University Times

Since 1980, the divorce rate in the United States has leveled off after more than a half-century of rising steadily.

But does that leveling off apply equally to major racial and ethnic groups in this country? And, if not, do differences in risk factors for divorce account for the differences in divorce rates between the groups?

Those were the key questions that sociologist/demographer Megan Sweeney discussed at an Oct. 16 lecture here. The lecture, part of a series sponsored by Pitt’s Center on Race and Social Problems, focused on research published in “Can Differential Exposure to Risk Factors Explain Recent Racial and Ethnic Variation in Marital Disruption?” which Sweeney co-authored with Julie Phillips, a sociologist at Rutgers University.

Relying on data from two national studies — the U.S. Census Bureau’s Current Population Survey (using data from 1985, 1990 and 1995) and the 1995 National Survey of Family Growth — the researchers looked at a national sample of three groups: non-Hispanic white women, non-Hispanic black women and Mexican-American women (both U.S.- and foreign-born).

The U.S. Census sample was large (more than 40,000 women participants across the three groups), but did not measure risk factors adequately, while the latter, smaller sample (3,222 whites, 751 blacks, 471 Mexicans) did measure divorce risk factors.

The data all were taken from surveys and interviews of women ages 15-30 in first marriages formed since 1975, said Sweeney, who is an associate professor at UCLA.

A cursory glance at divorce rates yields three general demographic facts, she said:

• The divorce rate over the past century, with a few blips and dips, shows a single trend: There was a steady increase in the divorce rate until 1980.

• That increase was particularly dramatic in the mid-1960s. “Social scientists really pay attention to this, because divorce has been associated with negative economic, emotional and health outcomes for adults and children,” Sweeney said.

• More recently, the divorce rate for the U.S. population as a whole has remained unchanged since 1980.

“If the plateau of the divorce rate continues, the children born in the late-1970s will, for the first time, face risks of divorce similar to their parents,” Sweeney noted. “This is an important finding, because as a trend it can help us predict what future families will look like, but it’s very unclear if that finding applies across the whole U.S. population. What made us think it may not apply is that there is considerable diversity across racial and ethnic groups in patterns of marital disruption, that is counting separation and divorce, instead of simply legal, formal divorce.”

Recent evidence suggests that almost one third (32 percent) of marriages among white women end in separation or divorce within 10 years, compared to almost half (49.4 percent) among black women.

And, although 27.5 percent of first marriages among Mexican-American women end within 10 years, this overall figure conceals important differences by birth status: Only 13.1 percent of foreign-born Mexican women’s marriages end in separation or divorce, compared to 41 percent of U.S.-born Mexican women’s marriages, Sweeney said.

“If you look at 1980, the difference in rate of marital disruption between black and white women was about 9 percent; by 1993, the difference expanded to about 29 percent. So we definitely have a widening gap in disruption rates,” she said.

Thus looking at divorce rates alone does not tell the whole story of marriages that are unsuccessful, she added. The three basic demographic facts regarding the divorce rate are, in fact, misleading, when they are applied to racial and ethnic groups.

“It turns out, when you look at race and ethnicity in marital instability, for example, there’s great diversity in the proportion of white and black women who go on to divorce after they separate,” Sweeney said. “About 97 percent of white women go on to divorce within five years after they separate from their spouses, but for black women it’s 67 percent, so there’s a big difference there and it shows that there is a high percentage of black women who end up in a long-term separation without divorcing.”

Moreover, the risk of marital separation has continued to rise for black women since 1980, whereas for white women it has been drifting downward, she said.

“So what we thought of as a demographic fact, this leveling of the divorce rate across the population, is not a demographic fact for African Americans,” Sweeney said. “It simply doesn’t apply.”

This conclusion led her to a second analysis that focused on two questions:

• Can differential exposure to key risk factors explain these overall group differences in levels of marital disruption?

• Do the major risk factors of divorce and separation apply equally to racial and ethnic groups?

“It’s important to identify the underlying causes of these large group differences regarding marital instability, given the evidence of adverse outcomes for adults and children,” Sweeney said.

Using data from the National Survey on Family Growth, the researchers looked at the following risk factors: age at first marriage; education; employment status; whether the woman had a birth or a conception before she got married; whether she had sex before marriage; whether she cohabited before marriage; the region of her residence and whether she lived in a major city; whether she was born in the U.S.; whether her parents divorced by the time she reached age 14; her mother’s educational attainment, and whether she came from a Catholic background.

