Healing Anxious Attachment: A Step by Step Guide

You heal anxious attachment step by step: you learn to notice when your attachment system fires, you resist the urge to chase constant reassurance, you build a fuller life outside your relationship, you ask for what you need directly, and you gather steady, corrective experiences over time. If you've ever reread a text twenty times hunting for a hidden meaning, or felt your stomach drop because a reply came an hour late, you already know anxious attachment from the inside. Here's the good news. It's one of the most changeable patterns in all of relationship psychology, and this guide walks you through exactly how the change happens.

What Is Anxious Attachment, and Can It Be Healed?

Anxious attachment, sometimes called anxious preoccupied attachment, is an insecure relational pattern built around a deep fear of abandonment and a strong hunger for closeness. And yes, it can absolutely be healed. Attachment styles were first mapped in adults by Cindy Hazan and Phillip Shaver, whose landmark 1987 study found that attachment patterns show up in roughly the same proportions in grown-ups as in children. Summarising that work, the University of Illinois notes that about 60 percent of adults classify themselves as secure and around 20 percent as anxious. So if this is you, you're in very large company.

Day to day, anxious attachment feels like a low hum of worry that never fully switches off. You need reassurance that your partner still cares. You read into small shifts in their tone or texting speed. You feel panicky when they want space, and you struggle to feel settled even when things are objectively fine. The good moments bring relief, but the relief rarely lasts. The anxiety just finds something new to grab onto.

The crucial thing to understand is that this isn't a character flaw or neediness. It's a nervous system that learned early on that connection is precious but unreliable, so staying alert felt like the safest way to hold onto it. That's a learned pattern. And anything learned can be relearned.

Why Do You Have Anxious Attachment?

Anxious attachment usually develops in childhood in response to caregiving that was inconsistent rather than absent. A parent who was warm and attuned some of the time, and distracted or overwhelmed at other times, teaches a child a confusing lesson: love is possible, but you can't count on it, so you'd better work hard to keep it close.

A child in that spot learns to amplify their distress, watch the caregiver's mood like a hawk, and stay vigilant for any hint of withdrawal. These were smart adaptations. They helped a small person hold onto connection in an unpredictable home. The trouble is that the nervous system carries those same strategies into adulthood and aims them at romantic partners who might be perfectly steady. The old alarm keeps ringing when there's no real fire.

This pattern is common, and by some measures it's becoming more so. A 2014 cross-temporal meta-analysis by Konrath, Chopik, Hsing and O'Brien in Personality and Social Psychology Review pooled 94 samples of American college students, 25,243 people in total, and found secure attachment fell from 48.98 percent in 1988 to 41.62 percent in 2011. Insecure styles climbed from 51.02 to 58.38 percent across the same window. Later experiences reinforce the pattern too. A painful breakup, a betrayal, or a relationship with an emotionally unavailable partner can deepen anxious tendencies in someone already prone to them. Understanding this shifts the question from "what's wrong with me" to "what did I learn, and what can I learn instead."

Does Anxious Attachment Change on Its Own?

Partly, yes. And that's more encouraging than it sounds, because it means the pattern isn't fixed to begin with. Chopik, Edelstein and Grimm tracked attachment across the longest span anyone has managed, publishing their results in the Journal of Personality and Social Psychology. They pooled three long-running studies, 628 people followed from age 13 to age 72, and found that attachment anxiety declined on average as people got older, with the steepest drop in middle age and later adulthood. Attachment avoidance fell too, in a steady line across the whole lifespan.

One detail in that study is worth pulling out. Being in a romantic relationship was linked to lower anxiety and lower avoidance across adulthood. Which lines up with everything else the research says about how attachment shifts: it moves through experience, not through willpower.

But don't read that as permission to wait it out. These are average trends measured across decades, and averages hide a lot of individual variation. Some people's anxiety softens on its own as their life gets steadier. Others spend thirty years repeating the same painful cycle because nothing in their environment ever contradicted the old expectation. Time on its own doesn't heal anything. What heals is accumulated evidence that connection can be trusted, and you can either wait and hope life hands you that evidence or you can go looking for it deliberately. The steps below are how you go looking.

Does Who You Date Keep the Pattern Alive?

Often, yes. This is the part most healing advice skips, and it's the reason a lot of people do the inner work and still feel stuck. You can practise every skill in this guide and make very little headway if you keep landing in relationships built to confirm your worst fear.

The evidence here is old and still striking. Kirkpatrick and Davis surveyed 354 heterosexual dating couples for a 1994 study in the Journal of Personality and Social Psychology and found that partners weren't paired randomly at all. Across all 354 couples there wasn't a single anxious-anxious pair and not one avoidant-avoidant pair. Anxious people had found avoidant partners, and the other way round.

Then comes the finding that should stop you in your tracks. The couples made up of an avoidant man and an anxious woman stayed together across the full three years of follow-up, and they were surprisingly stable given that both partners had rated the relationship poorly right at the start. Read that again. Three years of a relationship neither person was happy in, and it held.

That's the trap in one sentence. Stability and happiness are not the same thing, and the anxious-avoidant pairing is unusually good at producing the first without the second. It works because the two systems slot together. Their distance triggers your alarm, your pursuit triggers their retreat, and the whole thing keeps running because the intermittent moments of closeness feel enormous against the background of not enough. Your nervous system reads that intensity as love. It's actually just your attachment system being kept switched on.

