Relationship Anxiety: Why You Worry Even When Things Are Good

Relationship anxiety is worry about your relationship that keeps running even when your partner hasn't given you a reason for it. That's the part people find hardest to explain to anyone else. Things are fine. Your partner is kind, shows up, says the right things. And you're still lying awake at 1am wondering whether they're going to leave.

If that sounds familiar, the important thing to know early is that the anxiety usually isn't information about your relationship. It's information about your nervous system, and where it learned to expect connection to be unreliable. Here's what the research actually says, and what tends to shift it.

What Is Relationship Anxiety?

Relationship anxiety is persistent doubt, fear, or worry about a romantic relationship that's out of proportion to what's actually happening. It's not the same thing as having concerns about a relationship that genuinely has problems. That distinction matters, and we'll come back to it.

In practice it looks like a few recognisable things. Rereading messages for tone. Noticing your partner is quieter than usual and building a whole story out of it. Asking "are we okay?" more often than you'd like to admit. Mentally rehearsing the breakup conversation while nothing is going wrong. Some people describe it as waiting for the other shoe to drop, except the shoe never drops and the waiting never stops.

The worry also tends to be sticky. You get reassured, you feel better for an hour or a day, and then the doubt comes back wearing a slightly different outfit.

How Common Is Relationship Anxiety?

There's no clean prevalence figure for relationship anxiety specifically, because it isn't tracked as its own condition. What we do have is the surrounding picture, and it's not a small one.

According to the National Institute of Mental Health, an estimated 19.1 percent of US adults had an anxiety disorder in the past year, and 31.1 percent will experience one at some point in their lives. The past-year figure splits noticeably by sex: 23.4 percent of women against 14.3 percent of men. Those numbers come from the National Comorbidity Survey Replication, which collected its interviews between 2001 and 2003. They're old, but they're still the figures NIMH publishes for any anxiety disorder, so treat them as a rough scale rather than a current count.

Attachment research gets closer to the specific pattern. Drawing on Hazan and Shaver's 1987 study, summarised by attachment researcher R. Chris Fraley at the University of Illinois, roughly 60 percent of adults describe themselves as securely attached, about 20 percent as avoidant, and about 20 percent as anxious-resistant. That last group is the one most likely to experience what we're describing here.

So somewhere around one in five adults carries the attachment pattern that drives this. You're not unusual, and you're definitely not broken.

Why Does It Happen When Nothing Is Wrong?

Because the alarm system was calibrated a long time ago, in a situation where it made sense.

Anxious attachment typically develops when early caregiving was inconsistent. Not absent, and not necessarily unkind. Inconsistent. Sometimes warm and available, sometimes distracted or unresponsive, and hard to predict which you'd get. A child in that environment learns something specific: connection is possible, but you have to work for it and watch for it. So they develop heightened vigilance and turn the volume up on distress signals, because that's what reliably brought a caregiver back.

That strategy worked. That's the part worth sitting with. It wasn't a malfunction, it was an intelligent adaptation to the environment you were in.

The problem is that it doesn't switch off when you grow up and end up with someone perfectly reliable. The same system runs in adult relationships, and it looks like monitoring your partner's tone, reading meaning into a slow reply, and needing more reassurance than most people can supply. Attachment researcher Sue Johnson calls this a hyperactivated attachment system. It scans for threat constantly rather than responding to actual threat, which is exactly why a good relationship doesn't calm it down.

There's a specific mechanism worth naming here, because most people doing this don't realise they're doing it. Simpson and Rholes reviewed the adult attachment literature in Current Opinion in Psychology (2017) and describe how anxious people, when they sense a threat to the relationship, "get into the heads" of their partners and infer relationship-threatening thoughts that aren't necessarily there. You decide you know what they're thinking. Then you feel less close to them, based on something you invented. The review also notes that anxious people lean on hyperactivating coping strategies, which push attention toward the distress and encourage rumination, keeping the attachment system switched on more or less permanently.

So it's not only that the alarm is sensitive. It's that the way you try to soothe it keeps it running.

Past relationships matter too. If someone left abruptly, cheated, or ran hot and cold, the system updates accordingly and carries that forward into partners who had nothing to do with it. Our guide to attachment styles covers how each pattern forms in more detail.

How Is It Different From a Real Gut Feeling?

This is the question people actually want answered, and it deserves a straight response rather than reassurance. Sometimes the worry is right. Sometimes you're picking up on something real and being told it's just your anxiety.

A few things usually separate them.

• Intuition points at something specific. You could name it, describe it, give an example. Anxiety tends to be vague and hard to pin down, and gets more slippery the harder you look at it.
• Intuition holds still. Real concerns stay put. Anxiety moves targets. Resolve one worry and another appears within days, which tells you the worry wasn't really the point.
• Intuition can be satisfied. Ask yourself what you'd need to see to feel settled. If you have an answer, that's a concern. If nothing would fully do it, or the relief would last an afternoon, that's anxiety.
• Intuition doesn't need constant checking. If you already know something's off, you're not compulsively looking for proof. You're deciding what to do about it.

Those are useful rules of thumb, but there's a research finding here that complicates the picture in a way most articles skip. The assumption is usually that anxious people are bad at reading their partners. That turns out to be wrong, and the truth is more specific.

So the radar works. If you've spent years being told you're imagining things and it never quite fit, that's why. The problem isn't detection. It's calibration, and there's a separate study that pins down which direction the error runs.

Put the two together and you get something more useful than "trust your gut" or "it's just your anxiety." You're probably picking up the signal. You're just reading the meaning too dark. Your partner being quiet is real. Your partner being quiet because they're done with you is the part the bias adds.

Which gives you a practical move. Instead of asking whether the feeling is real, ask what you've actually observed versus what you've concluded. Write both down if you need to. The observation usually survives scrutiny. The conclusion often doesn't, and separating them is most of the work.

The honest complication is that you can have both at once. Anxious people do sometimes end up with unreliable partners, partly because the anxiety makes inconsistency feel familiar. If the concerns are specific and repeatable, take them seriously regardless of what your attachment style is.

What If Something Actually Happened?

Everything above assumes the alarm is going off when there is no fire. For a lot of readers that is exactly right. But some of you are not in that group, because there was a betrayal, and the watchfulness you are carrying is a response to something that genuinely occurred.

