Anxious Attachment in Relationships: Dating, Hard Conversations, and Conflict
By Mark Ryan, BACP Accredited Psychotherapist and NCPS Accredited Relationship Therapist
I've written before about where attachment theory comes from and what the four adult styles broadly look like. This piece skips that ground and goes somewhere more specific, the three places anxious attachment tends to show up and cause difficulty. In who you end up dating in the first place, in the harder conversations that come once you're together, and in outright conflict once things have properly settled in. Knowing the theory rarely changes anything on its own. What tends to help is recognising the pattern in the three moments it's most likely to be running the show.
In finding a partner
This is where anxious attachment often does its quietest and most consequential work, before a relationship has properly started. It's become common shorthand to say anxious people are attracted to avoidant ones, and real-world pairing patterns do skew that way more often than chance would predict. But it's worth being precise about what the research shows, because the popular version overstates the case. When people are asked directly to rate their attraction to hypothetical partners with different attachment styles, anxiously attached individuals don't consistently prefer avoidant partners, in some studies they're notably less drawn to them than secure alternatives. What the data shows more reliably is that anxious and avoidant partners end up paired together in real, ongoing relationships more often than pure chance would suggest, which points to something happening over time, in how the relationship develops, rather than a fixed initial attraction.
A large recent study combining five longitudinal datasets of newlywed couples found that attachment insecurity doesn't just affect a relationship on its own, it interacts with a partner's attachment insecurity in ways that compound over time. Two anxious partners together tended to follow what the researchers called a weak-link pattern, where one partner's anxiety undermines the other's stability rather than the two balancing each other out (Peters, Meltzer and McNulty, 2025). A partner who's calm, consistent, and doesn't create much emotional intensity can, to an anxious nervous system that's used to reading intensity as evidence of caring, register as slightly flat or unconvincing. A partner who runs a little hot and cold, without meaning to, can register as more compelling simply because the nervous system has more practice with that rhythm.
None of this means anxious attachment dooms someone to a string of unavailable partners. It means the early feeling of certainty, this one feels right, is a less reliable narrator than it seems in the moment, particularly early on, and is worth treating as information to sit with rather than a verdict to act on immediately. Slowing down that early certainty by even a few weeks tends to reveal far more than the certainty itself ever does.
A practical version of this is worth naming. Early dating naturally involves gaps in information, a message that takes a day to answer, plans that are loosely made rather than firmly confirmed, a new partner who hasn't yet earned the trust that comes with time. For someone anxiously attached, these ordinary gaps can generate a level of distress that feels proportionate to what's happening but usually isn't, because the nervous system is responding to the general shape of uncertainty rather than to anything the new partner has done. Noticing that distinction, is this feeling about them, or about uncertainty itself, tends to be one of the more useful early habits to build, well before a relationship is established enough to bear the weight of that conversation directly.
There's a further layer worth naming specifically, the tendency to notice, sometimes with real dismay, that this isn't the first time. The distant partner, the one who ran hot and cold, the one who needed more looking after than they gave back, has a familiar shape to it, and it's rarely the first relationship to take that shape. For someone anxiously attached, this repetition usually isn't bad luck or poor judgement. It tends to be the same nervous system, calibrated early to read a particular kind of unpredictability as what closeness feels like, doing what it was built to do, over and over, with different faces attached. I've written more fully about why this repetition happens and what shifts it, including why simply knowing your own pattern rarely changes it on its own. The anxious-specific version of that pattern is usually a pull toward partners who recreate a particular kind of unreliability, because unreliability, for this nervous system, is what got wired up as love in the first place.
In difficult conversations
Once a relationship is established enough to have something to lose, a second layer opens up, the harder conversations that aren't quite conflict. Raising a need, checking in about where the relationship is going, naming that something has felt off lately, these conversations often get either avoided for months or rehearsed obsessively beforehand, sometimes both in sequence. A recent study on how people share difficult news with partners found that attachment anxiety predicts both when people disclose distress and how, with more anxiously attached people disclosing sooner and more urgently, seeking attention and reassurance in the disclosure itself rather than simply passing on information (Snyder, Sun, Nie and Jakubiak, 2024). A conversation that hasn't happened yet can occupy as much nervous system real estate as one that's actively happening.
