attraction.lab

Do Couples Stay Together After Infidelity? The Famous 57% Is Eight Couples Out of Fourteen

The famous 57% comes from a five-year follow-up of 19 couples: eight of the 14 who disclosed the affair stayed, and one of the five who hid it.

Research-backed writing on attraction, dating and relationships. Drafted by AI against a named bibliography, published only when every claim links to its source.
28 min read
Editorial illustration for: Do Couples Stay Together After Infidelity? The Famous 57% Is Eight Couples Out of Fourteen

Before the article. If what is happening in your relationship involves fear, threats, being prevented from leaving, monitoring of your phone or money or movements, or any physical harm, the question on this page is not the one that matters, and those patterns are not something to negotiate because infidelity is also involved. Find A Helpline, in the sources at the bottom, lists free and confidential services by country. In the United States the National Domestic Violence Hotline is 1-800-799-7233, or text START to 88788, 24 hours a day. And if the weeks since finding out have left you unable to sleep, eat or think about anything else, and it is not shifting at all, that is a reason to talk to a professional rather than to read statistics.

With that said. Almost every page that answers this question answers it with a percentage, and the percentages run from 20% to 79%. This page is an audit of where each of them comes from, and the short version is that the spread is not a disagreement about the facts — it is a disagreement about who got counted.

The figure quoted most often, 57%, comes from a genuinely good study: a five-year follow-up, with assessments every six months, inside a randomised trial of 134 couples. Nineteen of those couples had experienced infidelity. Fourteen had disclosed it and eight of those were still married at five years, which is the 57%. Five had kept it secret and one of those was, which is the 20% that gets used as a warning. Nobody quoting either number says so.

The widest-circulating figure, 60-75%, turns out to be sourced — by the one page willing to say where it got it — to a self-help paperback from 2006. The largest study ever run, a 2026 survey of 3,429 people with John Gottman among its authors, found around three-quarters still together and states in its own abstract that the result applies to people who found it through affair-recovery social media. And the only dataset drawn from the general population rather than from people seeking help, 16,090 adults in the General Social Survey, found that more than half of those who reported extramarital sex separated or divorced.

So there are twelve sections. The safety case. The short answer. The nineteen couples. What was and was not randomised. The paperback. The 2026 survey and the sentence in its own abstract. The sampling pattern that runs through all of it. The one representative sample. The larger therapy study that measured something else. The trial nobody has run. Why the definition moves the number. And what is left that is actually usable. This page takes no position on whether you should stay or go. Our editorial policy explains why every empirical claim here carries a linked source.

#1

Before the numbers: when this is not a statistics question

If you are reading this a few days after finding out, the honest thing to say first is that no percentage on this page can tell you what will happen to your relationship, and nothing here is an argument for staying or for leaving. That decision is not a prediction problem, and this page does not take a side on it.

One thing has to go above the research, though, because it gets filed under the same heading and it is not the same situation. If what is happening involves being threatened, frightened, prevented from leaving, monitored, having your phone or your money or your movements controlled, or any physical harm, then the question of survival rates is not the question. Those are patterns of control, and the fact that infidelity is also in the picture does not make them something to negotiate. Find A Helpline, in the sources at the bottom, lists free and confidential services by country. In the United States the National Domestic Violence Hotline is 1-800-799-7233, or text START to 88788, 24 hours a day.

The same goes for the state a discovery can leave someone in. Intrusive images, not sleeping, not eating, and not being able to think about anything else are common in the weeks after, and none of that is a character flaw. If it is not moving at all after months, or it comes with thoughts of harming yourself, that is a reason to talk to a professional rather than to read further. A therapist can do something with it that an article cannot, and that is true whichever way the relationship goes.

With that said, what follows is an audit. The question in the title has a number attached to it everywhere it is asked, and the number turns out to be more interesting than the pages quoting it let on.

#2

The short answer: the range runs from 20% to 79%, and the reason is where people were recruited

Five sentences, each of which a later section sources and unpacks.

