Yes — with honest qualifications. Longitudinal research finds that attachment patterns show meaningful stability and real change: they are modestly stable across life, but genuinely malleable in response to new relationship experiences. Change is possible, usually slow, and usually relational — it happens between people, not by force of will.
This guide reviews what the evidence actually says: what “stability” means, what earned security research really found (including the nuance most summaries skip), how change happens, what it doesn’t mean — and how this question differs from whether personality can change (a related question about a different construct).
Table of Contents
What “Stability” Really Means
First, the baseline. Attachment patterns are not fixed traits — but they’re not weather vanes either. The key finding: Groh, Roisman, Booth-LaForce, and colleagues (2014) followed attachment security from infancy to late adolescence and found stability that was “significant but very modest” [3].
“Modest stability” is doing a lot of work in that sentence. It means early patterns predict later functioning better than chance — the foundation metaphor from developmental research holds — while leaving enormous room for change. The same research tradition that established that early experience matters also established its limits. Anyone selling you either “you’re stuck forever” or “just think positive and transform overnight” is ignoring half the evidence.
Earned Security: What the Research Actually Found
The most hopeful — and most misunderstood — finding in this literature is earned security: adults who describe difficult early attachment experiences yet whose current state of mind about attachment is secure, coherent, and balanced.
The landmark study is Roisman, Padrón, Sroufe, and Egeland (2002), who followed participants from infancy to age 21 [1]. Here’s the finding and the nuance most summaries skip: when the researchers identified “earned-secures” retrospectively (adults who recalled difficult early experiences but functioned securely), those adults were not more likely than continuously-secure adults to have been classified as insecure in infancy. In other words, the study does not show a clean pipeline from “insecure baby” to “secure adult” — the retrospective definition complicates the simple conversion narrative.
So what can we honestly say? That security can develop later in life, and that the pathways researchers have explored point consistently in one direction: alternative supportive relationships. Saunders, Jacobvitz, Zaccagnino, Beverung, and Hazen (2011) investigated pathways to earned security and found a meaningful role for alternative support figures — a responsive grandparent, a close mentor, a steady friend, a good partner — in revising internal working models [2].
The honest summary: earned security is real as a present-tense phenomenon (adults functioning securely despite difficult histories), the simple “rewrote my childhood” story is more complicated than it looks, and the mechanism that keeps showing up is accumulating experiences of reliable connection with other people. (For the secure pattern itself, see our guide to secure attachment.)
How Change Happens
Working models revise through repeated experience, not single epiphanies. The pathways the research points to:
- Corrective relationship experiences. A steady partner, close friendships, or family bonds that reliably show up — over months and years, not weeks. Each answered bid for connection is a data point against the old model. This is the mechanism behind most documented change: the environment stops confirming the alarm.
- Alternative support figures. The Saunders et al. finding deserves emphasis: change often comes not from the original relationships but from new ones — mentors, friends, chosen family — who provide the reliable-care experiences the early environment didn’t [2].
- Therapy as a secure base. A good therapeutic relationship can function as a practice ground for the patterns above: depending on someone, being accurately seen, experiencing repair after misunderstanding. This is one option among several — not a prescription, and not a claim that therapy “rewires” attachment on a schedule.
- Deliberate practice of secure behaviors. Naming needs directly, staying in uncomfortable conversations, tolerating uncertainty in small doses — the growth strategies from the style profiles. Behaviors practiced repeatedly become the new evidence the working model is built from.
Notice what’s missing from this list: willpower as a primary engine, insight alone, and timelines. The research supports direction, not speed.
What Change Doesn’t Mean
Honesty requires the other half:
- Not willpower alone. “Just choose to be secure” misunderstands the mechanism. Working models are built from experience, and experiences of reliable care can’t be willed into existence — though you can put yourself where they’re likelier to happen.
- Not erasing the past. Change means the old patterns lose their grip on the present; it doesn’t mean the history disappears. Someone can function securely and still recognize the old alarm when it sounds — the difference is what happens next.
- Not on a timeline. No credible research gives a schedule for attachment change, and any source that does is inventing it. Slow is normal; non-linear is normal.
- Not a different person. Growth toward security doesn’t erase temperament or history. The goal was never to become someone else — it’s to make the attachment system less hair-trigger, so ordinary closeness stops feeling like danger.
Attachment Change vs. Personality Change: Different Questions
It’s natural to wonder whether changing your attachment style is the same as changing your personality. They’re related but distinct constructs — and the distinction matters.
