Max Bazerman's mea culpa; a day late and a dollar short
Max Bazerman provides his perspective on the Gino Ariely "clusterfake" affair
Max Bazerman of Harvard has written an article about how he got embroiled in the Ariely-Gino “clusterfake” affair. It appears that Bazerman is even writing a book about this.
Bazerman is a senior scholar. He is clearly bothered by the taint on his reputation and legacy. I am guessing this article (and the book) is Bazerman’s attempt to rehabilitate his reputation. If so, then this is a day late and a dollar short. If the article is a good overview of what Bazerman has to say on this topic, then the book will not be worth reading.
Bazerman leaves out key pieces of information
In the abstract Bazerman he promises tantalizingly that he will reveal how the fraud unfolded. But no such details are provided in the article.
In talking about the original PNAS paper, Bazerman mentions “evasive responses” by Ariely. But then goes on to say that another co-author pointed out a resolution and this satisfied Bazerman. But this explanation doesn’t make any sense. For one thing, even if the other co-author’s explanation satisfied Bazerman, why did he not continue to be troubled by Ariely’s “evasion”?
Second, the same group of authors along with a few more co-authors then wrote a paper, again for PNAS, where they argued that their previous results were invalid; that signing first did not lead to greater dishonesty. (And for reasons I explain later, I will NOT provide a citation to this paper; who knows maybe this is fraudulent too!)
At this point, the editors of PNAS asked the authors if, given the second paper, the first paper should be retracted. The authors did nothing till Data Colada revealed the fraud. Why?
Did Bazerman not understand, while writing the second paper, that there were serious problems with the original study?
This appears to be the key to success in the big leagues: the ability to hold two mutually exclusive ideas as being simultaneously true and sell it to the same journal!
Here, Bazerman displays a total lack of introspection. This, in turn, suggests that Bazerman may not have much light to shed on these matters.
Bazerman’s remedies are mostly useless
Bazerman trots out the usual remedies to data fraud; pre-registration, greater transparency about how the data was collected, greater diligence by journals and institutions. Bazerman seems oblivious to the fact that institutions or even journals have little incentive to do any of this.
He has some mildly useful suggestions about requiring co-authors to check for data integrity issues even though it is unclear how this can ever be enforced. I will come back to this later because there may be a way to enforce this, just not in the way Bazerman is thinking of.
But Bazerman resolutely refuses to acknowledge the elephant in the room. What to do when researchers like Ariely, Gino, Stapel, Wansink, Hauser or more recently Asad Islam of Monash University fabricate entire datasets?
How can the safeguards suggested by Bazerman be of any use in such cases of brazen fraud?
No matter how many safeguards we erect, enterprising fraudsters will figure out ways to circumvent them.
I have not read the other articles in that issue of the journal. I am hoping that those other articles have more substantive suggestions than Bazerman’s mostly impractical ones.
What can we do?
Possibly not much. Like many other careers, academic success depends crucially on networks, which is a polite synonym for nepotism. The school you went to and the school you work at matters a whole lot more than they should. So, the people involved in these networks have little reason to call each other out.
But here is my question.
When it comes to academic fraud, somehow people seem to apply the criteria of “beyond reasonable doubt”. This is incorrect. The criteria should be the same as in all civil law cases.
What does the preponderance of evidence suggest?
In the case of Gino and Ariely the preponderance of evidence suggests that they have perpetrated fraud.
Here is another question. Is it reasonable to assume that the fraudsters do this for some papers while they are scrupulously honest about others? Most likely not. They probably approach all of them with the same mindset, except in some cases the results are good and do not need massaging.
I suggest we stop giving them at least one thing they hanker for: citations.
Stop citing these people. Not only the papers that have been retracted or have had questions marks placed against them, but ALL of them. Stop citing Ariely and Gino completely.
Also, bear in mind that most people out there are completely unaware of any of this, and they go about blithely trumpeting the findings of these fraudsters. Point out their mistake to them.
For those of you who write or plan to write popular books, don’t write up the results of single studies as being the truth. Allow for the possibility that the authors may be wrong or worse, may have completely made things up.
If you think the evidence is strong enough, then stop citing some of the co-authors of these fraudsters, at least the ones who have co-authored many papers with them. It is hard to believe that someone co-authored as many as eight papers with them but was completely unaware of the shenanigans. And if they were indeed unaware then what does that say about their diligence, integrity and self-reflection?
In Gino’s case, excluding Dan Ariely, this should include at least some of the following: Leigh Tost, Ovul Sezer, Jane Risen, Lamar Pierce, Don Moore (all with four co-authored papers), Ting Zhang, Lisa Shu and Julia Minson (five co-authored pieces), Juliana Schroeder, Adam Galinsky and Alison Brooks (seven papers), Maurice Schweitzer and Max Bazerman (eight papers), Bradley Staats (nine papers), Maryam Kouchaki (thirteen papers) and Michael Norton (seventeen papers). (Note: I have taken this information from the Many Authors page and take no responsibility for its accuracy.)
What level of preponderance of evidence are you comfortable with?
I suggest that if you are still citing people like Kouchaki or Norton (and possibly Staats, Bazerman, Schweitzer, Brooks, Galinsky and Schroeder), then maybe you should seriously question your own judgement.
Stop citing them and question those who unquestionably cite these people. If enough of us do so, hopefully it will make a difference.
As for me, I intend to remove all references to these people from my teaching and my books. I will scrub all references to these people as I work on the second edition of my behavioural economics textbook, which does sell quite a few copies in English and Chinese.
I think if enough people follow suit this may make a difference.
This is my opinion and I may well be wrong.