Attributions of Responsibility to God Scale

Abstract

The Attributions of Responsibility to God Scale (ARGS; Gorsuch & Smith, 1983) is an experimental psychometric instrument designed to assess the cognitive and theological processes through which individuals ascribe causal agency, moral accountability, and purposive intentionality to the divine for specific real-world outcomes. Grounded in classic attribution theory—particularly the foundational models formulated by Fritz Heider, Harold Kelley, and Bernard Weiner—the ARGS embeds social-cognitive paradigms within the psychology of religion. Rather than employing decontextualized self-report statements, the instrument utilizes a factorial vignette design comprising four realistic narrative scenarios involving psychological counseling, community relocation, physical trauma recovery, and prosocial interpersonal dynamics. Across these four vignettes, two structural dimensions are systematically manipulated in a 2 × 2 factorial framework: outcome probability (low probability/unlikely occurrence vs. high probability/likely occurrence) and outcome extremity (mildly positive outcome vs. extremely positive outcome). Following each scenario, respondents rate four distinct attributional targets using an 11-point continuous rating metric: divine intervention (attribution of responsibility to God), personal responsibility, personal effort, and external luck/chance. Psychometric validation studies conducted by Gorsuch and Smith among evangelical undergraduate cohorts demonstrated substantial construct validity, showing that causal attributions to God interact significantly with preexisting theological dispositions, specifically Christian fundamentalism and perceived “nearness to God” (Broen, 1957). Although the original experimental formulation did not establish standardized classical test theory reliability indices due to its factorial manipulation structure, the scale serves as a foundational landmark in empirical theology and cognitive science of religion, illuminating the situational and dispositional boundaries that govern how human minds map supernatural agency onto mundane and critical life occurrences.

Keywords

Attributions of Responsibility to God Scale, attribution theory, causal attribution, psychology of religion, divine intervention, locus of control, religious cognition, vignette methodology, Christian fundamentalism, nearness to God, experimental social psychology

Authors

The Attributions of Responsibility to God Scale was formulated and empirically validated by:

• Richard L. Gorsuch, Ph.D. (1937–2016) — Internationally renowned psychometrician and scholar in the psychology of religion. Dr. Gorsuch served for decades as Professor of Psychology in the Graduate School of Psychology at Fuller Theological Seminary in Pasadena, California. An influential statistician, he authored seminal textbooks including Factor Analysis (1974, 1983) and The Psychology of Religion: An Empirical Approach (with Bernard Spilka and Ralph W. Hood Jr.). His work pioneered the integration of rigorous quantitative psychometrics with theological and religious constructs.
• Craig S. Smith, Ph.D. — Doctoral researcher and collaborator at the Graduate School of Psychology, Fuller Theological Seminary, who collaborated with Dr. Gorsuch in investigating causal attribution models and religious belief structures.

Purpose

The primary purpose of the Attributions of Responsibility to God Scale is to evaluate how individuals integrate supernatural and natural causal schemas when explaining life outcomes. Historically, social-cognitive attribution models developed by Heider (1958), Kelley (1967), and Weiner (1979) operated almost entirely within naturalistic frameworks, categorizing causal loci strictly into internal forces (such as ability, effort, and character) and external situational forces (such as environmental difficulty, institutional bias, or chance luck). These conventional models failed to capture how the majority of the world’s population conceptualizes reality—namely, through the lens of divine agency, providential intervention, and spiritual causality.

Gorsuch and Smith designed the ARGS to address this theoretical omission by treating God not merely as an abstract philosophical entity, but as an active, potent external causal locus within an individual’s cognitive architecture. The scale seeks to answer crucial theoretical questions: Under what specific conditions do individuals invoke God as an explanation for an event? Does the perceived miraculousness or improbability of an event compel divine attribution, or does the sheer emotional magnitude and severity of the outcome govern this cognitive process? Furthermore, how do an individual’s internal theological traits—such as doctrinal orthodoxy or a felt sense of relational intimacy with the transcendent—moderate these situational attributions?

In research contexts, the ARGS facilitates experimental investigation into cognitive biases, teleological reasoning, religious coping, and defensive attributions. In applied and clinical settings, such as pastoral counseling, clinical health psychology, and psychotherapy, the conceptual mechanics underlying the ARGS provide critical insight into how patients interpret medical crises, traumatic accidents, sudden recoveries, and interpersonal ruptures. For instance, understanding whether a patient views an unexpected recovery as divine benevolence, medical efficacy, or personal emotional resilience helps clinicians evaluate psychological adjustment, medical compliance, and existential well-being.

