There is moaning, wailing and gnashing of teeth on Facebook this morning by political scientists who are troubled by Donald Green‘s request to Science to retract a co-authored article of his that it recently published. Political scientists are expressing concerns along the lines that this is horrible for the discipline, and so on. They are wrong. The detection of fraud is always good. Full stop.
Let’s consider a case that genuinely warranted such concern: the use of the seal of the state of Montana in a mailer sent to voters in a get out the vote experiment. The problem with the decision to use the seal is that it was politically tone deaf, drew the ire of elected officials, and played into the hands of anti-science legislators. Detection of academic fraud is none of these things.
If that’s the case, then why all the concern about this being bad for political science? My best guess is that folks do not recognize an implicit assumption they have about science that is Pollyannaish: like crime, vanity, sloth, and so on, fraud does not occur among my people. This fantasy is, of course, a bunch of unicorns and rainbows. Just as the optimal level of crime and terror attacks is NOT zero, the optimal level of academic fraud is not zero. And this implies that the hand wringing about hoping that deterrence will work next time is foolishness that professional political scientists should be embarrassed to post on social media. Right? Wrong. And I explain why (i.e., discuss norms) below.
Setting aside norms, academic fraud is banal: it would be really odd if it did not happen. Why? Because that would suggest that (1) there are no benefits from publishing influential work or (2) human beings are not involved in the production of scientific work. Neither is true. So let’s now consider deterrence systems. Political scientists study those, right? And the findings are that they work 100% of the time, right? And research into efficiency of deterrence suggests that we should spend infinite resources to design systems that are perfect, right? Dick Cheney is right: deter all threats! The NSA must monitor all communications! I trust I have snarked my point. When we teach and discuss the scientific process we really need to emphasize this.
And that is why the detection and detraction are such a good thing. This genuinely is good news. What is really, really bad is undetected fraud that influences future work. And the great thing about the scientific process is that it is structured such that the really bad outcome is very, very unlikely despite the banal existence of academic fraud. Detection is awesome because it demonstrates that science frigging works! And we all now have an outstanding example we can use to teach our students this. That. Is. Excellent.
OK, so if “everything is awesome,” what’s up with the concern on Facebook? That is norm reproduction. Human societies rely strongly upon norms, and social media produce a new mechanism for norm reproduction. Fables and similar tales were the oral standard, and when our species developed inexpensive means to produce and distribute texts, books, newspapers and magazines replaced the oral tradition. During the 20th Century radio, film and television rivaled and eventually displaced print. The World Wide Web has put individuals back in circulation, and social media has dramatically expanded our ability to cheaply signal our fealty to norms.
That said, what is undoubtedly troubling is the fact that shaming is the primary mechanism by which norms are enforced. And we know that people are low information consumers; that we are drawn to train wrecks and scandal; and that those two in conjunction will lead to lots of academics to connect “academic fraud” and “political science,” and to engage in norm reproduction by sharing their mutual disdain for academic fraud over coffee, lunch and cocktails. This will, of course, happen, and as political scientists our social stature within the academy will dip (and the extent of that dip will vary negatively with the level of information each academic consumes). This is the primary source of the bleating, and I may well value that social stature less than my median colleague, and may thus be inappropriately cavalier. What moves me to write this post, however, is that granting that cost, I think the benefit far outweighs it, and I want to urge my colleagues to push back and remind their shaming interlocutors that detection of academic fraud is good. Full stop.
To summarize, the bleating we see on Facebook is part and parcel of norm reproduction. Yes, our clan will pay some social costs, but such is the process of norm reproduction. More importantly, when we remove our social “hats” and replace them with our social scientist “hats” it should be apparent that a very influential researcher requesting that a very widely read journal retract a very widely reported finding this is nothing but good. The system worked! I may have more to say about this, but I’ll leave it there for now.
 Brief jab at Don Green: That’s theory, bro, it ain’t a casually identified inference.