Inflation of impact factors by journal self-citation in cardiovascular science

Recently, several major cases of fraud in science were exposed in the Netherlands [1–3]. Unfortunately, the case of Don Poldermans from Erasmus University outlined that this scientific misbehaviour has also affected cardiovascular science, [3] after previous shocking cases in medicine, [4] including the Woo-Suk Hwang debacle on cell stem technology published by Science [5] (see ref. [6] for an extensive analysis of the case). During the editorship of Michiel J. Janse (and my associate editorship) of Cardiovascular Research (1995–2002), there were no retractions, although some investigations were initiated. However, there were five retractions over the subsequent 10 years under the co-editorship of Michael H. Piper and David Garcia-Dorado [7]. An increase in retractions suggests that fraud in science is a problem of increasing importance. Of course, editors and reviewers alike have neither been educated to act as data police officers, nor is it their primary goal to search for fraud [6]. Still, they are the only gatekeepers we have for maintaining moral and ethical standards in science when authors fail in this respect. For this reason it is important that editors act responsibly. During the last years several journals were banned from the Journal Citation Reports, a product of Thomson Reuters, which implies that a journal is not assigned an impact factor for several years. The main reason for this was excessive self-citation by journals, which ‘inflates’ the impact factor of a journal compared with its value were it solely based on citations obtained from other journals. The fact that papers in a certain journal are more often cited by that same journal than would be expected from citation by other journals is not a new observation. It has been demonstrated and analysed previously [8, 9]. Of course, both author self-citation and journal self-citations are in themselves useful, relevant and functional when it comes to transfer of knowledge. It becomes a problem when journal self-citation serves a merchandising rather than a scientific goal, i.e. when authors are pushed to include citations with the primary goal of increasing the impact factor of the journal that is going to publish their paper. This could be labelled as ‘coercing citation’.