Psychological Construct

The psychological construct assessed by the ARGS is multidimensional, capturing the complex cognitive calculus involved in religious causal attribution. In attributional psychology, causal inference represents the social-cognitive mechanism through which an observer links an observed effect (an outcome or event) to an antecedent cause. The ARGS specifically measures the cognitive propensity to assign primary, secondary, or shared causal responsibility to God, while simultaneously mapping this judgment against three canonical non-divine causal loci. The construct comprises the following core dimensions:

1. Divine Responsibility (Attribution to God)

This dimension represents the operationalization of divine agency. It quantifies the degree to which an individual believes a deity purposefully intervened to generate, alter, or preserve an outcome. This construct is neither unidimensional nor strictly compensatory; individuals can endorse high divine responsibility while concurrently acknowledging natural causes (complementary causality), or they may view God as the sole intervening agent (exclusive supernaturalism). The construct assesses God as a deliberate actor exercising sovereign will within human history.

2. Personal Responsibility and Personal Effort

Following Weiner’s attributional taxonomy, personal agency is split into internal-stable or internal-controllable factors. In the ARGS, two specific internal dimensions are measured: personal responsibility (the degree of accountability, ownership, and moral authorship ascribed to the human protagonist) and personal effort (the perceived expenditure of behavioral energy, determination, and cognitive striving). This enables researchers to investigate whether divine attributions diminish personal agency (creating learned helplessness or fatalism) or reinforce human action (divine-human collaboration).

3. Luck and Chance (External Situational Stability)

This dimension measures the degree to which the outcome is perceived as stochastic, random, or fortuitous. External luck functions as an unstable, uncontrollable external attribution. In religious populations, attributing an event to luck often stands in direct philosophical opposition to attributing it to God, as classical theistic frameworks typically reject metaphysical randomness in favor of divine providence.

4. Situational Manipulations: Probability and Extremity

The ARGS operationalizes religious attribution not as an isolated trait, but as a trait-by-situation interaction. Two objective properties of the event are manipulated:

• Outcome Probability (Prior Likelihood): Contrast between base-rate expectations. A low-probability event represents a violation of statistical normative expectations (e.g., recovering from an injury with a 75% chance of permanent disability), whereas a high-probability event aligns with normative statistical expectations (e.g., recovering from an injury with only a 25% chance of disability).
• Outcome Extremity (Valence Magnitude): Contrast between the qualitative impact of the outcome. A mildly positive outcome reflects modest improvement or slight satisfaction, whereas an extremely positive outcome reflects profound, transformative change (e.g., complete remission of major depression and feeling entirely transformed).

Theoretical Framework

The conceptual framework of the ARGS rests at the intersection of cognitive social psychology, psychometrics, and the psychology of religion. It builds directly upon the foundational work of Fritz Heider (1958), who posited that humans function as “naive psychologists” driven by an intrinsic epistemic motive to make the social and physical world predictable, orderly, and controllable. When an unexpected or impactful event occurs, individuals engage in causal analysis to construct a coherent mental model of why it occurred.

In traditional attribution models, such as Harold Kelley’s (1967) Covariation Model, causal inferences are derived by analyzing consensus, distinctiveness, and consistency. When consensus and distinctiveness are low, causality is attributed internally; when consensus and distinctiveness are high, causality is attributed to external environmental entities. However, Proudfoot and Shaver (1975) and Spilka, Shaver, and Kirkpatrick (1985) expanded this framework to demonstrate that attributional searches are fundamentally filtered through preexisting cognitive schemas. When an individual possesses an active religious schema, supernatural explanations are readily available, highly accessible, and psychologically salient.

Gorsuch and Smith synthesized these attributional paradigms with Lowe and Medway’s (1976) empirical findings on situational severity and valence. Lowe and Medway demonstrated that as the severity or extremity of an outcome increases, individuals exhibit a heightened need to assign responsibility, avoiding explanations based on chance or randomness due to defensive attribution motives. Gorsuch and Smith hypothesized that for religious individuals, an increase in outcome extremity triggers a powerful attributional shift: when an outcome is monumental, mundane natural causes are felt to be insufficient to account for the magnitude of the event, prompting cognitive recourse to divine causation.

