The marriage statistics are familiar by now: divorce rates that peaked in the 1980s and have declined since; marriage rates that are at historic lows in most Western countries; average age at first marriage that has risen steadily, reaching 30 for women in the United States by 2022. The statistics are often presented as symptoms of cultural decline, as evidence that people no longer take commitment seriously, as the inevitable consequence of feminism eroding the structures that kept marriages intact.
None of these interpretations survive contact with the actual research.
What the research shows is more interesting and more hopeful: that the marriages that are happening are more stable than those of previous generations, that the characteristics of successful long-term partnerships are well-understood, that cultural assumptions about what makes marriages work are often wrong, and that the factors most associated with sustainable committed partnership have changed significantly in the last 50 years.
What the Declining Marriage Rate Actually Means
The decline in marriage rates in Western countries is not, primarily, a story about commitment phobia or cultural decay. It is, primarily, a story about delayed first marriage.
The median age at first marriage in the United States rose from 20.8 for women in 1970 to 28.6 in 2022. The equivalent rise in the UK was from 22 to 31. This reflects several converging trends: women’s significantly expanded educational and economic participation, reduced economic necessity of marriage for women’s financial security, and the availability of cohabitation as a precursor to or substitute for legal marriage.
The consequence for marriage quality is significant: marriages formed at older ages are, on average, more stable. The divorce risk for marriages contracted before age 20 is approximately 50% higher than for marriages contracted in the mid-to-late twenties. The research interpretation is consistent: people who marry later have more self-knowledge, more stable identities, more realistic expectations of partnership, and more experience to draw on in choosing partners.
The marriages that are happening are better marriages, on average, than those of previous generations. They are just fewer of them, started later, and more likely to be the second relationship architecture after a period of cohabitation.
What the Research Shows Actually Matters
Forty years of Gottman Institute research, supplemented by longitudinal studies from multiple institutions, have produced a picture of what distinguishes stable, satisfying marriages from those that fail. The findings consistently surprise people, because they are not about the things the culture focuses on.
Friendship is the foundation. The single strongest predictor of long-term marital satisfaction, across Gottman’s research, is the quality of the couple’s friendship — defined as mutual knowledge, genuine interest, and fondness. Not romantic passion. Not sexual chemistry. Not compatibility of values. Friendship, in the specific sense of knowing and liking the other person’s inner life, predicts long-term satisfaction more strongly than any other variable measured.
The conflict ratio matters more than the conflict frequency. Happy couples have as many conflicts as unhappy couples. What distinguishes them is the ratio of positive to negative interactions — what Gottman calls the “magic ratio” of 5:1. For every critical, contemptuous, defensive, or stonewalling interaction, five positive interactions are required to maintain relationship health. The implication is not that couples should avoid conflict but that they should invest in the positive interactions that create the emotional buffer around conflict.
Contempt is the specific killer. Among the “four horsemen” Gottman identified as predictors of divorce — criticism, contempt, defensiveness, and stonewalling — contempt is the most destructive and the most predictive. Contempt is treating your partner from a position of superiority: eye-rolling, mockery, sneering. Its antidote is building a culture of genuine admiration and respect, which is the positive investment that makes the relationship sustainable.
Division of domestic labour predicts female satisfaction. Research specifically on female marital satisfaction finds that perceived fairness in domestic labour division is one of the strongest predictors of women’s relationship satisfaction and sexual satisfaction. This finding has become more robust as the research has become more detailed: it is not simply about hours of housework but about the cognitive load of household management — the planning, tracking, and invisible coordination that is disproportionately borne by women. When this is distributed equitably, women report substantially higher relationship satisfaction.
What Has Changed
The marriage that research identifies as stable in 2026 looks different from the marriage that was stable in 1970, and the differences reflect genuine social change.
The importance of shared values has increased. In an era of high social polarisation — on politics, on social issues, on what kind of life is worth living — partners who share fundamental values show higher stability. This does not mean identical political views; it means agreement on the foundational questions about family, community, how to spend time and money, what matters.
The role of female economic independence has changed. In 1970, female economic dependence on marriage was a structural factor in marriage stability: women stayed in marriages partly because they had no independent economic foundation. The research now shows that women’s economic independence is associated with higher relationship quality rather than higher divorce rates — because women who can leave unsatisfying marriages also self-select into more satisfying ones.
Sexual satisfaction as a predictor has become more central. The research on sexual satisfaction in long-term relationships has become more sophisticated, and the finding is consistent: while sexual frequency naturally declines over the life of a long-term relationship, sexual satisfaction — which is determined by communication, reciprocity, and the quality of individual encounters rather than their frequency — is a significant predictor of overall relationship satisfaction for both partners.
Cross-Cultural Perspectives
The Western model of marriage — romantic love as the basis, individual choice as the mechanism, emotional fulfillment as the primary purpose — is historically recent and culturally specific.
Arranged marriages, which persist in modified forms across South Asia, parts of the Middle East, and parts of East Asia, show different patterns in research. Studies comparing marital satisfaction in arranged versus love marriages have produced complex findings: arranged marriages in communities where they are the cultural norm show comparable long-term satisfaction to love marriages, with some studies finding higher satisfaction in arranged marriages after 10 years — possibly because expectations are calibrated differently at the outset.
The finding that challenges Western romantic assumptions is this: the intense romantic love that Western culture identifies as the prerequisite for marriage is not reliably predictive of long-term relationship satisfaction. Fisher’s neuroscience research shows that the intense early-stage love response is associated with elevated dopamine, decreased serotonin (similar to OCD states), and reduced activity in regions associated with critical judgment. This is not the optimal state for assessing long-term partnership compatibility.
The cultures that build partnership gradually — that allow affection, respect, and intimacy to develop over time rather than requiring their pre-existence — may be approaching something the research vindicates, even if the mechanisms (arranged marriage) involve their own power dynamics and constraints.
What This Means Practically
The marriage that works, in 2026, based on the best available research, has several features:
It is entered later rather than earlier, with self-knowledge on both sides. It is built on genuine friendship and mutual curiosity rather than on the intensity of early passion. It maintains a high ratio of positive to negative interactions as a deliberate practice rather than a spontaneous state. It distributes domestic and cognitive labour equitably, specifically because this equity is a consistent predictor of female satisfaction. It maintains sexual communication, because assumption is the enemy of sexual satisfaction in long-term relationships. It handles conflict with tools rather than suppression or escalation.
None of this is unattainable. All of it requires intention. The culture that tells you that love should be effortless — that the right relationship just works — is, on this specific point, empirically wrong. The marriages that last and that are worth lasting are not the ones that require no work. They are the ones where the work is recognised as work, distributed fairly, and done with the fundamental goodwill that is the prerequisite for everything else.
Related reading: