Earlier this year, a leading researcher of the practice of science Daniele Fanelli published a rather provocative paper in Plos One. In it, he essentially maintains that, for all the claims about the perverse “publish or perish” philosophy (PoPP), if the publication rate is adjusted by the number of coauthors, there is no evidence of it actually pushing scientists to publish more.
It is not the point of this post to argue with the finding that the individual publication rate has not increased but there are potential other negative effects of the PoPP that Fanelli doesn’t touch upon, and that do not result in higher individual publication rates.
1) Publishing with co-authors may have substantial costs. For instance, one may have to adjust her ideas to achieve a compromise. It may also push the overall bias in research to less controversial or more easily verifiable hypotheses because it’s easier to convince potential coauthors to work on them.
2) Through creating an intense competition as to who can publish more, PoPP may result not in higher individual publication rate but in a higher rejected paper rate. Which would mean that a lot of effort is wasted.
3) PoPP may fuel proliferation of more sub-domains or interdisciplinary domains of study and more journals dedicated to publishing papers from them than is actually justified. A quick glance at the names of economic journals in the IDEAS/Repec ranking (with the likes of Journal of Accounting and Economics, Journal of Risk and Uncertainty and Small Business Economics) suggests that something might be amiss. Risk and uncertainty are inherent in any choice, and accounting and small business don’t seem specific enough to warrant issue-specific journals.
4) PoPP may result in promotion and glorification of more prolific researchers at the expense of ones with profound and truly original ideas that require a lot of work. This is another way in which PoPP may be pushing the overall direction of research towards more easily verifiable hypotheses and the cult of statistical significance.
Some people might say that (2) is implausible. Does not competition for producing a good normally result in getting more of it produced per producer? If we look at the history of computers or mobile phones, we can see that intense competition in both domains led to almost any current producer producing far more than the early ones. However, this has not been true in all domains. For instance, despite the fact that there are undeniably more restaurants today than, say, 50 years ago, this probably does not mean that an average restaurant today serves a lot more customers than back then.
Ultimately, PoPP is a beautiful manifestation of the profound problem with any non-profit activity brilliantly described by Ludwig von Mises in his works on bureaucracy. While for-profit activities have the imperfect gauge of profit to guide them, the only resort non-profit activities have are various sorts of bureaucratic criteria.
Individual publication rate is one such criterion. It is attractive because it is relatively easily quantifiable and unambiguous. But it is still fairly arbitrary in terms of determining whether the ideas and efforts of a particular scientists have created value. The way to partly remedy the problem may be in tying scientific research more tightly to the creation of real value, as Daniel Sarewitz suggests. Needless to say, though, that, having said what I said about bureaucracies above, I completely disagree with Sarewitz’s suggestion (bring back the Cold War-era DoD) as to how this should be achieved.