Dr. John Gottman can predict, with roughly 94% accuracy, which newly married couples will eventually divorce — based on watching how they talk to each other for as little as 15 minutes, according to the Gottman Institute’s own published research FAQ. That number comes from a series of longitudinal studies spanning more than four decades, built on earlier work like the 1992 divorce-prediction study Gottman conducted with Kim Buehlman and Lynn Katz, and later replicated by independent researchers — Rand Conger and Ron Simons’s team at Iowa State University, and Julia Babcock at the University of Houston. The predictive power doesn’t come from whether couples fight. It comes from what they do — or don’t do — about the relationship in between the fights, which is exactly the gap a specific, structured weekly practice was built to close.
Most couples only sit down and talk about the actual state of their relationship once something has already gone wrong — a missed anniversary, a resurfaced resentment, a fight that finally forces the conversation neither partner wanted to start. Therapists who work from the Gottman Method argue that timing is the core problem, not the conflict itself.
The Behaviors That Actually Predict Divorce
Gottman’s research identifies four communication patterns — criticism, contempt, defensiveness, and stonewalling — that reliably predict early divorce, with contempt standing out as the single strongest predictor. Couples showing these patterns divorced an average of 5.6 years after their wedding, per the Institute’s research summary. Crucially, the physiological arousal measured during these conflict conversations — elevated heart rate, a stress response — correlated with declining marital satisfaction over the following three years, even before either partner would describe the marriage as struggling. The problems were measurable in the body long before either person named them out loud. That’s the gap a crisis-driven approach to relationship conversations can’t close: by the time a couple feels compelled to talk, the patterns have often been compounding, unaddressed, for years.
A Weekly Structure Built Around That Gap
The Gottman Institute’s own answer to that timing problem is a practice it calls the State of the Union meeting — a weekly, structured check-in designed by Dr. Gottman specifically to surface small issues before they calcify into the criticism-contempt-defensiveness-stonewalling cycle his research flags as dangerous. The format has four parts. Partners start with appreciations: each names five specific things they valued about the other over the past week, tied to what those actions revealed about the partner’s character, not just the actions themselves. Next comes a positive reflection on what’s working in the relationship currently — a deliberate counterweight to the brain’s tendency to fixate on friction. Only then does the conversation turn to an actual issue, using what Gottman calls a “softened start-up”: a structured statement built as “I feel ___ about ___, and I need ___,” which keeps the complaint anchored to a specific situation rather than a character attack on the partner. The meeting closes with each partner naming one concrete action that would help them feel more loved in the coming week — a request, not a demand, and small enough to actually act on.
Why the Order of Operations Matters
The sequence is deliberate, not incidental. Leading with appreciation and reflection before addressing a problem primes the conversation with evidence that the relationship is fundamentally solid, which makes the issue that follows easier to raise and easier to hear — a structural defense against the contempt Gottman’s research identifies as the most corrosive of the four patterns. For the issue-discussion portion, the Institute also teaches an acronym, ATTUNE — awareness, tolerance, turning toward, understanding, non-defensive listening, empathy — as the posture each partner should hold while the other is speaking, distinct from strategies aimed at winning the disagreement.
None of this requires a relationship to be in trouble to be worth doing. That’s arguably the entire point: a weekly, low-stakes ritual with a fixed structure gives couples practice raising small friction points while the stakes are still low, instead of letting them accumulate into a single high-stakes conversation that only happens once resentment has already set in. The Gottman research suggests conflict itself isn’t the threat to a relationship — it’s inevitable, and even necessary. What the data links to divorce is conflict handled without structure, repeatedly, for years, until the physiological stress response to disagreement becomes the relationship’s default setting rather than an occasional weather event. A standing weekly meeting doesn’t eliminate disagreement. It just makes sure the relationship gets talked about on a schedule the couple controls, instead of one dictated by whatever finally breaks.
