Do most cheaters cheat again?
The short answer
The strongest evidence finds that people who were unfaithful in one relationship were about three times as likely to be unfaithful in the next. That is a real and substantial increase, but it is a statement about relative risk, not about a majority. Tripling a minority risk does not produce a majority, and the research does not establish that most people who cheat go on to cheat again.
What readers think
If a partner had been unfaithful in a previous relationship, would you consider it a dealbreaker?
Anonymous. One vote per reader.
The phrase "once a cheater, always a cheater" is old enough that most people treat it as folk wisdom rather than a claim that could be checked. It has in fact been checked, in a study that followed the same people out of one relationship and into the next, and the result is more interesting than either the saying or its usual rebuttal.
What makes this question hard is not that the evidence is missing. It is that the evidence answers a slightly different question than the one people are actually asking. Someone asking whether most cheaters cheat again is usually asking about one specific person, and wants to know what the odds are for them. The research reports how much a group's risk shifts, which is a related number but not the same one, and the gap between those two things is where nearly every article on this topic goes wrong.
What the study actually followed
The most directly relevant work comes from a longitudinal sample of 484 adults who were tracked across two successive mixed-gender romantic relationships[1]. That design is the reason this study carries more weight than the surveys that usually get cited on this topic. Asking people in one relationship whether they have ever been unfaithful tells you about their history. Watching the same people move from one relationship into another tells you whether the behaviour travels with the person, which is the thing the saying actually claims.
Three findings came out of it. People who had been sexually involved outside their first relationship were around three times more likely to do the same in their second. People who knew their first partner had been unfaithful were roughly twice as likely to report the same of their next partner. And people who merely suspected their first partner had been unfaithful were about four times more likely to suspect it again.
Those odds held across gender and across marital status, and survived controls for the usual demographic risk factors. That last detail matters: it means the pattern is not simply an artifact of who tends to end up in unstable relationships.
Why three times more likely is not the same as most
This is the point on which the popular coverage of this study consistently fails, and it is worth being precise about.
A relative risk tells you how one group compares to another. An absolute risk tells you what share of a group the thing actually happens to. They are different numbers and they can point in very different directions. If a behaviour occurs in a small minority of people, tripling that minority still leaves it a minority. Three times a small number is a larger small number.
So when a headline reports that cheaters are three times more likely to cheat again and captions it "once a cheater, always a cheater", it has substituted one claim for another. The study supports the first. It does not establish the second. Answering the question as asked, whether most people who cheat go on to cheat again, the honest response is that this research does not show that, and it was not designed to.
None of which makes the finding trivial. A threefold difference in risk is large by the standards of relationship research, and if you are weighing a specific decision it is real information. It is simply information about a shifted probability, not about a certainty or a majority.
The finding about suspicion is the strangest part
The result that deserves more attention than it gets is the fourfold increase in suspecting a new partner after having suspected a previous one.
There are at least two explanations and the study cannot separate them. One is that some people are better than average at reading the signs, and their suspicion in both relationships was accurate. The other is that having been deceived once changes how someone reads ordinary ambiguity afterwards, so that the same behaviour that would previously have gone unnoticed now registers as evidence.
Both are plausible and they are not mutually exclusive. What follows practically is narrower than it looks: a strong feeling of suspicion is not self-validating, and the strength of the feeling carries no information about whether it is correct. That cuts in both directions. It is not a reason to dismiss the feeling, and it is not a reason to treat it as proof.
What predicts infidelity better than a person's history
Treating past infidelity as the main predictor also misses that it is not the strongest one available.
Work on predictors of extra-dyadic sex in heterosexual couples has found that relationship-level and situational factors do substantial explanatory work: relationship satisfaction, sexual satisfaction, and individual differences in how strongly someone responds to sexual cues or to stress all contribute[3]. Broader reviews of the literature make a similar point, that infidelity is better understood as arising from a combination of individual, relational and contextual factors than from a stable personal trait[2].
