The decline of impact factors
Many, including myself, shall certainly welcome a scientific world in which it won’t be true any longer that “a research published in Nature is, by the very fact, of the highest quality” and that “a young scientist who has published in Nature has far higher chances of getting a job“. But we don’t have to forget that the issue is deeper.
In the past, careers in science were supposedly determined by a panel of wise men (I would like to add “and women”, but it would be an anachronism): as well known, oligarchy is fair only in the eyes of those who share the same wisdom as the oligarchs. Presently, the panel of wise persons is still required for hiring, promotions etc, but there is a request of control by an independent, supposedly neutral authority. This motivates the demand of metrics, may reduce the influence of the whims of some people but introduces other problems. I fear that we won’t hit the perfect system.
Back to the statistical survey: figures 4 and 5 are really intriguing: they indicate that only few of the most cited papers are published in the most cited journals, and the percent is declining since around the year 1990. I am not sure if this is an instance of Simpson paradox… What is even more intriguing is that each figure has two graphs, and it seems to me that, by the definitions used, the two graphs should add up to 100%; but they don’t. So either something is wrong with me, or with this analysis: better finish this post and go back to work.