In practice, this often looks like over-preparing what to say, running the likely responses in advance, and then finding that the partner's actual, real-time reaction, which is rarely as catastrophic as the rehearsed version, still doesn't fully land as reassuring. That's not a failure of the conversation. It's a sign that what's needed isn't one good conversation but a pattern of them, repeated enough times that the nervous system starts updating its prediction, rather than a single decisive exchange doing all the work at once.
A related pattern is difficulty tolerating a conversation that doesn't resolve immediately. Needing to know where things stand right now, rather than being able to sit with a partner who says they need a day to think, is one of the more common and more exhausting versions of this, for both people involved.
There's a specific version worth naming separately, the conversation about the relationship's future, whether that's moving in together, exclusivity, or something further down the line. These conversations carry extra weight for anyone, but for someone anxiously attached they can become almost impossible to raise casually, because the stakes feel high rather than proportionate. A single conversation about pace can feel like it's deciding the entire relationship's fate, which pushes people either toward avoiding the conversation for months, or toward raising it in a moment of heightened anxiety rather than calm, which rarely produces the outcome either partner wants to see.
Attachment insecurity doesn't just affect a relationship on its own, it interacts with a partner's attachment insecurity in ways that compound over time.
In conflict
Conflict is where all of this tends to come to a head. Anxious attachment doesn't just make conflict more emotional. It changes what the conflict is about before either of you has said much at all. A recent cross-cultural study, tracking couples in the UK, China, and New Zealand, found that attachment anxiety is consistently linked to a tendency to interpret a partner's ambiguous or mildly negative behaviour, a short reply, a cancelled plan, a distracted evening, as hostile or threatening, and that this pattern holds across very different cultures (Li, Carnelley, Crawford, Lin and Rowe, 2026). In plain terms, the argument that starts about the washing up is very rarely about the washing up. It's an old, well-rehearsed question, will you stay, arriving dressed as a disagreement about chores.
This shows up in a specific shape. Where a securely attached partner can disagree, feel annoyed, and let the annoyance pass once the issue is addressed, an anxiously attached partner often can't let the conversation end until the underlying fear has been directly soothed, not just the practical issue. Ending a fight before that soothing has happened can feel, physically, like being left mid-freefall, which is part of why the same argument can seem to resolve and then reopen an hour later. The practical issue was settled. The attachment fear underneath it wasn't.
This also changes the shape of an apology. A securely attached partner can usually accept a reasonably good apology and move on. An anxiously attached partner often needs the apology to include something closer to a direct reassurance about the relationship itself, not just the specific behaviour, because the underlying question was never about the behaviour to begin with. Partners who don't know this can end up baffled that a seemingly adequate apology hasn't landed, when what's missing isn't sincerity but scope, the apology addressed the incident and left the fear it triggered untouched.
This is also where protest behaviour tends to intensify, repeated attempts to reconnect, sometimes through pursuit, sometimes through escalation, because disengagement from a partner mid-conflict can register as a genuine threat rather than a normal pause. Jealousy during conflict often has the same root, a specific flavour of the same protest response rather than a separate issue. I've written more generally about breaking cycles like this, and a lot of what's in that piece applies directly here, slowing the moment down enough to notice the fear driving the behaviour, rather than just the behaviour itself.
What a partner's response changes
None of this happens in a vacuum. How a partner responds to anxious attachment in the moment is one of the strongest available levers on whether the pattern intensifies or settles over time. A partner who reads the protest behaviour correctly, as fear rather than manipulation, and responds with something concrete rather than reassurance in the abstract, tends to see the intensity ease over months, not because the anxious partner has fixed themselves, but because the nervous system is slowly collecting evidence that reaching out doesn't end badly.
Concrete tends to work better than abstract here. A vague, don't worry, I love you, is easy to discount in a moment of fear. A specific, I'm not going anywhere, I'll text you when I land, tends to land better, because it gives the anxious system something particular to hold onto rather than something to reinterpret. This isn't about walking on eggshells around a partner's anxiety. It's about understanding that specificity does more regulatory work than general reassurance, for reasons that have more to do with how threat-detection works than with anything either partner is doing wrong. Partners who understand this tend to find the adjustment fairly small in practice, a few more specific words here and there, rather than a wholesale change in how they communicate.