That distinction matters more than it might look, because the advice in this guide can misfire on you. Cutting back reassurance-seeking makes sense when you are training an oversensitive alarm to settle. It makes much less sense when the alarm already went off once and was right. That is not miscalibration, it is memory, and asking you to override it is asking for the wrong thing.

The shape of the two is different if you listen closely. Relationship anxiety asks what if. Post-betrayal vigilance asks again. One is bracing for something that has not happened, the other is tracking something that has. You can have both at once, which is common and confusing, but the second one does not respond to the same work.

Here is the encouraging part, and it comes from a decent-sized study rather than a slogan. Atkins and colleagues looked at 145 couples who named infidelity as a problem against 385 couples in therapy for other reasons, drawn from a community sample in Germany and Austria and published in the Journal of Family Psychology in 2010. The infidelity couples started out significantly more distressed and more depressed. They kept improving through therapy and on to the six-month follow-up, and by that point they were not statistically distinguishable from the couples who had never dealt with an affair.

So the gap closes. That is worth sitting with if you have been told, usually by someone who is not a clinician, that this never really heals.

And now the other half, because giving you only the hopeful number would be dishonest. An earlier study by the same lead author, published in the Journal of Consulting and Clinical Psychology in 2005, tracked a much smaller group of 19 infidelity couples inside a trial of 134. Two findings came out of it. Couples where the affair was disclosed did better than couples where it stayed hidden. But at the five-year mark, the infidelity couples had divorced at 43 percent against 23 percent for the rest.

Both of those are true at the same time. Recovery is genuinely possible and measurable at six months, and the elevated risk does not vanish at five years. Nineteen couples is a small sample and worth treating as a signal rather than a verdict, but it is the honest counterweight to the first number.

What changes in practice is mostly what you ask for. If you go looking for help, name it as trust repair rather than as anxiety management, because those lead to different rooms. Individual work on an oversensitive threat system is not the same as couples work on rebuilding a broken agreement, and the second one is what this situation calls for. Our guide to rebuilding trust covers what that repair actually involves. If the affair is still running, or still concealed, that is the thing to deal with first, since the 2005 finding on disclosure points the same way.

Is It Relationship Anxiety or Relationship OCD?

There's a version of this that goes past ordinary relationship anxiety, and it has a name. Relationship obsessive compulsive disorder, usually shortened to ROCD, was mapped out by Guy Doron and colleagues at Reichman University. It comes in two flavours. Relationship-centred, where the obsession is "do I really love them, is this right," and partner-focused, where you fixate on some feature of your partner, their looks, their intelligence, the way they chew.

The tell isn't the content of the thought. Plenty of people wonder whether they're with the right person. The tell is the shape of it. The doubt is intrusive rather than chosen, it arrives unwanted and won't leave, and you end up running rituals to settle it. Comparing your relationship to other people's. Testing whether you still feel a spark. Googling the same question at midnight for the fourth time this week.

And it's not a milder thing than textbook OCD, which is the part that tends to surprise people.

Same severity, different target. That matters practically, because ROCD responds to a specific approach. Exposure and response prevention, the standard OCD treatment, works on resisting the rituals rather than resolving the doubt. Standard reassurance-based couples work can actually make it worse, since reassurance is the compulsion. If the description above landed uncomfortably close, that's worth raising with a therapist by name rather than describing it as general anxiety.

Is It Relationship Anxiety or Rejection Sensitivity?

There's a second thing that looks almost identical from the inside, and separating them tells you where the work actually goes.

Geraldine Downey and Scott Feldman described it in the Journal of Personality and Social Psychology in 1996, volume 70, pages 1327 to 1343. They called it rejection sensitivity, and defined it as a disposition to anxiously expect rejection, readily perceive it, and react to it intensely. Notice the middle step. It isn't only that you fear being left. It's that you keep finding evidence of it in things that are genuinely ambiguous.

There's a study that shows this happening in real time, and it's the clearest illustration I've seen. Carolina Norona, Jessica Salvatore, Deborah Welsh and Nancy Darling worked with 386 adolescents, 193 couples, and published it in the Journal of Adolescence in 2014, available through PubMed Central. They recorded couples having two kinds of conversation, one neutral and one conflictual, then compared what each partner thought had happened against trained observers.

People low in rejection sensitivity rated the two conversations very differently, at 1.38 for the neutral one and 2.42 for the conflictual one. People high in rejection sensitivity rated them about the same.

Read that again, because it's not what most people assume. The problem wasn't that they overreacted to conflict. It's that a neutral conversation and a difficult one registered as roughly the same event. The instrument that's meant to tell you when something is wrong had stopped discriminating.

The wider picture holds up. Aditi Mishra and Rebecca Allen pulled together 60 studies into 21 separate meta-analyses in Personality and Individual Differences in 2023, volume 208. Higher rejection sensitivity went with lower relationship satisfaction, lower closeness, and lower perceived partner satisfaction, alongside more relationship concerns, more conflict, more jealousy, and more self-silencing. They found no evidence of publication bias, which matters when a finding is this tidy.

Self-silencing is the one worth flagging. It's the thing that makes rejection sensitivity so hard to spot from outside. You don't necessarily get louder. You go quiet, stop raising things, and manage yourself so nothing triggers the rejection you're braced for.

So how do you tell which you have?

• Relationship anxiety tends to be about the future. Will this last, is this right, what happens if it goes wrong.
• Rejection sensitivity tends to be about the last ten minutes. What did that tone mean, why was that reply short, what did that pause signify.
• Rejection sensitivity shows up outside the relationship too. With friends, at work, in group chats. If the same reading of ambiguous signals happens everywhere, it isn't really about this partner.

The practical difference is where you intervene. Relationship anxiety responds to work on tolerating uncertainty. Rejection sensitivity responds to work at the interpretation step, which means slowing down between the ambiguous thing and the meaning you assign it. That gap is usually less than a second, which is exactly why it needs deliberate practice rather than willpower. Our guide on hypervigilance in relationships covers the scanning that keeps it primed.

Is Low Self-Esteem the Real Cause?

