The vice-chancellors of Lancaster, Umist and York correctly point out that research league tables are usually based on misleading assumptions about staff not submitted to the RAE. Such people are either scored as zero or treated as having no effect on the rankings. The truth must lie between these extremes.
One might reasonably suppose that the unselected would have been rated, on average, two points below their officially research-active colleagues in the same department.
A league table constructed on this basis is available at www.shef.ac.uk/voas .