Journal of Psychology and Theology

2004, Vol. 32, No. 2, 92-103

Copyright 2004 by Rosemead School of Psychology

Biola University, 0091-6471/410-730

92

Can a relationship with God be describedwithin an attachment framework? The gen-eralization of Bowlbys (1969, 1973, 19 8 0 )Attachment Theory regarding infant-caregiver bondsto adulthood love relationships has generated awealth of research over the last two decades regard-ing the dynamics of personal relationships in adult-hood (see Simpson & Rholes, 1998, for an editedvolume providing an excellent overview of the field).However, one might question if Attachment Theorycan be generalized to relationships between an indi-vidual and a Deity. And yet, recent theoretical andempirical work done by Lee Kirkpatrick and othershas suggested that an attachment framework mayprovide an interesting line of inquiry in the psycholo-gy of religion literature.God as an attachment figureIn describing the attachment bond, Ainsworth(1985) delineated four criteria: Maintaining proximi-ty with the attachment figure, seeing the attachmentfigure as a secure base of explorative behavior, con-sidering the attachment figure as providing a havenof safety, and experiencing separation anxiety whenremoved from the attachment figure (leading to griefif the attachment figure is also lost). Using these cri-teria, Kirkpatrick (1999) has persuasively argued thatrelationship with God can be described as an attach-ment bond. (Although Biblical passages suggest thatGod has been viewed in the Judeo-Christian tradi-

ATTACHMENT TO GOD: THEATTACHMENT TO GOD INVENTORY,TESTS OF WORKING MODELCORRESPONDENCE, AND ANEXPLORATION OF FAITH GROUPDIFFERENCESRICHARD BECKAbilene Christian UniversityANGIE MCDONALDPalm Beach Atlantic University

Recent theoretical and empirical work by Lee Kirk-patrick and others has suggested that relationshipwith God can be fruitfully described as an attach-ment bond. However, this literature has been limitedby the lack of a sound theoretical and psychometricscale that operationalizes the attachment to Godconstruct. Toward that end, the paper presents datafrom three samples, two college and one communitysample, describing the psychometric properties of theAttachment to God Inventory (AGI) as well as pro-viding tests of the correspondence and compensationhypotheses. In general, the AGI subscales of Avoid-ance of Intimacy and Anxiety about Abandonmentdisplay good factor structure, internal consistency,and construct validity. Comparisons of the AGI withadulthood attachment measures appear to support,albeit weakly, a correspondence between workingmodels of romantic others and God.Authorship status was decided by the flip of a coin. The authorswould like to thank Jonelle Unruh for statistical assistance inStudy 1, Joe Toren for data collection assistance in Study 3, andJana Beck for editorial feedback. In addition, the authors wouldlike to thank members of our research teams who participatedduring various phases of this project: Chaise Camp, Vashti Carr,Erica Fo s t e r, Jenny Kincaid, Danelle Misch, and Angie Wi n k l e r.F i n a l l y, thanks to Dr. Todd Hall for helpful feedback concern-ing the framing of the correspondence versus compensationquestion. Correspondence concerning this article may be sentto either Angie McDonald, PhD, or Richard Beck, PhD, AC UBox 28011, Abilene, TX 79699. Email: mcdonalda@acu.edu orb e c k r @ a c u . e d u

BECK and MCDONALD 93

tion as a mother, father, and a lover, it is less clearhow an attachment model describes Deity/Personrelationships in other world religions, particularly ifthe Deity is not thought of as personal in nature.Consequently, in this series of studies we chose tofocus our attention on the Judeo-Christian tradition,Christianity in particular. We revisit this issue in oursubsequent discussions.)

Empirical research concerningattachment with God

The limited but growing empirical literature con-cerning attachment with God and the relationshipbetween attachment styles and religiosity has suggest-ed that attachment perspectives are a fruitful line ofinvestigation in the psychology of religion research.In two studies, Kirkpatrick and Shaver (1990, 1992)found relationships between attachment style andreligious variables such as religious belief, commit-ment, and involvement; God image; conversion expe-riences; and experiences of glossolalia. In addition,Kirkpatrick (1997, 1998) in two longitudinal studies,found evidence that God may serve as a compensato-ry attachment figure for individuals displaying inse-cure attachment patterns. Granqvist (1998) alsofound evidence that individuals may use God as a sub-stitute attachment figure; although Granqvist suggest-ed that this process may be more complex than previ-ously thought. Others have found relationshipsbetween adulthood attachment and spiritual maturity(TenElshof & Furrow, 2000).