It's the self-diagnosis most people land on, and it's about half right. The link is real and it's been measured carefully. But the direction it runs in isn't what the usual advice assumes, and that changes what's actually worth your effort.

Start with how tightly the two are bound together. Shen, Liu and Brat surveyed 1,708 adults for a study in The Professional Counselor (2021) and found self-esteem and adult attachment anxiety moving almost in lockstep, with a standardised coefficient of -.95. Self-esteem tracked psychological distress too, at -.68. What matters is where their model put self-esteem: in the middle of the chain, not at the start of it. Early attachment shaped self-esteem, self-esteem fed adult attachment anxiety, and that anxiety carried through into distress, with an indirect effect of -.24.

So low self-esteem isn't a separate problem sitting next to your relationship anxiety. It's a link in the same chain. That's part of why "just work on your confidence" sounds like useful advice and then goes nowhere.

Now the finding that flips the usual plan. Martin Pinquart pooled 202 studies covering 81,485 people for a meta-analysis in Psychological Reports (2023), looking at attachment security and self-esteem together. The two correlated at r = .34, a solid moderate link and no surprise. The interesting numbers are the cross-lagged ones, the analyses built to test which comes first. Attachment security predicted later gains in self-esteem at r = .19. Self-esteem predicted later gains in attachment security at r = .08. Same pair of variables, measured in both directions, and one direction came out more than twice as strong as the other.

Read that plainly and it says feeling securely attached builds self-esteem faster than self-esteem builds secure attachment. The arrow mostly runs the opposite way from how the advice is usually written.

One caveat before you take it too far. Pinquart was measuring attachment to parents, not to a partner, so this isn't a direct claim about your relationship. It's the foundation layer that the romantic patterns get built on top of. The direction it points is still worth taking seriously.

What it means in practice is that pouring months into esteem-building exercises while leaving the relational patterns untouched is working the weaker lever. The stronger one is the material this guide keeps circling: what happens in your body when you're triggered, whether you can sit through the gap before a reply arrives, whether reassurance can actually land when it's offered. Work on those and self-esteem tends to come along behind. It's also why healing anxious attachment usually gets further than a confidence project does.

Does Perfectionism Feed Relationship Anxiety?

Yes, and this is the version that hides best, because from the outside it looks like you're just trying hard.

The kind that matters here isn't the tidy-desk sort. Researchers call it perfectionistic concerns, and it covers harsh self-scrutiny, an outsized fear of making mistakes, sensitivity to how other people are judging you, and reactions to small failures that land far heavier than the failure deserved. The socially prescribed flavour is the one to watch. That's the belief that other people are demanding perfection from you and will withdraw if they don't get it.

Sit that next to relationship anxiety and the overlap is obvious. If you believe you're being kept on condition of performance, then every ordinary wobble reads as a scorecard.

Sean Mackinnon, Simon Sherry, Martin Antony, Sherry Stewart, Dayna Sherry and Nathan Hartling tested this in 226 heterosexual couples for a study in Journal of Family Psychology (2012, Vol. 26, Issue 2, pages 215 to 225). They used a mixed design, combining longitudinal measures with daily online questionnaires from both partners across days 2 to 15, so conflict was scored from what the couple actually reported day by day rather than from one person's summary at the end.

What they found supports what's called the social disconnection model. Perfectionistic concerns didn't run straight into depressive symptoms. They ran through the relationship first. The perfectionism generated interpersonal problems, the conflict piled up, and the low mood came out the other end. The damage travelled through the connection, not around it.

That's the part worth holding onto, because it reframes the whole thing. If you've been treating your standards as a private matter that only costs you, the evidence says they're not staying private. They're showing up in the room.

Here's the practical tell, and it's a useful one to check honestly. Relationship anxiety on its own has you monitoring your partner. Are they quieter than usual, did that reply take longer, what did that tone mean. Add perfectionism and you're monitoring yourself just as hard. Was that too needy, did I text first again, was I too much at dinner. Two surveillance jobs running at once, and the second one is the reason you often can't say what you actually want. It feeds straight into the self silencing habit covered further down.

The fix isn't lowering your standards, which never works as an instruction anyway. It's testing the belief underneath them. Being slightly disappointing and staying chosen is the experience that shifts it, and you can only get that by risking a small ordinary need out loud. Our guide on how to communicate needs is the practical version of that, and anxious attachment recovery covers what tends to come up when you start.

Why Does Reassurance Stop Working?

Because reassurance treats the symptom, and the relief it gives is exactly what keeps the loop alive.

The cycle runs like this. Anxiety spikes. You ask your partner if everything's okay. They say yes. You feel better. Your brain files that away as evidence that asking works. Next time the anxiety spikes, the urge to ask is a bit stronger and arrives a bit sooner. Over time the threshold drops, the relief gets shorter, and your partner starts to feel like nothing they say ever lands.

That last part is measurable. Evraire, Dozois and Wilde studied 110 couples in Europe's Journal of Psychology (2022), tracking daily reassurance seeking and trust. Higher attachment anxiety predicted significantly more excessive reassurance seeking (b = 0.11, p < .01), and the effect was more than twice as strong in men (b = 0.21) as in women (b = 0.09). More striking: for women high in attachment anxiety, seeking reassurance predicted lower trust the following day, not higher.

Read that again, because it's the whole problem in one finding. The thing you do to feel safer is the thing that leaves you trusting less tomorrow.

None of this means reassurance is bad or that you should never ask. It means asking alone doesn't fix it, and past a certain frequency it makes things worse for both of you. If you recognise this loop, how to stop overthinking in relationships goes further into interrupting it.

Does Talking It Through With Friends Help?

Often, yes. But there's a particular way of talking that makes you feel closer to your friends and more anxious at the same time, and it happens to be the default setting for most people with relationship anxiety.

Psychologists call it co-rumination, a term Amanda Rose introduced in 2002. It isn't just talking about a problem. It's going over the same problem repeatedly, speculating about causes, rehearsing what was said, and staying with how bad it feels rather than moving through it. If you've ever sent a friend a screenshot and asked what they think he meant by that, you already know the shape of it.

