What Is John Gottman's 'Love Lab'?
A lab apartment for observing real couples
John Gottman, psychologist and researcher at the University of Washington, spent several decades studying what sets couples who stay together and satisfied apart from those who end up separating. To do it, he didn’t settle for surveys or retrospective interviews.
He built what he himself calls the “Love Lab”: an apartment equipped like an ordinary home (kitchen, couch, TV) where volunteer couples spent a whole night going about everyday life while a team of researchers observed and recorded them.
The researchers didn’t just watch what the couples said. They also recorded their physiology, such as heart rate or sweating, while they ate, talked or discussed a sensitive topic. The underlying idea was unusual for the couples psychology of the time: instead of asking people how they think they communicate, observe how they actually communicate, with their bodies reacting in real time to what they say and hear.
Where the data come from: the published studies
Behind the lab there are specific scientific papers, and they’re worth knowing because they’re more modest, and more interesting, than the popular summary:
- Gottman and Levenson (1992) studied 73 married couples in 1983 and contacted them again in 1987. Couples in which negative interactions outweighed positive ones had poorer health, lower satisfaction and a higher risk of separation. This is where the “cascade model” first appears: low satisfaction, then thinking about separating, and finally separating, as successive steps.
- Gottman, Coan, Carrère and Swanson (1998) followed 130 newlywed couples who had passed through the apartment for six years. One of the most cited results is that the husband’s willingness to accept his wife’s influence predicted stability, while textbook “active listening,” the kind taught in many therapies, barely appeared in real arguments and predicted nothing.
- Carrère and Gottman (1999), using the same sample, looked at the first three minutes of an argument: couples who went on to divorce started them with more negative emotion and less positive emotion.
Five minutes, three studies, 91% accuracy
The finding Gottman is best known for outside academia comes from his book The Seven Principles for Making Marriage Work (1999, with Nan Silver): he says he can predict whether a couple will stay together or end up separating after listening to them in the lab for just five minutes, with an average accuracy of 91% across three studies.
This isn’t a clinical hunch: it’s the result of applying a coding system to observed interaction, looking for specific patterns (what tone is used, whether disagreement turns into contempt or stays in the realm of argument) that his team had isolated beforehand.
What that figure really means
Here’s the part popular accounts usually leave out. Those predictions were calculated on the same couples used to build the statistical model. It’s like designing an exam after knowing the answers: the fit comes out very high, but it doesn’t tell you how well it would do with new couples.
- Richard Heyman and Amy Slep (2001) analyzed exactly this problem with archival data and showed that when a divorce-prediction equation is tested on a different sample (cross-validation), accuracy drops drastically. They also pointed out that the actual divorce rate in the population has to be taken into account: if it’s low, many of the couples flagged as “at risk” end up not divorcing.
- Scott Stanley, Thomas Bradbury and Howard Markman (2000) warned of another leap: the fact that certain behaviors predict divorce doesn’t show that teaching people to change them prevents it. Between observational data and intervention there’s a bridge that has to be tested separately.
Neither critique invalidates Gottman’s work, which remains one of the richest observational foundations on couples. What they do is put the 91% in its place: a fit result within his own samples, not an ability to predict the future of any couple.
What the analysis actually looks at
The coding system doesn’t focus on the topic of the conversation, but on its form. Among other things, it analyzes:
- Whether a disagreement starts with what he calls a “harsh startup” (criticizing the other person’s character from the first sentence) or with a more careful framing of the problem.
- Whether, when one partner voices a complaint, the other responds to the facts or defensively.
- The tone, gaze and gestures accompanying each exchange, not just the words.
These seemingly tiny signals turned out to be more predictive of the outcome than the specific topic of the argument. In the book, Gottman illustrates the method with a couple he calls Dara and Oliver, observed during their stay in the apartment: what interests the team is how they handle a particular disagreement, more than the content of the disagreement.
