Adjective Ratings of God

1. Abstract

The Adjective Ratings of God (ARGS), developed by psychometrician and psychologist of religion Richard L. Gorsuch in 1968, is an established quantitative instrument designed to assess multidimensional representations and semantic conceptualizations of the divine. Moving substantially beyond simplistic, categorical survey queries such as “Do you believe in God?”, the ARGS operationalizes the specific qualitative, affective, and theological attributes that individuals ascribe to the concept of the deity. Grounded theoretically in the synthesis of Bernard Spilka and colleagues’ anthropomorphic factor models and Charles E. Osgood‘s semantic differential framework of affective meaning, the scale draws from an initial pool of 91 descriptive adjectives evaluated via a 3-point categorical rating format (1 = “The word does not describe ‘God'”, 2 = “The word describes ‘God'”, and 3 = “The word describes ‘God’ particularly well”).

Through higher-order and exploratory factor analyses, Gorsuch established a hierarchical structure comprising eight first-order factors, two second-order factors, and one overarching third-order factor, isolating five primary psychometric subscales: Traditional Christian, Deisticness, Wrathfulness, Omni-ness, and Irrelevancy. Across initial normative samples of university undergraduates ($N = 585$) and independent cross-validation cohorts, the subscales demonstrated robust internal consistency coefficients, ranging from $\alpha = .71$ to $.94$ for the primary conceptual dimensions, with the sole exception of the Irrelevancy subscale ($\alpha = .49$), which suffered from statistical range restriction and floor effects. Subsequent psychometric evaluations have further underscored the instrument’s structural stability, criterion-related validity, and utility in differentiating intrinsic versus extrinsic religious orientations, coping styles, attributional patterns, and psychological well-being. This article provides a comprehensive academic review of the scale’s historical foundations, theoretical paradigm, empirical psychometrics, scoring protocols, and complete verbatim questionnaire items.

2. Keywords

Adjective Ratings of God, God Concept, God Image, Psychology of Religion, Richard L. Gorsuch, Semantic Differential, Psychometrics, Anthropomorphism, Traditional Christian, Wrathfulness

3. Authors

The primary author and architect of the Adjective Ratings of God Scale is Richard L. Gorsuch, Ph.D. (1937–2016). At the time of the instrument’s foundational development and publication in 1968, Dr. Gorsuch was affiliated with the Department of Psychology at Vanderbilt University in Nashville, Tennessee. He later served as a Senior Professor of Psychology in the Graduate School of Psychology at Fuller Theological Seminary in Pasadena, California, where he directed pioneering research at the intersection of psychometrics, personality theory, multivariate statistics, and the empirical psychology of religion.

Dr. Gorsuch made monumental contributions to multivariate methodology, widely recognized for his authoritative text Factor Analysis (1974, 1983), and co-authored the widely used State-Trait Anxiety Inventory (STAI) alongside Charles Spielberger and Robert Lushene. His conceptual and psychometric insights regarding the ARGS integrated earlier developmental and lexical research conducted by Bernard Spilka, Philip Armatas, and J. Nussbaum (1964) at the University of Denver, as well as the psycholinguistic measurement paradigms pioneered by Charles E. Osgood, George J. Suci, and Percy H. Tannenbaum (1957) at the University of Illinois.

4. Purpose

The foundational rationale underlying the formulation of the Adjective Ratings of God Scale centers on addressing a profound methodological inadequacy in social science: the historical reliance on single-item, dichotomous inquiries regarding religious affiliation or the existence of a deity. In classical sociological and epidemiological surveys, researchers routinely relied on queries such as “Do you believe in God?” with simple binary (yes/no) response parameters. Gorsuch (1968) argued that such rudimentary measurements conflate nominal metaphysical assent with the deeply nuanced, phenomenological, and cognitive-affective internal models that govern religious life. Two individuals may provide identical affirmative responses to the question of theistic belief, yet one may conceptualize the divine as a compassionate, intimate, and benevolent protector, whereas the other views God as an austere, distant, cosmic judge or an impersonal philosophical construct.

