Adult Attachment Scale- Revised

1. Abstract

The Adult Attachment Scale- Revised (AAS-R), developed by social psychologist Nancy L. Collins in 1996, is a premier self-report psychometric instrument designed to assess individual differences in adult attachment styles within interpersonal and romantic relationships. Evolving from the foundational work of Collins and Read (1990) and inspired by the seminal categorical conceptualizations of attachment theory pioneered by John Bowlby, Mary Ainsworth, and Hazan and Shaver (1987), the AAS-R provides a continuous, multidimensional measurement framework. The scale consists of 18 items rated on a 5-point Likert response format ranging from 1 (“Not at all characteristic of me”) to 5 (“Very characteristic of me”). It is available in two parallel contextual forms: a romantic relationships version and a generalized close relationships version.

Psychometrically, the AAS-R can be scored according to two validated factor models: a triadic model yielding three six-item subscales—Close (comfort with intimacy and emotional closeness), Depend (confidence in relying on others and perceptions of partner availability), and Anxiety (fear of abandonment, rejection, and unreciprocated love)—and an alternative, highly robust dyadic model aligned with modern attachment taxonomies yielding Attachment Anxiety (6 items) and Attachment Avoidance (12 items, derived from composite scoring of the Close and Depend items). Empirical evaluations confirm that the AAS-R exhibits robust psychometric properties across diverse populations, with internal consistency reliability estimates typically ranging from α = .77 to .88 for Anxiety, .75 to .85 for Close, and .72 to .82 for Depend. Confirmatory factor analyses support structural validity, and extensive studies confirm strong convergent, discriminant, and predictive validity across studies investigating romantic relationship satisfaction, cognitive attributions, affect regulation, conflict resolution strategies, and mental health outcomes.

2. Keywords

Adult Attachment Scale- Revised, AAS-R, Nancy Collins, attachment theory, internal working models, attachment anxiety, attachment avoidance, relational intimacy, romantic relationships, psychometrics, relational trust, dependency

3. Authors

The Adult Attachment Scale- Revised was formulated and revised by Nancy L. Collins, Ph.D., Professor of Psychological and Brain Sciences at the University of California, Santa Barbara (UCSB). Dr. Collins received her doctorate in social psychology from the University of Southern California (USC), where she conducted foundational research on relationship science and adult attachment alongside Stephen J. Read.

Primary Affiliation and Inquiries:
Department of Psychological & Brain Sciences
University of California, Santa Barbara
Santa Barbara, CA 93106-9660, USA
Email: [email protected]
Laboratory: Close Relationships Lab at UCSB

4. Purpose

The primary purpose of the Adult Attachment Scale- Revised is to provide an empirically sound, continuous measure of individual variation in internal working models of self and others in the context of adult relationships. In early adult attachment research, attachment styles were conceptualized categorically, sorting individuals rigidly into secure, avoidant, or anxious/ambivalent classifications (Hazan & Shaver, 1987). However, categorical typologies suffered from notable methodological limitations, including loss of statistical power, inability to capture borderline or mixed styles, and the implicit assumption of homogeneity within types. The AAS-R was engineered specifically to overcome these constraints by measuring the underlying cognitive-affective dimensions that generate these stylistic variations.

In academic research, the instrument is widely deployed across social, developmental, clinical, and personality psychology. It allows investigators to operationalize attachment not as an all-or-nothing category, but as an interplay of continuous dimensions. Researchers use the AAS-R to explore how working models govern cognitive processing—such as cognitive appraisal, causal attributions, and memory retrieval regarding relationship events (Collins, 1996; Collins et al., 2006). Moreover, it serves as a critical variable in studies investigating dyadic caregiving, support-seeking, nonverbal behavioral expressions during distress, physiological stress responses (e.g., cortisol reactivity and cardiovascular dynamics), and vulnerability to psychological distress, such as depressive and anxiety disorders.