There is no single survival rate, and the spread is not a disagreement about facts. Published figures for the share of couples still together after infidelity run from about 20% to about 79%. They differ mainly because each study recruited a different kind of person, and in most cases recruited people who had already stayed.

The most-quoted figure, 57%, is eight couples. It comes from a five-year follow-up of 19 couples with infidelity inside a randomised trial of 134 couples. Of the 14 who disclosed the affair, eight were still married at five years. Of the five who kept it secret, one was.

The 60–75% that circulates most widely has no study behind it. The one page on the first results page that cites a source for it cites a self-help paperback.

The largest survey ever run on the question says in its own abstract that it does not generalise. Nickerson and colleagues surveyed 3,429 people in 2026 and found around three-quarters still together, and the people were recruited from social media accounts dedicated to affair recovery.

The one dataset that did not select on the outcome points the other way. In 16,090 adults from the General Social Survey, more than half of the men and women who reported extramarital sex had separated or divorced.

That last contrast is the whole subject. A help-seeking sample and a representative sample are answering two different questions, and the answers are roughly 57% and roughly 45%, in opposite directions, from the same pair of authors.

#3

Where the 57% comes from: nineteen couples, and the two groups inside them

The number traces to one paper, and it is a good paper. Rebeca Marín, Andrew Christensen and David Atkins published Infidelity and behavioral couple therapy: Relationship outcomes over 5 years following therapy in Couple and Family Psychology: Research and Practice, volume 3, issue 1, pages 1–12, in 2014.

What it is: inside a randomised clinical trial of 134 chronically and seriously distressed married couples, 19 couples had experienced infidelity. All 134 had been randomised to Integrative Behavioral Couple Therapy or Traditional Behavioral Couple Therapy, and they were assessed roughly every six months for five years after about eight months of treatment. That follow-up length is unusual and it is why this study carries the whole subject.

The abstract reports divorce rates for three groups, and it reports them in a format worth reading closely — a percentage followed by the number of couples it represents:

  • Secret infidelity: 80% divorced, n = 4.
  • Revealed infidelity: 43% divorced, n = 6.
  • No infidelity: 23% divorced, n = 26.

Do the arithmetic the pages quoting it do not. Four divorces at 80% means a group of five couples, so the 20% who stayed is one couple. Six divorces at 43% means a group of fourteen, so the 57% who stayed is eight couples. Twenty-six at 23% means about 113. Five plus fourteen is the nineteen the abstract names, and the three groups together come to 132 of the 134.

So the headline that answers this keyword across most of the first page of results — 57% of couples stay together when the affair is revealed — describes eight couples. The companion figure that gets used to warn people about secrecy describes one.

None of that makes the study bad. A five-year follow-up with assessments every six months is better designed than anything else in the area, the authors report their subgroup counts in the abstract rather than burying them, and their own conclusion is appropriately hedged: two pathways, with some infidelity couples becoming indistinguishable from couples without infidelity and others deteriorating markedly. The problem is downstream, in what happens to a percentage once it is separated from its denominator. The full text sits behind a paywall and Unpaywall lists no open copy, so the figures above are the abstract's own and the group sizes are arithmetic from them.

This is the same failure mode as the one in the research on whether living together before marriage leads to divorce, where the famous ratio survives the trip into the coverage and the conditions it was measured under do not.

#4

Nobody was assigned to tell the truth, and one page sells it as a choice

The comparison everybody draws from that study is causal: disclose and your odds improve, conceal and they collapse. Look at what was actually randomised.

The abstract is explicit. All couples were randomised to one therapy or the other regardless of infidelity status. The trial assigned the type of therapy. Whether a couple had experienced infidelity, and whether they disclosed it, were observed, not assigned. Couples sorted themselves into the secret and revealed groups by what they did, and then fourteen of them were compared with five.

That matters because the people who conceal an affair during couple therapy are plausibly different from the people who disclose it, in ways that would predict divorce on their own — how much of the relationship they had already given up on, whether the affair was still going, how much else was being withheld. Attributing the gap to the act of disclosure requires assuming those differences away, and five couples is not enough to check.

