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.