Skip to content
Psych Dictionary

Attachment styles

  • Concept · Developmental and personality psychology
  • Mixed or contested
  • Published October 18, 2026
  • 18 sources
  • 6 min read

In short

Attachment styles are research classifications of how people differ in seeking closeness and comfort from attachment figures, measured in infancy with the Strange Situation and in adulthood with interviews or questionnaires. They are not clinical diagnoses.

Key facts

Field
Developmental and personality psychology
Originators
Mary Ainsworth and colleagues (infant classifications, 1978); Mary Main and Judith Solomon (disorganized category); Cindy Hazan and Phillip Shaver, Kim Bartholomew, and Kelly Brennan and colleagues (adult measures)
First described
1978 onward (infant classifications); 1987 onward (adult romantic attachment)
Also known as
Attachment patterns; Attachment orientations

State of the evidence

Mixed or contested

The classifications are established research measures, but their status as distinct types is contested. Taxometric analyses of infant Strange Situation behavior and of adult self-reports find continuous dimensions rather than categories[1][2]. Infant classifications show small associations with later behavior problems[3][4], and the two main adult measures, the Adult Attachment Interview and self-report questionnaires, overlap only trivially (mean r = .09)[5]. Disorganized attachment, the classification most often misapplied, is not a validated individual-level clinical diagnosis[6].

Evidence labels summarise the research literature; they are not recommendations. How labels are assigned

Attachment styles are the classifications researchers use to describe individual differences in attachment: how a person seeks, or avoids, closeness and comfort from an attachment figure under stress. The term covers two research traditions that grew out of attachment theory. Developmental researchers classify infants from their behavior in Mary Ainsworth’s Strange Situation and classify adults from interviews about childhood; social and personality researchers measure adults’ feelings in close relationships with questionnaires[5][7]. The two traditions use similar labels for different measurements.

Infant classifications in the Strange Situation

Ainsworth’s first three classifications came from her Baltimore study of infants and mothers and were set out in Patterns of Attachment in 1978[9][10]. They rest mainly on what the infant does when the mother returns after the two brief separations of the Strange Situation, not on how upset the infant is when she leaves; about half of secure one-year-olds protest the separations and half do not[10][11]. At Bowlby’s suggestion the groups were given neutral letters rather than evaluative names[11].

Secure (B)

Attachment behavior rises as the procedure becomes more stressful; on reunion the infant seeks proximity, contact or interaction with the caregiver, is comforted and returns to play[10][11].

Insecure-avoidant (A)

Attachment behavior does not rise with stress; the infant directs attention and movement away from the caregiver, often to the toys, and avoids the caregiver on reunion despite often having searched for her during the separation[11][10].

Insecure-resistant or ambivalent (C)

The infant is markedly distressed by separation and on reunion mixes contact-seeking with angry resistance, or stays passive, and is not readily settled[11][10].

Disorganized (D)

The infant shows conflicted, disoriented or fearful behavior in the caregiver’s presence, such as contradictory movements, freezing or apprehension, intense enough to disrupt an organized pattern[6].

Ainsworth’s home observations showed that the avoidant and resistant infants had less harmonious relationships with their mothers than the secure infants, and the Strange Situation classifications were validated against those observations[10]. Ainsworth and colleagues summarized the contrast between the two insecure groups in a dictum: the resistant infant fears that it will not get enough of what it wants, and the avoidant infant fears what it wants[11].

Disorganized attachment

Some infants did not fit Ainsworth’s three groups: in a 1981 interim report on 46 families, Mary Main and Donna Weston found that 12.5% of infants could not be classified with her system[11]. Main and Judith Solomon went on to identify seven classes of behavior that, when shown with enough intensity in the parent’s presence, mark an infant as disorganized: contradictory behavior shown in sequence or simultaneously, undirected or interrupted movements, stereotypies and anomalous postures, freezing or slowed movements, direct signs of apprehension of the parent, and direct signs of disorientation[6]. Disorganized infants also receive the best-fitting organized classification as a secondary code[6].

A meta-analysis of nearly 80 studies with more than 6,000 infant–parent pairs estimated that about 15% of infants in middle-class families show disorganized attachment, and that the rate can be two or three times higher in other social contexts and in clinical groups, for example after maltreatment. Frightening parental behavior may play a part, but it did not appear to be the only cause[12].

