Monday, February 14, 2011

Religion = Romance? Divorce Rates vs. Church Attendance



I've lost track of how many times I've heard pastors and Christian authors quote the statistic that divorce rates are the same for church-goers as in the general population. This is misleading and disheartening to many devout Christians. The truth is, church attendance has a profound correlation with a reduced probability of divorce.

Sociologists and economists studying religion tend to use frequency of church attendance as a proxy for religiosity. (The notion being that those more serious about their faith are more likely to be involved in church than those less serious.) This measurement has far more predictive power on a whole host of social outcomes than simply asking if someone is a Christian.

To illustrate, above is a graph I just created in Stata showing the impact of church attendance on divorce rates. (Based on data from the 2006 General Social Survey.) The effect is particularly strong for men who attend church on a weekly basis.

Here are a few takeaways from the graph:
  • Just over half of non-weekly church attenders who have ever been married have been divorced at least once.
  • Approximately 41% of women who attend church weekly who have ever been married have been divorced at least once. One way to think of this is that weekly church attendance is correlated with an approximately 20% reduction in the probability of divorce for women.
  • Approximately 32% of men who attend church weekly who have ever been married have been divorced at least once. One way to think of this is that weekly church attendance is correlated with an approximately 40% reduction in the probability of divorce for men.
  • Part of the lessened effect for women is likely due to the fact that more women than men attend church on a weekly basis. Statistically, this means weekly church attending women are more likely to be married to a non-weekly attending man than vice-versa.
  • This data may actually understate the differences between weekly and non-weekly church attenders. Weekly attenders are more likely to get married, while non-weekly church attenders are more likely to cohabit as alternative to marriage. (Cohabitation has a much higher likelihood of dissolution than marriage.)
It is important to note that data shows a correlation only and does not necessarily imply causality. (It could be an attribute that makes an individual likely to go to church also makes that individual more likely to marry and less likely to divorce.) However, as a weekly church attender myself, I find this data highly encouraging.

See more of my posts on the economics of religion.

4 comments:

David said...

Suppose there is a causation; what I'd like to know is greater religiosity lead to better mate selection (e.g. through mixers and thus less chance of disagreement) or is greater religiosity encourage stubbornness about leaving an unhappy marriage? (These are the two obvious sources of causation I could think of.)

Brian Hollar said...

That's posible. Although you'd have to reconcile it with data that people with higher religiosity also score higher on measurements of happiness than less religious people do too.

I think what is more likely that:

1) Religious people view divorce more negatively that less devout people. Having an ex ante aversion to divorce prior to marriage could incentivize this group to take marriage more seriously and lessent their likelihood to divorce.

2) Greater commonality of shared belief would also likely reduce the probability of divorce. (EX: Cross-cultural and interfaith marriages have higher divorce rates than the general population.)

3) Increased monitoring by a share social group (churches) along with support from these groups to those who are struggling in marriage may also contribute to lower levels of divorce.

4) There could be a happiness effect. If religious people are more happy, it could also be that happy people are less likely to divorce.

All very interesting questions....

Kevin said...

Brian,

"I've lost track of how many times I've heard pastors and Christian authors quote the statistic that divorce rates are the same for church-goers as in the general population. This is misleading and disheartening to many devout Christians"

I believe this is coming from a Pew report, quoted on Focus on the Family. I've heard it since the mid-nineties (then US total population ~280M, Catholic ~70M)and I believed it was Evangelical Christians. So Catholic, Main-line Protestant and Orthodox could easily account for the difference between your graph and the Pew statistic in my mind.

Granted, I usually hear "Evangelical" thrown around like "we're the white hats," so if Pew is accurate and your is also, maybe a little less pride on the adjective is in order.

Sam said...

I think your four possibilities pretty well cover it