Grade awarding judgements in differentiated examinations

The consistency of grade awarders’ standards and the application of these standards to define comparable levels of achievement in differentiated examinations are investigated. An alternative statistical definition of comparability between grades from different versions of differentiated examinations is considered. Little variation in grading standards between teams of awarders was found for the examination as a whole, but some variation was apparent at individual grade boundaries on some components. In general, a higher proportion of candidates reached each grade boundary on easy components than on harder ones. It was thought probable that the standards that awarders apply to complete examinations were not always appropriate when applied to components which test only part of the syllabus. It is suggested that comparable standards of performance on different components should not be defined solely in terms of the awarders’ judgements; the need for statistical comparability based on the candidates’ responses should be taken into account.

Share this page