Randomized trials are the gold standard of dietary studies, but as I said above, they're really, really hard to conduct properly. You have to find a stable population of people. You have to pick half of them randomly and get them to change their diets. You have to trust them to actually do it. You have to follow them for years, not months. Virtually no trial can ever truly meet this standard.