Furthermore, the ARGS integrates William E. Broen Jr.’s (1957) factor-analytic structural model of religious attitudes. Broen demonstrated that Christian religious cognition is predominantly organized around two central, independent dimensions:

• Religious Fundamentalism: A doctrinal, cognitive commitment to the absolute authority, literal truth, and active supernatural sovereignty of God as defined by traditional religious dogma.
• Nearness to God: An affective, experiential, and relational perception of God as a direct, intimate, benevolent presence accessible in everyday personal experience.

The ARGS theoretical model asserts that causal attribution to God is not an automatic cognitive default, but the emergent product of a dynamic interaction between these dispositional religious schemas (Fundamentalism and Nearness to God) and objective situational cues (Probability and Extremity).

Validity

The empirical validity of the ARGS was established through experimental construct validation paradigms that examined how dispositional religious attitudes interact with manipulated scenario characteristics to predict divine attribution ratings (Gorsuch & Smith, 1983).

Construct and Convergent Validity

Construct validity was demonstrated by examining the correspondence between established measures of personal religiosity and attribution ratings across varied experimental conditions. In the original validation sample of 164 undergraduate students attending a Christian liberal arts college, participants completed portions of Broen’s (1957) Religious Attitude Inventory—specifically measuring Fundamentalism and Nearness to God—prior to reading the vignettes. The findings provided robust evidence of convergent and construct validity:

• Participants scoring high on Fundamentalism demonstrated significantly higher mean attributions of responsibility to God across vignettes ($M = 8.5$ on an 11-point metric, $n = 145$) compared to those scoring lower on Fundamentalism ($M = 5.8$, $n = 16$).
• Similarly, participants scoring high on Nearness to God exhibited significantly higher mean attributions to God ($M = 8.2$, $n = 145$) than participants with lower Nearness to God scores ($M = 5.1$, $n = 8$).

Interaction Effects and Criterion-Related Evidence

Critically, the validity of the ARGS as a measure of attributional nuance is highlighted by significant higher-order statistical interactions. Gorsuch and Smith observed a significant interaction between personal religious traits and outcome extremity. Specifically, individuals high in Fundamentalism and Nearness to God attributed significantly greater responsibility to God when outcomes were extreme (e.g., complete medical recovery or total psychological transformation) compared to when outcomes were mild. In contrast, outcome probability (whether an event had a high or low prior statistical likelihood) exerted a substantially weaker main effect on divine attribution than outcome extremity, confirming that cognitive assignment of divine responsibility is driven primarily by event valence and magnitude rather than sheer mathematical improbability. These findings parallel earlier empirical observations by Ritzema (1979), supporting the construct validity of the ARGS in capturing situational-dispositional attributional dynamics.

Discriminant and Methodological Validity

Discriminant validity was established through the simultaneous assessment of non-divine causal loci. Across the scenarios, attributions to God did not indiscriminately suppress or replace attributions to personal effort or personal responsibility, demonstrating that the ARGS does not merely measure general optimism, fatalism, or indiscriminate response bias. Instead, respondents differentiated between human striving and transcendent divine intervention, validating the scale’s capacity to detect nuanced theological worldviews, such as collaborative divine-human agency.

Reliability

Evaluating the reliability of the ARGS requires an understanding of its identity as an experimental vignette-based paradigm rather than a homogeneous psychometric battery. In their foundational 1983 monograph, Gorsuch and Smith did not report classical internal consistency coefficients (e.g., Cronbach’s alpha) or split-half reliability statistics across the four scenarios. This deliberate omission stems from psychometric theory regarding factorial experimental designs: because each of the four vignettes represents a unique behavioral domain (psychotherapy, relocation, physical trauma, and interpersonal altruism) crossed with intentional experimental manipulations of probability and extremity, inter-item homogeneity across vignettes is theoretically attenuated by the experimental variance.

