Marriage: Addition or division?

A new model for predicting divorce may help prevent split-ups

Valentine's Day is Saturday, and we are all thinking about true love and heart-shaped chocolate candy. Well, maybe not all of us. Some of us, actually, are considering the quantifiable aspects of divorce.

In America today, some 50 percent of marriages are predicted to end in divorce. And at the University of Washington in Seattle they say they can tell you exactly - well, almost exactly - which ones those will be.

A psychologist, a mathematician, and a pathologist have devised what they call a proven mathematical formula for detecting which relationships will go sour - thereby holding out hope that such couples can overcome their problems, and avoid divorce.

"We have been able to predict that divorce will happen before [it does]. That's old news," says John Gottman, emeritus professor of psychology. "But what we have now is a scientific model for understanding why we can predict it with such accuracy."

The work marks the first time a mathematic model is being used to understand such deep personal human interactions, adds James D. Murray, emeritus professor of applied mathematics. "It is totally objective. And our prediction of which couples would divorce within a four-year period was 94 percent accurate."

This is how it works: Couples face each other and discuss - each speaking in turn - a subject over which they have disagreed more than once in the past. They are wired to detect various physiological data, such as pulse rates, and they're also videotaped. A session lasts a mere 15 minutes.

The research team watches and analyzes the tapes and data, awarding plus or minus points depending on the type of interactions and according to a standard scoring system. Everything is then translated into equations and plotted on a graph, which the researchers have dubbed the "Dow-Jones Industrial Average for marital conversation."

Once this is done, different situations are simulated and analyzed from the equations and graphs, and predictions are made.

Over the past 16 years more than 700 couples (at different stages of their marriages) took part in the research.

But let's go back a moment. It all starts, say, with a chat about mothers-in-law - apparently one of the hot topics of contention among couples, along with money and sex, according to Dr. Murray. "The husband might say to his wife, 'Your mother really is a pain in the neck.' Well, that's a minus two points. A shrug, that's a no-no - so minus one. And rolled eyes - very negative; that's minus two."

If however, the husband were to say, "Your mother is a pain in the neck ... but she is sometimes funny," then, according to the researchers, you would take away two points and then give one back. If the husband cracked a smile, he would get another point.

At the end of all the additions and subtractions, a stable marriage is indicated by having five more positive points than negative ones. Otherwise, warns the team, the marriage is in trouble.

In trouble - but not doomed. The whole point of the model, says Dr. Gottman, is that it gives therapists new understanding with which they can help couples overcome patterns of interaction and prevent divorce.

"What we are suggesting," says Murray, "is that couples who take this experiment then be told the prediction and realize they are going to have to both change their behavior and repair what is wrong."

Not everyone buys into this model. Bonnie Jacobson, a clinical psychologist and adjunct professor at New York University, says it is "absolutely impossible" to understand the workings of a relationship via a one-size-fits-all model.

"For mostly every couple I have seen, it's hard to see how they got together in the first place," she says. "So unless you really get to know the nuanced dynamics, you will never 'get it' or be able to help."

Christine Fasano was married for only 14 months before getting a divorce last year. She agrees the dynamics of a relationship are nuanced and complex - but also sees merit in the University of Washington study's basic assumption that if one looks starkly at interaction between a couple, it is possible to ascertain whether the relationship is headed toward demise.

"I'm not surprised the model works," she says. "It's actually not that profound. My basic observation of couples that are happily married is that they treat each other well. That is basically what they are saying, and that is hard to argue with."

So, any final advice for Valentine's Day from the divorce research team out in Washington?

"I would never give advice on matters of the heart," says Murray the mathematician, who, incidentally, has been married 45 years. "But I suppose the bottom line is, yes, communication. And being good to one another. That is nice to quantify."