One page on the first results page goes further than the hedge allows. It lays the three groups out as a decision table headed If you choose this path…, with rows for "Secrecy / White-Knuckling" at ~80%, "Standard Discovery / No Plan" at ~45–75%, and "Honest Disclosure + Structured Recovery" at ~43%, and it closes by selling an intensive weekend. That page is better than most at the meta-level — it says plainly that there is no single true survival number, that survive can mean three different things, and it cites Marín and Atkins correctly by journal, volume and page. Then its own call to action reads: "the 43% success rate for structured recovery is a lifeline." The 43% was the divorce rate. The success rate in that row is 57%. The page flips its own number in the sentence asking for the booking.

The honest version of what that study supports is narrower and still worth knowing: among couples already in behavioural couple therapy, the ones who had concealed an affair did much worse over five years than the ones who had not, and there were five of them. The question of what happens when only one partner is willing to engage at all has its own evidence, and it is covered in what happens when a partner will not go to couples therapy.

#5

The 60–75% that half the internet repeats traces to a self-help paperback

The most widely circulated figure for this question is not the 57%. It is "60–75% of couples stay together," and it appears on the first results page three times.

One page states it as fact and attributes it to "studies showing 60-75% surviving" without naming one. A second attributes it to Divorce magazine. The third is the only one that gives a citation, and the citation is the finding of the day.

That page opens with: "Several clinical and population-based studies show that 60–75% of couples reconcile after an affair (Solomon et al., 2006)," and repeats it a few paragraphs later as a figure that "clinical help-seeking samples consistently show." Its reference list gives the source in full: Solomon, M. F., & Teagno, M. P. (2006). Rebuilding trust and repairing your relationship after an affair. New Harbinger Publications.

New Harbinger is a self-help publisher. The work is a trade paperback for couples, not a clinical or population-based study, and there are not several of them — there is one of it. It does not appear in Crossref under that title. So the number offered as a synthesis of the research literature is sourced, by the one page willing to say where it got it, to a book written for the general reader twenty years ago.

This is the pattern this site keeps running into and it is worth naming, because it is not dishonesty so much as drift: a figure gets a plausible citation, the citation gets dropped by whoever quotes the page next, and after a few hops the number is simply what everyone knows. The textbook case is the 7-38-55 split attributed to Albert Mehrabian, which does not say what it is used to say; we go through that one in what the research on body language actually predicts.

What can be said for the 60–75% range is that it is not crazy. It sits in roughly the right place for a help-seeking sample, which is what the Marín follow-up and the 2026 survey both are. But a range that happens to land near better-measured figures is still not a measurement, and nothing on the first page of results lets a reader tell the difference.

#6

The 2026 survey of 3,429 people, and the sentence in its own abstract

The largest thing ever run on this question was published in 2026, and it is the best news in the subject and the clearest illustration of the problem at the same time.

Kathleen Nickerson, Benjamin Rein, John Gottman, Rebecca Stone, Rece Davies, Mila Petrova and Ariana Johnson published Should I Stay or Should I Go After Infidelity: Pathways to Relationship Preservation After an Affair Amongst Individuals Accessing Social Media Contents on Post-Infidelity Recovery and Reconciliation in the Journal of Sex & Marital Therapy in June 2026.

The sample, from the abstract: analysable responses from 1,151 "commitment-breaking" partners — people who had an affair — and 2,278 "betrayal-experiencing" partners, for 3,429 in total. Analysis was descriptive statistics plus classification and regression trees. The strongest predictor of staying was love among those who had the affair and formalisation of the relationship — marriage, engagement, a long-term relationship — among those who had been betrayed. Past trauma, which the authors expected to matter, showed no measurable effect on whether the relationship was preserved.

The headline figures are not in the abstract, so they are quoted here from where they are attributable. A Psychology Today post citing the paper gives 75.5% of those who had been unfaithful ending the affair and reconciling, and 79% of those cheated on remaining with their partner. The first author, interviewed by PsyPost, gave 76% and 79% and added two more: 73% of those who had affairs reported feeling regret while the affair was still going on, and about 70% of straying partners and 36% of betrayed partners said the relationship was better afterwards — a gap of 34 points on the same relationship, which is the most quietly devastating number in the whole literature. The paper itself is paywalled; Taylor & Francis returns 403 to both an ordinary request and a headless browser, so nothing is quoted here from inside it.

