Can Your Attachment Style Change? What Longitudinal Research Says

The honest answer to "can my attachment style change?" is yes — with caveats. The popular framing on the question has swung in both directions over the last decade. One camp says attachment is fixed in childhood and adults are stuck with whatever they wired up at age four. Another camp says attachment is endlessly fluid and a six-week course will fix you. Neither is accurate. The peer-reviewed longitudinal data, which now spans 30+ years across multiple cohorts, sits in between — and the in-between is more useful and more hopeful than either extreme.

This piece walks through what the longitudinal research actually shows about adult attachment stability and change. The core sources are R. Chris Fraley's stability work and meta-analyses, Glenn Roisman's longitudinal studies on adult attachment continuity, the 20-year Minnesota Longitudinal Study of Risk and Adaptation led by Alan Sroufe and colleagues, the Berkeley Adult Attachment Interview studies originating with Mary Main, and the wider Bowlby–Ainsworth–Hazan–Shaver tradition that the field rests on. Where I use a specific number, it comes from one of those sources rather than from popular summaries. Where the evidence is mixed, I say so. (See attachment theory in dating for the wider primer.)

What the Headline Numbers Actually Say

The right way to read these numbers: attachment is a stable trait — more stable than mood, less stable than personality — that nonetheless shifts in a meaningful fraction of adults across normal life. The 25–30% number is striking because it implies that across a typical decade, roughly half to two-thirds of adults will have moved category at least once. The directions of those moves are roughly balanced — secure-to-insecure happens; insecure-to-secure happens. The thing that determines the direction is mostly what happened in between.

What Actually Drives Change

Across the longitudinal samples, four factors consistently emerge as the strongest predictors of attachment-style movement.

1. A current relationship with a securely-functioning partner

This is the single most-replicated predictor of insecure-to-secure movement in adults. Hazan and Shaver's original 1987 paper hinted at it; subsequent work by Phillip Shaver, Mario Mikulincer and others has replicated it; clinical literatures including Sue Johnson's EFT and Stan Tatkin's psychobiological approach are built on it. Sustained exposure to a partner who is reliably available, responsive, low in defensiveness, and capable of repair — across years rather than months — is one of the most powerful corrective experiences adult attachment can have. (See secure-functioning couples and secure attachment in love.)

2. Coherent autobiographical narrative

One of the more counter-intuitive findings from the Adult Attachment Interview tradition — Mary Main's work and subsequent replications — is that what most strongly predicts adult attachment classification is not what happened in childhood but how coherently the adult can talk about what happened. Adults with rough childhoods who can speak about them in a balanced, integrated, non-defensive way often qualify as secure on the AAI. Adults with relatively benign childhoods who cannot speak about them coherently often qualify as insecure. The implication is that the work of becoming earned-secure is partly a work of making the story of your own attachment coherent — which is one of the things good psychotherapy does best. (See becoming securely attached.)

3. Major life events

Longitudinal samples show that attachment is more likely to shift around major life transitions: starting or ending a long-term relationship, becoming a parent, bereavement, serious illness, redundancy. These events stress-test the existing attachment patterns and produce either confirmation or recalibration. The direction of movement depends on how the support system held during the transition — a partner, friend network, or therapist who held well tends to nudge toward security; isolation during a difficult transition tends to nudge toward insecurity. (See dating after a long marriage.)

4. Therapy — done long enough

The meta-analytic evidence on therapy and attachment change is real but specific. Attachment-informed therapies — EFT for couples, mentalisation-based therapy, schema therapy, and longer-term psychodynamic work — show measurable shifts in attachment-related measures across treatments lasting 6+ months. Short-term therapies show much smaller effects. The implication is not that all therapy is useless; it's that attachment change is one of the deeper kinds of change therapy can produce, and it takes time. The "six weekend workshops to fix your attachment" framing is generally overstated.

What Doesn't Reliably Change Attachment

The same body of evidence is reasonably clear about what doesn't move the needle, despite popular framing.

The Direction of Change — Toward and Away From Secure

The pop-psychology framing usually focuses on the insecure-to-secure direction. The longitudinal data show that secure-to-insecure also happens — and is informative.