The primary purpose of the ARGS is therefore to delineate, quantify, and differentiate the specific qualitative dimensions and structural components of an individual’s operational “God concept.” From a theoretical standpoint, distinguishing between an individual’s explicit theological doctrine (often termed the intellectualized “God concept”) and their experiential, emotionally laden schema (frequently operationalized as the “God image”) requires sensitive, multidimensional lexical assessment. The ARGS captures how semantic descriptors are organized within an individual’s cognitive architecture, mapping the psychological reality of God as an internalized object or systemic schema.

In clinical, counseling, and empirical research contexts, the ARGS serves several crucial applications:

• Clinical and Pastoral Assessment: The scale enables mental health clinicians, pastoral counselors, and clinical psychologists to evaluate how punitive versus benevolent divine representations correlate with clinical distress, moral injury, internalizing disorders, generalized anxiety, and clinical depression. Attributing high endorsement to wrathful, controlling, or terrifying divine characteristics often indicates pathological guilt, spiritual struggle, or scrupulosity.
• Coping and Health Psychology: In behavioral medicine and health psychology, divine attributes mapped by the ARGS serve as powerful predictors of religious coping strategies. Conceptions of God as supportive, personal, and loving consistently correlate with collaborative coping, resilience, lower physiological stress reactivity, and better recovery trajectories during chronic illness. Conversely, perceptions characterized by deistic distance or severe wrathfulness predict passive-deferring coping or spiritual alienation.
• Empirical Psychology of Religion: Researchers utilize the ARGS to examine how divine representations interact with core personality traits (such as the Five-Factor Model dimensions of Agreeableness, Neuroticism, and Openness), attachment styles (secure versus insecure attachment to God), developmental stages, and structural belief systems such as intrinsic versus extrinsic religious orientations.

5. Psychological Construct

The psychological construct assessed by the Adjective Ratings of God Scale is the multidimensional cognitive representation of God. This construct encompasses both the culturally transmitted propositional beliefs that individuals endorse regarding divine agency and the deeply internalized affective schemas through which the divine is understood. Factor-analytic extraction demonstrates that this semantic space cannot be reduced to a single evaluative continuum; rather, it occupies a complex multidimensional space reflecting specific theological, relational, and emotional dimensions.

Traditional Christian

The Traditional Christian dimension represents an orthodox, classically theistic, and predominantly Western Judeo-Christian conceptualization of God. This factor integrates divine benevolence, personal relationality, sovereignty, and metaphysical holiness. Adjectives loading heavily on this dimension—such as Loving, Kind, Forgiving, Fatherly, Holy, Redeeming, Just, and Merciful—capture an active, relational deity who intervenes in human history, exhibits deep ethical concern for humanity, and embodies unconditional agape. In Gorsuch’s hierarchical architecture, this factor often functions as a broad, higher-order evaluative dimension incorporating facets of supportive care, fatherly benevolence, and divine majesty.

Deisticness

The Deisticness dimension captures an impersonal, removed, and hands-off conceptualization of the transcendent. Rooted in Enlightenment-era Deism, this schema views the divine as a cosmic first cause, prime mover, or distant architect who established natural and moral laws but remains uninvolved in individual human affairs, daily prayer, or personal communion. Core marker items loading onto this construct include adjectives such as Distant, Formal, Inaccessible, Permissive, and Passive. Individuals scoring high on Deisticness typically reject notions of personal intervention, emotional intimacy with God, or anthropomorphic responsiveness, viewing God instead through an abstract, intellectualized lens.

Wrathfulness

The Wrathfulness dimension measures the perception of the deity as punitive, harsh, demanding, authoritarian, and severely judgmental. Rather than emphasizing benevolence or comfort, this construct reflects fearful expectations of divine retribution, moral condemnation, and strict discipline. Adjectives defining this subscale include Cruel, Damning, Avenging, Dangerous, Hard, Punishing, Stern, Severe, and Wrathful. Psychologically, high Wrathfulness scores frequently correlate with legalistic religious upbringing, anxious or avoidant internal working models, heightened state-trait anxiety, authoritarian social attitudes, and maladaptive perfectionism.