That framing matters for the question at hand. If infidelity were purely dispositional, a person's history would be close to destiny. If it is substantially relational and situational, then the same person in a different relationship, at a different stage of life, under different pressures, is genuinely facing different odds. The evidence supports the second picture more than the first, while still finding that history carries real predictive weight.
It depends, on what?
What changes the answer
Several things change this answer for a specific situation rather than a population.
Whether anything was addressed matters. The research tracks whether infidelity recurred, not whether anyone did any work in between. A person who understood what happened, and a person who has never examined it, both appear in the same column of the data. That is a limitation of what the numbers can tell you, not evidence that the work makes no difference.
The circumstances are not interchangeable either. A single incident during a period of separation and a sustained concealed relationship are recorded similarly in most research designs, and most people do not regard them as equivalent when deciding what to do next.
And the base rate you are reasoning from is probably not measured. Studies of infidelity depend on self-report about behaviour people have strong reasons to conceal, which means almost every figure in this area is likely to be an undercount by an unknown margin.
Where people disagree
Both sides of it
Reasonable people land in different places on this. Here is the strongest version of each case, not a straw man of the one we disagree with.
The case for yes
- The longitudinal evidence is real and reasonably strong. Following the same people across two relationships is a better design than the cross-sectional surveys this topic usually rests on, and a threefold difference is not a marginal effect.
- Some of the mechanisms plausibly do travel with a person rather than with a relationship. Individual differences in how strongly someone responds to sexual cues, and in how they behave under stress, are reasonably stable characteristics.
- For someone weighing a decision, an elevated risk is legitimate information even when it falls well short of a certainty. Declining to act on it because it is not a majority is its own kind of error.
The case for no
- The question asked is whether most cheaters repeat, and the research does not answer that. It reports a relative increase, which is a different quantity, and reading it as a majority is a straightforward misreading.
- Infidelity is substantially predicted by relationship and situational factors, which means the same person genuinely faces different odds in a different relationship under different conditions.
- The data cannot see whether anything changed in between. Everyone who did serious work on what happened and everyone who did none are pooled together, so the figure describes an unsorted group rather than a person's fixed prospects.
Evidence
What the research says
- Moderate evidence484 adults tracked across two relationships
Adults followed across two successive romantic relationships were roughly three times more likely to be sexually involved outside the second relationship if they had been outside the first. Knowing a first partner had been unfaithful was associated with about twice the likelihood of reporting it of a second partner, and suspecting a first partner with about four times the likelihood of suspecting again.
Knopp, K., Scott, S., Ritchie, L., Rhoades, G. K., Markman, H. J., & Stanley, S. M., Once a Cheater, Always a Cheater? Serial Infidelity Across Subsequent Relationships (2017)(opens in a new tab)What this does not show
Relies on self-reported disclosure of behaviour people have strong incentives to conceal, so the underlying rates are probably undercounts. The sample was drawn from a longitudinal study of unmarried adults in the United States and skews younger, which limits generalisation to long-established marriages and to other cultural settings. The design records whether infidelity recurred but not whether anything was addressed in between, so it cannot speak to whether intervention changes the odds. Reported as relative risk, which does not establish what share of the group repeated.
- Moderate evidence
Reviews of the infidelity literature describe it as arising from a combination of individual, relational and contextual factors rather than from a single stable disposition, and note persistent definitional inconsistency across studies about what counts as infidelity in the first place.
Fincham, F. D., & May, R. W., Infidelity in romantic relationships (2017)(opens in a new tab)What this does not show
A narrative review rather than a meta-analysis, so it summarises a literature rather than pooling effect sizes. The definitional inconsistency it identifies is itself a limitation on every prevalence figure in this area, because studies using different definitions are not measuring the same thing.
- Limited evidence
In a sample of heterosexual couples, extra-dyadic sexual involvement was predicted by relationship and sexual satisfaction alongside individual differences in sexual response and in behaviour under stress, with the pattern of predictors differing between men and women.
