Lessons from the Love Lab

Before scientists zeroed in on mental illness, we thought poor mothering was the cause of almost every disorder. We assumed that cold, distant mothers created autism, and mothers giving double messages spawned schizophrenia. But later, studies in genetics and neuroscience corrected these misconceptions. It turned out moms weren’t so bad; broken DNA and crossed wiring were more the problem.

Recently, there’ve been thousands of studies on what works and doesn’t in individual psychotherapy treatment. Now we know that for many people, cognitive therapy eases depression, behavioral desensitization cures phobias, and medication helps contain psychoses. As therapists, we look to these methods to help our clients because they’ve been scientifically validated and have proven helpful. But what methods do we use when it comes to treating couples? Let’s look at the history of our field first to understand the context in which scientific studies have been conducted and what they have taught us.

In psychology, it began in the mid-20th century. The 1950s and ’60s were revolutionary years in American intellectual history, a time of turmoil, upheaval, and immense creativity. Innovation replaced stale thinking in nearly every field, and psychotherapy was no exception. Up until then, psychoanalysis had been the gold standard for treatment. But now new visionaries like Fritz Perls, R. D. Laing, B. F. Skinner, Albert Ellis, Carl Rogers, and Victor Frankl broke free of Freudian constraints, creating new therapies that focused on the here-and-now instead of the there-and-then.

The identity of who needed treatment morphed as well. Pioneers like Virginia Satir, Murray Bowen, Paul Watzlawick, Salvador Minuchin, and Jay Haley birthed a new field—therapy for couples and families. These thinkers tried to pinpoint relationship dynamics that disrupted smooth family functioning and then designed interventions to change them. Each was brilliant. Yet none of these greats paused to scientifically study the relationships they treated. Instead, they observed as best they could, conceptualized what they saw, and plunged forward into formulating treatments. Minuchin was particularly prescient. He said, “Every marriage is a mistake. It’s how you deal with it that matters.” He presaged John’s later finding that all couples have perpetual unresolvable problems they must learn to live with. Minuchin thought misplaced boundaries within the family cause dysfunction, and therapy should focus on realigning them.

Paul Watzlawick, Gregory Bateson, and Don Jackson spotted hidden meanings and agendas lurking under the surface of family communications and figured these were the problem. They designed ways to fish these metacommunications out of the depths and into the open, where they could be more directly discussed. Ideas like these became the cornerstones of family therapy. But some contained serious flaws not spotted until later, when they were scientifically scrutinized. In their book The Mirages of Marriage, Lederer and Jackson based their couples’ treatment on a theory called “Quid Pro Quo.” They thought healthy marriages were those where one partner responds to the other’s good behavior with his or her own good behavior in turn. Marriages managed reciprocally would succeed, whereas those lacking in reciprocity would fail. Based on reciprocity theory, they proposed that the cure for ailing relationships was to help couples establish a contingency contract, an agreement in which each spouse “gives to get.”

For years no one tested this theory’s validity. Finally, a decade later Bernard Murstein put contingency contracting under the microscope. His research revealed that relationships based on reciprocity are actually ailing and failing, not healthy as Lederer and Jackson had presumed. The spouses who tally up the good they do versus the good their partners do are in fact quite unhappy. They’re like affect accountants, with thoughts like, “I did this great thing for her, but she never reciprocated.” In truth, couples in happy relationships rarely give reciprocity a second thought. Murstein found that people do good for their partners because they simply love them and want what’s best for them, not to get something in return.

When treating couples, if we embraced reciprocity theory and encouraged couples to create a contract with one another, imagine what could happen. Chances are good they’d remain miserable. Then we’d be left thinking either we were bad therapists (always a possibility) or our clients were impossible to help (less likely). The real glitch would probably be in the method itself.

In the mid-1960s, George Bach presented another treatment approach in his book The Intimate Enemy. Bach emphasized that couples should always air their resentments by “letting it rip” rather than allowing them to build up. This way they clear the air. He believed so strongly in this concept that he even encouraged partners to express themselves by hitting each other with foam rubber bats called batakas.

