The Wellcome Trust and the ³ÉÈËVRÊÓƵ Funding Council for England are among the bodies calling for the use of the journal impact factor in funding, appointment and promotion decisions to be scrapped.

The metric - which ranks journals by the average number of citations their articles attract in a set period, usually the preceding two years - has become ¡°an obsession in world science¡±, says a coalition of academics, editors, publishers and research funders, in a declaration published on 16 May.

¡°The Journal Impact Factor was developed to help librarians make subscription decisions, but it¡¯s become a proxy for the quality of research,¡± said Stefano Bertuzzi, executive director of the American Society for Cell Biology (ASCB).

¡°The ¡®high-impact¡¯ obsession is warping our scientific judgement, damaging careers, and wasting time and valuable work.¡±

³ÉÈËVRÊÓƵ

The San Francisco Declaration on Research Assessment, which stems from the ASCB¡¯s annual meeting in December, notes some of the metric¡¯s well- documented deficiencies and calls for a greater focus on the content of papers rather than where they are published.

Deficiencies include how easily the metric can be skewed by just a few papers, how it fails to account for the differences between fields, and its potential to be ¡°gamed¡± by editors and authors.

³ÉÈËVRÊÓƵ

The declaration, also signed by Bruce Alberts, editor-in-chief of Science, and Sir Paul Nurse, president of the Royal Society, says that journals should reduce their emphasis on impact factors and make available a range of article-level metrics.

Research assessment should also consider a variety of research outputs, such as datasets, and effects such as influence on policy and practice, while organisations that supply metrics should be open about their data and methods, it adds.

Michael Marks, one of the four editors of the journal Traffic and a declaration signatory, said the group realised that the scientific world had been using impact factors inappropriately. ¡°Initially our gut reaction was to blame the [metric] itself, but it¡¯s not the JIF¡¯s fault,¡± he said. ¡°It¡¯s our use of [it] that¡¯s the problem.¡±

elizabeth.gibney@tsleducation.com