Insecure-to-secure ("earned secure")

The earned-secure category was first articulated formally in the Adult Attachment Interview research. It describes adults who would have been classified as insecure on standard developmental measures but who, by adulthood, function attachment-wise like adults who were secure all along. The earned-secure trajectory typically involves some combination of long-term relationships with secure partners, sustained therapy, deliberate work on the coherence of one's own attachment narrative, and luck of timing — being in a relatively stable life period during which the corrective experiences could land. (See becoming securely attached.)

Secure-to-insecure (the loss of secure functioning)

Less talked about, equally real. Adults who tested secure earlier in life can shift toward insecure under certain conditions: prolonged exposure to an unsafe or destabilising relationship, untreated trauma layered onto previously good functioning, bereavement without adequate support, sustained financial or health stress, or simply long-term isolation. The shift is usually not catastrophic — secure adults under these conditions often move from secure to insecure-leaning rather than to fully insecure. It is reversible in the same ways the original move was caused: the underlying stressor changes, support improves, and the function returns.

The Distinction Between Style and State

One thing that confuses the question of whether attachment "changes" is that the field uses two slightly different measures. The trait-level measures (used in most longitudinal research) ask about general patterns across relationships. The state-level measures ask about attachment within a specific current relationship. The two move at different rates.

State-level attachment can shift within months as a relationship stabilises or destabilises. Someone who is generally avoidant can feel measurably more secure within a specific relationship that has lasted three years with a reliable partner. Trait-level attachment moves slower — usually across years — and is what longitudinal samples track. When a friend tells you they've become "more secure" in their new relationship, they may be reporting a real state-level shift even if their trait-level pattern hasn't fully moved yet. Both are real; the timescales differ. (See anxious attachment in dating and avoidant attachment.)

A Practical Read on "Working On Your Attachment Style"

If You're Insecurely Attached and Dating

The hopeful version of the picture, accurately stated: insecure attachment in early adulthood is not destiny. The longitudinal data show that roughly half of adults will move attachment classification at least once across a decade, and the moves are reasonably balanced between toward-secure and away-from-secure. The lever you have — what you choose to spend years inside — is one of the most important variables. Choosing a partner whose attachment style supports security, then doing the relational work of repair and check-in across years, is one of the highest-impact things you can do for your own attachment trajectory. (See dating anxiety vs anxious attachment and love vs attachment.)

The cautious version: you cannot rely on the relationship to do all the work. Adults who shift from insecure to earned-secure usually combine the corrective relationship with some form of focused inner work — therapy, narrative-coherence work, group or community support. The combination is what produces durable change. The relationship alone often produces state-level improvement within the relationship but not trait-level improvement across relationships.

Why This Matters for Matching

Because attachment styles can shift, particularly in the direction of "earned secure," we don't treat attachment as fate at LoveCertain. We weight it at 20% of compatibility — alongside values (40%), life stage (25%) and communication style (15%) — but we read it as the current snapshot rather than the final answer. The reason attachment still matters in matching: state-level functioning within a specific relationship is shaped by the fit between the two partners' styles. Two anxious-attached partners can produce a high-conflict relationship; an anxious–avoidant pair can produce the classic pursue-distance cycle; a secure partner paired with an insecure one often (but not always) raises the insecure partner's state-level functioning. We screen for the combinations where the fit is workable and the corrective trajectory is plausible. (See how matching works and compatibility science.)

For an external authoritative primary-source overview of attachment theory and adult attachment, see Simply Psychology's attachment styles overview, drawing on Bowlby, Ainsworth, Hazan and Shaver and the wider tradition.

The Honest Encouragement

If you identify with an insecure style today, you are not stuck with it. You also will not be free of it by next Wednesday. The shape of the change is years rather than weeks, and the lever you have is a combination of corrective relationship, narrative coherence, sustained therapeutic or relational work, and patience with yourself. Most adults who do this work report that the change, when it arrives, arrives less dramatically than they expected — not a single transformation, but a quiet realisation, somewhere across the third or fourth year, that the old patterns no longer have the same grip. The relationship is built in these moments, and the years that produce them.