The problem is that while on average there are significant differences between men and women, the individual variation within men and within women is so much larger than these sex-based differences that sex is not a good predictor for behavior in the majority of the cases.
Well, I'm not so sure how constructive it is to try making behavioral predictions about people based on general categories (race, sex, etc.) in the first place. I think it's far more useful to seek understanding and respect.