Attachment Style Measure

Abstract

The Attachment Style Measure, developed by Jeffry A. Simpson in 1990 (frequently designated in the psychometric literature as the Adult Attachment Questionnaire or Simpson AAQ), represents a landmark transition in the operationalization of romantic attachment theory. Evolving from the original single-item categorical typology introduced by Cindy Hazan and Phillip Shaver in 1987, Simpson decomposed the three classic prototypical paragraphs—Secure, Avoidant, and Anxious/Ambivalent—into 13 discrete, self-report Likert-type items. Respondents evaluate their feelings, expectations, and behavioral tendencies in close relationships on a 7-point continuum ranging from 1 (Strongly Disagree) to 7 (Strongly Agree). Psychometrically, the instrument was conceived to yield either three continuous subscale scores corresponding to Hazan and Shaver’s primary styles or two orthogonal higher-order dimensions: Attachment Avoidance and Attachment Anxiety. In initial psychometric evaluations, internal consistency was robust for the Avoidant dimension (α = .79), whereas the Secure (α = .51) and Anxious/Ambivalent (α = .59) subscales demonstrated lower alpha coefficients, reflecting the constraints of decomposing categorical text without initial exploratory factor derivation. Subsequent empirical work has confirmed that reconfiguring the instrument into continuous dimensions markedly improves structural stability and predictive power. The scale has demonstrated profound predictive validity across experimental stress paradigms, longitudinal relationship dissolution studies, and dyadic observational research, laying the empirical groundwork for modern dimensional psychometrics in adult romantic attachment.

Keywords

Attachment Style Measure, Adult Attachment Questionnaire, Jeffry A. Simpson, Hazan and Shaver, adult attachment, romantic relationships, attachment anxiety, attachment avoidance, internal working models, psychometrics

Authors

The Attachment Style Measure was developed by Jeffry A. Simpson, Ph.D., currently a Distinguished University Teaching Professor and Chair of the Department of Psychology at the University of Minnesota Twin Cities. Simpson is a preeminent scholar in social and personality psychology whose theoretical and empirical contributions have profoundly shaped contemporary understanding of adult romantic attachment, interpersonal close relationships, evolutionary psychology, and social development across the lifespan. Contact information and institutional affiliation: Department of Psychology, University of Minnesota, 75 East River Road, Minneapolis, MN 55455, USA; Email: [email protected].

Purpose

The primary purpose of the Attachment Style Measure is to provide a continuous, nuanced self-report assessment of adult romantic attachment orientations. Prior to Simpson’s 1990 study, adult attachment research was heavily dominated by Hazan and Shaver’s (1987) discrete-choice paradigm, in which participants were forced to read three comprehensive descriptive paragraphs—representing secure, avoidant, and anxious/ambivalent styles—and self-categorize into a single best-fitting category. Although revolutionary, this forced-choice categorical approach suffered from severe psychometric limitations: it assumed taxonic distribution of relational styles, precluded the assessment of individual differences within categories, masked ambivalent or mixed attachment profiles, and severely restricted statistical power by preventing the use of standard continuous multivariate analyses such as multiple regression, structural equation modeling, and hierarchical linear modeling.

Simpson (1990) designed the 13-item measure to break these monolithic descriptions into their constituent cognitive, affective, and behavioral assertions. By measuring these items on an expanded 7-point Likert scale, the purpose was threefold:

• Dimensional Precision: To capture the exact degree to which individuals endorse secure, avoidant, or anxious features in romantic relationships, acknowledging that adult attachment exists along continuous gradients rather than discrete qualitative types.
• Predictive and Behavioral Granularity: To enable fine-grained prediction of relationship outcomes—including emotional distress during breakup, dyadic support-seeking and support-giving under stress, partner perception, and relationship satisfaction—over longitudinal intervals.
• Methodological Flexibility: To allow researchers to evaluate attachment either through classic tripartite subscale scores (Secure, Avoidant, Anxious) or through higher-order dimensional scores (Avoidance and Anxiety), or by algorithmic assignment to discrete categories for retrospective comparison with early categorical literature.