Why this method changed the study of relationships
Before this kind of research, much of what was known about couples came from therapists describing clinical cases, not from systematic observational data on couples without diagnosed problems. The Love Lab made it possible to accumulate hours of real interaction, compare couples who stayed stable with those who separated, and statistically isolate which behaviors set them apart.
That work produced concepts that now circulate well outside academia: “love maps” (knowing the other person’s inner world, which we cover in quality time), the four horsemen, or the idea that most relationship conflicts never get fully solved.
Relationships, the immune system and health
One of the most striking findings Gottman reports in his book has nothing to do with communication, but with the body. According to his account, when his team analyzed blood samples from fifty couples who had passed through the lab, they found differences in white blood cell response between couples very satisfied with their relationship and those who were neutral or dissatisfied. In the same book he cites data showing that married people live, on average, four years longer, and attributes part of that effect to shared health habits.
Later research confirms the general direction, with a more modest size. The meta-analysis by Theodore Robles and colleagues (2014), which pooled 126 studies with more than 72,000 people, found that better relationship quality is associated with better health, with small effects (correlations between 0.07 and 0.21), similar to those of habits such as diet. The authors themselves caution that most of the designs don’t allow claims of causation.
If you’re curious how these ideas translate into the way each person gives and receives affection, you can continue with our love language test.
Sources
- Carrère, S. and Gottman, J. M. (1999). Predicting divorce among newlyweds from the first three minutes of a marital conflict discussion. Family Process, 38(3), 293-301. https://doi.org/10.1111/j.1545-5300.1999.00293.x
- Gottman, J. M., Coan, J., Carrère, S. and Swanson, C. (1998). Predicting marital happiness and stability from newlywed interactions. Journal of Marriage and the Family, 60(1), 5-22. https://doi.org/10.2307/353438
- Gottman, J. M. and Levenson, R. W. (1992). Marital processes predictive of later dissolution: Behavior, physiology, and health. Journal of Personality and Social Psychology, 63(2), 221-233. https://doi.org/10.1037/0022-3514.63.2.221
- Gottman, J. M. and Silver, N. (1999). The Seven Principles for Making Marriage Work. Crown.
- Heyman, R. E. and Slep, A. M. S. (2001). The hazards of predicting divorce without crossvalidation. Journal of Marriage and Family, 63(2), 473-479. https://doi.org/10.1111/j.1741-3737.2001.00473.x
- Robles, T. F., Slatcher, R. B., Trombello, J. M. and McGinn, M. M. (2014). Marital quality and health: A meta-analytic review. Psychological Bulletin, 140(1), 140-187. https://doi.org/10.1037/a0031859
- Stanley, S. M., Bradbury, T. N. and Markman, H. J. (2000). Structural flaws in the bridge from basic research on marriage to interventions for couples. Journal of Marriage and Family, 62(1), 256-264. https://doi.org/10.1111/j.1741-3737.2000.00256.x
Want to know more about yourself?
Discover your result with one of our related tests.
Frequently asked questions
Is it true that Gottman predicts divorce with 91% accuracy?
That's the figure he himself reports, averaged across three studies with couples observed in his lab. Those predictions were calculated on the same couples used to build the model, without testing it on new samples, and Heyman and Slep (2001) showed that in that case accuracy drops sharply once validated. It's an interesting research finding, not a guarantee that applies to any couple.
How is the Love Lab different from a simple couples survey?
It doesn't ask couples what they think they do well or poorly; it directly observes them living together for hours, including their physiology, instead of relying only on what each partner says about themselves.
Can I use this to find out if my relationship will last?
Not directly: the analysis depends on expert coding of hours of real interaction, not a self-assessment. It's more useful for understanding what the research looks at than as a personal diagnosis.
Who worked with Gottman on this research?
His closest collaborator on the physiological studies was psychologist Robert Levenson. The newlywed studies involved, among others, James Coan, Sybil Carrère and Catherine Swanson. Many of the ideas he later popularized were written with journalist Nan Silver.