If we practiced his methods, a session of his might look like this:

Jill: “I hate it that you never help me with the housework!” Whap!

Jack: “Well, we never have enough sex and you’re a lousy lover!” Whap!

Not surprisingly, later studies countered Bach’s approach. It turned out that airing resentments has no cathartic effect. Venting doesn’t diminish anger; it increases it. Bach did have a point, however. We know that suppressing anger isn’t the answer either. Bottling up anger can lead to depression, withdrawal, bitterness, even a dampened immune system. But anger isn’t the enemy. Anger is hard-wired into our brains. It’s a normal and natural response when we’re treated unfairly or blocked from achieving a goal. Minus physical or verbal attacks, couples do need ways to raise and listen to each other’s complaints. Bach’s mistake was to only focus on airing resentments (plus perhaps the batakas). Research has revealed that much more is needed.

Murray Bowen embraced the opposite approach. He thought all negative emotions like anger are destructive and interfere with problem solving. Couples must remain calm, cool, and collected—in short, completely rational, for if they let loose their feelings, all hell will break loose. Then nothing can be achieved. In other words, emotion is the enemy that mucks up the marital work of problem solving. In Bowen’s world, rationality is good and emotion not so good. The therapist’s job is to make sure partners remain calm by helping them rein in their unruly emotions. Then a healthy rational dialogue can follow.

Bowen also pointed out that the healthiest couples are less “enmeshed” and more “differentiated.” The more separate and independent they are from each other, the better. Dependency, he claimed, is dysfunctional. It goes hand in hand with being too emotional. So according to Bowen, couples should work on being less dependent, simmer down their emotions, and become more rational.

Again later studies undermined these ideas. Research in neuroscience showed that emotions are central to problem solving, not an obstacle to it. In fact, millions of neurons connect the brain’s limbic system or center for emotions with the frontal cortex, or the executive problem-solving area of the brain. Kahneman and Tversky showed there are two kinds of thinking. One enables rapid intuitive thinking (they called it Type 1 thinking) and the other, a slower rational thinking (Type 2 thinking). The neurologist Antonio Damasio wrote that a patient of his had a tumor removed from his frontal cortex and was afterwards unable to prioritize his options, think creatively, or problem solve. These scientists demonstrated that without access to our emotions, our rational processes are handicapped, and we become little better than a conglomeration of arbitrary and random behaviors.

Regarding Bowen’s disdain for dependency, how could our species have survived through the mists of time without it? Biologically, we’re pack animals. For at least 250,000 years we’ve lived collectively. Communal cooperation has enabled us to survive centuries of ice, famine, and the teeth of hungry predators. Without depending on each other, our species would’ve perished long ago. With good reason we say, “No man or woman stands alone.”

Science has substantiated the value of interdependency. Medical research has demonstrated that when patients depend on others and aren’t alone, they recover faster and better from a whole host of problems, including heart attacks, strokes, and cancer. Psychological studies have also shown that people who are in committed relationships tend to live longer and are happier and more successful than those who live alone. In addition, sociologists have found that people who possess greater social capital—that is, networks of others they depend on—report greater life satisfaction than those with little social capital. There’s little doubt that interdependency has distinct life-giving benefits.

Scientific studies did well at exposing the myths about relationships. But it wasn’t until the 1970s that anyone actually observed couples’ relationships to discover the truth of why some relationships succeeded while others failed. Up until then, most studies had relied on partners filling out questionnaires or supplying self-reports. But these methods were littered with flaws and bias. There were still no valid answers to explain the course of relationships.

John had recently earned his doctorate and was teaching psychology at Indiana University, where he met a colleague who became his best friend, Robert Levenson. At the time, their respective romantic relationships weren’t going so well. For the time being, they figured it might be better to study romantic relationships rather than have them.

