[quinnchr]

Except we do know, we have a pretty darn good idea. We know how many people self-report mental illness. We know how many people actually seek treatment for mental illness. We know the self-reported rate of mental illness amongst other populations and can compare the two.

It's literally like any other disease. We don't know how many people actually get the flu every year. We only know how many people self-report getting the flu, and how many people seek treatment. Self-reported surveys are used all the time in epidemiology and we manage to form effective public policy based on that.

So sure, there's definite bias when it comes to self-reporting, but your suggestion is to ignore what data we do have and form public policy based on what? How you feel? First principles?