In clinical and applied settings, the instrument provides clinicians with an immediate, easily administered diagnostic snapshot of a client’s relational vulnerabilities. It reveals whether relational distress stems primarily from defensive emotional distancing, intimacy avoidance, and hyper-reliance on the self (Avoidant orientation), or from chronic fears of abandonment, hypervigilance toward partner unavailability, and excessive demands for closeness (Anxious orientation).

Psychological Construct

The Attachment Style Measure assesses the cognitive, emotional, and behavioral dynamics that govern adult romantic partnerships, anchored in the construct of Internal Working Models (IWMs) of self and others. Simpson’s operationalization targets three primary relational configurations, each defined by distinct relational expectations, defense mechanisms, and affective responses:

1. Secure Attachment (S)

The Secure attachment construct reflects positive models of both self and other. Secure individuals possess high self-worth, an internalized belief in their own lovability, and an enduring expectation that relational partners are trustworthy, responsive, and well-intentioned. Within the 13-item measure, items such as “I find it relatively easy to get close to others” and “I rarely worry about being abandoned by others” reflect an intrinsic comfort with emotional intimacy and mutual interdependence. Secure individuals do not fear self-disclosure, manage interpersonal conflict with constructive communication, and freely offer and solicit social support without experiencing crippling vulnerability or prideful resistance.

2. Avoidant Attachment (Av)

The Avoidant construct is defined by a negative model of others paired with a defensively self-reliant model of self. Avoidant individuals defensively suppress vulnerability, minimize the importance of emotional intimacy, and maintain psychological distance from relational partners. Captured by items such as “I’m somewhat uncomfortable being too close to others” and “I find it difficult to trust others completely,” this dimension indexes defensive deactivating strategies. When faced with relational stress or demands for deeper commitment, avoidant individuals pull away, experience acute discomfort when partners attempt to establish intimacy, and exhibit a pronounced aversion to depending on others or having others depend on them.

3. Anxious/Ambivalent Attachment (Ax)

The Anxious/Ambivalent construct reflects a negative model of self coupled with an idealized, hypervigilant focus on relational partners. Individuals scoring high on this dimension suffer from pervasive self-doubt, chronic insecurity regarding their partner’s affection, and intense fears of rejection and abandonment. This construct is manifested in items such as “I often worry that my partner(s) don’t really love me” and “I often want to merge completely with others, and this desire sometimes scares them away.” Anxious attachment employs hyperactivating affect-regulation strategies, wherein the attachment behavioral system is perpetually energized. High scores correspond to emotional volatility, intrusive monitoring of the partner, insatiable demands for reassurance, and severe distress during separations or relational transitions.

Theoretical Framework

The Attachment Style Measure is rooted in the evolutionary-developmental paradigm of John Bowlby (1969/1982) and the empirical traditions established by Mary Ainsworth and colleagues (1978). Bowlby posited that human infants possess an innate behavioral system designed to maintain physical and emotional proximity to primary caregivers (attachment figures) to maximize survival under threat. Based on parental responsiveness, infants construct internal working models—enduring cognitive schemas containing generalized representations of relational partners as responsive or rejecting, and the self as worthy or unworthy of care.

In 1987, Cindy Hazan and Phillip Shaver translated Bowlby and Ainsworth’s paradigm to adult romantic love. They hypothesized that the romantic bond between adult romantic partners functions via the exact same biological and psychodynamic attachment system that binds infants to caregivers. Adults exhibit identical prototype patterns: Secure adults experience love as trusting and enduring; Avoidant adults experience love with fear of intimacy and emotional distance; and Anxious/Ambivalent adults experience love as an overwhelming obsession characterized by jealousy, insecurity, and desire for extreme union.

