Investigation discovers fraud in more than half of first tranche of psychologist’s research papers

The first in a series of academic investigations into the “unprecedented” scientific fraud committed by Dutch former professor of social psychology Diederik Stapel has judged more than half of his publications to be fraudulent.

The investigating committee chaired by emeritus professor in psycho-linguistics Pim Levelt has so far only investigated a small part of his research published while employed by Tilburg University (www.commissielevelt.nl). The information is being published online as soon as the results become available as a service to the scientific community.

Out of 20 publications dating from 2008 to 2011, fraud was determined in 12. These were published in eight scientific journals. In addition, three PhD dissertations, supervised by Stapel, were also judged to be fraudulent.

However he published approximately 130 research papers and wrote a further 24 book chapters. The conclusions of the Levelt Committee to date are just a small “first tranche” of investigations. More results are expected to be published online this month including those from two further academic investigations covering his earlier work while at the universities of Amsterdam and Groningen.

A final report containing an overview of all the results plus some general conclusions and recommendations is due to be presented this summer. This will also consider the responsibility of any of the co-authors of the fraudulent papers.

The Levelt committee, supported by expert statisticians, has considered criteria such as: “conspicuous oddities” in the data; whether the datasets compared with the reported results; or whether questionnaires matched with the content of the articles. Its final judgment, based on a combination of these factors, determines whether the publications were based on fictitious data or studies.

The website then lists all the publications investigated and that it judges are fraud together with a summary of the main reasons such as: “fake data collection”; “highly implausible” results; an “impossible to realise experimental set up,” and an admission of fraud by Stapel.

A spokesperson for Tilburg University, which set up the independent the committee, said: “Levelt wants to make public all available information as soon as possible because of concern for those academics involved. We are concerned about informing them as quickly as possible which of the articles they have published with Stapel are “infected” (by fraud). Also, we want to achieve clarity for others who have worked with him but whose articles are not fraud. The work, careers and reputation of these researchers, some of them very young, are at stake, thus the urgency.”

The new president of the Royal Netherlands Academy of Arts and Sciences, professor in molecular genetics Hans Clevers is quoted this week in the Dutch press describing Stapel as the “proverbial bad apple,” just one among 20 000 Dutch researchers who can bring down what the others have built up. Stapel admitted fabricating data and apologised last October after a report concluded the degree of fraud was “unprecedented” (BMJ 2011;343:d7201, doi:10.1136/bmj.d7201).