Attachment to God Inventory

Abstract

The Attachment to God Inventory (AGI), developed by Richard Beck and Angie McDonald (2004), is a premier 28-item self-report psychometric instrument designed to operationalize and quantify individual differences in perceived relational dynamics with God through the lens of contemporary attachment theory. Grounded in the structural framework of the Experiences in Close Relationships (ECR) inventory developed by Brennan, Clark, and Shaver (1998), the AGI conceptualizes attachment to the divine along two orthogonal, continuous dimensions: Anxiety about Relationship with God (14 items) and Avoidance of Intimacy with God (14 items). Respondents rate items on a 7-point Likert scale ranging from 1 (Disagree Strongly) to 7 (Agree Strongly). The Anxiety dimension captures hyperactivating relational strategies, characterized by fear of abandonment, acute hypersensitivity to perceived divine disapproval, and emotional volatility regarding God’s presence. Conversely, the Avoidance dimension captures deactivating relational strategies, characterized by compulsive self-reliance, discomfort with emotional intimacy, and a minimization of the importance of communion with God.

Extensive psychometric investigations have established the AGI as a robust, structurally sound measure. Exploratory and confirmatory factor analyses consistently substantiate the two-factor orthogonal model across diverse theological, denominational, and cultural cohorts. Reliability evaluations demonstrate high internal consistency, with Cronbach’s alpha coefficients typically ranging between .83 and .88 for both subscales across collegiate, community, and clinical samples. Construct, convergent, and discriminant validity have been confirmed through systematic associations with adult interpersonal attachment styles, spiritual well-being, religious coping strategies, psychological distress (depression, trait anxiety), and global indices of life satisfaction. The AGI bridges the intersection of the psychology of religion and clinical science, providing clinicians, pastoral counselors, and empirical researchers with a standardized tool to evaluate religious working models, spiritual struggle, and their downstream implications for psychological well-being.

Keywords

Attachment to God Inventory, AGI, Attachment Theory, Psychology of Religion, Religious Coping, Spiritual Well-Being, Attachment Anxiety, Attachment Avoidance, Internal Working Models, Richard Beck

Authors

The Attachment to God Inventory was developed and validated by:

• Richard Beck, Ph.D. — Professor and Chair of the Department of Psychology, Abilene Christian University, Abilene, Texas, United States. Dr. Beck is an internationally recognized scholar in the psychology of religion, social psychology, and theology, with extensive empirical work investigating the intersection of cognitive-affective working models, religious deconstruction, and interpersonal dynamics.
• Angie McDonald, Ph.D. — Associate Professor of Psychology, Abilene Christian University and Dallas Baptist University, Texas, United States. Dr. McDonald specializes in clinical psychology, developmental attachment paradigms, family systems, and the empirical measurement of religious and spiritual functioning.

Correspondence regarding the original psychometric validation of the AGI was directed to the Department of Psychology, Abilene Christian University, ACU Box 28011, Abilene, TX 79699-8011.

Purpose

The primary purpose of the Attachment to God Inventory (AGI) is to systematically operationalize, assess, and differentiate individual differences in the affective, cognitive, and behavioral dimensions of an individual’s personal relationship with God. In the broader landscape of the psychology of religion, early psychometric instruments largely focused on institutional religiosity, denominational orthodoxy, religious orientation (e.g., intrinsic vs. extrinsic religiosity), or broad measures of spiritual well-being. Although valuable, these conceptualizations frequently obscured the subjective, relational quality of how believers internally experience and emotionally negotiate their perceived relationship with the divine.

Drawing on developmental and social-psychological paradigms, Lee Kirkpatrick (1992, 1999) argued that for many believers, God functions psychologically as an authentic attachment figure—serving as a safe haven in times of distress, a secure base from which to explore the world, a target of proximity seeking, and a source of separation protest. Consequently, Beck and McDonald (2004) constructed the AGI to operationalize this specific dynamic, moving away from categorical classifications (e.g., secure, anxious-ambivalent, avoidant) toward a continuous dimensional model. The scale evaluates how an individual’s internal working models of the self (worthiness of love) and internal working models of others (trustworthiness and availability of the attachment figure) manifest within their relationship with God.

