Staying Friends With an Ex: What the Research Measured, and Which Reasons Went Badly
The pages that rank cite the right studies and not one reports a number. In the study that separated them, staying connected and watching ran opposite ways.
Before the article. The language of staying friends is also the language used to keep someone reachable who has asked to be left alone, so the first section deals with that rather than a footnote at the bottom. If someone is monitoring your location or your phone, turning up where you did not say you would be, using a child or a shared account to keep contact open, threatening to share images, or alternating affection with punishment until you manage your own behaviour around the reaction, that is coercive control and nothing on this page addresses it. In the United States the National Domestic Violence Hotline is 1-800-799-7233, or text START to 88788, 24 hours a day. Find A Helpline, in the sources at the bottom, lists free and confidential services by country.
With that said. Every page on the first screen of results for this question answers the same question — should you? — and the research stopped asking it years ago, because nobody assigns people to stay friends and any comparison between those who do and those who do not is a comparison of everything else about them.
What has been measured instead is the motive. Griffith, Gillath, Zhao and Martinez built and validated a scale for it in 2017 and found four reasons: security, practical, civility, and unresolved romantic desires. Unresolved romantic desire was the one associated with negative outcomes. Security and practical were associated with better ones. And practical or civility made the friendship less likely to last — which means the version most often recommended as the mature option is the one that mostly dissolves anyway.
The second finding is the one the genre collapses. Tara Marshall studied 464 people and separated two things that normally get discussed as one: remaining connected to an ex, and going and looking at their page. They came apart in opposite directions. Surveillance went with more distress, more longing and more sexual desire. Remaining connected went with less of all three. Both went with lower personal growth, measured on a standard inventory.
Fourteen sections follow: the safety case and the real base rate for post-breakup pursuit, what the ranking pages actually cite, the four reasons and how each one went, the second instrument and why its famous claim cannot be checked, the study that compared an ex-friendship against an ordinary friendship inside the same people, the contact-versus-surveillance split, the growth each costs, why avoiding distress has a measured price, where watching stops being watching, what people actually do and in what proportions, what the "it will stop you moving on" study really measured, who ends up friends, and what the evidence will support you doing. Our editorial policy explains why every empirical claim here carries a linked source — and why, where the source turned out to be unreadable, this page says so instead of borrowing a number from somewhere else.
Before anything else: when the contact is not a friendship
This page is written for the ordinary version of the question, where two people who used to be together are trying to work out what they are to each other now. There is a version that is not that, and it has to go at the top rather than in a footnote, because the vocabulary of staying friends is exactly the vocabulary that gets used to keep someone reachable who has said they want to be left alone.
The measured version of the ordinary end of this is worth knowing before the alarming version. Olivia De Smet and colleagues surveyed 631 adult ex-partners about unwanted pursuit behaviour after a separation — the repeated, unwanted pursuit of contact or intimacy, which is the common and usually milder end of what stalking laws describe. Ex-partners reported on average five to six such behaviours after the breakup. Men and women reported an equal number, and so did same-gender and opposite-gender ex-partners. What predicted the count was not gender: it was who ended it (the pursuer was more often the one who was left), rumination and cognitive preoccupation with the ex-partner, anxious attachment in the former relationship, borderline traits, and past delinquent behaviour. Rumination was a stronger predictor in women than in men.
Read that number the way the authors frame it. Five or six is not a portrait of a stalker; it is a portrait of how ordinary low-level pursuit is in the weeks after a breakup, and that is useful if you have been treating one or two awkward messages of your own as evidence of something pathological.
The line that matters is somewhere else entirely. If someone is monitoring your location or your phone, showing up where you did not tell them you would be, using a child, a shared lease or a shared account to keep contact going, threatening to release images, or alternating between affection and punishment until you manage your own behaviour to avoid the reaction, that is not an unresolved friendship. It is coercive control, and none of the research below addresses it. In the United States the National Domestic Violence Hotline is 1-800-799-7233, or text START to 88788, 24 hours a day. Find A Helpline, in the sources at the bottom of this page, lists free and confidential services by country. If what kept you in the relationship is what is now keeping you in contact, what the research on trauma bonding does and does not support is the more relevant page.
With that said, the rest of this is about the ordinary case, and the ordinary case has been measured better than the search results suggest.
What the pages that rank actually cite
Measured on 2 November 2026. Google, DuckDuckGo, Brave and Mojeek each returned a bot challenge from this connection, so the first page was read from a search API rather than from a rendered results page, and no challenge was solved. The ranking set it returned matches the one recorded here on 30 October: TIME, two Psychology Today posts, NPR, Forbes, Vice, Verywell Mind, Healthline and a handful of therapy practices.
The useful question about a results page is not who is on it. It is what they cite, and that is only visible if you open them.
TIME (10 July 2018, updated 14 February 2024) does the best job on the page, and it is worth saying so. It links five real papers: Schneider and Kenny on ex-partner friendships, Spielmann and colleagues on longing for an ex, Marshall and colleagues on growth after a breakup, Mogilski and Welling on motivations, and Griffith and colleagues on reasons and outcomes. Those are, more or less, the correct five. It names no author, no year, no journal, no sample size and no effect size for any of them. A reader who wants to know how many people were studied, or how large any of these effects were, has nowhere to go except the links.
Psychology Today, "Should Exes Remain Friends?" (March 2023) cites no empirical study at all. No researcher, no year, no number.
