Can Attachment Styles Change? What the Research Says

9 minute read

A client sits down for a first session and hands you her framework before you can offer one. She took a quiz, she explains, or watched a thirty-second reel, and she is fearful-avoidant, which explains why her relationships keep collapsing, and what she wants to know is whether therapy can fix that or whether she is simply wired this way. Somewhere between the app and your office, a probabilistic research construct became a personality verdict.

Underneath the label she is asking a real question, and it deserves a real answer. Can attachment styles change? The short version is yes. The longer version is more interesting, because the research suggests the question itself carries a hidden assumption: attachment was never the fixed trait the quizzes imply. Understanding what actually stays stable, what shifts, and what moves it matters for anyone doing trauma work, and it changes the story clients get to tell about themselves.

Where the Four Styles Came From

The styles began as observations of babies. Mary Ainsworth's Strange Situation procedure classified infants by how they used a caregiver for comfort after brief separations, yielding the secure, avoidant, and ambivalent patterns, with Main and Solomon later adding the disorganized category for infants whose behavior would not resolve into any strategy at all. A detail from that original work gets lost in nearly every popular retelling: the classification described a relationship rather than a child. The same infant could be secure with one parent and avoidant with the other, because the pattern captured what that infant had learned to expect from that specific person.

Hazan and Shaver carried the framework into adult life with a deceptively simple study, asking adults to pick which of three paragraphs best described their experience of romantic love. The findings were striking, the measure was easy to administer, and the adult attachment styles were born. Four decades later they have escaped the laboratory entirely. They live in captioned infographics, dating app bios, and the vocabulary clients bring to intake, usually with an unstated assumption attached: that a style is something you have, the way you have a blood type.

What the Boxes Get Wrong

The stability data tell a more layered story than the type framing suggests. Fraley's meta-analysis of longitudinal studies found that attachment security from infancy to adulthood shows modest stability at best, a result more consistent with an early prototype that keeps some influence while leaving substantial room for change. Waters and colleagues followed infants for twenty years and found that most kept their classification, while the ones who changed had usually lived through significant negative life events, which means the changes were lawful rather than random. A later meta-analysis found that stability weakens considerably as the interval between assessments grows. Early experience matters. It also keeps getting revised.

The categories themselves have not survived scrutiny either. Modern measurement finds that adult attachment varies along continuous dimensions of anxiety and avoidance rather than sorting people into discrete types, so most people are moderately something rather than purely anything. And the relationship-specific finding from infancy holds in adulthood: people carry different expectations into different relationships, secure with a best friend, anxious with a partner, avoidant with a parent, because what got learned was a set of predictions about particular kinds of closeness rather than a single global setting.

From Fixed Types to Shifting States

The most clinically useful development in attachment research reframes the whole question. Rather than asking which box a person belongs in, current thinking tracks states: where a nervous system sits right now on a spectrum that runs from shut-down self-protection through regulated openness to pressured vigilance, and how information processing changes at each position. Sue Marriott and Ann Kelley, in their book Secure Relating, describe this as moving along a spectrum with an implicit map, built from relational history, that biases where you tend to sit and how quickly you leave the regulated middle when something threatens. Everyone visits every part of the spectrum. Security, in this framing, is a state you can occupy and return to rather than a club you either joined in infancy or didn't.

That reframe explains things the type model never could. It explains why your most avoidant client has moments of open, undefended contact, and why a secure-seeming person can flood without warning when one particular subject surfaces. It explains why attachment behavior changes with stress, with sleep, with the person across the table, something we watched play out in reading autonomic responses in the couples room. And it dissolves the fatalism clients absorb from the quiz version, because a state you move through is workable in a way a fixed identity never is.

So Can Attachment Styles Change?

They can, and the evidence goes beyond optimism. The clearest demonstration comes from research on earned security: adults who describe genuinely difficult, sometimes harsh childhoods and nonetheless show coherent, secure states of mind when interviewed as adults. Roisman, Padrón, Sroufe, and Egeland documented this prospectively, following people from infancy and finding that earned-secure adults were parenting warmly and functioning well in relationships despite early histories that predicted otherwise. Whatever their beginnings wrote, it was not the final draft.

