digital rhetoric

the long tail of research citations

Picked this up from the WPA listserv: a recent article from the Boston Globe cites research that indicates that contemporary scholarly work cites a fewer range of sources than work in the past.

James Evans, a sociologist at the University of Chicago, analyzed a

database of 34 million articles in the sciences, social sciences, and
humanities, and determined that as more journal issues came online, new
papers referenced a relatively smaller pool of articles, which tended
to be more recent, at the expense of older and more obscure work.
Overall, Evans says, published research has expanded, due to a
proliferation of journals, authors, and conferences. But the paper,
which appeared in July in the journal Science, concludes that the
Internet's influence is to tighten consensus, posing the risk that good
ideas may be ignored and lost – the opposite of the Internet's promise.

"Winners are inadvertently picked," says Evans. "It drives out diversity."

I haven't read the Science article (subscription required–talking about ignoring the "internet's promise"!). And the Boston Globe article indicates that there are others with research that counters Evans' claims. So in short, no one knows. However I think there are other forces at play here, one of which is mentioned above.

There has been an explosion of publication, fueled largely by increased pressures to publish at institutions of all varieties. So if there is 3X, 5X, 10X the amount of research being published annually in my field. It quickly becomes difficult to read even narrowly in my specialty, let alone browse through related fields. The network allows this information to be published but then serves to create the long tail effect as we make choices as readers.

That means that essentially 80% of the article reading events in a field will be of the top 20% of the available articles. So let's say that over a certain period of time 100 articles are published. Let's hypothesize that there are 1000 scholars in the field and that they will read 25 of these articles during this period. That's 25000 article readings. 20000 of those readings will be of 20 articles. The remaining 80 articles will be read a total of 5000 times, with a long tail were the last 30 or 40 articles are read only a handful of times.

Now let's say you have 1000 articles instead of 100 but still only 25000 readings. Now you have 20000 readings of 200 articles in the "head" but you have to look closer because 20% of those articles are getting 80% of the readings within the head. So 40 articles are getting 16000 readings, while the remaining 160 are getting 4000 readings.

So in short, it would appear that the long tail effect, if it works strictly by mathematical formula, could work to have this narrowing effect. But it isn't really the long tail that is at work here so much as the brutal forces of an attention economy. The long tail business model assumes that through the network you can reach a larger market. That is the number of available readers should expand with the networking of yoru product.

So let's change the example.

Let's say that in the print world only 1000 interested scholars would have access to your research and the limits of the print economy mean that only 100 articles can be published. That's the first example I gave above. Now let's say that in the online world we can publish 1000 articles (again as above) but this time the increased access of the Internet means there are 50,000 possible readers. (You may scoff at this, but we know more people than this are teaching FYC at this moment–are they not all potential readers of rhet/comp scholarship?.) Suddenly we can say 900 articles see the light of day that would never have been published (or possibly even attempted) under the old print economy and tens of thousands of new readers have access to them.

Now it's quite likely that 99% of those 50,000 readers will never write an article that cites your work. Surprisingly, that's the way it tends to work with most writing. Many people read Time magazine; few people cite it in their research. Still I wouldn't mind being asked to write an article for Time, would you? That's not to say that citations can't be one useful measure of the value of academic work. But it might mean that if more and more academics are being asked to publish more and more, it might make sense for us to think about valuing writing to broader audiences.

In the interim, if you feel like breaking this long tail curve, feel free to be the first to cite this blog post.