In clinical and counseling settings, the AAS-R provides practitioners with a nuanced diagnostic profile of a client’s relational tendencies. Rather than labeling an individual solely as “insecure,” the scale reveals whether the insecurity stems primarily from hyperactivating strategies (intense fear of abandonment, vigilance, and emotional hyper-reactivity captured by the Anxiety dimension) or deactivating strategies (defensive self-reliance, discomfort with intimacy, and emotional suppression captured by the reverse of Close and Depend, or the overarching Avoidance dimension). This dimensional profiling informs evidence-based couples therapy (such as Emotion-Focused Therapy), individual psychotherapy, and relational interventions, enabling therapists to tailor relational interventions to target specific maladaptive internal representations.

5. Psychological Construct

The Adult Attachment Scale- Revised measures the cognitive, emotional, and behavioral components that constitute adult attachment representations. These internal representations, originally conceptualized by Bowlby as internal working models, are mental schemas reflecting expectations regarding the self’s worthiness of love and the responsiveness and reliability of relational partners. The instrument captures these models across distinct yet interconnected dimensions.

The Triadic Dimensions (Close, Depend, Anxiety)

• Close (Comfort with Emotional Closeness): This subscale measures the degree to which an individual feels comfortable with intimacy, emotional proximity, and vulnerability. High scores indicate an ease with sharing personal feelings, a lack of defensive withdrawal, and an absence of distress when others seek closeness. For example, endorsing statements like “I find it relatively easy to get close to people” or rejecting “I am somewhat uncomfortable being close to others” reflects a strong comfort with interpersonal closeness. Low scores capture intimacy avoidance, marked by self-protective distancing and distress when partners seek deeper emotional engagement.
• Depend (Confidence in Partner Availability and Trust): This subscale assesses the extent to which an individual feels confident relying on other people for emotional and instrumental support in times of need. It measures the fundamental belief that attachment figures are reliable, dependable, and benevolent. High scores reflect trust in partner responsiveness (e.g., “I know that people will be there when I need them”), whereas low scores indicate deep cynicism regarding others’ availability (e.g., “I find that people are never there when you need them”) and compulsive, defensive self-reliance.
• Anxiety (Fear of Rejection and Abandonment): This dimension assesses chronic concerns regarding unrequited love, abandonment, and relationship instability. It operationalizes the working model of the self as deficient, unlovable, or easily discarded. Individuals scoring high on this dimension display relational hypervigilance, frequently questioning their partner’s emotional investment (e.g., “I often worry that romantic partners don’t really love me” and “When I show my feelings for others, I’m afraid they will not feel the same about me”).

The Alternative Two-Dimensional Model (Anxiety and Avoidance)

In alignment with the integrative psychometric synthesis of attachment scales presented by Brennan, Clark, and Shaver (1998), Collins provided an alternative two-dimensional scoring system for the AAS-R. In this framework:

• Attachment Anxiety: Remains identical to the original 6-item Anxiety subscale, measuring chronic relational insecurity, hyperactivating strategies, and fear of abandonment.
• Attachment Avoidance: Formed by aggregating and reverse-scoring the 12 items comprising the Close and Depend subscales. This unified 12-item dimension captures deactivating strategies, emotional detachment, behavioral independence, and active resistance to relational dependency and intimacy.

By mapping these two orthogonal dimensions onto a Cartesian coordinate system, researchers can re-derive the four classic attachment prototype categories defined by Bartholomew and Horowitz (1991): Secure (Low Anxiety, Low Avoidance), Preoccupied (High Anxiety, Low Avoidance), Dismissing-Avoidant (Low Anxiety, High Avoidance), and Fearful-Avoidant (High Anxiety, High Avoidance).

6. Theoretical Framework

The theoretical bedrock of the Adult Attachment Scale- Revised is Bowlby’s attachment theory (1969, 1973, 1980), originally developed to describe the biobehavioral bond between human infants and primary caregivers. Bowlby postulated that infants possess an innate psychobiological system—the attachment behavioral system—designed to maintain proximity to supportive caregivers (attachment figures) to ensure survival. Through repeated interactions with caregivers, infants construct cognitive-affective representations, termed internal working models. These models include a model of self (reflecting whether one is worthy of love and support) and a model of others (reflecting whether social partners are accessible, trustworthy, and supportive).

