What are the statistical odds your marriage actually lasts?
Enter your circumstances to see your estimated divorce probability, based on ACS, NCFMR, and Gottman Institute research. Most couples significantly overestimate their risk.
Querying population data…
How long do most last?
Median relationship duration before divorce or breakup.
What percentage of marriages end in divorce?
The widely-cited claim that half of all marriages end in divorce was a cross-sectional projection made in the 1980s that was never borne out by longitudinal data. Pew Research Center analysis from 2023 found that roughly one third of ever-married Americans have experienced divorce, a significantly lower figure than the 50% that entered popular culture.
The US refined divorce rate was 7.1 per 1,000 married women in 2022, down from a peak of 10.0 in 2008, according to National Center for Health Statistics data. In England and Wales, the Office for National Statistics recorded 103,816 divorce dissolutions in 2023, with a median marriage duration of 12.7 years for divorcing couples.
These population-level figures mask enormous variation by demographic factors. Age at marriage, education, prior marriages, religious participation, and courtship duration each independently shift individual probability, sometimes substantially. Our relationship longevity calculator covers typical durations for both dating relationships and marriages.
Does age at marriage affect divorce risk?
Age at marriage remains one of the most reliable predictors of divorce risk. Research from Westrick-Payne and Lin at the National Center for Family and Marriage Research (NCFMR, 2023) confirms that marriages entered at ages 15–24 carry the highest hazard. The mechanism is well understood: younger marriages often precede full identity formation, career stability, and the financial independence that supports partnership longevity.
After the mid-twenties the protective effect of age is real but diminishing, marrying at 35 rather than 30 adds little additional statistical protection. Our age gap calculator examines a related variable: whether the age difference between partners affects divorce risk.
| Age at marriage | Approximate relative risk |
|---|---|
| Under 20 | Highest hazard |
| 20–24 | Elevated |
| 25–29 | Below average |
| 30–34 | Low |
| 35–39 | Low |
| 40+ | Low to moderate |
Does the age gap between partners affect divorce?
Research by England, McClintock, and Shafer (2016) found strong evidence that partner age gaps correlate with dissolution risk. A 1-year age gap is associated with roughly 3% higher odds of dissolution compared to same-age couples. A 5-year gap raises the figure to approximately 18%. A 10-year gap reaches approximately 39%, and a 20-year gap is associated with approximately 95% higher odds of dissolution.
The mechanisms are life-stage mismatch and power dynamics. Couples with large age differences often face divergent priorities, one partner may want children while the other already has grown children, or retirement timelines may differ significantly. Power imbalances that accompany large age gaps are also associated with lower relationship satisfaction.
What are Gottman's Four Horsemen?
John Gottman and Robert Levenson's research at the University of Washington identified four communication patterns they termed the Four Horsemen of the Apocalypse, predictors of relationship dissolution with 87–94% accuracy across multiple studies.
Criticism involves attacking a partner's character rather than their behaviour, "You're so irresponsible" rather than "I was frustrated you forgot to pay that bill." Contempt is the single greatest predictor of dissolution, expressing superiority through sarcasm, eye-rolling, sneering, or mockery. Defensiveness is treating a partner's complaint as an attack and responding with cross-complaints or playing the victim. Stonewalling is shutting down, going silent, and withdrawing from the conversation entirely.
Gottman's research also identified the repair that prevents dissolution: a 5:1 ratio of positive to negative interactions during conflict discussions. Couples who maintain five positive exchanges for every negative one, even during disagreements, are significantly more stable than those whose ratio falls below this threshold.
Frequently asked questions
No. The "half of all marriages" figure was a 1980s projection based on cross-sectional divorce rate data, not longitudinal tracking of actual marriages. Pew Research 2023 analysis of ever-married Americans found roughly one third had experienced divorce. The US refined rate has fallen substantially since 2008 and current evidence does not support the 50% claim.
It depends on timing. Research by Rhoades and Stanley at the University of Denver, drawing on 1,600 Americans tracked between 2010 and 2019, found that couples who cohabited before engagement had a divorce rate of 34%, compared to 23% for those who cohabited only after becoming engaged. The "cohabitation effect" appears to be specific to pre-engagement cohabitation, possibly because some of those couples slide into marriage without a deliberate commitment decision.
