For the last few years, I have been reluctant to publish in academic journals and edited volumes, where the content that I contributed would not be available on the web. I have now some doubts about the wisdom of this position.
If one assumes that citation is a proxy indicator that someone out there is reading what one writes, one could conclude that the more cited articles are the most read. I dislike publishing in platforms that do not make content available on the web because I think that fewer people can read what I publish there. So, it would be reasonable to hypothesize that my work that is freely available on the web, on platforms such as SSRN and in open journals, should be read more and should generate more citations.
A scan of my Scholar.Google profile suggests that this may not necessarily be true. The articles that have generated the highest number of citations are from “not-open” publication platforms, such as conventional journals. But obviously, this has to be corrected for the longevity of the publications as well. The longer something has been around, the more chance it will be read (and cited). So it would not be fair to compare my recent work (after I got on the open publication kick) with the old work written when I still not a slave of university culture. So this would require some correction for age of publication.
Then we have to consider the quality of what we write. Difficult to operationalize, but perhaps my recent work, written under harsher time constraints than when I was in the university, is inferior to what I wrote then. The only way people operationalize this is by taking the quality of the journal as a proxy, which would not quite work with a study on closed and open journals.
The way to get rid of the extraneous factors that could be influencing the work of an individual scholar would be to work with large numbers. This should technically cancel out individual idiosyncrasies.
I do not have the time to do this study, but it seems like there is something for a researcher to work on. The data is available (if one wants additional sophistication, one could do both Scholar.Google and ISI).