I dislike the two-year window of the Impact Factor calculation because it does not reflect the longevity of a paper’s citation record. Here is an interesting chart of the number of cites to JGR papers in 2012, both from itself (the yellow bars) and from other journals (the blue bars):

This chart is straight from the Journal Citation Report for JGR for 2012. The mode on this chart is 2008, four years back (okay, followed very closely by two years back). The citations to JGR show a relatively fast rise to the peak, but then there is an extended tail lasting for decades. The median of the citations to JGR papers is 10.0 years (what the chart labels as “half-life”), meaning that half of all citations to JGR papers are to those published 10+ years ago. The two lighter colored columns are those used to calculate the 2012 Impact Factor. You can see just how much of the “impact” of JGR is missed by the Impact Factor; only 11% of the citations to JGR are included in the Impact Factor calculation. Even the Five-Year Impact Factor captures less than half of the citations to JGR papers (29%, to be specific).

By the way, as you might expect, the 2012 citations to GRL papers has a lower median age than JGR, of only 7.1 years. This is the anomaly in the field, though…nearly all journals in which space physicists publish have median years back of citations to their papers of 10 or higher. In addition, most have modes out in the 4-7 year range (including GRL, which has a mode in 2007, five years back).

A related plot to show you, just because I have it from the same JCR report, contains the number of cites from JGR to other papers (that is, to papers published in other journals (blue bars) or to papers published in JGR (yellow bars)):

In general, it is a very similar picture to the one above. The mode is in 2010, only two years back, but the median is 9.0 years.

Another interesting fact: in 2012, JGR had an average of 53.1 references per paper published that year. This is well above the GRL average of 26.6 citations per paper, and in general is on the big side for space physics journals. In fact, it is tied for second with Icarus and about ten references per paper behind ApJ.

What does all of this mean? While we (the collective “we” of all JGR authors, not just space physics) cite a few newer papers, we really like to cite old papers, and others like to cite our old papers as well. This is a good thing; I am completely in favor of each author citing the papers that are most relevant to your study, however old the papers may be. I’m just giving you information about the distribution of what we cite and how it is reflected in the Impact Factor and other measures of journal importance.

A 5 year window option is also available. But studies show that journal rankings by IF within a discipline are not changed much by using 2 or 5 years, probably because the citation latency distributions tend to be pretty similar. Maybe JGR v. GRL is an exception. The “sliding max” 2 year window seems like a better way to normalize across variations in the distribution (ref. http://arxiv.org/ftp/arxiv/papers/1304/1304.5101.pdf).

Pingback: New JGR Impact Factor | Notes from the JGR-Space Physics Editor-in-Chief