Psychology Today, "Staying Friends With an Ex: Smart or Risky?" (25 November 2025) does better and cites Griffith and colleagues in full, with authors, year, journal, volume and page range — and reports no sample size, no percentage and no effect size from it. It also leans on a 1978 paper in the Journal of Divorce, and repeats the unsourced line about loneliness being as harmful as smoking fifteen cigarettes a day.
Forbes (July 2024) and NPR (31 May 2024) could not be retrieved from here — a 403 and a timeout — so this page makes no claim about what they cite.
That is the shape of the gap, and it is the same shape found on this site's last three subjects: the ranking pages are not ignorant of the literature. They point at it and then decline to open it. Nobody on the first page tells you how many people any of this was measured in. The rest of this article is mostly those numbers.
One honest limit before they start. Two of the five papers the genre leans on are behind paywalls with no open copy anywhere, and that is reported in place rather than papered over. Our editorial policy explains why every empirical claim here carries a linked source, and that promise cuts both ways: where the source cannot be read, this page says so instead of quoting a number it found on a blog.
The research question is not whether, but why
Search for this subject and every page asks the same question: should you? The research stopped asking that some time ago, for a reason that is easy to miss and changes everything downstream.
"Should you stay friends with an ex" has no measurable answer, because the people who stay friends and the people who do not are not two conditions anyone assigned. They differ beforehand — in how the relationship went, in who ended it, in what they want now — and any comparison between them is a comparison of those differences. What is measurable is the motive, because motives vary within the group that stayed friends, and you can ask people about theirs.
So that is what Rebecca Griffith, Omri Gillath, Xian Zhao and Richard Martinez did in 2017, in Personal Relationships. Across two studies they identified what predicts staying friends with an ex-partner, built and validated a measure of the reasons people give, and then looked at how those friendships went. The measure exists as an instrument in its own right: the Reasons for Staying Friends with an Ex-Partner Scale, deposited in APA's PsycTESTS in June 2018 and linked in the sources below.
Their analysis found four reasons, and the four are worth holding separately because the rest of the article turns on them:
- Security — the ex-partner is a source of emotional support, advice, trust.
- Practical — the friendship keeps something concrete attached to it.
- Civility — politeness, avoiding a confrontation, not wanting to hurt them.
- Unresolved romantic desires — the friendship is the form the wanting has taken.
The item wording that defines each one sits behind the journal's paywall, so this page describes the factors by their published names and does not invent examples for them. What the four are for is the next section, and it is the one result in this subject that answers the question people are actually asking.
It is worth noticing what has happened to the question in the move. "Should you?" asks for a rule. The research hands back something more awkward and more useful: the same arrangement, staying in touch with someone you used to love, goes differently depending on what it is doing for you. That is the same reframe that runs through the research on why people stay in relationships that are not making them happy, where the reasons for staying turn out not to be the mirror image of the reasons for leaving.
Which reasons went badly, and which friendships quietly did not last
Here is the published result, and it is the most useful paragraph on this subject anywhere.
Staying friends out of unresolved romantic desires was associated with negative outcomes. Staying friends for security or practical reasons was associated with more positive outcomes. And — this is the part nobody quotes — staying friends for practical or civility reasons made the friendship less likely to last in the long term.
Read those two sentences together, because they are not saying the same thing and the difference is the whole lesson. Practical appears in both, pointing opposite ways: it went well for the person and badly for the friendship's survival. A friendship held together by a lease, a dog, a shared department or a mutual friend group does its job and then expires when the thing holding it stops existing. That is not a failure. It is what that kind of friendship is.
Civility is the one most people are actually in without admitting it. Staying friends because ending it properly would be a scene, because they would be hurt, because it is easier to be vague — that was the reason associated with the friendship not lasting. Which is to say: the version most likely to be sold as the mature option is the version that mostly dissolves anyway, just slowly and with more ambiguity along the way. The site has a page on what sporadic, non-committal contact does when it keeps a connection alive without advancing it; the overlap is not accidental.
And unresolved romantic desire is the one worth being honest with yourself about, because it is the only one the study put on the negative side on its own. The friendship is not the problem there. The friendship is the shape the hope is taking, and a shape that keeps hope alive is not a neutral arrangement.
Now the limit, stated in the same type size. The paper is paywalled and no open copy exists in any repository. What is reported above is the direction of each finding, taken from the published abstract, which is all the publisher makes readable. The sample sizes, the effect sizes and the confidence intervals are behind the paywall, so this page cannot tell you how large any of it is — and neither could TIME or Psychology Today, which is presumably why neither of them tried. Direction without magnitude is weaker evidence than it looks. Treat these four as a vocabulary that has been validated, and a set of signs that have been measured, rather than as a quantity.
The other instrument, and the claim you cannot check
There is a second paper in this area, and almost every popular article about it has read only the headline.
Justin Mogilski and Lisa Welling published "Staying friends with an ex: sex and dark personality traits predict motivations for post-relationship friendship" in Personality and Individual Differences in 2017 (volume 115, pages 114–119). It is the source of the line that circulates every few years in the form staying friends with your ex means you might be a psychopath — a reading the title does not quite support even on its own terms, since predicting a motivation is not the same as predicting the friendship.
Here is what can be verified about it, and it is less than it should be. The citation is real and checks out in Crossref: those authors, that journal, that volume, those pages. Beyond that, the paper is closed. Unpaywall reports no open-access location and no repository copy. OpenAlex, Semantic Scholar and Crossref all return the record with the abstract withheld by the publisher. The usual mirrors either require an account or returned a bot challenge.
