Source: Getty
Op. cit. effect: departmental citation counts correlate with RAE ratings
The research excellence framework could be replaced by citation analysis, but only in the hard sciences and only if results are not used to rank quality, a study has concluded.
Four physicists and mathematicians compared the results of the 2008 research assessment exercise - the REF’s forerunner - with citation data from Thomson Reuters.
In a published on the arXiv pre-print server, “Comparison of a citation-based indicator and peer review for absolute and specific measures of research-group excellence”, the authors find a high correlation between the total citation count for papers submitted by a department to the RAE and that department’s quality rating determined by the assessment panels.
The correlation is particularly high for larger groups in the hard sciences.
The REF is intended primarily to inform the distribution of quality- related research funding based on an aggregation of peer-review scores at departmental and institutional level. For this reason, the paper says that “the use of citation-based indicators may offer a much cheaper, and less intrusive alternative to [peer review] for large research groups in the hard sciences”. But it cautions that citation analysis would be “far less reliable for the social sciences and humanities”.
One of the paper’s authors, Ralph Kenna, deputy director of Coventry University’s Applied Mathematics Research Centre, told Times Higher Education that in practice it would be “unacceptable” to replace peer review with citation counts in some subjects and not others because departments might try to choose which panel to submit to based on whether they expected to do better via peer review or citation analysis.
“It would stymie curiosity-driven research and further distort the academic landscape,” he said.
The paper also warns against using citations as a measure of departmental or institutional quality. This is because quality should be a per-head measure, but “almost paradoxically” citations-per-head correlate only weakly with RAE quality scores.
The formal inclusion of citation analysis in the 2014 REF was rejected after a feasibility study on the grounds that citations were “not sufficiently robust” to replace peer review - including in the sciences. According to Graeme Rosenberg, REF manager for the Higher Education Funding Council for England, worries included anomalies such as negative citations, the possibility of self-citation and “citation clubs” and the time lag between publication and the accumulation of citations.
However, the issue is likely to be revisited ahead of the next REF, which is expected in 2020.
