And they did use cognitive tests that were correlated with job performance. If the people did not perform differently on those tests, or performed better on those tests… well there you go. Is it the most rigorous study in the world? Obviously not. Does it indicate what it purports to indicate? It sure does.

That sort of method has proven problematic, as the thing that matters could be something correlated with breakfast, and the breakfast itself could be irrelevant.

You usually have to ask people to change their behavior. Pretty straight forward in this case though.