The short answer
Yes, it can change, but usually slowly and not on command. The research describes attachment as a pattern with a fairly stable core and real room for movement around it. You are not the same at 40 as you were at 18, and you are not a different person every week either. Both halves matter if you want to work on it.
A note on what researchers measure: they mostly track the two dimensions, anxiety about being left and discomfort with closeness, rather than four labels. Change usually shows up as a score drifting over months or years, not a sudden jump from one style to another, which is worth remembering before reading too much into one result.
How stable is it from childhood?
A 2013 meta-analysis by Pinquart, Feussner and Ahnert pooled 127 studies covering more than 21,000 attachments to mothers and fathers, from infancy to early adulthood. Overall stability was moderate, a correlation of about .39. It was higher over short gaps and lower over long ones, and over intervals longer than 15 years they found no significant stability at all.
That last finding is worth sitting with. Early attachment matters, but it does not lock in adult relationships. The same review found that children in higher-risk circumstances who started out secure were less likely to stay secure, which suggests that life conditions keep shaping the pattern long after infancy.
What changes across adulthood
The longest view comes from Chopik, Edelstein and Grimm, who combined three long-running samples covering ages 13 to 72. On average, attachment anxiety declined with age, particularly in middle and older adulthood, and avoidance declined steadily across the lifespan. Being in a relationship predicted lower anxiety and avoidance, and men scored higher on avoidance at every age.
These are averages, so they say nothing certain about any one person, and the study cannot show whether relationships lower insecurity or more secure people are simply more likely to be partnered. But the data cut against the idea that insecurity only hardens with time. For most people, the trend ran toward more security, not less.
Week-to-week swings around a steady core
Fraley and colleagues measured adults daily for 30 days in one study and weekly for a year in another. The pattern that fitted best was a stable underlying factor with temporary ups and downs around it. In plain terms, you have a home base, and a bad week or a difficult partner can move you away from it for a while without rewriting it.
Some people swing more than others. Davila, Burge and Hammen followed 155 young women over two years and found that change in attachment style behaved partly like a trait of its own: some people were more prone to fluctuate, and the authors linked this to adverse earlier experiences. So if your results move around, that itself is useful information.
What seems to help
Two routes have the most support. Relationships are one: in the lifespan data above, being in a relationship went with lower anxiety and avoidance. Therapy is the other: a systematic review found that attachment security tended to rise and anxiety to fall during psychological therapy, across different approaches, though the authors called for more controlled trials.
Researchers also talk about earned security, meaning people who became secure despite an insecure start. In a 23-year study, Roisman and colleagues found that people who were insecure as infants but secure as young adults went on to do well in their close relationships. They also found that judging this from adults' memories of childhood was less reliable than expected.
What is still not known
Be wary of anyone promising to change your attachment style in a set number of days. Several basic questions are still open, and most of the evidence is correlational, which means it shows what tends to go together rather than what causes what. These are the gaps worth knowing about before you read any bold claim.
- Whether avoidance changes much in therapy: the review found the evidence unclear.
- How much change comes from a new relationship and how much from simply getting older.
- Which specific exercises or self-help approaches cause lasting change.
- How long changes last after a relationship ends or therapy stops.
Tracking your own pattern
The honest version is slower and less dramatic: patterns shift with repeated experience, and you can put yourself in the way of the right experiences. If you want to see change, start with a baseline. The free test gives your anxiety and avoidance scores in 20 statements and about three minutes, plus a Big Five snapshot. Retake it in a few months and compare the scores, not only the label.