So this page is not going to report its numbers. The secondary accounts that do report them do not agree with each other — one says the authors polled 860 people, another describes 348 people listing reasons, another a second study of 513 — and when three summaries of one paper give three sample sizes, the honest move is to say that nobody writing about it from outside a university library knows which is right. That includes us.
What is worth taking from its existence is a structural point rather than a finding. There are two instruments in this literature and they do not agree on how many reasons there are. Griffith and colleagues' analysis landed on four factors. Mogilski and Welling's, by every account of it, landed on seven. That is not a scandal; factor solutions depend on the items you start with and the sample you run them in, and both teams were doing something reasonable. It is a reason to hold the four categories loosely. A list of reasons recovered from a questionnaire is a description of how a particular group of people talk about their lives, not a discovery about the structure of the world — the same caution this site applies to what a personality or attachment score can and cannot tell you.
A friendship with an ex is not the same object as a friendship
Before the question of how it goes, there is a prior question that one study settled unusually cleanly, and it did so with a design worth describing because it dodges the problem that ruins most of this literature.
Carl Schneider and David Kenny, in the Journal of Social and Personal Relationships in 2000, did not compare people who stayed friends with people who did not. They asked each participant about two friendships at once: a cross-sex friend who had once been a romantic partner, and a cross-sex friend with whom there had never been any romantic involvement. Both ratings come from the same person, on the same day, using the same scales — rules of friendship, costs and benefits, romantic desires, overall quality. Whatever is idiosyncratic about how that person rates friendships cancels out, because it is in both ratings.
The comparison came out one-sided. More positive qualities were associated with the never-romantic friend. More negative qualities, and more romantic desire, were associated with the friend who used to be a partner. The authors' conclusion is the sentence to keep: the relationship between ex-romantic partners is qualitatively different from a platonic cross-sex friendship.
That is not a verdict on whether to have one. It is a correction to the framing that makes the decision confusing in the first place. People ask whether they can "just be friends" as though friendship were a known destination they might or might not reach. In the one study that put the two side by side inside the same people, it was not the same destination. It carries more cost, more friction and more residual wanting than the friendship it is being compared to, on average — and knowing that in advance is more useful than being told it is possible.
One caveat the design cannot escape: this is cross-sectional and self-reported, published in 2000, and like Griffith's paper it sits behind a paywall, so the sample size and the size of each difference are not readable from outside. The direction is clear. The magnitude is not stated here because it cannot be checked.
The split that changes the question: staying connected, and keeping watch
This is the study the subject deserves and the one almost nobody cites properly, including the pages that link to it.
Tara Marshall, in Cyberpsychology, Behavior, and Social Networking in 2012, looked at 464 people after a breakup and did something the rest of the literature mostly does not: she separated two things that get discussed as one. Remaining Facebook friends with the ex-partner is one variable. Surveillance — going and looking at their page — is a different one. People do both, either, or neither.
The two came apart, and they came apart in opposite directions.
Surveillance was associated with greater current distress over the breakup, more negative feelings, more sexual desire for the ex-partner, more longing for them, and lower personal growth.
Remaining Facebook friends, compared with not remaining friends, was associated with less negative feeling, less sexual desire and less longing for the former partner — and also with lower personal growth.
All of this held after controlling for offline contact, personality traits, and the characteristics of the former relationship and the breakup that are known to predict how people adjust. That control set matters: the obvious objection to any result like this is that people who were fine anyway are the ones who stay connected, and the analysis was built to take some of that out.
Notice what this does to the question. The genre argues about whether to stay friends. The one study that pulled the two behaviours apart found that the connection was not where the distress lived. The watching was. Someone who deletes the ex and then checks their page from a logged-out browser three times a week has chosen the arrangement that in these data looks worse, not better, and they have done it believing they made the clean choice.
The mechanism is not mysterious once you say it plainly. A friendship is a channel with a person on the other end of it who knows you are there. Surveillance is a channel with nobody on the other end: all the exposure, none of the conversation, and no limit on how often you can use it. The same asymmetry runs through what monitoring a current partner has been measured to do, where watching reliably produces more material to watch.
Two honest limits. This is correlational, so the arrow could run the other way: people who are already in more pain may be the ones who look more often. And it is Facebook in 2012, which is not the set of channels anyone uses now. What there is no reason to think changed is the distinction itself.
Both groups paid the same price, and the price was growth
Go back to the paragraph above and read the one result the two groups shared, because it is the one that gets left out of every summary.
People who remained Facebook friends reported lower personal growth. People who conducted surveillance reported lower personal growth. The two behaviours disagreed about distress, desire and longing; on growth they agreed.
The author's own framing of the whole paper is that exposure to an ex-partner through Facebook may obstruct the process of healing and moving on. That is the sentence TIME is gesturing at when it says staying friends can hold you back, and on this paper the gesture is fair — but it is worth being precise about what it costs you, because the paper is precise.
Staying connected, in these data, is associated with feeling better and growing less. Those are not the same axis, and the genre treats them as though they were. The advice to cut contact is usually sold as the thing that will make you feel better faster. In the only study that measured both, cutting contact was associated with more negative feeling, more longing and more sexual desire — and more growth.