In empirical research, the AGI serves as an essential instrument for testing theoretical hypotheses regarding the developmental origins of religious belief. Researchers employ the AGI to investigate whether an individual’s attachment to God mirrors their human parental attachments (the correspondence hypothesis) or functions as a compensatory mechanism to substitute for early relational neglect or trauma (the compensation hypothesis). Furthermore, the scale is widely utilized to delineate patterns of religious coping, spiritual deconstruction, theological scrupulosity, and existential distress.

Clinically, the AGI provides licensed mental health practitioners, marriage and family therapists, and pastoral counselors with a standardized metric for identifying pathological or distressing relational patterns directed toward the divine. Believers experiencing acute psychological disorders often express their pathology through religious idioms; for instance, high attachment anxiety toward God frequently correlates with scrupulosity, obsessive fears of divine retribution, and chronic spiritual guilt. High attachment avoidance, on the other hand, frequently manifests as emotional detachment, cynical alienation from faith communities, and an inability to experience comfort during acute life crises. By identifying these continuous coordinates, clinicians can design spiritually integrated cognitive-behavioral or emotion-focused interventions aimed at fostering a more secure theological and emotional working model.

Psychological Construct

The Attachment to God Inventory measures two distinct, orthogonal dimensions of adult attachment applied directly to the divine: Anxiety about Relationship with God and Avoidance of Intimacy with God. These two dimensions correspond directly to the modern two-dimensional model of adult interpersonal attachment pioneered by Brennan, Clark, and Shaver (1998), reflecting underlying internal working models of self and others.

1. Anxiety about Relationship with God (Anxiety Subscale)

Attachment anxiety in the context of the divine reflects an individual’s pervasive fear of abandonment, rejection, or emotional disapproval from God. This dimension is rooted in a negative or defective “working model of self.” Individuals scoring high on this subscale tend to experience deep-seated self-doubt regarding their spiritual worthiness, accompanied by a relentless preoccupation with whether they are in God’s favor or under divine condemnation. Behaviorally and affectively, high attachment anxiety triggers hyperactivating attachment strategies:

• Hypervigilance and Reassurance Seeking: The individual exhibits chronic hypervigilance regarding signs of divine absence or disfavor, continuously seeking subjective emotional reassurance or supernatural validation of God’s love (e.g., Item 21: “I crave reassurance from God that God loves me”).
• Emotional Volatility and Spiritual Distress: The perceived relationship is experienced as turbulent, unpredictable, and unstable (e.g., Item 15: “My emotional connection to God feels like a roller coaster ride”).
• Jealousy and Resentment: High anxiety frequently fosters spiritual jealousy when observing the seemingly untroubled spiritual lives of others, manifesting as frustration that God appears to favor other believers (e.g., Item 5: “I sometimes feel that God loves other people more than Me”).
• Anger and Protest Behavior: When prayers appear unanswered or life circumstances become adverse, anxious individuals exhibit active protest behavior, characterized by anger and feeling abandoned (e.g., Item 11: “I often feel angry with God for not being there for me when I need Him”).

2. Avoidance of Intimacy with God (Avoidance Subscale)

Attachment avoidance reflects an individual’s discomfort with emotional intimacy, psychological vulnerability, and interpersonal dependency within their relationship with God. This dimension is rooted in a defensive or distrustful “working model of others,” wherein the attachment figure is viewed as fundamentally unreliable, intrusive, or distant. Consequently, high-avoidance individuals employ deactivating attachment strategies designed to minimize distress by suppressing relational needs:

• Compulsive Self-Reliance: The individual exhibits a marked preference for psychological autonomy, resisting the concept of leaning upon God for emotional sustenance or existential direction (e.g., Item 6: “I prefer not to depend too much on God”).
• Discomfort with Emotional Vulnerability: Prayers and spiritual rituals are characterized by cognitive, intellectualized, or transactional interactions rather than deep affective disclosure. The person actively avoids emotional expression or tears during prayer (e.g., Item 12: “I am uncomfortable being emotional in my communication with God”; Item 24: “I am uncomfortable sharing my deepest thoughts and feelings with God”).
• Suppression of Intimacy Needs: High-avoidance individuals often downplay the importance of having an emotionally close bond with the divine, exhibiting indifference toward spiritual intimacy (e.g., Item 2: “I just don’t feel a deep need to be close to God”).

