Stevan Harnad complains about "so much noise and so little sense" in regard to the post-research assessment exercise metrics debate, but validation of a metric is not achieved by demonstrating correlation.
When the Treasury suggested replacing peer review by metrics, we pointed out that many factors correlated with the average RAE grade of UK universities and colleges. My favourite was the numbers of gardeners each higher education institution employed, but the number of nursery nurses was also a good indicator of research quality. That no longer seems to be the case, which shows how easily these signals can drift!
Director, Evidence Ltd.