Which raises the obvious question: growth in what sense, and is it worth having? The measure is not vague. In the companion paper described in the next section, personal growth is the 21-item Posttraumatic Growth Inventory, rewritten to refer to the breakup — a standard instrument that asks about changed priorities, new possibilities, a different sense of one's own strength, and closer relationships with other people. It is not a euphemism for feeling fine.
So the trade is real and it is a trade. A connection that keeps the person reachable appears to soften the loss and slow the reorganisation. Whether that is the right trade is not something a study can tell you, and this page is not going to pretend otherwise. What it can tell you is that there is a trade, which is more than the confident version of this advice admits in either direction. The marker of having finished the reorganisation is set out in what the research actually measures when it measures having moved on, and it is not the absence of feeling.
Which is only bad news if distress is the enemy
The assumption underneath almost every page on this subject is that the goal is to hurt less, and that anything which reduces the hurting is therefore progress. One study took that apart with enough numbers to be worth reading slowly, and it is open access, so unlike most of what is cited here you can check every figure yourself.
Tara Marshall, Kathrine Bejanyan and Nelli Ferenczi, in PLoS ONE in 2013, ran two online studies of people recalling their most distressing breakup. Study 1 recruited 411 people and modelled 364 of them (236 women, 126 men, 2 unspecified). Study 2 had 465 after removing eight duplicate IP addresses and modelled 370 (308 women, 62 men). They measured breakup distress with the 16-item Breakup Distress Scale, rumination split into brooding and reflection, a tendency to rebound — a construct with a small and counterintuitive literature of its own, set out in whether rebound relationships work — and personal growth with the 21-item Posttraumatic Growth Inventory. Indirect effects were tested with 1,000 bootstrap samples.
The result that matters here is the one about avoidance.
Attachment avoidance predicted less distress — β = −.19 in Study 1, β = −.14 in Study 2 — and it predicted less growth: β = −.30. Attachment anxiety ran the other way on both: more distress (β = .44 and β = .42) and, with avoidance in the model, slightly more growth (β = .12).
And the route from distress to growth was not direct. It ran through rumination, and through both halves of it: distress to growth via brooding, β = .26 (CI .17 to .35), and via reflection, β = .14 (CI .08 to .22). Even brooding — the half of rumination that is supposed to be the bad one — carried a positive indirect path to growth in these data.
Put beside Marshall's Facebook paper, the two assemble into one picture. The thing that reliably reduces how much a breakup hurts is also the thing that reliably reduces what comes out the other side, whether the mechanism is a dispositional pull away from the feeling or a channel that keeps the person half-present. Feeling less and processing less travel together.
The limits here are real and the authors state them. Both studies are cross-sectional and retrospective — people remembering a breakup, not followed through one — so the arrows are modelled, not observed. Both samples are self-selected and online, and Study 2 is 83% women. A β of .26 is an indirect effect in a structural model, not a promise about your week. And none of this says that more pain is better, which would be a stupid reading of it. It says that avoiding the feeling has a measured cost, and that the cost shows up in a standard instrument rather than in a metaphor.
Where watching stops being watching
The surveillance finding needs a boundary drawn around it, because there is a point on that continuum where it is no longer a question about your own recovery.
Amy Lyndon, Jennifer Bonds-Raacke and Alyssa Cratty surveyed 411 valid participants about how they communicated with ex-partners on Facebook. The responses resolved into three factors: covert provocation, public harassment and venting. Each of the three was related to perpetrating cyber obsessional pursuit and obsessive relational intrusion — the research categories that sit underneath stalking and cyberstalking statutes. Participants who engaged in cyber obsessional pursuit were almost six times more likely to also perpetrate the offline kind. The authors' summary of their own result is the line to keep: there is a kernel of truth to the popular term Facebook stalking.
The practical content of that is a distinction, not an accusation. Looking at someone's page is passive and the earlier section is about what it does to you. Covert provocation — posting things calculated to be seen and read a particular way by one person, keeping a shared audience in play, using mutual friends as instruments — is not passive. It is contact, conducted at a distance in a way that cannot be declined, and in these data it travelled with behaviour the law has names for.
It is worth saying which direction the risk runs, because the genre almost never does. The person most likely to be hurt by this is the ex-partner, not the person doing it. If the honest description of your last month includes composing posts with one reader in mind, this is the paragraph that is for you, and the correct next step is not a better strategy for being seen. It is removing the channel.
And if you are on the receiving end of it, the first section of this page applies, including the helplines. The ordinary version of post-breakup pursuit is common — five to six behaviours on average in the De Smet sample — and being on the receiving end of a common thing does not mean you have to absorb it. The more general pattern of being kept half-present in someone's life without a place in it is covered in what the research on being pocketed and on concealed relationships shows.
Four ways people actually handle it, and how common each is
Everything above is about associations. This one is about base rates, which are rarer and more grounding, and it is the only study here that asks what people actually do rather than what it correlates with.
Brandon McDaniel, Michelle Drouin, Jayson Dibble, Adam Galovan and Madison Merritt surveyed 256 people who had been through a breakup within the previous year, asking about monitoring, interacting, deleting posts and photos, deleting the former partner, deleting their family and friends, stopping social media use altogether, and keeping digital possessions. About 38% reported distress over the breakup sometimes or more often. A latent profile analysis found four groups:
- Clean breakers — 61.3%. Zero to very little monitoring, interacting or deleting. Unlikely to delete the ex-partner, unlikely to stop using social media, unlikely to keep digital possessions.
- Ritual cleansers — 15.6%. Very little monitoring or interacting, but deleting: their own history, the ex's family and friends, the ex.
