Choosing a journal V: impact factor

8th Feb 2012

This the fifth post in my series on choosing a journal, following posts on getting your paper published quickly, getting it noticed, practicalities, and peer review procedure.

It is all very well getting your paper seen by lots of people, but will that lead to an increase in your reputation? Will it lead to that all-important grant, promotion or university rating?

The impact factor of a journal is a measure of the average number of citations of papers published over the previous two years in the year being measured. A very common view among academics is that having your paper published in a journal with a high impact factor is the most important thing they can do to ensure tenure, funding, promotion and generally success. And in fact the impact factor of the journals your papers are in still has a big influence on many of those whose job it is to assess scientists (as discussed recently on Michael Eisen’s blog). It is also a factor in whether librarians choose to subscribe to a journal, which will affect how widely your paper is seen.

So even if the impact factor has flaws, it is still important. However, remember the following caveats:

  • Citations are only a proxy measure of the actual impact of a paper – your paper could have an enormous influence while not being cited in academic journals
  • Impact doesn’t only occur in the two years following the publication of the paper: in slow moving fields, in which seminal papers are cited five or ten years after publication, these late citations won’t get counted towards the impact factor so the journal’s impact factor will be smaller than justified
  • The impact factor measures the average impact of papers in the journal; some will be cited much more, some not at all
  • There are ways for journals to ‘game‘ impact factors, such as manipulating types of article so that the less cited ones won’t be counted in the calculation
  • The methods used for calculating the impact factor are proprietary and not published
  • Averages can be skewed by a single paper that is very highly cited (e.g. the 2009 impact factor of Acta Crystallographica A)
  • Although impact factors are calculated to three decimal places, I haven’t seen any analysis of the error in their estimation, so a difference in half a point may be completely insignificant
  • New journals don’t get an impact factor until they have been publishing for at least three years.

So although it is worth looking at the impact factor of a journal to which you are considering submitting your paper, don’t take it too seriously. Especially don’t take small differences between the impact factors of different journals as meaningful.

Other new metrics are being developed to measure average impact of journals, such as the Eigenfactor and Source Normalized Impact per Paper (SNIP) and SCImago Journal Rank (SJR). These might be worth looking at in combination with the impact factor when choosing a journal.

Your experience

How important is the impact factor of a journal in your decision to submit there? Have you taken other measures of impact into account? Do you think the impact factor of journals you have published in has affected the post-publication life of your papers?

And journal editors, how much difference does the impact factor of your journal make to how many papers are submitted to it, or to your marketing? Do you know the Eigenfactor, SNIP or SJR of your journal?

Tags: 

2 Comments on “Choosing a journal V: impact factor

Actually, IMHO, the most important caveats of the IF are just three:
The IF is negotiable and doesn’t reflect actual citation counts
http://dx.doi.org/10.1371/journal.pmed.0030291
The IF cannot be reproduced, even if it reflected actual citations
http://www.jcb.org/cgi/content/full/179/6/1091
The IF is not statistically sound, even if it were reproducible and reflected actual citations
http://www.mathunion.org/fileadmin/IMU/Report/CitationStatistics.pdf

sharmanedit says:

Thanks for providing these links, Bjoern. They give a lot of background and detail relevant to problems with the impact factor.

One quote from the first (PLoS Medicine) strikes me as particularly relevant: “…a journal’s impact factor can be substantially affected by the publication of review articles (which usually acquire more citations than research articles)…” This might explain why review journals generally have large impact factors, and it would also mean that the impact factor of a journal that has both reviews and research will be artificially inflated compared with one with just research.

It is clear that journal editors can ‘game’ impact factors a lot, but also that they cannot control or even find out how their final impact factor is calculated. I commend the editors at PLoS and at Rockefeller University Press (your second link) for resisting the temptation to work the system and for instead publishing articles about its problems.

Leave a Reply