October 7, 2010

The Reversal of the College Marriage Gap

IV. College and Marital Stability

Marriage does not always last a lifetime, and a prominent fault line in marital stability is education. College-educated adults are less likely to experience divorce and multiple marriages than are less-educated adults (Bramlett and Mosher, 2002).

In 2008, the Census Bureau started to collect information on whether an individual had divorced in the past year. Among all 25- to 59-year-olds who had married once and were married within the past year, 2% report divorcing within the past year. College-educated individuals in first marriages were less likely to divorce (1.4%) than less-educated individuals (2.3%).

Since those obtaining college degrees used to delay marriage to a greater extent than less-educated individuals, it is possible that the college-educated in first marriages are older than the lesser educated in first marriages and thus less likely to divorce. Yet, even within narrow age groups, first marriages of college-educated individuals are less likely to have dissolved in the past year than the first marriages of less-educated individuals (Figure 10).

Similarly, the greater marital stability of the college-educated does not solely reflect basic demographic differences between college-educated adults and adults without a college diploma. Women’s first marriages are more likely to dissolve than men’s, and black adults have higher divorce rates than white adults. Table 2 reports first-marriage dissolution rates within gender and race groups, and the college-educated are less likely than the less educated to experience divorce.

In regard to marital length, longer first marriages are not necessarily more stable than shorter first marriages. The annual risk of divorce increases during the first five years of marriage and then smoothly declines (Kreider, 2005). Yet, the lower divorce probabilities of the college-educated are not solely due to the fact that they are in longer marriages. Table 2 reports the likelihood of marital dissolution by length of first-marriage intervals.9 For most age groups and marital intervals, the college-educated display lower risk of divorce than the less educated.

A second crude indicator of marital instability is the rate of being married more than once. In 2008, among 25- to 59-year-olds who have ever married, about 23% have married more than once. The college-educated ever-married adults were less likely to have married at least twice (17%) than less-educated ever-married adults (26%). Because older adults are more likely than younger adults to have ever been divorced or widowed, the multiple marriage rate rises with age. In any age group in 2008, however, the college-educated ever-marrieds were less likely than their less-educated counterparts to have married more than once. For example, in 2008, among college-educated 55- to 59-year-old ever-marrieds, 28% had multiple marriages. In comparison, 37% of their less-educated counterparts had multiple marriages (Figure 11).

The college-educated are less likely to be married multiple times in part because they are less likely to experience divorce.10 They are also less likely to experience the death of their spouse. Once divorced or widowed, it is not clear that the more educated are any more or less likely than the less educated to remarry. Some studies find no clear association between education and remarriage propensities (Bramlett and Mosher, 2002).

The college-educated have been less likely to be married multiple times for much of the 20th century. The 1990 and 2000 censuses did not ask respondents how many times they were married, but censuses up to 1980 also show that the multiple marriage rate is lower for college-educated adults (Table 3).

  1. The 2008 American Community Survey does not have a direct question on length of marriage. It does, however, ascertain year of marriage, so for persons currently in their first marriage, one can infer marriage length. For the divorced and widowed, however, duration of marriage is not available. For those who divorced in the past year, I assumed that duration of marriage equals 2008 minus year of marriage minus 0.5 years.
  2. But the likelihood of a first marriage ending in widowhood in the past year is pretty minute for any spouse, given that this analysis examines 25- to 59-year-olds.