And now the part that no page using these figures carries. The abstract's own closing sentence scopes the result: "Amongst individuals accessing social media contents about recovery and reconciliation after infidelity, relationship preservation is likely and predictable." The recruitment, also from the abstract, was "through the social media profiles of three mental health professionals enabling post-infidelity recovery and reconciliation." The first author put it even more plainly to PsyPost: the sample consists of people already highly motivated to recover, "so the high reconciliation rates found in this group do not apply to the general population."

That is a research team stating the limitation of their own headline in the abstract and again in the press interview. It is the right way to publish a number like this. It is also the sentence that falls off first.

Further illustration for: Do Couples Stay Together After Infidelity? The Famous 57% Is Eight Couples Out of Fourteen
#7

Every study behind the first page sampled people who had already stayed

Line the studies up by who was recruited and the pattern is not subtle. This is the single most useful thing to know about the subject, and no page ranking for the query says it.

Mitchell and colleagues, 2026. The paper sitting in second position for this query — open access, in the Journal of Marital and Family Therapy — interviewed 11 injured partners who chose to remain with their partner after infidelity. It is a careful qualitative study and it produces nine themes about how that decision gets made. It cannot produce a survival rate at all, because 100% of its sample stayed by definition.

Fife and colleagues, 2023. A grounded-theory study in the Journal of Social and Personal Relationships of 16 heterosexual couples who experienced sexual infidelity, chose to stay together, and self-identified as having experienced meaningful healing. Three filters, all on the outcome. (A Psychology Today write-up of it lists an ethnicity breakdown that adds to 25 couples, while the paper reports 16 and the same write-up's other two breakdowns — nine married and seven not, ten male strayers plus three female plus three both — each add to 16.)

Nickerson and colleagues, 2026. 3,429 people recruited from affair-recovery social media accounts, as the previous section sets out.

Marín, Christensen and Atkins, 2014, and Atkins and colleagues, 2005. Nineteen couples who had enrolled together in a randomised trial of couple therapy — which means both partners agreed to eight months of treatment and five years of follow-up assessments.

Atkins, Marín, Lo, Klann and Hahlweg, 2010. 145 couples reporting infidelity in community couple therapy in Germany and Austria. Still a help-seeking sample, and a much larger one.

There is a name for this. Selecting participants on the outcome you intend to measure tells you about the people who reached that outcome and nothing about how often it is reached. Every one of these studies is legitimate and most are explicit about their scope; several are the best work available on how couples recover. None of them is capable of answering the question in the title, and the first page of results treats their percentages as though they were.

The mirror image of this problem turned up when we went through whether the no-contact rule works, where the biggest site in the niche computed success rates over the subset of its clients for whom it had worked. A rate measured only on the people who got the result contains no information about the result.

#8

The one sample that did not select on the outcome, and it points the other way

There is exactly one dataset in this literature drawn from the general population rather than from people seeking help, and it is cited by one page out of ten.

Elizabeth Allen and David Atkins published The Association of Divorce and Extramarital Sex in a Representative U.S. Sample in the Journal of Family Issues, volume 33, issue 11, pages 1477–1493, in 2012. They used self-reported extramarital sex, divorce and separation from 16,090 individuals assessed between 1991 and 2008 as part of the General Social Survey.

Relative to people married and never divorced, a history of extramarital sex raised the odds of being divorced but remarried (odds ratio 2.6), divorced and not remarried (4.1), and separated (5.8). Their summary: the data "suggest that more than half of men and women who engage in EMS also separate or divorce from their spouse."

And the paper opens by naming the exact tension this article is about. In its own words: "Research based primarily from clinical or help-seeking populations suggests that most couples who have experienced infidelity do not divorce within the time frames assessed." The authors knew they were producing the other half of the picture, and they discuss their results explicitly in terms of methodological differences between studies.