John hoped to understand relationships well enough to predict their future course. Why did some couples end up on the high road toward marital success? What made others fall to the doomed road below that dead-ended in divorce? He began his research in 1972. Before then only seven studies had tried to predict the future of couples’ relationships. One investigator had examined partners’ wardrobes to see if their clothing choices predicted future marital satisfaction. The study found that wives with nice wardrobes would have happier relationships, but nice clothes made no difference for husbands. Imagine a therapy based on this study: “Jane, you should go shopping. But for you, Jack, those sweats are fine.”

Clearly there was more to learn. But John had to fight an uphill battle to do observational research. His colleagues argued it was hard enough to observe reliable patterns of behavior in one person. Observing interactions in two would only square the data’s statistical unreliability, thus making the study of couples pointless. John persisted anyway.

First he observed university student couples. While they discussed a problem they were having for 15 minutes, he videotaped and coded their interactions using a system he and his graduate student Cliff Notarius invented, the Couples’ Interaction Scoring System (CISS). Later on, another student of his replicated the study with couples in rural Indiana. In both studies, the types of communication patterns observed were nearly identical.

Next John built a “talk table” with a rating dial that ranged from positive to negative. While discussing an issue, the device enabled partners to rate the intention of their own responses as well as the impact of their partner’s responses. Using statistics designed to study sequences of interaction, again John found extremely reliable patterns in couples’ thoughts and interactions.

In 1976, John and Robert Levenson teamed up. Levenson brought to the research a specialized knowledge of psychophysiology and its measurement that was added to the rating dial procedures. In the lab, couples were wired up to instruments that measured heart rate, sweat gland production, blood velocity, and overall bodily movement. These measurements were synched to the video timecode. Then the couples were asked to discuss a conflict issue for 15 minutes. For some couples, as they became upset their physiological measurements rocketed into the stratosphere. They might have looked calm on the outside but inside, their heart rates would jump to 100–150 beats per minute. Their hands would sweat, their blood would race, and their bodies would jiggle. These data indicated that they were experiencing an attack. In the moment the partner facing them resembled a saber-toothed tiger with fangs bared. In response, they shifted into diffuse physiological arousal (DPA), a state of fight-or-flight.

Afterward, the couples watched their videotapes and rated the emotional responses they recalled having during their discussion. Then they returned home. To find out how these couples fared, three years later they were contacted again and asked to return to the lab where the same procedures were repeated.

John, Levenson, and their colleagues were astounded. In other studies that tried to predict behavior, correlations between the best personality measures and behavior hovered around 0.10 to 0.30. This meant that by measuring personality at Time 1, scientists could predict Time 2 behavior with only 9 percent accuracy at best. But by measuring couples’ behavior, rating dials, and physiology at Time 1, Gottman and Levenson could accurately predict the changes in the couples’ marital satisfaction three years later with 90 percent accuracy. Numbers like that were unheard of in psychology research. It turned out that diffuse physiological arousal during conflict discussion was an especially powerful predictor of relationship demise in the future, along with several other factors.

In the same study, couples were also asked to have an events-of-the-day discussion before talking about a conflict. By analyzing the two discussions together, more patterns emerged. When husbands showed disinterest or lacked positive emotions in the events-of-the-day discussion, their wives tended to complain more harshly in the conflict discussion that followed. The quality of the couple’s friendship, especially how well the husband maintained it, predicted the quality of the couple’s ability to manage conflict.

For the next series of studies, John fine-tuned his observational coding methods and analytic processes. Based on Paul Ekman and Wallace Friesen’s Facial Affect Coding System, he created the Specific Affect Coding System (SPAFF) plus new methods of sequential and time-series analysis that enabled deeper description and understanding of change over time.

The original studies were then replicated and expanded to follow newlyweds through their transition into parenthood. Together with Levenson, John also studied the relationships of gay and lesbian couples for 12 years and older heterosexual couples through retirement for as long as 20 years, thanks to Levenson’s tenacity. In addition, John joined with his colleague, Neil Jacobson to study violent couples for nine years. Once again, most of the findings replicated.

