Wednesday, May 23, 2007

Christians and promuscious sexual behavior

(Post 5 in a series)

In the last several posts, I looked at pretty standard measures of sexual behavior--premarital and extramarital sex. Here I look at a measure that perhaps gets at promiscuity. The General Social Survey asks respondents how many sexual partners they have had in the previous year, and below are the percentage of respondents who reported that they had 3 or more sexual partners. I chose the number three because with two it could be someone getting a divorce and remarrying. Three implies something more.

% of respondents with three or more sexual partners in the past year:


13%, No religious affiliation (n=2,369)
12%, Non-active Black Protestants (580)
9%, Non-Active, other religions (740)
9%, Non-Active Catholics (3,166)
8%, Jewish (406)
6%, Non-active Mainline Protestants (2,695)
6%, Active Black Protestants (580)
6%, Non-active Evangelicals (2,795)
3%, Active, other religions (454)
2%, Active Evangelicals (2,361)
2%, Active Catholics (3,166)
1%, Active Mainline Protestants (1,070)

Caveats: I am not controlling for age... maybe it's just that young people sleep around more and young people are also less religious. I'm also not controlling for marriage. Maybe religious people are more likely to be married and married people are less likely to have multiple sexual partners.

GSS data, 1985-2004.
Next in series: Christians and pornography

6 comments:

Jay Livingston said...

Brad,
What are the N's? When you get down to something that's relatively rare, broken down into so many categories, you may need a very large sample. For example, your data show that twice as many Active Catholics as Active Protestants have had 3+ partners -- 2% vs. 1%. In some studies with large samples, that kind of difference might be important -- suppose, for example, these were rates of some rare but serious disease -- but without the N's, it's hard to know what to make of these data.

Unknown said...

Or, maybe atheists are more honest about such things than religious respondents. Brueckner's work on the virginity pledge is relevant here.

Jay-- if Brad used the coital frequency variable I think he used, his N would be around 640 respondents (not inc. missing values on level of religious activity). If so, I'd guess none of the differences between the "active" categories are significant, nor is the difference between atheists and agnostics or non-active black protestants.

SARA said...

I lack the knowledge on how such surveys are carried out, so depending on the level of privacy a survey like this is conducted, i'd tend to agree with Kim that perhaps athiests would be more honest than religious folk.
I'm not sure if a survey regarding pre-marital sex would ever be accurate unless conducted in a "secret ballot" similar to voting proceedure..this topic is private/personal.

Brad Wright said...

I'll post the Ns soon... Usually I try to minimize the number of numbers I present, since I'm aiming these things at the lay person, but you make a good point as to why the Ns would be important here.

Underreporting bias is a distinct possibility (as is supurious and reverse causation). Does anyone know of any studies of religion and survey responses?

Brad Wright said...

Since I'm using 20 years of the GSS, the total N is over 20,000. Even within categories there is still plenty of "N". Typically, I don't like to report cells that have less than n=100 (which is why I don't report active vs. inactive Jews).

I'm not too concerned about 1% vs. 2%. Since I'm not going past the decimal, it could be 1.48 vs. 1.51, with rounding. Rather, I'm trying to portray the data in rather broad strokes.

Glen Davis said...

Love the series of posts. Very stimulating. There's an interesting related article at http://www.slate.com/id/2167293 "Even Evangelical Teens Do It"

I'd love to hear what you think about it.