Dan Ariely, alleged fraudster

I've read several books by Dan Ariely, and came to appreciate him as the seeming dean of a younger generation of behavioral psychology researchers, standing on the shoulders of giants like his mentor, Dan Kahneman and Danny's collaborator, Amos Tversky.

I'd even referenced him in an newspaper column about 18 months ago in the early days of coronavirus, specifically citing his "Predictably Irrational." Extended thoughts in that vein are here. A non-Goodreads/Amazon review is here.

The one book of his that I have reviewed at Goodreads is ...  wait for it below ...

And, now, he has been accused of research fraud. And, if this finding is correct, simple and childish research fraud. (For a deeper dive, here's the blog, with the particular blog post, that first discovered this.)

"WHY?" comes immediately to mind.

My best guess is that he'd gotten himself up a creek without a paddle on some previous claims, hitting the rapids or even the full waterfall version of Replication Bias, and decided to fix things.

But, really? Just adding numbers to a column in an Excel spreadsheet? Wow.

There's also a fat-assed Irony Alert. The allegedly fraudulent research was used in a 2012 co-authored paper about ... 

HOW TO REDUCE DISHONESTY!

And, that one book of his I reviewed? "The Honest Truth about Dishonesty." AND, a paragraph of the review needs quoting:

He also notes that, short of sociopath types, while most of us cheat, most of us don't cheat that much. For instance, on matrix-completion tests where people are paid based on personal claims of number of tests completed? They'll claim to have done a couple of extra, but not double what they actually did.

AND AND AND!

The alleged fraud was in 2012. Guess what year that book was published. So, that just confirms things. He had said shit in the book that wasn't true, I presume, had now learned that it wasn't true, and was in CYA mode.

And, the #fail was indeed childish. Beyond just adding numbers to an Excel column, a different font was used for these new numbers. (Sadly, it was Calibri, then Cambria; no Comic Sans involved. That said, that's like a modern version of one of the classical manuscript errors. I most font sets, Calibri and Cambria are adjacent to each other in a scroll list; this is like a medieval scribe either skipping or reduplicating a word.)

And, it's not like Ariely was a scrubbeenie. He was a tenured prof at Duke; he'd taught there before and came back, per Wiki.

Beyond the link above? Story is now in the Economist, which also notes the 2012 hypocrisy squared book publication date. Wiki says he's asked for the paper to be retracted.

So, the "alleged" is in the title and the body only for purely legal reasons.

Per Proverbs?

"The guilty flee when no one pursues."

 

==

 

As far as his overall writings? "The Honest Truth about Dishonesty" is obviously trash. I assume that anything later also is. "Predictably Irrational" is enough earlier it may well be OK. But, maybe not. And, shouldn't we just start calling Ariely himself "Predictably Irrational"?

 

Since Ariely wants the paper retracted, after first tweeting Retraction Watch during the day yesterday, I posted a comment on their most recent post last night. It went into moderation, I guess because of the link. And, as of this afternoon, it had NOT posted. So, I posted a comment on their FAQ / comments policy page. And, it went into moderation.

Anyway, the story has now metastasized, whether Retraction Watch runs anything or not. Samuel Goldman, going beyond the replication bias, argues that what we have is a clear case of "star professor" sucking up the oxygen of grants, etc., and having incentive to Lather.Rinse.Repeat.

Manipulation to produce favorable results is exactly what behavioral economics would predict in this situation. So it's ironic that one of the leading practitioners of the field seems caught in a trap that his own ideas help explain.

Goldman also goes "there" with Ye Old Petard Hosting reference:

The solution isn't gimmicks like honesty statements. Reforming social science means changing funding and professional incentives that encourage dishonest behavior by concentrating rewards on the biggest names, most charismatic personalities, and splashiest arguments. In other words, the benefits enjoyed by Ariely himself.

Well, it's true!

And, backed up by the Daily Mail. (If they're doing a story, we're officially in dogpile time.) It says that one co-author expressed concern about the data at the time, but Ariely said "We're all good."

And, NEW shit has been dragged up. His Wiki page, newly updated, references an Israeli journal that says while Ariely was at MIT, without institutional review board approval, he used electric shock on participants in an experiment.

And, via the "fonts" link above, I don't need to be harshing Retraction Watch's mellow. It reported earlier this year about ANOTHER Ariely issue: a 2004 paper that got an "expression of concern."

 

BuzzFeed has a piece that documents more replication issues. And, per that, we probably should just cut to the chase and call Ariely an outright liar. Ariely is, indirectly, trying to blame the insurer for whom he did this study. But, he's done that in the past — until busted. In a study a little over a decade ago about dentists allegedly themselves not always knowing when a cavity is a cavity, when issues arose, he blamed Delta Dental. Problem? Delta said it had never given him the information he was talking about.

And, with all of this new information? The note of concern is for work from before Predictably Irrational was published. So is the unauthorized electroshock. So is the fake cavities research. So, we can't trust it, either, can we? And, personally, having referenced it in a column, this is kind of like a source going bad.

Per Goldman, academia is also very political, and I suspect some long knives have been laying in wait for Ariely, re why this is so popular. And, I think a lot of people have probably had suspicions for a long time. That said, will Duke put him on some sort of probation?

That said, maybe we should thank Ariely in a way. He's shown that, if the emperor of behavioral economics and psychology isn't totally naked, it's probably about as thinclad as that of evolutionary psychology. After all, others, including one of his aforementioned mentors, Dan Kahneman, have had their own replication problems (though they didn't make up data to try to hide that, as far as we currently know).

That said, outside of academia, through jobs like "chief behavioral officer" at Lemonade, he's been laughing all the way to the bank, and the statute of limitations for suits over some of this have certainly expired.