Of Baseball Pitches and Fad Diets: When science gets it wrong

An article by John Kay in The Financial Times (and thank you, TheBrowser.com, for highlighting it) recalls economics icon Milton Friedman’s and a colleague’s observation that we can often get very good at something (Friedman used billiards as an example) without knowing much about the complex principles behind it. I myself always appreciated the fact that physicists used to argue that a true curveball in baseball was an impossibility and that catching a fly ball required solving trigonometric equations, while baseball players (and in the latter case even dogs) effortlessly accomplish these feats without much regard for their impossibility or mathematical demands. And it will be a long time before our best scientists and engineers can design a fighter plane whose multi-computerized flight systems can begin to approach what a fly can do evasive-flight-wise with a teeny speck of a brain–a mere 100,000 neurons, or one-millionth as many as an aeronautical engineer. The prodigious pattern-recognition capabilities and other elegant programming of our brains endow us (and other creatures) with the ability to intuit or otherwise toss off solutions that elude formal analysis.

This notion that analytically daunting phenomena are often simply accomplished by other means brings to my mind some of the interchanges I’ve had in recent months with metabolic experts in both the ultra-low-carb and ultra-low-fat weight-loss camps. They’ve got the science theory, they’ve got the rodent studies, they’ve got the diet studies, and voilà: the only way to lose weight is cut out carbs or cut out fat, depending on which camp you’re listening to. The two camps come to nearly opposite conclusions, and both are absolutely unwavering and uncompromising in their belief that the science completely and clearly proves they’re right. And of course there are many other scientifically sure-of-themselves diet-gimmick camps.

Back in college my math and physics professors occasionally inflicted on me and my classmates a certain type of killer problem–well, they were killers to me, anyway–in which we were presented with a step-by-step proof that started with perfectly valid assumptions, and ended up with an indisputably mistaken conclusion. We had to figure out where the proof had gone wrong. I usually had a great deal of trouble doing so–the errors were very subtle, and sometimes were semi-errors spread out among several steps. These problems left me with a deep and permanent appreciation of how right something could look in every aspect of every step, and yet end up flat-out wrong.

Though scientists and other highly trained experts should know better–for one thing, I would think most of them got that same kind of what’s-wrong-with-this-picture? problem in college–these folks often seem oblivious to the way that very right-sounding arguments can leave them impaled on silly conclusions. Of course, one of the main points of science is supposed to be to test theories to avoid this problem, but, as I’ve discussed endlessly elsewhere to the point where I’m tired of hearing myself repeat it (though fortunately others have been smartly chiming in on this subject lately) these tests tend to be terrible, and typically end up merely reflecting researchers’ biases rather than getting at any real truths. In the end, the scientists who do the best work tend to be not the ones who avoid bias–that’s just about impossible–but rather the ones who for whatever reasons end up with the right biases. When an expert’s instincts run off the rails, you get, well, for example, claims that fat or that carbs are entirely responsible for obesity. And these experts become so convinced that the science and data back them up that their beliefs become unreasonably unshakable

This willingness of intelligent, knowledgeable people to let solid-seeming scientific reasoning make them absolutely certain about what is obviously not so is remarkable. There’s so much easily observable evidence in the world all around us that people are able to lose weight and be healthy while still eating carbs, or still eating fat, that it just feels silly to have to make a point of arguing it–if someone doesn’t see it at this point, they’ve just decided not to see it. Tens of millions of people have tried the Atkins-style ultra-low-carb diet–it was initially a fad diet in the 1970s, and was reignited as a much bigger fad in 2002 after a New York Times Magazine cover story outlined the science that seems to back it up, an article written by the brilliant science journalist (and my friend) Gary Taubes. Meanwhile, tens of millions more have tried very-low-fat diets. How has that worked out for America? We all know the sad answer to that question. Out of the dozen or so people I know personally who went on the Atkins diet, not a single, solitary one kept the weight off for more than two years, and most lasted on the order of a few months. (Some years back, after getting a very convincing science lecture one-on-one from Taubes, I tried it, and lasted three weeks–I felt like crap the whole time, though I don’t seem to be typical.) I know a few people who lost weight on very-low-fat diets and actually kept it off, but they didn’t really stick with the diet–they added back modest amounts of fats back in, and they also happen to be very physically active people.

