That seems like a separate topic. The question at issue is whether people who can be clearly classified as "men" or "women" according to general consensus show some cognitive differences on average. I have not seen anyone deny this.

In a similar way, one might argue that "races" don't really have any objective existence, and yet still acknowledge that e.g. African-Americans are less likely to go to college than Asian Americans. Such stats aren't invalidated by the lack of any objective biological definition of "race", or by the existence of some people with mixed African/Asian heritage, etc. etc.