Can attachment styles change?

Kasia Staniaszek, Psy.D. · August 12, 2026

Four patterns, in the shortest version anyone can give. Secure: closeness is basically safe, and an argument doesn’t threaten the whole thing. Anxious: closeness feels precarious, so you monitor it. Avoidant: needing someone feels unsafe or pointless, so you handle it alone. And a fourth, where both of those run at once and the same person is who you want and who you brace against.

Almost nobody asks which one they are without also asking whether they are stuck with it. So: can attachment styles change? Yes. The route people want, out of the anxious or avoidant pattern and toward the secure one, is a real route and people do travel it. The part worth knowing is the shape of the trip. How far you can expect to move, over what kind of time, and what makes a change hold instead of fade.

What tends to stay put

Researchers argued for years about whether the early pattern persists or gets overwritten by later experience. One study went at it by measuring people intensively: 203 adults every day for a month, and 388 adults every week for a year. Attachment scores moved around constantly. Underneath the movement sat a stable factor that kept drawing people back toward their own average.

Which changes how you should read your own progress. Six good weeks with a steady partner is not yet change. It may be ordinary week-to-week variation that happens to be pointing up. It is also why people decide they have finally sorted this out, then meet the whole old pattern intact the first time something goes badly wrong.

Change over time, and change that lasts

The largest test of what actually moves the baseline followed more than 4,000 people through ordinary life events: new relationships, breakups, moves, job changes. Half of those events shifted attachment right away. Then most people drifted back to the trajectory they had been on before it happened.

Roughly a quarter of the events produced change that endured. And the averages conceal the interesting part. Even for events where the group as a whole ended up back where it started, some individuals became lastingly more secure and others lastingly less. What separated them had a lot to do with how they came to understand the event, rather than the event itself.

Change in adulthood, and the direction it runs

Widen the window and there is a drift. A study following 628 people from age 13 to 72 found both attachment anxiety and attachment avoidance declining across adulthood, anxiety most steeply in middle age and beyond. Being in a romantic relationship predicted lower levels of both.

That is a span of decades rather than months, and it is an average with plenty of individuals travelling the other way. It is still worth saying plainly: an ordinary adult life bends slightly toward security on its own.

Where therapy fits

A review of studies that measured attachment before and after psychological therapy found security rising and attachment anxiety falling, consistently enough to show up across different patient groups, therapy types and settings. Results for avoidance were unclear, and the authors say directly that controlled trials are still needed. I would rather hand you that hedge than a tidier claim.

One thing shows up in all of it: repetition, over a long stretch. An old forecast gets revised by turning out wrong many times over, inside a relationship that lasts long enough for the corrections to add up. That is the mechanism, whether it happens in a marriage, a long friendship, or a consulting room. None of them do it in six weeks.

So the honest version has two halves. Your pattern is more stable than the self-improvement shelf admits, and considerably less fixed than the deterministic reading of childhood claims. Most of what you would actually want to change is reachable. It is measured in years.

Related: what your attachment style is, and what it isn’t, and, on doing the changing, how to become securely attached. On working this way at all, does online therapy work as well as in person.