The classification is easily over-read. A review signed by more than 40 attachment researchers concluded that it is false to treat disorganized attachment as a definitive assessment of an individual in child-protection settings, as a reliable indicator of maltreatment, as a strong predictor of pathology or as a fixed trait of the child[6]. Disorganization is relationship-specific, and rates in families facing five or more socioeconomic risks are similar to those in maltreated samples[6].

Adult romantic attachment styles

In 1987 Cindy Hazan and Phillip Shaver proposed that romantic love is an attachment process, governed by the same behavioral system that binds infants to caregivers, and that adults show analogues of Ainsworth’s three groups: secure, avoidant and anxious-ambivalent[13][7]. Their measure consisted of short multisentence descriptions of each style[7]. Adults who described themselves as secure, avoidant or ambivalent reported different childhood relationships with their parents[10]. The anxious-ambivalent category is the forerunner of what later research calls attachment anxiety, the dimension behind the term anxious attachment[7].

Kim Bartholomew argued that this scheme merged two kinds of avoidance. Fearful-avoidant adults avoid closeness to protect themselves from being hurt or rejected; dismissing-avoidant adults avoid it to maintain a defensive self-reliance[7]. With Leonard Horowitz she proposed four prototypes defined by the combination of a positive or negative image of the self and a positive or negative image of others; secure is positive on both, and the anxious-ambivalent group became preoccupied[14][7] . In their studies, each style had a distinct profile of interpersonal problems, according to both self-reports and friends’ reports[14].

The anxiety and avoidance dimensions

Brennan, Clark and Shaver gathered more than 320 items from existing self-report measures, gave them to a large sample, and found that the variation could be organized in two dimensions. They called the first anxiety: vigilance about rejection and abandonment. They called the second avoidance: discomfort with closeness and with depending on others[7]. Their questionnaire, the Experiences in Close Relationships scales, was among the inventories that Fraley, Waller and Brennan analyzed with item response theory to show how scales with better measurement properties could be built; the revised scales are known as the ECR-R[15][16].

The two dimensions of adult attachment and Bartholomew's four categoriesA square plane divided into four quadrants by dashed lines. The horizontal axis is attachment-related avoidance, discomfort with closeness and dependence, rising from left to right. The vertical axis is attachment-related anxiety, fear of rejection and abandonment, rising from bottom to top. Low anxiety with low avoidance, bottom left, is secure: a positive model of self and of others. High anxiety with low avoidance, top left, is preoccupied: a negative model of self and a positive model of others. Low anxiety with high avoidance, bottom right, is dismissing-avoidant: a positive model of self and a negative model of others. High anxiety with high avoidance, top right, is fearful-avoidant: negative models of both. The boundaries are illustrative, because scores vary continuously.PreoccupiedNegative model of self,positive model of othersFearful-avoidantNegative model of self,negative model of othersSecurePositive model of self,positive model of othersDismissing-avoidantPositive model of self,negative model of otherslowhighAVOIDANCE · DISCOMFORT WITH CLOSENESSANXIETY · FEAR OF REJECTIONlowhighDashed boundaries are illustrative: scores vary continuously.

Figure 1Self-report research describes adult attachment with two continuous dimensions, anxiety and avoidance, and places Bartholomew’s four categories in the quadrants. After Bartholomew and Horowitz (1991) and Brennan, Clark and Shaver (1998), as summarized by Fraley and Shaver (2000). The categories are regions of a continuous space, not types, and the drawing is not a diagnostic tool.

Text description

A square plane divided into four quadrants by dashed lines. The horizontal axis is attachment-related avoidance, discomfort with closeness and dependence, rising from left to right. The vertical axis is attachment-related anxiety, fear of rejection and abandonment, rising from bottom to top. Low anxiety with low avoidance, bottom left, is secure: a positive model of self and of others. High anxiety with low avoidance, top left, is preoccupied: a negative model of self and a positive model of others. Low anxiety with high avoidance, bottom right, is dismissing-avoidant: a positive model of self and a negative model of others. High anxiety with high avoidance, top right, is fearful-avoidant: negative models of both. A note under the drawing states that the dashed boundaries are illustrative because scores vary continuously.

The two dimensions map onto Bartholomew’s: anxiety corresponds to the model of self, and avoidance to the model of others[7]. Her four categories become regions of the space. Secure and dismissing adults are low in anxiety; preoccupied and fearful adults are high in it; secure and preoccupied adults are low in avoidance; dismissing and fearful adults are high in it[7]. The same two functions appear in infancy: in Ainsworth’s original data, one combination of behavior scales separated resistant infants from the rest, and another separated avoidant infants from the rest[7].

