Future UK RAEs to be Metrics-Based

[Stevan Harnad]

On Tue, 19 Sep 2006, Sylvan Katz wrote:

> > The question at hand is a specific and purely empirical one: Is there a
> > significant and sizeable correlation between RAE ranking, as currently
> > determined by the RAE panels, and citation counts?
> 
> Linear correlations can be misleading see
> Katz, J. S., 2006: Indicators for complex innovation systems. Research 
> Policy, 35, 893-909.

You are quite right that linear correlations can be misleading, either because of
an underlying nonlinear relation or because of violation of distribution
assumptions. But that does not mean that they *are* misleading in the case of the
RAE/citation correlations. (Someone should visualise the relationships in each
case, to see if there is any indication of nonlinear components, and of course the
distribution assumptions should be checked.)

Stevan Harnad