4 ms·

A lot of nutritional studies have a common flaw: 1. Various health outcomes and indicators are caused by a food or combination of foods that you regularly eat.

A lot of nutritional studies have a common flaw: 1. Various health outcomes and indicators are caused by a food or combination of foods that you regularly eat. 2. A diet is studied where these foods get removed from your diet, to be replaced with whatever the diet under study is. 3. The participants in the diet study have poorer than average health outcomes and indicators (otherwise, why would they try changing their diet?) 4. Regression to the mean happens, and all sorts of dietary changes appear to be good for your health. Basically, if you're fatter than average, and it's caused by the foods you eat, you can fix that by making a list of everything you eat and not eating those foods. "Whatever you don't eat" is a more average diet than what you're currently eating, so you'll get more average results. (This idea shamelessly stolen from HN user jimrandomh. Hopefully I didn't butcher the details too badly.)

coldtea 10y ago
>3. The participants in the diet study have poorer than average health outcomes and indicators (otherwise, why would they try changing their diet?) Because they're asked to, as part of the experiment? >4. Regression to the mean happens, and all sorts of dietary changes appear to be good for your health Why would the mean be better, since in (3) you mentioned that those people had "poorer than average health outcomes and indicators" to begin with? For regression to the mean, they should have started the diet after some extreme readings. But who said those are what prompted them to the diet? People don't start on a diet when they get extremely heavy, they start and stop diets all the time, while being continuously obese. >Basically, if you're fatter than average, and it's caused by the foods you eat, you can fix that by making a list of everything you eat and not eating those foods. "Whatever you don't eat" is a more average diet than what you're currently eating, so you'll get more average results. Well, if you eat the same calories from your new "Whatever you don't eat", then you'll get as fat.
- ThrustVectoring 10y ago
  >Because they're asked to, as part of the experiment? Let me try another way to word the same concept: why would an actor in the best shape of their life with a personal chef and nutritionist participate in the experiment? Similar logic goes for less extreme examples on this end of the curve. >For regression to the mean, they should have started the diet after some extreme readings. They're obese, their diet is an extreme reading from the space of possible diets. They're choosing foods that only satisfies their various needs at a caloric quantity that causes weight gain. >But who said those are what prompted them to the diet? People don't start on a diet when they get extremely heavy, they start and stop diets all the time, while being continuously obese. Caloric restriction diets and food choice diets aren't the same thing here. People usually start and stop diets all the time because caloric restriction diets have huge user experience issues. Specifically, the very common yoyo dieting; you feel bad about being fat so you decide to do something about it, then you do caloric restriction in a way that fails to meet many of your other needs, then you abandon the diet because the stress of not meeting your needs starts to outweigh the negative feelings from being fat. >Well, if you eat the same calories from your new "Whatever you don't eat", then you'll get as fat. If your current diet meets your needs at an above-average number of calories, you won't eat the same number of calories (unless you try to manually control caloric intake through some variant of calorie counting).

coldtea 10y ago

>3. The participants in the diet study have poorer than average health outcomes and indicators (otherwise, why would they try changing their diet?) Because they're asked to, as part of the experiment? >4. Regression to the mean happens, and all sorts of dietary changes appear to be good for your health Why would the mean be better, since in (3) you mentioned that those people had "poorer than average health outcomes and indicators" to begin with? For regression to the mean, they should have started the diet after some extreme readings. But who said those are what prompted them to the diet? People don't start on a diet when they get extremely heavy, they start and stop diets all the time, while being continuously obese. >Basically, if you're fatter than average, and it's caused by the foods you eat, you can fix that by making a list of everything you eat and not eating those foods. "Whatever you don't eat" is a more average diet than what you're currently eating, so you'll get more average results. Well, if you eat the same calories from your new "Whatever you don't eat", then you'll get as fat.

ThrustVectoring 10y ago
>Because they're asked to, as part of the experiment? Let me try another way to word the same concept: why would an actor in the best shape of their life with a personal chef and nutritionist participate in the experiment? Similar logic goes for less extreme examples on this end of the curve. >For regression to the mean, they should have started the diet after some extreme readings. They're obese, their diet is an extreme reading from the space of possible diets. They're choosing foods that only satisfies their various needs at a caloric quantity that causes weight gain. >But who said those are what prompted them to the diet? People don't start on a diet when they get extremely heavy, they start and stop diets all the time, while being continuously obese. Caloric restriction diets and food choice diets aren't the same thing here. People usually start and stop diets all the time because caloric restriction diets have huge user experience issues. Specifically, the very common yoyo dieting; you feel bad about being fat so you decide to do something about it, then you do caloric restriction in a way that fails to meet many of your other needs, then you abandon the diet because the stress of not meeting your needs starts to outweigh the negative feelings from being fat. >Well, if you eat the same calories from your new "Whatever you don't eat", then you'll get as fat. If your current diet meets your needs at an above-average number of calories, you won't eat the same number of calories (unless you try to manually control caloric intake through some variant of calorie counting).