Attachment style percentages: how common secure, anxious, avoidant and fearful styles are

When your attachment style comes out anxious or avoidant, it is natural to wonder how many people share it. The short answer: in large US studies, a little over half of people were secure, the biggest group, and the rest split between the avoidant and the anxious side. How large each share is, though, depends a good deal on which measure was used, who was asked and where the lines were drawn.

This guide keeps to studies that can be named and looked up. It goes through surveys that sort people into three styles, studies that classify people by interview, and a Korean study with four styles, then explains why the numbers differ and what to keep in mind when reading a result from the mind.io attachment style test.

One point first. An attachment style is less a fixed kind of person than a position on two dimensions: how much you worry about being abandoned in a relationship (anxiety) and how uncomfortable closeness feels (avoidance). Style percentages depend on where those positions are cut and named, so where your scores sit on the two dimensions tells you more than whether your style is common or rare.

Hazan and Shaver’s 1987 study, widely credited with launching research on adult romantic attachment, analyzed 620 readers aged 14 to 82 who answered a “love quiz” printed in a newspaper. Asked to pick which of three descriptions fitted them best, 56% chose secure, 25% avoidant and 19% anxious-ambivalent (anxious). A second study that asked 108 university students the same way found much the same: 56%, 23% and 20%. The authors noted that these adult shares were roughly the same as those reported in studies of infant attachment.

A newspaper quiz, however, reaches people who chose to send in their answers, so it is not a representative sample. That is why a 1997 study by Mickelson, Kessler and Shaver is cited alongside it. Working with a nationally representative sample of American adults, it found 59% secure, 25% avoidant and 11% anxious, a distribution similar to earlier studies. The authors also reported that attachment style was linked to background factors such as income and age, and to interpersonal adversity in childhood.

Read together, the three-style US surveys put secure people at a little over half, avoidant people at about one in four and anxious people at one or two in ten. What stands out is that the order stayed the same whether the people were newspaper readers or a national sample.

Attachment can also be classified by interview instead of questionnaire. The Adult Attachment Interview (AAI) listens to how people talk about their childhood experiences with caregivers and classifies their current state of mind from the way they tell it. In 2009, Bakermans-Kranenburg and van IJzendoorn pooled more than 10,500 AAI classifications from over 200 studies spanning 25 years.

In the combined samples of North American non-clinical mothers that the analysis used as its benchmark, 58% were secure (autonomous), 23% dismissing and 19% preoccupied, and a separate 18% were additionally coded as unresolved about loss or trauma. The authors reported that the distribution was largely independent of language and country of origin, and found no gender difference in the use of dismissing versus preoccupied strategies. Clinical groups, by contrast, showed more insecure and unresolved classifications.

Interview classifications and relationship questionnaires measure different things. The AAI looks at the state of mind with which people recall their childhood, not at their current romantic relationship, while questionnaires ask people to report how they feel in close relationships. Even so, the broad picture is similar: secure people make up more than half, and two insecure styles share the rest.

In 1991, Bartholomew and Horowitz combined a person’s view of self with their view of others to describe four styles. Positive on both makes secure; a negative view of self with a positive view of others, preoccupied; a positive view of self with a negative view of others, dismissing; negative on both, fearful. mind.io sorts its four styles by the anxiety and avoidance dimensions, whose four cells broadly correspond to these: its anxious style is close to preoccupied, its avoidant style to dismissing and its fearful-avoidant style to fearful. The three-style studies offered no fearful option, so people who would have been fearful in a four-style scheme were counted in one of the three other styles.

A study by Jeong and Lee (2005) sorted 265 Korean university students into these four styles and found 32.8% secure, 6.8% fearful, 17.2% preoccupied and 34.7% dismissing. The four shares add up to 91.5%, and the abstract does not explain the remaining 8.5%. Unlike the US representative sample and the interview studies, dismissing came out slightly ahead of secure.

That does not mean “Koreans tend to be avoidant.” This was a student sample, and the questionnaire and the way styles were assigned differed from the US studies. Even for the same group of people, the shares can move a lot depending on the questionnaire used and where the cut-offs are drawn.

First, the measures differ. Picking one of three descriptions, scoring items on two dimensions and classifying an interview can sort the same person differently. Some ask about romantic relationships, others about childhood experiences with caregivers.

