Review article
Adult attachment measures: A 25-year review

https://doi.org/10.1016/j.jpsychores.2009.08.006Get rights and content

Abstract

Objective

Over the past 25 years, attachment research has extended beyond infant–parent bonds to examine dyadic relationships in children, adolescents, and adults. Attachment has been shown to influence a wide array of biopsychosocial phenomena, including social functioning, coping, stress response, psychological well-being, health behavior, and morbidity, and has thus emerged as an important focus of psychosomatic research. This article reviews the measurement of adult attachment, highlighting instruments of relevance to—or with potential use in—psychosomatic research.

Methods

Following a literature search of articles that were related to the scales and measurement methods of attachment in adult populations, 29 instruments were examined with respect to their utility for psychosomatic researchers.

Results

Validity, reliability, and feasibility were tabulated on 29 instruments. Eleven of the instruments with strong psychometric properties, wide use, or use in psychosomatic research are described. These include the following: Adult Attachment Interview (George, Kaplan, and Main); Adult Attachment Projective (George and West); Adult Attachment Questionnaire (Simpson, Rholes, and Phillips); Adult Attachment Scale (and Revised Adult Attachment Scale) (Collins and Read); Attachment Style Questionnaire (Feeney); Current Relationship Interview (Crowell and Owens); Experiences in Close Relationships (Brennan, Clark, and Shaver) and Revised Experiences in Close Relationships (Fraley, Waller, and Brennan); Parental Bonding Instrument (Parker, Tupling, and Brown); Reciprocal Attachment Questionnaire (West and Sheldon-Keller); Relationship Questionnaire (Bartholomew and Horowitz); and Relationship Scales Questionnaire (Grifiin and Bartholomew).

Conclusion

In addition to reliability and validity, investigators need to consider relationship focus, attachment constructs, dimensions or categories of interest, and the time required for training, administration, and scoring. Further considerations regarding attachment measurement in the context of psychosomatic research are discussed.

Introduction

Attachment theory [1], [2], [3], [4] is a broad theory of social development that describes the origins of the patterns of close interpersonal relationships. The interaction of environmental (especially parental) and genetic factors in early development leads to individual differences in patterns of attachment behavior. Attachment behaviors are interpersonal actions that are intended to increase an individual's sense of security, particularly in times of stress or need. These interpersonal patterns are quite stable and, in adulthood, are known as adult attachment styles. It is beyond the scope of this article to describe the developmental origins of attachment security and insecurity or to explore the psychological construct in depth [1], [5], [6].

Adult attachment is becoming increasingly important in psychosomatic research because attachment influences many biopsychosocial phenomena, including social functioning, coping, stress response, psychological well-being, health behavior, and morbidity [7], [8], [9], [10], [11], [12], [13], [14]. Research that incorporates measurement of attachment provides a unique perspective because attachment constructs are theoretically and empirically distinct from other personality and social constructs such as neuroticism, global distress, self-esteem, defensiveness, dysfunctional beliefs, and support seeking [5]. The purpose of this review is to describe and evaluate methods of measuring adult attachment style.

Although there are many approaches to measuring and classifying attachment styles, all instruments differentiate patterns of secure attachment and subtypes of insecure attachment. The first instrument to measure patterns of infant–parent attachment was Ainsworth et al.'s strange situation paradigm [1]. This procedure assesses an infant's attachment–exploration balance, or the degree to which the infant uses the caregiver as a “secure” base from which to engage the environment [1]. The prototypical secure infant is distressed by separation from the caregiver, signals this distress directly upon the caregiver's return, and immediately calms with contact. The ambivalent/resistant infant also shows distress on separation, but signals for and resists contact upon the caregiver's return. The avoidant infant may or may not manifest behavioral signs of distress upon separation from the caregiver, although physiological indices suggest high reactivity [15], [16]. The infant essentially ignores the caregiver on re-union and shows little outward indication of distress. Ainsworth et al.'s recognition and coding of these patterns influenced the development of many measurement instruments for infants, children, and, more recently, adults.

Measurement of adult attachment began with the Adult Attachment Interview (AAI) [17], [18]. It was originally developed to predict the attachment pattern of infants to caregivers and was subsequently applied to numerous other research questions. The coding scheme focuses on predictive clues in the interview narrative such as narrative coherence in secure adults and idealization of caregiver in avoidant/dismissing adults. The AAI yields three categories that are similar to infant attachment categories: (a) secure/autonomous, (b) avoidant/dismissing, and (c) anxious/preoccupied (the adult version of ambivalent/resistant), and a fourth “unclassifiable” category. Where applicable, individuals can also be “unresolved” with respect to loss, trauma, or abuse.

