First, the good news:an investigation into Brian Wansink’s research practices “found no fraud, no theft, no plagiarism, and no sexual misconduct or Title IX issues.” We ought to recognize the reality of that, that most men have no incident in their past of wrestling unwilling women down and trying to rape them, so we should notice that, especially when some men are trying to pretend that attempted rape and sexual assault are just a phase that all boys go through. By all accounts I’ve seen, Wansink seems to be good, collegial, helpful person, and not a US senator.
But now the bad news: he’s not a very good scientist.
Cornell University has been investigating his research since November. In a statement, the university told BuzzFeed News that Wansink was found to have “committed academic misconduct in his research and scholarship, including misreporting of research data, problematic statistical techniques, failure to properly document and preserve research results, and inappropriate authorship.”
The news came a day after six of Wansink’s papers were retracted, giving him a total of 13 retractions.
Now it’s all over. Wansink has announced his retirement. Just as well, since his idea of research was to take ideas about diet that people wanted to be true, do lots of experiments and observations, and then when they turned out not to be true, finagle the statistics until he got the results everyone wanted to hear.
Under his leadership, the Food and Brand Lab produced studies that reinforced a theme: Simple environmental cues can help people lose weight and eat healthier, without the need for rigorous dieting and intense exercise. It was a theme that earned him coverage everywhere from Good Morning America to O, the Oprah Magazine, to the New York Times. He once led the USDA committee on dietary guidelines. He oversees a $22 million federally funded program to promote “smarter lunchrooms” in nearly 30,000 schools.
But for years, the Food and Brand Lab massaged shoddy data into published, peer-reviewed studies in a brazen ploy for media coverage, as BuzzFeed News has reported.
I guess winning the approval of Oprah and not trying to rape anyone isn’t the same as doing good science.
But but but peer review!?
Hmm. Oprah and Dr Oz are his peers, after all.
@2 In that case he’ll have his own TV show in no time.
How did they find ‘no fraud’? How do the investigators define ‘fraud’?
One paper being retracted could very well be an honest and embarassing mistake from which the investigator and lab learn from. Maybe years later a second paper would need to be retracted after internal due diligence discovers a critical error in the study. After that, I would hope everyone would find any work coming out of the PI’s lab to be highly suspect. How does a PI put out 13 papers that are retracted without committing fraud? What amazing skill set does he have to accomplish this?
I think “fraud” would be something like fabricating data or outright lying about methodology. Sounds to me like he was massaging real data. I guess the distinction is that as long as you can’t prove deliberate fraud, he can always just claim to be innocently incompetent.
But, yeah, 13 papers retracted is bad. Reputation-wise, there’s not much difference.
@whywhywhy #4:
Lots of money and/or low friends in high places?
He oversees a $22 million federally funded program to promote “smarter lunchrooms” in nearly 30,000 schools.
I remember someone on another blog suggesting that “no one takes social science research seriously. Guess he was wrong.
In general , the statement ““found no fraud, no theft, no plagiarism, and no sexual misconduct or Title IX issues.” sounds reasonable (well, I am Canadian and have no idea what “Title IX issues” means).
Wansink seems to have been someone who would “data mine” a data set till he found something with a p => 0.05 even if he had to strangle the data single handed. His results also seemed to defy the basic rules of arithmetic and logic.
The finding that “found no fraud, no theft, no plagiarism” is not surprising. He was an incompetent researcher spinning tales that had no basis plus being a total idiot in the research area ( I was going to say incompetent but that implies some knowledge of the issues and I am not sure about that).
Still he did invent his own fairy tales.
The word I’m hearing here at Cornell is that the food scientists are pissed because they are catching the flack for this even though Wansink and the Food & Brand Lab were in the marketing department, not food science.
I walk by the lab on my way into work. Today the board outside the lab normally filled with papers written by the students in the lab, with group photos and lines connecting the papers to the pictures of the authors, was empty, stripped from what it was just yesterday. The glass case with his books in it was untouched. Sadly, the display glorifying his students, who didn’t do anything wrong except follow the instructions of their professor and mentor, was destroyed while his work still stands.
jrkrideau @ #7: What tipped people off to the problem was him commenting on a blog that he was telling his students to data mine the data until they found something with a p < 0.05, as if that was a good way to generate papers. Many times, a single data set would generate multiple papers based on different things they found with p < 0.05, without retesting specifically for that thing, or even mentioning in the papers where the data came from.