> Maybe 1 in 20 studies don't have obvious confounding factors that cause us to have no real base facts, but we can find a lot of correlation.
Even worse: Many studies are done on mice, who do not eat the same things as humans
And worse still: Many are done with "food surveys" which lean too much on the poor memory and faulty assumptions of participations (eg what some Americans think of as a serving of "meat" is really a serving of fried breaded bread with condiment levels of meat mixed in. Most Americans do not know that chicken nuggets are half corn. etc )
The people with the worst data may be the scientists, unfortunately. You can draw somewhat accurate conclusions from your own personal testing and anecdotes, but you might draw mightily wrong conclusions from survey data which appears to say something that simply isn't true.
Even worse: Many studies are done on mice, who do not eat the same things as humans
And worse still: Many are done with "food surveys" which lean too much on the poor memory and faulty assumptions of participations (eg what some Americans think of as a serving of "meat" is really a serving of fried breaded bread with condiment levels of meat mixed in. Most Americans do not know that chicken nuggets are half corn. etc )
The people with the worst data may be the scientists, unfortunately. You can draw somewhat accurate conclusions from your own personal testing and anecdotes, but you might draw mightily wrong conclusions from survey data which appears to say something that simply isn't true.