Hazan and Shaver considered how adults with different attachment histories would classify themselves according to the ways they think, feel, and behave in close relationships. They argued that the three attachment patterns seen during infancy would emerge as three primary interpersonal styles during adolescence and adulthood. Their original approach presented adults with three patterns of attitudes towards romantic relationships and asked subjects to classify themselves based on perceived similarity to the description [19]. A large US national comorbidity survey that used this technique found prevalence rates as follows: 59%, secure; 25.2%, avoidant; 11.3%, anxious; 4.5%, unclassifiable [20]. A higher prevalence of insecure attachment is generally found in clinical populations, and findings are dependent on the instruments used. Following Hazan and Shaver, more sophisticated self-classification methods and extensive questionnaire-based scales have evolved. A great deal of research has demonstrated the utility of these self-report measures in testing and confirming fundamental predictions about attachment theory. However, there are distinctions between instruments that merit careful consideration before an assessment method is chosen.

Self-report measures probe conscious attitudes towards relationships and memories of experiences in current relationships; therefore, they cannot detect when defenses distort responses. Self-report questionnaires are also criticized for being passive (i.e., that they do not detect those attachment phenomena that need to be activated to be manifested). Self-report instruments most often focus on views that individuals currently hold about themselves and others in close relationships. This distinction is relevant to psychosomatic research, where the focus of interest is usually on the contemporary state of the individual in terms of health behavior, course of illness, and impact on future health. Some investigators use interviews about contemporary relationships to reduce response bias and to increase attachment activation while focusing on current relationships [21], [22]. Others have used projective tests in the form of line drawings depicting various attachment situations [23] to increase the activation of thoughts and feelings linked to attachment experiences.

Although all attachment measures probe emotional regulation, interpersonal awareness, and behavioral strategies in close relationships, they have been employed by two distinct professional silos: social psychologists, who have developed and used self-report measures, and psychodynamic and developmental psychologists, who have preferred tests that do not rely on conscious self-evaluation [5], [24]. Different methods of assessing attachment style emphasize different attachment phenomena [25]. The AAI and other interview methods may be used to assess narrative coherence as a marker of secure attachment [17] or to assess a person's ability to reflect on his/her inner world and the perceived intentions or subjective experiences of others [26]. Projective tests may be used to assess a person's capacity to maintain self–other boundaries and to demonstrate self-agency in resolving attachment dilemmas [23]. Self-report measures directly assess conscious attitudes towards, or awareness of behaviors in, experiences of separation, loss, intimacy, dependence, and trust [24].

Measures of attachment either assign individuals to categories of attachment style or measure the degree to which various dimensions of attachment style are present. Dimensional models of adult attachment converge on two dimensions of insecurity: attachment anxiety (negative sense of self) and attachment avoidance (negative sense of others). Attachment anxiety is characterized by an expectation of separation, abandonment, or insufficient love; a preoccupation with the availability and responsiveness of others; and hyperactivation of attachment behavior. Attachment avoidance is characterized by devaluation of the importance of close relationships, avoidance of intimacy and dependence, self-reliance, and relative deactivation of attachment behavior. If standard and acceptable cutoff points are defined, categories can be derived from dimensional scales. Categorical measures of attachment are criticized theoretically, for assuming that differences among people within a category are “unimportant or do not exist” (Mikulincer and Shaver [24], p. 85), and analytically, for their limited statistical power compared with dimensional measures [27]. However, for clinical use, the categorical approach of recognizing phenomena according to their similarity to prototypic “textbook cases” is often preferred [25]. Moreover, there is no consensus as to whether attachment phenomena are inherently categorical or dimensional. Analytically, when a categorical construct is measured using a dimensional scale, part of the observed variance is spurious.

Bartholomew and Horowitz's four-category model helpfully reconciles categorical and dimensional models by defining categories that correspond to combinations of extreme positions on the dimensions of attachment anxiety and attachment avoidance. Thus, secure attachment is conceptualized as a relative absence of attachment anxiety and attachment avoidance; preoccupied attachment is conceptualized as high attachment anxiety and low attachment avoidance; dismissing attachment is conceptualized as high attachment avoidance and low attachment anxiety; and fearful attachment is the combination of high insecurity on both dimensions of attachment avoidance and attachment anxiety.