So what do you do with this? Not swear off avoidant people, which isn't realistic and treats them as the problem. Two things instead. Notice when intensity is doing the work that steadiness should be doing, because the pull you feel toward someone who leaves you guessing is information about your wiring, not about your compatibility. And pay attention to how calm you feel around someone, not just how much you want them. Secure feels boring at first if chaos is what you've been calibrated to. That flatness isn't the absence of chemistry. It's your alarm being quiet, which is the thing you've been trying to achieve. If this dynamic sounds like every relationship you've had, our guide to the anxious avoidant trap breaks the cycle down in more detail.

What Are the Steps to Heal Anxious Attachment?

Healing isn't one big leap. It's a stack of small, repeatable steps that slowly retrain your nervous system. Here are the five that matter most.

1. Learn to Spot Your Activation

Everything starts with awareness. Attachment researchers describe the anxious system as becoming hyperactivated, flooding you with urgency to restore closeness right now. The first skill is simply catching that state as it happens instead of getting swept away by it. The racing thoughts, the tight chest, the compulsion to text or check, the story that you're about to be left. When you can name it, you're no longer fully inside it. Try labelling it out loud: "my attachment system is activated. This is a feeling, not a fact." That tiny bit of distance is where choice lives.

And there's early evidence that training this noticing skill pays off directly. A 2025 systematic map by Gazder, Drummond, Gelegen and Stanton in Behavioral Sciences gathered every study that has tested meditation against adult attachment and found just five. In those five, mindfulness practice was linked to reductions in attachment anxiety, though Kelly and Garland's work found it moved anxiety without shifting attachment avoidance. Five studies is a thin base, so treat this as promising rather than settled. It does match what therapists see in the room, which is that watching your own state without acting on it is a skill you can build.

2. Break the Reassurance Loop

Reassurance feels like the cure, but it's often part of the trap. When anxiety spikes and you seek reassurance, the relief teaches your brain that reassurance is what made you okay. So next time the urge is stronger, and over time the relief gets shorter while the demand grows. The fix isn't to bottle everything up. It's to sometimes let the wave of anxiety rise and fall without immediately acting on it. Each time you tolerate the discomfort without chasing reassurance, you teach your nervous system that the feeling passes on its own. That's how the alarm slowly recalibrates.

But you can't just remove the reassurance and leave a hole. Something has to fill it, and the research points hard at self-compassion. A 2025 meta-analysis by Hill, Vasiliou, Sirois, Hughes and Thompson in Psychology and Psychotherapy: Theory, Research and Practice pooled 37 studies and found self-compassion correlated at r = .40 with secure attachment and r = -.28 with anxious attachment. Put simply, the more anxious people were, the harder they tended to be on themselves. Their conclusion was that compassion-focused therapy may be an effective option for people whose attachment runs anxious or avoidant. In practice this means the voice you use on yourself at 2am, when the text still hasn't come, matters more than you think. Talking to yourself the way you'd talk to a frightened friend isn't a soft extra. It's the thing that makes riding out the wave survivable.

3. Build a Fuller Life Outside the Relationship

Anxious attachment tends to make one person the centre of your whole emotional universe, which makes every interaction feel life or death. One of the most grounding moves you can make is widening your world. Invest in friendships, hobbies, work that means something, and a relationship with yourself that doesn't rise and fall on one person's mood. This isn't about playing it cool as a tactic. It's about genuinely building enough sources of steadiness that no single relationship has to carry the entire weight of your wellbeing.

4. Ask for Needs Directly Instead of Protesting

Anxious attachment often expresses needs sideways, through what therapists call protest behaviour. Going quiet to provoke a response, hinting instead of asking, testing a partner, or escalating distress to get attention. These moves tend to push people away, which confirms the very fear underneath them. The healthier path is direct, vulnerable communication. Saying "I'm feeling anxious right now and some reassurance would really help" is far more likely to bring you closer than withdrawing and hoping to be chased. It takes courage, because it means showing the soft feeling under the protest. But it's one of the most powerful shifts you can make.

5. Gather Corrective Experiences and Support

Attachment changes most powerfully through relationships, both with steady partners and with therapists. Being consistently met with reliability, warmth, and respect slowly rewrites the old expectation that connection can't be trusted. Therapy can speed this up dramatically. Emotionally Focused Therapy, developed by Sue Johnson, has strong evidence behind it. The 1999 meta-analysis by Johnson, Hunsley, Greenberg and Schindler in Clinical Psychology: Science and Practice found that 70 to 73 percent of distressed couples moved into recovery after EFT, with an effect size around 1.3. That finding has held up. A 2019 systematic review by Beasley and Ager in the Journal of Evidence-Based Social Work, pooling nine randomised trials from the 19 years after that first meta-analysis, reported a Hedge's g of 2.09 for improvement in marital satisfaction, with gains still present at follow-up. You don't have to untangle this alone, and the right support is often the difference between insight and lasting change.

One finding is worth knowing because it tells you where to aim your effort. A 2013 study by Benson, Sevier and Christensen in the Journal of Marital and Family Therapy followed 134 married couples through up to 26 sessions of behavioural couple therapy. Early shifts in attachment only hinted at later gains in satisfaction, a trend rather than a solid result. The other direction was strong. Early gains in how satisfied people felt predicted lower attachment anxiety through the end of treatment, and that link was still there at the two year follow-up. The authors' read was that satisfaction mostly shifts first, and attachment follows. So the goal isn't to attack your attachment style head on. It's to build enough genuinely good relational experience that your attachment style quietly updates itself.

What Can You Do in the Moment You're Triggered?

Bring a specific person who makes you feel safe vividly to mind, and hold them there for a couple of minutes. That sounds far too simple to be a real technique. It's one of the most heavily tested interventions in all of attachment research, and it has a name: security priming.

