It that's true, it would be trivial to design a survey to get whatever sexual assault statistics that you want. Even after the survey has been taken, by having

It that's true, it would be trivial to design a survey to get whatever sexual assault statistics that you want. Even after the survey has been taken, by having questions ambiguous, and then reclassify them as either sexual assault or not.