Attachment patterns describe how you relate to close others: your expectations of care, your comfort with dependence, your strategies under relational stress. Personality traits (like the Big Five) describe broader, more general dispositions — how outgoing, organized, or emotionally reactive you tend to be across all of life, not just in close relationships. Research finds the two are associated but not identical: attachment dimensions relate to Big Five traits without being reducible to them.
So the questions are cousins, not twins. If you’re curious about the personality side, our Personality cluster’s guide to whether personality can change takes up that distinct question with its own evidence. What follows here stays with attachment: patterns of closeness, built in relationship, revisable in relationship.
A note on what this article is (and isn’t)
This is educational content about research on malleability — not a promise, a program, or a diagnosis. Nothing here guarantees change on any timeline, and recognizing the possibility of change is not the same as achieving it. If relationship distress is persistent or overwhelming, a qualified mental health professional is the right next step — not a webpage.
Frequently Asked Questions
Can you change your attachment style on your own?
Partly. The deliberate practices — noticing patterns in real time, communicating needs directly, tolerating uncertainty in small doses — are genuinely yours to do, and they matter. But the research keeps pointing at the same mechanism: working models revise through experienced reliable care, which means other people are usually part of the story. “On your own” can start the process; relationships tend to complete it. That’s not a weakness in the plan — it’s the theory working as described.
How long does it take to change attachment patterns?
Honestly: research doesn’t give a timeline, and you should distrust anyone who does. The longitudinal studies measure change across years, not weeks — and change is typically gradual and non-linear, with old patterns resurfacing under stress even after long stable stretches. The absence of a timeline isn’t a gap in the science; it’s the science telling you this isn’t the kind of thing timelines describe well.
Does therapy change attachment style?
Therapy is one documented pathway, not a guaranteed mechanism. The research on earned security points to supportive relationships generally — and a therapeutic relationship can be one such relationship: a place to experience being accurately seen, to practice depending on someone, to have ruptures repaired. But therapy isn’t the only route (friendships, partnerships, mentoring bonds all count in the literature), and no ethical therapist promises attachment transformation on a schedule.
Is earned security the same as being secure all along?
Functionally, in the present, the research treats them similarly — an earned-secure adult’s current state of mind about attachment looks secure: coherent, balanced, flexible. But the histories differ, and Roisman et al.’s nuance matters here: the retrospective “earned” label doesn’t map cleanly onto documented insecure infancies [1]. The useful takeaway isn’t the label’s purity — it’s that secure functioning in the present is what predicts how relationships go, whatever road led there.
Keep Exploring
- Secure Attachment — the dedicated profile: signs, development, and earned security.
- Attachment Styles — all four patterns compared side by side.
- Attachment Theory Explained — the full framework: Bowlby, Ainsworth, and adult attachment research.
References
- Earned-secure attachment status in retrospect and prospect — Roisman, G. I., Padrón, E., Sroufe, L. A., & Egeland, B. (2002), Child Development, 73(4), 1204–1219. https://doi.org/10.1111/1467-8624.00467 (citation verified via ERIC record: https://eric.ed.gov/?id=EJ661625)
- Pathways to earned-security: The role of alternative support figures — Saunders, R., Jacobvitz, D., Zaccagnino, M., Beverung, L. M., & Hazen, N. (2011), Attachment & Human Development, 13(4), 403–420. https://doi.org/10.1080/14616734.2011.584405 (citation verified per research pack)
- Attachment — From Theory to Practice (Fearon et al.; UCL-hosted PDF summarizing attachment theory progress, including Groh et al., 2014 on “significant but very modest” stability of attachment security from infancy to late adolescence). https://discovery.ucl.ac.uk/id/eprint/1543228/1/Fearon_Attachment%20theory_progress.pdf (opened and verified)
- Attachment in Young Adulthood — Lifespan Human Development: A Topical Approach (open textbook), Baylor University (adult attachment; stability and change). https://openbooks.library.baylor.edu/lifespanhumandevelopment/chapter/chapter-17-2-attachment-in-adulthood/ (opened and verified)
- Attachment style — APA Dictionary of Psychology (describes adult attachment styles). https://dictionary.apa.org/attachment-style (opened and verified)
- What Do Self-Report Measures of Attachment Assess? — P. R. Shaver (chapter; attachment dimensions; attachment across relationships and time). https://adultattachment.faculty.ucdavis.edu/wp-content/uploads/sites/66/2015/09/Shaver_2004_What-Do-Self-Report-Measures-of-Attachment-Assess.pdf (opened and verified)