The idea comes straight out of attachment theory. You carry mental representations of the people who've made you feel safe, and those representations can be switched on deliberately rather than waiting for someone to soothe you. Researchers have been testing this in labs for over two decades by having people picture a supportive figure, recall a time they felt cared for, or read words like "love" and "hug" before measuring what happens to their mood, their thinking, and their behaviour.

The numbers are better than you'd guess. Gillath, Karantzas, Romano and Karantzas pooled 120 studies from 97 articles covering 18,949 participants for a 2022 meta-analysis in Personality and Social Psychology Review. The overall effect came out at d = .51. Broken down by outcome, affect showed the strongest response at d = .62, with cognitive outcomes at d = .45 and behavioural outcomes at d = .44. For context, an effect of that size in psychology is a solid, real-world-visible one, not a statistical curiosity.

Here's the detail that matters most if you're anxiously attached. The researchers checked whether people's existing attachment style changed how well priming worked, and for behavioural outcomes it did. Larger effects showed up in people scoring higher on attachment anxiety and avoidance. So this isn't a technique that works best on the people who need it least. If anything, the opposite.

On what to actually do, a 2019 systematic review by Gillath and Karantzas in Current Opinion in Psychology, covering 20 studies, found that the version you consciously do yourself works well, and singled out guided imagery and visualisation as the methods with the most consistent benefits. You don't need a lab or a subliminal word flashing past your eye. You need to close your eyes and picture someone properly.

How that looks in practice when your chest is tight and the reply hasn't come:

• Pick one specific person, not a general idea of support. A grandparent, a friend, an old teacher, a therapist. It doesn't have to be a partner, and it doesn't have to be someone still in your life.
• Go sensory rather than abstract. Their voice, what the room smelled like, where you were sitting, what they actually said. Detail is what makes the representation switch on.
• Give it two or three minutes. This isn't a quick affirmation. The studies had people sit with the imagery properly.
• Use it as a gap-widener, not a cure. Its job is to buy you the few minutes between the spike and the text you'd regret, which is the same gap step one and step two are training.

So what happens if you don't just do it once? Someone has tested that. Zhang, Tao, Chen, Zhang and Liu ran a randomised controlled trial published in Behavioral Sciences in 2023, splitting 56 college students into a priming group of 30 and a waitlist control group of 26. The priming group did seven guided imagination sessions across two weeks, one every other day, each a 17 minute audio recording asking them to imagine themselves in secure relationships and being loved and supported. Their scores on the Interaction Anxiety Scale dropped significantly from before to after. The control group didn't budge.

Two things in that trial are worth taking seriously, and they point in opposite directions. The useful one is the dose. Seven sessions, seventeen minutes, done alone with a recording. That's a fortnight of homework, not a treatment programme, and it was enough to move a real measure. The sobering one is that the researchers also tracked attention bias, meaning whether participants' eyes and focus still snagged on threatening faces, and that didn't shift in either group. So the practice changed how anxious people felt in social interaction without changing the underlying threat radar. Feeling better and being rewired aren't the same result.

Two honest caveats, because the internet version of this research skips both. The effects measured in these studies are short-term by design, so priming is a tool for getting through a bad hour rather than a substitute for the slower work of gathering real corrective experiences. And the meta-analysis authors themselves reported mixed evidence of asymmetry in their data, which is the statistical hint that studies finding nothing may be underrepresented in the literature. The repeated-priming trial is small too, 56 people, and its authors flag plainly that they collected no follow-up data, so nobody knows whether the benefit was still there a month later. Take the direction seriously and hold the exact size loosely.

That follow-up gap has since been closed, and by a trial that also settles a question this guide raises twice. Stevenson, Matthews and Akram randomised 122 UK undergraduates into three arms, published in Applied Psychology: Health and Well-Being (Vol. 18, Issue 4, 2026, e70202). One group did repeated attachment security priming, one did mindfulness practice, one waited. Five consecutive days of digital practice, then measurements taken immediately, a week later, and four weeks later. Both active arms beat the waitlist on stress, insomnia, emotion regulation and even intention to drop out of university. Priming produced the most consistent and substantial benefits across the time points.

That's the first time anyone has run the two main self-guided tools in this guide against each other, and the imagery work edged the meditation. Don't over-read it. These were undergraduates, the practice ran a single week, everything was self-reported, and the headline outcomes were stress and sleep rather than attachment anxiety itself. But it answers the Zhang trial's biggest weakness. The benefit was still measurable four weeks after people stopped practising, which is the first sign that this does something more durable than getting you through a bad hour.

Still, as a thing to reach for at 2am instead of your phone, it's about as well-evidenced as self-help techniques get.

Does Meditation Actually Help Attachment Anxiety?

Yes, and the interesting part is that it seems to help anxiously attached people more than it helps everyone else. Which is the opposite of what most people assume when they're told to try meditating and their first thought is that sitting alone with their own head sounds like a punishment.

The clearest evidence comes from a 2022 study by Taylor West, Brian Don and Barbara Fredrickson, published in the journal Emotion. They followed 113 midlife adults through ten weeks of daily reports, with six of those weeks given over to meditation training, and randomly assigned people to either mindfulness meditation or loving-kindness meditation. Then they tracked emotional patterns day by day rather than just measuring before and after.

People scoring higher on attachment anxiety who got the mindfulness training showed significant increases in positive emotions over the study alongside decreases in negative ones. There was also a clean dose-response pattern at the daily level. On days when participants meditated more than usual, they reported more positive emotion that same day (B = 0.01, p < .001) and fewer negative feelings. More practice, better day. Not dramatic, but consistent.