Needless to say, what happens to my personal acquaintances hardly constitutes convincing evidence. But if you look at the range of diet studies out there and talk to a lot of experts in the field and others about what has and hasn’t worked–something I’ve been doing quite a bit of for the past year–it becomes utterly clear that neither of these extreme diets achieves particularly good long-term adherence levels, and without long-term adherence a diet is at best a waste of time. To be sure, some people do permanently lose weight on one or the other approach, and in the Internet era a small minority of fad-diet beneficiaries can band together on websites and blogs and sound like a movement. More power to you if you’re one of these beneficiaries. But clearly for the vast majority of people cutting out carbs or cutting out fat is simply not going be a lifelong habit. It doesn’t matter whether cutting carbs or cutting fat will actually help you lose weight and be healthy–and it certainly doesn’t matter if the science appears to back the approach up–if you can’t stick with it. What we need, of course, is eating and exercise habits that we can stick with, and ways of being helped to adopt and stick with healthy eating and exercise habits. Cutting back at least a bit on carbs and/or fat is usually part of it, but it’s not typically the main part, and it’s rarely the entire shooting match. Having to drastically cut back on a major food group usually makes an eating plan nearly impossible to stick with long-term for most people. (But for those of you who insist on doing so, you might want to cover your bases with the Dukan Diet, a big fad diet in France that is now making its way to the US, and which advocates cutting out both fat and carbs, leaving you eating mostly protein. Good luck! In its defense, I’ll point out that it does also advocate behavior-oriented measures. On the other hand, many medical experts caution that eating very-high-protein can be dangerous over the long term.)

I’ve tried talking to various advocates of extreme diets about these seemingly obvious problems with their claims and it’s like talking to someone about deeply held religious beliefs. (Actually, some deeply religious people I’ve spoken with are much more open-minded about their beliefs than are some people who cling to dubious scientific beliefs.) Taubes is one of the smartest people I know, and he has surely become one of the most knowledgeable people on the planet with regard to how the human metabolism deals with sugar, but I just have to scratch my head over his (and his hard-core supporters’) response to the observation that working on behavior change seems to be a critical factor when it comes to permanent weight loss, the response essentially being, “All eating-related behavior is driven by carbs and their effect on metabolism, exercise doesn’t matter, the science proves all this, and that’s that.” Low-fat proponents, at least the ones I’ve interviewed, tend for whatever reasons to be much more aware of the role of behavior change than the somewhat fanatical very-low-carb crowd, but can be as tenacious in their patently implausible claim that a single type of food is largely responsible for obesity. Both low-fat and low-carb fanatics are of course very good at pointing out the holes in the scientific reasoning and evidence of the other camp, but blind to the holes in their own claims.

One of the things I really like about experts who focus on behavior change, and especially applied behavior analysis researchers–the folks who study ways of applying what is essentially B.F. Skinner’s science of behaviorism–is that most of them rarely seem to get hung up on theories. They just try to figure out through observation of people what best works in real life–either for an individual, or for the greatest percentage of people in a group–and then that’s what they push, while they continue to look for something that works even better, or on a broader range of people. It’s very hard to picture a behaviorist insisting that a pitcher can’t possibly make a baseball curve, or that eliminating carbs is a necessary and sufficient condition for shedding weight.

8 thoughts on “Of Baseball Pitches and Fad Diets: When science gets it wrong

  1. Anonymous says:

    As someone who studies neuroendocrinology, I find Seth Robert's ideas not farfetched. "Fad" is often a word for the cynics who don't like to think.