“We also looked at some characteristics of the marriage,” Sweeney said. “There is evidence that when spouses are not well-matched on some major characteristics that that can contribute to marital instability.”

Those characteristics included whether the wife was older, younger or about the same age as her spouse; whether there were significant differences in educational attainment; whether the husband had been married previously and whether he had children, and whether the couple were of the same race and ethnicity.

“What we’re doing is what’s called a decomposition,” Sweeney said. “We were looking at group differences in levels of marital instability, and we’re asking, ‘What proportion of that difference, comparing white and black risks of divorce, for example, is attributable to differences in exposure to the risk factors I just mentioned?’”

Sweeney’s study analysis concluded that about 30 percent of the white-black gap is attributable to group differences in these characteristics or risk factors.

“Our results suggest that if differences in population composition between groups were removed, the white/black differentials in disruption would be reduced by approximately 30 percent; the black/Mexican differentials by about 50 percent,” she said.

“However, when you try to explain the white/Mexican gap, the story gets entirely muddled, because of the wide gap within the Mexican group based on nativity status.”

The bottom line, she said, is that differential exposure to risk factors only partially explains the gaps in marital disruption between groups. “There is a lot more going on that needs much more profound analysis.”

Sweeney’s research also examined whether risk factors actually operate in the same way across groups.

“For example, cohabiting with a partner before you marry is associated with a substantially elevated risk of divorce or separation after you marry. Your risk of marital disruption is about 50 percent greater than if you didn’t cohabit.”

There is no single reason to explain why couples cohabit, she pointed out. “One reason might be as a trial marriage: Is this the guy I want to spend the rest of my life with? Maybe the best way to figure that out is to live with him for a while, really get to know him,” Sweeney said. These unions are characterized as having no firm plans for marriage and point to a concern that the relationship might not work out permanently, she said.

A second reason to cohabit is as a substitute marriage, where couples are committed, but they do not feel the need to make the arrangement legal. A third model is cohabiting as an extended engagement, a precursor for marriage where the couple is committed to the relationship but unable or unwilling to marry due to outside factors such as economic instability.

“All those models of cohabitation are operating across all groups,” Sweeney said. “But how frequently are they represented? There’s some evidence from sociological and demographic literature that they may be differentially represented among these three groups. For example, Mexican Americans and blacks generally are more likely than whites to use cohabitation as a context for child-bearing.”

Attitudinally, black and Mexican women are no more accepting of cohabitation than white women. “It’s not that they think cohabiting is better than marriage; it’s an alternative that might be more attractive or necessary based on economic factors,” Sweeney said.

“Also, while white and black women are equally likely to have plans to marry when they begin cohabiting, white women are much more likely to realize those plans.”

With respect to Mexican Americans, culturally there is more acceptance of consensual unions, so that they often function as a surrogate marriage for lower-economic populations.

“So for Mexican Americans, there is no association between cohabitation and marital disruption, for black women there is almost no effect statistically, and yet what we see is that the risk of disruption is 48 percent greater for white women who cohabit than for those who don’t,” Sweeney said.

“In a nutshell, the major demographic facts simply do not apply to some of our larger racial and ethnic groups: The fact that the divorce rate stopped rising in the United States after 1980 does not apply for black women at all when separation is factored in. Or, the fact that cohabitation before marriage is one of the strongest risk factors for disruption, that does not apply at all for blacks or Mexicans,” she said.

Sweeney acknowledged that her research is much more successful at debunking demographic “facts” than in explaining the group differences.

She said demographers, generally, also were hamstrung by a decreasing amount of relevant data. “One thing to note is that these data are up only to the mid-1990s, so they’re 10 years old now,” Sweeney said. “What’s happened in the last 10 years? Well, the answer is we don’t know; we actually can’t know because the data are not available to us for the more recent years.”

The Current Population Survey, for instance, in 1995 stopped compiling marital histories, such as the age of first marriage. The U.S. Census no longer asks whether an individual has divorced and remarried, only “married or single?”

“In all fairness to us, that puts us right in line with those who try to explain racial and ethic differences in family patterns generally. We’re just not doing a good job at that,” Sweeney said.

—Peter Hart