Meanwhile, in 1986, John and I serendipitously moved from different corners of the country to Seattle. Shortly afterward, we met each other and in 1987 we married. Now we could benefit from all that research of John’s not only professionally, but personally too.

We joined forces to build an apartment lab at the University of Washington so we could observe more than 15-minute lab discussions. Newlywed couples were asked to spend 24 hours in the apartment as they might in a bed-and-breakfast, as they pleased. Then they were followed up and observed annually for the next six years.

The atmosphere in the apartment was soothing with just a few exceptions: there were three cameras bolted to the walls, a one-way window, staff personnel who observed the couple from behind the window, and other staff who took occasional blood and urine samples from each partner. Other than that, it was like a perfectly restful vacation spot.

Most couples ate, read, watched TV, talked, and slept. John’s student Janice Driver spent years working in the apartment lab, trying to ferret out what predicted sustained friendship and intimacy in these newlyweds. Driver was particularly intrigued by the minutiae of their smallest exchanges—the turn of a head, a mumbled monosyllable, a focused eye gaze. She wanted to know, when one partner made a bid for connection by calling the other person’s name or commenting on something, would that elicit the other partner’s interested response, or something else? After a decade, she and John figured out that the smallest moments told an important story. The particular ways partners responded to each other’s bids for connection forged the relative strength of their future friendship and intimacy, which in turn shaped how well they managed conflict. It was like the effect of water on rock. If waves of water emptily lapped away at the base of a rock, eventually it would weaken and collapse with erosion. But if the waves continually deposited new silt at the rock’s base, over time the rock would grow stronger and capable of withstanding the big storms ahead.

The newlyweds who remained happily married six years later turned toward each other’s bids an average of 86 percent of the time compared with 33 percent of the time for those destined to divorce. It only took a few words to make all the difference. If one partner said, “It’s tough reading all the bad news in the paper,” and the other said, “Yeah, it sure is,” that relationship was much more likely to succeed than one with no response. Silence or turning away from a bid was like a death knell for the relationship. Worst still were the marriages where the partner turned against the other’s bid with exclamations like “Be quiet! Can’t you see I’m busy?” In sum, these small moments of turning toward each other’s bids for connection were crucial for relationship happiness. Even if a hurricane assaulted a marriage, these moments built the foundation that kept it from toppling into the sea.

Taken together, the couples we studied who generously gave us their time put a crystal ball into our hands. Across all the studies, we’d learned enough to watch a couple at Time 1 and accurately predict with better than 90 percent accuracy whether they’d separate or divorce six years later. Once that news came out, people we didn’t know very well stopped inviting us for dinner.

The Four Horsemen

All in all, we’ve studied more than 3,000 couples and participated in studies of 3,500 more. Here’s a summary of everything our couples have taught us.

First, there are four big predictors of relationship demise. We call them the Four Horsemen of the Apocalypse, after the biblical harbingers of doom. The first Horseman is criticism. If partners regularly use criticism to voice their complaints where one partner blames a problem on the other partner’s character flaws, the relationship will slowly sink. Words like “You never wash the dishes!” or “You’re so selfish” only inspire resentment, not cooperation or care.

The second Horseman is contempt. This one leads couples to gallop over a cliff. Partners who are contemptuous act superior and punctuate their criticisms with a sneer, a left lip corner raise, or an eye roll that signifies their superiority and disgust. They may also mock their partner or use sarcasm, like “Aw, your pinkie hurts? Poor baby. Guess that gets you out of doing the dishes . . . again.” When partners bludgeon each other with contempt, it not only destroys relationship happiness, it also shreds the listening partner’s immune system. The number of times one partner hears the other partner’s contempt during a 15-minute conflict discussion predicts how many infectious illnesses the listener will have the following year. Contempt quickly destroys relationships.