The research on it is unusually clean, because it finds both things at once. Denson, Moulds and Grisham surveyed 495 UK adults in PLOS ONE (2026), a sample stratified by age, gender and ethnicity to be representative, with a mean age of 46.65 and a range from 18 to 87. Co-rumination predicted more anxiety, with a standardised coefficient of .20 and p below .001. In the very same model, it predicted less relationship distress, at minus .14 with p of .002, and it correlated with better relationships wellbeing and lower loneliness.

Read those two numbers together, because that's the whole problem. Co-ruminating genuinely does bring you closer to the people you do it with. It also genuinely winds you up. You aren't imagining the closeness, which is exactly why the habit is so hard to see as a habit.

And it's the same loop described in the reassurance section above, just pointed at a different person. Lisa Starr ran a 14-day daily diary study of 51 young adults, published in the Journal of Social and Clinical Psychology and available through PubMed Central, and found co-rumination and excessive reassurance seeking correlated at .44. The detail that matters most is what happened day to day. For people high in trait co-rumination, days with more problem talk predicted higher depressed mood, a coefficient of .46 with p below .001. For people low in co-rumination, talking about problems had no such effect at all.

So it isn't the talking. It's the kind of talking. Which is good news, because it means the answer was never to stop telling your friends things.

A few things separate processing from rehearsing:

• Are you after a new angle or the same verdict? Processing is open to an answer you didn't expect. Rehearsing is shopping for the one you already have.
• Have you told this exact story before? Once is context. The fourth time, to the third friend, is the loop.
• How do you feel when you hang up? Settled is a good sign. More wound up than when you started is the clearest signal there is, and it's easy to miss because the conversation itself felt good.
• Try asking for something specific. "Can you help me work out what I actually want to say to him" gives a friend a job. "Can you believe he said that" gives them a script.

There's one more thing worth knowing about the friends you lean on. They're working from a skewed sample. They mostly hear about your relationship on the days it hurts, because those are the days you reach for your phone. Over a year that builds a picture of your partner assembled almost entirely from their worst moments, and then that picture gets handed back to you as advice. Your friends aren't wrong to worry. They're just reading from a file you filled in selectively.

This is one of the clearer arguments for talking to someone outside your circle. A therapist has no stake in your partner being the villain, no accumulated file of grievances, and a professional reason to slow down a story you've told fluently forty times rather than nod along to it. If waiting lists or cost are what's stopping you, online therapy with licensed therapists is the lower friction route in. And if the loop shows up in your own head as much as in your group chat, how to stop overthinking in relationships covers the solo version of it.

Does Social Media Make Relationship Anxiety Worse?

Yes, and the newest research suggests it might be doing more of the work than your attachment style is.

Checking is easier than it's ever been. You can see when they were last online, who liked the photo, who they followed this week, whether the reply took eleven minutes when it usually takes two. None of that information existed twenty years ago. Now it's free, infinite, and available at 1am.

That last sentence is the one to hold onto. Attachment anxiety on its own didn't do the damage. The behaviour it produced did, and the behaviour sits closer to the outcome than the personality trait does.

Which is unusually good news, because behaviour is the more changeable of the two. You can't decide to have a different nervous system by Thursday. You can decide not to open their profile tonight, and that turns out to be the lever that's actually connected to something.

Worth being precise about why scrolling makes it worse rather than better. Surveillance produces ambiguous information almost every time. A like, a follow, a slightly delayed reply. Ambiguity is exactly what an anxious system is worst at, because it fills gaps with the worst available reading. So you go looking for relief and come back with fresh raw material. It's the reassurance loop again, except your partner isn't even in the room to interrupt it.

The practical version isn't a digital detox. It's narrower than that. Notice which specific check you run when the anxiety spikes, because it's usually one habit rather than social media generally, and put friction in front of that one thing. Log out. Move the app off your home screen. Mute the account you keep visiting. You're not trying to develop willpower. You're trying to make the twenty seconds between urge and action a bit longer.

Does Relationship Anxiety Affect Your Body Too?

Yes, and the size of the effect surprises most people. This gets treated as a purely emotional problem, something happening in your head that you should be able to think your way out of. The measurements say otherwise.

Stanton and Campbell studied 116 married heterosexual couples, 232 people in total, and published the results in the peer-reviewed journal PLoS One (2014). Higher attachment anxiety predicted more medical symptoms (b = 0.42), more bodily pain (b = 0.33), worse overall health perceptions (b = -0.28), and poorer social functioning (b = -0.23), all at p below .01.

That's a body under sustained load, not a mood. If you've been putting the headaches, the shallow sleep, and the permanently tight shoulders down to something unrelated, they might not be unrelated.

But the finding that should genuinely change how you approach this is the second one. The researchers tested whether feeling well supported protected people's health. For people low in attachment anxiety, it did. For people high in attachment anxiety, it didn't. Their health stayed poor even when they perceived support as high, and on some measures the picture looked worse at high support, not better (health perceptions b = -0.49, p < .001; social functioning b = -0.42, p < .001).

Which lines up uncomfortably well with the reassurance research above. Being given support doesn't help much if the system receiving it can't take it in. That's not an argument for asking your partner to try harder. It's an argument for working on the receiving end, which is the part therapy actually targets.

Does Poor Sleep Make Relationship Anxiety Worse?

Yes, and it's one of the few inputs here you can change this week. The section above treats bad sleep as a symptom of relationship anxiety. It's also a cause. That arrow runs both ways, and the return direction is better evidenced than most people realise.

Start with the brain. Yoo, Gujar, Hu, Jolesz and Walker scanned sleep-deprived people and published the result in Current Biology (2007) under a title that gives away the finding: the human emotional brain without sleep, a prefrontal amygdala disconnect. After a night without sleep, amygdala reactivity to negative images was about 60 percent greater than in rested controls. And the connection between the amygdala and the medial prefrontal cortex, the region that normally applies the brakes, weakened.

So the threat detector gets louder and the part of you that says "hang on, that's probably nothing" gets quieter. If you've noticed that a slightly flat reply from your partner is survivable on Tuesday and unbearable on Thursday, that's not you being inconsistent. Different hardware settings, same message.

Then there's what happens between two people. Gordon and Chen ran a 14-day daily diary study plus a lab study of couples, published in Social Psychological and Personality Science (2014). People reported more conflict in their relationships after poor nights of sleep. In the lab, when one partner had slept badly, the couple showed a lower ratio of positive to negative feeling during a conflict conversation, and empathic accuracy dropped for both of them, not just the tired one.