Simpson (1990) recognized that while Hazan and Shaver’s categorical theory provided intuitive face validity, it artificially dichotomized continuous psychological traits. Drawing from cognitive schema theory and social psychometrics, Simpson posited that adult attachment orientations operate as continuous internalized beliefs. As Kim Bartholomew and Leonard Horowitz (1991) and later Brennan, Clark, and Shaver (1998) definitively demonstrated, adult attachment variations universally resolve into two fundamental, intersecting Cartesian axes:

• Attachment Anxiety (Model of Self): The degree to which an individual’s self-worth depends on interpersonal validation, reflecting the threshold of vulnerability to perceived abandonment.
• Attachment Avoidance (Model of Others): The degree to which an individual distrusts others and utilizes behavioral and cognitive avoidance strategies to circumvent intimate reliance.

Simpson’s 13-item measure served as the critical transitional bridge in attachment theory, moving the field from crude categorical taxonomy to continuous multidimensional measurement, enabling researchers to test systemic, dyadic interactions across stressful ecological contexts.

Validity

The Attachment Style Measure has undergone extensive construct, predictive, convergent, and discriminant validation across dozens of empirical investigations since its introduction in 1990.

Construct and Convergent Validity

Simpson (1990) examined the convergent validity of the continuous 13-item scores against Hazan and Shaver’s original forced-choice categorical measure in a sample of 144 heterosexual dating couples. When respondents completed both instruments, those categorized as Secure on Hazan and Shaver’s forced-choice item scored significantly higher on Simpson’s continuous Secure subscale than those categorized as Avoidant or Anxious (p < .001). Similarly, individuals categorized as Avoidant scored highest on the continuous Avoidant scale, and Anxious respondents scored highest on the Anxious scale. The continuous scales also demonstrated high convergent validity with the Adult Attachment Scale developed concurrently by Nancy Collins and Stephen Read (1990), displaying strong correlations between corresponding subscales measuring closeness, dependency, and anxiety.

Predictive and Behavioral Validity

The ultimate strength of Simpson’s measure lies in its predictive validity, established in several milestone observational and longitudinal studies:

• Relationship Dissolution and Emotional Distress: In Simpson’s (1990) original longitudinal investigation, couples were tracked over a six-month period. For relationships that dissolved, attachment styles measured at baseline robustly predicted post-breakup functioning. Anxious individuals experienced substantially greater emotional trauma, post-dissolution distress, and lingering obsession than secure individuals. Avoidant men displayed an emotionally blunted response immediately following breakups, consistent with deactivating defenses.
• Stress and Dyadic Support Interaction Paradigm: In a classic laboratory study, Simpson, Rholes, and Nelligan (1992) subjected dating couples to high anticipatory anxiety (informing women they would be exposed to an anxiety-provoking experimental procedure). Women’s scores on Simpson’s attachment measure dramatically predicted their support-seeking behavior: secure women sought more physical proximity and emotional reassurance from their partners as anxiety mounted. In contrast, avoidant women actively withdrew, seeking less comfort as their fear escalated. Furthermore, avoidant male partners responded to their partner’s fear with anger and emotional coldness rather than empathy.
• Transition to Parenthood: In longitudinal assessments of couples transitioning to parenthood, Simpson and colleagues demonstrated that pre-partum attachment avoidance and anxiety predicted post-partum declines in marital satisfaction, maladaptive parenting beliefs, and depressive symptomatology (Simpson, Rholes, Campbell, & Wilson, 2003).

Discriminant Validity

Research evaluating Simpson’s measure confirms that attachment dimensions are distinct from general personality traits such as the Big Five personality traits (Extraversion, Neuroticism, Agreeableness, Conscientiousness, and Openness). While Attachment Anxiety correlates moderately with Neuroticism (typically r ≈ .35 to .42), structural equation modeling confirms that attachment dimensions account for substantial unique variance in dyadic conflict, intimacy, and caregiving behaviors after controlling for Neuroticism and general trait anxiety.

Reliability

The internal consistency and temporal stability of the Attachment Style Measure have been rigorously evaluated across multiple samples:

Internal Consistency

In Simpson’s (1990) foundational validation study, Cronbach’s alpha coefficients varied across the three decomposed subscales:

• Avoidant Dimension (4 items): Demonstrated acceptable internal reliability with α = .79.
• Secure Dimension (5 items): Exhibited modest internal consistency with α = .51.
• Anxious/Ambivalent Dimension (4 items): Exhibited modest internal consistency with α = .59.