Omni-ness

The Omni-ness subscale assesses the traditional classical theistic metaphysical attributes of absolute transcendence and boundless capability. Distinct from purely emotional or relational qualities, this construct encapsulates the boundless philosophical limits of divinity: Omnipotent (all-powerful), Omniscient (all-knowing), Omnipresent (present everywhere), Infinite, and All-wise. This factor captures cognitive adherence to standard theological doctrines regarding cosmic supremacy, metaphysical totality, and epistemic perfection, often functioning independently of whether an individual feels personal warmth toward the divine.

Irrelevancy

The Irrelevancy dimension evaluates the perceived insignificance, weakness, or complete unimportance of the divine concept in an individual’s phenomenological world. Defined by items such as Feeble, Weak, Worthless, and Mythical, this factor taps into non-theistic, atheistic, or thoroughly secular orientations where the idea of God is seen as an obsolete cultural artifact or an impotent illusion. Notably, in normative religious samples, this subscale often exhibits extreme positive skewness, floor effects, and attenuated variance, as practicing believers categorically reject these descriptors.

6. Theoretical Framework

The conceptual foundation of the Adjective Ratings of God Scale rests upon the integration of two distinct theoretical paradigms: the anthropomorphic-theological factor approach developed by Bernard Spilka and colleagues (1964), and the psycholinguistic Semantic Differential framework formulated by Charles E. Osgood, George Suci, and Percy Tannenbaum (1957).

Anthropomorphic Factor Model (Spilka, Armatas, & Nussbaum, 1964)

Spilka and his collaborators argued that human beings inherently conceptualize divine, intangible entities through projected human qualities and relational schemas. Drawing from psychodynamic object-relations theory, social cognition, and developmental psychology, Spilka asserted that individuals construct mental models of God using the same cognitive categories they use to perceive human interpersonal figures, particularly parental archetypes. By administering a 63-adjective checklist to religious participants, Spilka et al. derived five primary anthropomorphic factors:

1. Kindly Father: Benevolent, loving, protective parental attributes.
2. Stern Father: Punitive, controlling, demanding patriarchal attributes.
3. Impersonal: Remote, philosophical, abstract qualities.
4. Omni-ness: Cosmic metaphysical perfections.
5. Supreme Ruler: Sovereign, kingly, majestic authority.

This theoretical stance posits that religious schemas are emotionally grounded, functional social-cognitive representations that reflect an individual’s personal history, familial dynamics, and internalized relational expectations.

The Semantic Differential Framework (Osgood, Suci, & Tannenbaum, 1957)

Concurrently, Charles E. Osgood’s psycholinguistic theory postulated that human affective and connotative meaning across cultures can be organized within a three-dimensional semantic space: Evaluation (Good vs. Bad), Potency (Strong vs. Weak), and Activity (Active vs. Passive) — often abbreviated as the EPA framework. Working under this rubric, semantic meaning is captured through continuous ratings across pairs of antonyms or descriptive adjectives. Osgood’s theory implies that any concept, including the divine, is mediated by universal neuro-affective evaluation mechanisms.

Gorsuch’s Integrative Synthesis

Richard L. Gorsuch (1968) recognized a vital theoretical tension between these two models. If Osgood’s semantic differential framework fully explained religious verbal behavior, ratings of God would collapse cleanly into generic EPA dimensions, rendering domain-specific religious theories redundant. Conversely, if religious language represents a unique psychological domain, empirical ratings would yield substantive, doctrinally and relationally specific factors (e.g., Omni-ness, Wrathfulness, Grace) that resist reduction to simple Evaluation or Potency dimensions.