One finding is worth flagging because it cuts against the obvious advice. The longitudinal benefit showed up for mindfulness meditation and did not show up for loving-kindness meditation. Loving-kindness practice asks you to generate warmth toward yourself and others, which sounds like exactly what an anxiously attached person needs. But if your baseline is already straining toward connection, a practice built on reaching outward may not be the one that settles you. Plain attention training might do more.

And if you're choosing between this and the priming exercise above, the one head-to-head test so far gave priming the edge on stress and sleep over four weeks. That's a single trial on students, so treat it as a tiebreaker rather than a ruling. Doing either beat doing nothing by a wide margin, which is the finding that actually matters.

Why would this work at all? Anxious attachment runs on a body that's braced. Your threat detection is tuned high, so a delayed reply doesn't register as information, it registers as danger, and by the time you've noticed the feeling you're already drafting the message. Mindfulness training doesn't remove the alarm. It buys you the gap between the alarm going off and you acting on it, and that gap is where every other strategy in this guide becomes possible.

Start smaller than you think you need to. Five minutes counts, and the daily-level data suggests consistency matters more than duration. If sitting still makes the spiralling worse rather than better, that's real and fairly common early on, and it usually means starting with something that gives your attention an anchor, like a body scan or focusing on your breath moving, rather than open awareness. UC Berkeley's Greater Good Science Center has a readable piece on prioritising your own needs while healing anxious attachment that pairs well with this, because the practice only sticks once you've accepted you're allowed to take the time.

How Do You Get Through a Breakup With Anxious Attachment?

Badly at first, and then better than most people expect. Breakups hit anxiously attached people harder than almost anyone, and that's measured rather than just felt. But the same research that documents the extra pain also shows where it goes, and it doesn't go nowhere.

Tara Marshall, Kathrine Bejanyan and Nelli Ferenczi ran two studies published in PLOS One in 2013, with 411 people in the first and 465 in the second, all recently out of a relationship. Attachment anxiety predicted heightened breakup distress at a beta of .44 in the first study and .42 in the second. Attachment avoidance ran the other way, at -.19 and -.14. So the ex who seemed fine inside a fortnight probably wasn't a stronger person than you. They were a differently wired one, and the research on avoidance suggests the bill arrives later.

Here's the part nobody mentions. In the same data, attachment anxiety was indirectly linked to more personal growth afterwards, and the route it travelled was the distress itself, at a beta of .06. Avoidance was indirectly linked to less growth, through suppressed distress, at -.03. The second study traced the same thing through rumination. Breakup distress predicted growth through brooding at .26 and through reflection at .14, with anxiety linked to growth via brooding at .12.

Read that carefully before you take the wrong lesson from it. It isn't saying suffering is good for you, and it definitely isn't permission to marinate. It's saying that the thing your brain insists on doing after a breakup, turning it over and refusing to leave it unexamined, is the same machinery that produces the update. Skipping the pain works in the short run and costs you the learning.

The rebound question, answered honestly

The same study looked at the tendency to jump straight into something new, which anxious attachment predicts fairly reliably. Rebounding was indirectly linked to personal growth at a beta of .06, but only among participants whose breakups had happened longer ago. Which is a careful way of saying the growth showed up with distance, not in the rebound itself.

The practical read is not that rebounding is forbidden. It's that a new person arriving three weeks in tends to end the processing rather than support it. Your nervous system reads the fresh attention as the threat being cancelled, the distress drops, and the part of you that was about to learn something goes quiet. If you've noticed you've never actually been single between relationships, that's the pattern our guide on why relationships end keeps running into.

A real warning about journaling

This guide, like most of them, suggests writing things down. That advice needs a caveat, and it comes from a study specifically designed to test it.

David Sbarra, Adriel Boals, Ashley Mason, Grace Larson and Matthias Mehl recruited 90 recently separated adults for Clinical Psychological Science in 2013. Average age 40.4, average relationship length 13.5 years, and on average 3.8 months past the separation. Everyone wrote for 20 minutes on three consecutive days, randomly assigned to traditional expressive writing, a narrative version, or a neutral control task. Then the researchers followed them out to eight months.

The people who were actively searching for meaning in what had happened did worse on expressive writing than on the neutral control task. The effect was a statistical trend at three months and significant by eight, at b = .23, p = .03. Inside the control condition, the meaning-searchers reported significantly less distress than everyone else. The authors put it plainly: writing about an experience while it's still raw and unfolding can intensify the distress rather than settle it.

If you're anxiously attached, you are almost certainly a meaning-searcher. That's what the hyperactivating strategy does. So in the first weeks, use writing for logging rather than excavating. Note what happened and what you felt, then stop. Save the why for later, ideally with someone else in the room. Structured support makes a real difference here, and online therapy is a low-friction way to get it while everything is still loud.

What actually helps in the first month

• Treat contact like a substance, not a decision. Every check of their profile resets the clock on the withdrawal. This is the same reassurance loop covered earlier, just pointed at someone who is no longer there to answer.
• Get the story straight later, not now. Coherence is the thing that predicts security, and you can't build a coherent account of something you're still standing inside.
• Borrow other attachment figures on purpose. Friends, family, a therapist. The anxious instinct after a loss is to funnel everything toward one replacement person. Spread it deliberately instead.
• Expect the timeline to be uneven. Grief for an attachment doesn't fade on a schedule, and a bad Thursday in month three isn't evidence you're going backwards.
• Watch for the returning urge. Wanting them back is not information about whether the relationship was good. It's the hyperactivating system doing what it was trained to do. Our guide on ending things cleanly covers the separating part in more detail.

