Adding to the research excellence framework complexity will be the work involved in making sure that citation counts are accurate. Thomson ISI has two different tools for determining citation numbers, "citation analysis" and "cited reference search", and they can return quite different results. The former counts citations only when they refer to a paper in the ISI database; the latter is more inclusive.
In a citation analysis of my publications, one 2005 article apparently has been cited only three times. But a search shows an additional 13 citations, 12 of which correctly list the digital object identifier. The reason for the discrepancy appears to be that ISI indexes the publication as Journal of Geophysical Research - Solid Earth whereas the 13 missing citations refer to the Journal of Geophysical Research .
Will authors have to check all their citations and try to get mistakes corrected by ISI? Or will institutions have to provide citation counts per paper backed by evidence that the citations exist?
Regardless of the mechanism, the time commitment across the community will be substantial.
Sandy Steacy, Professor of earthquake physics Ulster University.