The limitations are real and the page that cites it handles them well, which deserves saying. A lifetime snapshot is not a post-affair follow-up: extramarital sex "at any point" could mean an affair in a marriage that ended for unrelated reasons, or in an earlier marriage entirely, and being currently divorced does not establish which event came first. The survey measures individuals, not couples, and it asks only about extramarital sex, which leaves out everything that the broader definitions count.

So it is not the true number either. What it is, is the only figure here measured on people who were not selected for having stayed or for having sought help — and it lands below half rather than above three-quarters. Put the two side by side and you get the actual state of knowledge: among couples who go looking for help together, most are still together a few years later; among adults in the population who report an affair, most are not. Both can be true at once, and nothing in this literature tells you which group a given reader belongs to.

#9

The largest therapy sample measured something else entirely

If the help-seeking figures are the ones that apply to someone about to book an appointment, the obvious next question is what the biggest help-seeking study found. The answer is that it did not measure survival, and that is worth understanding rather than skipping.

David Atkins, Rebeca Marín, Tracy Lo, Notker Klann and Kurt Hahlweg published Outcomes of couples with infidelity in a community-based sample of couple therapy in the Journal of Family Psychology, volume 24, issue 2, pages 212–216, in 2010. It is a secondary analysis of couple therapy as actually delivered in the community in Germany and Austria: 145 couples who reported infidelity as a problem, compared with 385 couples who sought therapy for other reasons — nearly eight times the infidelity sample of the trial that produces the famous percentage.

What it reports: infidelity couples were significantly more distressed and reported more depressive symptoms at the start of therapy, but kept improving through the end of treatment and to six months afterwards. At follow-up they were not statistically distinguishable from the non-infidelity couples, replicating the earlier trial result. Sexual dissatisfaction did not depend on infidelity status. There was substantial missing data, and the authors ran sensitivity analyses suggesting the main findings were not an artefact of it.

Notice what the outcome variable is. Distress and depressive symptoms over time — not whether the couple was still together. The same is true of the 2005 paper from the same group: Atkins, Kathleen Eldridge, Donald Baucom and Christensen's Infidelity and behavioral couple therapy: Optimism in the face of betrayal, in the Journal of Consulting and Clinical Psychology, 73(1), 144–150, followed the same 19 couples to the end of treatment and found that those who had revealed the affair showed greater improvement in satisfaction than the non-infidelity couples. The authors call it exploratory in the abstract.

So the best-powered study in this area answers a different and arguably more useful question — do couples who come in after an affair get the same benefit from therapy as everyone else, and the answer is that they start worse and end up in the same place — while the question of how many stay together rests on the 19-couple cell.

This also explains an absence readers of this site may have noticed. Our page on whether couples therapy works runs through the whole outcome literature and never gives a figure for infidelity, because the trials and reviews it draws on excluded couples presenting with infidelity, along with intimate partner violence and infertility, on the grounds that they are distinct clinical populations rather than a relationship-distress sample. The number this page is about lives in the gap that exclusion creates.

#10

What has never been run, and what gets sold as though it had

At this point the question worth asking is the one that settles arguments in other parts of psychology: has anyone run a randomised controlled trial of treatment specifically for recovering from an affair?

As far as can be established, no. A PubMed search for a randomised controlled trial of couple therapy for infidelity or affair recovery returns nothing. What exists is the affair-specific protocol developed by Kristina Coop Gordon, Donald Baucom and Douglas Snyder, and it is worth stating exactly what its evidence base is, because it is cited on the first page of results as "evidence-based protocols for structured disclosure."

Gordon, Baucom and Snyder published An Integrative Intervention for Promoting Recovery from Extramarital Affairs in the Journal of Marital and Family Therapy, volume 30, issue 2, pages 213–231, in 2004. The design, in the abstract's own words, is a replicated case-study design. Six couples entered and completed treatment. The majority were less emotionally or maritally distressed at the end, and the injured partners reported greater forgiveness. The authors frame it as exploring efficacy and discuss areas for future research.

Six couples in a case-study series is a reasonable first step for a new treatment and it is how treatments get developed. It is not what the phrase "evidence-based protocol" implies to someone deciding whether to spend money on a weekend intensive, and twenty years later there is still no trial behind it.

