Women bow down before the male god

On another weblog someone alluded to the sex difference in religious belief among black Americans, to the effect that it was more pronounced than among whites. Is this true? I decided to check the GSS, and found something interesting, though not too surprising. It's a robust cross-cultural finding that women are more religious than men, and it holds for the United States. But how does it break down across demographics?

Know God Exists
  Male Female Difference Proportional Difference
All 55.3 70.7 15.4 28%
Whites 52.4 67.8 15.4 29%
Blacks 73.7 86.6 12.9 18%
College+ 44.7 58.9 14.2 32%
No college 58.9 74.1 15.2 26%
Liberal 41.6 56.8 15.2 37%
Conservative 63.9 79.3 15.5 24%

As you can see women are more religious than men across various population segments in a rather consistent manner. Contrary to the intuition of some there is actually less difference among blacks than among whites. In fact inspection seems to show that the more secular a class is, the greater the sex difference. In other words, secularization seems to effect males to a greater extent than females within any give class. The diagram below illustrates it for the population segments above:

i-7ab6069f9b8b25f892b8b86ced440c8c-sexdiff.png

The pattern held when I did log-transformations. In any case, below are the data broken down by year. Some of the N's are small, so with caution, and for "entertainment purposes only." All the years show the percentage who "Know God Exists."

i-814a744c86fb6b688ccee7f0fc3fec34-allknowgod.png
i-00f520e2c307658a1c7c6c631ad252ad-whitesknowgod.png
i-9c49431f0c3c6dfb33e31a43a4ab6efd-blacksknowgod.png
i-2d1577f58c76ec78bab630f15faa2a8a-collegeplusknowgod.png
i-286a3de58d54fd3388f389cd954fbd90-nocolknowgod.png
i-7049c84f5123f30134f1953f4048ba9a-libknowgod.png
i-6148b7e32122383ae9f7dae3a01b2787-consknowgod.png

More like this

This is one of the more depressing gender gaps.

Well, good for you for getting me to click through by using an interesting post title. But how do you know women who "know god exists" aren't assuming a female god?

My wife once remarked that once you've experienced the miracle of giving birth, it must be easier to believe in other miracles. Can this be tested by comparing women who have and haven't given birth?

By bioIgnoramus (not verified) on 07 Aug 2009 #permalink

Interesting workup. Perhaps the propensity to logical thought is inversely correlated with belief? Seems a rational conclusion. The gender effect seems to confound the correlation between education and religiosity too, given more women today are achieving advanced degrees than men. Perhaps this trend hasn't existed long enough to be broadly represented in the population though. I wonder how the numbers would look if adjusted for education level and broken out by age?

But how do you know women who "know god exists" aren't assuming a female god?

oh, i assumed basic facts like the reality that women are more christian fundamentalist than men as a whole was common knowledge. but i guess people are ignorant about that too. perhaps i should post on that....

bio, children makes a huge difference. the male female gap grows children vs. no children. but the issue might be reversed.

alpha, i'll look.

My wife once remarked that once you've experienced the miracle of giving birth, it must be easier to believe in other miracles.

I've been told that pregnant women lose brain cells at a rate much higher than normal, but never have I encountered such a striking example of the effects of that loss on cognition.

Oh, look, I found a penny! Another miracle!

I'm surprised you made the first graph, treating the difference as a proportion, since the absolute number is so shockingly constant.

By Douglas Knight (not verified) on 08 Aug 2009 #permalink

What might have happened that caused the specific dips and peaks in certain years? Esp in 1993 and 1998.

I don't understand why you created the "proportional difference" column. Why did you calculate the difference as a proportion to the number of believers? Wouldn't it be just as meaningful to calculate it as a proportion of the unbelievers? But then the rankings would be reversed!

Similarly, I question your conclusion that "there is actually less difference among blacks than among whites." An equally valid way to describe the same data would be to say that "among blacks, twice as many men as women are unbelievers, a larger difference than among any other ethnic group."

Comparing differences in proportions is tricky. Maybe you need to think about a model of some underlying continuous latent variable.

> once you've experienced the miracle of giving birth,
> it must be easier to believe in other miracles.

or may be it's the other way around: believing in miracles is a precondition for women to have children...