Tests · attachment
Attachment Style Test
Your attachment style is the pattern you fall into in close relationships: how much you worry about losing the connection, and how much distance you need to feel safe. This 12-question test measures both scales and plots you on a chart, so you can see which of the four styles you're closest to right now, secure, anxious, avoidant, or fearful-avoidant. Answer with your closest relationships in mind, romantic ones first.
- 12 questions
- Written for this site
- Not a published scale
- About this test
← All tests · Read about anxious attachment · AI assistant
About the Attachment style test
How much you worry about losing closeness (anxiety) and how much you pull away from it (avoidance) in close relationships, placed on a chart with four labeled corners.
- Answers
- 0 (not me at all) to 3 (definitely me)
- Chart
- Four corners, split at 9 on each scale
- Where the questions come from
- Written for this site. They aren't the wording of the ECR-S or the ECR-R
- Two-scale idea from
- The Experiences in Close Relationships (ECR) questionnaires, Brennan, Clark and Shaver, 1998
Where it comes from
These 12 questions aren't a published questionnaire. They cover the two themes of the Experiences in Close Relationships (ECR) scales: worry about being left or unloved (anxiety), and discomfort with closeness and depending on others (avoidance). Brennan, Clark and Shaver built the 36-item ECR in 1998, with 18 items for each dimension. Fraley, Waller and Brennan published a revised version, the ECR-R, in 2000, and Wei and colleagues published a 12-item short form, the ECR-S, in 2007. The wording here matches neither.
The four labels come from different work. In 1991 Bartholomew and Horowitz proposed four adult attachment patterns, defined by whether a person's image of self and of others is positive or negative. Shaver and Fraley describe two dimensions under those patterns and prefer to call them anxiety and avoidance. Researchers usually name the four patterns secure, preoccupied, dismissing and fearful. This page uses secure, anxious, avoidant and fearful-avoidant.
What the research found
This evidence comes from the published scales and doesn't transfer to these 12 questions. The ECR-S paper reports six studies in which two factors, anxiety and avoidance, fit well after statistically removing the influence of response sets, and its validity matched the full ECR's. A 2020 study of Chilean adults found poor fit for three short ECR forms, the ECR-S among them. The ECR-R paper analyzed answers from 1,085 people on four attachment questionnaires and chose items using item response theory.
Whether styles are real types is a separate question. Shaver and Fraley's overview says Fraley and Waller showed in 1998 that there's no evidence for a true attachment typology, only regions in a two-dimensional space. A 2015 analysis of two samples (Fraley, Hudson, Heffernan and Segal) found individual differences more consistent with a dimensional than a categorical model, for general attachment and for relationships with parents and peers. Fraley's FAQ for the ECR-R recommends against sorting people into categories from their scores, though he says a label can help explain one person's scores.
We found no published study of these 12 questions. Their reliability and the cut-off of 9 are untested.
How scoring works
| Score | Label | What it means |
|---|---|---|
| Anxiety 0 to 8, avoidance 0 to 8 | Secure | Both scores under 9 of 18. The matching pattern in the four-pattern model is secure. |
| Anxiety 9 to 18, avoidance 0 to 8 | Anxious | Higher anxiety, lower avoidance. Researchers call the matching pattern preoccupied. |
| Anxiety 0 to 8, avoidance 9 to 18 | Avoidant | Lower anxiety, higher avoidance. Researchers call the matching pattern dismissing. |
| Anxiety 9 to 18, avoidance 9 to 18 | Fearful-avoidant | Both scores 9 or higher. Researchers call the matching pattern fearful. |
No study set the cut-off of 9. It's the midpoint of each 0 to 18 scale, a choice made for this page. The published ECR scales give continuous scores on each dimension and come with no cut-offs. Anxious, avoidant and fearful-avoidant are popular names for the corners, not terms from the scale authors.
What it can't tell you
The chart splits each scale at 9 of 18, the midpoint. A person who scores 8 on anxiety and 9 on avoidance lands in a different corner from someone who scores 9 on both, though one point separates them. The published ECR scales give continuous scores. Fraley suggests that anyone who must sort people into groups use the median of their own sample.
The ECR scales ask about romantic partners. These questions mix "someone close to me" with "my partner," so your answers may blend several relationships, and researchers study attachment to specific people separately from general attachment. A self-report score can't show why a pattern formed, and it isn't a diagnosis.
How to use your result
Read your two scores first and the corner second. A high anxiety score points to more worry about being left or unloved, a high avoidance score to more discomfort with closeness, and both can be high. Retake it after a few weeks, keeping one relationship in mind, and see what moves.
If these patterns are hurting your relationships, a therapist can help you look at them. For more, see our pages on anxious attachment and avoidant attachment.
More on this topic: Anxious attachment, Avoidant attachment, Fear of abandonment, Fear of commitment.
Your answers
Your answers are scored in your browser. If you sign in and choose to save a result, we store the score and its label, not your individual answers. Without an account, nothing is saved.
Sources 8
- Fraley, FAQ for the ECR-R (University of Illinois): the 36 ECR-R items, the definitions of attachment anxiety and avoidance, the terms of use, the advice against classifying people, and the median-split method. The developer's own page.
- Shaver and Fraley, Self-Report Measures of Adult Attachment: history of the self-report scales, Bartholomew's four patterns, the two dimensions and the point that there's no true typology. An overview by two researchers in the field, not a study.
- Wei, Russell, Mallinckrodt and Vogel, Journal of Personality Assessment, 2007 (PubMed): abstract only: the 12-item ECR-S, six studies, two factors that fit after removing response sets.
- Fraley, Waller and Brennan, Journal of Personality and Social Psychology, 2000 (PubMed): abstract only: 1,085 people, four inventories, item response theory.
- Bartholomew and Horowitz, Journal of Personality and Social Psychology, 1991 (PubMed): abstract only: the four-group model defined by self-image and image of others. The abstract doesn't name the four patterns.
- Fraley, Hudson, Heffernan and Segal, Journal of Personality and Social Psychology, 2015 (PubMed): abstract only: two samples, dimensional model fits better than a categorical one.
- ECR-S form as posted by a clinician (drrebeccajorgensen.com): a copy of the 12 ECR-S items with the 1 to 7 answer scale and the scoring key, credited to Wei et al. 2007. Not the authors' own site.
- Guzman-Gonzalez, Rivera-Ottenberger, Brassard, Spencer and Lafontaine, Psicologia: Reflexao e Critica, 2020 (PubMed Central): open-access full text: a Spanish-language study in Chile that reports poor fit for three existing short ECR forms in its samples, and summarizes critiques of the ECR-S. Secondary source.