One last thing worth saying. The growth findings above came from people whose breakups were months and years behind them, not weeks. Nobody in that data grew on schedule, and nothing here is asking you to extract a lesson while you're still bleeding. The lesson tends to arrive on its own, later, usually when you notice you'd handle the same situation differently now.

Can You Heal Anxious Attachment While You're Single?

Yes, and if you've been reading this guide feeling like every instruction assumes a partner, this section is for you. A lot of attachment advice quietly requires someone else to practise on, which leaves single people stuck waiting for the relationship that's supposed to fix them. That's backwards, and it's also the moment the pattern does the most damage, because waiting to be chosen is the anxious script in its purest form.

Here's the honest version. Some of the research does point at partnership. Chopik, Edelstein and Grimm found being in a romantic relationship was linked to lower attachment anxiety across adulthood, and the corrective-experience step in this guide leans on real relational evidence accumulating over time. So no, being single isn't neutral. But look at what actually moved in the studies that measured change directly, and a lot of it needed nobody at all.

The repeated priming trial above is the clearest example. Fifty-six students, a 17 minute audio file, seven sessions alone over two weeks, and interaction anxiety came down. No partner involved. Gillath and Karantzas found the same thing in their review of 20 studies: the version you consciously do yourself works, and guided imagery is the most reliable form of it. And the self-compassion finding is entirely internal by definition. Hill and colleagues pooled 37 studies and found self-compassion correlated at r = -.28 with anxious attachment, which is a relationship between you and your own inner voice, not between you and anyone else.

What that means practically, if you're not seeing anyone right now:

• Your friendships are attachment relationships. Corrective experience means repeated evidence that connection holds. A friend who shows up consistently supplies that evidence. It doesn't only count if you're sleeping with them.
• Practise the in-the-moment tools before you need them. Security priming is easier to reach for at 2am if you've already run it on a calm Tuesday. Skills built under pressure for the first time tend not to hold.
• Watch the pattern in how you date, not just how you partner. The spike when someone takes a day to reply is the same spike. Being single gives you a lower-stakes place to notice it.
• Do the therapy now if you can. Being single is arguably the easier time for it, because you're working on the pattern instead of firefighting a live relationship.

There's a real advantage hiding in this, too. In a relationship, it's genuinely hard to tell your own progress apart from your partner's behaviour. A calm week might be your regulation improving or just them being attentive. Alone, that confound disappears. Whatever steadiness you build is unambiguously yours, and it's still there when the next person shows up. If dating is where this bites hardest, our guide to dating with anxious attachment covers the early stages in more depth.

Does the Story You Tell About Your Childhood Matter?

More than what happened in it. That's not a motivational line, it's the single most replicated finding in adult attachment research, and it's probably the most useful thing in this guide for anyone doing the work alone.

Mary Main and her colleagues at UC Berkeley built the Adult Attachment Interview to test this. It's a structured interview about your childhood, but the scoring ignores the content almost entirely. It doesn't matter whether you describe a warm home or a bleak one. What gets coded is coherence: whether your account hangs together, whether the general claims match the specific memories, whether you can stay with a hard topic without either shutting down or getting swept away.

People who can tell a difficult story coherently score secure. People who had a fine childhood but describe it in vague, idealised, contradictory terms often don't. The story is the marker, not the history.

Which raises the obvious question: can you get there from an insecure start? Glenn Roisman, Elena Padrón, Alan Sroufe and Byron Egeland went looking, in a study published in Child Development in 2002 that drew on a 23 year longitudinal sample, indexed on PubMed. They studied "earned secures," adults who describe negative childhood experiences coherently. The prospectively defined earned secures went on to have successful close relationships in young adulthood. The coherence wasn't cosmetic. It tracked how their actual relationships went.

But the same study is worth reading honestly, because it complicates the tidy version. Retrospective earned secures were not more likely than continuously secure people to have been anxiously attached as infants, and in childhood they'd had some of the most supportive parenting in the sample. So "earned security" is messier than the internet version, where you simply narrate your way out of a bad start. Some of what looks like earning is a hard adult perspective applied to a childhood that had more support in it than the person remembers.

That's not bad news. It means two things worth holding. Adult judgment gets harsher than the evidence sometimes warrants, and how you organise the account is doing real work independent of the facts.

So here's the practice, and it's genuinely free. Write your childhood out. Not a vent, and not a list of grievances. Aim for the thing the AAI actually rewards:

• Match the general to the specific. If you write "my mum was unpredictable," follow it with an actual memory. If no memory comes, that gap is information worth sitting with.
• Hold both sides in one sentence. "She was often overwhelmed and she also sat with me every night I was ill." Coherence lives in the and, not in picking a side.
• Notice where you go vague or flip register. Sudden jokes, sudden formality, sudden fog. Those shifts mark where the story is still costing you something.
• Write it more than once, months apart. Coherence builds by revisiting, which is a lot of what therapy is doing when it circles the same ground.

If your account keeps landing on the same handful of scenes and you can feel it not resolving, that's a reasonable thing to bring to a therapist rather than grind at alone. And if you want the origins mapped out first, our guide on childhood relationship patterns covers where these templates form.

How Do You Know It's Working?

This trips people up, because the early signs of progress don't look like what you'd expect. You assume healing means the anxiety disappears. It usually doesn't, not at first. What changes is your relationship to it.

Here's what actually shifts, roughly in the order it tends to show up.

• The gap gets longer. You still feel the spike, but there are now a few seconds between the feeling and the reaching for your phone. That gap is the whole game. It widens over time.
• Recovery speeds up. A trigger that used to wreck your entire weekend takes an afternoon. Then an hour. The intensity may not drop much before the duration does.
• You catch the story. You notice yourself building a narrative out of a short reply, and you can label it as a story rather than reading it as evidence.
• Protest behaviour drops. You ask instead of testing. You say "I'm anxious" instead of going quiet and waiting to be chased.
• Your world stops shrinking. Plans with friends survive contact with relationship stress instead of getting cancelled.