That last detail matters. Your bad night degrades your partner's ability to read you accurately. So you're less able to interpret them generously, and they're less able to interpret you correctly, at the same time. The researchers also found conflict actually got resolved most often when both partners were rested.

The part that makes this genuinely useful is what the effect survived. Gordon and Chen report it wasn't explained away by stress, anxiety, depression, low relationship satisfaction, or by the partner being the reason for the poor sleep. Sleep was doing its own work.

What to actually do with this:

• Stop having the conversation at midnight. The talk that feels urgent at 1am is the same talk at 10am with better equipment running it. Almost nothing in a relationship genuinely needs resolving before sleep.
• Treat a bad night as a known distortion. Not "ignore your feelings," but "I slept four hours, so my read on this is running hot." You'd apply the same discount to a judgement made while drunk or starving.
• Look at sleep before you look for reassurance. When a spike arrives, check the last two nights first. If they were bad, that's a live explanation worth ruling out before you go looking for evidence about the relationship.
• Protect their sleep too, not just yours. The lab data says one person's bad night costs the couple. Late-night processing sessions tax the person you're trying to feel closer to.
• Watch the loop. Anxiety wrecks sleep, wrecked sleep amplifies anxiety, and round it goes. If you can't break in at the anxiety end, breaking in at the sleep end works too.

None of this means your worries are only tiredness. It means tiredness sets the volume. Fix the volume first, then see what the worry actually sounds like.

Does Relationship Anxiety Look Different If You Are Neurodivergent?

Often yes, and a lot of standard advice quietly assumes a nervous system that isn't yours.

People with ADHD in particular tend to describe something more intense than ordinary worry. It gets called rejection sensitive dysphoria, usually shortened to RSD. Worth being straight about the label first. RSD is not a formal diagnosis, it doesn't appear in the DSM, and it's a descriptive term that emerged from clinical practice rather than from a validated measure. That doesn't mean the experience isn't real. It means you should treat the name as shorthand rather than as a finding.

The research that does exist is worth reading carefully. Rowney-Smith, Sutton, Quadt and Eccles ran focus groups with undergraduates who had ADHD diagnoses and published a thematic analysis in PLOS One. It's a small qualitative study, five participants, so treat it as a description of experience rather than as a measurement of how common anything is. Three themes came out of it: withdrawal, masking, and bodily sensations.

One participant's description of waiting on a reply is the most relationship-relevant thing in the paper. They reported being awake until four in the morning, constantly checking their phone, wondering why the person had not responded. If you've done that, you'll recognise that it isn't really a decision you make.

Two findings from that study change how the usual advice lands.

The first is anticipation. Participants described a general baseline expectation of being rejected, and said the anticipating was worse than the actual rejection. Advice built around coping with a bad outcome misses that, because the damage is happening before anything has occurred.

The second is masking, and it's the one that matters most for a partner. Participants described putting on a mask of toughness, and one described losing track of their own state entirely, saying that at some point you're masking with yourself and can no longer identify whether you're bothered. So the partner sees someone who seems fine, and the person masking cannot report accurately on themselves either. Nobody in the room has good information.

What to do differently if this is you:

• Externalise the check. If you can't read your own state reliably, a written note beats introspection. Two lines at the time, not a reconstruction at midnight.
• Treat the anticipation as the target. Not the imagined rejection. The waiting is the part doing the damage, so that's what needs a plan.
• Say the thing about masking out loud once. Telling a partner that you go quiet and composed when you're most distressed gives them a signal they'd otherwise read as contentment.
• Be sceptical of advice that assumes a steady baseline. Guidance to simply sit with discomfort assumes a system that settles on its own. If yours doesn't, that isn't a failure of effort.

If this reads like your experience and you have never been assessed, that is worth raising with a professional rather than self-labelling from articles, this one included. A therapist who understands ADHD will treat the emotional intensity as part of the picture instead of as an overreaction to be managed. Online therapy is one way to start that conversation without a long wait. And our guide on how to stop overthinking in relationships covers the rumination side, which is the piece that tends to run hardest at 2am.

How Does Relationship Anxiety Affect Your Partner?

Most people carrying this worry about it constantly, usually in the form of "am I too much." So here's the actual picture, which is more nuanced than the fear suggests.

Li and Chan pooled 73 studies covering 118 independent samples and 21,602 people for a meta-analysis in European Journal of Social Psychology (2012, Vol. 42, Issue 4). Both anxious and avoidant attachment were linked to worse relationship quality across cognitive, emotional and behavioural measures. But the sizes weren't equal. Avoidance was more strongly tied to lower general satisfaction, less connectedness, and less support than anxiety was.

That's worth sitting with if you've decided you're the problem in your relationship. The pattern that does more damage to satisfaction, on average, is the one that pulls away, not the one that reaches out too hard. Anxious partners tend to be highly invested, and that investment counts for something.

A larger and more recent pooling adds the piece most people get backwards.

Your attachment anxiety costs you more than it costs them. That's not permission to ignore the effect on your partner, and it isn't nothing for them either. But if you've been running a private narrative where you're ruining someone else's life, 132 studies say the person paying the highest price is you.

What does cost you is the behaviour the anxiety produces, not the anxiety itself. The checking. The testing. The questions asked in a tone that makes any answer sound wrong. Partners rarely withdraw because you feel anxious. They withdraw because they feel monitored, or because they've run out of ways to prove something that can't be proven.

And that's where it turns into a loop with two people in it. You reach, they pull back a little, the pulling back confirms exactly what you feared, so you reach harder. Neither of you is acting in bad faith. The dynamic just runs itself once it starts, which is why our guide to the anxious avoidant trap is worth a read if that shape sounds like your relationship.

The useful move here isn't apologising for having needs. It's separating the need from the delivery. "I'd love some time with you this week" and "you never make time for me anymore" come from the same place and get very different results.

What Can Your Partner Actually Do to Help?

If you're the partner reading this, you've probably worked out that more reassurance isn't landing. The research above explains why. So here's what does appear to move the needle, and it's less effortful than what you're likely doing now.