The Four-Quadrant Typology

By crossing these two continuous dimensions, individuals can be conceptually mapped onto four distinct attachment quadrants, mirroring Bartholomew and Horowitz’s (1991) adult attachment typology:

• Secure Attachment (Low Anxiety, Low Avoidance): Individuals view themselves as loved and worthy of care, while viewing God as dependable, loving, and emotionally available. They comfortably use God as a safe haven and secure base.
• Preoccupied Attachment (High Anxiety, Low Avoidance): Individuals yearn for deep intimacy with God but remain plagued by constant fears of abandonment, scrupulosity, and feelings of divine rejection.
• Dismissing Attachment (Low Anxiety, High Avoidance): Individuals defensively disavow the need for emotional closeness to God, emphasizing rigorous self-reliance, intellectualism, or ritualistic compliance without emotional engagement.
• Fearful-Avoidant Attachment (High Anxiety, High Avoidance): Individuals crave divine acceptance and protection but intensely fear vulnerability and rejection, resulting in paralyzing ambivalence, chronic spiritual alienation, and high psychological distress.

Theoretical Framework

The Attachment to God Inventory is conceptually anchored in John Bowlby’s attachment theory (1969, 1973, 1980) and its empirical operationalization in developmental and social psychology by Mary Ainsworth (1978), Hazan and Shaver (1987), and Brennan, Clark, and Shaver (1998).

Bowlby’s Attachment Paradigm and the Extension to the Divine

Bowlby posited that humans possess an innate, evolutionarily derived behavioral system designed to maintain proximity to primary caregivers (attachment figures) to ensure physical protection and psychological survival. Over repeated interactions with caregivers, individuals construct internalized cognitive-affective representations, termed internal working models (IWMs), of the self and social others. These working models organize expectations, guide emotional appraisals, and dictate interpersonal behavioral strategies throughout the lifespan.

Lee Kirkpatrick (1992, 1999) extended Bowlby’s evolutionary and cognitive framework into the psychology of religion, arguing that the relationship between a monotheistic believer and God functions as a genuine attachment relationship rather than a mere metaphor. Kirkpatrick demonstrated that for believers, God satisfies Ainsworth’s four behavioral criteria that define an attachment bond:

1. Proximity Maintenance: Believers actively seek closeness to God through prayer, contemplation, worship, and engagement with sacred scriptures.
2. Safe Haven: In moments of acute physical threat, grief, existential terror, or personal crisis, believers turn to God as a refuge for comfort, solace, and emotional stabilization.
3. Secure Base: A felt sense of God’s abiding presence and protection empowers believers to explore the world, take vocational or moral risks, and confront existential uncertainties with confidence.
4. Separation Protest: The perceived absence, silence, or withdrawal of God elicits intense emotional distress, grief, panic, and protest behavior, as evidenced across historical spiritual literature (e.g., the “dark night of the soul”).

Operationalizing Attachment: From Categories to Two Dimensions

Early attempts to measure attachment to God (e.g., Kirkpatrick & Shaver, 1992) utilized single-item categorical measures adapted from Hazan and Shaver’s (1987) infant-derived categories (Secure, Avoidant, Anxious-Ambivalent). However, psychometricians in personality and social psychology demonstrated that categorical typologies suffer from poor statistical power, low test-retest reliability, and artificial categorization of continuous constructs (Fraley & Waller, 1998). In a landmark psychometric synthesis, Brennan, Clark, and Shaver (1998) subjected all existing adult attachment items to large-scale factor analysis, demonstrating that adult attachment is defined by two orthogonal, continuous dimensions: Attachment Anxiety and Attachment Avoidance.

