Except most of those studies are bullshit bad science done with fundamentally flawed methodology. It got a throwaway line in the article, but just about the only honest way to make scientific conclusions about nutrition would be to take a human population of statistically significant size and diversity, move them underground to a controlled environment, and control and monitor their food intake for some period of years. That would be a data set that might not be spurious (although I bet even within that, people would find ways to eat stuff the researchers can't record). That's not going to happen, so nutrition will continue to mostly be junk science that utilizes crappy methods and statistical manipulations to create sensationalist headlines and keep selling books.