Attachment researchers have spent two decades testing something called security priming. The idea is simple: deliberately bring to mind an experience of being cared for reliably, and see whether the attachment system settles. Hudson and Fraley ran a repeated version of this over four months and published it in the Journal of Research in Personality (2018, Vol. 74, pages 147 to 157). Compared with a control group, repeated priming raised security over time and lowered trait attachment anxiety.

Here's the honest wrinkle, and it's the interesting bit. Priming people with anxious memories worked about as well as priming them with secure ones. The authors' reading is that the active ingredient may be reflecting on close relationships at all, rather than the specific flavour of the memory. So the mechanism is less magic than the headline suggests. But the direction holds: attention to relationships, repeated over months, shifts the setting.

There's a second line of research that points the same way, and it produced a result nobody expected.

That's a strange and slightly uncomfortable finding, so it's worth being clear about what it does and doesn't say. It isn't a licence to outsource the work. It's evidence that in a two-person system, steadiness is contagious in a specific direction. The calmer person sets the temperature more than the anxious person can, which means a partner who stays present during a spike is doing something measurable rather than just being patient.

It also explains why "just relax" has never once worked. The anxious person can't reach their own brake from inside a spike. The other person can, by not escalating.

What that translates to in practice, for the partner:

• Be predictable rather than intense. A reliable text at a predictable time does more than a grand gesture on a random Tuesday. The system updates on repetition, not on volume. This is the single highest-value thing you can do and it costs almost nothing.
• Say when you're coming back. Most spikes happen in gaps. "I'm heading into a long meeting, I'll call at six" removes the ambiguity that the anxiety would otherwise fill in. Then actually call at six.
• Answer the fear, not the question. "Are we okay?" isn't really asking for a yes. Try "we're okay, and I can tell you're anxious tonight, do you want to talk about that instead." It interrupts the loop rather than feeding it.
• Don't take over the management of it. Becoming the sole regulator of someone else's nervous system burns out fast and quietly teaches them they can't do it alone. Support the work, don't do the work.
• Name the withdrawal if it happens. If you need space, say that it's about capacity and not about them, and say when you're back. Unexplained distance is the exact input this pattern is worst at handling.

And one thing to skip. Don't try to argue someone out of the anxiety with evidence. You'll lose, not because you're wrong but because the doubt regenerates faster than you can address it. Steadiness beats persuasion here every time. Our guide to communicating needs covers the wider version of that conversation.

Which Habits Keep the Loop Running for Both of You?

Three of them, and reassurance seeking is only the one everybody talks about. The other two are quieter, which is exactly why they last so long.

Christine Paprocki and Donald Baucom treated relationship-based anxiety as a set of behaviours rather than a feeling, and named three that tend to travel together. Their paper is in Family Process (2017, Vol. 56, Issue 1, pages 45 to 58).

Excessive reassurance seeking. You already know this one, and the section above covers why the relief keeps shrinking.

Self silencing. This is the one that catches people off guard. You stop saying what you actually want, because you've already decided you're too much. So you ask about the relationship constantly and almost never ask for anything inside it. That gap sits behind a complaint a lot of partners make badly: you seem anxious all the time, but I never find out what you actually need. The worry is loud. The request underneath it never gets made.

Partner accommodation. Your partner starts rearranging their life around the fear. They stop mentioning the friend you're funny about. They skip the work drinks. They text more than they'd choose to. Every piece of it is kind, and all of it quietly teaches the system that the danger was real enough to need managing. Accommodation is love doing the wrong job.

Can one conversation really shift it?

They built a single psychoeducational session aimed at those three habits and ran it with 21 couples. Reassurance seeking dropped. Self silencing dropped. And maladaptive accommodation dropped in the partners, which is the part worth noticing, because nobody had asked the partners to work on themselves.

Keep the size of it in view, though. Twenty one couples is small, every couple got the session so there was no comparison group, and the paper doesn't tell you where those numbers sat a year later. This isn't a cure with a receipt on it.

What it does give you is a better map. If you've been treating this as one bad habit you need willpower to stop, you've been working on a third of the problem. And the two habits you probably haven't named are the two that don't require anyone to win an argument about whether the fear is reasonable. You can start saying what you want. Your partner can stop rearranging. Neither of those has to wait for the anxiety to be gone first.

It's also the shape of work a couples therapist does, so if this loop is well established, going in together tends to be more honest than framing it as one person's thing to fix. If a local waiting list is the barrier, online therapy with licensed therapists is the lower friction way in.

What Actually Helps Relationship Anxiety?

Several approaches have real evidence behind them, and they work on different parts of the problem.

Cognitive Behavioural Therapy targets the thinking. Catastrophising, mind reading, all-or-nothing conclusions. CBT teaches you to catch a thought like "they've gone quiet, they're done with me" and check it against what you actually know. Reassurance seeking shows up across the anxiety disorders, and reducing it during CBT is associated with better outcomes. How strong is the evidence, honestly: Bhattacharya and colleagues, including Stefan Hofmann, pooled ten randomised placebo-controlled trials covering 1,250 people in Current Psychiatry Reports in 2023 and found CBT beat placebo on target symptoms with a Hedges' g of 0.24. That's a small effect, and worth saying plainly rather than dressing up. But note what it was measured against. Placebo controls are a much harder test than a waiting list, and CBT still came out ahead. One more honest caveat: seven of those ten trials were about PTSD, and the authors say themselves that the result is largely driven by PTSD research. So it tells you CBT holds up against a tough comparison in anxiety broadly. It doesn't tell you how big the effect is for the specific worry this article is about.

Emotionally Focused Therapy works on the attachment dynamic underneath. Developed by Sue Johnson, EFT helps couples name what they're actually afraid of rather than expressing it through protest or withdrawal. Beasley and Ager reviewed nine randomised controlled trials in the Journal of Evidence-Based Social Work (2019) and found a Hedge's g of 2.09 for improvement in relationship satisfaction, with gains holding at follow-up. That's a very large effect, and it's the number you'll see quoted everywhere.

But a stricter read is worth having. Rathgeber and colleagues ran a randomised-trial-only meta-analysis in the Journal of Marital and Family Therapy (2019), pooling 33 studies and 2,730 participants. They got g = 0.73 for EFCT at post-treatment and g = 0.66 six months out. Solid, useful, and nowhere near 2.09. The pooled gains across couple therapies had faded by twelve months, though the only twelve-month data came from behavioural couple therapy, so nobody has clean long-run numbers for EFT specifically.

