Are You a Statistician?

If we could do randomized experiments for everything, then science would be better in the sense that randomization would take care of any confounds. We don't have the blinded trials that we'd like the RCTs with, you know, large groups of people to measure outcomes. In the case of nutrition in particular, these associations have not proven very strong and when they are able to be tested in even smaller ways and then they tend to do that to not hold up very well at all.