attraction.lab

Does the No Contact Rule Work? What Has Actually Been Measured About Cutting Contact

The phrase returns nothing in PubMed. The one diary study the pages half-cite is 58 people, and nearly half of exes reconcile with no technique at all.

Research-backed writing on attraction, dating and relationships. Drafted by AI against a named bibliography, published only when every claim links to its source.
40 min read
Editorial illustration for: Does the No Contact Rule Work? What Has Actually Been Measured About Cutting Contact

Before the article. There are two reasons to stop speaking to an ex-partner and they are not the same question. If a former partner is threatening you, following you, monitoring your phone or your location, threatening to share images, or using a child or a shared account to keep open a channel you have asked to close, cutting contact is a safety measure and this page is not the one you need: Find A Helpline, in the sources at the bottom, lists free and confidential services by country, and in the United States the National Domestic Violence Hotline is 1-800-799-7233, or text START to 88788, 24 hours a day.

With that said. The first screen of results for this question contains no Healthline, no Cleveland Clinic, no Verywell Mind and no Psychology Today — it is coaching sites, lifestyle media, a quiz and an ebook storefront. Seven of those pages were downloaded and read rather than judged by their titles, and across all seven there are zero journal names, zero sample sizes and zero effect sizes. Two of them link exactly one study each. Neither says how many people were in it: one is a brain-imaging study of fifteen people looking at a photograph, and the other is a 28-day diary of 58 undergraduates.

The phrase itself returns nothing in the research literature. A PubMed title-and-abstract search for "no contact rule" gives zero records, and in Europe PMC the phrase matches eleven papers about concussion protocols and graphene. The relevant work exists under other names, and four studies have measured post-breakup contact directly: a 28-day beeper diary of 58 people, a five-month study of 122 separating adults whose contact was coded from ambient audio recordings rather than self-report, a survey of 137 separating adults, and a 2026 survey of 438 recent breakups. They do not all point the same way, and the one that pointed the other way found contact associated with better adjustment among people who had accepted the ending.

Sixteen sections follow: the safety case, what the ranking pages contain counted rather than characterised, the prescribed durations side by side, the fifteen-person brain scan doing the persuading, the diary study the genre half-remembers, the audio-recording study nobody has read, the study where contact ran the other way, a 2026 paper none of them knows exists, the reconciliation base rates, what the reunited relationships looked like, whose numbers the biggest site in the niche is publishing, the one experiment on turning the feeling down, what people say they actually do, the only randomised design in the field, where the rule is quietly right, and the limits. Our editorial policy explains why every empirical claim here carries a linked source, and why the figures that could not be read are named as unreadable instead of borrowed from a summary.

#1

Before anything else: cutting contact is sometimes a safety decision, and that is a different page

There are two completely different reasons to stop speaking to a former partner, and the internet writes about them as if they were one thing.

The first is the one this page is about: a breakup hurt, contact keeps hurting, and someone wants to know whether going quiet for a set number of days helps them recover or brings the other person back. That question has an evidence base, it is thinner than the search results suggest, and the rest of this article is a careful reading of it.

The second is not a recovery technique at all. If a former partner is threatening you, turning up where you did not say you would be, monitoring your phone or your location, threatening to share images, or using a child or a shared account to keep a channel open that you have asked to close, then stopping contact is a protective measure and the thing you need is not an article about how many days to wait. Find A Helpline, listed 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.

It is worth saying why this distinction is at the top rather than in a footnote. The research reviewed further down includes the one dataset that compares people who stayed broken up against people who broke up and reunited, and in that dataset the reunited group reported physical violence at twice the rate of either the stably together or the stably broken up. That is a correlation in a cross-sectional survey and it does not say that reuniting causes anything. It does mean that the outcome the no contact rule is most often sold as producing — the ex coming back — is the one outcome in this literature that has been measured alongside harm, and the result is not a selling point.

If the pattern you recognise is affection alternating with punishment until you start managing your own behaviour around the reaction, what the research behind the phrase trauma bonding actually supports is the more relevant page, and it is honest about which parts of that phrase have evidence and which do not.

#2

What the pages that rank actually contain, counted rather than characterised

The search was run on 6 November 2026 for the exact question — does the no contact rule work — and the first screen of results is unusual even for this niche. In order of appearance: romantific.com, The Good Men Project, a blog attached to an ebook storefront, a marriage.com quiz literally titled Will The 'No Contact Rule' Work Quiz?, a page hosted on a `pages.dev` subdomain, The Everygirl, a dating coach's personal site, two pages from a course-selling blog, eharmony, a breakup coach's Substack, developgoodhabits.com, two pages from a coaching marketplace, and a parenting blog.

No Healthline, no Cleveland Clinic, no Verywell Mind, no Psychology Today. This site has now measured enough of these to recognise the pattern: when a dating idea has an invented name — a test, a rule with a number — the authorities do not turn up, and the first page fills with coaching, lifestyle media and forums. It happened with the three-day rule and it happened with the 3-6-9 rule, and both of those are the pieces on this site that did best in search.

The next step is the one that matters, and it is the one the genre skips: seven of those pages could be downloaded as plain text and read rather than judged by their titles. Across all seven, searching the body text for the markers of a sourced claim returns almost nothing. Zero instances of "et al". Zero journal names. Zero sample sizes. Zero reported effect sizes. Zero percentages, apart from the prescribed waiting periods. The word study appears once or twice per page, attached to no particular study.

