The Marriage-Preservation Debate

Reexamining the research on divorce

Magazine Issue
March/April 2006
The Marriage-Preservation Debate

During the past decade or so, a movement of “marriage savers” comprised of therapists, marriage-education programs, and religious and secular promarriage organizations has been pushing back against what many feel is a divorce culture run amok. To cement their case against divorce, many marriage savers, particularly social conservatives, quote liberally and publicly from what they call “definitive” research on the damaging effects of broken marriages on individual spouses, families, and, most of all, the “children of divorce.”

You’d never know, listening to this polemicized use of evidence, that the social-science research on divorce actually presents a far more nuanced and less pessimistic picture. What then does the research about divorce tell us?

The Impact of Divorce

To understand the controversy about the impact of divorce, it’s helpful to take a look at the history of research into the issue. Perhaps most famous among the early research were the longitudinal studies of Judith Wallerstein, now of the Judith Wallerstein Center for the Family in Transition in California. In 1976, she conducted a study in which she interviewed a sample of 131 young children and their parents from middle-class, white, urban, northern California families who’d recently gone through a divorce. In the years following, she interviewed the children of these divorced families at several junctures in their lives into young adulthood, although only 93 of the original 131 children were reinterviewed at the 25-year mark.

In their initial study, Wallerstein and her then collaborator, Joan Kelly, found a range of reactions among the children in their sample just after the divorce. Some seemed to be handling the disruption without major emotional difficulty, while others were struggling with depression, school difficulties, and other types of psychopathology.

After the first study of these children, Kelly (who came to doubt Wallerstein’s conclusion that children were “scarred for life” by divorce) left the project. Over the next 20 years, Wallerstein alone conducted three follow-ups of the children from these divorces, and came to increasingly dark conclusions about the long-term effects of divorce. By the time she’d produced her last follow-up study in 2001, she believed that divorce not only was harmful to children when it happened, but led to what she termed a “sleeper effect,” which crippled their ability to form romantic relationships as adults.

Wallerstein’s view was adopted wholesale by socially conservative “family values” proponents, and widely trumpeted as proving beyond a doubt that virtually all children whose parents divorce suffer traumatic and lasting emotional injury.

Yet the methodology underlying Wallerstein’s famous study is today considered primitive by most social-science researchers.

In the first place, she’s criticized for basing her conclusions on interviews with a small sample of divorced families representing only one particular kind of family–urban, middle class, and Caucasian. The study, critics contend, is thereby low in what’s termed “external validity”–the ability to represent the general population.

A more serious objection is that Wallerstein’s version of the effects of divorce may not accurately represent even the population she studied. There was never a control group–a set of comparable individuals at the same junctures in life who hadn’t experienced divorce in their family–which is now virtually de rigueur in this kind of study. Therefore, it’s impossible to know how many of the difficulties Wallerstein observed in the children of divorce might also hold true for those whose parents remained together.

Additionally, Wallerstein’s data-gathering methods have been challenged. Rather than being a dispassionate observer, she clearly developed relationships with many of the children she followed, and encouraged them to read her books. Some critics suggest she actively cocreated with them the idea that they were irreparably damaged by their parents’ divorce. Thus the threats to the validity of her findings render the Wallerstein reports a good read, but poor science.

Wallerstein’s study did have the positive effect of spurring some of the most prominent social-science researchers in America to check out whether her findings applied to other populations. Of the many projects that have assessed the impact of divorce on families, the most prominent have been those conducted by Mavis Hetherington of the University of Virginia, who followed divorced families over time; Christine Buchanan, Eleanor Maccoby, and Sanford Dornbusch of Stanford University, who examined the effects of divorce on adolescents; Constance Ahrons of the University of Southern California, who examined the effects on the adult partners over time; and Robert Emery of the University of Virginia, who examined the feelings of young adults whose parents had divorced when they’d been children.

So what does this later research tell us? A great deal, as it turns out, and it’s a far more complex, and even hopeful, story than Wallerstein’s research suggested.

(1) For almost everyone experiencing divorce, there are indeed short-term negative consequences. Both children and adults show more role strain, a greater number of emotional and behavioral difficulties, and, in Mavis Hetherington’s provocative words, a feeling of not “being me”– that is, of acting and feeling in ways that are atypical.

(2) But after a period of one to two years, most members of divorced families do well and can’t be distinguished on measures of functioning, symptoms, or happiness from their counterparts in families that didn’t go through divorce. Most members of divorced families don’t suffer from depression, anxiety, and school or work problems, nor do they experience grave disruptions in their lives.