Assessing attachment to God and thecompensation or correspondence hypothesis

Although the empirical research has suggestedintriguing relationships between attachment vari-ables and religious constructs, this literature hasbeen limited by the lack of a psychometrically soundinstrument to assess attachment to God. This voidhas limited researchers from addressing one of themore intriguing questions in this literature. The cor-respondence or compensation question is anattempt to determine if attachment to God basicallymirrors the persons caregiver and lover attachmentstyle (the correspondence hypothesis), or rather, ifrelationship with God helps the person compensatefor deficient caregiver bonds, where a relationshipwith God fills an attachment void (the compensationhypothesis). As noted above, some evidence suggeststhat the compensation hypothesis may be correct

(Granqvist, 1998; Kirkpatrick, 1997, 1998). Howev-er, other evidence (e.g., Brokaw & Edwards, 1994;Hall & Brokaw, 1995; Hall, Brokaw, Edwards, &Pike, 1998), building upon Object Relations theory,suggests that the correspondence hypothesis may becorrect. Specifically, this work has shown that posi-tive relationships with caregivers are associated withmore loving and nurturing God images. Conversely,it appears that negative relations with caregivers areassociated with God being experienced as moredemanding and authoritarian.

These conflicting lines of evidence suggest thatresearchers must be careful when framing the issueof correspondence versus compensation. Specifical-ly, there is a distinction between compensatorybehavior (e.g., conversion, religious practices) andhow an individual experiences God (i.e., Is God per-ceived as loving and kind, or distant and judgmen-tal?), and one cannot necessarily be inferred fromthe other. Within the attachment to God literature,this issue is even more vexing due to the lack of a psy-chometrically sound instrument assessing attach-ment to God. Consequently, comparisons betweenattachment to God, God imagery, and compensato-ry religious behavior cannot proceed until the psy-chometric issues are resolved.

The Attachment to God Inventory

Building upon attachment pattern classificationschemes for childhood bonds with caregivers(Ainsworth, Blehar, Waters, & Wall, 1978) andadulthood love relationships (Bartholomew, 1990;Hazan & Shaver, 1987, 1990), Brennan, Clark, andShaver (1998) have recently argued that two dimen-sions underlay most attachment classification mod-els: Avoidance of Intimacy and Anxiety about Aban-donment. Consequently, this model is dimensionalin nature allowing individuals to vary along the twocontinuous dimensions of Avoidance and Anxiety.Yet, should one choose to use a typological model,these dimensions can be dichotomized to generatethe classic fourfold typology of Secure, Preoccu-pied, Fearful, or Avoidant attachment(Bartholomew, 1990). The flexibility of this classifi-cation model is clear in that it can incorporate bothdimensional and typological schemes of attachmentclassification. To synthesize the wide variety ofadulthood attachment measures used byresearchers, and to operationalize the Avoidanceand Anxiety dimensions, Brennan et al. (1998) have

94 ATTACHMENT TO GOD INVENTORY

developed, through factor analytic work with exist-ing attachment scales, the Experiences in CloseRelationships (ECR) scale.

Following Brennan et al. (1998), we wanted todevelop a measure that assessed the attachmentdimensions of Avoidance of Intimacy and Anxietyabout Abandonment as they apply to relationshipwith God. Consequently, the Experiences in CloseRelationships scale became a model for our Attach-ment to God Inventory (AGI).

Our conceptualizations of the Avoidance andAnxiety dimensions as they apply to relationshipwith God were straightforward and paralleled Bren-nan et al.s (1998) descriptions. Specifically, Avoid-ance of Intimacy with God involves themes such as aneed for self-reliance, a difficulty with dependingupon God, and unwillingness to be emotionally inti-mate with God. In contrast, Anxiety over Abandon-ment involves themes such as the fear of potentialabandonment by God, angry protest (resentment orfrustration at Gods lack of perceived affection), jeal-ousy over Gods seemingly differential intimacy withothers, anxiety over ones lovability in Gods eyes,and, finally, preoccupation with or worry concerningones relationship with God.

Of the three studies that follow, the first twooverview the scale construction and validation of theAGI. The AGI is then used to test hypotheses con-cerning correspondence or compensation in both acollege and community sample (Studies 2 and 3).Finally, in the adult community sample, faith groupdifferences concerning attachment to God areexplored (Study 3). Since relationship to God isoften fostered within diverse religious communities,we wanted to determine if the construct was stableacross religious affiliation.

