[Pulcinella]

We (and by we I mean the government) just need to bite the bullet and and pay for a multi billion dollar study where we pay tens of thousands of people to control* and monitor what they eat, exercise, etc for 20 years and just figure this out. None of this “well we had 20 grad students journal what they did for 6 months and this is the result” low quality studies.

*Morally correct and ethically of course. I’m not saying we force people into large hamster cages.

[nradov]

Unfortunately such a study still wouldn't produce valid results. The problem is that study subjects don't report accurate data. They lie because they're ashamed of their choices, or they forget, or they just don't know how to accurately count what they eat.

Consider a hypothetical study subject who goes to watch the game at his local bar, drinks 7 beers, eats a plate of hot wings, and then stops at Taco Bell for nachos on the way home. When he wakes up the next morning what do you suppose he'll enter on the form?

[lazyasciiart]

You want to look at the Framingham studies. Not what you're describing, but it's far far above the 20 grad student model https://www.ncbi.nlm.nih.gov/projects/gap/cgi-bin/study.cgi?...