In 1987, Cindy Hazan and Phillip Shaver revolutionized social psychology by demonstrating that the attachment behavioral system remains operative across the lifespan, serving as the foundational organizing principle of adult romantic relationships. They argued that romantic love functions as an attachment process wherein adult partners serve as primary attachment figures who provide a “safe haven” during acute distress and a “secure base” from which to explore the social and physical environment.

Building upon Hazan and Shaver’s three-category classification, Nancy Collins and Stephen Read (1990) conducted factor-analytic investigations of the original descriptions of attachment styles. Their research demonstrated that adult attachment is best decomposed into continuous dimensions reflecting comfort with closeness, trust in dependency, and anxiety over abandonment. In her landmark 1996 publication, Collins refined these items into the AAS-R to articulate a comprehensive model of working models as dynamic social-cognitive architectures. According to Collins (1996), internal working models operate via four interconnected components:

1. Memories of Attachment Experiences: Episodic and autobiographical records of past caregiving and romantic histories.
2. Beliefs and Expectations: Core generalized cognitive schemas regarding the benevolence of others and the self’s relational worth.
3. Attachment-Related Goals and Needs: Motivational orientations driving individuals toward proximity-seeking or self-preservation.
4. Strategies for Regulating Distress: Regulatory coping mechanisms, divided broadly into hyperactivating strategies (intensified proximity-seeking, emotional amplification, and constant vigilance) and deactivating strategies (suppression of attachment needs, emotional numbing, and compulsive autonomy).

7. Validity

The AAS-R has undergone extensive psychometric validation across thousands of participants in varied empirical investigations.

Construct and Convergent Validity

Extensive research demonstrates robust convergent validity with alternative attachment assessment batteries. In their comprehensive psychometric synthesis, Brennan, Clark, and Shaver (1998) found that the AAS-R subscales correlate strongly with corresponding dimensions of the Experiences in Close Relationships (ECR) inventory. The AAS-R Anxiety subscale correlates highly (r > .80) with the ECR Anxiety subscale, while the AAS-R composite Avoidance score (or inverted Close/Depend scores) correlates strongly (r > .75) with the ECR Avoidance subscale. Furthermore, Bartholomew and Shaver (1998) established significant convergence between AAS-R self-report dimensions and peer-report ratings, interview-coded adult attachment classifications (such as the Adult Attachment Interview), and romantic partner reports.

Predictive and Criterion Validity

The predictive utility of the AAS-R has been confirmed across diverse domains of interpersonal functioning:

• Cognitive Attributions: In experimental and longitudinal designs, Collins (1996) and Collins et al. (2006) showed that high Anxiety and low Depend/Close scores predict maladaptive causal attributions for hypothetical and real partner transgressions. Highly anxious individuals consistently attribute partner unresponsiveness to intentional malice or lack of love, fostering cyclical distress.
• Relationship Satisfaction and Stability: In dating and married couples, scores on the Close and Depend subscales positively predict relationship commitment, trust, dyadic consensus, and sexual satisfaction, whereas high Anxiety scores predict relationship dissolution, chronic jealousy, and low dyadic adjustment (Collins & Read, 1990).
• Caregiving and Support Interactions: Observational studies of couples undergoing laboratory-induced stress tasks indicate that individuals with low scores on Close and Depend exhibit defensive behavioral withdrawal and fail to offer responsive caregiving when their partner is in distress (Collins et al., 2006).

Discriminant Validity

The AAS-R dimensions demonstrate clear discriminant validity from generalized personality traits, such as the Big Five (Collins & Read, 1990). Although Attachment Anxiety shares modest variance with Neuroticism (r ≈ .35 to .45), it accounts for significant unique variance in relationship-specific cognitions and behaviors above and beyond neurotic traits. Similarly, while Avoidance and Close relate modestly to Extraversion and Agreeableness, they uniquely predict intimacy avoidance and proximity-seeking behaviors in close dyads.

8. Reliability

The Adult Attachment Scale- Revised demonstrates consistently high internal consistency and longitudinal stability across community, collegiate, and clinical samples.

