John and Mary seemed the perfect pair, loving and committed, yet they ended up in divorce court. Bob and Carol's union was always volatile and rocky, but they have somehow managed to stay together. Can anyone say which marriages will survive in a time when half of all couples break up before death do them part?
Perhaps. A team of psychologists led by Dr. John Gottman at the University of Washington has recently developed a procedure that has predicted with 94 percent accuracy which couples will divorce within three years. Gottman's team interviewed 56 couples in their homes, asking 11 questions about the history of their relationship. The couples were asked to discuss how they met, courted and married, the good and bad times, how their marriages had changed and their philosophy of what made them work. The couples were also videotaped and hooked up to polygraph equipment as they talked in the laboratory, so researchers could gather physiological data. Three years later seven of the 56 couples had divorced. The researchers, as they explained in an article published in the Journal of Family Psychology, had predicted all the divorces, although three couples that they believed would divorce were still married.
The signs of trouble, it turned out, could be identified years before a couple split up. Surprisingly, the researchers found that the husband's disappointment with the marriage—not the wife's—was the single most reliable predictor of divorce. These findings, says Gottman, will be invaluable in helping marriage counselors "tell very early in a relationship if a couple needs help to head off problems."
Gottman, 50, a professor of psychology, has written 15 books and dozens of articles on marriage and divorce. He lives in Seattle with his second wife, Julie, a clinical psychologist, and their 20-month-old daughter, Moriah. He shared his findings with reporter Joan DeClaire.
At 94 percent accuracy rate is quite a feat. How did you achieve this?
We use different methods than previous studies, which relied chiefly on personality inventory questionnaires. People keep their problems to themselves when they fill out a questionnaire. Our oral history interview gets at the couple's perceptions of their relationship. We look at factors such as whether a couple express disappointment, whether they show affection toward each other, whether they have a lot of memories about their courtship and marriage. We also look at couples' behavior on videotape as they talk about how their day went as they try to resolve conflicts. While they are interacting, we take an electrocardiogram. We monitor how fast their blood is flowing. We monitor their palms and feet, where there are sweat glands that respond to psychological stimuli. We generally find that people who are going to divorce are more aroused physically than people whose marriages are stable.
So you get more objective data?
Right. It's much harder to conceal the truth from the camera. While reviewing the videotape, we can turn down the sound and look at facial expressions. Our observers are trained in a system of facial coding in which all the facial muscle groups are numbered and associated with emotions like fear and anger. A wife's facial expressions of disgust—such as raising the upper lip—are a very good predictor of whether the couple will separate in the next four years. If she displays more than one a minute, there's a good likelihood they will separate. Facial expressions are just one observation system. The other four systems measure expression in areas other than the face, such as the emotional content of words and gestures.
Are there common perceptions held by couples headed for divorce?
There is a trajectory—what I call a cascade toward marital dissolution—that has five characteristics. The first is flooding—feeling overwhelmed by a partner's negative emotions. So if you see your partner's anger, irritation and distress and say, "Hey, let me out of here," that's a very bad sign. Second, these couples see their marital problems as quite severe. Third, they don't see much advantage in trying to work them out. Fourth, they have arranged "parallel" lives—they don't eat together, they don't have friends in common. Fifth, they say they're lonely.
Are there also certain behaviors indicating divorce is likely?
There is a behavioral cascade I call the four horsemen of the apocalypse—criticism, contempt, defensiveness and withdrawal. If I say to my wife, "I am upset that you didn't balance the checkbook. Three checks have bounced, and that embarrasses me," that's a healthy way to complain. But if I say, "You keep embarrassing me. You are irresponsible and inconsiderate," that's a criticism—one of the four horsemen. Now I can escalate that to contempt by saying, "You're really a thoughtless, cruel person," which insults somebody's self-concept. Contempt can also be conveyed in facial expressions and voice tone, like repeating what the other person says mockingly or rolling your eyes. The third horseman is defensive-ness—the feeling that you are under attack, that you have to deny responsibility for the problem—which will fuel arguments. And the fourth is withdrawal. People don't give the speaker the usual signals that they are tracking what's being said. They act like a stone wall.
When these four horsemen take over the interaction in a marriage, repair mechanisms—like recognizing you're off the track and trying to talk things out—don't work anymore, and it all goes downhill.
How can troubled couples change such behavior?
If somebody is thinking "that dope, that jerk, that inconsiderate person....," it's important to stop and say, "Wait, I love this person. He really does care. Why am I putting him down this way?" You have to short-circuit contempt in your brain as well as in your behavior. It's also important for couples to know when they need to calm down. When people are physiologically aroused, hearts beat fast, blood flows quickly and adrenaline is secreted. You can't process information very well. Couples ought to master one tool—good listening, compromise, negotiating an agreement—to help them when they are really upset.
Do men and women view trouble in a marriage differently?
Women tend to see it as their responsibility to do something about it. Men tend to withdraw—they'll work harder, do things with friends instead of family. It's important for couples to understand this so they don't attribute problems to one another when the problems really have to do with gender differences. There's a danger, for example, that a woman would feel a man should be like her. And when he's not, she feels it's because he's unwilling, doesn't love her, doesn't care. And that's really a shame.
What makes a marriage stable?
We have found there are three types of stable marriage. The validating couple doesn't fight a lot, and when they have an issue they face it, negotiate and compromise. That's the ideal. But only a third of stable marriages we studied were like that. The rest engage in patterns of behavior that some psychologists would call pathological. In the conflict-avoiding marriage, the couple never fight, but it's not very intimate, and there isn't a lot of emotion. In the volatile marriage, the partners are terrible listeners—one doesn't wait for the other to get their point of view out before they say, "You think that!? Baloney!"
But what's most surprising is that in all stable marriages the ratio of positive to negative behaviors is the same—5 to 1. Volatile couples, for example, are very affectionate and quite romantic. Whereas in unstable marriages negative behaviors outweigh positive by 1.25 to 1. It's really remarkable that these numbers are so unchanging.
What advice do you hare for troubled couples? Is there a certain point at which they should seek Intervention?
My research has no shoulds. Maybe that's because I don't do therapy. But I would say that since flooding is the basic thing driving people toward divorce, one approach would be to try to understand and accept each other's style of expressing negative emotions. One might say, "I'm really overwhelmed by your anger." And the other might respond. "OK. let me tone it down." Hang in there.
Your Reaction