Notice that none of these require your partner to behave differently. That's deliberate. Progress you can only measure through someone else's reassurance isn't progress, it's just a calmer week. And expect this to move in layers. A stressful period can bring back old reactions, and it doesn't undo the work. You're not back at the start, you're going round the same bend with better brakes.

Is a Bad Week a Setback or Just Normal Variation?

Almost always normal variation, and the research here is unusually reassuring. Attachment anxiety isn't a single number you carry around and slowly file down. It moves within a single day.

Jaakko Tammilehto and colleagues measured exactly how much, in work published in the Journal of Personality in 2024. Two samples of adults, 122 and 127 people, were pinged seven to ten times a day for a week and asked how they felt about closeness right then. That produced 4,629 and 5,322 separate readings.

Your attachment state was different at 11am and at 4pm. Different again after that phone call. So if you've been treating one bad evening as evidence that six months of work came apart, you've been reading a single data point as a trend line.

Trait anxiety does shape where you sit on average. People with higher trait anxiety had higher baseline state anxiety in both samples, at beta = 0.294 and 0.282, both p < 0.001. The pattern is real and it does follow you around. But a baseline is not a fixed point, and it isn't what you're feeling at any given moment.

Then comes the finding that should change how you read a bad week. People with high baseline state anxiety showed, in the authors' words, "highly unpredictable state attachment responses." The unpredictability is part of the pattern itself. So a week that swings around isn't proof the healing failed. It's the thing you're healing, doing what it does.

The usual caveat applies, and the authors raise it themselves. Measuring a moving target is hard, and state attachment anxiety showed modest within-person reliability, with omega coefficients between 0.41 and 0.47. Take the shape of the finding as solid and the precision as rough.

Your bad evening is probably about the evening

The second piece comes from Aleksandra Kaurin, Paul Pilkonis and Aidan Wright, writing in Affective Science in 2022. They had 263 people log 3,971 individual social interactions across a week, recording each one as it happened rather than reconstructing it afterwards.

What they found is that momentary attachment tracked the interpersonal context of each interaction rather than sitting still as a trait. Warmth in the room moved it. Coldness moved it the other way.

Which reframes the thing you do at midnight. When the old panic arrives after a specific conversation, that feeling is partly a reading of that conversation. It isn't a status report on your progress, and it's definitely not a verdict on whether you're fixable.

What to do with a bad week

• Judge over months, not days. The markers in the section above are trend measures. Checking them after a hard weekend is like weighing yourself straight after a big dinner.
• Expect it under load. Stress, illness, bad sleep and a new relationship all lift the baseline. A newer relationship has more room to swing than an established one, because there's less accumulated evidence to reassure you.
• Ask what happened, not what's wrong with you. A spike usually has a cause somewhere in the previous 48 hours. Finding it beats grading yourself.
• Don't restart the programme. The strongest pull after a bad week is to decide the approach isn't working and swap it for a different one. Switching resets your evidence and buys you nothing.

If the bad weeks are stacking up rather than passing, that's a different thing and worth saying plainly. A pattern that keeps deepening despite the work is a reason to bring someone in rather than push harder on your own. Online therapy makes that a smaller step than hunting for a local waiting list, and a therapist can tell normal variation apart from something needing a different approach faster than you can from the inside.

Does Your Phone Keep the Anxiety Going?

Probably more than you'd like to think. Most of the advice on anxious attachment was written for a world where you couldn't see whether someone had read your message, when they were last online, or who they'd just followed. You can see all of it now, and the checking is free, invisible, and available at three in the morning.

Researchers call it interpersonal electronic surveillance, and anxious attachment is one of its strongest predictors. Jesse Fox and Katie Warber studied 328 college students for Cyberpsychology, Behavior, and Social Networking in 2014 and found attachment style predicted how much people monitored a partner online, with preoccupied and fearful styles reporting the highest levels of all. Uncertainty on its own didn't do it. The attachment pattern did.

The longer-term picture is less comfortable. Métellus and colleagues followed 322 adults aged 18 to 29 across two years for the Journal of Marital and Family Therapy in 2025. Social media jealousy predicted more electronic surveillance, and that surveillance predicted lower relationship satisfaction a full year later. Here's the part worth sitting with. Once the researchers accounted for the jealousy and the checking, the direct link between attachment anxiety and satisfaction stopped being significant. The behaviour was carrying the damage, not the trait.

Which is good news, oddly. You can't switch off a nervous system that learned its job in childhood. But you can absolutely change what your thumbs do at 11pm.

The checking feels like information gathering. It isn't. It's reassurance seeking wearing a disguise, and it works exactly the same way as asking "are we okay" for the fourth time. You get twenty seconds of relief, the threshold rises, and next time you need more. Every scroll through their tagged photos teaches your brain that the anxiety was right to fire.

What actually helps is boring and structural. Turn off read receipts and last-seen, on both sides. Mute rather than unfollow, so you're not making a statement you'll have to explain. Put the phone in another room during the hours you tend to spiral, usually late evening. And when you catch yourself mid-check, name it plainly: "that's the checking." No verdict attached. Our guide on how to stop overthinking in relationships goes deeper on interrupting that loop once it's already running.

One caveat. If the urge to monitor is constant, or you're doing it to someone who has actually given you reason not to trust them, that's a different problem and it needs a different conversation. Chronic checking is sometimes anxiety, and sometimes it's an accurate read on a situation that isn't safe.