Internal Consistency

Across validation studies by Collins (1996), Collins and Read (1990), and subsequent independent psychometric evaluations, the Cronbach’s alpha coefficients for the AAS-R subscales fall within acceptable to excellent ranges:

• Close Subscale (6 items): α = .75 to .85 (e.g., α = .77 in Collins, 1996; α = .82 in Collins et al., 2006).
• Depend Subscale (6 items): α = .72 to .82 (e.g., α = .78 in Collins, 1996; α = .80 in Collins et al., 2006).
• Anxiety Subscale (6 items): α = .77 to .88 (e.g., α = .85 in Collins, 1996; α = .86 in Collins et al., 2006).
• Avoidance Composite (12 items): α = .83 to .89 across multiple independent cohorts.

Test-Retest Reliability

Longitudinal investigations confirm that adult attachment working models exhibit marked stability over time, consistent with theoretical expectations of trait-like schemas. Collins and Read (1990) reported test-retest correlations over a 2-month interval of r = .68 for Close, r = .71 for Depend, and r = .52 for Anxiety. Over longer intervals ranging from 8 months to 2 years, test-retest coefficients remain stable between .50 and .65, reflecting both foundational cognitive schema preservation and meaningful sensitivity to significant relational life events (such as relationship dissolution, marriage, or therapeutic intervention).

9. Factor Analysis

The structural integrity of the AAS-R has been extensively tested using both Exploratory Factor Analysis (EFA) and Confirmatory Factor Analysis (CFA).

Exploratory Factor Analysis

The initial derivation by Collins and Read (1990) utilized principal components analysis with varimax and oblimin rotations on the expanded item pool derived from Hazan and Shaver’s work. The extraction clearly revealed three dominant factors accounting for approximately 52% to 58% of the total variance:

1. Factor 1 (Depend): Items reflecting the perceived dependability and reliability of others, with primary factor loadings ranging from .54 to .78.
2. Factor 2 (Anxiety): Items reflecting chronic fear of rejection and being unloved, with factor loadings between .58 and .82.
3. Factor 3 (Close): Items reflecting comfort with intimacy, closeness, and vulnerability, with factor loadings ranging from .50 to .75.

Confirmatory Factor Analysis and Model Comparison

Subsequent psychometric studies have tested competing structural models using CFA. A three-factor first-order model (Close, Depend, Anxiety) demonstrates adequate to good fit across multiple fit indices: χ²/df < 2.5, Comparative Fit Index (CFI) ≥ .92, Tucker-Lewis Index (TLI) ≥ .90, and Root Mean Square Error of Approximation (RMSEA) ≤ .06.

When evaluated against the two-dimensional model of adult attachment (Anxiety vs. Avoidance), higher-order CFA models indicate that the Close and Depend latent factors load heavily onto a generalized higher-order Attachment Avoidance construct (loadings > |.70|). Consequently, structural equation modeling literature supports both representations: the three-factor model is valuable when researchers seek to distinguish between comfort with intimacy (Close) and interpersonal trust (Depend), whereas the two-factor model provides optimal parsimony and direct comparability with instruments such as the ECR (Brennan et al., 1998).