Second, styles are names given to cut points on continuous scores. An analysis pooling AAI data from 3,218 people in 40 studies (Raby and colleagues, 2022) found that individual differences in state of mind fit continuous dimensions better than a few categories, and a study of self-report questionnaires (Fraley and colleagues, 2015) also favored dimensions over categories. Where the scores are cut changes the style shares, even for the same people.

Third, different people take part. Newspaper readers, students, national samples and people in counseling or treatment show different distributions; in the interview analysis above, clinical groups had more insecure classifications. A person’s current situation, such as being in a relationship or having just gone through a break-up, may also color their answers.

So there is no single answer to “what percentage of people are secure.” It is more accurate to say that in the US representative sample and the interview studies’ North American norm sample, secure was the most common style and insecure patterns were far from rare. How to read your attachment style result explains how to read your own result.

The mind.io attachment style test asks 28 questions, 14 each on anxiety and avoidance, turns the two scores into 0–100 and splits the four styles at 50. Scores within 5 points of that line are flagged as close to the boundary. The 50 is simply the middle of the scale, not a line that a set share of the population crosses, so the style shares among mind.io results cannot be compared directly with the studies above.

The attachment style ranking anonymously collects only the result styles of people who tapped the join button on their own result page. What is sent is the test, the result style and the page language; answers, scores and names are not, and IP addresses and cookies are not read. The numbers count entries, not people, and entries are anonymous and unverified, so automated ones may be mixed in.

The ranking stays hidden until 30 entries have come in. It reflects people interested enough in attachment to take the test on mind.io themselves, so it would not be surprising for insecure styles to come out more or less often than in representative samples. For a closer look at where you stand, check the anxiety and avoidance scores on your result page.

Q. What is the most common attachment style? A. In the US nationally representative sample (Mickelson and colleagues, 1997) and the North American non-clinical mothers’ norm sample in the pooled interview analysis (Bakermans-Kranenburg and van IJzendoorn, 2009), secure was the largest group, at 59% and 58%. Some samples differ, though: in the Korean student study (Jeong and Lee, 2005), dismissing came out slightly ahead of secure.

Q. Which is more common, anxious or avoidant? A. In every study in this guide, the avoidant side (avoidant in the three-style studies, dismissing in the four-style study and the interviews) outnumbered the anxious side (anxious or preoccupied): 25% to 11% in the US representative sample, 23% to 19% in the interview studies’ North American norm sample and 34.7% to 17.2% in the Korean student study. The three-style avoidant group may also include people who would have been fearful in a four-style scheme.

Q. How common is the fearful-avoidant style? A. In the Korean four-style student study, fearful was the smallest group at 6.8%. The “unresolved” classification in the interview studies (18% of the North American norm sample) is a separate code tied to loss or trauma, so it is hard to treat it as the same thing as the fearful-avoidant style.

Q. Can my attachment style change? A. Attachment style is not regarded as fixed like an inborn trait. Researchers speak of earned security when people who lacked consistently safe early experiences later develop more secure patterns through relationships and self-reflection; becoming more securely attached covers what may help. Read your result as your current tendency.

Q. Why does the mind.io ranking differ from these studies? A. The questions and the cut-offs differ, and the people who join are a self-selected sample. Figures from representative-sample studies aim to estimate the population; the mind.io ranking adds up the results of people who chose to take part.

References: Hazan & Shaver (1987), Journal of Personality and Social Psychology 52(3), 511-524, https://doi.org/10.1037/0022-3514.52.3.511; Mickelson, Kessler & Shaver (1997), Journal of Personality and Social Psychology 73(5), 1092-1106, https://doi.org/10.1037/0022-3514.73.5.1092; Bakermans-Kranenburg & van IJzendoorn (2009), Attachment & Human Development 11(3), 223-263, https://doi.org/10.1080/14616730902814762; Bartholomew & Horowitz (1991), Journal of Personality and Social Psychology 61(2), 226-244, https://doi.org/10.1037/0022-3514.61.2.226; Jeong & Lee (2005), Relationship among Adult Attachment Styles, Negative Cognitive Distortion, and Depression, Korean Journal of Psychology: General 24(1), 167-184, https://accesson.kr/kpageneral/v.24/1/167/15578; Raby et al. (2022), Development and Psychopathology 34(1), 307-319, https://doi.org/10.1017/S0954579420000978; Fraley, Hudson, Heffernan & Segal (2015), Journal of Personality and Social Psychology 109(2), 354-368, https://doi.org/10.1037/pspp0000027 (accessed October 2026).