What happened? He had all the advantages, he was a deservedly-successful academic researcher, he coauthored a bestselling book . . . and it’s just done downhill from there.
I’m not talking about Levitt retiring from his academic post or deciding not to do research anymore. Doing research is a choice, and unless you’re involved in some urgent project—curing a disease or winning a war or righting some injustice or raising living standards or whatever—or some interesting project—baseball statistics or the theory of random walks or whatever—you shouldn’t do it. I say this in the same way that I say you shouldn’t write a novel unless you really feel you have something to say. Write a novel on a frivolous topic, that’s fine with me, but do it because you have something to say in that format, or just because you think it will be fun. There are enough novels out there. Write your novel or do your research because you have that sense of urgency or curiosity—or if you need to do it to pay the bills. Levitt doesn’t feel the urgency, and his bills are paid by other means, so that’s fine. For that matter, I don’t go around proving lots of theorems. I could spend my time doing that, but it’s not something I feel like doing.
Levitt’s worked hard and now he’s in the stage of his life where he’d like to relax a little more. Fair enough; I don’t work as hard as I used to either. And, in his judgment, he can make the largest contribution to the world as a public communicator rather than as a teacher, adviser, or academic researcher. Fair enough; here I am right now blogging rather than working on statistics research.
So, yeah, my problem with Levitt is not that he’s moved from being a researcher to being a communicator. Communication can be great. My problem is with the content of what he’s communicating, content such as junk climate science (“The PDO cool mode has replaced the warm mode in the Pacific Ocean, virtually assuring us of about 30 years of global cooling” . . . ummmm, uhhhh, maybe not, dude!) and junk celebrity psychology (“And the data keep supporting your model” . . . ummm, no, actually!).
Regarding that latter example, I wrote that it frustrates me when bad science takes up the space that could be occupied by good science. And Steven Levitt . . . he’s got the training and experience to evaluate scientific claims! He could read the relevant papers, he could download and reanalyze what data are available, he could google search for the easily accessible replications and criticisms, etc. He doesn’t have to do that work, but, if he’s gonna ask, “How do you know whether to believe surprising results?”, then he should. Otherwise, why bother? What’s the point of one more puff piece? I just don’t get it. Levitt can play a useful role in the conversation here, and he chooses not to. Really frustrating.
Linguist Mark Liberman picked up on this story and answers my question as follows:
My [Liberman’s] hypothesis about why “Levitt can play a useful role in the conversation here, and he chooses not to” is a simple one.
Levitt isn’t in the business of evaluating scientific experiments, he’s in the business of mass-audience journalism. He wants clicks, and people will be interested, for good reason, in a Freakonomics episode that promises to tell them about how we can “improve our physical health by changing our mind”.
That makes sense. If your goal is to get clicks, whether through happy talk (as with Levitt) or scary talk (as with various purveyors of misinformation), credulity is a kind of superpower. As the saying goes, you can be much more persuasive if you first fool yourself. And as we know from economics (or “freakonomics”), incentives matter! If the goal is clicks, there’s an incentive to be credulous, and you should be able to find a way to get there.
OK, but that’s just sad.
Levitt was a University of Chicago professor, he has (by academic standards) nearly infinite resources, he could study whatever he wants, and he’s going for . . . clicks?
OK, sure, if you have a message you want to share with the world, then it makes sense to go for those clicks. I pour my heart and soul into these blog entries, and I indeed want as many readers as possible. But, the first step here is having a message you want to share with the world. If you have nothing to say, then the clicks aren’t getting you anything.
So, maybe Liberman’s hypothesis is correct. If so, it just makes me sad.
Also Levitt pushing climate change denial . . . that sort of thing can cause real damage; see here, for example.
It’s not too late for him to turn all this around, to confront his mistakes and learn from them. Realistically, I doubt this will to happen, for the same reasoning-from-experience that reputedly persuaded Laplace to assign a very high probability that the sun will rise tomorrow, but it’s possible! We can still hope.
P.S. I’m not saying I think Levitt’s a bad person, nor that I’m in the position to judge such a thing. He’s a free man and can do what he wants! His career trajectory just makes me sad. He’s talented and well-situated in both academia and the media, there are all sorts of interesting things he could be doing, or he could even take a well-earned break, but instead he invests his time and reputation in . . . promoting junk science? What a waste.
