You'd be forgiven for thinking this is a cheap shot, but nope— Wikipedia lists the difference in mean height for Americans as about 5" (~5'10" for men, ~5'5" for women). If you get three guys together in a room, you've probably already got more variation than that.
The general point here is that there is much, much more overlap between the sexes than there is difference between them. That's why although you certainly can make interesting statistical inferences about populations, applying the same statistics on an individual level (this person is a woman, she's probably [whatever]) is very likely to be false and is equivalent to blind prejudice.
Look at page 11. 19 year old women in the 90th percentile are still shorter than 19 year old men in the 50th percentile. Now, while I was never very good at statistics, I'm pretty sure if we had a running bet over who was taller in any random male/female pairing, I would drain your bank account dry.
The general point here is that there is much, much more overlap between the sexes than there is difference between them. That's why although you certainly can make interesting statistical inferences about populations, applying the same statistics on an individual level (this person is a woman, she's probably [whatever]) is very likely to be false and is equivalent to blind prejudice.