To rigorously test these competing paradigms, Gorsuch synthesized the 63 items from Spilka et al. (1964) with 28 representative items from Osgood’s EPA semantic differential lexicon, generating a comprehensive 91-item measurement battery. His empirical findings revealed that while broad evaluative and potency aspects operate within religious language, human perceptions of God fundamentally demand domain-specific theological and relational constructs. Thus, the ARGS validated a hierarchical cognitive architecture: broad affective evaluations anchor the higher-order factors, while culturally and psychologically distinct schemas (Traditional Christian, Deisticness, Wrathfulness) govern lower-order cognitive organization.

7. Validity

Extensive empirical investigation across several decades has established the construct, convergent, discriminant, and criterion-related validity of the Adjective Ratings of God Scale.

Construct Validity

Gorsuch (1968) demonstrated construct validity by confirming that the scale’s empirical factor structure reflects distinct theoretical conceptualizations rather than mere semantic artifacts. Although semantic differential dimensions (Evaluation, Potency, Activity) were present, they merged into specific religious constructs. Construct validity was corroborated by showing that the Traditional Christian dimension accounted for the largest proportion of common variance among church-attending samples, reflecting prevailing Judeo-Christian religious social learning.

Convergent and Discriminant Validity

The ARGS exhibits strong convergent validity when benchmarked against standardized measures of religious commitment, motivation, and coping:

• Intrinsic vs. Extrinsic Religious Orientation: Research utilizing Gordon Allport‘s Religious Orientation Scale has demonstrated that intrinsic religious orientation correlates positively and robustly with the Traditional Christian ($r \approx .55$ to $.70$) and Omni-ness subscales, while correlating negatively with Irrelevancy and showing nonsignificant or inverse associations with Deisticness (Schaefer & Gorsuch, 1992). Conversely, extrinsic-social religiosity frequently associates with external, authoritarian, or deistic representations.
• Prayer Behavior and Religious Practice: Fairchild et al. (1993) documented strong convergent concordance between ARGS profiles and prayer modalities. Participants with elevated Traditional Christian and Kindliness scores engaged significantly more in contemplative, petitionary, and conversational prayer. In contrast, those scoring high on Deisticness reported negligible personal prayer, aligning with the theoretical expectation that a distant god is not engaged as an interpersonal dialogue partner.
• Perceptions of Divine Control and Locus of Control: Sundin, Ladd, and Spilka (1995) investigated the relationship between ARGS dimensions and perceived divine agency. The Traditional Christian and Omni-ness scales correlated positively with collaborative and active theistic control beliefs, whereas the Wrathfulness scale correlated with fatalistic religious attributions and external locus of control.

Criterion-Related and Clinical Validity

The scale possesses criterion-related validity in differentiating clinical and psychological adjustment indices. High scores on the Wrathfulness subscale are linked with elevated shame, scrupulosity, guilt-proneness, and symptoms of clinical depression and generalized anxiety. Conversely, individuals who view God primarily through the Traditional Christian lens exhibit higher subjective well-being, life satisfaction, and purpose in life, providing clear predictive validity across health psychology contexts.

8. Reliability

The internal consistency and structural stability of the Adjective Ratings of God subscales have been evaluated across multiple independent investigations using Cronbach’s alpha ($lpha$) and split-half reliability coefficients.

Original Psychometric Findings (Gorsuch, 1968)

In the original derivation study comprising 585 undergraduate students, Gorsuch reported the following internal consistency reliabilities across the five extracted subscales:

• Traditional Christian (50 items): $\alpha = .94$
• Omni-ness (5 items): $\alpha = .89$
• Wrathfulness (13 items): $\alpha = .83$
• Deisticness (5 items): $\alpha = .71$
• Irrelevancy (4 items): $\alpha = .49$

The exceptionally high reliability of the Traditional Christian subscale reflects both its extensive item pool and the strong coherence of orthodox theistic descriptors among college respondents. The attenuated reliability observed for the Irrelevancy scale ($\alpha = .49$) was explicitly identified by Gorsuch as a statistical artifact resulting from severe restriction of variance; religious respondents overwhelmingly assigned scores of “1” (“does not describe”) to terms like Feeble or Worthless, generating floor effects that truncated covariance.

