ÁñÁ«ÊÓƵ

We at Evidence strongly agree with Andrew Oswald that the quality of articles withinjournals is a critical factor in the use of citations ("Journals study raps snobbery", May 19).

For that reason, when we supply bibliometric databases to institutions, we always provide both the observed citation count for an article and the expected citation count for that journal in the year of publication. The relative impact can then be judged in context.J We believe that it is vital to get each paper in the right vehicle for the right audience - and that may mean using journals that are widely read by research users as well as researchers, even if these have lower "impact factors".

Jonathan Adams Director, Evidence Ltd