In the RAE 2001, widely differing proportions of staff were judged internationally excellent in different academic subjects. This has important financial and other implications, and in 2008 a hierarchical structure will aim to increase consistency. We look at those subjects that will be part of a super-unit with social policy in 2008, and consider what objective rationale may explain the wide differences in ratings in 2001. A quantitative analysis found no such rationale, indeed some variations appear perverse. This highlights the need for greater consistency of judgements, perhaps steered by a ‘super-panel’.