It’s the curse of the contrarian. He got famous with some iconoclastic theses in Freakonomics. Then, he got addicted to it, chasing bigger and bigger upheavals of conventional wisdom. Thing is, those are harder to find, and most conventional wisdom is true. So, he ends up stretching too far and arguing against climate change.
Interesting, do you have any other examples in mind? Ioannidis?
Ioannidis was one of the first to point out over-certainty and p values misuse in medicine and biology, not only pop Psychology.
Peter Duesberg was a virologist a Berkeley who claimed that HIV didn’t cause AIDS. My wife knew him from journal clubs and says he always wanted to be the center of attention.
Bob Curry was a geomorphologist at UC Santa Cruz who suffered the same affliction. He studied botany before he went into geomorphology, and got a lot of attention by pointing out that repeated logging would deplete soil nutrients, something that foresters apparently hadn’t thought of before. Ever after, he kept dropping similar bombshells that fizzled.
Basically all climate contrarians started out as decent (good) scientists who then got more addicted to contrarianism than they were to truth. Lindzen, Curry are two obvious examples. It’s a sadly predictable pathway.
Yeah…well the Left has some loopy contrarians also. Look at the Peak Oilers for instance. Matt Savinar literally moved on from Peak Oil grifting to selling astrology forecasts.
https://peakoildebunked.blogspot.com/2013/07/429-oil-drum-bites-bag.html
Alexey Guzey was seemingly on the verge of falling into this curse and becoming the “Sleep is bad, actually” guy, but he seems to have walked back from that.
How are the people in this thread judging who is actually mistaken in the end?
I don’t see any predictions, replications, etc being mentioned. Do you just use how many people currently believe something? Is there anything beyond consensus/authority heuristics being applied?
Anon:
There is literally a graph at the top of this post demonstrating that Levitt was wrong in his notorious post pushing global-warming denial.
Your OP had one but thats not the “thread”.
Also, this reminds me to update that presidency chart. Biden had an even bigger “democrat spike” than Obama, followed by apparently plummeting temps since Trump’s election: https://www.drroyspencer.com/wp-content/uploads/UAH_LT_1979_thru_August_2025_v6.1_20x9-scaled.jpg
> Levitt was wrong in his notorious post pushing global-warming denial.
His notorious post written by someone else? (Or maybe I’ve missed the link to some other post from him.)
Carlos:
Levitt is the face of Freakonomics and, yes, he has responsibility for what’s posted there by his coauthors. Not in every detail, but in the big things, yeah.
Ok.
(ChatGPT tells me that the face of Freakonomics is Dubner though. A second time it gave me both names, the public/media face and the academic/intellectual origin. I tried a few times and it was either Dubner or both… but I’m sure it would eventually agree with you. Anyway, maybe being one of the faces is bad enough.)
Anoneuoid –
Ignoring ocean heat content is bad enough, but tracking SATs in 4- (or even 8-) year intervals is really bizarre.
Oh, wait. You’re actually tracking it in 10-month intervals. Wow.
I really enjoy Andrew’s critiques of Freakonomics. Here is a direction that I haven’t seen him discuss. For me the whole genre was dumb from the get-go. How did they come up with this stupid name “Freakonomics”? What is “freakish” about this project (other than the naivete)? And why is it that doing bad applied statistics is suddenly a branch of economics?
Rodney:
I dunno, I thought that Levitt’s work that went into the original book—the sumo wrestlers and all the rest—was kinda cool. The use of statistical methods for applied microeconomics was not “sudden” and I don’t think Levitt ever claimed it was. I took the point of the original Freakonomics book as being a popularization, with interesting recent examples, of an interesting convergence between basic microeconomic theory (“incentives matter”) and basic statistics (regression analysis, etc.). Freakonomics wasn’t just an pop-econ textbook talking about supply, demand, and the miraculous free market, nor was it a pop-stat textbook talking about all that you can learn from data. It was an interesting hybrid with cool examples and good stories. Yes, it was always ridiculous for an Ivy-credentialed insider to frame himself as a “rogue,” but if that’s what it took to sell it . . . well, there are worse ways to sell books.