STUDY 1

The focus of Study 1 was the construction of theAttachment to God Inventory (AGI). As mentionedabove, we attempted to model the AGI after theExperiences in Close Relationships (ECR; Brennanet al., 1998) scale which operationalizes the attach-ment dimensions of Avoidance and Anxiety in adult-hood love relationships. Specifically, we sought tocreate a self-report measure that would assess Avoid-ance of Intimacy with God and Anxiety about Aban-donment by God. From a psychometric stance, weaimed to create a two-dimensional scale (corre-sponding to the Avoidance and Anxiety dimensions)

that displayed simple factor structure, internal con-sistency, and minimal shared variance between sub-scales. From a theoretical standpoint, we wanted toachieve a balance of items within each subscale toadequately sample the various themes involved in thehigher-order dimensions of Avoidance and Anxiety.

METHOD

Item development

Items were initially developed to assess a varietyof sub-domains subsumed under the Avoidance andAnxiety dimensions. As noted above, for the Avoid-ance dimension, we generated items assessingthemes of difficulty depending upon God, unwilling-ness with expressing intimacy with God, and needfor self-reliance. For the Anxiety dimension, itemswere generated for themes concerning angr yprotest, preoccupation with the relationship, fears ofabandonment by God, anxiety over ones lovability,and jealousy. Using these themes, an initial item poolof 70 items was generated. When appropriate, someitems closely followed the wording of items from theExperiences in Close Relationships scale. Each itemwas rated along a 7-point Likert scale (1 = Disagreestrongly, 4 = Neutral/Mixed, 7 = Agree strongly).

Participants and procedure

Participants were 507 undergraduate and graduatestudents from Abilene Christian University. AbileneChristian University is a small (approximately 4500-5000 students), private, Christian institution. Sixty-two percent of the sample was female. The mean ageof the sample was 20.13 (SD = 2.89). Eighty-five per-cent of the sample was Caucasian, 6.3% African-American, and 3.9% Hispanic. Sixty-seven percent ofthe sample was affiliated with the Churches of Christ,11.0% Baptist, 6.5% Non-Denominational, 3.4%Catholic, and 2.2% as Methodist. Participants wereasked to complete the 70-item measure. Course creditwas offered for participation.

Results and Discussion

As mentioned previously, the final item selectionwas based on achieving the best balance betweenthree psychometric properties simultaneously: factorstructure (dominant factor loadings .40, cross fac-tor loadings .25), internal consistency (alpha >.80), and minimal shared variance between subscales(r2 < .10). We also made theoretical decisions to

BECK and MCDONALD 95

ensure balanced content within each subscale. Basedon these criteria, 28 items (14 Avoidance items, 14Anxiety items) were ultimately selected to comprisethe Attachment to God Inventory (see Appendix forthe entire scale and scoring instructions). The finalprincipal components analyses with Varimax rota-tion for these 28 items indicated that, indeed, twofactors best fit the data. Factor 1 accounted for23.2% of the variance and was labeled Avoidance.Factor 2 accounted for 13.9% of the variance andwas labeled Anxiety. Table 1 presents the loadingsof each of the items on the two higher-order factors.

Oblique rotations also were investigated but did notimprove factor structure. The subscales exhibitedgood internal consistency, with an alpha coefficientof .86 for the Avoidance items and .84 for the Anxi-ety items. Finally, after summing the subscale totals,Avoidance and Anxiety were found to share only6.1% of their variance (r = .248).

STUDY 2

There were two goals for Study 2. First, it wasimportant to ascertain whether the factor structureand internal consistency estimates of the AGI sub-

Table 1Principal components with Varimax rotation factor loadings for AGI items

AGI Item Factor 1: Factor 2:Avoidance Anxiety

My experiences with God are very intimate and emotional. (R) .74 .05I prefer not to depend too much on God. .68 .15My prayers to God are very emotional. (R) .66 -.16I am totally dependent upon God for everything in my life. (R) .65 .19Without God I couldnt function at all. (R) .64 -.01I just dont feel a deep need to be close to God. .64 .08Daily I discuss all of my problems and concerns with God. (R) .61 .25I am uncomfortable allowing God to control every aspect of my life. .60 .21I let God make most of the decisions in my life. (R) .59 .22I am uncomfortable with emotional displays of affection to God. .54 .15It is uncommon for me to cry when sharing with God. .53 -.23I am uncomfortable being emotional in my communication with God. .50 .05I believe people should not depend on God for things they should do for themselves. .50 .07My prayers to God are often matter-of-fact and not very personal. .47 .24I worry a lot about my relationship with God. -.05 .65I often worry about whether God is pleased with me. -.04 .65I get upset when I feel God helps others but forgets about me. .24 .63I fear God does not accept me when I do wrong. .11 .62I often feel angry with God for not responding to me. .17 .61I worry a lot about damaging my relationship with God. -.09 .61I am jealous at how God seems to care more for others than for me. .19 .60I am jealous when others feel Gods presence when I cannot. .16 .56I am jealous at how close some people are to God. .17 .56If I cant see God working in my life, I get upset or angry. .05 .55Sometimes I feel that God loves others more than me. .20 .55Almost daily I feel that my relationship with God goes back and forth from hot to cold. .08 .50I crave reassurance from God that God loves me. -.25 .50Even if I fail, I never question that God is pleased with me. (R) .09 .43