Two of the seven do link exactly one paper each, which is more than the audit of several earlier topics on this site turned up, and both are worth taking seriously. Neither of them tells you how many people were in the paper it links. The next two sections are about those two papers, because what the pages leave out of them is the whole argument.

#3

The durations, laid side by side, which is the fastest way to see that nobody sourced them

Collect the numbers from the pages themselves and the problem becomes visible without any research at all.

  • developgoodhabits.com: "Some people say the no contact rule takes 30 days, some say six months, and others say a year is needed in order to properly grieve the end of a relationship." It then recommends committing to whichever one you picked.
  • eharmony: "The recommendation is a minimum of 21 days," stretching to "60 or even 90" after a bad breakup, followed two paragraphs later by "We'd suggest a minimum of three months, or even six."
  • The course-selling blog: a 30-day timeline, with 90 days mentioned as the longer version.
  • The largest breakup-coaching site in the niche, measured separately because it does not rank for this exact question: 30 days as the default, 21 and 45 as variants.

So the recommended dose ranges from three weeks to a year, the same page sometimes gives two different answers, and not one of these numbers is attributed to anything. This is the same structure as the other invented numbers this site has traced: the figure is specific enough to feel like a finding and has no origin anyone can point to.

It is worth being precise about what is and is not being claimed here. Nobody has shown that 30 days is wrong. Nobody has tested it. No study in this literature assigned people to a waiting period, compared one length against another, or followed a group instructed to stay silent for a month against a group who was not. The number is a convention, and a convention is a legitimate thing for a coach to work with — as long as it is not presented as a dose.

The honest version of the question is therefore not "how long" but "what happens when there is contact, measured how, in whom". There are four studies that have measured that directly, and they do not all point the same way. That is the substance of this page.

#4

The 1989 brain scan that is doing most of the heavy lifting, and its sample size

The single linked study on developgoodhabits.com sits behind the sentence that does the persuading: "research done through MRI scans shows that experiencing romantic rejection or going through a breakup has similar effects on the brain as an addict's drug withdrawal does. So using the no contact rule will certainly be a struggle — but the truth is that this is the fastest and most effective method."

The link goes to Helen Fisher, Lucy Brown, Arthur Aron, Greg Strong and Debra Mashek, in the Journal of Neurophysiology in 2010. The study is real, it is well known, and it is good. Here is what it is.

Ten women and five men. Fifteen people who had recently been rejected by a partner and reported still being intensely in love, placed in a scanner and shown, alternately, a photograph of the person who rejected them and a photograph of a familiar neutral acquaintance, with a distraction task in between. Looking at the rejecter activated the ventral tegmental area bilaterally, the ventral striatum, medial and lateral orbitofrontal and prefrontal cortex, and the cingulate gyrus — regions associated with gains and losses, with craving, and with emotion regulation, and overlapping with areas identified in work on cocaine craving. The authors' own suggestion is that this overlap may help explain the obsessive behaviour that follows rejection.

What the study did not do, and did not claim to do, is measure contact. Nobody in it was assigned to stop speaking to anyone. There is no waiting period, no follow-up, no comparison of people who cut contact against people who did not. It is a cross-sectional imaging study of fifteen people looking at a picture, and the inference from "rejection engages reward circuitry" to "therefore thirty days of silence is the fastest and most effective method" is an argument the paper does not contain and cannot support.

The addiction analogy is also doing more rhetorical work than the data justifies. Overlapping activation is not equivalence of mechanism, and the paper says so in the hedged language that journals require and websites strip out. If the leap from a brain region to a confident behavioural prescription sounds familiar, it is the same move audited in what the popular body-language signals of attraction are actually built on.

#5

The one diary study in the field, and what the page that cites it leaves out

eharmony's single footnote is the better citation, and it points at the right paper. The sentence it supports is modest — "time is a key factor in recovery from a breakup, basically, the more time passes, the less it hurts" — and the link goes to David Sbarra and Robert Emery, Personal Relationships, 2005.

That paper deserves better than a footnote about time passing, because it is the only study anyone has run that watched the emotional aftermath of a breakup day by day while it was happening.

Fifty-eight young adults. They were recruited while still in a serious dating relationship — which is the design detail that makes the study unusual, because it means the pre-breakup measures were taken before anyone knew there would be a breakup. When a relationship ended, the participant was given a beeper and signalled at random times for 28 days to complete an emotions diary. Compared with people whose relationships stayed intact, the dissolution group reported more emotional volatility, most of all immediately after the breakup. Multilevel growth modelling showed a linear decline in love and curvilinear patterns for sadness, anger and relief.

And then the finding that every page on this subject is reaching for and none of them quotes: contact with a former partner slowed the decline for love and sadness. Attachment style and the rated impact of the breakup predicted both where people started emotionally and how fast they changed.

So there is a real result here, and it is the closest thing the no contact rule has to an empirical foundation. It is also the second of four separate clocks that the research runs on a breakup, which is laid out in how long it actually takes to get over a breakup — and the reason a figure for one of them never transfers to the others. Four things have to be said about this one in the same breath, and the pages that gesture at it say none of them.