- Wistful reminiscers — 12.9%. Very little deleting, but frequent monitoring of the ex-partner and frequent interaction with them and with their family and friends.
- Impulsives — 10.2%. High amounts of all of it.
On distress, the ordering was: impulsives highest, then wistful reminiscers, then ritual cleansers, then clean breakers.
The first number is the one to sit with. Nearly two thirds of people did essentially nothing. They did not stage a purge and they did not keep watch; they left it alone and it receded. That is the modal experience of a breakup in this sample, and it is invisible in a search results page, because nobody writes an article for the person who is not doing anything.
The second thing worth noticing is that the two middle groups are the two strategies the internet argues about — delete everything, or stay in touch and keep looking — and both sat above the people who did neither. Which is a useful corrective to the idea that there is a technique here waiting to be adopted.
And the limit, which is the same as everywhere else on this page: this is one cross-sectional survey of 256 people, and nothing in it establishes direction. Impulsives may be distressed because they are doing all of it, or doing all of it because they are distressed, and a profile analysis cannot separate those. The percentages are a map of what people do. They are not a ranking of what works.
"It will stop you moving on" — what that study actually measured
The single most repeated claim about staying friends with an ex is that it will hold back your next relationship, and it is usually supported with a link to one paper. The paper is real, it is good, and it did not measure that.
Stephanie Spielmann, Samantha Joel, Geoff MacDonald and Aleksandr Kogan published "Ex Appeal" in Social Psychological and Personality Science in 2013. They followed people who were currently in relationships at three points over six months, measuring current relationship quality, emotional attachment to ex-partners, and how good the available alternatives looked.
The variable is longing: emotional attachment to a former partner. Not friendship. Not contact. Not whether anyone was still speaking to anyone. A person who has not seen their ex in four years can score high on it; a person who has coffee with theirs every fortnight can score low.
What they found runs mostly in the direction nobody quotes. Declines in current relationship quality predicted increased longing for ex-partners over time — above and beyond attention to relationship alternatives in general, which is the control that makes the result interesting. The reverse path existed too, but it was conditional: increased longing predicted decreased relationship quality only among people considering recent ex-partners.
So the headline the genre takes from this paper is at best half of it, and the weaker half. The dominant arrow goes from the current relationship souring to the ex becoming appealing again. The authors' own reading is that ex-partners get used in a substitution process, propping up belongingness when a relationship starts to fail.
That is a more useful thing to know than the version it gets flattened into. If an old relationship has started to look better lately, the measured bet is that something changed in the present one, not that the past one has reasserted its claim. The same inversion appears in what the research says about why people stay in relationships that are not making them happy — what looks like attachment to an option is often information about the alternatives, and the relationship satisfaction test scores alternatives and investment separately from satisfaction for exactly that reason.
The part of the claim that does have support behind it is narrower and sits in the previous sections: staying connected was associated with lower personal growth, and unresolved romantic desire as a reason for the friendship was associated with negative outcomes. Neither of those is "being friends with your ex will ruin your next relationship."
Who ends up friends in the first place
One more piece, because it answers a question people arrive with and do not ask out loud: whether wanting to stay friends says something about them.
Melinda Bullock, Jana Hackathorn, Eddie Clark and Brent Mattingly, in the Journal of Social Psychology in 2011, asked whether satisfaction during the relationship that ended predicts friendship after it. It did. The more satisfied people had been during the dissolved romance, the more likely they were to remain friends afterwards, and the more likely they were to engage in friendship maintenance behaviour — the ordinary upkeep of a friendship, checking in, making plans, keeping it going.
And the mediation is the interesting part: friendship maintenance fully mediated the association between past romantic satisfaction and current friendship satisfaction. The happy relationship did not leave behind a good friendship by itself. It left behind two people inclined to do the work, and the work is what produced the friendship.
That is worth putting next to Griffith's list of predictors, which is broader: sex, sexual orientation, attachment styles, personality traits, time since the breakup, and the reasons for the breakup all predicted whether people stayed friends.
Three things follow, and they are the closest this subject comes to practical advice that is actually derived from evidence rather than asserted over it.
Staying friends is not a neutral default that happens when nobody objects. It is maintained, and when nobody maintains it, it stops. Which reframes the civility finding from earlier: a friendship kept out of politeness is one nobody is doing upkeep on, and that is a decent guess at why it was the one that did not last.
Attachment is in the predictor list twice. It predicts whether people stay friends at all in Griffith's data, and it predicts post-breakup pursuit in De Smet's, where anxious attachment in the former relationship was among the things that explained the count. If you want to know which way you are likely to lean on this, the attachment style test scores the two dimensions in about two minutes and says plainly on its own page what a score like that cannot tell you — and whether those dimensions move over a lifetime has been measured, slowly and in years rather than months.
Time since the breakup is a predictor, not a prescription. It predicted staying friends. It does not follow that there is a correct interval to wait, and nobody has measured one. Any specific number you have been given for how long to go without contact was made up, in the same way the three-day rule and the 3-6-9 rule were made up, and this site has the receipts on both. The thirty days in particular has now been traced: what has actually been measured about cutting contact with an ex found no study that assigned anyone a waiting period, and the pages prescribing one disagree with each other by a factor of seventeen.
What the evidence will support you doing
The defensible list is short. Everything on it traces to something above, and the things that are not on it are not here because the evidence does not reach them.