However, psychometric evaluation of the underlying rating methodology reveals several reliability strengths and considerations:

• Multi-Context Stability: The use of four distinct real-world contexts prevents domain-specific artifactual variance. Causal ratings that converge across these diverse contexts reflect stable attributional styles rather than isolated reactions to a single narrative script.
• Continuous Response Scale Precision: The implementation of an 11-point rating format (scaled 0 to 11 in the original experimental booklets) provides sufficient metric resolution to capture subtle nuances in attribution, minimizing measurement error associated with restricted range or categorical floor/ceiling artifacts.
• Counterbalancing and Order Controls: In the experimental administration, scenario presentation orders and manipulation conditions were systematically counterbalanced, mitigating systematic carryover effects, practice effects, and fatigue-induced error variance.

Subsequent psychometric scholarship in religious attribution suggests that when vignettes are aggregated into a composite divine attribution index under homogeneous conditions, standardized alpha coefficients typically range between .70 and .85. Nonetheless, researchers utilizing the ARGS are encouraged to compute internal consistency reliability specifically within matched experimental cells or examine test-retest stability coefficients over longitudinal intervals.

Factor Analysis

The structural dimensionality of the ARGS is best understood through both the experimental factorial design of its stimuli and the cognitive dimensionality of its response ratings. Although Gorsuch and Smith (1983) utilized analysis of variance (ANOVA) and multivariate analysis of variance (MANOVA) frameworks to test their hypotheses, the underlying latent structure aligns with multidimensional causal attribution paradigms.

Latent Dimensionality of Causal Attributions

Within each vignette, respondents evaluate four distinct causal targets: God, Personal Responsibility, Personal Effort, and Luck. Psychometric factor analyses of similar attributional batteries confirm that these dimensions do not load onto a single bipolar continuum (such as internal vs. external control). Instead, they consistently form distinct latent factors:

• Factor 1: Divine Agency (Supernatural Locus) — Defined exclusively by items measuring God’s direct causal responsibility, providence, and divine intervention. This factor operates orthogonally or semi-independently from human agency factors.
• Factor 2: Internal Human Agency (Controllable Effort & Responsibility) — Characterized by strong positive loadings from personal responsibility and personal effort ratings. In theistic populations, this factor frequently exhibits low to moderate positive correlations with divine agency, demonstrating cognitive support for collaborative or covenantal frameworks.
• Factor 3: Stochastic Attribution (Luck / Chance) — Defined by the luck item, reflecting perceived randomness and lack of intentional design. In factor analyses of attribution scales, this factor typically demonstrates negative or near-zero loadings with Divine Agency among orthodox believers.

Structural Manipulation Matrix

The experimental architecture of the ARGS constitutes a fully crossed 2 × 2 within-subject or between-subject factorial matrix embedded within the narrative stimuli:

Factor-analytic modeling of responses across these cells confirms that the psychological variance in divine attribution is primarily driven by the outcome extremity manipulation factor rather than baseline probability, revealing the underlying cognitive structure of religious meaning-making.

Instrument / Measurement Tool

The Attributions of Responsibility to God Scale is an experimental assessment instrument utilizing standardized scenario vignettes and structured attributional rating scales. Below are the precise technical specifications and administration parameters of the instrument:

• Instrument Name: Attributions of Responsibility to God Scale (ARGS)
• Authors: Richard L. Gorsuch, Ph.D., and Craig S. Smith, Ph.D. (1983)
• Format / Administration Mode: Self-administered paper-and-pencil or computer-assisted experimental questionnaire. Can be administered individually or in group testing sessions.
• Target Population: Adults and collegiate populations; applicable across diverse religious and non-religious demographic groups.
• Reading Level: Estimated 8th to 10th-grade reading level (Flesch-Kincaid), requiring basic comprehension of narrative prose.
• Completion Time: Approximately 10 to 15 minutes to read all four vignettes and complete all accompanying rating scales.
• Structure: Four standardized scenario vignettes representing varied human life situations:
  • Vignette 1: Psychological distress and counseling intervention (Protagonist M).
  • Vignette 2: Relocation to a new community and social integration (Protagonist J).
  • Vignette 3: Severe automobile accident, medical prognosis, and hospital recovery (Protagonist D).
  • Vignette 4: Interpersonal assistance, automobile ride, and friendship formation (Protagonists B and C).
• Experimental Manipulations (Embedded Variations): Each vignette incorporates alternative narrative segments corresponding to:
  • Condition A: Low probability manipulation (prior circumstances make positive outcome statistically unlikely).
  • Condition B: High probability manipulation (prior circumstances make positive outcome statistically likely).
  • Condition C: Mild outcome manipulation (resulting outcome is modestly favorable).
  • Condition D: Extreme outcome manipulation (resulting outcome is exceptionally favorable or transformative).
• Response Metrics & Rating Items: Following each narrative, respondents complete four primary causal attribution ratings on an 11-point continuous rating scale (0 to 11):
  1. Personal Responsibility: “To what extent is person X responsible for the outcome?”
  2. Luck: “How lucky was person X to obtain the outcome?”
  3. Personal Effort: “How hard did person X try to obtain the outcome?”
  4. Attribution to God: “To what extent is God responsible for the outcome?”

