Bryan Caplan  

Endogenous Sexism Explained

An Internet secession so far... Happy "Almost Assassinate Hitl...
Several people in the comments got the point of my endogenous sexism scenario.  Namely: Friends pass a stricter selection filter than spouses of friends.  If you think poorly of someone, you won't be their friend.  But if you think poorly of the spouse of your friend, you'll probably put up with your friend's spouse to preserve your relationship with your friend.  As long as people tend to make more same-sex friends, then, men's male associates will seem better than their female associates, and female's female associates will seem better than their male associates. 

Two lessons:

1. When a man doesn't like his wife's friends, or a women doesn't like her husband's friends, it's not surprising.

2. By itself, #1 does not imply sexism.  But #1 combined with statistical naivete readily leads to sexism.

How strong should we expect this effect to be in the real world?  Hard to say, but friendship is strongly segregated by gender.  The General Social Survey, for example, asks about the gender of your best friend.  Same-sex besties outnumber opposite-sex besties by 4:1 for men and 6:1 for women.


P.S. Don't these results imply quite a bit of unrequited best friendship?  Hmm.

COMMENTS (4 to date)
NZ writes:

#2 is true only if one does not understand (A) what one personally looks for in a close friend, and never connects that with (B) behavioral/mental traits that are statistically common to each sex.

However, most people do have at least some intuitive grasp on both (A) and (B), so most men understand that the reason their wives' best friends would make poor close friends for themselves is simply because these friends are women, and they don't have enough in common with most women to support a non-romantic friendship. When my wife asks what I talk about with my friends and I tell her, she has to try hard not to either yawn or roll her eyes and grin--and that goes the other way too.

Therefore, that #2 "readily leads to sexism" is unlikely.

Instead, it's simply more likely that if a person is sexist (and what does that mean, by the way?) he will continue to be sexist towards his spouse's close friends.

Ditto NZ: "...what does [sexist] mean...?"

From Bryan's usage, I guess that the normative content of "sexist" is negative. If Bryan says something is sexist then that thing is bad in Bryan's view. Correct?

Now, if we try to induce the empirical meaning which Bryan attributes to "sexist", we have examples. That "men hunt together and women gather together" is perhaps a fact, but I believe Bryan does not perceive this as sexist.

I guess that Bryan is making a point about "sexism" as that word seems to be used in the mainstream -- but not as that word is used in Bryan's own judgments.

Emily writes:

Sure. But this also assumes that you're spending time with the opposite-sex spouses of your friends. I gave the extreme example of not-doing-this the comments: you know all of the same-sex people and no opposite-sex people except your own spouse. But there are of course less extreme examples.

For instance, if you only know the spouses of your close friends, and the correlation between peoples' traits and the traits of their spouses is high (which seems likely), the average opposite-sex people you're spending time with (spouses of close friends) could still be better than the same-sex people you're spending time with (close friends and not-close friends.) Also, if you have fewer opposite-sex friends, and distributions are the same among opposite-sex as same-sex people, your opposite-sex friends should on average be better (that is, align more closely with your preferences) than your same-sex friends, unless you have a really weird friend-finding algorithm.

NZ writes:


Algorithm? Distributions are the same? Only spouses of close friends? I don't understand what you're talking about. Who makes friends this way?

Even when all parties are unmarried (i.e. when there is no proxy friend selection via spouse), cross-sex platonic friendships are uncommon. Here are two reasons that have nothing to do with "native statistical reasoning":

-the inevitable cropping up of romantic feelings, creating awkwardness and frustration
-males and females tend to have very divergent interests, patterns of behavior, and styles of friendship

Comments for this entry have been closed
Return to top