Thursday, November 30, 2006

Divorce rates by religious affiliation I

(Post four of a series on Christian divorce rates)

Perhaps the best known series of sociological data is the General Social Survey (GSS), which contains data on religion and marital outcomes. Below I present the percentage of ever-married respondents who had ever divorced or were currently separated by religious affiliation and frequency of attendance.

Divorce rates by religious affiliation & attendance. General Social Survey, 2000, 2002, 2004 (N= 5,963)

58% Non-active Black Protestants
54% Non-active Evangelicals
51% No religious beliefs (e.g., atheists, agnostics)
48% Non-active other religions
48% All non-Christians
47% Active Black Protestants
42% All non-Christian religions
42% Non-active Mainline Protestants
41% All Christians
41% Non-active Catholics
39% Jewish
38% Active other religions
34% Active Evangelicals
32% Active Mainline Protestants
23% Active Catholics

Data from 1985 – 1999 show nearly identically religion-divorce patterns with an across-the-board increase in divorce rates. For these data.

Technical notes
The coding of the religious variable was taken from Steensland et al. (2000). Click here for full citation and description of coding.

“Active” was defined as attending church services about once a week or more often.

The GSS has various weighting variables for household composition and race oversampling. For simplicity sake, I have analyzed unweighted data, but others use weighted data.

I thank Christine Zozula for very-able research assistance. These analyses were informed by the work and advice of W. Bradford Wilcox, a sociologist at the University of Virginia who is an expert in this area. Thank you!

These data are from a working draft. I think that they’ve been done correctly, but I could be wrong. Any mistakes, and they do happen on a somewhat regular basis, are mine, and I will correct them as they come to my attention. I welcome replications of these and all of my analyses. Let me know what you find.

Tomorrow: Divorce rates by religious affiliation II


Anonymous said...

Some thoughts:

1) I don't understand why you condition on being currently married. Does that mean that currently divorced folks are excluded from your numbers?

2) Would the numbers look different if this weren't a household survey? What about a random sample of American adults? Presumably, married couple households are contributing twice to these numbers (but could practice different religions) while separated married households may be contributing just one person to the numbers.

3) What about the marriage as a unit of analysis, rather than currently married people? What's the survival rate of marriages disaggregated by religious affiliation of the officiant?

4) Can you reproduce prior results and tell us what they did wrong? I suppose that's in a future installment. Interesting stuff!

Brad Wright said...

Very good questions Gary.

1) I conditioned on having ever been married & have clarified that in the text. Good catch.

2) I'm not sure that not being household would make a different beyond weighting issues. The GSS picks an single individual in the household, so it wouldn't report if either spouse has had a divorce. But... there may be more subtle sampling problems here that I haven't thought about. Thoughts from others?

3)You're right that a more sophisticated analysis of marriage would want to use the couple as the unit of analysis. Then you could add relationship variables as well as individual.

My analyses here are part of a larger project on empirical stereotypes of Christianity, so I don't go that deep into marriage. I think that the sociology of family folks like to use the National Survey of Family and Households for its partner data.

4) I can't reproduce Barna's findings, for he doesn't release his data. (For a discussion of this, see the comments on my post on Tuesday).

I think that I know why the different substantive conclusions, and I'll post my ideas next week.

Thanks for posting!


Knumb said...

Even figure 3 or four points one way or another on each finding, due to error or missampling, these are very interesting findings.

54% Non-active Evangelicals... some kinda slingshot effect, it'd be interesting to know the cause.

No active/non active Jewish, or is the percentage the same for each?

23% Active Catholics Hanging on like a crow landing in windy weather *cough* *cough*.

Jay Livingston said...

Obviously the big difference is between Active and Nonactive (no surprisethere). As for Catholic/Protestant, I just ran a quick-and-dirty controlling for region. It looks as though the difference shrinks to insignificance in the South and East.

Brad Wright said...

Very few Jewish respondents attended services once a week or more, so I didn't distinguish between active and non-active Jews.

Interesting about regional differences between Protestant & Catholic.

The sample size is so large here that just about any difference will be statistically significant, but I'm not sure how robust the Catholic vs. Protestant difference is, so for me that's not the big story as much as active Christian versus non-active or atheist.

Jay Parmar said...

According to estimates, almost one half of all marriages will ultimately end in divorce. Since increasing the well-being of children is the alleged goal of the marriage promoters, it would seem logical that some effort would be made to help the divorce process to be less of a traumatic experience for children. Children brought up by separate parents or divorced parents are always unhappy and seems to vent their anger on the society. Even though they pretend to be happy, innermost they seem to have missed a greatest thing-the joy of living with both parents. Sometimes,divorce is not the solution to the dispute in the family but one of the partners tend to seek liberty for his or her infidelity.Parents should try and think twice for the sake of the children and their future. Visit Divorce UK for more information.