Beck and McDonald (2004) recognized that previous religious attachment scales (such as Rowatt & Kirkpatrick’s 2002 instrument) were constrained by truncated length, high skewness, and ceiling effects, as religious individuals predominantly categorized themselves as “secure.” To overcome these psychometric limitations, Beck and McDonald constructed the AGI using the ECR’s structural paradigm. They carefully drafted items that captured the full range of hyperactivating (anxious) and deactivating (avoidant) attachment strategies specifically enacted within the cognitive-affective domain of an individual’s personal communion with God.

Validity

The construct, convergent, discriminant, and predictive validity of the Attachment to God Inventory have been rigorously evaluated across numerous empirical investigations.

Construct and Convergent Validity

In their initial validation studies involving over 800 Christian university students and adults, Beck and McDonald (2004) demonstrated robust convergent validity by examining the interplay between the AGI and interpersonal attachment measures. The AGI Anxiety subscale correlated significantly and positively with adult interpersonal attachment anxiety as measured by the ECR (r = .38 to .46, p < .001). Similarly, the AGI Avoidance subscale correlated significantly and positively with interpersonal attachment avoidance (r = .32 to .41, p < .001). These moderate, statistically significant correlations provide strong empirical support for the correspondence hypothesis—confirming that internal working models formed in human social contexts systematically generalize to the divine realm, while simultaneously demonstrating that attachment to God is not merely redundant with interpersonal attachment.

Convergent validity was further corroborated through correlations with established spiritual and psychological constructs:

• Spiritual Well-Being: The AGI Avoidance subscale demonstrates substantial negative correlations with the Religious Well-Being (RWB) and Existential Well-Being (EWB) subscales of the Spiritual Well-Being Scale (SWBS; Paloutzian & Ellison, 1982), with coefficients ranging from r = -.55 to -.70. Individuals avoiding intimacy with God consistently report lower spiritual fulfillment.
• Religious Coping: Research using Pargament’s RCOPE demonstrates that AGI Anxiety is strongly associated with Negative Religious Coping (spiritual discontent, pleading for divine intervention, feeling punished by God; r = .45 to .58), whereas AGI Avoidance is negatively related to Positive Religious Coping (seeking divine guidance, benevolent religious reframing).
• Psychological Symptomatology: AGI Anxiety exhibits consistent positive associations with neuroticism, trait anxiety, depression, and scrupulosity. Conversely, secure attachment to God (low scores on both AGI dimensions) is strongly associated with emotional resilience, adaptive emotion regulation, and higher self-esteem.

Discriminant and Incremental Validity

Beck and McDonald (2004) confirmed the discriminant validity of the AGI by assessing its relationship with social desirability and general religious participation. Correlations between the AGI subscales and the Marlowe-Crowne Social Desirability Scale were negligible to weak (|r| < .15), demonstrating that the scale is resistant to impression management. Furthermore, the AGI subscales successfully discriminated between distinct theological traditions: charismatic and evangelical groups scored significantly lower on avoidance than mainline Protestants, reflecting theological emphases on personal, emotionally expressive communion with God. Crucially, multiple regression analyses have consistently demonstrated that the AGI accounts for unique, incremental variance in predicting psychological distress, loneliness, and life satisfaction above and beyond global measures of religious commitment (e.g., church attendance, frequency of prayer).

Reliability

The Attachment to God Inventory exhibits excellent psychometric reliability across diverse empirical studies, including undergraduate cohorts, non-academic community adult samples, clinical outpatient populations, and diverse ecclesiastical traditions.

Internal Consistency

During its initial development, Beck and McDonald (2004) administered the preliminary instrument to multiple independent samples to establish internal consistency:

• Sample 1 (College Undergraduates, N = 268): The 14-item Anxiety subscale yielded a Cronbach’s alpha of α = .86. The 14-item Avoidance subscale yielded a Cronbach’s alpha of α = .88.
• Sample 2 (Cross-Validation Cohort, N = 227): The internal consistency remained exceptionally stable, with α = .84 for the Anxiety subscale and α = .86 for the Avoidance subscale.
• Independent and Replicative Studies: Subsequent research across various global populations has consistently replicated these values. Studies examining adult church congregations and theological students typically report alpha coefficients between .83 and .89 for Anxiety, and between .85 and .90 for Avoidance. Item-total correlations for both subscales consistently exceed .40, with the vast majority falling between .50 and .72, indicating high item homogeneity within each subscale.