Ask what the friendship is for, and answer honestly. This is the one piece of advice with a validated instrument behind it. Of the four published reasons, one — unresolved romantic desire — was the reason associated with negative outcomes, and it is also the reason people are most able to describe as something else. If the honest answer is that you are in it because leaving entirely would end the possibility, you are not deciding whether to be friends. You are deciding how to wait, and what the research on the mechanism of intermittent hope actually found is the relevant page.
Separate contact from surveillance, and treat them as different decisions. This is the most useful finding on the page and the one the genre collapses. In the only study that measured both, remaining connected was associated with less longing, less negative feeling and less sexual desire; looking at their page was associated with more of all three. Deleting someone and then checking on them anyway is the combination with nothing to recommend it.
Expect the trade, and pick a side of it deliberately. Staying connected was associated with feeling better and growing less, on a standard growth inventory. You are allowed to choose the softer landing. What is not available is choosing it while believing it costs nothing.
Do not treat distress as the enemy. Avoidance predicted less pain and less growth, and the path from distress to growth ran through rumination rather than around it. A bad fortnight is not evidence that you have chosen wrong.
If it is practical, let it be practical. A friendship held together by a lease, a pet, a workplace or a mutual group did well by the people in it and did not last, and that combination is a reasonable thing to accept in advance rather than discover as a disappointment.
And if you are composing posts for one reader, stop. That behaviour sat with cyber obsessional pursuit in the one study that measured it, and the person it harms is not you.
The limits, in the same type size as the findings. Two of the three papers this subject rests on are paywalled with no open copy, so their effect sizes are unreadable from outside a university. The Facebook studies are from 2011 and 2012 and describe one platform. The growth study is cross-sectional and retrospective and 83% women in its second sample. The profile study is 256 people in one survey. Nothing here was randomised, because nobody can randomise this, so every arrow in this article is modelled rather than observed.
What survives all of that is narrower than a rule and more useful than one. Whether to stay friends with an ex has no measured answer. What it is doing for you does, and the research has been pointing at that question for nine years while the search results have been answering the other one. If the pattern you recognise in yourself is the one where the same arrangement keeps reappearing with different people, why people repeatedly choose the same kind of wrong person is the next thing to read, and the rest of this section sits under Mindset.
Pulling it together
If one thing survives from all of this, it is that the question has been asked wrong for a long time. Whether to stay friends with an ex has no measured answer, and it cannot have one, because nobody can assign people to the two conditions. What the motive is doing for you does have one, and the research has been pointing at it since 2017 while the search results have gone on answering the other question with more confidence than any of the studies contain.
The second thing is the split. Staying connected and keeping watch are different behaviours with opposite associations, and the common move — cutting someone off and then checking on them anyway — is the one combination the evidence has nothing good to say about.
The third is the trade, and it is the one worth sitting with. Remaining connected was associated with feeling better and growing less. Avoidance was associated with less pain and less growth. Both halves of that are measured, and neither tells you which to want. What they rule out is the version where the softer landing is also the faster recovery.
And the limits are part of the finding. Two of the three central papers are paywalled with no open copy anywhere, so their magnitudes are unreadable from outside a university library, and that is the honest state of a subject every website on the first page writes about with no number in it at all.
Frequently asked questions
Should you stay friends with an ex?
There is no measured answer to that question and there probably cannot be, because nobody assigns people to stay friends or not — the two groups differ beforehand in how the relationship went, who ended it and what each person still wants. What has been measured is the motive. Griffith, Gillath, Zhao and Martinez built and validated a scale for it in 2017 and found four reasons: security, practical, civility, and unresolved romantic desires. Staying friends out of unresolved romantic desire was associated with negative outcomes; security and practical reasons were associated with more positive ones. So the useful version of the question is not whether, it is what the friendship is doing for you. One caveat worth knowing: that paper is paywalled with no open copy, so the published abstract gives the direction of each finding and not its size.
Is it healthy to be friends with an ex?
The closest thing to an answer is a trade rather than a verdict. Tara Marshall studied 464 people after a breakup and separated remaining Facebook friends with an ex from conducting surveillance on their page. Remaining friends was associated with less negative feeling, less sexual desire and less longing for the ex-partner than not remaining friends. It was also associated with lower personal growth — measured, in the companion study, with the 21-item Posttraumatic Growth Inventory. Surveillance was associated with the opposite on the first three: more distress, more longing, more desire, and lower growth as well. All of it held after controlling for offline contact, personality and the characteristics of the relationship and the breakup. So staying connected looks like feeling better and reorganising more slowly, and these are correlations rather than experiments.
Does wanting to stay friends mean you still have feelings?
Sometimes, and there is a measured version of this rather than a guess. Carl Schneider and David Kenny asked each participant about two friendships at once — a cross-sex friend who had once been a romantic partner, and one who never had — so each person served as their own comparison. More romantic desire, and more negative qualities, were associated with the ex-partner friendship; more positive qualities with the never-romantic one. Their conclusion was that a friendship with an ex is qualitatively different from a platonic friendship rather than a slightly harder version of one. That is not the same as saying wanting it means you want them back. It means residual wanting is common enough in these friendships that assuming it is absent is the riskier bet.
Should I stay friends with my ex if I still want to get back together?