  2. I don't find Roberts' ideas all that far-fetched either, in terms of the explanations of why it should work. I don't find low-carb diets, or low-fat diets, or the Zone diet, or many other somewhat gimmicky, faddish diets far-fetched based on their explanations. They're all backed by some pretty impressive-sounding theory and studies. But there are two bottom-line problems with all of them: 1) They generally contradict one another, and all these supersmart, highly trained scientists and other experts end up claiming they're right and all the others have it completely wrong. It's possible that one of them (or one set of them) has it right and all the others are wrong, but I happen to think it's much more likely they all basically have it somewhat wrong. (To put it another way, they kind of cancel each other out.) 2) When you turn a bunch of overweight people loose on any one of these diets in the real world (meaning not in formal studies, which are highly subject to researcher and subject bias and many other distortions), most people just don't end up sticking with them, for whatever reasons. As far as I'm concerned, that means the diets don't work, whether or not people who stick with them end up losing weight and being healthier. I see no convincing evidence that Roberts' oil-swilling diet does any better than the others in this regard. As with all the diets, there are proponents who swear it has changed their lives, but I believe that this is a real minority of people who try it, and we can't know if even most of these proponents really end up sticking with it for many years. As for my use of the word "fad," I probably am being a little careless with it. For one thing, it doesn't have a precise meaning. But I don't think I'm being especially cynical, or that I'm not thinking carefully about this stuff. I get accused of it all the time, though–from people who, apparently like you, think I'm right to criticize all these diets but I'm an unthinking cynic to criticize the particular one they happen to believe in. I agree with the old saw that being fair means that in the end you've equally pissed off all conflicting parties, so maybe I'm on the right track after all. Having said all this, I should point out that, as with all the various diets, there are particular aspects of Roberts' diet that I find especially appealing and plausible, and that make it easy to understand why the diet seems to work for some people. But for me, that's beside the point. What I want to know is: How do you help a large percentage of overweight people to do what it takes to become healthier over the long term? Sorry, but I'm pretty sure chugging oil ain't it.

  3. Anonymous says:

    Thanks for the response. I very much agree with most of your points. I suppose my bias comes from my training and interest in the neuroendo aspects of weight gain, which makes Seth's ideas appear to me particularly attractive and plausible.I started this interest in weight management through Gary Taubes. I became very convinced of the low-carb ideology. Since then I've become more and more agnostic about specific ideologies, and more interested in what all this says about our success/failure in trying to model and understand complex systems.

  4. It's easy to see everyone else's biases, we tend to be pretty blind to and forgiving of our own. (I certainly don't exclude myself.) My view of obesity, which focuses on behaviorism, sounds compatible with yours. A key point in behaviorism (or some versions of it) is that our internal mental processes (and in the case of obesity our metabolic processes) are both complex and nonobservable directly in any precise way, which makes them poor targets for intervention. We can in theory be much more verifiably effective in identifying environmental and behavioral factors related to obesity, and then focusing on altering those. One of the things I like about Roberts' approach is that it is to some extent based on his observations of his environment and behavior. I just think he happened to get locked into one particular effect that's kind of interesting but doesn't really come close to helping bring on the sort of permanent, broad environmental and behavioral changes needed to tackle obesity. The theory he's tacked onto it is interesting, too, but everyone's theories sound interesting. As for Taubes, he did a phenomenal job in showing why cutting out fat isn't the answer everyone claimed it was, and why carbs can be a big problem. I think he goes way, way too far when he claims cutting out carbs is the whole solution and the only possible solution. In fact, I think he's basically succumbing to some of the same sorts of biases that the cut-out-fat people did.

