Verified by 
Personality
What to Do About Scientific Fraud in Psychology Part II
Power corrupts in science too.
Posted November 2, 2011
Psychological science has made the headlines in the newspapers in the Netherlands again....but for all the wrong reasons.
In September, the esteemed Dutch psychologist, Diederik Stapel, was removed from his professorship position at Tilburg University because of allegations of serious academic fraud. An independent committee (chaired by the retired professor Levelt) was established to investigate these allegations and examine the scale and nature of the fraudulent behavior.
This committee published an interim report last Monday 31 October, based partly on information from Stapel himself, interviews with many of his junior and senior colleagues as well as analyses of his publications. It is not a nice read I am afraid to say. Here are some impressions and conclusions from the report.
The first conclusion is that the scale of the fraud is massive. It has been labeled by some science journalists as the worst fraud case in science history, yet I am not sure whether or not that is true. In any case, it has been proven that Stapel fabricated datasets in several dozens of studies stretching out over a period of many years at both Tilburg University and Groningen University (where he worked previously). Some of these articles with fake data have appeared in high profile journals such as Science and the Journal of Personality and Social Psychology. These articles are now being withdrawn. Many of the PhD-students he supervised have also been affected because he provided them with fake datasets that they subsequently (and unknowingly) put in their dissertations and in their publications with him.
The second important conclusion is that there is no evidence that Stapel had any accomplices. He basically committed the fraud by himself and although there were a lot of co-authors involved in publications with him there is no reason to assume that they knew anything about the fraud nor that they could have known. The latter is important. According to the committee the fraud was quite sophisticated.
For instance, in working with his collaborators such as PhD or Masters students, Stapel had them prepare all the materials for the studies, including the surveys and gifts for the participants, which they then put in the trunk of his car. But rather than drive to the high schools where he allegedly distributed the surveys among the students, he simply fabricated the datasets and gave his collaborators the data or in some cases only the statistical analyses to write the paper.
His "justification" for this rather peculiar practice was that (a) only he had privileged access to high schools, (b) he had many research assistants to help (paid from the grants that he received to do research), and (c) that writing the publication was a more important activity for the collaborator than putting in the efforts to collect the data. The committee's report concludes that in fact there were (a) no schools involved (b) no research assistants, and (c) no data collected. Yet there were always wonderful results coming out of these "school" studies which, in retrospect, were simply too good to be true.
The third important conclusion from the report has to do with the scientific climate in which the fraud could take place. How could he get away with this and for so long?
The committee's answer is twofold. The first answer has to do with the way Stapel worked and the way he abused the power and prestige associated with his professorship position. In the Netherlands university system, the professor is usually the head of a research group and wields almost absolute power over hiring and promotion decisions. So this was a good deterrent for junior researchers who would dare to criticize his working methods.
Stapel was perceived as a highly successful international scholar and due to his prestige, and even charisma, no one really questioned his integrity. There were a few cases, according to the report, in which juniors dared to question him, asking to see the original data that he had given them, but in a private conversation he accused them of questioning his integrity and told them it would be in their own interests to get this stuff published.
According to the committee, the second reason that the fraud could go undetected for so long has to do with the failure of the scientific peer review system.
After looking through the published articles that Stapel authored, the committee noticed some rather odd things in his publications (which could have alerted co-authors or reviewers).
For instance, it was quite remarkable that (now I quote from the report):
1. "Stapel (supposedly) collected the data all by himself, and processed and coded all the data himself or under his direct supervision by unknown assistants."
2. "The experiments were too complicated for school children, there were no completed questionnaires, and schools were never mentioned by name."
3. "The data are too good to be true, almost all the hypotheses are confirmed, the effects are very unlikely and impossible and there are no or hardly any missing or out-of-range data.
4. "Strange, improbable or impossible data patterns, odd correlations, exactly identical means and standard deviations, univariate distributions of odd variables."
The final part of the interim report contains some recommendations for the future. I am not convinced about whether these really help to avoid such practices in the future but they are a start at least.
So what now? There is a final report to be published with a verdict on all Stapel's publications which is due to be published early next year, I believe.
My first impressions as a social psychologist is that Stapel's scientific fraud reads like the classic case of the abuse of power by a smart and very ambitious man with a dark and Machiavellian personality who has done terrible damage to the field in general and to the people who trusted him as their supervisor or colleague (perhaps even without him realizing the consequences).
From a science viewpoint it seems that the field of psychology lacks some of the checks and balances that are available in other disciplines which could help to avoid such extreme cases of fraud such as sharing data, encouraging whistleblowers, identifying the contributions of co-authors to publications and rewarding teams instead of individuals (see my previous blog "What to do about scientific fraud in psychology"). The field would be well advised to take some steps in those directions.
As a scientist, I am also encouraged that this committee has done such a thorough and fast job, which suggests that maybe the system is working in some way.
At a personal level, I am encouraged that the committee has recommended that the many PhD students who have been affected in their academic careers by "Stapelgate" can keep their PhD's. They are suffering enormously from the withdrawal of articles that they have built their academic careers on (at least if they were based on fabricated data from Stapel) and deserve time and opportunity to rebuild their careers.
There is going to be a lot of soul searching in academic psychology in the Netherlands and abroad about the Stapel case and maybe the field will eventually get stronger as a result.