Mark, K. P., Janssen, E., & Milhausen, R. R., Infidelity in Heterosexual Couples: Demographic, Interpersonal, and Personality-Related Predictors of Extradyadic Sex (2011)(opens in a new tab)What this does not show
Cross-sectional, so it establishes association rather than direction: lower satisfaction may precede infidelity, follow from it, or both. The sample was a convenience sample of heterosexual couples and is not representative. Predictors of this kind explain a modest share of the variation, which means they describe tendencies across a group rather than supporting prediction about an individual.
How we rate evidence strength
- Strong evidence
- Multiple independent studies, including replications or meta-analyses, point the same way.
- Moderate evidence
- Several studies support this, but samples are limited or findings vary by population.
- Limited evidence
- Early, small, or single-study evidence. Treat this as a reasonable hypothesis, not a settled fact.
- Contested
- Credible researchers disagree, or the evidence points in conflicting directions.
Comparative view
Different perspectives
Traditions and disciplines answer this differently. We describe what each one teaches. We are not telling you which is right, and no tradition speaks with a single voice.
What the Evidence Can and Cannot Settle
Researchers in this area are generally more cautious about these numbers than the coverage of them suggests. Three constraints are worth stating plainly.
The first is measurement. Every figure here depends on people disclosing behaviour they have strong reasons to hide, and the size of the resulting undercount is unknown rather than estimated. This affects absolute rates far more than it affects comparisons between groups, which is part of why researchers tend to report the comparison.
The second is definitional. Studies disagree about whether infidelity requires sexual contact, whether it includes emotional involvement, and whether it depends on the couple's own agreement. Prevalence figures from studies using different definitions are not directly comparable, and pooling them produces a number that describes nothing in particular.
The third is the limit of what any population statistic does. A relative risk describes how two groups differ on average. It does not license a prediction about a particular person, and the confidence with which it is often applied to individuals is not supported by the design that produced it.
Therapist Perspective
Clinicians who work with couples after infidelity tend to be sceptical of the question in the form it is usually asked, because a probability is not the thing the decision actually turns on. What they generally attend to instead is whether the person who was unfaithful can give an account of what happened that goes beyond circumstance. An explanation resting entirely on external factors, that there was alcohol involved or a difficult period at work, tends to be treated as incomplete, not because those things are irrelevant but because they were equally true of many people who did not do the same thing.
The second thing commonly attended to is what happened after discovery rather than before it. Whether the full account came at once or was extracted in stages over months is often treated as more informative than the original incident, because it is a direct demonstration of how the person handles being in trouble.
A third observation from clinical practice is that the recovery is rarely linear, and that couples who are doing well often report distress returning periodically for a year or more. Reading that as evidence of failure causes a fair amount of unnecessary damage.
Not a settled view
Therapists differ substantially on whether the details of an affair should be disclosed in full. Some regard complete disclosure as a precondition for rebuilding trust; others hold that specific detail supplies images that intrude for years without aiding recovery. There is no settled professional consensus, and the disagreement is genuine rather than a matter of style.
Islamic Perspective
Sunni jurisprudence, broadly across the four schoolsWithin Islamic ethical thought, infidelity is treated as a serious wrong, and the framing differs from the secular one in a way that is relevant here. The emphasis falls less on predicting whether someone will repeat an act and more on tawba, the turning away from wrongdoing, which is generally understood to require recognising the wrong, ceasing it, sincere regret, and a settled intention not to return to it. Where a person has been wronged, scholars commonly hold that restitution or the seeking of pardon from that person is also owed, since the obligation is not only to God.
A second theme with practical bearing is the strong discouragement of tajassus, spying on others and pursuing their concealed faults, which is addressed directly in the Qur'anic prohibition in Surah al-Hujurat. This sits in some tension with the impulse to investigate a partner, and jurists have generally treated suspicion as something to be restrained rather than acted upon.
On the marriage itself, the classical schools recognise both reconciliation and separation as legitimate, and none obliges a wronged spouse to remain. Scholars have also consistently held that concealment of a person's past wrongdoing is generally preferable to its exposure, which shapes how these matters are often handled in practice.
