Andrew Oswald is right to decry the banishment by many REF panels of data such as citations. No doubt some will respond by pointing to the pilot study that compared the results of peer review with metrics. However, the inference drawn from the lack of correlation between the results of these methods - that it is metrics that are "wrong" - has never been justified and is completely irrational.
Robert Barton, Department of anthropology, Durham University
Have your say