It is 58 people. That is a diary study with thousands of observations, which is the point of the design, but it is 58 participants, mostly undergraduates, in one place, twenty-one years ago.

Contact was observed, not assigned. Participants reported whether they had been in touch. Nobody was told to stop. People who keep talking to an ex differ from people who do not in every way that also predicts how a breakup goes — who ended it, how much is unresolved, how much they still want the relationship back — and a correlation between contact and a slower decline in love is exactly what you would expect if the people still in love are the ones still making contact. The arrow is not identified.

The window is 28 days and that is the length of the diary, not a finding about 28 days. Nothing in the paper says what happens in month two. The popular 30-day rule is not derived from this study; it merely happens to be about as long.

The full text is paywalled with no open copy anywhere. Unpaywall reports no repository version, Semantic Scholar reports the same, and the publisher's page is closed. So the sample size and the direction of the effects come from the abstract, which is public, and the effect sizes cannot be read from outside a university library. This page is not going to invent them. That constraint is why our editorial policy exists in the form it does.

#6

The study that recorded people instead of asking them, and found a two-month lag

The strongest design in this literature is fifteen years newer, and nothing on the first page of results mentions it.

Karey O'Hara, Austin Grinberg, Allison Tackman, Matthias Mehl and David Sbarra published it in Clinical Psychological Science in 2020, and the method is the reason to care. One hundred and twenty-two recently separated adults — 35 men and 87 women, aged 24 to 65, average age 43.8, married an average of 13.4 years and entering the study on average 3.8 months after physically separating — wore the Electronically Activated Recorder on three occasions across five months. The EAR is a device that samples ambient audio in short snippets throughout the day. Contact with the ex-partner was therefore coded from recordings of the person's actual life, not from a questionnaire asking them to recall how often they had seen their ex. That removes, at a stroke, the most obvious problem with every other study on this page.

The results split in a way that is easy to misreport, so here they are in order.

Contact and distress measured at the same time: nothing. The between-person association between in-person contact with the ex and separation-related psychological distress was not reliably different from zero, and neither was the time-varying effect. On a snapshot, people who saw their ex more were not more distressed.

Contact and distress two months later: something. More frequent in-person contact predicted higher distress at the next assessment, holding concurrent contact, demographics, relationship characteristics and attachment orientation constant. For people without children with the ex, a one standard deviation increase in contact relative to their own average was associated with a 0.28 unit increase in distress two months later — which the authors calculate as roughly 112% of the amount of change in distress typically observed between visits. Put plainly: in these data, a burst of extra contact was associated with cancelling out slightly more than the recovery that two months would otherwise have brought.

And it only appeared in people without children. The interaction was reliable, and for participants who shared children with their ex the relation was flat — p = .953, which is as close to nothing as a number gets. The authors do not pretend to know why, and neither will this page, but the practical consequence is clear and it is the opposite of what the advice industry says: the co-parenting contact that every page tells you to minimise is the contact where this effect was absent.

The limits. This is a separation and divorce sample with an average age of 43 and marriages averaging more than thirteen years, so it is not the twenty-two-year-old six weeks out of a two-year relationship that the search query usually belongs to. Contact was not assigned here either, and a lagged association is still an association. And the cleanest reading of the result is not "contact is bad" but "contact today is unrelated to how you feel today and related to how you feel in two months" — which is a far stranger and more interesting claim than anything in the coaching literature, and one that nobody writing about the no contact rule has noticed exists.

#7

And the study where contact ran the other way, which is where the rule falls apart

If the previous section were the whole story, the rule would come out of this reasonably well. It is not the whole story, and the study that complicates it comes from the same research group.

Ashley Mason, David Sbarra, Amanda Bryan and Lauren Lee published it in the Journal of Social and Clinical Psychology in 2012. One hundred and thirty-seven community adults — 50 men, average age 40.6 — recruited through newspaper advertisements, divorce recovery groups and the local family court, an average of 3.77 months after separating from a marriage that had lasted on average just over thirteen years. Psychological adjustment was measured with the 22-item Impact of Event Scale–Revised. Two behaviours were recorded separately: non-sexual contact with the ex, and sexual contact with the ex. And one more variable was measured that turns out to run the whole result: acceptance of the separation.

First, a base rate worth sitting with. One hundred and thirteen of the 137 participants — 82.5% — reported having contact with their ex-partner. Whatever the internet recommends, four out of five separating adults in this sample were in touch.

Second, the headline: there was no main effect of contact on adjustment at all. Not a weak one. Contact with the ex was not significantly correlated with adjustment, and in the regression with covariates its coefficient was not reliably different from zero. Sexual contact with the ex: also no main effect. What did predict adjustment, and strongly, was acceptance of the separation — r = −.63, and β = −.65 in the model. People who had accepted that the relationship was over were doing better, and that single variable swamped everything about who was talking to whom.

Third, and this is the finding that should end the blanket version of the rule. Contact and acceptance interacted (β = −.28, p = .02), and when the interaction is unpacked the direction is not the one the genre assumes. Among participants high in acceptance, those who were in contact with their ex reported significantly better adjustment than those who were not (B = −5.95, p = .01). Among participants low in acceptance, there was no significant difference between being in contact and not being in contact (p = .23).