  5. David Brown says:

    "Both low-fat and low-carb fanatics are of course very good at pointing out the holes in the scientific reasoning and evidence of the other camp, but blind to the holes in their own claims."Exactly so. My take on the obesity epidemic is that the focus should be on healing the metabolism, not on losing weight. In this regard, I think Diana Schwarzbein comes as close as anyone to addressing the central issues.Some have said the obesity "causes" disability and early death. This is like saying that smoke generated during a forest fire sickens and kills trees. Well, that might possibly be true for a tiny percentage of forest flora. However, most trees are injured or killed by the flames. Likewise, it is chronic inflammation that damages tissue, sickens, and eventually kills humans. Deal with that problem and many people would revert to normal weight and those who remain large would be healthy.What I'd like to see is a lot more publicity about the connection between omega-6 seed oil intake and the chronic inflammatory diseases.(1) A little more than a year ago I realized I was slowly doing myself in by consuming a peanut butter sandwich almost daily for lunch. Two months after I quit eating peanut butter, the pain in my legs subsided. I've since been researching the omega-6 hazard and I'd say the evidence suggests that the introduction of omega-6 seed oils into the food supply about 100 years ago constitutes the major public health disaster of the 20th Century and beyond. To be sure, excessive sugar consumption plays a major role in obesity and chronic disease. And really, it's difficult to say which is worse. However, it's clear that scientists need to pay more attention to omega-6 (2) and stop demonizing saturated fat. (3)Google these phrases:1. Omega-3 Omega-6 imbalance2. 1 of 4 Bill Lands on Cardiovascular Disease3. Controversial saturated fat

  6. Iain says:

    I am 52, 5'11", weigh 175, 9% body fat, provincially competitive bicycle racer, have been weight training and running and/or cycling since I was 14. I just had my annual physical, BP was 100/66, all the blood work came back fine. And I have been eating 2 peanut butter and jam samdwiches for lunch 5 days out of 7 since I was about 5. That continues to this day. That is a lot of Omega 6, and I certainly don't feel like I am killing myself. Current evidence suggests quite the opposite. IMHO, lifestyle and genetics matter more than anything else. OTOH, bombarded as we are by endless, contradictory advice on diet and health, to the point where everything is bad for you, I have come up with my own saying: "Living is bad for your health" 🙂

  7. Conflicting theories are abundant among scientists (not to mention the public) with regard to which nutrients and substances may be helpful or harmful to health in particular quantities. Sadly, it has proven nearly impossible to assemble consistent, clear, reliable evidence to back up any particular theory about almost any of these nutrients and substances. I discuss this problem a bit in my Atlantic article about Dr. John Ioannidis, and will be writing more about it in the future. Part of the problem may be that we are all different in many ways, so there really may be no consistent answer to chase after for many substances. What does seem fairly (but not completely) clear is that being at a healthy weight and regularly exercising (even if it's just walking) have real potential benefits for most people, and doing whatever it takes to establish habits that over time keep you from carrying too much fat, and keep you moving, is worth going after. So yes, lifestyle is a big deal–but the key question is what it takes to make permanent changes in lifestyle in order to adopt a healthier one. I don't think that a few specific foods and substances often make a very big dent in that situation one way or the other, though of course there may be exceptions for some people.

  8. Keith Curley says:

    Thank you for injecting some sanity into this discussion in your last two posts (the only ones I've read so far.) And thank you for writing Wrong, which I'm almost done with as well. I was very impressed with both of Taubes' books, and I think he's got it half-right: namely his criticism of the established nutritional wisdom and the studies that were used to support it. Unfortunately, his own prescriptions as you say seem to highly biased.His pooh-poohing of exercise is just half-baked. But more abstractly, it seems highly plausible that certain diet compositions might be more satiating than others–I'd put my money with yours that the sugar and insulin spikes from high glycemic foods might be a real problem here. What the dietary composition should be…probably depends a lot on a person's individual characteristics and also behaviors as you rightfully point out. But granted the above, which seems common sense, wouldn't one also expect certain exercises and or activities might lead to a greater calorie expenditure than increase in apetite? Why doesn't Taubes see that?His whole neglect of behavioral issues seems really blind now that you point it out. (It's anecdotal I know but I think the fact that I exchanged my heavy-drinking habit for a habit of going to the gym every night had something to do with how I lost 70 pounds last year.)

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: