The Impact Factor (IF) of our journal for 2004 has risen to 1.030. Does this matter? Yes, of course it does, but unfortunately the IF has in recent years attained an undeservedly large significance in all imaginable and unimaginable contexts. I have read an excellent survey of the topic Citation[1], and in this editorial I would like to discuss the use, and above all the abuse, of the IF.
A 50-year history
In 1955 Eugene Garfield designed his IF, and his Institute for Scientific Information (ISI) is still responsible for today's IF. The original idea was that it would be easier to choose journals for coverage in Current Contents and the Science Citation Index, both of them products from ISI. Libraries would also receive help in selecting appropriate journals to subscribe to. The IF was thus designed to be a useful bibliometric measure.
Back in 1873, however, a citation index had already come into use on the subject of law. Frank Shepard was interested in how often cases decided by the Illinois Supreme Court affected subsequent cases.
Simple calculation
So how has ISI calculated the IF for our journal for 2004? In the numerator they have placed the number of known citations in 2004 of articles published in the SJPHC during 2002 and 2003, and in the denominator they have placed the number of articles published in SJPHC in 2002 and 2003. However, the fact is that while the numerator contains citations of all documents, including editorials and letters, the denominator contains only original articles and review articles.
The fact that in this editorial I do not cite the last two years of our journal does therefore not affect the IF negatively, since editorials do not occur in the denominator. If, on the other hand, I had found two relevant references in years 2003 and/or 2004 from the SJPHC, and had referred to them, the denominator would have increased by 2, and thus the journal's IF for 2005 would have risen by 0.025, assuming that we publish 40 original articles a year.
Many limitations
There is plenty of research on the use and abuse of the IF, and the review article alone Citation[1] contains 95 references. Here I shall consider briefly only the most important limitations and draw conclusions about what they mean in practice. Anyone who wants to read more is referred to the original publications.
ISI's IF favours subject fields with a high scientific turnover, as a result of its two-year window, as well as subject fields encompassing many journals. It is thus not possible to compare journals from different subject fields with the aid of the IF.
Furthermore, the IF favours English-language journals, primarily major American ones, although it is a quotient, and review articles; 15 of the 25 journals with the highest IF are review journals. In 1997 the ISI's Science Citation Index was estimated to cover only 2.5% (!) of the world's scientific journals.
It goes without saying that the journal's IF is not representative of the individual article; 15% of the articles thus account for 50% of the citations, and 50% for 90%. So, there are many reasons “Why the Impact Factor of journals should not be used for evaluating research” Citation[2].
It is thus obvious that one cannot compare the scientific performance of individual institutions with the aid of the IF, and even less so the performance of individual researchers. Yet this is done all over the world, and many researchers achieve scientific fame, and are also awarded resources, on the basis of this controversial, 50-year-old bibliometric measure.
Journals of family medicine
presents current IFs, both for journals of family medicine and, by way of comparison, for the more famous general medical journals. According to the arguments above, one can possibly compare journals in the same subject field with each other.
Table I. Impact Factor (IF) for 2004 for journals of family medicine with IF >1, and, by way of comparison, for a selection of general medical journals.
Perhaps it is the case that, of the European journals in the field of family medicine, the British Journal of General Practice is most important, closely followed by Family Practice and then by our own Scandinavian Journal of Primary Health Care. On the other hand, I find it difficult to judge the importance of the three American journals with an IF of the same size, especially since I know them only by name. Based on the reasoning above, it is possible that their IF is overestimated, at least when viewed from a European perspective.
Hints for editors
What can you do as an editor to improve the IF of your journal, even if you are aware that the IF is a dubious measure of scientific quality? Well, in the first place you should publish more articles, especially review articles, and more editorials and letters, but you should eliminate any case reports. In addition, you should increase the total number of references permitted, and ensure that authors do not forget relevant references to your own journal. Finally, you should make your journal immediately available on the Internet.
Anders Håkansson
Anders Håkansson, Professor, Department of Clinical Sciences, General Practice/Family Medicine, Lund University, Malmö University Hospital, Malmö, Sweden, [email protected]
New Danish Assistant Editor
We welcome Thomas Drivsholm as new Assistant Danish National Editor. Thomas has embarked on education to become a general practitioner after eight years of full-time research. His main research area is diabetes epidemiology, his interests also extending to the management of other chronic diseases. In his PhD thesis in 2001 Thomas looked at predictors of type 2 diabetes and at the close relationship between type 2 diabetes and macrovascular disease. Thomas has been active in the Danish College of General Practitioners in various working groups for the last 5 years. Jacob Kragstrup Editor-in-Chief
References
- Cameron BD. Trends in the usage of ISI bibliometric data: uses, abuses, and implications. Libraries and the Academy 2005; 5: 105–25
- Seglén PO. Why the impact factor of journals should notbe used for evaluating research. BMY 1997; 314: 498–502