(3) The incidence of problems in both children and adults who’ve gone through divorce remain higher, over time, than nondivorced families, but only slightly higher (e.g. about 20 to 25 percent in children from divorced families vs. 10 to 15 percent in other children). Thus the vast majority of children whose parents divorce, even when they have problems, fall within the normal range on all measures of functioning and symptoms.

(4) There’s no specific evidence that family members who go through divorce are any worse off than those headed by an unhappily married couple who stay married. But high conflict in marriage is a major risk factor (and many researchers would say the greatest risk factor), both for children of parents whose marriages last and for children of parents who divorce. It isn’t known, however, whether high conflict is worse for children when it’s an aspect of divorce (either before or after the divorce) or when it’s chronically present in a marriage that remains intact. In either context, though, high conflict hurts everyone.

Even though children from divorced families don’t necessarily do worse in life than those whose parents stay together, children of divorce almost universally experience emotional suffering at the time of the divorce, and have strong feelings about the event long afterward. When Robert Emery surveyed college students at the University of Virginia about their experiences with divorce in their families, he found that even these academically successful young people typically told stories of a family life filled with pain. For example, nearly 50 percent believed they’d had a harder childhood than others (compared to 14 percent among adults whose parents’ marriage remained intact), and 28 percent wondered whether their fathers loved them (compared to 10 percent among adults whose parents hadn’t divorced).

5) Also, according to some researchers, one group of children is, in fact, worse off after divorce than they were when their parents were together. These are children whose predivorce family life had been relatively untroubled and who’d never witnessed any overt signs of unhappiness or conflict between their parents before the divorce. Sociologist Paul Amato of the University of Nebraska, who wrote The Post-Divorce Family with Ross Thompson, for instance, considers the children of low-conflict parents particularly vulnerable. Others, however, including Constance Ahrons, point out that there’s simply no way of knowing how their lives would have progressed if their parents had stayed together. Obviously, it isn’t possible to compare how the same children do when their unhappily married parents divorce and when their unhappily married parents don’t divorce, nor is it possible to project how those unhappy marriages would evolve over time.

In sum, divorce is undoubtedly painful, and we simply don’t know whether it’s better or worse for children if their parents divorce or stay together in an unhappy marriage. There are indications, however, that the ongoing happiness of children has less to do with the simple fact of divorce than with the nature of their family life–whether it’s basically stable, calm, and without open conflict and emotional turmoil, or whether there’s a lot of fighting, anxiety, and depression. It’s certainly possible that children of divorced parents who grow up in a peaceful, stable, financially secure environment without a lot of parental conflict may well be better off than children living with parents who don’t divorce but fight all the time.

Predicting Divorce

Another line of research with direct implications for the practice of psychotherapy is research predicting who’ll divorce. It’s been possible for some time to predict which marriages will be happy and which will be unhappy by looking at factors like the ratio of positive to negative interactions, presence of spousal violence, psychopathology in one or both spouses, communication skills, and level of shared expectations. But until the last decade, it was a truism that predicting which couples would actually stay together and which would divorce was impossible–there were just too many seemingly random variables.

John Gottman’s groundbreaking research has radically changed this view, however, by uncovering the actual factors that foreshadow divorce. What Gottman and his collaborator, Richard, Levinson, isolated as predictive factors weren’t aspects of individual personality, but certain interactive traits. Gottman and Levinson found that the presence in the marriage of what they termed the “four horsemen” (criticism, defensiveness, contempt, stonewalling, and a recently added fifth, belligerence) are very strong signs that a divorce is on its way.

It appears, for example, that when one partner treats the other with contempt or stonewalling–acting as if the partner isn’t present, isn’t worth listening to, or isn’t worthy of respect, even in subtle ways–the marriage is probably on its way out. Surprisingly, Gottman found that the extent to which couples argued was unrelated to whether they divorced. Instead, what mattered was how couples treated each other, especially during arguments. He describes how these patterns of mutual mistreatment (one or more of the “five” horsemen) tend to produce what he calls a “cascade” effect–a downward spiral of increasingly bad behavior by both spouses, leading with a kind of inevitability to divorce.

Gottman’s insights are invaluable to clinicians, providing them with empirically validated information they can use to great effect with couples who are on a slippery slope. In my own practice, I’ll say directly to clients exhibiting one of Gottman’s “horsemen” that their behavior corresponds with patterns that research has shown lead to the demise of marriage. Hearing this can be a loud wake-up call for couples, making them more aware of what they’re doing and promoting a real sense of urgency about changing the way they treat each other.