Read that twice, because it inverts the advice. In the one study that measured contact, acceptance and adjustment together, going quiet was not associated with doing better in either group, and in the group who had made peace with the ending, having contact was the arrangement associated with the better outcome.

There is a fourth result in the paper and this page is going to report it rather than quietly drop it, because dropping inconvenient findings is the habit that produced the genre this article is auditing. Sexual contact with the ex also interacted with acceptance, and among participants low in acceptance, those having sex with their ex reported better adjustment than those not (B = −3.56, p = .05, which is the boundary and should be read as such). No advice follows from that. It is cross-sectional, it is one borderline interaction in one sample of 137 people, the obvious alternative explanation is that people who feel better are the ones who end up in bed with an ex, and the same authors' other data point the other way about what unresolved wanting does to people. It is in here because it is in the paper.

What survives from all of this is a reframe rather than a rule: the thing that moved the outcome was not the contact, it was whether the person had accepted the ending. That is the same variable, under a different name, that carries the only other randomised result in this field, and it is the subject of the last third of this page.

#8

A 2026 paper none of the pages has read

The literature on this question is still moving, which is itself an argument against confident advice, and the most recent entry was published nine months ago.

Erin McClung, Ryan Pecore, Elisa Stragapede, Paola Michelle Garcia Mairena and Serena Corsini-Munt published it in the Journal of Sex & Marital Therapy in February 2026. Four hundred and thirty-eight people, predominantly young, cisgender, heterosexual undergraduates who had recently been through a breakup, in a retrospective cross-sectional survey. The paper's main interest is something else entirely — whether being autonomy-supportive while ending a relationship is associated with the other person coming out of it better — and the answer there was that autonomy support during the breakup went with more positive mood and higher subjective vitality afterwards.

But the second finding is directly on this subject: ongoing contact with an ex-partner was associated with increased anxiety and depression.

That is three studies now — a 28-day diary of 58 undergraduates, a five-month audio-recording study of 122 separating adults, and a survey of 438 recent breakups — in which contact lines up with worse outcomes on at least one measure, against one study of 137 separating adults in which it lines up with better outcomes for the people who have accepted the ending. Which is what "it depends" looks like when you read the actual papers instead of using the phrase to avoid them.

The limits on the new one are the ones its own authors state. It is retrospective and cross-sectional, so people were remembering a breakup rather than being followed through one. The sample is undergraduates. And the full text is paywalled, with no repository copy, so the sample size and the direction come from the abstract and the effect sizes are not readable from outside a subscription — the same constraint as the 2005 diary study, reported the same way.

It is also worth noting what finding it into this article required. The phrase that the entire internet uses returns nothing: a PubMed title-and-abstract search for "no contact rule" gives zero records, and in Europe PMC the phrase matches eleven papers, every one of them about concussion protocols, volleyball refereeing or the contact angle of graphene. The relevant literature exists under other words — post-breakup contact, ex-partner contact, separation-related psychological distress, relationship churning — and it only appeared when the same question was asked of several databases rather than one. The counts are written out in the sources at the bottom so they can be re-run rather than taken on trust.

Further illustration for: Does the No Contact Rule Work? What Has Actually Been Measured About Cutting Contact
#9

The claim the rule is actually sold on: that they come back

Half the pages on the first screen are not really answering the question in the query. They are answering a different one — whether thirty days of silence will make a former partner reappear — and they answer it in the affirmative, with varying degrees of coyness. eharmony puts it plainly: "if you want your ex back... implementing the no contact rule for 30 days can be a great way to remind them of just what they're missing."

Nobody has run that experiment. You cannot: it would require assigning half a group of heartbroken people to stay silent and the other half not, and then waiting to see whose ex-partner got in touch. There is no such study, in any database, under any of the search terms above.

What does exist is the base rate, and the base rate is the single most useful number on this page, because it tells you what happens with no technique at all.

Sarah Halpern-Meekin, Wendy Manning, Peggy Giordano and Monica Longmore analysed the Toledo Adolescent Relationships Study — 792 young adult daters and cohabitors — and reported it in the Journal of Adolescent Research in 2013. 44.8% reported a reconciliation: a breakup in the relationship they were describing, followed by getting back together. Among the 427 who had broken up at least once, 216 reported having had sex with the ex while broken up. And reconciliation was not usually a single event: the average number of reconciliations across the whole sample was 1.096, but among people who had experienced at least one, the average was 2.445.

René Dailey and colleagues put the same thing a different way in Personal Relationships in 2009. In Study 1, 445 American college students: almost two thirds had been in an on-again/off-again relationship at some point. In Study 2, 236 participants with quantitative measures, confirming the pattern and adding that a greater number of renewals went with more negatives and fewer positives.

And it does not stop at dating. Amber Vennum, Rachel Lindstrom, J. Kale Monk and Rebekah Adams, in the Journal of Social and Personal Relationships in 2014, used a nationally representative sample of 323 cohabiting and 752 married couples and found that over a third of cohabiters and a fifth of spouses had broken up and renewed within their current relationship.

