“The scientific literature must be cleansed of everything that is fraudulent, especially if it involves the work of a leading academic”

November 30, 2012

(This article was originally published at Statistical Modeling, Causal Inference, and Social Science, and syndicated at StatsBlogs.)

Someone points me to this report from Tilburg University on disgraced psychology researcher Diederik Stapel. The reports includes bits like this:

When the fraud was first discovered, limiting the harm it caused for the victims was a matter of urgency. This was particularly the case for Mr Stapel’s former PhD students and postdoctoral researchers . . . However, the Committees were of the opinion that the main bulk of the work had not yet even started. . . . Journal publications can often leave traces that reach far into and even beyond scientific disciplines. The self-cleansing character of science calls for fraudulent publications to be withdrawn and no longer to proliferate within the literature. In addition, based on their initial impressions, the Committees believed that there were other serious issues within Mr Stapel’s publications . . . This brought into the spotlight a research culture in which this sloppy science, alongside out-and-out fraud, was able to remain undetected for so long. . . .

The scientific literature must be cleansed of everything that is fraudulent, especially if it involves the work of a leading academic.

Sounds familiar? I think it also applies to recipients of the Founders Award from the American Statistical Association. There’s more:

The most important reason for seeking completeness in cleansing the scientific record is that science itself has a particular claim to the finding of truth. This is a cumulative process, characterized in empirical science, and especially in psychology, as an empirical cycle, a continuous process of alternating between the development of theories and empirical testing. . . .

My first reaction was that all seems like overkill given how obvious the fraud is, but given what happened with comparable cases in the U.S., I suppose this “Powell doctrine” approach (overwhelming force) is probably the best way to go.

Meanwhile, I’ve been told that George Mason University has appointed Edward Wegman to its Promotion and Tenure Committee. I’d love to hear what goes on in those meetings. Scary to think that someone’s career is in the hands of someone who writes papers by cutting and pasting from Wikipedia. Perhaps one difference is that in the U.S., everyone’s afraid of a lawsuit. I hear about cases of fraud or theft at Columbia and other universities but the alleged perps seem to stay around forever.

P.S. Again, for those of you who wonder why I keep bringing up these cases, I agree that my reaction seems too strong. After all, who cares about the late-in-life follies of a mediocre time-server, a man whose contributions to research were always more administrative than scientific? As I discussed earlier, I think my big problem with Wegman is that, as Dr. Anil Potti might say, he’s pooping in my sandbox. Although I have a sneaking admiration for his refusal to back down or admit anything, I have no admiration at all for his colleagues at his university and elsewhere who try to push the problem away. The report from Tilburg (quoted above) shows how it should be done.

P.P.S. I agree with commenter Ryan Shaw below that “cleansing” should mean labeling the fraud everywhere it occurs, not expunging it from the record as if it had never happened. Indeed, such expunging would not be possible. If fraudulent article A is cited in legitimate article B, then the citation for A will always be there. The idea is that a search for A should lead to a prominent statement that A was based on fraud. To cleanse here is to label, not to remove all traces of.

Please comment on the article here: Statistical Modeling, Causal Inference, and Social Science

Tags: , ,