Fearful-avoidant attachment is a region of this self-report space. It is a different construct from infant disorganization, which is coded from observed behavior in the Strange Situation[6][7].

Interview and questionnaire measures

The Adult Attachment Interview asks adults about their childhood attachment relationships and classifies their current state of mind from the coherence of the account, not from whether the experiences were good or bad: autonomous-secure, dismissing or preoccupied, with an additional code for unresolved loss or trauma[10][17]. In combined samples of North American non-clinical mothers, 58% were classified secure, 23% dismissing and 19% preoccupied, with 18% additionally coded unresolved. Men and women did not differ in the use of dismissing versus preoccupied strategies, and distributions were largely independent of language and country; clinical groups showed more insecure and unresolved classifications[17].

The interview and the questionnaires measure different things. A meta-analysis of 10 studies (combined N = 961) found a mean correlation of .09 between interview security and self-reported attachment dimensions, trivial to small by conventional criteria, and the two kinds of measure predicted somewhat distinct aspects of adult relationships[5]. A person classified as secure on one measure is therefore not necessarily secure on the other.

Three ways of measuring attachment, aligned by patternA table with three columns, one per measurement tradition, and four rows, one per pattern. The Strange Situation observes infants about one year old; the Adult Attachment Interview codes an adult's transcript about childhood; two-dimension self-report asks adults about close relationships. Secure: secure (B) in infants, secure on the interview, low anxiety and low avoidance on questionnaires. Avoidant: avoidant (A), dismissing, and high avoidance, called dismissing-avoidant when anxiety is low. Anxious: ambivalent or resistant (C), preoccupied, and high anxiety, called preoccupied when avoidance is low. Disorganized: disorganized (D) in infants and unresolved on the interview, which is coded in addition to a main classification; the two-dimension model has no counterpart. Under the table, a badge between the first two columns reads 75% match: a parent's interview classification and the infant's classification agree on secure versus insecure in 75 percent of pairs. A badge between the second and third columns reads r = .09: security measured by interview and by questionnaire correlates only weakly.StrangeSituationInfants aboutone year old;observedbehaviorAdultAttachmentInterviewAdults; codedtranscript aboutchildhoodTwo-dimensionself-reportAdults;questionnaire oncloserelationshipsSecureSecure (B)SecureLow anxiety,lowavoidanceAvoidantAvoidant (A)DismissingHighavoidanceDismissing-avoidant whenanxiety is lowAnxiousAmbivalentor resistant(C)PreoccupiedHigh anxietyPreoccupiedwhen avoidanceis lowDisorganizedDisorganized(D)UnresolvedCoded inaddition to amainclassificationNocounterpartThe model hastwo dimensionsonly75% matchr = .0975% match: A parent’s interview classification andthe infant’s Strange Situation classification agreeon secure versus insecure in 75% of pairs.r = .09: Security measured by interview and byquestionnaire correlates only weakly (meta-analysis,N = 961).

Figure 2Attachment research uses three main measurement traditions, and their category labels line up more neatly than their results. A parent’s Adult Attachment Interview predicts the infant’s security in the Strange Situation, but interview and questionnaire measures of adult security barely correlate. After Ainsworth, Blehar, Waters and Wall (1978), van IJzendoorn, Schuengel and Bakermans-Kranenburg (1999), Bakermans-Kranenburg and van IJzendoorn (2009) and Fraley and Shaver (2000); figures from van IJzendoorn (1995) and Roisman and colleagues (2007).

Text description

A table with three columns, one for each measurement tradition, and four rows, one for each pattern. The Strange Situation observes the behavior of infants about one year old; the Adult Attachment Interview codes the transcript of an adult describing childhood; two-dimension self-report questionnaires ask adults about close relationships. In the secure row the entries are secure (B), secure, and low anxiety with low avoidance. In the avoidant row they are avoidant (A), dismissing, and high avoidance, called dismissing-avoidant when anxiety is low. In the anxious row they are ambivalent or resistant (C), preoccupied, and high anxiety, called preoccupied when avoidance is low. In the disorganized row they are disorganized (D) and unresolved, which the interview codes in addition to a main classification; the two-dimension model has no counterpart. Under the table, a badge between the first two columns reads “75% match”: a parent’s interview classification and the infant’s classification agree on secure versus insecure in 75 percent of pairs. A badge between the second and third columns reads “r = .09”: across studies with 961 participants, security measured by interview and by questionnaire correlates only weakly. The rows align labels; they do not show that the categories measure the same thing.