Test-Retest Stability

Test-retest reliability evaluations over intervals ranging from 4 to 10 weeks demonstrate satisfactory temporal stability, reflecting the theoretical assumption that internal working models are enduring trait-like cognitive-affective structures while retaining responsiveness to profound developmental or life-crisis transitions. Observed test-retest correlation coefficients have routinely ranged between r = .75 and .84 for both dimensions, indicating robust stability across time without psychometric drift.

Factor Analysis

The structural dimensionality of the Attachment to God Inventory was established through rigorous exploratory factor analysis (EFA) and subsequently confirmed via confirmatory factor analysis (CFA).

Exploratory Factor Analysis (EFA)

In Study 1 of the initial scale development (Beck & McDonald, 2004), an initial pool of candidate items was subjected to principal-components analysis followed by oblique and orthogonal (Varimax) rotations. Examination of the scree plot and eigenvalues decisively revealed a clear two-factor solution, mirroring the ECR’s structural architecture:

• Factor 1 (Avoidance of Intimacy with God): Accounted for the largest proportion of total variance (approximately 21.4%). Items loading on this factor explicitly addressed discomfort with emotional prayer, self-reliance, reluctance to depend on God, and lack of perceived intimacy.
• Factor 2 (Anxiety about Relationship with God): Accounted for approximately 18.2% of the variance. Items loading on this factor captured intense worry over God’s love, emotional volatility, spiritual jealousy, and fear of divine abandonment.

Factor loadings for the retained 28 items were robust, with all target items exhibiting primary factor loadings exceeding .45 (ranging from .48 to .78) and minimal cross-loadings (consistently below .25). Crucially, the correlation between the two factors was virtually zero (r = -.04 to .06), confirming that Anxiety and Avoidance operate as truly orthogonal, independent dimensions of attachment to God.

Confirmatory Factor Analysis (CFA) and Fit Indices

In subsequent validation cohorts, Beck and McDonald (2004) performed Confirmatory Factor Analysis to compare alternative structural configurations, including a one-factor unidimensional model, a two-factor oblique model, and the hypothesized two-factor orthogonal model. The two-factor orthogonal model provided a superior and highly acceptable fit to the empirical data:

• Root Mean Square Error of Approximation (RMSEA): Values typically range from .052 to .064 (90% CI [.046, .071]), falling well within the standard threshold for good model fit (< .08).
• Comparative Fit Index (CFI): Ranged from .91 to .94 across validation samples, demonstrating robust fit against null baseline models.
• Tucker-Lewis Index (TLI): Consistently exceeded .90.
• Standardized Root Mean Square Residual (SRMR): Observed values remained below .065.

Subsequent psychometric invariance testing has demonstrated structural metric and scalar invariance across biological sexes and moderate invariance across diverse Christian denominations (e.g., Protestant, Catholic, Orthodox), affirming that the underlying latent constructs are measured equivalently across demographic groups.

