A few weeks ago, I proposed that one could rank journals based on **what percent of a journal’s authors have also published in a top journal**. I calculated this statistic for economics and for finance, using the top 5/top 3 journals as a reference point.

Of course, one does not have to give top journals such an out-sized influence. One beauty of this statistic is that it can be calculated for any pair of journals. That is, we can ask, **what percent of authors that publish in journal X have also published in journal Y**? This “journal connectedness” measure can also be used to infer quality. If you think journal X is good and you want to know whether Y or Z is better, you can see which of these two journals has a higher percentage of authors from X publishing there. Of course, with the additional flexibility of this ranking come more caveats. First, this metric is most relevant for comparing journals from the same field or general-interest journals. If X and Y are development journals and Z is a theory journal, then this metric will not be very informative. Additionally, it’s helpful to be sure that both Y and Z are worse than X. Otherwise, a low percentage in Z may just reflect more competition.

With those caveats out of the way, I again used Academic Sequitur‘s database and calculated this connectedness measure for 52 economics journals, using all articles since 2010. Posting the full matrix as data would be overkill (here’s a csv if you’re interested though), so I made a heat map. The square colors reflect what percent of authors that published in journal X have also published in journal Y. I omitted observations where X=Y to maximize the relevance of the scale.

A few interesting patterns emerge. First, the overall percentages are generally low, mostly under 10 percent. The median value in the plot above is 3 percent and the average is 4.3 percent, but only 361 out of 2,652 squares are <0.5 percent. That means that a typical journal’s authors’ articles are dispersed across other journals rather than concentrated in some other journal. This makes sense if the typical journal is very disciplinary or if there are many equal-quality journals (eyeballing the raw matrix, it seems like a bit of both is going on, but I’ll let you explore that for yourself).

There are some notable exceptions. For example, 41% of those who have published in *JAERE* have published in *JEEM*, 54% of those who published in *Theoretical Economics* have published in *JET*, and 35% of those who have published in *Quantitative Economics* have published in the *Journal of Econometrics*. These relationships are highly asymmetric: only 13% of those who have published in *JEEM* have published in *JAERE*, only 16% of those who have published in *JET* have published in *Theoretical Economics*, and only 4% of those who have published in the *Journal of Econometrics* have published in *Quantitative Economics*.

There is also another important statistic contained in this map: horizontal lines with many green and light blue squares indicate journals that people seem to be systematically attracted to across the board. And then there’s that green cluster at the bottom left, with some yellows thrown in. Which journals are these?

I had the benefit of knowing what the data looked like before I made these heat maps, so I deliberately assigned ids 1-5 to the top 5 journals (the rest are in alphabetical order). So one pattern this exercise reveals is that authors from across the board are flocking to the top 5s (an alternative interpretation is that people with top 5s are dominating other journals’ publications). And people who publish in a top 5 tend to publish in other top 5s – that’s the bottom left corner. In fact, if you omitted the top 5s, as the next graph does, the picture would look a lot less colorful.

But even without the top 5, we see some prominent light blue/green horizontal lines, indicating “attractive” journals. The most line-like of these are: *Journal of Public Economics*, *Journal of the European Economics Association*, *Review of Economics and Statistics*, *Economics Letters*, and *JEBO*. Although *JEBO *was a bit surprising to me, overall it looks like this giant correlation matrix can be used to identify good general-interest journals. By contrast, the *AEJ*s don’t show the same general attractiveness.

Finally, this matrix illustrates why Academic Sequitur is so useful. Most authors’ articles are published in more than just a few journals. Thus, to really follow someone’s work, one needs to either constantly check their webpage/Google Scholar profile, go to lots of conferences, or subscribe to many journals’ ToCs and filter them for relevant articles. Some of these strategies are perfectly feasible if one wants to follow just a few people. But most of us can think of way more people than that whose work we’re interested in. Personally, I follow 132 authors (here’s a list if you’re interested), and I’m sure I’ll be continuing to add to this list. Without an information aggregator, this would be a daunting task, but Academic Sequitur makes it easy. Self-promotion over!

If you think of anything else that can be gleaned from this matrix, please comment.