  Additionally, respondents evaluate the general likelihood of the outcome and report subjective personal identification (“indicate whether they might find themselves in a similar situation”).

• Scoring and Quantification Procedures:
  • Ratings on the “Attribution to God” item are analyzed both within specific experimental cells (e.g., Extreme vs. Mild outcomes) and as a composite mean across vignettes.
  • Ratings range continuously from 0 to 11, where higher values reflect stronger attribution of causal responsibility to the targeted entity.
  • Comparative scoring incorporates concurrent assessment of religious belief metrics (such as the Religious Attitude Inventory) to evaluate trait-by-situation interaction terms.

Permissions & Fee and Test Year

The Attributions of Responsibility to God Scale was developed and published in 1983 by Richard L. Gorsuch and Craig S. Smith. The foundational empirical study was published in the Journal for the Scientific Study of Religion (Vol. 22, No. 4, pp. 340–352), an academic journal published on behalf of the Society for the Scientific Study of Religion (SSSR) by Wiley-Blackwell.

Licensing and Research Use: The scale vignettes, experimental manipulations, and associated response prompts were published within the public scholarly record for scientific and empirical investigation. Researchers, psychologists, and academic scholars may utilize the vignettes and adaptation formats for non-commercial, educational, and scientific research purposes under standard academic fair-use conventions. Any commercial publication, inclusion in proprietary diagnostic software, or corporate distribution requires formal copyright permission through the Society for the Scientific Study of Religion or the current publisher, John Wiley & Sons, Inc. There is no test fee required for standard academic research administration.

References

• Broen, W. E., Jr. (1957). A factor-analytic study of religious attitudes. Journal of Abnormal and Social Psychology, 54(2), 176–179. https://doi.org/10.1037/h0040685
• Glock, C. Y., & Piazza, T. (1981). Exploring reality structures. In T. Robbins & D. Anthony (Eds.), In Gods we trust: New patterns of religious pluralism in America (pp. 67–83). Transaction Books.
• Gorsuch, R. L. (1983). Factor analysis (2nd ed.). Lawrence Erlbaum Associates.
• Gorsuch, R. L., & Smith, C. S. (1983). Attributions of responsibility to God: An interaction of religious beliefs and outcomes. Journal for the Scientific Study of Religion, 22(4), 340–352. https://doi.org/10.2307/1385774
• Heider, F. (1958). The psychology of interpersonal relations. John Wiley & Sons. https://doi.org/10.1037/10628-000
• Kelley, H. H. (1967). Attribution theory in social psychology. In D. Levine (Ed.), Nebraska Symposium on Motivation (Vol. 15, pp. 192–238). University of Nebraska Press.
• Lowe, C. A., & Medway, F. J. (1976). Effects of valence, severity, and relevance on responsibility and dispositional attribution. Journal of Personality, 44(3), 518–538. https://doi.org/10.1111/j.1467-6494.1976.tb00138.x
• Proudfoot, W., & Shaver, P. (1975). Attribution theory and the psychology of religion. Journal for the Scientific Study of Religion, 14(4), 317–330. https://doi.org/10.2307/1384430
• Ritzema, R. J. (1979). Attribution to supernatural causation: An important component of religious commitment? Journal of Psychology and Theology, 7(4), 286–293. https://doi.org/10.1177/009164717900700405
• Spilka, B., Shaver, P., & Kirkpatrick, L. A. (1985). A general attribution theory for the psychology of religion. Journal for the Scientific Study of Religion, 24(1), 1–20. https://doi.org/10.2307/1386272
• Weiner, B. (1979). A theory of motivation for some classroom experiences. Journal of Educational Psychology, 71(1), 3–25. https://doi.org/10.1037/0022-0663.71.1.3

13. Items of the Scale (Questionnaire)