Diederik Stapel, the Dutch social psychologist who admitted to faking data in numerous published papers, has retracted a 2011 study published in the journal Science, which concluded that disorderly environments leads to discrimination. It is the first of what are likely to add up to dozens of retractions.
Science issued an editorial expression of concern for the paper, “Coping with chaos: How disordered contexts promote stereotyping and discrimination,” on 1 November. A day earlier, the Dutch university committees investigating Stapel issued a preliminary report that indicated that Stapel had fabricated or manipulated data in at least several dozen publications, but the report did not name specific papers (see Report finds massive fraud at Dutch universities).
The committees, at the universities of Amsterdam, Groningen and Tilburg where Stapel studied and worked between 1994 and 2011, plan to identify tainted papers in a final report that will not be completed until mid-2012 at the earliest, says Pim Levelt, head of the Tilburg committee and director emeritus at the Max Planck Institute for Psycholinguistics in Nijmegen, Netherlands.
The Science retraction is signed by Stapel, who was a professor at Tilburg until he was dismissed in September, and his one coauthor Siegwart Lindenberg, a professor of cognitive psychology at Tilburg and the University of Groningen. Lindenberg is not implicated in the fraud.
“Stapel’s doing had caught me as much by surprise as it did anybody else,” Lindenberg wrote in an e-mail to Nature. “I never had any suspicion. He was a very trusted man, dean of the faculty, brilliant, successful, no indications for me to be distrustful. In this, I was not the only one. I also had no trouble with the results of the experiments.”
Lindenberg says he and Stapel together designed the experiments, interpreted data and co-wrote the paper, but that Stapel alone provided the data in the form of tables. Nature covered the paper in a news article at the time (see Chaos promotes stereotyping).
This description fits in with a broad sketch of one of Stapel’s tactics that is outlined in the 31 October report. In these cases, co-authors left Stapel to run experiments and collect data with the help of a network of contacts at other universities. Stapel never performed the experiments and instead manufactured plausible data that he presented to his coauthors.
The retracted paper includes five experiments that were said to have been conducted at a train station in Utrecht, in “an affluent neighbourhood in a Dutch city,” and in an unspecified laboratory. The 31 October report noted that Stapel’s papers included vague descriptions of study populations and locations, such as “Universities in the North of the Netherlands.” The Science paper makes no mention of where the lab studies were conducted and mentions only “students” who were “all Caucasian” as the participants.
The field experiments “found” that people were less inclined to sit next to a Dutch-African person in a dirty train station than a clean one, and that people were less willing to donate money to a charity for minorities when they were surrounded by a disordered neighborhood street than a neat one. The three lab experiments, meanwhile, suggested that people form stereotypes to provide structure to a disorganized environment.
Lindenberg says he does not yet know if all of these experiments were fabrications, or just some of them. “[Stapel] did not indicate which of the experiments contained faked data. I would love to know myself but he is not reachable, and I cannot find out any other way,” he wrote to Nature. He plans to repeat the study.
The committees investigating Stapel are sifting through each data set in more than 150 papers to identify those that were fabricated or otherwise manipulated, Levelt says. The process involves gathering data and details about how they were collected from Stapel’s co-authors. “We have to go through each paper and come to a conclusion,” he says. “It’s an immense job,” Levelt tells Nature.
Nature has published two opinion articles on the case. Jenny Crocker, a social psychologist at Ohio State University and past president of the Society for Personality and Social Psychology, looks at the case in the context of research on moral transgressions (see The road to fraud starts with a single step). While, Jelte Wicherts, a psychologist at the University of Amsterdam, says his field needs to be more open with its raw data (see Psychology must learn a lesson from fraud case).
Crocker has another op-ed in the latest issue of Science, which contains an interesting insight: Of the 40 papers Stapel submitted to journals published by the American Psychological Association since 2003, 16 were rejected, 24 accepted. The papers were handled by 25 different editors. “Under such circumstances,” she writes, “it would be almost impossible to detect a pattern of data fabrication.”