“ Clear evidence that one group differs from another doesn't provide any useful information for dealing with individuals “

This isn’t necessarily true.

For example, I know that two-year-olds are far more sensitive to missing naps then eight-year-olds are.

This changes how I structure time and environment around a two year old so that he is the statistically more likely to behave. If I don’t do that, I set them up for individual behavioral failures.