The lower alpha coefficients observed for the Secure and Anxious subscales in the 1990 study were directly attributable to Simpson’s methodological decision to preserve Hazan and Shaver’s original paragraph sentences verbatim rather than expanding the item pool to optimize internal consistency. Recognizing this limitation, Simpson, Rholes, and Phillips (1996) expanded the instrument to a 17-item revision (Simpson’s revised AAQ), which yielded substantially enhanced internal consistency metrics, typically α ≥ .81 for Attachment Avoidance and α ≥ .78 for Attachment Anxiety across collegiate and community cohorts.

Test-Retest Stability

Longitudinal evaluations of Simpson’s scale indicate moderate to high stability over time, consistent with the theoretical premise that adult attachment schemas are stable trait-like dispositions that remain sensitive to major life transitions. Over periods ranging from 2 to 6 months, test-retest correlations consistently range between r = .68 and r = .78 for both Avoidant and Anxious dimensions. Over multi-year intervals (e.g., 2 to 5 years), stability coefficients remain statistically significant (r ≈ .50 to .60), with changes systematically correlated with major life events such as severe relational trauma, infidelity, or marriage to an exceptionally secure partner.

Factor Analysis

The factor structure of the 13-item measure has been the subject of substantive psychometric inquiry and constructive critique:

Exploratory Factor Analysis (EFA)

In the original 1990 study, Simpson elected not to conduct an exploratory factor analysis prior to establishing subscales, relying instead on the conceptual authority of Hazan and Shaver’s three prototypical paragraphs to group items a priori into Secure (5 items), Avoidant (4 items), and Anxious/Ambivalent (4 items) groupings. However, when Charles Carver (1997) and other psychometricians subjected the 13 items to exploratory factor analysis using principal components analysis with varimax and oblimin rotations, the tripartite model failed to replicate cleanly.

Instead, factor analytic investigations routinely yielded a dominant two-factor solution accounting for over 50% of the total variance:

• Factor 1 (Avoidance): Capturing items reflecting discomfort with closeness, reluctance to depend on others, and difficulty with trust (Items 2, 5, 6, 7, 8, 9). Item factor loadings on this dimension consistently range from .62 to .81.
• Factor 2 (Anxiety): Capturing items indexing fear of abandonment, obsession with partner love, and overwhelming desires for total enmeshment (Items 10, 11, 12, 13). Item factor loadings range from .58 to .76.
• Behavior of Secure Items: Items originally designated as “Secure” (e.g., Items 1, 2, 3, 4, 5) did not load onto a distinct, third independent factor. Rather, they demonstrated negative cross-loadings onto the Avoidance or Anxiety factors (for example, Item 1 and Item 5 loaded inversely on Avoidance, while Item 4 loaded inversely on Anxiety).

Confirmatory Factor Analysis (CFA)

Subsequent confirmatory factor analyses comparing a 3-factor model against a 2-factor orthogonal/oblique model demonstrated superior fit indices for the two-dimensional model (RMSEA < .06, CFI > .92, TLI > .90) after dropping or reassigning complex cross-loading items. These factor analytic outcomes contributed directly to the broad consensus in contemporary attachment research (formalized by Brennan, Clark, & Shaver, 1998) that adult romantic attachment is fundamentally two-dimensional (Avoidance and Anxiety) rather than three-dimensional.