Not a settled view
The schools differ on the procedural consequences of infidelity within marriage and on what a wronged spouse may claim, and contemporary scholars differ further on how classical rulings apply in jurisdictions where civil law governs marriage and divorce. Nothing here is a legal ruling, and a person seeking one should consult a qualified scholar familiar with their own context.
Practical steps
What you can actually do
Separate the two questions you are actually asking. What the population risk is, and what you have observed in this specific relationship, are independent sources of information, and the second one is usually more informative about your situation than the first.
If you are deciding whether to stay, the more useful question than whether it will happen again is whether the account you have been given explains what happened. An explanation that leaves you with nothing you did not already know is not much to rebuild on.
Treat a persistent feeling of suspicion as a signal that something needs resolving, without treating the intensity of the feeling as evidence about the facts. Those are genuinely separate, and the research on repeated suspicion is a good reason to hold them apart.
Resist the temptation to resolve uncertainty through surveillance of a partner's devices or accounts. It reliably produces more ambiguity rather than less, and it changes the relationship into something neither of you chose.
Give it time before treating any decision as final. Most people are poor forecasters of their own feelings in the weeks immediately following a discovery, and very few decisions here actually have to be made that week.
Worth unlearning
Common misconceptions
That the research proves "once a cheater, always a cheater". It reports that risk roughly triples, which is a statement about relative likelihood and not a claim that repetition is typical or inevitable.
That cheating is a fixed personality trait. The evidence points more towards a combination of individual, relational and situational factors, which is why the same person can face materially different odds in a different relationship.
That a strong suspicion is self-validating. People who suspected a previous partner are about four times more likely to suspect a new one, and the research cannot tell us how often that suspicion was accurate.
That published infidelity rates are precise. They rest on self-reported disclosure of concealed behaviour, and are likely to understate it by an unknown amount.
The short version
Key takeaways
People who were unfaithful in one relationship were about three times as likely to be unfaithful in the next, in a study that followed the same 484 adults across both.
ResearchA threefold increase in relative risk does not establish that most people who cheat repeat, and the study was not designed to answer that question.
By definitionSuspicion repeats even more strongly than infidelity does, and the research cannot distinguish accurate perception from changed interpretation.
ResearchRelationship satisfaction and situational factors carry substantial predictive weight, which is why history is informative without being decisive.
Research
Worth saying
When to get professional help
Couples therapy is worth considering if you have decided to stay and the conversation keeps restarting without moving, or if months have passed and one of you is still checking for reassurance many times a day. Individual therapy is often the better starting point if the suspicion is affecting your sleep, your work or your health, whether or not the suspicion turns out to be founded, because that level of distress is worth addressing on its own terms. If you are being pressured to accept an account of events that you do not believe, or being told that your memory of what happened is wrong, that is worth raising with a professional who is not also being told the same story.
Still wondering
Related questions people ask
Three times more likely, which is a comparison between two groups rather than a share of one. If the behaviour occurs in a minority of people, then tripling that risk produces a larger minority, not a majority. The study reports how much the risk shifts and does not report what proportion of people who were unfaithful went on to repeat it.
Community
What other people say
These are readers describing their own relationships, not professionals. Take them as experience, not advice, and please do not name anyone.
No one has shared their experience yet. If you have been here, you would be the first, and probably the most useful thing on this page.
Check our work
Sources
- [1]Once a Cheater, Always a Cheater? Serial Infidelity Across Subsequent Relationships(opens in a new tab)
Knopp, K., Scott, S., Ritchie, L., Rhoades, G. K., Markman, H. J., & Stanley, S. M. · Archives of Sexual Behavior, 46(8), 2301-2311 · 2017
- [2]Infidelity in romantic relationships(opens in a new tab)
Fincham, F. D., & May, R. W. · Current Opinion in Psychology, 13, 70-74 · 2017
- [3]Infidelity in Heterosexual Couples: Demographic, Interpersonal, and Personality-Related Predictors of Extradyadic Sex(opens in a new tab)
Mark, K. P., Janssen, E., & Milhausen, R. R. · Archives of Sexual Behavior, 40, 971-982 · 2011