One more absence belongs here, and the person best placed to name it names it herself. The author of the third-ranking page — who is also the first author of the 3,429-person survey — writes on that page: "We have no large, long-term studies of couples who reconcile, so I can't give you statistics about couple outcomes ten, twenty or even thirty years into the future." That is correct, it is the most important sentence on the first page of results, and it sits in section four of a page whose headline promises to tell you how long couples stay together.

#11

Even the denominator is unstable: what counts as infidelity moves the number

Everything above assumes we know which couples belong in the group "couples who experienced infidelity." That assumption does not survive contact with the measurement literature, and it is the reason prevalence figures for this subject range by a factor of three.

Adrian Blow and Kelley Hartnett published the canonical pair of reviews in the Journal of Marital and Family Therapy in 2005: part I, a methodological review, at 31(2), 183–216, and part II, a substantive review, at 31(2), 217–233. Part I exists because the measurement problem is the field's central difficulty. Definitions used in studies range from intercourse only to much broader categories of extradyadic activity, and the broader the definition, the higher the prevalence figure that comes out of it.

The spread that produces is wide. The 2026 open-access study in second position cites recent large representative US estimates of 13% to 40% of individuals engaging in infidelity, and attributes the variability directly to how infidelity is defined. For married people specifically, a Psychology Today post citing the General Social Survey data file puts it at 16% who had at least one affair. The Marín paper's own opening line says prevalence rates for American marriages range from 20% to 40%.

So the survival rate has an unstable numerator sitting on an unstable denominator. If a study counts emotional affairs it is describing a larger and differently composed group than a study counting intercourse, and there is no reason to expect the two groups to stay together at the same rate.

That definitional fight is not academic, and it has been measured on the people living it. When seven experienced couple counsellors were interviewed about where they draw the line, they disagreed with each other, and the researchers concluded that infidelity is socially constructed. We go through that study, and the attempt to measure the ambiguous end of the spectrum, in what has actually been measured about micro-cheating. The research on how people react to the varieties of it is in the psychology of jealousy, where the single most famous finding in the area turns out to depend on the answer format.

#12

What the evidence does support, and what it cannot tell you

Put all of it together and the useful output is not a percentage.

What is reasonably well supported. Staying together after infidelity is common enough that the stereotype of automatic divorce is wrong — every help-seeking sample agrees on that, across very different designs. Among couples who stayed, the Marín follow-up found marital stability and satisfaction that did not differ from couples with no infidelity, and satisfaction rising over time regardless of infidelity status; that is the most encouraging finding in the literature and it is about the couples who got there, not about your odds of getting there. Concealment went with much worse five-year outcomes in the one study that looked, in a group of five couples, for reasons that may or may not be the concealment itself. And couples who come to therapy after an affair start more distressed and end up, by follow-up, where everyone else ends up — that one has replicated across 19 couples in a trial and 145 in community practice.

What is not supported. Any specific probability for a specific couple. Any claim that disclosure causes the gap in the five-year figures. The 60–75% range as a research finding. And any version of the question that treats stayed together and recovered as the same outcome — the one page on the first results page that makes that distinction explicitly is right to, and it is the distinction the percentages all blur.

What has never been measured at all. Ten- and twenty-year outcomes for couples who reconcile. Whether an affair-specific treatment beats ordinary couple therapy. And anything about same-sex couples: not one study above reports on them, and Fife's sample and Hall and Fincham's were explicitly heterosexual.

One finding is worth closing on because it points somewhere other than the odds. Julie Hall and Frank Fincham followed 87 people who had experienced infidelity in a romantic relationship and found that forgiveness fully mediated the link between how people explained the affair and whether the relationship ended — and that those who initiated the breakup reported lower forgiveness than those whose partner initiated it. Whatever that is, it is not a survival rate. It is a description of a process that happens inside one person over months.

