J. Gowrishankar

In the last several years, the entity ‘journal impact factor’ has begun to take on a larger than life dimension in the assessment of the quality of journals and also of the research output of nations, their constituent institutes and universities and departments, and even of individual scientists therein. So much so that I know of several scientists who base their decisions to send manuscripts to particular journals because of decimal differences in the impact factors between the different journals.

The impact factor is defined as the mean number of citations received in the current year by papers published in the journal in the preceding two years. I have recently come across two rather unrelated articles that discuss the relevance and value of the impact factor in assessment of research quality, which I wish to share with the readers.

The first¹ is a laudatory review published in Nature of a book authored by J. M. W. Slack, ‘Eggs and ego: an almost true story of life in the biology lab’. Unfortunately, I have not had the opportunity to read the book itself, but I quote from the review by Lawrence¹: ‘Nowadays, as Slack points out, assessment of researchers is not by the content of the articles, not even by their titles, but just by the names of journals in which they are published (sic). Slack picks the top ‘‘fashion journals’’ in biology as Cell, Nature and Science, which have high impact factors. Yet the impact factor is determined not by the bulk of the papers in the journal, but by a few heavily cited ones. Thus, most of what the fashion journals contain is actually the same sort of thing the specialist journals carry but dolled up to look a bit special. Slack’s [own] papers published in specialized journals were just as good as those [he] published in the fashion journals, both in his opinion and as measured by the citations they attracted’. Lawrence also states that ‘in a creative industry like research, where real discoveries are always ahead of their time, these measures [such as the impact factor] are at best crude’.

The second article I wish to cite is a commentary² by Graham Walker, editor-in-chief of the Journal of Bacteriology and I quote some excerpts: ‘It is strikingly obvious how biasing one’s editorial decisions to favor only fast-moving areas of research  . . . could greatly increase a journal’s impact factor . . . [On the other hand,] we publish truly the best papers in the field but do not bias our decisions by considering the perceived popularity of the topic . . . . The enduring legacy of this editorial tradition has been the publication of important papers that continue to generate citations for many years’. Walker then provides data which show that: (i) approximately 2200 articles published in Journal of Bacteriology in 1995 and 1996 garnered 7700 citations in 1997, for an impact factor of 3.5; (ii) these 7700 citations represent only 17% of the citations in 1997 to all papers published in the journal, with the remaining 83% (which one may call the ‘forgotten citations’) citing papers published in 1994 and the years preceding; and (iii) there were, respectively, around 2000 and 2400 citations in 1997 to the 900 papers published in 1988 and the 1000 in 1989, for an ‘endurance impact factor’ (my own jargon) of 2.3. So then, should we say that there are the hares as well as the tortoises among the different journals?

In our country, some scientometrists and science administrators have gone one step beyond the impact factor to employ derivatives, such as average impact factor per paper published (which in my opinion is nonsensical), to compare the research outputs of different institutions. One wonders whether to such a science establishment (as Lawrence¹ puts it), ‘the real purpose of the endeavours becomes forgotten and it devotes itself, not to making the important measurable, but to making the measurable important’.

Lawrence, P. A., Nature, 1999, 397, 487–488.
2. Walker, G. C., J. Bacteriol., 1999, 181,
   1–3.

J. Gowrishankar is at the Centre for Cellular & Molecular Biology, Uppal Road, Hyderabad 500 007, India.