The trouble with these kinds of studies (and with empirical evidence in general) is that there are often omitted correlating factors left out of the analysis. D

The trouble with these kinds of studies (and with empirical evidence in general) is that there are often omitted correlating factors left out of the analysis. Doctors would be sure to control diet and other medications, economists would control socioeconomic status, sociologists would look for cultural factors, etc. However, they're not going to do two or three of these sets at once so the result is a biased study no matter who performs it. In this case I think a good potential omitted factor is the level of parental discipline. Suppose parents A are Dr. Spock laissez-faire with their kids. This is most likely correlated with: Not sitting still for the family meal Watching a lot of television Getting their way in public places So Junior gets lots of TV dinners, quick snacks, and time with the power rangers so he wants to do karate instead of read Fox in Socks Parents B run a tight ship and consequently their vegan, free-range fed wonder-child is happy to sit on the magic carpet and hear about anything so long as it's not The Wealth of Nations again. Applying the overused correlation != causation mantra is a bit cheap, to be fair we must follow the immediate next step, "If correlation != causation, why?" I think here we have plenty of compelling falsifiable narratives to try before we make green key-lime pie illegal.