If you are trying to decide, the evidence here is nearly useless and the honest framing is a different one: not what are the odds but what would have to be true for this to work, and is it. That is a question about the present rather than a forecast, and a snapshot of the present is something that can actually be taken: our relationship satisfaction test scores satisfaction and commitment separately, which is useful here precisely because the two can come apart badly after an affair, and it says on its own page that it is an educational tool and not a diagnosis. The related questions this site has gone through against the primary sources are what longitudinal data says about thinking of ending it, why people stay in relationships that are not working, and, if it ends, what is actually known about how long recovery takes. The rest of this section is under Mindset.

Pulling it together

If one thing survives from all of this, it is that the question has no answer in the form it is asked. "Do couples stay together after infidelity" sounds like it has a rate attached, and the rates that exist describe particular groups of people who were recruited in particular ways: nineteen couples who both agreed to eight months of therapy, sixteen couples who said they had healed, eleven partners who had stayed, 3,429 people who follow affair-recovery accounts, and 16,090 adults answering a national survey about their lives. The first four land between 57% and 79%. The last one lands below half. Nothing about that is contradictory, and nothing about it tells an individual reader anything about their own relationship.

The second thing is the one genuinely hopeful finding, stated at its real size. Among the couples in the five-year follow-up who stayed, stability and satisfaction looked like couples with no infidelity at all, and satisfaction rose over time. That is a real result about a real, small group of people who got there. It is not a forecast.

The third is what the better pages on this subject already say and the percentages obscure: staying together and recovering are two different outcomes, and a number that counts the first tells you nothing about the second.

And the thing most worth carrying away is the question to replace the one in the title. Not what the odds are, because nobody can tell you, but whether the specific things this literature associates with couples who did well — disclosure rather than drip-fed revelation, both people actually in the room, forgiveness as a process rather than an announcement — are present or possible here. That is answerable, and it is answerable by you rather than by an average.

Frequently asked questions

What percentage of couples stay together after infidelity?

There is no single figure, and the honest answer is that the number depends almost entirely on who was recruited into the study. The most-quoted figure is 57%, from Marín, Christensen and Atkins's five-year follow-up inside a randomised trial of 134 couples: of the 14 couples who had disclosed an affair, eight were still married at five years. Of the five who had kept it secret, one was, which is where the 20% figure comes from. A 2026 survey of 3,429 people recruited through affair-recovery social media found around 76% of unfaithful partners and 79% of betrayed partners still together, and its abstract states the result applies to that population rather than to the general one. Against all of those, Allen and Atkins's analysis of 16,090 adults in the General Social Survey found that more than half of men and women who reported extramarital sex separated or divorced. Help-seeking samples cluster above half; the one representative sample lands below it.

Where does the 57% figure actually come from?

From one paper: Marín, Christensen and Atkins (2014) in Couple and Family Psychology: Research and Practice, 3(1), 1-12. It followed couples for five years after about eight months of behavioural couple therapy, with assessments roughly every six months. Nineteen of the 134 couples in the parent trial had experienced infidelity. The abstract reports divorce rates of 80% (n = 4) for secret infidelity, 43% (n = 6) for revealed infidelity and 23% (n = 26) for no infidelity — percentages paired with the number of divorces, from which the group sizes follow: five couples, fourteen couples and about 113. So 57% still together means eight couples out of fourteen. The study is well designed and reports its subgroup counts in the abstract; the problem is that no page quoting the percentage passes the denominator along.

Is it true that 60 to 75% of couples survive an affair?

That range circulates very widely and no study has been found behind it. Of the three pages on the first results page that use it, one attributes it to unnamed "studies," one attributes it to Divorce magazine, and the third — the only one that gives a reference — cites Solomon and Teagno's Rebuilding Trust and Repairing Your Relationship After an Affair, published by New Harbinger in 2006, while describing the figure as coming from "several clinical and population-based studies." New Harbinger is a self-help publisher and the work is a trade paperback, not a study. The range is not implausible for a help-seeking sample, which is roughly where the better-measured figures land, but it is not a measurement and nothing on that page lets a reader tell the difference.

Does telling the truth about an affair improve the chances of staying together?