This is the one case where the research is least ambiguous, and it points the same way the honest answer already does. Unresolved romantic desire was the single reason, out of the four in Griffith and colleagues' measure, that was associated with negative outcomes on its own. The friendship in that case is the shape the hope has taken, and an arrangement that keeps hope alive on an irregular schedule is the same mechanism described in the research on sporadic, non-committal contact. Nothing in the literature says you cannot. It says this is the version that went badly, and that the people in it are usually able to describe it as something else.
How long should you wait before becoming friends with an ex?
Nobody has measured a waiting period, and any specific number you have been given was invented. Time since the breakup does appear in Griffith and colleagues' list of things that predicted whether people stayed friends — along with sex, sexual orientation, attachment styles, personality traits and the reasons for the breakup — but a predictor is not a prescription, and no study has compared people who waited different lengths of time and followed what happened. This site has traced two of the popular numerical rules in dating back to their origins and neither came from research: the three-day rule came from a 1996 film, and the 3-6-9 rule has no identifiable source at all.
Is it bad to unfollow, delete or block an ex?
It is a normal thing to do and it is not the majority thing to do. McDaniel and colleagues surveyed 256 people who had broken up within the previous year and found four patterns: clean breakers, who did almost nothing, were 61.3%; ritual cleansers, who deleted history, the ex and the ex's family and friends but did not monitor, were 15.6%; wistful reminiscers, who monitored and interacted frequently but deleted little, were 12.9%; and impulsives, who did high amounts of everything, were 10.2%. Distress was highest in impulsives, then wistful reminiscers, then ritual cleansers, then clean breakers. The survey cannot say which way that runs — deleting may follow distress rather than cause it. What it does show is that nearly two thirds of people neither purged nor kept watch.
Should I worry that my partner is still friends with their ex?
Nothing in this literature measured that, so anyone giving you a confident answer is giving you an opinion. What is adjacent, and worth knowing, is the direction of the arrow in the one longitudinal study on longing for ex-partners. Spielmann, Joel, MacDonald and Kogan followed people in relationships at three points over six months and found that declines in current relationship quality predicted increased longing for an ex, above and beyond attention to alternatives generally; the reverse path held only for people considering recent exes. Their reading is that ex-partners get used to prop up belongingness when a relationship starts to fail. Which suggests that the more informative thing to look at is the state of the relationship you are in rather than the friendship you are worried about.
Does staying friends with an ex stop you from moving on?
The claim is usually supported with a link to Spielmann and colleagues' 2013 paper, which measured emotional attachment to ex-partners and not friendship, contact or whether anyone was still speaking. The part of the claim with direct support behind it is narrower: in Marshall's study, remaining Facebook friends was associated with lower personal growth, and so was surveillance — while remaining friends was associated with less longing, not more. So the measured cost of staying connected is a slower reorganisation of the self, not a stronger pull back towards the person. The marker of having finished that reorganisation, in the research on recovery, is being able to describe what you want from the next year without the account routing through them.
Sources
What this article draws on. For the broader bibliography see our sources page.
- Griffith, R. L., Gillath, O., Zhao, X. & Martinez, R. (2017) — Staying friends with ex-romantic partners: predictors, reasons, and outcomes. Personal Relationships, 24(3), 550-584 — two studies identifying predictors of staying friends, developing and validating a measure of the reasons, and examining outcomes. Four reasons: Security, Practical, Civility, Unresolved Romantic Desires. Unresolved Romantic Desires was associated with negative outcomes; Security and Practical with more positive ones; Practical or Civility with the friendship being less likely to last long term. Sex, sexual orientation, attachment styles, personality traits, time since breakup and reasons for breakup all predicted staying friends. Paywalled with no open-access copy in any repository, so the direction of each finding is readable from the published abstract and the sample sizes and effect sizes are not.
- Griffith, R. L., Gillath, O., Zhao, X. & Martinez, R. (2018) — Reasons for Staying Friends with an Ex-Partner Scale. PsycTESTS Dataset, deposited 11 June 2018 — the instrument developed in the paper above, catalogued by the American Psychological Association as a measure in its own right. Linked so that the existence of a validated scale for this can be checked rather than taken on trust.
- Marshall, T. C. (2012) — Facebook surveillance of former romantic partners: associations with postbreakup recovery and personal growth. Cyberpsychology, Behavior, and Social Networking, 15(10), 521-526 — 464 participants. Surveillance of an ex-partner's page was associated with greater current distress, more negative feelings, more sexual desire, more longing and lower personal growth. Remaining Facebook friends, relative to not remaining friends, was associated with LESS negative feeling, sexual desire and longing, and also with lower personal growth. All results held after controlling for offline contact, personality traits, and characteristics of the former relationship and the breakup. The study this article is built around, and the one the ranking pages cite without separating its two halves.
- Marshall, T. C., Bejanyan, K. & Ferenczi, N. (2013) — Attachment styles and personal growth following romantic breakups: the mediating roles of distress, rumination, and tendency to rebound. PLoS ONE, 8(9), e75161 — open access. Study 1: 411 recruited, 364 modelled. Study 2: 465 after removing duplicate IPs, 370 modelled, 83% women. Avoidance predicted less distress (β = -.19 and -.14) and less growth (β = -.30); anxiety predicted more distress (β = .44 and .42). Indirect paths from distress to growth ran through brooding (β = .26, CI .17-.35) and reflection (β = .14, CI .08-.22). Cross-sectional and retrospective; the arrows are modelled rather than observed.