Attachment phenomena have been described as “state-dependent traits.” This description refers to the fact that attachment behaviors are not always on display but are activated by specific events such as situations of danger, threat, or isolation. On the other hand, there is a trait-like consistency to the patterns of behavior that are triggered in such situations. Furthermore, attitudes towards relationships (e.g., expectations of others' trustworthiness or one's own lovability) guide attachment behavior and have a trait-like consistency. Attachment measures may vary in their sensitivity to the activation or inactivation of attachment phenomena; thus, some measures may be more sensitive to state-dependent changes.

While developmental attachment experiences do give rise to stable conscious attitudes and preferences in adulthood, it is also true that some observations of adult attachment style are specific for the circumstances of a particular type of relationship or for a particular dyad. Thus, measures of (a) an adult's memories of attachment to his/her parents, (b) an adult's attitudes and experiences in a current romantic relationship, (c) general attitudes towards adult romantic relationships, and (d) an adult's parenting attitudes and behaviors towards one's children are not interchangeable. Furthermore, since patterns of attachment are fundamentally oriented towards dyadic interactions, patterns of attachment may differ for the same individual in different relationships (e.g., secure with respect to mother, but insecure with respect to father; or varying from one romantic relationship to the next). Therefore, the researcher will need to decide which dyad is most important to examine for their particular research question.

The nomenclature used in the measurement of adult attachment is complicated. Starting with the AAI, adults are assigned to four categories: “free and autonomous with respect to attachment” (a.k.a. secure), “enmeshed and preoccupied with attachment” (a.k.a. anxious), “dismissing of attachment” (a.k.a. avoidant), and “cannot classify” or “unresolved with respect to trauma” (a.k.a. disorganized). The earliest categorical self-report instrument assigned people to three categories (secure, preoccupied, and avoidant), but Bartholomew and Horowitz's [21] subsequent four-category scheme has secure, preoccupied, avoidant/dismissing, and avoidant/fearful categories. Dimensions of “attachment avoidance” and “attachment anxiety” are measured by many of the instruments described in this review. It is important to appreciate, as described above in Consideration II: Categorical Versus Dimensional Measurement, how these differently named attachment categories or dimensions overlap and differ.

None of the measures of adult attachment in current use was developed for psychosomatic research. Despite good evidence to support the theoretically derived links between the quality of patient–provider relationship, health care utilization, and other medical outcomes [9], [11], [12], [14], [15], [28], [29], [30], [31], [32], it may be off-putting for patients with serious medical conditions to report on attitudes towards romantic relationships, which, on the face of it, seem to have little relevance to their most pressing concerns. Thus, the apparent relevance of questions or items in an attachment measurement to the situations of the persons who are being evaluated will be a consideration in choosing an attachment measure.

Section snippets

Literature review

We utilized two complementary search strategies. In the “bottom-up” strategy, we identified all publications on attachment and reviewed method sections to compile a list of common attachment instruments. In the “top-down” strategy, we used published reviews of attachment measures to identify seminal psychometric reports and then searched for articles that cited these reports.

Results

The characteristics of 29 adult attachment instruments are summarized in Table 1. The original papers which describe the instruments are cited in the table in addition to subsequent studies which provide further psychometric support. We describe and review 11 instruments based on: (a) their strong psychometric properties, (b) wide use, or (c) potential utility for psychosomatic research. Usage was determined with citation searches. Additionally, we reviewed psychosomatic studies in which adult

Discussion

The importance and relevance of attachment to clinical populations in psychosomatic medicine is an exciting research frontier. A partial survey of findings to date demonstrates its relevance to stress response [9], [14], [15]; relationship with care providers in diabetic patients [10], [70]; pain [12], [32], [94]; chronic diseases, including ulcerative colitis [9]; alopecia, leg ulcers, and breast cancer [13], [95]; somatization [11], [96]; hypochondriacal concerns [31]; and health care

References (161)

  • BuchheimA et al.

    Neural correlates of attachment trauma in borderline personality disorder: a functional magnetic resonance imaging study

    Psychiatry Res Neuroimaging

    (2008)
  • MeredithPJ et al.

    The relationship of adult attachment to emotion, catastrophizing, control, threshold and tolerance, in experimentally-induced pain

    Pain

    (2006)
  • StuartS et al.

    Attachment and interpersonal communication in somatization

    Psychosomatics

    (1999)
  • AinsworthM et al.

