> The 485 didn't sound very controlled ("asked to eat larger meals at a certain time") This is typically how it's done though: if you were to make a subselecti

> The 485 didn't sound very controlled ("asked to eat larger meals at a certain time") This is typically how it's done though: if you were to make a subselection of the treatment group to only include those that adhered to the intervention then (a) you risk undoing the randomization, because people who are more likely to adhere to the protocol are likely different in other ways as well and (b) the effect size you'd report would be a best case scenario, not realistically the kinds of results a dietician would see when helping people.