Read the gap between those two figures as a lesson in how to hear therapy statistics. Tighter inclusion rules give smaller, more trustworthy effects. EFT still works. Just expect something closer to a real improvement than a transformation, and expect to need a refresher at some point.

Regulation skills handle the body. Learning your early physical signals, the tight chest or restless hands that arrive before the thoughts do, gives you a window to act before the spiral starts. Grounding practices work here, and they don't require your partner to do anything.

Two practical things worth trying on your own:

1. Delay the ask. When the urge to seek reassurance hits, wait twenty minutes before acting. Not forever, just twenty minutes. Most spikes drop on their own, and each time you ride one out you're teaching your system it can.
2. Say the fear instead of the accusation. "I'm feeling anxious and I think it's mine, not yours" gets a completely different response than "you've been distant lately." One invites closeness, the other invites defence.

These work best combined. Thinking, attachment, and body are the same problem viewed from three angles.

Does Self-Compassion Actually Work Here?

Yes, and there's a good reason it fits this particular problem better than most advice does. It works on the receiving end, which is exactly where the research says the blockage sits.

Think back to the Stanton and Campbell finding earlier. Support from a partner didn't protect the health of people high in attachment anxiety, even when they perceived that support as high. Something was stopping the good input from being absorbed. That something is usually the voice running underneath the anxiety, and it's rarely kind. "I'm too much." "They're going to get tired of me." "Why can't I just be normal about this." Reassurance from outside has to get past all of that before it lands, and mostly it doesn't.

The link is more than a nice theory. Bolt, Jones, Rudaz, Ledermann and Irons surveyed 342 people for a study in the Journal of Relationships Research (2019, Vol. 10) and found that low self-compassion mediated the path from attachment anxiety to poorer relationship quality. Avoidance ran through a different route entirely, low compassion toward the partner rather than toward the self. Two insecure patterns, two different missing ingredients.

Worth flagging the limit the authors themselves report. The mediation showed up on relationship quality measures but not on relationship satisfaction. So this is a real pathway, not a cure-all, and it's honest to say so.

Compare those numbers to the CBT figure from the section above, where a placebo-controlled pooling landed at g = 0.24. Different comparisons and different populations, so it isn't a fair head to head. But self-compassion training isn't a soft add-on to the real work. It holds its own.

Now the correction most people need before trying it, because self-compassion has an image problem. It isn't letting yourself off the hook, and it isn't telling yourself you're great. Kristin Neff, who built the standard measure, frames it as three parts: treating yourself kindly instead of harshly, recognising that struggle is a shared human thing rather than a personal defect, and holding painful feelings in awareness without either suppressing them or drowning in them. That middle piece matters here more than people expect. A lot of relationship anxiety comes bundled with the belief that nobody else is this needy, which is measurably untrue when roughly one in five adults carries the same pattern.

What it looks like in practice, during an actual spike:

• Talk to yourself the way you'd talk to a friend in the same state. Not cheerfully. Just without contempt. Most people are far harsher to themselves than they'd ever be to someone they love, and they don't notice the gap until they look for it.
• Name it as anxiety, not as evidence. "This is my attachment system firing" is a different sentence from "something is wrong." One creates a bit of distance. The other closes it.
• Drop the second layer. Being anxious is uncomfortable. Being ashamed of being anxious is the part that turns twenty minutes into a whole evening. You can usually put the second one down even when the first one won't budge.
• Give yourself the reassurance you were about to request. Not as a trick to avoid your partner. As practice at being a source of steadiness rather than only a recipient of it.

And the honest caveat. If self-criticism is severe, self-compassion work can feel awful at first, and that reaction is common enough to have a name in the literature. Backdraft. It usually means you've found the right spot rather than the wrong one, but it's a good reason to do this alongside a therapist rather than alone. Our anxious attachment healing guide covers the longer version of this work.

Can Relationship Anxiety Actually Change?

Yes. And the long-range evidence on this is more encouraging than almost anything else in this article, which is why it's worth ending on rather than burying.

The fear underneath most of this is that you're stuck. That anxious attachment is a permanent setting, installed in childhood, and the best you can hope for is managing it forever. That's not what the longitudinal data shows.

Fifty-nine years of data, and the direction of travel is toward security. Not away from it.

Now the honest caveats, because this finding gets oversold. These are mean-level changes across a population, so they describe a trend rather than a promise about you specifically. The timescale is decades, not months. And the bit about relationships lowering anxiety is a correlation, not a prescription. Being in a relationship doesn't automatically calm the system, as anyone reading this from inside a perfectly good relationship already knows.

What that finding probably reflects is accumulated evidence. Years of someone staying, being reliable, and not confirming the thing you brace for. Your nervous system updates slowly, and it updates on repetition rather than on reassurance. Which is the same lesson the reassurance research delivers from the other direction.

The practical read is this. The slow drift toward security happens on its own, roughly, over a very long time. Therapy is what compresses it. That's the actual argument for getting help, and it's a better one than "you're broken and need fixing." You're not going the wrong way. You're going the right way slowly, and there are ways to speed that up.

What If the Relationship Actually Ends?

This is the question underneath most of the worry, and it usually goes unasked. You're not really lying awake wondering whether they'll leave. You're wondering whether you'd survive it.

So here's the honest answer. Breakups hit anxiously attached people harder, and the research is specific about why.

Gehl, Brassard and colleagues followed 196 young adults through a breakup and published the results in Emerging Adulthood (2024). They measured people at one month and again at three months afterwards. Higher attachment anxiety predicted more depressive and anxiety symptoms at both points. No surprise there.

The interesting part is what carried the effect. It wasn't the attachment style doing the damage directly. It ran through two specific coping habits.

• Self-punishment. Turning on yourself. Replaying what you did wrong, deciding you were too much, treating the ending as a verdict on you. This pathway linked attachment anxiety to both depression (beta = .110 at one month, .095 at three) and anxiety symptoms (beta = .099 and .079).
• Low accommodation. Struggling to adjust, reframe, or let the situation be what it is. This one linked to depressive symptoms (beta = .056 and .048) but not to anxiety.