Instrument / Measurement Tool

• Formal Tool Name: Attachment Style Measure (Adult Attachment Questionnaire [AAQ])
• Primary Author: Jeffry A. Simpson (1990)
• Construct Assessed: Adult romantic attachment orientations across tripartite styles (Secure, Avoidant, Anxious/Ambivalent) and bivariate continuous dimensions (Attachment Avoidance and Attachment Anxiety)
• Test Type: Self-report psychological scale
• Item Count: 13 items
• Response Scale: 7-point Likert-type scale ranging from 1 (Strongly Disagree) to 7 (Strongly Agree)
• Target Population: Adults (ages 18+) currently in or with past experience in romantic relationships
• Administration Time: Approximately 3 to 5 minutes
• Subscale Item Breakdown:
  • Secure Subscale (5 items): Items 1, 2, 3, 4, 5
  • Avoidant Subscale (4 items): Items 6, 7, 8, 9
  • Anxious/Ambivalent Subscale (4 items): Items 10, 11, 12, 13
• Reverse-Scored Items: Items 2, 5, and 12 are worded in a negative or reverse direction relative to their assigned dimension and must be reversed prior to scoring (e.g., 1 → 7, 2 → 6, 3 → 5, 4 → 4, 5 → 3, 6 → 2, 7 → 1).
• Scoring and Indexing Options:
  • Three-Subscale Continuous Scoring: Calculate the mean score of the items for each subscale (Secure = [Item 1 + (8 – Item 2) + Item 3 + Item 4 + (8 – Item 5)] / 5; Avoidant = [Item 6 + Item 7 + Item 8 + Item 9] / 4; Anxious = [Item 10 + Item 11 + (8 – Item 12) + Item 13] / 4).
  • Two-Dimensional Scoring (Avoidance & Anxiety): Compute composite scores where Avoidance combines items reflecting discomfort with closeness and trust (Items 2, 5, 6, 7, 8, 9), and Anxiety combines items reflecting fears of abandonment and unrequited love (Items 10, 11, 12 reversed, 13).
  • Categorical Assignment: Respondents can be classified into one of three discrete categories (Secure, Avoidant, Anxious/Ambivalent) based on their highest standardized subscale score, facilitating direct comparison with Hazan and Shaver’s early typology.

Permissions & Fee and Test Year

The Attachment Style Measure was first published in 1990 in the Journal of Personality and Social Psychology by the American Psychological Association (APA). Under standard academic conventions, the 13 items are in the public domain for non-commercial educational, scientific, and empirical research purposes without royalty or licensing fees. Researchers and clinicians utilizing the scale should provide appropriate bibliographic attribution to Simpson (1990). Commercial exploitation, incorporation into proprietary digital platforms, or distribution within fee-for-service commercial assessment suites requires formal permission from the copyright holder and the author.

References

• Ainsworth, M. D. S., Blehar, M. C., Waters, E., & Wall, S. (1978). Patterns of attachment: A psychological study of the Strange Situation. Lawrence Erlbaum Associates.
• 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
• Bowlby, J. (1982). Attachment and loss: Vol. 1. Attachment (2nd ed.). Basic Books. (Original work published 1969)
• 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.
• Carver, C. S. (1997). Adult attachment and personality: Converging evidence and a new measure. Personality and Social Psychology Bulletin, 23(8), 865–883. https://doi.org/10.1177/0146167297238007
• 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
• Mikulincer, M., & Shaver, P. R. (2007). Attachment in adulthood: Structure, dynamics, and change. Guilford Press.
• Simpson, J. A. (1990). Influence of attachment styles on romantic relationships. Journal of Personality and Social Psychology, 59(5), 971–980. https://doi.org/10.1037/0022-3514.59.5.971
• Simpson, J. A., Rholes, W. S., Campbell, L., & Wilson, C. L. (2003). Adult attachment, the transition to parenthood, and depressive symptoms. Journal of Personality and Social Psychology, 84(6), 1172–1187. https://doi.org/10.1037/0022-3514.84.6.1172
• Simpson, J. A., Rholes, W. S., & Nelligan, J. S. (1992). Support seeking and support giving within couples in an anxiety-provoking situation: The role of attachment styles. Journal of Personality and Social Psychology, 62(3), 434–446. https://doi.org/10.1037/0022-3514.62.3.434
• Simpson, J. A., Rholes, W. S., & Phillips, D. (1996). Conflict in close relationships: An attachment perspective. Journal of Personality and Social Psychology, 71(5), 899–914. https://doi.org/10.1037/0022-3514.71.5.899

Items of the Scale