Factor eigenvalue: 6.51 3.88

Note: (R) = reverse scored item

96 ATTACHMENT TO GOD INVENTORY

scales could be maintained in a replication sample.The second goal was to compare the Attachment toGod Inventory to the adult romantic attachmentscale that it was closely based on, the Experiences inClose Relationships scale (Brennan et al., 1998).This comparison would provide an initial explo-ration into the compensation versus correspondencehypotheses.

METHOD

Participants and procedure

Surveys were administered to 118 students (89females and 29 males) in undergraduate and gradu-ate courses at Abilene Christian University. Studentsranged in age from 18 to 46 with mean of 20.66 (SD= 3.98). Seventy-two percent of the sample was Cau-casian, 6.9% Hispanic, 6% Asian, and 8.6%African/African American. Over half of the sample(55.1%) described themselves as Church of Christ,18.6% Baptist, 6.8% Catholic, 6.8% Non-Denomina-tional, and 1.7% Methodist. Course credit wasoffered for completion of the measures.

Measures

In addition to the 28-item Attachment to GodInventory, participants completed the 36-item Expe-riences in Close Relationships scale (Brennan et al.,1998). As previously mentioned, the ECR is an adultromantic attachment inventory that assesses theattachment dimensions of Avoidance and Anxiety.

RESULTS AND DISCUSSION

Replication of AGI psychometrics

As discussed above, a main agenda of Study 2 wasa replication of the AGIs psychometric properties inan independent sample. Toward that end, we con-ducted a principal components factor analysis withVarimax rotation for the 28 AGI items. Based uponthe scree plot for this analysis, a two-factor solutionwas deemed to be the best solution for the data. Fac-tor 1 was labeled Anxiety and accounted for 17.9%of item variance. Factor 2 was labeled Avoidanceand accounted for 15.4% of the item variance.

The AGI items performed well in the replicationsample. Specifically, 26 of the 28 AGI items loadedmost strongly with their original dimensions, Anxietyor Avoidance (Items 14 and 16, originally drafted andselected for the Avoidance dimension, had higherloadings on the Anxiety dimension). Overall, despite

the performance of two items, the factor structure ofthe AGI remained stable in the replication sample.Further, good internal consistency estimates wereobserved for both the AGI-Anxiety (alpha = .80) andAGI-Avoidance (alpha = .84) subscales.

Finally, based upon the strong two-factor solutionfor the AGI, we expected that the summed Anxietyand Avoidance subscales would share little of theirvariance. This was the case with the AGI subscalessharing only 1.4% (r = .12) of their variance. By con-trast, in this sample the ECR Anxiety and Avoidancesubscales shared 5.3% of their variance.

Comparison of AGI and ECR:Correspondence or compensation?

The additional goal of Study 2, beyond replicat-ing AGI psychometrics, was a comparison of ourattachment to God measure with the ECR, a mea-sure of adulthood attachment in love relationships.Table 2 presents the zero-order correlationsbetween the ECR and AGI subscales. As can beseen in Table 2, the pattern of correlations is moreambiguous than what either the correspondenceor compensation hypotheses would have anticipat-ed. Specifically, although AGI-Avoidance was unre-lated with ECR-Anxiety its relationship with ECR-Avoidance, a lthough posit ive, was alsonon-significant. However, there was some evi-dence of correspondence between the AGI-Anxi-ety and ECR-Anxiety ratings.

These correlations are intriguing in that they pre-sent one of the first direct tests of associationbetween an attachment to God measure and anadulthood attachment measure. If any trend isapparent in this data, it is towards correspondence(particularly for the attachment dimension of Anxi-ety). However, the present results are by no meansclear. Specifically, the pattern of associations regard-ing the Anxiety dimensions may be due to a develop-mental characteristic of the sample. That is, collegestudents may display a pervasive preoccupation withtheir relationships. There is some evidence to sup-port this in that ECR-Anxiety scores (M = 61.89, SD =19.69) were significantly higher (t 117 = 6.82, p