As strange as it sounds, if you want to lose weight, it is worthwhile to learn something of the philosophy and practice of science. And more diet and health books are opening with a short discourse on the topic, as they try to explain how so many well-intentioned researchers can be violently at odds on what the best diet is.
The best known and best written book of this caliber is Good Calories, Bad Calories by Gary Taubes, and if you are new to science, research and the philosophies and practices of science, I suggest that you don’t give up and read the first section twice. If you have no interest in diet, but are interested in science, I also recommend this book.
In it, Taubes goes backwards in time to prevalent beliefs, diets, and studies, and then comes back to the present, adding new believes, studies, recommendations and controversies. As each new element is added, he studies the conflicting studies, and notes how they are never resolved, disappearing slowly into the background, with studies of that time ultimately becoming “unobserved”- published but not read or referenced, and, ultimately, unfunded.
Also impacted is how studies are constructed on what they focused on. Certain seminal studies used to drive national nutritional recommendations turned out to be extremely poorly designed. A common flaw is not have a comparable control due to calorie restriction on one diet, but not the other. Another was to measure heart disease, but not overall mortality. A lot of studies raised the spectre of reduced heart disease but increased cancer. That was blithely ignored as “likely due to other unknown factors”.
Good science tries to actively falsify it’s own theories. Scientists who move past all the warning markers because of some positive markers move deeper and deeper into the quagmire, where all results are suspect to do compounding assumptions, confounding factors, and ongoing confirmation bias.
Some great insights in Good Calories, Bad Calories:
1. the heart disease epidemic post WW-II didn’t actually happen. It was poorly interpreted data.
2. the definition of the “classic” American diet pre-WWII used to identify causes of the non-existent epidemic was largely wrong.
3. most recent drops in heart disease mortality is from better treatment.
4. high fat diets don’t cause weight gain due to caloric density (9 calories for fat due to 4 for proteins and carbs). This was a strong factor in “it can’t hurt to recommend low fat anyway, since fat causes weight gain” rationalizations. It turns out to be very wrong. Fat is strongly correlated to satiety and less calorie consumption.
5. much of the supposed benefit of a low-fat diet is at the population level. i.e. the net benefit to an individual’s personal risks is pretty low, with researches predicting 3 days to 3 months of additional life span IF the low-fat theory is in fact correct.
6. you need to consider that when you remove something from a diet, something is added. By removing fats and increasing carbs, what is the result? Diabetes? More heart disease?
So the epidemic that pushed people to quick action in making the low-fat diet the national standard never existed, and main supporting facts that helped to move it along were false. However, once these were accepted, other studies had been cherry picked to move the movement along, with highly disturbing other facts being ignored.
The most disturbing point for me was that even the tenuous correlation between blood cholesterol and heart disease was mostly in men. If fact, a strong reverse correlation between cholesterol and mortality was often observed in women. This, amazingly, seems to have been ignored by many researchers.
The sad truth that Taubes brings us to is that there is a desperate need to do a lot of good science that wasn’t done previously, and that some seriously bad science has been used to turn all of us into guinea pigs.
Another book that is a bit lighter on the science, but has a more pointed introduction to some of the impact of complexity and scientific process on diet, is Body By Science, By Doug McGruff and John Little. It has a great overview on genetics and considerations of that for “what works for 95% of the population”, and makes a great case for the benefits of weight training over low-intensity aerobic type exercises. The science can be overwhelming, but the end recommendations are very compelling.
And finally, there is Michael Pollan’s books. They are lighter on the science overall, but give a good survey to the sociology of science and the bad suggestions that can result.
Pollan also makes the suggest of considering what you add to your diet more then what you remove. If you remove fats, what are you adding? If it’s not vegetables, you may ruining your health with refined carbs.
Pollan’s books are, also, ironically a good example of the “nutritionism” issues he mentions. He does an admirable job of knocking it down in the beginning of “In Defence of Food” but lapses quit heavily into it himself in the second half. Whereas Taube’s takes excess carbs and refined foods and switches them back to fat and protein, Pollan moves them over the vegetables, even though he does identify many successful hight fat traditional diets. He also makes nutrient recommendations from the current state of science, which is exactly what all previous purveyors of “nutrionism” did. However, he does so by excluding processed foods and supplements.
We all want to share our beliefs and insights, and it becomes almost irresistible to hold back. But that is where the best scientists shine, actively trying to disprove their own theories, and not ignoring the warning markers that exist in their data. Good science, like good quality, is conservative and slow, because complexity is a wild beast that can strike in many hidden and insidious ways. The only way to combat complexity and our own nature is a strong process, with a strong, proven, philosophy and practice of science. If you are fed up with all the conflicting diet messages out there, consider these books are gain a little comfort from understanding why things are as crazy as they are.
