> Yes, there is a difference on average, but the overlap is way bigger than the difference

Unless you have evidence to back this up, it sounds an awful lot like a "mere belief"