What helps
The work here is split between two things, and both matter. One is individual, learning to notice the physical signature of the fear as it arrives, a tightening in the chest, the urge to check a phone, the pull to send one more message, and building enough of a pause between the urge and the action that a choice becomes possible. The other is relational, learning to name the underlying fear directly, I'm worried this means you're pulling away, rather than performing it through escalating bids for reassurance that a partner often can't correctly interpret in the moment.
Timing matters more than people expect. The self-regulation work is harder to do in the middle of the fear than it sounds when described calmly afterward, which is why it tends to start smaller than people want it to, noticing the pattern after the fact at first, then catching it partway through, and only later managing to interrupt it while it's actively happening. Expecting to master this in the middle of a real moment before you've practised it anywhere else tends to set the work up to fail, and that's a normal part of the process, not a sign it isn't working.
Neither of these happens quickly, and neither happens through insight alone. Knowing why you feel what you feel is a start, not an ending. What shifts the pattern is enough repeated experience of a fear arising, being named, and being met, that the nervous system slowly revises its prediction about what happens when you reach for someone.
How therapy can help
Some of this work happens individually, particularly the self-regulation side, learning to tolerate the physical intensity of the fear without immediately acting on it. Some of it happens in the room with a partner, particularly where the pattern has become a recognisable cycle between you, since the partner's response is very often what determines whether the fear settles or intensifies.
You won't be asked to arrive with a tidy account of your attachment history, and you don't need one. What I'm listening for is the live pattern, what a recent hard moment looked like moment to moment, what you felt in your body before you acted, and what happened in the gap, or the absence of a gap, between the fear arriving and the response leaving your mouth or your thumbs. That gap, however small, is usually where the work starts.
Where a partner is involved, I'm also listening for their side of the cycle, what it's like to be on the receiving end of the protest behaviour, and what tends to happen in them that makes reassurance land or fail to land. Anxious attachment rarely exists in isolation from a partner's own pattern, often an avoidant one, and understanding both halves of the cycle tends to matter more than working on either partner's pattern alone. A partner who feels endlessly blamed for triggering the anxiety, without any acknowledgement of their own part in how they respond to it, tends to disengage from the work, which then reinforces exactly the fear the anxious partner was afraid of in the first place.
Where both partners are willing, working on this together tends to move faster than either partner working on it alone, because the cycle that keeps the fear activated is relational, not purely individual. An anxious partner practising self-regulation in isolation, while their partner remains unaware of what specifically helps or worsens the pattern, is doing half the available work. Bringing both halves into the same room, so each partner understands what's happening in the other during a hard moment, tends to shift things faster than either piece done separately, over time.
Either way, the aim isn't to become a different kind of person, or to stop having needs for closeness and reassurance, which are entirely legitimate needs. It's to find ways of meeting them that don't cost you the relationship you're trying to protect in the first place.
If you feel like now might be the right time to talk this through with a therapist, I offer a free 30-minute consultation in person (in London) or online.
About the Author
Mark Ryan is a BACP Accredited Psychotherapist and NCPS Accredited Relationship Therapist working in person across central London, from rooms in Pimlico, Kensington, and Angel. He works with couples, individuals, and polycules on relationship issues including jealousy, affairs and betrayals, sex problems, and considered separation. His practice is integrative, sex-positive, and affirming of all relationship structures and identities.
References
Peters, S. D., Meltzer, A. L., & McNulty, J. K. (2025). Own and partner attachment insecurity interact to predict marital satisfaction and dissolution. Social Psychological and Personality Science. https://doi.org/10.1177/19485506241237294
Snyder, M. D., Sun, E. R., Nie, J., & Jakubiak, B. K. (2024). When and how do people share? Attachment anxiety predicts support-seeking strategies. Personal Relationships, 31(4), 1113–1132. https://doi.org/10.1111/pere.12573
Li, D., Carnelley, K. B., Crawford, M. T., Lin, X., & Rowe, A. C. (2026). How hostile attribution bias affects attachment anxiety/avoidance and relationship satisfaction: A cross-cultural perspective during the COVID pandemic. The Journal of Social Psychology. https://doi.org/10.1080/00224545.2026.2626036