Categories versus dimensions

The labels suggest types, but the evidence favors dimensions. Fraley and Waller applied taxometric methods, which can distinguish latent types from latent continua, to adult romantic attachment and found the data more consistent with a dimensional model[7]. Fraley and Spieker applied the same methods to Strange Situation data from 1,139 fifteen-month-olds and found variation in attachment patterns largely continuous, not categorical[1]. A later analysis of two samples reached the same conclusion for general attachment and for attachment in specific relationships with parents and peers[2].

Duschinsky suggests that the categorical look of Ainsworth’s original data may partly reflect how her Baltimore families were recruited, since a sample chosen for contrast makes extremes more likely[11]. Under a dimensional model, a label such as avoidant names a region of a continuous distribution, and two people on either side of a boundary differ by degree, not in kind.

Associations with later development

Infant classifications predict later development only modestly. Across 69 samples, insecure attachment was associated with children’s externalizing problems (d = 0.31), with the largest effect for disorganization (d = 0.34) and weaker effects for avoidance and resistance[3]. Across 42 samples, the association with internalizing symptoms was smaller (d = 0.15) and significant for avoidance but not for resistance or disorganization[4]. Security itself is moderately stable from infancy to early adulthood, and stability is lower over long intervals and in samples facing adversity[18].

Insecure attachment and reactive attachment disorder

Reactive attachment disorder is a psychiatric diagnosis. It describes children who lack attachments despite the developmental capacity to form them; its companion diagnosis, disinhibited social engagement disorder, describes children who lack normal reticence with unfamiliar adults[8]. Both are defined for children reared in conditions of severe neglect and are expressed across settings and with different adults[6].

An insecurely attached child, by contrast, has an attachment: the classification describes how the child organizes behavior toward a particular caregiver, and a child can show different patterns with different caregivers[6][11]. Zeanah and Gleason note that deciding when aberrant behavior is better explained by an attachment disorder than by insecure attachment has caused confusion, in part because attachments are nearly universal in humans[8].

Sources18

Numbered in order of first citation.

