Musings on this article: http://fivethirtyeight.com/features/you-cant-trust-what-you-read-about-nutrition/
I work for the Center for Human Nutrition at Washington University Medical School. I have at various points crunched food diary data (when the article says food diaries generate a lot of data, they aren’t kidding), filled out food diaries of my own, and processed samples of folks who participate in our diet and exercise studies to see what happens to their metabolic processes when they eat (or don’t eat) certain foods. Sometimes we control exactly what they eat; sometimes they report it themselves.
Misrepresentation in reporting food is easy to do, and most people don’t do it on purpose. How do you know exactly what’s in that yummy fancy restaurant food? Is it summer when your favorite foods are in season so you eat them all the time? Do you not want to look bad on paper so you don’t eat that second slice of pie when you normally would (or eat it and not report it)?
Basically this means that you can gather up whole reams of data and then compare it with whatever you want to, and come up with some pretty silly conclusions, like the picture that accompanies this entry: if you eat cabbage, you probably have an innie belly button. What? While it’s possible that cabbage has some effect on the position of your belly button, this is probably not true. So don’t trust what the media says is the next new food thing guaranteed to make you lose weight and avoid cancer forever, because it may be based on faulty interpretation of data. Do your own research.