Run Game Conversation

Attachment styles and dating: what the theory says and where it stops

6 min read

"I'm anxious, she's avoidant" is one of the most common ways people describe a dating problem online. The labels come from real research. The research is also narrower, and less settled, than the quizzes and videos make it sound. Here's what the theory says, what the evidence supports, and how to use the vocabulary without turning it into a verdict on someone.

Where it came from

Attachment theory began with the British psychiatrist John Bowlby, whose trilogy Attachment and Loss appeared between 1969 and 1982. It was about children: how infants bond with caregivers and what happens when that bond is disrupted. Mary Ainsworth then developed a lab procedure, the Strange Situation, in which an infant is observed as a caregiver leaves and returns. Infants were sorted into secure, avoidant and ambivalent (also called resistant) patterns, and a fourth, disorganized, was added later by Mary Main.

Adults entered the picture in 1987. Cindy Hazan and Phillip Shaver published a paper in the Journal of Personality and Social Psychology arguing that romantic love can be understood as an attachment process. They used a "love quiz" printed in a local newspaper. In a summary of the study, 620 people aged 14 to 82 responded (205 men, 415 women), and respondents answered questions about their most important romantic relationship and their childhood. Roughly 56% were classed as secure, 25% as avoidant and 19% as ambivalent. Those classed as secure reported longer relationships on average (about 10 years against roughly 5 to 6 for the other groups) and described their love as trusting and friendly. Avoidant respondents described fear of closeness and jealousy, and ambivalent respondents described preoccupation with the partner and a strong wish for union.

Two things about that study are worth knowing. The sample chose itself, because anyone who read the paper and felt like answering could. And the relationship lengths came from people describing one past or present relationship, and length also depends on how old the respondent is, a possible confound.

What the labels mean in practice

Many popular versions use four: secure, anxious-preoccupied, dismissive-avoidant and fearful-avoidant. Another common research framing uses two dimensions instead of boxes. Roughly, attachment anxiety is worry that a partner won't be there or won't care enough, and attachment avoidance is discomfort with closeness and a preference for self-reliance. You can be high or low on each, in any combination.

Some behaviors that fit the descriptions:

These are patterns of behavior, not types of people, and how they show up can vary with the situation.

How strong is the evidence?

Parts of it are solid, and parts are thin or disputed.

Types or dimensions. In 1998, R. Chris Fraley and Niels Waller used statistical tests on data from 639 young adults to ask whether adult attachment is made of categories or of continuous dimensions. Their abstract states the question; it doesn't say which answer won, and the debate has continued. A 2021 study of 515 adults in Spain (all university students, about 77% women) argued for a mix: continuous differences, with two broad groups, about 77% in a normal range and 23% insecure.

Stability. A 2020 meta-analysis of early-childhood studies (Opie and colleagues, 63 studies) cites earlier lifespan meta-analyses that put the secure-versus-insecure correlation across time at about 0.39, and found a lower figure, 0.28, for early childhood. The authors call that moderate stability: patterns carry over, and plenty of people change. The Wikipedia entry on the theory also notes that security can shift with later relationships and experiences.

Limits raised about the theory. As summarized on Wikipedia, critics say it was built largely on infants and mothers, gives less weight to temperament, genetics, fathers, peers and wider family, and doesn't travel evenly across cultures. Those are limits on the research base, not claims that the patterns don't exist.

Dating-specific claims. Online posts often say a given pairing, such as an anxious person with an avoidant one, is doomed or inevitable. Treat that as a popular framing; the studies cited here don't test it. One 2010 study (Mikulincer and colleagues) reported that higher anxiety was linked to stronger ambivalence toward a partner, and higher avoidance to ambivalence about distance in negative situations. That describes feelings, not who ends up with whom.

Why a quiz isn't a diagnosis

The 1987 study relied on a self-report questionnaire in a newspaper. Most online quizzes are self-reports too: a few questions, a label, no clinician. They capture how you describe yourself on the day, in whatever mood you're in. A single result can't tell you about your childhood, your partner, or your future. Where a quiz doesn't say who built it or how it was tested, treat it as entertainment.

Applying a label to someone else is a bigger leap. You can't diagnose a partner from their texting habits, and "you're avoidant" tends to end conversations. If patterns are causing real distress, a licensed therapist can assess them properly. A quiz can't.

Using the vocabulary well

Used carefully, the words help you describe a specific moment. "When I don't hear back for a day, I get anxious and start imagining the worst. What would help?" is clearer than "you're avoidant." Similarly, "After a big conversation I need a few hours alone, and it isn't about you" gives your partner something to work with.

On the early dating side, jealousy early in dating and stop needing approval cover the day-to-day habits that tend to come up with anxiety. Esther Perel is a therapist whose work looks at how desire and closeness coexist in long relationships, a different angle on the same questions, and boundaries in a relationship is a practical guide to asking for space or reassurance directly.

What to do next

  1. If you take a quiz, take it once, note the result, and treat it as a starting question, not an answer.
  2. Describe your own pattern in one sentence of behavior, such as "I check my phone after sending a message."
  3. Write down one thing that helps you when that pattern kicks in: a walk, a call with a friend, waiting an hour before replying.
  4. Tell a partner one specific request: "If I'm quiet, it's often about needing space, not about you," or "A quick text when you're running late helps me a lot."
  5. Don't apply labels to your partner. Ask about situations instead: "What do you need when things feel stressful?"
  6. If the same pattern repeats across relationships and bothers you, book a session with a licensed therapist.

Personality types in dating is a natural next read.

Sources