Partners who respond to each other with the third Horseman, defensiveness, are also riding roughshod toward demise. Defensiveness goes hand in hand with criticism and contempt, since few partners can withstand being trampled without wanting to defend themselves. When being defensive, partners may either play the innocent victim, as in “I do the dishes all the time. Why are you being so mean?” or they can counterattack, as in “You’re a fine one to talk. When was the last time you paid the bills?” This is the toughest communication habit to eradicate.

The fourth Horseman we call stonewalling. Gottman and Levenson found that when partners become physiologically aroused during conflict discussions with heart rates above 100 beats per minute, they often shut down all verbal responses, divert their gaze, and turn their bodies away, thus blocking out their partners and becoming in effect a stone wall. Slamming into walls is terrible for a relationship.

These Four Horsemen bear bad tidings for a relationship. In fact, couples plagued by them divorce an average of 5.6 years after the wedding. On the other hand, we find that almost everyone uses them from time to time (including us). But the difference between happy couples who are relationship masters and unhappy couples who are relationship disasters is that master couples make repairs; disaster couples don’t. Relationship masters don’t sweep bad fights or regrettable incidents under the rug and pretend they never happened. Instead, they return to them, talk about them, and try to understand them. They listen to each other’s feelings and points of view. They figure out what they each did wrong. Then they take responsibility for what they regret saying or doing and apologize. When regrettable incidents are processed like this, they lose their destructive force, like a typhoon that’s grounded and halted by moving inland.

Our Four Horsemen aren’t the only powerful predictors we found. Positive interactions count, too, especially during conflict. When discussing a problem, if partners express understanding and empathy or smile or make a small repair or say something funny, their relationships are likely to succeed. Relationships fare much better when their ratio of positive to negative interactions during conflict is at least 5 to 1; that is, their positive interactions outnumber their negative interactions by 5 to 1. Relationships destined to fail have an average ratio of 0.8 to 1, or around one positive interaction for every negative one. During nonconflict times, the ratio that predicts success is much higher, around 20 to 1. Clearly relationships need rich bank accounts of positive interaction to survive and thrive. During conflict, husbands and wives each have important roles to play for their relationships to do well. Wives who raise their complaints gently without blame and criticism tend to have sunnier futures, while husbands who accept influence from their wives benefit, too. The Los Angeles Times coined this finding for husbands the “Yes, dear” phenomenon.

As they collected data over longer time spans, John and Levenson also found another pattern that predicted divorce. There was a group of couples who didn’t have the Four Horsemen during conflict. Instead, they showed an emotional disengagement, a lack of responsiveness, and low levels of positive emotions during conflict. Couples who were emotionally disengaged divorced an average of 16.2 years after the wedding.

Now that all the moving parts were identified, the predictors of relationship success or failure, it was time to build a theory for what it takes to make a relationship succeed. But for a theory to be valid, it has to be testable or disconfirmable. That’s the hallmark of good science. Testing theory in our field requires clinical interventions.

In 1996, we began to explore what would happen if we taught distressed couples the same skills successful couples displayed in their relationships. We started with simple steps called proximal change studies. This tested the power of one intervention to change a poor conversation to a better one the second time around.

We first tried out a couple’s intervention for physiological arousal. After one or both partners moved into fight-or-flight responses during a conflict discussion, a staff person entered the room, told the couple there was a problem with the equipment, and asked them to move temporarily to the waiting room. They were instructed to read a magazine and not talk to each other. After 20 minutes, they were summoned back to the lab and asked to continue their conflict discussion. In reality, nothing was wrong with the equipment. The break in the discussion was designed to slow their heart rates and dissipate their stress to see if a break and self-soothing could convert an escalated quarrel into a calm and constructive discussion. The intervention worked. Physiologically aroused couples with no break listened badly and ended up trampling each other with the Four Horsemen. But when another group of couples took a break, in the second discussion they spoke with more care, listened to one another, expressed empathy, and worked toward compromise. The differences between the pre- and post-break discussions were so dramatic it looked like partners had undergone a brain transplant.