Temporal trends in the Journal Diversity Index

Warning: astonishingly trivial

Three weeks ago I showed you my Journal Life List, and I invented the Journal Diversity Index (J/P, where my P papers have appeared in J different journals).  A lot of you liked that and calculated your own JDIs, and I don’t know that we learned anything profound, but it was fun and there’s nothing wrong with that.

But I can never leave well enough alone.

A few folks pointed out that when you publish your first paper, your JDI has to be 1.0 (one paper, one journal).  That’s true!  A few more suggested that as a consequence, your JDI should only decline through your career. That isn’t strictly true, although it’s a tempting mistake.  But it got me wondering about the temporal trend in my JDI, and since I had the data and a graphing package (not R, in case you’re wondering), off I went.  Here’s the result: my JDI from the date of my 1st publication to my most recent (69th, omitting monographs, book chapters, and nonreviewed contributions).  My data are in blue, bounded by the theoretical upper and lower limits in grey (1.0 and 1/P, respectively).

The overall trend in my JDI is indeed downward, but there’s more to the pattern than that.  To start off, my JDI stayed at 1.0 for what seems to me a remarkably long time (through my first 13 publications*).  Then the downward trend started, although it was punctuated by several apparent increasing runs.  I thought those increases would be associated with shifts in my research program – I’ve had quite a few.  But they aren’t; they seem to be nothing more than chance (just like most streaks in sports).  Humans are really good at seeing pattern when it isn’t really there, which is why we need inferential statistics.  While I haven’t done a runs test on my JDI data, I’m pretty sure one would fail to detect significant runs up or down.  Finally, around publication #40, my JDI flattened out; since then it’s ticked up or down a little but stayed between 0.6 and 0.65.

That long period of JDI stasis interests me, because of the suggestion I mentioned that JDI should always trend down.  Of course, if you publish enough papers, it has to.  If there are about 30,000 scientific journals being published**, then once I’ve published 30,000 papers I have no choice but to repeat.  In practice, of course, nobody will ever exhaust all the journals.  That means the temporal movement of your JDI will reflect three factors:

  • the increasing pool of journals in your discipline already on your life list, which makes it harder and harder to submit to a journal not already represented, and hence tends to push JDI down;
  • the advent of new journals, which (assuming that you publish in them) tends to push JDI up; and
  • your disciplinary drift, as you move to new projects in new study systems and new subdisciplines, which gives access to previously unavailable journals and so also tends to push JDI up.

At least in my case, the combination of new journals and a limited academic attention span have been plenty to balance any tendency towards saturation in my journal life list.

OK, what have we learned today?  That JDI needn’t decline monotonically; that JDI can (and I hypothesize tends to) reach a quasi-steady state, but that it takes quite a while to do so; and (mostly) that when I tag a post “Warning: astonishingly trivial”, I’m not kidding. Sorry about that.

© Stephen Heard  August 29, 2017


*^Can anyone beat my record?  I mean, it isn’t a competition or anything.  Unless I win 🙂

**^This is a guess.  There were about 28,000 in 2014, not counting the thousands of fake journals, but I haven’t been able to find more recent stats.

 

Advertisements

3 thoughts on “Temporal trends in the Journal Diversity Index

  1. Pavel Dodonov

    Damn you, Stephen, you just made me waste twenty minutes of my life to make my own plot.
    😛
    But I’m curious, do you know the order of your papers as well as the years and journals they were published in? And is it the order of acceptance or of publication? Some journals take a long, long time to publish a paper once it’s been accepted.
    I don’t remember the order my papers came out, just the years, so I calculated JDI per year since my first paper in 2011: 1.0 – 0.8 – 0.86 – 0.9 – 0.91 – 0.86 – 0.82 – 0.77 (lower bound values of 0.5, 0.2, 0.14, 0.1, 0.09, 0.07, 0.06, 0.05). I have values up to to 2018 because I have a paper that’s been accepted but will probably take a long, long time to become published. In any case, my JDI dropped below 1 at my third, fourth or fifth paper, so you’re in the lead.
    And my comment was even more trivial than your post, sorry for wasting the time of all who read it 🙂

    Like

    Reply
    1. ScientistSeesSquirrel Post author

      The order (within years) is the order they’re listed on my CV, which is the order of acceptance – because when I get final acceptance, my weird little celebration is to immediately add the paper to my CV as “in press”. There might be a case somewhere, where one paper got accepted first but ended up with a later date of publication than another – when the gap between acceptance and print spans year-end. But that would be rare.

      Happy to have wasted time together! 🙂

      Like

      Reply

Comment on this post:

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s