September 23, 2018

Brian Wansink, a researcher we reference in our book Influencer: The New Science of Leading Change and corresponding training course, has come under intense scrutiny this year. We want to share our perspective on the different critiques Brian has received.

We use Brian’s research as one, among many examples to demonstrate different sources of influence and specifically how physical factors (Source 6) such as plate size can influence eating behavior. Brian has resigned from Cornell University after a faculty committee found he had “committed academic misconduct in his research and scholarship dating as far back as 2008, including misreporting of research data, problematic statistical techniques, failure to properly document and preserve research results, and inappropriate authorship.”

Summary of Critiques

Misreporting of research data: In two papers, Wansink reported the wrong ages for subjects (reporting them as being ages 7-9, when they were actually preschoolers). While subsequent studies by others have shown that age does not affect the results of these particular studies, this kind of mistake indicates oversights that are likely the result of a factory mentality in Wansink’s lab.

Problematic statistical techniques: Here there are two kinds of problems. First, Wansink has made his share of typos, transposition errors, and statistical mistakes. But that’s not the big deal. The big deal is that he has done what’s called “p-hacking.” He explored his datasets after the fact, looking for significance, instead of confining himself to predictions he made in advance. Technically, this is data mining, not statistical analysis. It’s impossible to prove that this practice affected any of his published studies, but his emails and blog postings suggest that it did.

Failure to properly document and preserve research results: Most journals don’t ask researchers to preserve their raw data (original paper and pencil surveys or coding sheets), but some journals do. Wansink published six articles in JAMA (Journal of the American Medical Association), which requires researchers to keep their raw data, but he failed to preserve this data. As a result, these articles have been retracted.

Inappropriate authorship: Wansink was found to be overly generous with authorship, especially at conference presentations. He gave some of his post-docs and graduate students more credit than they deserved—though this is often a judgment call.

Our Perspective

Wansink is an amazingly clever researcher who has singlehandedly created the field of research called “eating behavior.” He has published 899 scholarly articles in reviewed journals, studies that have been cited in more than 25,000 other scholarly articles. Our own conclusion is that, for a time, Wansink became too focused on publishing. His lab became something of a factory, with post-docs running multiple experiments, and churning out articles as if on an assembly line. We believe his academic misconduct was the result of this “quantity over quality” mindset.

As we conducted our own research for Influencer, we visited Wansink’s lab; we’ve spent dozens of hours talking with him and his collaborators; and we’ve carefully reviewed and even replicated many of his studies. So, while serious, this doesn’t cancel or negate the contributions he has made. And we don’t believe it reflects the Brian we know and continue to respect.

We are not planning to remove or replace any of the training videos of Brian we use in the Influencer course at this time. These videos have proven themselves to be effective teaching tools in helping people understand the sources of influence that impact their behavior.

In summary, when we reference Wansink’s work, we need to acknowledge his misconduct. And at the same time, we believe his research continues to demonstrate the impact one’s environment has on influencing behavior (Source 6). It’s also important to note that our body of research on influencing change stands independent of Brian’s research.

For more information, and links to articles and responses, please reference the sources below:

Washington Post
The Cornell Daily Sun

David Maxfield
Coauthor of Influencer & VP of Research at VitalSmarts