So: somewhere between a fifth and two thirds of people, depending on which population and which question, break up and get back together. It is close to a coin flip in the young-adult dating data. Which means any technique at all can take credit for it. Wait thirty days and a reunion follows: that is roughly what you would expect from doing nothing, and the rule has no way to show it did better than nothing, because the comparison has never been made. The only random assignment anywhere near this question went somewhere else entirely: the experiments behind the rebound idea manipulated how easily people believed they could find somebody new, and never put anybody in a relationship at all.

#10

What the reunited relationships looked like, which is the part nobody sells

There is one place in this literature where the prize the rule promises has actually been measured, and it is the same 792 people.

The same four authors published the second analysis in the Journal of Marriage and Family in 2013, and it sorted everyone into three groups. 40.97% were stably together. 14.87% were stably broken up — they broke up once and that was that. 44.16% were churners: on and off. Across the whole sample, 39.84% reported physical violence in the relationship and 52.51% reported verbal abuse.

The comparison is the finding. Churners were twice as likely as either the stably together or the stably broken up to report physical violence, and half again as likely to report verbal abuse, and the association held net of demographics, family background, parental conflict history, adolescent temper, cohabitation, relationship duration, self-esteem, sense of control, self-disclosure and love.

And the quieter half of the result is the one worth carrying away: the stably broken up looked like the stably together. The authors' own sentence is that the people experiencing the most conflict are not the ones choosing to permanently exit their relationships. Ending it and leaving it ended was not the distressed condition in these data. Going back and forth was.

The limits are serious and they are the usual ones. This is cross-sectional, from one metropolitan county in Ohio, and it establishes no direction: relationships with violence in them may be the ones that churn, rather than churning producing the violence, and both are plausible. Nothing here predicts what would happen to any particular couple.

What it does do is put a price tag on the question the rule is usually asked. Dailey's own review of the field, in the International Encyclopedia of Interpersonal Communication, summarises it without hedging: compared with partners whose relationships do not break up and renew, on-off partners report lower satisfaction, more relational uncertainty, greater conflict and more aggression, along with greater structural constraints. Vennum's data add that cycling while dating predicts cycling in cohabitation and in marriage, and that cyclers report more constraints to ending the relationship, more uncertainty about its future, and lower satisfaction.

None of that says a reunion cannot work. Dailey's review notes explicitly that some partners use the transitions to improve the relationship. It says that the average reunited relationship in the measured samples is worse than the average relationship that was never interrupted, and worse than ending it once, and that a technique whose advertised output is a reunion is advertising something with a known downside that no page in the search results mentions. The mechanism by which hope gets rationed out in small doses until someone cannot leave has its own small literature, and that is what the research on breadcrumbing actually found.

#11

The numbers the biggest site in the niche does give, and whose numbers they are

One site in this industry does publish statistics, in quantity, and it deserves a section of its own — partly because it is the most visible page on the subject anywhere, and partly because its numbers are a short course in how to read a number.

Its own text, quoted:

  • "Last year I wrote that the no contact rule success rate has been close to 75% for our clients." Success is then defined by the author, in three parts, the first of which is that your emotions have settled down.
  • "Three years ago when I polled our success stories for the year we found that 70% of them used some form of no contact. Two years ago... 90%... Last year when I ran the same poll we found that 98% of them used some form of no contact."
  • "I polled our private facebook support group to see how often exes actually reached out during a no contact rule and I was shocked to learn that 60% of the time your ex WON'T reach out to you during no contact."
  • "In 2012 a graduate student researching the impact Facebook has on breakups found that 89% of exes admitted to Facebook stalking their significant others." No name, no journal, no sample.
  • "Research has shown that only about 7% of the population falls into this category," of a fearful attachment style. No source.

The 70-90-98 sequence is the instructive one, because it is a measurement that cannot come out any other way. The population polled is the success stories. The question is whether they used the technique the site sells. A number computed on the people for whom things worked, asked whether they did the recommended thing, will rise towards 100% as the business gets better at recommending it, and it contains no information whatsoever about whether the technique works — because the people for whom it did not work are not in the denominator. The same arithmetic would prove that lottery tickets win, if you only polled winners.

The 75% has the matching problem from the other side: the population is paying clients of the service, the outcome is defined by the service, and there is no comparison group of people who did nothing.

This is a sharper finding than the one this site has been getting used to. For the last several topics the gap was that nobody gave a number. Here somebody gives a dozen, and the question that had to be asked instead was who is in the denominator. It is the same question that decided whether the five love languages quiz measures what it claims and how one study of the ick ended up circulating with three different sample sizes: the number is rarely the problem, the sample is.

And it should be said that the best page in these results is honest about exactly the thing the others are not. One small test-and-article site states plainly that the 30-day rule "is a popular theory, but it's not grounded in psychological research," and correctly summarises the 2005 diary finding about contact slowing the decline in love and sadness, including the caveat that the study only covered 28 days. It gives neither paper's sample size — the diary study is 58 people and the brain-imaging study it also cites is fifteen — but it is the one page on this subject that tells the reader the number of days was made up. Credit where it is due.

#12

If the goal is to stop loving them, there is an actual experiment

Underneath "does no contact work" there is usually a more specific wish: to stop feeling this. That wish has been tested experimentally, which almost nothing in this article has, and the result is a trade rather than a technique.