10. Instrument / Measurement Tool

• Instrument Name: Adult Attachment Scale- Revised (AAS-R)
• Original Author: Nancy L. Collins, Ph.D. (1996)
• Target Population: Adults and adolescents (ages 16 and above) involved in or with experience in romantic or close interpersonal relationships.
• Administration Format: Self-report paper-and-pencil or computerized questionnaire.
• Administration Time: Approximately 3 to 5 minutes.
• Total Item Count: 18 items.
• Contextual Versions:
  • Romantic Relationships Version: Items explicitly reference romantic partners (e.g., “romantic partners don’t really love me”).
  • Close Relationships Version: Items reference interpersonal relationships broadly (e.g., “other people don’t really love me”).
• Response Scale: 5-point Likert-type scale with numerical anchors:
  • 1 = Not at all characteristic of me
  • 2 = Slightly characteristic of me
  • 3 = Moderately characteristic of me
  • 4 = Substantially characteristic of me
  • 5 = Very characteristic of me
• Subscales and Primary Scoring Rules (Triadic Model):
  • CLOSE Subscale (6 items): Measures comfort with emotional closeness. Compute the mean score of Items 1, 6, 8*, 12, 13*, and 17*. (*Items marked with an asterisk are reverse-scored: 1→5, 2→4, 3→3, 4→2, 5→1).
  • DEPEND Subscale (6 items): Measures trust and comfort in depending on others. Compute the mean score of Items 2*, 5, 7*, 14, 16*, and 18*. (*Items marked with an asterisk are reverse-scored).
  • ANXIETY Subscale (6 items): Measures chronic anxiety over rejection, unreciprocated love, and abandonment. Compute the mean score of Items 3, 4, 9, 10, 11, and 15 (no reverse scoring).
• Alternative Scoring Rules (Dyadic Model – Anxiety vs. Avoidance):
  • ANXIETY Dimension (6 items): Mean of Items 3, 4, 9, 10, 11, and 15.
  • AVOIDANCE Dimension (12 items): Formed by combining the reverse of the Close and Depend items. Specifically, reverse-score items that indicate comfort with closeness/trust, and retain raw scores for items that indicate discomfort/distrust. Formula: Compute the mean of Items 1*, 2, 5*, 6*, 7, 8, 12*, 13, 14*, 16, 17, and 18.

11. Permissions & Fee and Test Year

The Adult Attachment Scale- Revised was formally established and published in its revised iteration by Nancy L. Collins in 1996, emerging from earlier versions developed in Collins and Read (1990). Dr. Collins placed the instrument into the academic public domain for scientific research and educational purposes. No licensing fees or purchase royalties are required for non-commercial academic research, institutional teaching, or clinical evaluation.

Researchers wishing to use the AAS-R in scholarly investigations are permitted to duplicate and administer the scale freely, provided that proper bibliographic citation is given to the primary validation publications (Collins & Read, 1990; Collins, 1996). Inquiries regarding commercial administration, software translation licensing, or requests for unpublished research manuals may be addressed directly to Dr. Nancy L. Collins via email ([email protected]).

12. References

• 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
• Bartholomew, K., & Shaver, P. R. (1998). Methods of assessing adult attachment: Do they converge? In J. A. Simpson & W. S. Rholes (Eds.), Attachment theory and close relationships (pp. 25–45). Guilford Press.
• Bowlby, J. (1969). Attachment and loss: Vol. 1. Attachment. Basic Books.
• Bowlby, J. (1973). Attachment and loss: Vol. 2. Separation: Anxiety and anger. Basic Books.
• Bowlby, J. (1980). Attachment and loss: Vol. 3. Loss: Sadness and depression. Basic Books.
• Brennan, K. A., Clark, C. L., & Shaver, P. R. (1998). Self-report measurement of adult attachment: An integrative overview. In J. A. Simpson & W. S. Rholes (Eds.), Attachment theory and close relationships (pp. 46–76). Guilford Press.
• Collins, N. L. (1996). Working models of attachment: Implications for explanation, emotion, and behavior. Journal of Personality and Social Psychology, 71(4), 810–832. https://doi.org/10.1037/0022-3514.71.4.810
• Collins, N. L., Ford, M. B., Guichard, A. C., & Allard, L. M. (2006). Working models of attachment and attribution processes in intimate relationships. Personality and Social Psychology Bulletin, 32(2), 201–219. https://doi.org/10.1177/0146167205280907
• Collins, N. L., Ford, M. B., Guichard, A. C., & Feeney, B. C. (2006). Responding to need in intimate relationships: Normative processes and individual differences. In M. Mikulincer & G. S. Goodman (Eds.), Dynamics of romantic love: Attachment, caregiving, and sex (pp. 149–189). Guilford Press.
• Collins, N. L., & Read, S. J. (1990). Adult attachment, working models, and relationship quality in dating couples. Journal of Personality and Social Psychology, 58(4), 644–663. https://doi.org/10.1037/0022-3514.58.4.644
• 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

13. Items of the Scale (Questionnaire)