DSM-V, medicalization of bereavement, and bias

courtesy Flickr user focus2capture

The ever-excellent Mind Hacks blog has a good post covering a Washington Post story on the elimination of the bereavement exemption from the diagnosis of major depressive disorder. The DSM-V panel has significant ties to big pharma, making it easier for normal sadness to be medicalized and treated with drugs. It also points out the importance of understanding implicit bias. Check it out here.

Stapel, Fraud, Inflated Type-I Error, and the Future of Social Psychology

This week the final report on the Diederik Stapel debacle was released. The press release is here, and you can download the whole report here. From the press release:

The Committees identified 55 publications in which it is certain that Stapel committed fraud during his time in Groningen and Tilburg. In addition, eleven older publications by Stapel published when he worked in Amsterdam and Groningen show indications of fraud. The earliest dates from 1996. A total of ten doctoral dissertations supervised by Stapel are ‘contaminated’ (seven in Groningen and three from recent years in Tilburg).

I thought this might be a good place to compile some links to some of the articles and interesting pieces in this case.

The web page of the Joint Tilburg/Groningen/Amsterdam investigation of the publications by Mr. Stapel is here.

A recent article in The Atlantic titled “The Data Vigilante” covers Uri Simonsohn, who has developed an algorithm to detect anomolies in data that might indicate fraud or inflated Type-I error. (Dec 2012)

A recent post in Science magazine’s ScienceInsider about the affair: Final Report: Stapel Affair Points to Bigger Problems in Social Psychology (Nov 2012)

Here is a link to Simonsohn’s recent article in Psychological Science titled “False-Positive Psychology: Undisclosed Flexibility in Data Collection and Analysis Allows Presenting Anything as Significant.” Here, he details a number of intentional and unintentional ways that results can achieve statistical significance as a product of experimenter effects (he calss this “researcher degrees of freedom”) rather than the experimental effect itself. (Nov 2011)

Here is a good article in the NY Times by Benedict Carey on the initial findings in the case. (Nov 2011)

An article from Nature on the initial report. (Nov 2011)

In addition to Stapel, at least two other social psychologists have resigned their positions amid allegations of fraud. Derek Smeesters, also in the Netherlands resigned after Uri Simonsohn’s data analytic technique was applied to his data. An article about that is published in Science magazine’s ScienceInsider. (July 2012)

This algorithm has raised concern that a new witch hunt may be underway where individuals are selected for unknown reasons to subject their data to this technique in order to discover more fraudulent findings. In fact, a second social psychologist was implicated by Simonsohn’s technique, Mark Sanna at the University of Michigan. Sanna resigned and asked JESP to retract three of his papers. Nature has an article on that. (July 2012)

To add to our paranoia, a post on the Society for Personality and Social Psychology’s listserv came out earlier this year which described emails being sent to individual members requesting their data. These emails originated from “Jay Zimmerman” and “Laurie Rhodes,” both of whom do not appear to be real researchers via internet searches. It turns out this was part of a research project that included deception about the intention of the data collection. This is covered in this Google Groups post by social psychologist Brian Nosak.