Sandra Langeslag and Michelle Sanchez, in the Journal of Experimental Psychology: General in 2018, took people who were upset about a recent breakup and had them perform one of three strategies before looking at a photograph of their ex, with EEG recording throughout. The three were negative reappraisal of the ex-partner, reappraisal of the love feelings themselves, and distraction. The results:

  • Negative reappraisal decreased love feelings — and made participants feel more unpleasant. It worked, and it cost something.
  • Love reappraisal did nothing to how in love or how pleasant participants felt.
  • Distraction did not change love feelings at all, but made participants feel more pleasant. It did not touch the attachment; it improved the mood.
  • All three reduced motivated attention to the ex, measured as a smaller late positive potential in the EEG — an objective index rather than a self-report, which is why the study is worth the space.

That maps onto the question better than any number of days does. If what you want is to stop loving them, the strategy with evidence behind it is the one that makes you feel worse while it works. If what you want is to feel better this afternoon, the strategy with evidence behind it does not reduce the love at all. The two goals the no contact rule promises to deliver simultaneously came apart in the one experiment that measured them separately.

The companion paper is open access, so its numbers can be checked rather than quoted: Langeslag and Jan van Strien, in PLOS ONE in 2016, established that deliberate love regulation is possible at all, with 32 participants in Study 1 and 20 plus 20 in Study 2. That is the size of the experimental literature on regulating feelings for an ex. Seventy-two people, in two papers, by one research group. Everything else on this subject is a survey.

The 2018 paper is paywalled with no repository copy, so the strategies and their directions come from the abstract and the sample size is not reported there. That is the third of the central papers on this page that cannot be fully read from outside a university, and it is the honest state of the field rather than a failure of searching.

#13

What people actually say they do, and the strategy almost nobody names

The open-access half of that work contains a table that is, quietly, the most awkward number in this article.

Langeslag and van Strien asked participants, in open questions with no options supplied, what they do to feel better when they are broken-hearted. The answers were coded into categories. Across the three groups — 32 in the first study, then 20 people currently in a relationship and 20 who had recently broken up — the counts came out like this:

  • Distraction: 59%, 75% and 80%. The most common answer in every group, by a wide margin.
  • Social support: 53%, 20% and 5%. Second in the first study.
  • Undertaking new activities, reappraisal of various kinds, expressing emotions: scattered through the middle.
  • Avoidance: 3%, 10% and 10%. One person out of 32 in the first study.

Asked separately what they do specifically to decrease love feelings — which is closer to what the rule is for — avoidance rose only to 19%, 15% and 15%, while focusing on the negative aspects of the former partner came out top at 31%, 45% and 45%.

So the behaviour that an entire advice industry prescribes as the first and most important step is the one that, when people are asked openly and in their own words, almost nobody spontaneously reports using. They distract themselves and they talk to people. That is not evidence that avoidance does not work — these are small samples and self-reports of strategy, not outcomes. It is evidence that the rule is not describing what people do, and it is a reason to be suspicious of a prescription that has to be taught because nobody arrives at it by themselves.

It also lines up with a base rate from much further up this page: 82.5% of 137 separating adults were in contact with their ex. The modal response to a breakup, in the data, is to stay in some kind of touch and distract yourself. The modal response in the search results is a thirty-day silence. Those are different populations: one is people, the other is content.

#14

The only randomised study in breakup recovery points at talking, not silence

There is exactly one study in this area where people were assigned to conditions, and it is the reason the last two sections of this article look the way they do.

Grace Larson and David Sbarra published it in Social Psychological and Personality Science in 2015. Two hundred and ten recently separated young adults were assigned either to a measurement-intensive condition — 120 people, four visits with multimethod assessments over nine weeks — or to a pre-post condition — 90 people, intake and exit assessments only, across the same nine weeks. Nobody was given therapy. Nobody was given advice. The only difference between the groups was how often somebody asked them about the breakup.

The measurement-intensive group showed larger decreases in self-concept disturbance over time. No other main effect of condition appeared. And the mediation is where it gets interesting: improvement in self-concept clarity explained the decreases in breakup-related emotional intrusion, in loneliness, and in the use of first-person plural words — we, us, our — when describing the separation. The authors' own framing is that research assessing coping can produce change without creating any expectation of doing so.

Two things follow, and they are the two most practical things on this page.

The variable that moved was the self, not the feeling. Repeatedly being asked who you are now, and answering, was associated with reassembling a sense of self faster. That is the same finding, from the same research group, that the whole of what the research actually measures when it measures having moved on is built around: poorer self-concept recovery preceded poorer well-being, rather than following it. If you only take one thing from this article, take that the marker is not how often you think about them.

The direction is the opposite of the rule's logic. The no contact rule is, in its popular form, a programme of not engaging — not with the ex, often not with the subject. The one assigned intervention in this literature was more engagement with the subject, nine weeks of it, and it helped. Whether that generalises to talking to a friend, writing, or a therapist is not established by this study and this page is not going to claim it is. What is established is that being asked about it four times instead of twice was enough to show up in a measured outcome, which is a strange and encouraging result, and that it had nothing to do with whether anyone was speaking to their ex.

#15

Where the rule is quietly right, and it is not about days

It would be a bad article that audited this subject and came out with nothing to say for the thing. There is a defensible core, it is narrower than the rule, and almost all of it has already been measured on this site rather than this one.