Cross-Validation and Subsequent Replications

Gorsuch evaluated an independent cross-validation sample of 85 randomly selected males set aside from the primary dataset, confirming that the factor structure and item-scale correlations remained invariant. Subsequent independent empirical studies confirmed strong reliability metrics across diverse cohorts:

• Fairchild et al. (1993) ($N = 116$):
  • Traditional Christian: $\alpha = .83$
  • Wrathfulness: $\alpha = .89$
  • Deisticness: $\alpha = .65$
  • Omni-ness: $\alpha = .69$
  • Kindliness (heuristic primary factor): $\alpha = .94$
• Sundin, Ladd, and Spilka (1995) ($N = 149$):
  • Traditional Christian: $\alpha = .94$
  • Wrathfulness: $\alpha = .79$
  • Deisticness: $\alpha = .71$
  • Omni-ness: $\alpha = .86$
  • Kindliness (heuristic primary factor): $\alpha = .95$

Across investigations, the primary conceptual subscales (Traditional Christian, Wrathfulness, Omni-ness, and Deisticness) consistently meet and exceed the standard psychometric threshold of $\alpha ge .70$ for empirical research.

9. Factor Analysis

The mathematical extraction and structural refinement of the ARGS represents a classic application of hierarchical multivariate factor analysis. Gorsuch utilized exploratory factor analysis (EFA) procedures designed to disentangle overlapping variance and evaluate higher-order cognitive schemas.

First-Order Factor Solutions

The initial 91 adjectives, augmented by 9 experimental marker variables (8 random items and 1 sex indicator variable, which were excluded from the final instrument), were subjected to principal axis factoring followed by oblique rotation (promax/quartimin criteria), allowing extracted factors to correlate realistically:

• The initial extraction identified eight first-order factors. These factors captured distinct semantic clusters, including Kindliness, Wrathfulness/Sternness, Omnipresence/Omniscience, Impersonality/Distance, and Social Evaluation.
• Rather than collapsing into the simple tripartite semantic differential model (Evaluation, Potency, Activity), the observed factor structure largely replicated the anthropomorphic schema identified by Spilka et al. (1964), confirming that domain-specific religious meaning structures organize the cognitive lexicon.

Second-Order and Third-Order Hierarchical Structure

Recognizing substantial intercorrelations among first-order factors, Gorsuch performed higher-order factoring on the first-order correlation matrix:

• Second-Order Factors: Factor analysis of the first-order factors yielded two broad second-order factors. The first represented an overarching, benevolent theistic commitment dimension combining warmth, omnipotence, and conventional worship attributes. The second represented an aloof, deistic, or harsh dimension.
• Third-Order Factor: Factoring the second-order correlations produced a single third-order general factor representing an overarching general evaluative orientation toward the reality and positive worth of the divine concept versus its total secular rejection.

Subscale Construction and Item Overlap

To balance clinical utility with empirical fidelity, Gorsuch isolated items across five of the most stable, theoretically coherent factors to build the operational subscales. Notably, consistent with classical hierarchical measurement models, certain core theological attributes display dual factor loadings across closely aligned constructs. For instance, item 42 (Fair) and items 56 and 57 (Loving and Passive / theological perfection items) reflect shared variance between broad traditional Judeo-Christian theology and absolute cosmic attributes (Omni-ness), while item 76 (Strong) cross-loads between Traditional Christian majesty and Wrathful/Potent authority.

10. Instrument / Measurement Tool

The Adjective Ratings of God Scale (ARGS), originally presented in research literature as the Religious Concept Survey, is a standardized, self-report psychological inventory designed for group or individual administration.

Structural Specifications

• Instrument Type: Self-report adjective rating checklist / semantic questionnaire.
• Total Item Count: 91 substantive descriptive adjectives (evaluated across five distinct subscales).
• Administration Time: Approximately 5 to 10 minutes.
• Target Population: Adolescents and adults (typically validated with university-level reading competence).
• Response Scale: 3-point categorical rating format:
  • 1 = The word does not describe “God.”
  • 2 = The word describes “God.”
  • 3 = The word describes “God” particularly well.