Freakonomics was a big first step in the dumbing down of economics and the extensive use of some not-very-good tools for all kinds of topics where the authors knew nothing. It is the lineal grandparent of lots of bad research on crime done by economists. It is the lineal grandparent of much of the terrible “research” put out during the pandemic where people claimed to estimate the effect of mask-wearing or bleach-drinking (just kidding) using their favorite claim about some cut-off etc. So while Levitt may not be a bad person, his self-promotional activities did a lot of damage.
The sumo stuff encapsulates the problem. You can claim this is amenable to economic reasoning. Perhaps. But Levitt knew zero about sumo wrestlers. And that shows up in all of what he, and his students, and his followers do: a quick-and-dirty dif-in-dif (or whatever) about something they have not bothered to understand. You cannot understand identification (think of the exclusion restriction in IV if you are ignorant of the situation).
Heckman is right: Levitt is a plague.
Levitt mentioned that the idea of the name originated from his sister. It does go well with the shtick of the “rogue economist” though.
It you feel so strongly against Freakonomics, why not write a book yourself with “proper” statistics used?
Being an armchair pundit here won’t make the bad science behind pop-psych/econ/stat books go away.
> Doing research is a choice, and unless you’re involved in some urgent project—curing a disease or winning a war or righting some injustice or raising living standards or whatever—you shouldn’t do it. I say this in the same way that I say you shouldn’t write a novel unless you really feel you have something to say. There are enough novels out there.
I’m surprised to see this. My research cures no diseases, wins no wars, has not even a tangential relationship to justice, and does little for living standards. I’ve always been really annoyed by the smarmy and tenuous game of grantsmanship where we pretend that our research should be funded as a cost-effective solution for the great humanitarian project. I do research because I adore discovery and careful thinking. I believe that basic science is valuable—not because it will cure cancer, but because it is good to pursue curiosity and wonder. I always resent when I’m advised that I should pretend my work will cure cancer. I study bugs. This will not cure cancer. Maybe that makes me selfish. Maybe we should reallocate labor away from the wonderful and to the industrious. Turn off the James Webb and go work the food bank. There are enough novels out there. There are enough pictures of stars.
Entomophagist:
Yeah, I didn’t write this so clearly. My point is not that you shouldn’t do research on frivolous topics or write novels on frivolous topics, it was that you shouldn’t do it unless you have some personal or societal need to do so (or because you need to get paid, but that’s not an issue with Levitt at this point). For example, I’ve done research on baseball because it’s interesting to me. So I guess my mistake was in the term “urgent”—I should change it to something weaker, such as “interesting.” I’ll go into the above post and fix this to be more clear.
“I study bugs. This will not cure cancer. ”
Probably it will not, but you don’t actually know if your research will lead to a cancer cure or not. Surely the 19th century mathematicians who developed linear algebra did not anticipate that their work would become a foundation of quantum mechanics, which would lead to development of the transistor, which would revolutionize computing machinery, which would lead to vastly improved medical imaging which has saved many lives from many diseases.
Yes, it is irritating to have to pretend that your work will make some immediate contribution to the great humanitarian project. Unfortunately, the public does not widely understand how science evolves over time and how the solution to important human problems sometimes rests on something that starts out very obscure and superficially useless. It is precisely because of that that we should support the funding of research based on a pure quest for knowledge (“curiosity,” if you wish). So, in order to sustain public support for scientific research, government agencies need something they can “sell” to the public as in their short-term interest.
Yes, ultimately, the solution to this problem is to improve the public’s understanding of these issues, but that doesn’t seem likely to come about in the foreseeable future.
Although it’s true that one never knows what seemingly useless research will turn out to be useful someday, I don’t think that’s the main reason to study the life cycles of various kinds of beetles, or the formation of galaxies, or lots of other research. I think the main reason to do those things is because they are interesting to some people, and that’s enough. Art is not “useful” either — or at least, a lot of art is not useful — but the world would be a much less interesting place without art.
One of the best books I’ve read in the past five years is “A Sting In The Tale”, by David Goulson. The book is about bumblebees. As it happens, some bumblebee research has been very important to people, since tomatoes can’t be pollinated by honeybees but bumblebees do it and there’s a lucrative trade in bumblebees for pollinating greenhouse tomatoes, which has prompted research into how to treat bumblebee mites and how to keep the bees alive etc. etc., but even without the book chapter about that I would still have loved the book, because I was fascinated by the bees (how they live, how they behave), and sometimes by the methods the researchers used to study them. I wouldn’t have liked the book any less even if I somehow knew that there would never be a practical application of anything in it.