Their model accounted for 49.7% of depressive symptoms and 51.7% of anxiety symptoms at the one month mark. That's a lot of the picture explained by how someone copes rather than who they are.

And that distinction matters more than it might sound. Coping habits are trainable. Self-punishment is a response you can catch and interrupt, and it's exactly the sort of thing attachment recovery work targets. Your attachment history sets the starting difficulty. It doesn't write the ending.

One more thing worth saying plainly. The fear of not surviving a breakup is itself part of the anxiety, and it tends to be wrong. People underestimate how well they'll cope, and that mistake has a name in the research literature: the impact bias. You are almost certainly more durable than the 1am version of you believes.

When Should You Talk to a Therapist?

Self-guided work genuinely helps a lot of people, and there's no need to gatekeep it. But some signs point toward getting support rather than pushing harder on your own.

• The anxiety is affecting your sleep, work, or health.
• You're checking your partner's phone, location, or social media.
• Your partner has said the reassurance seeking is wearing them down.
• You've ended good relationships before because the anxiety got unbearable.
• The worry has an intrusive, compulsive quality, which can point toward relationship OCD and responds to a specific treatment approach.

A therapist trained in CBT, EFT, or attachment-focused work will have a structure for this instead of improvising. If cost or waiting lists are the barrier, and for plenty of people they are, online CBT therapy with licensed therapists is a reasonable route in.

If you want a clearer read on your own pattern first, the free relationship anxiety quiz and attachment style quiz are good starting points, and anxious attachment recovery covers the longer arc of the work.

What Else Do People Ask About Relationship Anxiety?

Sources: National Institute of Mental Health, Any Anxiety Disorder statistics (National Comorbidity Survey Replication). Hazan, C. and Shaver, P. (1987), Romantic love conceptualized as an attachment process, Journal of Personality and Social Psychology 52(3). Evraire, L.E., Dozois, D.J.A. and Wilde, J. (2022), The Contribution of Attachment Styles and Reassurance Seeking to Trust in Romantic Couples, Europe's Journal of Psychology. Simpson, J.A. and Rholes, W.S. (2017), Adult Attachment, Stress, and Romantic Relationships, Current Opinion in Psychology. Li, T. and Chan, D.K.-S. (2012), How anxious and avoidant attachment affect romantic relationship quality differently: A meta-analytic review, European Journal of Social Psychology 42(4), 406-419. Stanton, S.C.E. and Campbell, L. (2014), Perceived Social Support Moderates the Link between Attachment Anxiety and Health Outcomes, PLoS One 9(4). Beasley, C.C. and Ager, R. (2019), Emotionally Focused Couples Therapy: A Systematic Review of Its Effectiveness over the past 19 Years, Journal of Evidence-Based Social Work. Doron, G., Derby, D., Szepsenwol, O., Nahaloni, E. and Moulding, R. (2016), Relationship Obsessive-Compulsive Disorder: Interference, Symptoms, and Maladaptive Beliefs, Frontiers in Psychiatry 7:58. Candel, O.-S. and Turliuc, M.N. (2019), Insecure attachment and relationship satisfaction: A meta-analysis of actor and partner associations, Personality and Individual Differences 147, 190-199. Chopik, W.J., Edelstein, R.S. and Grimm, K.J. (2019), Longitudinal Changes in Attachment Orientation over a 59-Year Period, Journal of Personality and Social Psychology 116(4), 598-611. Métellus, S., Vaillancourt-Morel, M.-P., Brassard, A. and Daspe, M.-È. (2025), Attachment Anxiety and Relationship Satisfaction in the Digital Era: The Contribution of Social Media Jealousy and Electronic Partner Surveillance, Journal of Marital and Family Therapy. Paprocki, C.M. and Baucom, D.H. (2017), Worried About Us: Evaluating an Intervention for Relationship-Based Anxiety, Family Process 56(1), 45-58. Mackinnon, S.P., Sherry, S.B., Antony, M.M., Stewart, S.H., Sherry, D.L. and Hartling, N. (2012), Caught in a bad romance: Perfectionism, conflict, and depression in romantic relationships, Journal of Family Psychology 26(2), 215-225. Hudson, N.W. and Fraley, R.C. (2018), Moving toward greater security: The effects of repeatedly priming attachment security and anxiety, Journal of Research in Personality 74, 147-157. Gazder, T. and Stanton, S.C.E. (2020), Partners' Relationship Mindfulness Promotes Better Daily Relationship Behaviours for Insecurely Attached Individuals, International Journal of Environmental Research and Public Health 17(19). Bolt, O.C., Jones, F.W., Rudaz, M., Ledermann, T. and Irons, C. (2019), Self-Compassion and Compassion Towards One's Partner Mediate the Negative Association Between Insecure Attachment and Relationship Quality, Journal of Relationships Research 10. Ferrari, M., Hunt, C., Harrysunker, A., Abbott, M.J., Beath, A.P. and Einstein, D.A. (2019), Self-Compassion Interventions and Psychosocial Outcomes: a Meta-Analysis of RCTs, Mindfulness 10, 1455-1473. Journal of Contextual Behavioral Science (2023), Characterizing the effects of self-compassion interventions on anxiety: Meta-analytic evidence from randomized controlled studies. Neff, K., Self-Compassion: Theory, Method, Research, and Intervention, Annual Review of Psychology. Johnson, S., Hold Me Tight. Gehl, K., Brassard, A., Dugal, C., Lefebvre, A.-A., Daigneault, I., Francoeur, A. and Lecomte, T. (2024), Attachment and Breakup Distress: The Mediating Role of Coping Strategies, Emerging Adulthood 12(1). Rose, A.J. (2002), Co-rumination in the friendships of girls and boys, Child Development 73(6). Starr, L.R. (2015), When Support Seeking Backfires: Co-Rumination, Excessive Reassurance Seeking, and Depressed Mood in the Daily Lives of Young Adults, Journal of Social and Clinical Psychology 34(5), 436-457. Denson, T.F., Moulds, M.L. and Grisham, J.R. (2026), Investigating links between co-rumination and personality, social functioning, and emotional well-being in a representative sample of adults, PLOS ONE 21(5), e0349391.