Rapid Responses to:

LETTERS: Gianluca Castelnuovo Time for the single researcher impact factor BMJ 2008; 336: 789-b [Full text]

Rapid Responses: Submit a response to this article

Rapid Responses published:

Single Researcher Impact Factor

Andrzej Grzybowski   (4 October 2008)

Single Researcher Impact Factor 4 October 2008
Andrzej Grzybowski, Assistant Prof., Dept. of History of Medicine Poznan University of Medical Sciences, Bukowska 77, 60-812 Poznan, Poland Send response to journal: Re: Single Researcher Impact Factor	I have read a letter of Castelnuovo recently published in BMJ with a great interest. [1] Evaluation of a researcher’s work has been warming up discussions among scientists for a long time. The only widely used “solid” indicator of a researcher’s performance is based on citation count and calculation of journals’ impact factor invented by Thomson-Reuters. Since its introduction in the middle of the 20th century, Journals Impact Factor has become widely used as an indicator of "scientific performance" and at the same time widely criticized and contested. [2,3] It is often forgotten that the Journals Impact Factor was initially designed for evaluation of popularity and ranking scientific journals as an aid to university librarians to help them decide which of them to subscribe to the local library. Due to lack of any other “objective” measuring tool, scientists adopted journals impact factor as an indicator of their own "performance". Just by reminding simple facts that only 20-30% of published articles "work" for a journal's impact factor, the rest of them receiving only single citations or no citations at all, one can draw a conclusion that the journals impact factor says nothing about the quality of individual articles, and definitely cannot be an indicative of the “quality” of their authors. There are also more and well known fundamental problems with citation-based impact factor. It can be easily manipulated by journals and authors. Most common manipulations include self-citations, negative citations and out-of-context citations. Publishing special types of articles like reviews for example, also influences the impact factor. But the most important disadvantage of journals impact factor is that it does not reflect directly the "quality" of the authors. Graczynski [4] has recently proposed an interesting and simple solution - the personal impact factor, which is associated with a scientist rather than with a journal. Basically, what he proposes is a very similar formula of calculation of a “personal” impact factor with a single yet a significant difference introduced to the methodology used by Thomson-Reuters: the subject of the calculation is a scientist [an author] rather than a journal. Hence the personal impact factor of a scientist for a given year is the number of citations in that year in papers published in the two previous years divided by all papers published by this author in those years. Such a calculation reflects a truly personal impact factor. One may ask why this simple idea has not been introduced earlier by Thomson. The answer lies behind the approach to validation of data. Thomson’s publication-oriented databases have not been able to separate articles published by authors of the same last name and initial. The methodology proposed by Graczynski seems to have overcome this problem. Bibliography: 1. Castelnuovo G. Ditching Impact Factors: Time for the single researcher impact factor. BMJ. 2008 12; 336 (7648): 789. 2. Rossner M, Van Epps H, Hill E. Show me the data. J Exp Med. 2007;204(13):3052-3. 3. Brown H. How impact factors changed medical publishing--and science. BMJ 2007; 334(7593): 561-4. 4. Graczynski M. Personal Impact Factor, Med Sci Monit 2008;14(10): ED1-2 Competing interests: None declared