Change also shows up on ordinary timescales. Attachment anxiety tends to soften across adulthood. Long, stable relationships with reliable partners shift working models slowly in a secure direction. And psychotherapy moves the needle directly: a meta-analysis of 36 studies covering more than 3,000 patients found that attachment security is meaningfully tied to psychotherapy outcome, and that gains in security over the course of treatment tend to travel with better results. None of this happens through insight alone. Maps built from repeated relational experience get revised the same way, by accumulating experiences that contradict the old predictions, inside therapy and outside it.

What Actually Moves the Maps

If patterns change through experience, the clinical task is to become the kind of experience that changes them, and that asks for more than warmth. It asks for consistency a client's nervous system can eventually believe, delivered across enough repetitions to outweigh the original learning. It asks for regulation before insight, because a map only revises when the person is in a state that can take in disconfirming information. And it asks for language that removes shame from the pattern itself: the old strategy was a solution to the situation it was built in, and it deserves respect on its way out the door.

The hardest revisions involve disorganized patterns, where the original learning was that closeness itself carried danger. Those clients test every prediction against you, often without knowing they are doing it, and the work moves at the speed of demonstrated safety. Progress there rarely announces itself. It looks like a client staying present through a moment that used to end the conversation, and noticing, sometime later, that the moment passed.

For the Therapist Reading This

Clients will keep arriving with the quiz version, and the correction you offer is genuinely good news: the styles were always a snapshot of learned expectations, expectations keep learning, and the state in front of you today is workable. For the clinician, the deeper opportunity is learning to track those states in real time, including the mixed and contradictory patterns the four boxes were never able to hold. That skill set has a current research base and a growing clinical literature, and it changes what you listen for in every session, with individuals and with the couples whose stuck cycles run on maps neither partner can see, the pattern we mapped in attachment injuries in couples therapy.

References

Ainsworth, M. D. S., Blehar, M. C., Waters, E., & Wall, S. (1978). Patterns of attachment: A psychological study of the Strange Situation. Erlbaum.

Chopik, W. J., Edelstein, R. S., & Fraley, R. C. (2013). From the cradle to the grave: Age differences in attachment from early adulthood to old age. Journal of Personality, 81(2), 171–183. https://doi.org/10.1111/j.1467-6494.2012.00793.x

Fraley, R. C. (2002). Attachment stability from infancy to adulthood: Meta-analysis and dynamic modeling of developmental mechanisms. Personality and Social Psychology Review, 6(2), 123–151. https://doi.org/10.1207/S15327957PSPR0602_03

Fraley, R. C., & Roisman, G. I. (2019). The development of adult attachment styles: Four lessons. Current Opinion in Psychology, 25, 26–30. https://doi.org/10.1016/j.copsyc.2018.02.008

Hazan, C., & Shaver, P. (1987). Romantic love conceptualized as an attachment process. Journal of Personality and Social Psychology, 52(3), 511–524. https://doi.org/10.1037/0022-3514.52.3.511

Levy, K. N., Kivity, Y., Johnson, B. N., & Gooch, C. V. (2018). Adult attachment as a predictor and moderator of psychotherapy outcome: A meta-analysis. Journal of Clinical Psychology, 74(11), 1996–2013. https://doi.org/10.1002/jclp.22685

Main, M., & Solomon, J. (1990). Procedures for identifying infants as disorganized/disoriented during the Ainsworth Strange Situation. In M. T. Greenberg, D. Cicchetti, & E. M. Cummings (Eds.), Attachment in the preschool years (pp. 121–160). University of Chicago Press.

Marriott, S., & Kelley, A. (2024). Secure relating: Holding your own in an insecure world. Harper Horizon.

Mikulincer, M., & Shaver, P. R. (2016). Attachment in adulthood: Structure, dynamics, and change (2nd ed.). Guilford Press.

Pinquart, M., Feussner, C., & Ahnert, L. (2013). Meta-analytic evidence for stability in attachments from infancy to early adulthood. Attachment & Human Development, 15(2), 189–218. https://doi.org/10.1080/14616734.2013.746257

Roisman, G. I., Padrón, E., Sroufe, L. A., & Egeland, B. (2002). Earned-secure attachment status in retrospect and prospect. Child Development, 73(4), 1204–1219. https://doi.org/10.1111/1467-8624.00467

Waters, E., Merrick, S., Treboux, D., Crowell, J., & Albersheim, L. (2000). Attachment security in infancy and early adulthood: A twenty-year longitudinal study. Child Development, 71(3), 684–689. https://doi.org/10.1111/1467-8624.00176