Data of dubious value

You reported the headline conclusion of a Research Fortnight article ("Russell Group gets smaller share of council cash", 1 May) without referring to our letter of correction regarding this article where we showed that the analysis did not support such a conclusion.

Not only did this year's Higher Education Statistics Agency data show that the Russell Group's average share of research council grants increased, but the methodology used in the analysis was very problematic.

Comparison of research council grants for just two consecutive years will never accurately describe an institution's success or otherwise. Figures for research grants are volatile because a few large individual grants for an institution will inevitably skew the figures for that year. The data reported can in no sense be understood to represent a trend.

Research Fortnight suggests that the data "tend to raise questions, rather than answer them". We agree, and find it all the more difficult to understand the conclusions drawn.

Malcolm Grant, chair, and Wendy Piatt, director general, Russell Group.