I was surprised when I read this passage in a Mark Regnerus op-ed two years ago, about the well-known instability of early marriages:
[W]hat is considered “early marriage” by social scientists is commonly misunderstood by the public. The best evaluations of early marriage — conducted by researchers at the University of Texas and Penn State University — note that the age-divorce link is most prominent among teenagers (those who marry before age 20). Marriages that begin at age 20, 21 or 22 are not nearly so likely to end in divorce as many presume.
Are later marriages a bad thing? No. Economic stability and higher education are leading predictors of successful marriages, and both take time to acquire. Couples who marry before the age of 25 are twice as likely to divorce as couples who marry after 25. Avoiding divorce is good for both children and adults. Duh.
Now that statistical claim could in theory be correct, and compatible with Regnerus’s claim. Let’s say that divorce rates are off-the-charts high for women who were 18 at marriage but fall sharply as women turn 21. Those high divorce rates for the youngest brides could still make the under-25/over-25 numbers look as different as Goldstein says. The number would then be true but a bit misleading.
The thing is, though, that I don’t think Goldstein’s claim is correct. Follow her link and you find this claim, which I think is the one to which she must be referring: “Divorce is more likely when women marry at a younger age (48% of brides married before age 18 divorce in 10 years, compared to 24% married at age 25 or later).” So the twice-as-likely figure is actually a comparison of brides over 25 to brides under 18.
Why have I spent so much time on what might have been just a typo on her part? Partly just out of curiosity. Also because I’m open to the possibility that the age of first marriage in the U.S. has risen high enough to be a cause for concern. If I’m right about these numbers, then Goldstein’s counterargument is weaker than she thinks: Further increases in that average age will not bring big gains from a lower divorce rate; decreases in that average will not bring big losses. And last because I think that one reason the average age of marriage has increased may be that people have exaggerated views about the instability of marriages formed in the early 20s (and in some circles may even have exaggerated views about the instability of marriages formed in the mid-20s).
Naturally, I shot an e-mail about this question to W. Bradford Wilcox, with whose response I’ll close:
The data generally suggest that marrying in mid-twenties brings you to a low-divorce plateau, and that waiting beyond that doesn’t help your divorce risk. Moreover, couples who wait much longer than this are less likely to be happy in their marriages (but still enjoy high levels of marital stability). So mid twenties is the sweet spot–on average.
More specifically, I’d say that couples who marry in early 20s have a slightly higher divorce risk–on average. (But I’d say couples who are both religious will do fine, insofar as they have community & normative support for their marriage.)
One problem, though, is that Americans without college degrees are still having their first children-on average-in their early 20s, often before they get married. College-educated Americans, by contrast, have postponed childbearing into late 20s and 30s, usually after marrying-so their risk of nonmarital childbearing is very low.