How much is worth spending to study any particular thing is a question, of course — should the Superconducting Supercollider have been cancelled, is the Webb Telescope worth the money, etc. — and potential practical applications can tilt the scales a long way if they are legitimate. But potential practical applications should not be the only motivation for research.
A former colleague where I used to work, Nelson Pomeroy, has also done a lot of research into bumblebees and their role in crop pollination.
https://www.researchgate.net/profile/Nelson-Pomeroy
One of my colleagues was worried about perjury (not technically, as we weren’t in a court) at one point during a DARPA application we were putting in. The question was whether we could get in trouble by claiming that we would hit 95% on some NLP benchmark that we knew was impossible to hit. At the same time, we knew the DARPA program manager would just cook the eval in the 11th hour so the project hit 95% because their job was on the line, too. And that’s just what happened—there were meetings about how to structure the eval to get the results we needed. It drove the people who’d optimized for the previous criteria up the wall as DARPA runs its evaluations competitively, not cooperatively among the grantees.
I find this as depressing as the Freakonomics issues.
I have no opinion about Steven Levitt, but I think if you’re going to accuse him of “pushing climate change denial” I’d like to see a fairer treatment than your repeated reporting that he had this view in 2009. (Not that is was justified in 2009, but people are allowed to learn and grow.) From a minute of searching, 2021: “I’m more favorable towards climate science than I used to be.”
Raghu:
I’m not actually claiming that Levitt or his collaborator Dubner ever had that climate-change denial view. I’m just saying he promoted that view, quite likely without ever thinking about it, just cos he thought it was cool to be contrarian. It’s good for people to learn and grow, not so good that he has not to my knowledge ever confronted his past failures. That 2009 article was not just stupid in retrospect; it was clearly stupid at the time, and perhaps Levitt’s lack of interest in examining his past failure in that area gives a clue to how he continues to be promoting flashy junk science to this day.
My standards are clearly much lower than some. I find “Freakonomics Radio” refreshing in that, unlike other programs on public radio, it is devoid of incessant “like”s, “you know”s and giggles. On occasion, I have even indicated publicly that a particular episode was worth listening too. Then again, I even enjoy TED Talks.
Climate change is out of Levitt’s expertise, but the abortion-crime link is his main academic claim to fame. If he is right, then there are some public policy implications. Some people say that he is wrong. If he wrong about that, then his academic stature is seriously diminished.
Why would academic stature be diminished if he is wrong about something? People seem to think that scientists have to be right to be respected. That’s the wrong way to think about theorizing in science.
Levitt has bet his reputation on the result. It is what he is mainly known for. It is a big part of how he got a high academic reputation.
Roger:
I don’t think Levitt has bet his academic or his popular reputation on any single claim. His scholarly reputation comes from having written several noteworthy papers of that sort as a young researcher. Not everyone things these papers are so great–I don’t really care so much, actually, as I’m more interested in his popular reputation–and I think there are reasonable cases to be made to give him a high or a low scholarly reputation, but in any case it’s not just one paper. It’s possible to say he did that paper wrong without that destroying the rest of his rep.
This post reminds me of the James Patterson model. Maybe Levitt has a staff that do the “research” and give him materials. If money is the ultimate goal, Patterson proves that having a famous name on the book cover sells books, and they will sell even if he didn’t write a word of it.
Kaiser:
I don’t think it’s quite like that. My impression is that Levitt as a young researcher did a bunch of applied micro research by himself and his colleagues, then at some point he encountered magazine reporter Dubner who wrote an article about Levitt and then a book with him. Through some combination of foresight and luck, Dubner transitioned from being a writer to being a new-media entrepreneur, and Levitt followed a path into public speaking followed by many successful pop-academics. The Freakonomics brand was valuable but Levitt did not have a stream of new research coming out, so Levitt and Dubner filled the content maw with research from Levitt’s friends and from various things Levitt and Dubner heard about on social media. Unsurprisingly, quality diminished, but it wasn’t on brand for them to criticize work that they were featuring. They had no problem criticizing clueless academics in general terms, but the Freakonomics brand was all about them talking about cool research. Beyond all this, Levitt appears to have a contrarian streak (“rogue economist” and all that), and Dubner is a political centrist (which, compared to most of academia and the news media, reads as conservative, as in his promotion of climate change denial, gender essentialism, and various random culture-war things like the claim that drunk driving isn’t so bad and that swimming pools are more dangerous than guns). Maybe it makes more sense to think of “Freakonomics” not as a Levitt product or even a Levitt/Dubner product but rather as a media channel, a sort of CNN of science promotion that gets filled with all sorts of material of uneven quality.