  1. 1.Fraley, R. C., & Spieker, S. J. (2003). Are infant attachment patterns continuously or categorically distributed? A taxometric analysis of strange situation behavior. Developmental Psychology, 39(3), 387–404. https://doi.org/10.1037/0012-1649.39.3.387
  2. 2.Fraley, R. C., Hudson, N. W., Heffernan, M. E., & Segal, N. (2015). Are adult attachment styles categorical or dimensional? A taxometric analysis of general and relationship-specific attachment orientations. Journal of Personality and Social Psychology, 109(2), 354–368. https://doi.org/10.1037/pspp0000027
  3. 3.Fearon, R. P., Bakermans-Kranenburg, M. J., van Ijzendoorn, M. H., Lapsley, A. M., & Roisman, G. I. (2010). The significance of insecure attachment and disorganization in the development of children’s externalizing behavior: A meta-analytic study. Child Development, 81(2), 435–456. https://doi.org/10.1111/j.1467-8624.2009.01405.x
  4. 4.Groh, A. M., Roisman, G. I., van Ijzendoorn, M. H., Bakermans-Kranenburg, M. J., & Fearon, R. P. (2012). The significance of insecure and disorganized attachment for children’s internalizing symptoms: A meta-analytic study. Child Development, 83(2), 591–610. https://doi.org/10.1111/j.1467-8624.2011.01711.x
  5. 5.Roisman, G. I., Holland, A., Fortuna, K., Fraley, R. C., Clausell, E., & Clarke, A. (2007). The Adult Attachment Interview and self-reports of attachment style: An empirical rapprochement. Journal of Personality and Social Psychology, 92(4), 678–697. https://doi.org/10.1037/0022-3514.92.4.678
  6. 6.Granqvist, P., Sroufe, L. A., Dozier, M., Hesse, E., Steele, M., van Ijzendoorn, M., Solomon, J., Schuengel, C., Fearon, P., Bakermans-Kranenburg, M., Steele, H., Cassidy, J., Carlson, E., Madigan, S., Jacobvitz, D., Foster, S., Behrens, K., Rifkin-Graboi, A., Gribneau, N., … Duschinsky, R. (2017). Disorganized attachment in infancy: a review of the phenomenon and its implications for clinicians and policy-makers. Attachment & Human Development, 19(6), 534–558. https://doi.org/10.1080/14616734.2017.1354040
  7. 7.Fraley, R. C., & Shaver, P. R. (2000). Adult romantic attachment: Theoretical developments, emerging controversies, and unanswered questions. Review of General Psychology, 4(2), 132–154. https://doi.org/10.1037/1089-2680.4.2.132
  8. 8.Zeanah, C. H., & Gleason, M. M. (2015). Annual Research Review: Attachment disorders in early childhood – clinical presentation, causes, correlates, and treatment. Journal of Child Psychology and Psychiatry, 56(3), 207–222. https://doi.org/10.1111/jcpp.12347
  9. 9.Ainsworth, M. D. S., Blehar, M. C., Waters, E., & Wall, S. (1978). Patterns of attachment: A psychological study of the strange situation. Lawrence Erlbaum.
  10. 10.Bretherton, I. (1992). The origins of attachment theory: John Bowlby and Mary Ainsworth. Developmental Psychology, 28(5), 759–775. https://doi.org/10.1037/0012-1649.28.5.759
  11. 11.Duschinsky, R. (2020). Cornerstones of attachment research. Oxford University Press.
  12. 12.van Ijzendoorn, M. H., Schuengel, C., & Bakermans-Kranenburg, M. J. (1999). Disorganized attachment in early childhood: Meta-analysis of precursors, concomitants, and sequelae. Development and Psychopathology, 11(2), 225–250. https://doi.org/10.1017/S0954579499002035
  13. 13.Hazan, C., & Shaver, P. (1987). Romantic love conceptualized as an attachment process. Journal of Personality and Social Psychology, 52(3), 511–524. https://doi.org/10.1037/0022-3514.52.3.511
  14. 14.Bartholomew, K., & Horowitz, L. M. (1991). Attachment styles among young adults: A test of a four-category model. Journal of Personality and Social Psychology, 61(2), 226–244. https://doi.org/10.1037/0022-3514.61.2.226
  15. 15.Fraley, R. C., Waller, N. G., & Brennan, K. A. (2000). An item response theory analysis of self-report measures of adult attachment. Journal of Personality and Social Psychology, 78(2), 350–365. https://doi.org/10.1037/0022-3514.78.2.350
  16. 16.Ravitz, P., Maunder, R., Hunter, J., Sthankiya, B., & Lancee, W. (2010). Adult attachment measures: A 25-year review. Journal of Psychosomatic Research, 69(4), 419–432. https://doi.org/10.1016/j.jpsychores.2009.08.006
  17. 17.Bakermans-Kranenburg, M. J., & van IJzendoorn, M. H. (2009). The first 10,000 Adult Attachment Interviews: distributions of adult attachment representations in clinical and non-clinical groups. Attachment & Human Development, 11(3), 223–263. https://doi.org/10.1080/14616730902814762
  18. 18.Pinquart, M., Feussner, C., & Ahnert, L. (2013). Meta-analytic evidence for stability in attachments from infancy to early adulthood. Attachment & Human Development, 15(2), 189–218. https://doi.org/10.1080/14616734.2013.746257

Further reading

  • Fraley, R. C., & Shaver, P. R. (2000). Adult romantic attachment: Theoretical developments, emerging controversies, and unanswered questions. Review of General Psychology, 4(2), 132–154. https://doi.org/10.1037/1089-2680.4.2.132
  • Granqvist, P., Sroufe, L. A., Dozier, M., Hesse, E., Steele, M., van Ijzendoorn, M., Solomon, J., Schuengel, C., Fearon, P., Bakermans-Kranenburg, M., Steele, H., Cassidy, J., Carlson, E., Madigan, S., Jacobvitz, D., Foster, S., Behrens, K., Rifkin-Graboi, A., Gribneau, N., … Duschinsky, R. (2017). Disorganized attachment in infancy: a review of the phenomenon and its implications for clinicians and policy-makers. Attachment & Human Development, 19(6), 534–558. https://doi.org/10.1080/14616734.2017.1354040
  • Cassidy, J., & Shaver, P. R. (2016). Handbook of attachment: Theory, research, and clinical applications (3rd ed.). Guilford Press.Edited volume; Cassidy and Shaver are the editors.

Cite this entry

  • APA 7

    Psych Dictionary. (2026, October 18). Attachment styles. In Psych Dictionary. https://www.psychdictionary.com/concepts/attachment-styles

  • MLA 9

    “Attachment Styles.” Psych Dictionary, 18 Oct. 2026, www.psychdictionary.com/concepts/attachment-styles.

Written by Editorial team. First published October 18, 2026. Editorial standards

Educational reference material. Nothing on this site is a diagnosis, a treatment recommendation or a substitute for care from a qualified clinician.