The five-year data is consistent with that and cannot establish it, and the distinction matters. In the Marín follow-up, couples who had concealed the affair divorced at 80% against 43% for those who disclosed. But the trial randomised couples to one type of therapy or the other regardless of infidelity status — disclosure was observed, not assigned. Couples sorted themselves into those groups, and people who conceal an affair during couple therapy may differ from those who disclose in ways that independently predict divorce: whether the affair was still ongoing, how much else was being withheld, how much of the relationship they had already given up on. With five couples in the secret group there is no way to check. Pages that present the three groups as paths a couple can choose, with the percentages as the odds attached to each choice, are reading a causal claim into an observational comparison of fourteen couples against five.

Why do the studies disagree so much?

Mostly because they recruited different people, and most of them recruited people who had already stayed. The 2026 qualitative study ranking second for this question interviewed 11 injured partners who had chosen to remain — 100% of its sample stayed, by design. Fife and colleagues' 2023 grounded-theory study used 16 couples who had experienced sexual infidelity, chosen to stay together and self-identified as having healed: three filters, all on the outcome. The 2026 survey recruited through affair-recovery social media accounts. The therapy studies recruited couples who had both agreed to months of treatment. Selecting participants on the outcome you want to measure tells you a great deal about how people reach that outcome and nothing about how often it is reached. A secondary reason is definitional: prevalence estimates for infidelity range from 13% to 40% depending on whether a study counts intercourse only or a broader set of behaviours, so even the group being counted is not the same group across studies.

Is there a therapy proven to help couples recover from an affair?

No randomised controlled trial of affair-specific couple therapy appears to exist; a PubMed search for one returns nothing. The best-known affair-specific protocol is Gordon, Baucom and Snyder's integrative intervention, published in the Journal of Marital and Family Therapy in 2004. Its design is a replicated case-study series and six couples entered and completed treatment; most were less distressed at the end and the injured partners reported more forgiveness. That is a reasonable first step for a new treatment, and it is what gets described on commercial pages as an "evidence-based protocol." What has better support is ordinary couple therapy for couples who happen to have had an affair: across 19 couples in a trial and 145 in community practice in Germany and Austria, they started more distressed than other couples and were statistically indistinguishable from them by follow-up.

Do couples who stay together after infidelity end up happy?

In the one study that followed them for five years, the couples who stayed looked like couples who had never had an affair: Marín and colleagues report that infidelity couples who remained married did not differ from non-infidelity couples in marital stability or relationship satisfaction, and that couples who remained married reported satisfaction increasing over time regardless of infidelity status. That is the most encouraging result in this literature and it comes with an important caveat — it describes the couples who got there, not the chances of getting there, and the same study found that the infidelity couples who did divorce reported the highest instability of any group. The authors describe two pathways rather than one. Separately, around 70% of unfaithful partners and 36% of betrayed partners in the 2026 survey said the relationship was better after the affair, which is both encouraging and a 34-point gap between two people describing the same relationship.

How long do couples stay together after infidelity?

Nobody knows beyond five years. That is the clearest gap in the subject, and the person best placed to say so says it: the author of one of the top-ranking pages on this query, who is also first author of the largest survey on it, writes that there are no large, long-term studies of couples who reconcile and that she therefore cannot give statistics about outcomes ten, twenty or thirty years out. The longest follow-up that exists is the five-year one with 19 infidelity couples in it. Everything beyond that horizon is unmeasured.

When should someone see a professional about this rather than read about it?

Two situations are different in kind from the decision this page is about. The first is any pattern of control or danger — threats, being frightened or prevented from leaving, having your phone, money or movements monitored, or physical harm. That is not a question about survival rates and it is not changed by infidelity being part of the picture; Find A Helpline lists free confidential services by country and the National Domestic Violence Hotline in the US is 1-800-799-7233. The second is the state a discovery can leave someone in. Intrusive thoughts, not sleeping and not being able to focus are common in the first weeks. The signal worth acting on is distress that is not moving at all over months, or getting worse, or coming with thoughts of harming yourself. Neither of those waits for a decision about the relationship.

Sources

What this article draws on. For the broader bibliography see our sources page.

Keep reading