What Slows the Healing Down?

Most people doing this work aren't failing at it. They're doing four or five things that quietly cancel out the progress. Here are the common ones.

Beating yourself up for being anxious. This is the big one, and it's almost universal. You catch yourself checking their location for the third time and the self-attack starts: pathetic, needy, why can't I just be normal. That attack is not a correction. It's more threat, aimed inward, at a nervous system that is already reading threat everywhere.

The correlational evidence for the alternative is in this guide already. But there's trial evidence too, which is stronger. Collado-Navarro and colleagues, working with Montero-Marin's team, ran a randomised controlled trial published in Depression and Anxiety in 2021, available via Wiley. They took 90 patients across three mental health units in Castellón, Spain, with depressive, anxious and adjustment disorders, and randomised them into attachment-based compassion therapy plus usual treatment, mindfulness-based stress reduction plus usual treatment, or usual treatment alone. Two findings matter here. Compassion therapy reduced affective distress, though it was not superior to plain MBSR. And in the path analysis, self-compassion came out as a significant mediator of the effect. Not a nice side effect. Part of the mechanism.

Practically: when you catch the behaviour, name it without the verdict. "There's the checking" does the job. "I'm pathetic" adds a second problem to the first.

Using insight as a substitute for practice. You can name your triggers, cite your own attachment style, and explain protest behaviour to your friends, while nothing in your actual Tuesday changes. Understanding the pattern is genuinely satisfying, which is exactly why it can become the thing you do instead of the work.

Testing instead of asking. Going quiet to see if they notice. Waiting the same number of hours they waited. Mentioning the ex. Every test you run gives you information you can't trust, because you engineered the situation, and it costs you the thing you actually wanted. Our guide on self sabotage in relationships goes into why the testing feels so necessary in the moment.

Expecting a straight line. A bad fortnight after three good months reads as proof that none of it worked. It isn't. Attachment work moves in layers, and stress reliably brings old responses back to the surface. The measure isn't whether the spike returns. It's how long it now takes you to come down.

Staying somewhere the fear keeps getting confirmed. This one is uncomfortable and it needs saying. You can do every exercise in this guide flawlessly while dating someone whose behaviour genuinely is unpredictable. Then the anxiety isn't a distortion to be regulated, it's an accurate reading. No amount of self-soothing fixes a real inconsistency, and trying to regulate your way into tolerating it will just teach you to distrust your own signal.

When Should You Get Professional Help?

Earlier than most people do. There's a tendency to treat therapy as something you earn by getting bad enough first, and that logic costs people years.

Some clearer signals it's time. The anxiety is affecting your sleep, your work, or your health. You've noticed the same painful pattern across three or more relationships. You're staying in something that hurts because leaving feels unsurvivable. Your reactions are frightening you, or frightening the people you love. Or you understand all of this intellectually and it's changing nothing, which is one of the most common and most frustrating places to get stuck.

That last one deserves a note, because insight and change are genuinely separate things. You can read every book on attachment and still feel your chest lock up when a text goes unanswered. Attachment lives in the body and in relationships, so it tends to shift through relational, experiential work rather than through understanding alone. That's the case for working with someone rather than just reading about it.

Look for a therapist who works with attachment specifically. Emotionally Focused Therapy has the strongest evidence base for couples, and attachment-focused or schema-based approaches are common for individual work. If cost, waiting lists, or just the thought of sitting in a room is the barrier, online therapy lowers it considerably and gives you structured worksheets between sessions, which suits attachment work well.

One thing worth expecting before you start, because nobody warns anxiously attached people about it. The therapy relationship is still a relationship, and your pattern will show up in it.

Notsu, Blansfield, Spina and Levy pooled 33 studies for a 2025 meta-analysis in Psychotherapy Research looking at how attachment insecurity relates to the working alliance, which is the bond and sense of shared purpose between client and therapist. Both forms of insecurity were linked to a slightly weaker alliance: r = -0.09 for attachment anxiety and r = -0.13 for avoidance.

Now look at the size of those numbers, because that's the actual message. Those are small correlations. Anxious attachment nudges the therapy relationship a little, and that's all it does. It is not a reason to think therapy won't work for you, and the same research group's earlier work found insecurely attached people did well in approaches that put relationships at the centre.

What it does mean in practice:

• Don't judge the fit on session one. Early awkwardness or a rush of "they probably think I'm too much" is your pattern activating, not evidence about the therapist.
• Say it out loud when it happens. "I noticed I felt anxious after last session" is not a derailment. For attachment work it's the material, and a good therapist will be glad you brought it.
• Give it a handful of sessions before you decide. The urge to quit early, or to over-please so they keep liking you, are both the pattern. Naming which one you're doing takes most of its power away.

Which Therapy Approaches Actually Change Attachment?

The honest answer is that the specific method matters less than whether it puts relationships at the centre of the work. That's not a dodge. It's what the largest analysis of this question actually found.

Levy, Kivity, Johnson and Gooch pooled 36 studies covering 3,158 patients for a 2018 meta-analysis in the Journal of Clinical Psychology. Two results stand out. People who arrive in therapy already more securely attached tend to get more out of it, which is a bit unfair but worth knowing. The more useful finding is the second one: patients who started with low attachment security did better in therapies that deliberately focused on interpersonal interactions and close relationships. So if you're anxiously attached, an approach that spends its sessions on your actual relationships will probably serve you better than one that treats your symptoms as though they exist in isolation.