The separation that actually earns its keep is contact versus watching. Tara Marshall studied 464 people after a breakup and did the thing the rest of the field does not: she separated remaining connected to an ex-partner from going and looking at their page. The two came apart in opposite directions. Surveillance went with more distress, more negative feeling, more sexual desire and more longing. Remaining connected went with less of all three. Both went with lower personal growth on a standard inventory. That is set out in full, with its limits, in what the research on staying friends with an ex measured, and the consequence for this page is blunt: blocking someone and then checking on them from a logged-out browser is the one combination in these data with nothing to recommend it, and it is also the most common way a thirty-day silence is actually lived.

Acceptance is the moderator, and it is not a number of days. The 137-person study found contact mattering only through whether the person had accepted the ending. Nothing about waiting produces acceptance. Thirty days of counting down to a text is not acceptance; it is a deadline, and a deadline keeps the relationship live in the only place it still exists.

If contact is pulling you back into monitoring, reducing it is maintenance, not strategy. Cues drive intrusive thought. Removing cues is a reasonable thing to do for its own sake. It stops being that the moment it is framed as a lever on another person's behaviour.

The duration is yours to pick and it is not a dose. Nobody has tested 21, 30, 60 or 90 days against each other or against nothing. Picking a length because it helps you commit is fine. Believing the length is doing the work is the part with no support.

And the thing with the best evidence is not silence, it is reorganising the self. That is the randomised result, and it is the one finding on this page that comes from an assigned condition rather than an observed one. Describing what you want from the next year without routing any of it through them is the practical version of the measure, and whether those patterns shift at all over a lifetime is measured in years, not in weeks, so a bad fortnight in month two is not evidence of anything.

#16

The limits, in the same type size as the findings

This page has made a number of specific claims and the reasons to distrust each of them belong here rather than in a disclaimer.

Nothing in this literature was randomised except one study about being interviewed. Nobody can assign people to stop speaking to an ex-partner, so every arrow about contact in this article is modelled from observation. People who maintain contact differ from people who do not in exactly the ways that also predict recovery, and no amount of covariate adjustment fixes that.

Three of the central papers cannot be fully read. The 2005 diary study, the 2018 love-regulation experiment and the 2026 survey are all paywalled with no repository copy in Unpaywall, OpenAlex or Semantic Scholar. Their sample sizes and the direction of their findings come from public abstracts. Their effect sizes are not quoted here because they are not readable, and a number borrowed from a secondary summary is how the three different sample sizes for one study of the ick got into circulation.

The samples do not match the reader. The two contact studies with the best designs are separating and divorcing adults with an average age in the forties and marriages averaging more than thirteen years. The diary study and the love-regulation experiments are undergraduates. The churning data are from one county in Ohio. Someone six weeks out of an eighteen-month relationship at twenty-four is not represented well by any of them.

The base rates are not forecasts. That 44.8% of 792 young adults reported a reconciliation does not mean your probability is 44.8%. It means reunions are common enough that their occurrence is uninformative about whether anything you did caused one.

The violence finding is cross-sectional and the direction is unknown. Churners reporting twice the physical violence is consistent with churning being harmful and equally consistent with violent relationships being the ones that churn.

And the honest summary of the whole question is unsatisfying. There is no measured answer to "does the no contact rule work", because the rule as stated — a fixed number of days of silence, undertaken to recover or to produce a reunion — has never been tested. What has been measured is contact, in four studies that disagree with each other in interesting ways; acceptance, which predicted adjustment far more strongly than contact did; reunion, which is common and goes with worse relationships on average; and self-concept reorganisation, which is the one thing anybody has manipulated and seen move.

If the pattern you recognise is that the same arrangement keeps reappearing with different people rather than that this particular one hurt, why people repeatedly choose the same kind of wrong person is the next thing to read. If the question underneath is whether you should have left at all, what the research says about why people stay in relationships that are not making them happy treats it as the alternatives-and-investment problem it has been measured as, and the relationship satisfaction test scores those two separately from satisfaction for that reason. The rest of this section sits under Mindset.

Pulling it together

The question has no measured answer, and it is worth being clear about why rather than hiding behind "it depends". The rule as stated — a fixed number of days of silence, undertaken either to recover faster or to make someone come back — has never been tested. No study assigned anyone to a waiting period. No study compared one length against another. The number of days is a convention that ranges, across the pages selling it, from three weeks to a year, and two of those pages give two different answers in the same article.

What has been measured is contact, and in four studies it behaves in a way no rule captures. Measured at the same moment, contact and distress were unrelated. Measured two months apart, more contact predicted more distress — but only in people who did not share children with their ex. And measured alongside acceptance of the ending, contact had no main effect at all, and went with better adjustment among the people who had accepted it.

The outcome the rule is most often sold as producing has a base rate: somewhere between a fifth and two thirds of people break up and get back together, with no technique involved. The relationships that came back reported lower satisfaction, more uncertainty, more conflict and twice the rate of physical violence, while the people who broke up once and left it broken up looked like the people who never broke up at all.

And the one thing anybody has actually manipulated was not silence. It was being asked about it: 210 separated adults, half of them interviewed four times instead of twice, and the group that was asked more reassembled a sense of themselves faster. The marker worth watching is not the day count. It is whether you can describe what you want from the next year without any of it routing through them.

