Attachment Styles: What the Research Supports and What Social Media Gets Wrong

Spend ten minutes on relationship content online and you will be told that you are anxious, avoidant, secure or some mixture of the three, that your type was stamped on you before you could talk, and that it explains why the last person you dated stopped replying. Some of this rests on decades of careful research. Some of it has drifted a long way from what that research found.

The honest picture is more interesting than the viral one, and in some ways more hopeful.

From babies to adult relationships

Attachment theory began with infants. John Bowlby and Mary Ainsworth described how babies form bonds with their caregivers and how those bonds come in recognisably different patterns: some infants are secure, some anxious and clingy, some seemingly indifferent.

In 1987 two researchers, Cindy Hazan and Phillip Shaver, asked a bold question: could romantic love be the same kind of bonding process, running in adults? They translated the three infant patterns into three short descriptions of how a person feels about closeness, and asked adults to pick the one that fitted them best. In two questionnaire studies, the three adult styles appeared in roughly the same proportions as in infancy, and people who chose different descriptions reported different experiences of love and different memories of their relationships with their parents.

It was a landmark paper, and it is also the origin of the habit that social media has inherited: sorting people into types. A few years later, Kim Bartholomew and Leonard Horowitz proposed four categories instead of three, built from how a person sees themselves and how they see others. That four-way version is where much of today's vocabulary comes from.

Two sliding scales, not four boxes

Most researchers now measure adult attachment on two continuous dimensions. Attachment anxiety is how much you worry about being rejected or abandoned, and how much you need reassurance. Attachment avoidance is how uncomfortable you are with closeness and with depending on others. Everyone sits somewhere on each scale. The familiar labels are really corners of this map: low on both is what people call secure.

Is there a real boundary between the corners, or are the types just lines drawn on a continuum? In 2015, R. Chris Fraley and colleagues tested this directly with taxometric analysis, a family of statistical methods designed to tell whether a trait comes in natural kinds or varies by degree. Across two samples, the results fitted a dimensional model better than a categorical one, both for attachment in general and for specific relationships such as those with parents and friends.

This matters more than it sounds. If attachment is a pair of sliding scales, then a person just below a cut-off and a person just above it are nearly identical, even though one gets called anxious and the other secure. Small shifts in mood or circumstance will move people across the line.

That is exactly what earlier work had seen. Mark Baldwin and Beverley Fehr looked at how stable people's answers to the original one-question measure were. About 30% of respondents changed their classification over periods from one week to a few months, and among those who had called themselves anxious-ambivalent, most changed. The authors suggested that the answer may reflect whichever way of relating is active in your mind at that moment, rather than a fixed trait.

Is it set in stone in childhood?

The popular story says your parents wrote your attachment style and you are now reading it aloud. The evidence supports a softer version.

In a 2002 meta-analysis of longitudinal studies, Fraley found that attachment security is moderately stable from infancy to age 19, and that the pattern fitted a model in which early experiences keep some lasting influence. Moderate is the key word: early relationships leave a trace, but they are far from a sentence.

Later life keeps moving the dial. William Chopik, Robin Edelstein and Kevin Grimm combined three long-running studies following 628 people from age 13 to 72. On average, attachment anxiety declined with age, especially in middle and older adulthood, and avoidance declined steadily across the lifespan. Being in a relationship went along with lower levels of both. These are averages from American samples using one particular measure, and they cannot show what caused the change, but they fit poorly with the idea that your style is frozen.

What it predicts, and what the viral version gets wrong

None of this means the concept is empty. The two dimensions do relate to real outcomes. A 2019 meta-analysis by Octav-Sorin Candel and Maria Nicoleta Turliuc pooled 132 studies and found that both attachment anxiety and avoidance go with lower relationship satisfaction. The link was stronger for a person's own satisfaction than for their partner's, and its size varied with age, relationship length and region.

Two cautions keep this in proportion. These are correlations, so they cannot tell us whether insecurity lowers satisfaction, an unhappy relationship raises insecurity, or both. And an average association across thousands of couples says little about what will happen in any single relationship. The viral version goes wrong in the leaps it makes from here.

• Types as identities. Calling yourself an avoidant as if it were a blood group ignores the finding that attachment varies by degree and shifts with context.
• Diagnosing other people from a few messages. A slow reply is weak evidence about anyone's inner life, and attachment style is not a clinical diagnosis in the first place.
• Destiny. Moderate stability means change is common. Age, experience and supportive relationships all appear to move people towards security on average.
• One style for every relationship. Research on relationship-specific attachment suggests you can feel secure with a friend and anxious with a partner at the same time.

What this means for you

If you take our attachment style test, read the result as a position on two scales today, not a label for life. Look at where you sit on anxiety and on avoidance separately, and ask which situations push each one up. Worry about rejection overlaps with what our rejection sensitivity test measures, and our relationship satisfaction test can help you see whether the patterns you notice are actually affecting how you feel about your relationship.

Above all, the research suggests curiosity rather than fatalism. The same person can feel very different in different relationships, and on average people become more secure over time. If patterns of fear or distance in relationships are causing you persistent distress, a qualified psychologist or therapist is a far better guide than any quiz or video.