One study is worth putting next to that, because it complicates the tidy version. Newman, Castonguay, Jacobson and Moore ran an additive trial published in the Journal of Consulting and Clinical Psychology in 2015. They randomised 83 people with generalised anxiety disorder to CBT plus supportive listening, or CBT plus interpersonal and emotional processing therapy, then tracked them out to two years. On average the interpersonal add-on was the more obvious fit for attachment problems. But for one subgroup it was the worse option. People who came in high on attachment-related anger got less symptom change from the interpersonal version than from plain supportive listening, at B = .010, p = .015, with a small effect size of d = .38.

Two caveats keep that from being a rule. The effect showed up straight after treatment and then disappeared, with no moderation at the 6 month, 12 month or 2 year follow-ups. And attachment anger is a specific thing, not the same as attachment anxiety in general. Still, the practical read is useful. If your anxious attachment runs hot, if what comes up when you're triggered is resentment rather than only fear, then diving straight into intensive relational work may be more than you can use in the first few months. That's an argument for pacing, not for avoiding it.

There's also direct evidence that attachment anxiety shifts during treatment, not just alongside it. The SOPHO-Net trial, reported by Strauß and colleagues in PLoS One in 2018, randomised 495 people with social anxiety disorder into three groups: 209 to cognitive behavioural therapy, 207 to psychodynamic therapy, and 79 to a waiting list. CBT produced a significant reduction in attachment anxiety from before to after treatment, with an effect size of -0.30. The psychodynamic group showed no significant change against the waiting list over the same stretch.

But two details in that trial complicate the tidy version of the story. Both groups kept improving after therapy finished, and by the 12 month follow-up the difference between them had largely closed. So the psychodynamic patients got there, just more slowly. And the effect size is modest, which fits everything else here: attachment moves in increments, not jumps.

One caveat you should hold onto. Those were people diagnosed with social anxiety disorder, not people who simply identify as anxiously attached, so stretching the result to your situation is exactly that, a stretch. What it does establish is the thing that matters most: attachment anxiety is measurably movable inside a course of therapy. It isn't a fixed trait you're stuck negotiating around for life.

Does Couples Therapy Work if You're the Anxious One?

It's the obvious question if you're already with someone, and the research answer is a qualified yes with a genuine catch at the end.

Emotionally focused couples therapy is the modality built directly on attachment theory, so it's the natural fit here. Rathgeber, Bürkner and colleagues pooled 33 randomised controlled trials covering 2,730 participants for a 2019 meta-analysis in the Journal of Marital and Family Therapy. At the end of treatment, emotionally focused therapy produced a medium effect on relationship satisfaction, g = 0.73, against g = 0.53 for behavioural couples therapy. Six months later the gains had shrunk but held, at g = 0.66 and g = 0.35.

Now the catch, and it's a real one. By the twelve month mark the improvement was no longer being maintained. The researchers also found no significant difference between the two approaches once you account for everything else, which undercuts the common claim that the attachment-based option is clearly superior.

So what do you actually do with that? Two things. Couples therapy reliably helps in the window you're in it and shortly after, which is worth a lot if you're currently drowning. But it isn't a one-off repair you complete and file away, and treating it that way is probably why the twelve month numbers sag. The couples who hold their gains tend to be the ones who keep practising the skills after the sessions stop, or who go back for a tune-up rather than waiting for the next crisis.

One more thing worth saying plainly. Couples work is not a substitute for your own. If your partner is steady and you're still checking their face for evidence you're about to be left, that's your nervous system running an old programme, and a joint session is a slow and expensive place to work on it alone. Do both if you can. Individual work on the trigger, couples work on the pattern between you.

Can Group Therapy Help Anxious Attachment?

It's the option almost nobody considers first, and there's a decent argument it should be higher on the list.

Here's the logic. Your pattern doesn't live in your head. It fires in the presence of other people, which is why you can understand it perfectly on a Tuesday afternoon and still spiral on a Friday night. Individual therapy gives you one relationship to practise in. A group gives you six or seven at once, in real time, with a clinician watching how you actually behave rather than how you report behaving. If you go quiet when someone seems to withdraw, or over-explain when you sense disapproval, it happens right there in the room where it can be worked with.

There's early evidence the format moves the needle. Zbigniew Wajda and colleagues published a 2022 study in Frontiers in Psychiatry following 22 patients through ten weeks of online group psychodynamic psychotherapy. Attachment anxiety dropped significantly across treatment with a moderate effect size of -0.557, and the reduction was still there at the six week follow-up. Attachment avoidance fell by a similar margin.

Now the honest caveats, because 22 people is not a lot. There was no control group, the follow-up was short, and the participants were being treated for neurotic, stress-related and depressive conditions rather than recruited as anxiously attached. The authors themselves called the result somewhat surprising, given the relationships were formed entirely over video and the course ran only ten weeks. Treat it as a promising signal, not proof.

What makes it interesting anyway is the direction it points. Ten weeks is short. Online is supposedly the weaker format for building a bond. And attachment anxiety still shifted. That fits the thread running through all of this: the thing that changes attachment is repeated relational experience, and a group is a concentrated dose of exactly that.

A practical note if you try it. The first few sessions will probably feel awful, and that's not a sign it's wrong for you. A room of people whose approval you can't yet read is close to a worst case scenario for an anxious nervous system. The discomfort is the material. If you tend to bail on things right before they get useful, our guide on self sabotage in relationships is worth a read first.

Healing anxious attachment isn't about becoming someone who never feels a flicker of insecurity. It's about building enough inner steadiness that the old fear stops running the show. It takes time, patience, and a lot of self compassion. But the research is clear and genuinely hopeful. This pattern can change, and the relationships on the other side of that change are calmer, warmer, and far more nourishing.