    Patterns of attachment: a psychological study of the strange situation

    (1978)
  • BowlbyJ

    Attachment and loss: vol. 1. Attachment

    (1969)
  • BowlbyJ

    A secure base

    (1988)
  • MainM

    Security in infancy, childhood, and adulthood: a move to the level of representation

    Monogr Soc Res Child Dev

    (1985)
  • MikulincerM et al.

    Attachment in adulthood: structure, dynamics, and change

    (2007)
  • MaunderRG et al.

    Attachment and psychosomatic medicine: developmental contributions to stress and disease

    Psychosom Med

    (2001)
  • MaunderRG et al.

    Attachment insecurity moderates the relationship between disease activity and depressive symptoms in ulcerative colitis

    Inflamm Bowel Dis

    (2005)
  • WallerE et al.

    Attachment representation and illness behaviour in somatoform disorders

    J Nerv Ment Dis

    (2004)
  • GunnarMR et al.

    Stress reactivity and attachment security

    Dev Psychobiol

    (1996)
  • George C, Kaplan N, Main M. Adult Attachment Interview. Unpublished manuscript,...
  • HesseE

    The Adult Attachment Interview: protocol, method of analysis, and empirical studies

  • HazanC et al.

    Romantic love conceptualized as an attachment process

    J Pers Soc Psychol

    (1987)
  • MickelsonKD et al.

    Adult attachment in a nationally representative sample

    J Pers Soc Psychol

    (1997)
  • BartholomewK et al.

    Attachment styles among young adults: a test of a four-category model

    J Pers Soc Psychol

    (1991)
  • BifulcoA et al.

    The Vulnerable Attachment Style Questionnaire (VASQ): an interview-based measure of attachment styles that predict depressive disorder

    Psychol Med

    (2003)
  • GeorgeC et al.

    The development and validity of a new measure of adult attachment: the Adult Attachment Projective

    Attach Hum Dev

    (2001)
  • MikulincerM et al.

    Measurement of attachment-related constructs in adulthood

  • Fonagy P, Steele M, Steele H, Target M. Reflective-function manual: version 5.0. For application to the Adult...
  • FraleyRC et al.

    Adult romantic attachment: theoretical developments, emerging controversies, and unanswered questions

    Rev Gen Psychol

    (2000)
  • DozierM et al.

    Clinicians as caregivers: role of attachment organization in treatment

    J Consult Clin Psychol

    (1994)
  • DozierM et al.

    Attachment and psychopathology in adulthood

  • SchmidtS et al.

    Subjective physical complaints and hypochondriacal features from an attachment theoretical perspective

    Psychol Psychother Theory Res Pract

    (2002)
  • Bakermans-KranenburgMJ et al.

    A psychometric study of the Adult Attachment Interview: reliability and discriminant validity

    Dev Psychol

    (1993)
  • RoismanGI et al.

    The Adult Attachment Interview and self-reports of attachment style: an empirical rapprochement

    J Pers Soc Psychol

    (2007)
  • van IjzendoornMH et al.

    The distribution of adult attachment representations in clinical groups: a meta-analytic search for patterns of attachment in 105 AAI studies

  • BenoitD et al.

    Stability and transmission of attachment across three generations

    Child Dev

    (1994)
  • SagiA et al.

    Stability and discriminant validity of the Adult Attachment Interview: a psychometric study in young Israeli adults

    Dev Psychol

    (1994)
  • FonagyP et al.

    The relation of attachment status, psychiatric classification, and response to psychotherapy

    J Consult Clin Psychol

    (1996)
  • PatrickM et al.

    Personality disorder and the mental representation of early social experience

    Dev Psychopathol

    (1994)
  • Holtzworth-MunroeA et al.

    Violent versus nonviolent husbands: differences in attachment patterns, dependency, and jealousy

    J Fam Psychol

    (1997)
  • Kobak RR. The Adult Attachment Interview Q-Sort. Unpublished manuscript,...
  • Fyffe CE, Waters E. Empirical classification of adult attachment status: predicting group membership. Unpublished...
  • FonagyP et al.

    The capacity for understanding mental states: the reflective self in parent and child and its significance for security of attachment

    Infant Ment Health J

    (1991)
  • FonagyP et al.

    Maternal representations of attachment during pregnancy predict the organization of infant–mother attachment at one year of age

    Child Dev

    (1991)
  • GrossmannK et al.

    Maternal attachment representations as related to patterns of infant–mother attachment and maternal care during the first year

  • Crowell J, Owens G. Current Relationship Interview and scoring system. Unpublished manuscript,...
  • Cited by (446)

    View all citing articles on Scopus
    View full text