Inductivist

I showed recently that Americans overwhelming select someone of the same race for a best friend.  Friendship is a good measure of how well groups are mixing, or, on the other hand, how much they cluster.

In this post, let's focus on religion. Does your best friend share your religious affiliation?  I used General Social Survey (GSS) data to answer this question (sample size = 1,932).  Here are the percentages who indicated a religious match between themselves and their best friend:

Percent whose best friend belongs to the same religion

Jews  75.6
Protestants  74.5
Catholics   59.4
None  42.3

Jews emerge as the group that clusters the most.  You might respond that Protestants have basically the same number, but we would expect a high number for a such a large group.  According to the latest GSS, 48.9% of Americans are Protestant, 23.3% are None, 21.2% are Catholic, and 1.7% are Jewish. If Protestants picked their friends at random, their best friend would also be Protestant around half the time. For Jews, a random process would give a Jewish best friend less than 2 percent of the time, and yet the number is over 75%.  This shows an intense level of clustering among Jews.

Males N = 547

Females = 593

My grandmother told me when I was a kid that men and women can't just be friends. Like so many, I suffered from the idealism of youth and didn't want to believe her. Now, I think that it's pretty much true, unless the friends are so hideous, they couldn't imagine sleeping together.

But let's look at the General Social Survey to see how often we see cross-gender best friends. The top graph shows what male respondents had to say about who their best friend is--other than their current partner. I've also broken it down in terms of marital status. I've done the same thing for women in the lower graph.

For men, having a girl as a best friend is quite rare for married men and never married men (8.5% and 10.4%, respectively) but it's close to 20% for separated, divorced, and widowed guys. Younger guys have more (male) friends than older guys, which might explain the low number of female best friends among the never-married. Older men who have been through a breakup or the death of a spouse probably don't have a group of pals anymore and so are more likely to make a close friendship with a women.

Still, I'm a bit skeptical of at least some of these friendships--I imagine one or both friends have thoughts of something more, but don't report it because it hasn't developed that far yet. And of course, cross-sex friendships are rare among married guys. What wife with any sense would let her husband get socially intimate with another woman?

It's the same for married women: only 3.1% have a man as a best friend. Hmm, isn't that estimate also about the same as the number of gay men in the population? It's doubtful that a husband would allow any other type of close male friend. It's also rare for widowed women: perhaps that is due to the fact that men are comparatively scarce among older people. It's a little more common among separated and divorced women. Fourteen percent of never-married girls have a male best-friend. Once again, I wonder if some of these are possible future relationships or one-sided romantic/sexual ones.

Basically, I think Grandma was right.