Yes, particularly for women. Research by Rotz (2012) found that rising female educational attainment accounted for approximately 15% of the total decline in the US divorce rate between 1980 and 2003. Education is associated with later marriage age, greater financial independence, and improved partner selection, all of which independently reduce dissolution risk. The protective effect is strongest when both partners hold degrees.
Research by Vaaler, Ellison, and Powers (2009), published in the Journal for the Scientific Study of Religion, found that weekly Catholic attendance was associated with 52% lower likelihood of divorce, and weekly Protestant attendance with 32% lower likelihood, compared to non-attenders. The protective effect is linked to shared values, social network accountability, and access to community support structures, not to religious belief per se.
The CDC National Center for Health Statistics reported a refined divorce rate of approximately 2.3 per 1,000 total population in 2022, down from a peak of 4.0 per 1,000 in the early 1980s. The more meaningful measure is the refined rate per 1,000 married women, which stood at 7.1 in 2022, also well below its historical peak of 10.0 in 2008. Both measures reflect a genuine long-term decline in divorce rates, driven primarily by later marriage age, higher educational attainment, and shifting norms around which couples choose to marry.
Longitudinal data from the US Census ACS · NCFMR Bowling Green · Gottman Institute · CDC NCHS
The evidence supports approximately doubling of risk for marriages entered before 25 compared to those entered at 25 to 29. NCFMR research by Westrick-Payne and Lin (2023) confirms the elevated hazard, and the mechanism is well established: marriages before the mid-twenties often precede full financial independence, career establishment, and identity consolidation. A Bowling Green State University analysis found that 48% of marriages entered at ages 18 to 20 ended in divorce within 10 years, compared to 24% for marriages entered at 25 to 29. The effect is real but not deterministic: many early marriages are stable, and many late marriages end.
The relationship between children and marital stability is more complex than a simple stabilising or destabilising effect. Research by Waite and Lillard (1991) found that the presence of young children modestly reduced dissolution risk in the short term, but that the effect diminished as children aged. Pre-marital childbearing, by contrast, is associated with elevated divorce risk. A 2014 study in the Journal of Marriage and Family (Dew and Wilcox) found that couples with children reported lower marital satisfaction than childless couples on average, but higher commitment to staying together. The net effect on divorce probability is roughly neutral to modestly stabilising for jointly-conceived children within marriage.
Financial stress is one of the most consistently identified predictors of marital instability. Dew (2011, Journal of Family and Economic Issues) found that financial disagreements were stronger predictors of divorce than disagreements about children, sex, or in-laws. A 2009 study in Family Relations found that financial strain was associated with a 30% increase in divorce hazard in the following three years, and that the effect was not simply a proxy for income: the stress response to financial uncertainty, not income level alone, drove the association. Couples who disagreed frequently about money showed elevated risk even at middle-income levels.
The US has one of the higher divorce rates among developed nations but is not the highest. OECD data shows that Russia, Belgium, and Luxembourg consistently report higher crude divorce rates than the US. The UK rate is lower: the ONS recorded 7.1 divorces per 1,000 married people in England and Wales in 2022. Japan and Italy have substantially lower rates, at approximately 1.7 and 1.4 per 1,000 population respectively, reflecting cultural and legal differences in how separation is handled. International comparisons are complicated by differences in legal process: some countries where divorce is rare have high rates of informal separation that do not appear in divorce statistics.
Research supports a protective effect of longer pre-marital courtship. A 2012 study by Huston, Niehuis, and Smith in the Journal of Marriage and Family found that couples who dated for more than three years before marriage had substantially lower divorce rates in the first five years than couples who married after less than one year of dating. The proposed mechanism is partner knowledge: longer courtship provides more information about compatibility, conflict style, and life goals before irreversible commitment. The Rhoades and Stanley research also found that relationship inertia, sliding into marriage without deliberate evaluation, was associated with lower satisfaction and higher dissolution rates regardless of courtship duration.
Second marriages have a higher divorce rate than first marriages in the US. NCFMR data estimates that approximately 60% of second marriages end in divorce, compared to approximately 40 to 45% for first marriages. Third marriages show a further elevated hazard of approximately 70 to 75%. The reasons are debated: selection effects (people willing to divorce once are more willing to do so again), the complexity added by stepchildren and blended family dynamics, and potentially inadequate post-divorce reflection before re-partnering. However, remarriages that follow longer courtships and occur at older ages do not show the same elevated risk as remarriages entered quickly after dissolution.
- US Census ACS · NCFMR Bowling Green · Gottman Institute · CDC NCHS