- Schneider, C. S. & Kenny, D. A. (2000) — Cross-sex friends who were once romantic partners: are they platonic friends now? Journal of Social and Personal Relationships, 17(3), 451-466 — the within-person design: each participant rated both a cross-sex friend who had once been a romantic partner and one who never had, on rules of friendship, costs and benefits, romantic desires and overall quality. More positive qualities were associated with the platonic friend; more negative qualities and more romantic desire with the ex-partner friend. The authors conclude the two are qualitatively different relationships. Paywalled.
- Spielmann, S. S., Joel, S., MacDonald, G. & Kogan, A. (2013) — Ex appeal: current relationship quality and emotional attachment to ex-partners. Social Psychological and Personality Science, 4(2), 175-180 — three waves over six months. Longitudinal declines in current relationship quality predicted increased longing for ex-partners, above and beyond attention to relationship alternatives in general; increased longing predicted decreased relationship quality only among those considering recent ex-partners. Cited across this subject as evidence that staying friends holds you back, which is not what it measured: the variable is emotional attachment to a former partner, not friendship or contact.
- McDaniel, B. T., Drouin, M., Dibble, J., Galovan, A. M. & Merritt, M. (2021) — Are you going to delete me? Latent profiles of post-relationship breakup social media use and emotional distress. Cyberpsychology, Behavior, and Social Networking, 24(7), 464-472 — 256 participants who had broken up within the previous year; about 38% reported distress sometimes or more often. Four profiles: clean breakers 61.3%, ritual cleansers 15.6%, wistful reminiscers 12.9%, impulsives 10.2%. Distress ran highest in impulsives, then wistful reminiscers, then ritual cleansers, then clean breakers. The only base rate in this article.
- Lyndon, A., Bonds-Raacke, J. & Cratty, A. D. (2011) — College students' Facebook stalking of ex-partners. Cyberpsychology, Behavior, and Social Networking, 14(12), 711-716 — 411 valid participants; three factors in how people communicated with ex-partners on Facebook: Covert Provocation, Public Harassment and Venting. Each was related to perpetrating cyber obsessional pursuit and obsessive relational intrusion, and participants who engaged in cyber obsessional pursuit were almost six times more likely also to perpetrate the offline form.
- De Smet, O., Uzieblo, K., Loeys, T., Buysse, A. & Onraedt, T. (2015) — Unwanted pursuit behavior after breakup: occurrence, risk factors, and gender differences. Journal of Family Violence, 30(6), 753-767 — 631 adult ex-partners, who reported on average five to six unwanted pursuit behaviours after separation, with no difference by gender or by same- versus opposite-gender pairing. Predicted by the ex-partner having initiated the breakup, rumination and cognitive preoccupation, anxious attachment in the former relationship, borderline traits and past delinquent behaviour. Rumination was a stronger predictor in women than men.
- Bullock, M., Hackathorn, J., Clark, E. M. & Mattingly, B. A. (2011) — Can we be (and stay) friends? Remaining friends after dissolution of a romantic relationship. The Journal of Social Psychology, 151(5), 662-666 — the more satisfied people had been during the relationship that ended, the more likely they were to remain friends and to engage in friendship maintenance behaviours; friendship maintenance fully mediated the association between past romantic satisfaction and current friendship satisfaction.
- Mogilski, J. K. & Welling, L. L. M. (2017) — Staying friends with an ex: sex and dark personality traits predict motivations for post-relationship friendship. Personality and Individual Differences, 115, 114-119 — the source of the recurring "staying friends with your ex means you might be a psychopath" headline. Listed here because the article discusses it, and listed with a warning: Unpaywall reports no open-access location and no repository copy, and OpenAlex, Semantic Scholar and Crossref all return the record with the abstract withheld by the publisher. Its sample sizes and effect sizes are not reported in this article because they could not be verified, and the secondary summaries that do report them disagree with one another.
- TIME (2018, updated 2024) — "Should You Stay Friends With an Ex? Here's What Experts Say" — the best-sourced page on the first screen of results, linked here because this article audits it: it points at five of the right papers and names no author, year, sample size or effect size for any of them, and it uses the Spielmann paper for a claim about friendship when that paper measured longing.
- Find A Helpline — Free, confidential support services listed by country — the referral used in the first section of this article, which covers the case the rest of the page does not address.
Keep reading
- MindsetIn-Laws and Divorce: Half of the Famous 20% Is in the Paper, Half Was Never ThereA husband close to her parents: 20% lower divorce odds, published. A wife close to his: that figure is in no paper, and the real result is about race.
- MindsetBest Age to Get Married: The Famous 28-to-32 Window Came From a Blog PostThe 11%-a-year figure comes from a 2015 blog post, and its own model turns at 28, not 32. In the same survey, in a journal, the curve flattens instead.
- MindsetMoney Arguments and Divorce: The "Top Predictor" Line Came From a Press ReleaseThe famous line is a quote in a 2013 press release. The paper is paywalled, and a co-author's other 2012 paper calls money only mildly predictive of divorce.
- MindsetSplitting Housework and Divorce: The Famous 50% Comes From a Report Almost Nobody Has ReadThe report behind "couples who share chores divorce 50% more" is 228 pages in Norwegian. Read in full, the gap is 3.1% against 4.6% over four years.
- TextingTexting Habits That Quietly Push People AwayAlmost every rule in this genre has never been tested. The handful of things that have been measured point somewhere else entirely.
- Body LanguageWhat Body Language Actually Predicts, and What It Does NotThe largest synthesis found eleven behaviours that track liking and a list of famous ones that track nothing. Three of those used to be on this page.