Preventing Divorce

A great deal of, maybe most, marital therapy is conducted with couples hovering on the brink of divorce. With the divorce rate now at around 50 percent of marriages in the United States, a fundamental goal of marital therapy and of the marriage education and enrichment programs available these days is to help keep couples together in at least reasonably satisfactory marriages. And yet, very little actual research has been done to answer the most fundamental questions about the impact of marital therapy and marital education and enrichment programs on divorce. Do these therapeutic and educational interventions actually reduce the frequency of divorce?

Up to now, the large sample sizes and lengthy follow-up required to examine this question, coupled with the limited funds available to study such interventions, have left us with few relevant studies. Although it’s been established that both marital therapy and marital preparation/ enrichment programs are effective in improving relationship skills and marital satisfaction, their impact on divorce is as yet unknown.

And we haven’t even begun to have research that examines complex questions related to the marital therapy field, such as whether marital therapies do a better job of improving marriages and keeping couples together when they demonstrate an active bias in favor of staying married or when they assume a neutral stance. According to the marriage preservers, the answer is obvious: the more a therapist “fights” for the marriage, the better chance the marriage has. But there’s no research clearly supporting this position.

Nonetheless, there have been some interesting research findings related to therapy, marriage education and enrichment programs, and divorce. For instance, Kurt Hahlweg, Howard Markman, and their colleagues showed that after taking a six-session Couples Learning program, 9 percent of the recipients divorced after three years, compared to 22 percent in the control group, who hadn’t taken the course.

In another study, Andrew Christenson of UCLA, the late Neil Jacobson of the University of Washington, and their colleagues studied the impact on couples’ intentions to divorce of two treatments–traditional, behavioral marital therapy and a version of behavioral therapy accenting the acceptance of one’s partner called Integrative Behavior Couple Therapy. They found that both treatments decreased clients’ thoughts about divorce to a level matching that of typical nondistressed couples, even though many couples in this study had begun treatment with a high level of marital dissatisfaction and were close to ending their marriages. An interesting finding in this study was that changes in couples’ thoughts about divorce most often came early in treatment (before the 12th session), suggesting that changes, if they’re going to occur at all, happen in the early stages of therapy.

Thus, it seems that at least some marital therapies and some education programs do have an impact on divorce rates. Yet the research is still too limited to give us much insight into the complex questions of when marriages can and can’t be improved, when the therapist’s commitment to preserving the marriage is most helpful, or when divorce might be the best choice for everyone in a family. If researchers can find the money to do the research (and for whatever reason, there’s almost no money available to study marital therapy aimed at distressed marriages), these are among the questions that need examination.

We know a lot more about marriage and divorce that we did even 10 years ago, and this knowledge has real clinical value, particularly in educating our clients about what behavior ruins marriages and what they can expect to experience during and after a divorce. However, research can only accurately depict some of the realities of divorce; it can’t tell us what we, as individuals, value.

Some value the preservation of marriage above all. Others value individual happiness, or think in terms of promoting what they see as being best for family development. We need to continue debating these issues about competing values. But if these debates are to have any substantive meaning, they must be informed by solid research, rather than by the biased reporting of questionable “findings” serving a particular social, moral, or political agenda. By becoming informed consumers of this research, we can begin to separate solid, rational arguments from moralistic exhortation. n


Amato, Paul R. “Children of Divorce in the 1990s: An Update of the Amato and Keith (1991) Meta-Analysis.” Journal of Family Psychology 15, no. 3 (Sept. 2001): 355-70.

Gottman, John M., and Robert W. Levenson. “Marital Processes Predictive of Later Dissolution: Behavior, Physiology, and Health.” Journal of Personality and Social Psychology 63, no. 2 (Aug. 1992): 221-33.

Grych, John H., and Frank D. Fincham. “The Adjustment of Children from Divorced Families: Implications of Empirical Research for Intervention.” In Robert M. Galatzer-Levy and Louis Kraus (eds.), The Scientific Basis of Child Custody Decisions. New York: Wiley, 1999.

Hetherington, E. Mavis, Margaret Stanley-Hagan, and Edward R. Anderson. “Marital Transitions: A Child’s Perspective.” American Psychologist 44, no. 2 (Feb. 1989): 303-12.

Hetherington, E. Mavis, and John Kelly. For Better or Worse: Divorce Reconsidered. New York: W. W. Norton, 2003.

Jay Lebow

Jay Lebow, PhD, is a former contributing editor to the Psychotherapy Networker and clinical professor at Northwestern University. He’s also senior therapist and research consultant at the Family Institute at Northwestern University.