Subscale Item Composition and Scoring Protocols

Subscale scores are computed by summing the numerical ratings (1, 2, or 3) assigned to the specific items designated for each dimension. Alternatively, mean item scores (ranging from 1.00 to 3.00) can be calculated to facilitate direct comparisons across subscales of varying lengths:

• Traditional Christian (50 Items): Items 1, 3, 5, 7, 8, 9, 10, 11, 19, 20, 21, 22, 23, 26, 29, 30, 32, 33, 34, 35, 37, 38, 40, 42, 44, 45, 46, 48, 50, 51, 53, 54, 56, 57, 58, 59, 62, 63, 65, 66, 68, 74, 75, 76, 78, 79, 82, 85, 86, 88. Evaluates conventional Judeo-Christian perceptions of divine benevolence, grace, majesty, and relational closeness.
• Deisticness (5 Items): Items 18, 39, 41, 55, 60. Assesses the conceptualization of God as distant, impersonal, uninvolved, and passive.
• Wrathfulness (13 Items): Items 4, 6, 12, 13, 14, 36, 43, 64, 70, 71, 76, 81, 90. Quantifies representations of God as punitive, harsh, severe, avenging, and frightening.
• Omni-ness (5 Items): Items 42, 46, 50, 56, 57 (in conjunction with the classical philosophical “omni” stems: Omnipotent [80], Omnipresent [46], Omniscient [50], Infinite [36], All-wise [9]). Measures belief in divine metaphysical perfections.
• Irrelevancy (4 Items): Items 24, 28, 87, 89. Gauges perceptions of the deity as weak, obsolete, mythical, or inconsequential.

11. Permissions & Fee and Test Year

The Adjective Ratings of God Scale was originally published in 1968 by Dr. Richard L. Gorsuch in the Journal for the Scientific Study of Religion (Vol. 7, No. 1, pp. 56–64). Under standard academic fair-use doctrines and psychometric evaluation conventions in the psychology of religion, the 91-item inventory is considered an open-access, public-domain instrument for non-commercial scholarly research and educational assessment.

No royalty fees or formal commercial licensing agreements are mandated for academic, clinical, or dissertation use. Researchers utilizing the scale are expected to provide full academic attribution by citing Gorsuch’s seminal 1968 publication. Those intending to adapt, digitize, or re-publish the instrument in commercial assessment batteries should consult the Society for the Scientific Study of Religion (SSSR) or the contemporary copyright holders of the original journal article.

12. References

Fairchild, D., Roth, H., Milmoe, S., Gotthard, C., Fehrmann, L., Richards, S., Kim, B. H., Sedlmayr, J., Carely, B., Pan, P., & Spilka, B. (1993). God images and prayer behavior: Consonance in the psychology of religion [Paper presentation]. Joint convention of the Rocky Mountain and Western Psychological Associations, Phoenix, AZ, United States.

Gorsuch, R. L. (1968). The conceptualization of God as seen in adjective ratings. Journal for the Scientific Study of Religion, 7(1), 56–64. https://doi.org/10.2307/1385110

Gorsuch, R. L. (1983). Factor analysis (2nd ed.). Lawrence Erlbaum Associates.

Osgood, C. E., Suci, G. J., & Tannenbaum, P. H. (1957). The measurement of meaning. University of Illinois Press.

Schaefer, C. A., & Gorsuch, R. L. (1992). Dimensionality of religion: Belief and motivation as predictors of behavior. Journal of Psychology and Christianity, 11(3), 244–254.

Spilka, B., Armatas, P., & Nussbaum, J. (1964). The concept of God: A factor analytic approach. Review of Religious Research, 6(1), 28–36. https://doi.org/10.2307/3510880

Sundin, H., Ladd, K. L., & Spilka, B. (1995). The relation between God images and perceptions of God’s control [Paper presentation]. Annual convention of the Rocky Mountain Psychological Association, Boulder, CO, United States.

13. Items of the Scale (Questionnaire)