Instrument / Measurement Tool

• Test Type: Standardized self-report psychometric rating scale.
• Construct Measured: Individual differences in internal working models of attachment to the divine (God), operationalized across two orthogonal dimensions: Attachment Anxiety and Attachment Avoidance.
• Target Population: Adolescents and adults (ages 16+) within monotheistic or personal-theistic religious traditions.
• Administration Format: Paper-and-pencil questionnaire, online assessment platform, or integrated clinical interview battery. Self-administered.
• Administration Time: Approximately 7 to 10 minutes.
• Total Item Count: 28 items.
• Subscale Structure:
  • Anxiety about Relationship with God: 14 items (Items 1, 3, 5, 7, 9, 11, 13, 15, 17, 19, 21, 23, 25, 27).
  • Avoidance of Intimacy with God: 14 items (Items 2, 4, 6, 8, 10, 12, 14, 16, 18, 20, 22, 24, 26, 28).
• Response Scale: 7-point Likert scale:
  • 1 = Disagree Strongly
  • 2 = Disagree
  • 3 = Disagree Slightly
  • 4 = Neutral/Mixed
  • 5 = Agree Slightly
  • 6 = Agree
  • 7 = Agree Strongly
• Scoring and Transformation Rules:
  • Step 1 (Reverse Scoring): Seven items on the Avoidance subscale are positively worded indicators of intimacy and emotional security. These items must be reverse-scored prior to calculating subscale scores. Specifically, reverse items: 4, 8, 14, 18, 22, 26, and 28 using the formula: Score_Reversed = 8 - Score_Original.
  • Step 2 (Subscale Mean Calculation): Compute the arithmetic mean of the respective 14 items for each subscale:
    • Anxiety Score = Sum of items (1, 3, 5, 7, 9, 11, 13, 15, 17, 19, 21, 23, 25, 27) / 14
    • Avoidance Score = Sum of items (2, 4_R, 6, 8_R, 10, 12, 14_R, 16, 18_R, 20, 22_R, 24, 26_R, 28_R) / 14
  • Step 3 (Interpretation): Subscale scores range continuously from 1.0 to 7.0. Higher scores on the Anxiety dimension indicate heightened fear of abandonment, emotional volatility, and hyperactivating strategies. Higher scores on the Avoidance dimension reflect marked discomfort with intimacy, compulsive self-reliance, and deactivating strategies. Median splits or sample means may be used to categorize respondents into the four quadrants (Secure, Preoccupied, Dismissing, Fearful), although analyzing the continuous dimensions is strongly recommended to preserve statistical power.

Permissions & Fee and Test Year

The Attachment to God Inventory was formally published in 2004 by Richard Beck and Angie McDonald in the Journal of Psychology and Theology (Vol. 32, Iss. 2, pp. 92–103). The copyright is held by the Rosemead School of Psychology, Biola University.

The AGI is an open-access psychometric instrument for non-commercial academic research, empirical dissertation work, and educational training. Researchers and clinicians may administer and reproduce the inventory without paying licensing fees, provided that appropriate scholarly attribution and citation are accorded to the original authors (Beck & McDonald, 2004). Commercial applications, incorporation into proprietary digital mental health software, or integration into fee-for-service testing batteries require explicit written permission from the copyright holders and original authors.

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
• Beck, R., & McDonald, A. (2004). Attachment to God: The Attachment to God Inventory, tests of working model correspondence, and an exploration of faith group differences. Journal of Psychology and Theology, 32(2), 92–103. https://doi.org/10.1177/009164710403200202
• 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.
• Fraley, R. C., & Waller, N. G. (1998). Adult attachment security: Psychometrics of conceptualization and measurement. In J. A. Simpson & W. S. Rholes (Eds.), Attachment theory and close relationships (pp. 89–114). Guilford Press.
• 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
• Kirkpatrick, L. A. (1992). An attachment-theory approach to the psychology of religion. The International Journal for the Psychology of Religion, 2(1), 3–28. https://doi.org/10.1207/s15327582ijpr0201_2
• Kirkpatrick, L. A. (1999). Toward an evolutionary psychology of religion and personality. Journal of Personality, 67(6), 921–952. https://doi.org/10.1111/1467-6494.00078
• Kirkpatrick, L. A., & Shaver, P. R. (1992). An attachment-theoretical approach to romantic love and religious belief. Personality and Social Psychology Bulletin, 18(3), 266–275. https://doi.org/10.1177/0146167292183002
• Paloutzian, R. F., & Ellison, C. W. (1982). Loneliness, spiritual well-being and the quality of life. In L. A. Peplau & D. Perlman (Eds.), Loneliness: A sourcebook of current theory, research and therapy (pp. 224–237). John Wiley & Sons.
• Rowatt, W. C., & Kirkpatrick, L. A. (2002). Two dimensions of attachment to God: Additions to investment model. Journal for the Scientific Study of Religion, 41(4), 637–651. https://doi.org/10.1111/1468-5906.00143

13. Items of the Scale (Questionnaire)