The Generalizability of Cross‐Moderation

Publisher Website
Google Scholar
Abstract
In response to increased public interest in the question of the comparability of the various examining boards' grading standards in the same subjects, the boards have in recent years published numerous reports on their own inter‐board comparability studies. The results of these studies have, in general, been disappointing in terms of the tentativeness of their conclusions about relative standards. This paper discusses the problems involved in attempting to compare grading standards across different examining boards, suggests reasons for the inevitable inconclusiveness of recently conducted cross‐moderation exercises, and, finally, suggests a strategy which would increase the future effectiveness of the cross‐moderation approach to comparability. The strategy involves the application of generalizability theory to the cross‐moderation method, leading to a programme of carefully designed subject‐studies from which the relative effects of some of the known influences on grade variation can be quantified. The results of such studies would then be available for the informed planning of future routine cost‐effective cross‐moderation exercises.