Skip to 0 minutes and 0 secondsThere's been this rash of much more serious violations of the scientific ethos in social psychology in the last four or five years. There have been fraud cases in other fields but it's gotten a lot of attention in social psychology just because numerous high-profile, tenured famous professors have basically been busted for just making up data in social psychology.
Skip to 0 minutes and 28 secondsFolks at the University of Michigan and leading Dutch universities. And so this is one example, a researcher from Tilburg, who when he finally admitted that he had actually fabricated the data in dozens of projects says, "I have failed as a scientist." I only put this up because it so closely links to this notion of identity and the ethos of being a scientific researcher. This was a superstar researcher in psychology. He was the Dean of Social and Behavioral Sciences at Tilburg. He won multiple career prizes by the time he was in his 30s, had 150 publications, was a superstar in the field.
Skip to 1 minute and 3 secondsAnd as he later admitted, he actually wrote a biography, he traces the point at which he started just making up data, he just started changing his data sets. So he's retracted at least 55 articles. He had all these important articles on racial bias and all these very socially important topics, including from leading journals. He had a retraction from Science, a retraction from all the leading psychological work. The work of all of his co-authors and grad students has been discredited, because he would often handle a lot of the data. In fact, there were warning signs. He never let anybody see the data, even his own grad students. They could help run the labs but he wouldn't share the data with them.
Skip to 1 minute and 47 secondsThat seems like a problem. That seems like a warning sign. He was forced to resign. They took away his Ph.D. They opened criminal proceedings against him, which were eventually settled for community service and some other things. And it's interesting, he wrote this biography where he talks about his descent into doing this, and some of the reviewers– actually, it was apparently beautifully written and some of the reviewers later discovered that whole sections had been plagiarized from James Joyce and other authors.
Skip to 2 minutes and 21 secondsSo it was this just like beautifully written thing about his emotions. So how could this have happened? It turns out data sharing is very rare in psychology. Very few psychologists post their data. Very few psychologists will share their data with you if you ask them, even though the American Psychological Association and other professional societies state that it is an obligation of a researcher to share data when possible. So there was a norm of secrecy in the field as a whole. I mean, it's an incredible departure from Mertonian norms, and any attempt to get data was seen in a very negative light. So it makes you wonder how many other cases there might be in other fields.
Skip to 3 minutes and 2 secondsFor those of us in economics, we're kind of a mixed field. Some journals require data sharing and some don't. The American Economic Review does, the Quarterly Journal of Economics doesn't. Some researchers post all their data; some post none of their data. There's no clear pattern. No clear norm one way or the other. So that opens up the possibility that there are a lot of things that may not be robust. How can we prevent this from happening in the future? I think the easiest solution would be for people to share data and post data, for journals to mandate it, for funders to mandate it, although there are some practical complications there.
Skip to 3 minutes and 44 secondsSo even say in the American Economy Review, even though for ten years people have been meant to post their data for the AER, if you actually try to go to the data archive and download datasets, many of them are basically incomprehensible. So the AER hasn't done a very good job of actually checking that the data is meaningful. So people have sometimes just posted variables with weird names and it's very difficult to actually back out what exactly is going on; other people do a better job. So we're in a period of transition.
Skip to 4 minutes and 10 seconds20 years ago before the internet, before everybody had a supercomputer in their pocket basically with smartphone etcetera, it was actually just logistically much harder to access lots of data and replicate. Now it's a lot easier. So we're in a transitional phase where ultimately we hope we get to a point where data is freely shared, but we're very far from that point right now.
Violations of the scientific ethos in social psychology
In this video, Professor Miguel gives us a particularly astounding example of fraud in social science research – perpetrated by Diederick Stapel, formerly the Dean of Social and Behavioral Sciences at Tilburg University, a prestigious Dutch university. This incident – or rather series of incidents – is unfortunately just one symptom of broader transparency and accountability issues in the scientific community. Professor Miguel asks us to think about how a case like Stapel’s could have happened and introduces us to the importance of open data.