Your description makes sense. What’s troubling is that we can come to this blog and say it’s “uneven quality” but out there, this stuff is considered high quality by virtue of the Freakonomics brand! It’s what paul said above, it’s like TED talks. It’s getting that stamp of approval.
Kaiser:
Or anything with the “Harvard” (or “Stanford,” or “Columbia,” etc.) stamp of approval.
Dubner at this point sees it as a small media company. He has 5 podcasts running at a any given time but leaves the flagship name for himself. I like Dubner a lot because he is a good journalist and isn’t tied to economics. He did excellent series on the growth of Dallas, the changing role of whales, and Richard Feynman. He does indulge himself like a recent episode about an off-Broadway play.
The Leavitt podcast, called People I Mostly Admire, seems like hagiography for Leavitt’s academic buddies. I stopped listening. He did do an interesting interview with the record producer Rick Rubin.
Hi Andrew,
It seems like you have a beef with Levitt.
Yes, I know the Freakonomics books oversimplifies many things. They leave out a lot of nuances when applying the concepts of economics to society. They take a Malcolm Gladwell-esque approach to popularize economics, behavior science, and psychology. They cherry pick stories to drive home their points.
BUT Levitt did took a risk here in venturing to write Freakonomics. And I think that’s commendable given the risk averse nature of ivory tower academics.
If you and so several of the commentators here are allergic to the Freakonomics books and podcasts, why not write a book for the popular audience but with rigor and “academically correct” style. Show the world it can be done.
Even better, self-publish the book yourself so you don’t have to be perceived as having to chase book sales if it was picked up and promoted by a major publisher.
For those who don’t listen to Levitt’s show, he is ditching the podcast to focus his efforts on developing his new school with Arizona State University. If done well and with luck it can be more impactful than his research, books, and podcast efforts.
Rj:
My beef with the Freakonomics team is not that they “leave out nuances” or “cherry pick stories” or that they are insufficiently “risk averse” or that they don’t have “academically correct” style.
My beef is that they uncritically promote junk science–for example, the ridiculous claim that beautiful parents were 36% more likely to have girls, the ridiculous claim that we were virtually assuring us of about 30 years of global cooling, and ridiculous claims of mind-body healing–and then, when people call them on it, they just move on. They’re littering the discourse with stupid, and, yes, I think that’s a problem.
I also think the Freaknomics team have done lots of good things. I like a lot of the original Freaknomics book, and I’ve enjoyed much of their writing since. Indeed, one reason I keep coming back to the problems of Freakonomics is that I know that they can do better, so I’m sad and frustrated when they make avoidable mistakes.
See the P.S. of my post above. You write of Levitt’s success selling many books etc. I’d like to think that he could do this without promoting junk science.
Then write that book that can be a counterpoint to the Freakonomics books.
The best way to counter books “promoting junk science” is with books that are
rigorous, fact based, not sensationalist. There are too many pop – psych / econ/ statistics books doing that former.
P.S. This blog should be called “The Mean Statistician.”
Rj:
1. I have no idea if a book with all of Freakonomics’s positive qualities and none of its negative qualities could be written. Or, if it could be written, that it could be written by me. Or, if it could be written, that it would be such a success. I suspect that part of the success of the Freakonomics franchise comes from its authors’ willingness to promote exciting ideas, and that includes some bad ideas, unfortunately.
2. You can write your own blog, then you can call it what you want. You could call it The Nice Statistician and you could fill it with positive articles about the awesome scientist who proved that each citation is worth $100,000, and the awesome scientist who discovered that beautiful parents are 36% more likely to have girls, and the courageous scientist who discovered that we were in for 30 years of global cooling, and the amazing scientist who discovered that people can heal better just by thinking that time was passing faster, and the superstar scientist who discovered that signing a form at the top is better than signing at the bottom . . . all sorts of things. Do it right and I’m pretty sure you could get many more readers than I do–I’m not kidding about that!
Andrew,
I think I will write that book someday … and you can skewer my work in this blog afterwards!