Frequently asked questions

Does the no contact rule work?

There is no study of it. The rule as stated — a set number of days of silence, undertaken to recover or to bring someone back — has never been tested against anything, in any database: a PubMed title-and-abstract search for the phrase returns zero records. What has been measured is contact with an ex-partner, in four studies that disagree. A 28-day diary of 58 people found contact slowing the decline in love and sadness. A five-month study of 122 separating adults found no association at the same moment and a two-month lagged one, present only in people without children. A 2026 survey of 438 recent breakups found contact associated with more anxiety and depression. And a study of 137 separating adults found no main effect of contact at all, with contact going alongside better adjustment in people who had accepted the ending. The variable that predicted adjustment most strongly in that last study was not contact, it was acceptance.

How long should the no contact rule last — 30 days, 60 days, 90?

Nobody has tested any of them, and the pages recommending them do not agree. Collected on 6 November 2026 from the first screen of results: 21 days as a minimum rising to 60 or 90, then "a minimum of three months, or even six" two paragraphs later on the same page; 30 days; 30 days or six months or a year, your choice; 30 days with 90 as the long version. No study in this literature assigned people to a waiting period or compared one length against another, so there is no basis for preferring any number. Picking a length because a deadline helps you stick to a decision is reasonable. Believing the length is the active ingredient is not supported by anything.

Does no contact make your ex miss you or come back?

That experiment has never been run, and the base rate makes it very hard to run informally. In a study of 792 young adult daters and cohabitors, 44.8% reported a breakup followed by getting back together in the relationship they described, and among those who had broken up at least once the average number of reconciliations was 2.4. In a nationally representative sample, over a third of cohabiters and a fifth of spouses had broken up and renewed within their current relationship. Reunions are common enough that one following thirty days of silence is roughly what you would expect from doing nothing. The worthwhile finding is about what came back: in the same 792-person sample, people in on-off relationships were twice as likely to report physical violence and half again as likely to report verbal abuse, while those who broke up once and stayed broken up looked like those who never broke up. That is cross-sectional and does not establish direction, but it is the only place the promised prize has been measured.

Will blocking my ex help me get over them faster?

The most useful distinction in this area is between staying connected and going to look. A study of 464 people after a breakup separated the two and they ran in opposite directions: surveillance of an ex-partner's page went with more distress, more longing and more sexual desire, while remaining connected went with less of all three. Both went with lower personal growth on a standard inventory. The practical consequence is that deleting someone and then checking on them anyway is the one combination with nothing to recommend it in these data, and it is a common way a blocking decision is actually lived. Removing cues that trigger intrusive thinking is a reasonable thing to do as maintenance. It is not a lever on the other person, and the study was correlational, so the arrow could run the other way: people in more pain may be the ones who look more often.

Does no contact apply if we have children together?

The one study that recorded contact rather than asking about it found the opposite of what the advice would predict. Among 122 recently separated adults whose contact with their ex was coded from ambient audio recordings over five months, more contact predicted higher distress two months later — and that effect was present only for people who did not share children with their ex. For participants who did, the relation was flat, at p = .953. The authors do not explain why and neither will this page. What follows is that the co-parenting contact the genre tells people to minimise is the contact where the measured association was absent, and that organising co-parenting around a rule derived from a different population has no support behind it.

Does the no contact rule work differently on an avoidant or an anxious partner?

Nothing has been tested in that form. Attachment does appear in this literature, and where it appears it predicts how the person going through the breakup fares rather than how a technique lands: in the 28-day diary study, attachment style predicted both the emotional starting point after the breakup and the rate of change over the following month. No study has assigned contact or non-contact and compared the result across attachment styles, so claims about what an avoidant ex will do after thirty days are not derived from data. Our 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 it is a description of patterns, not a category to assign another person to in their absence.

Is it normal to feel worse after going no contact?

It is consistent with the one experiment that measured this directly. In a study where people upset about a recent breakup were assigned to regulate their feelings in different ways, negative reappraisal of the ex-partner did reduce love feelings — and made participants feel more unpleasant. Distraction did not reduce love feelings at all but made them feel more pleasant. The two things the rule promises to deliver together came apart: the strategy that reduced the attachment was the one that cost mood, and the strategy that improved mood did not touch the attachment. Separately, recovery is not linear in any of this work, so a bad fortnight in month two is not evidence that a decision was wrong. What would be worth attention is not slowness but flatness — no movement at all across months, or distress that is deepening rather than fluctuating, which is a reason to talk to a doctor or a therapist.

What does have evidence behind it, if not the rule?

One thing, and it is the only assigned condition in this literature. Two hundred and ten recently separated young adults were allocated either to four research visits over nine weeks or to intake and exit assessments only. Nobody received therapy or advice; the only difference was how often somebody asked them about the breakup. The group asked more often showed larger decreases in self-concept disturbance, and the improvement in self-concept clarity explained reductions in breakup-related emotional intrusion, in loneliness, and in how often they used we and us when describing the separation. The variable that moved was the self rather than the feeling, which matches the other finding this site has written up at length: poorer self-concept recovery came before poorer well-being rather than after it. The practical form of the measure is whether you can describe what you want from the next year without any of it routing through them, and what the research measures when it measures having moved on sets out the rest.

Sources

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

Keep reading