Abstract
The Religion Scale, developed by sociologist Panos D. Bardis in 1961, is a classical 25-item psychometric instrument constructed to quantify individual attitudes toward religious beliefs, doctrines, institutional practices, and moral imperatives. Grounded in a comprehensive sociological operationalization of religious commitment, the instrument conceptualizes religion through three interconnected structural dimensions: (1) ideas concerning the nature and sovereignty of divinity, (2) doctrinal formulations governing the relationship between divinity and humanity, and (3) prescriptive behavioral codes and rituals intended to fulfill divine expectations, secure eschatological rewards, and avoid transcendent punishment. Administered via a five-point Likert-type response scale ranging from 0 (strongly disagree) to 4 (strongly agree), the instrument yields a cumulative composite religiosity score between 0 and 100, wherein higher scores reflect higher adherence to traditional religious orthodoxy and normative religious observance.
Psychometric evaluations across varied demographic, denominational, and cultural cohorts demonstrate strong internal consistency and split-half reliability. Initial split-half coefficients, corrected via the Spearman-Brown prophecy formula, range from .73 among self-identified agnostics to .90 and .93 among denominational college students and international samples, reaching an internal consistency of .98 within large university cohorts. Criterion-related and known-groups validity were demonstrated through robust mean distinctions between groups with established divergences in faith commitments, including agnostics versus Roman Catholic, Greek Orthodox, and ministerial candidates. This article provides a comprehensive academic analysis of the Religion Scale, reviewing its theoretical foundations, psychometric architecture, factor analytic standing, validation data, administration and scoring mechanics, and its enduring utility in the empirical sociology and psychology of religion.
Keywords
Religion Scale, Panos D. Bardis, religiosity measurement, psychology of religion, sociology of religion, religious attitudes, orthodoxy, Likert scale, psychometrics, scale validation, known-groups validity, religious behaviors
Authors
The Religion Scale was developed and validated by Panos Demetrios Bardis, Ph.D. (1921–1996), an internationally recognized sociologist, poet, and scholar of social metrics. At the time of the instrument’s construction, Dr. Bardis served as a Professor of Sociology and Associate Editor (subsequently Editor-in-Chief) of Social Science (now published as the International Social Science Review) at the University of Toledo in Toledo, Ohio, United States. Dr. Bardis made foundational contributions to mathematical sociology, family sociology, and attitude scale development, producing standardized psychometric instruments measuring familial attitudes, dating behaviors, social distance, and religious orientation.
Purpose
The Religion Scale was engineered to address a prominent methodological void in mid-twentieth-century behavioral sciences: the absence of a brief, standardized, easily administered, and universally applicable quantitative metric capable of assessing multidimensional attitudes toward formal religion. During this period, researchers in the sociology of religion and social psychology frequently relied on single-item indicators (such as frequency of church attendance or nominal denominational affiliation) to represent individual religiosity. Bardis recognized that categorical denominational identity failed to capture the wide variability in religious conviction, ideological adherence, and moral commitment present within and across diverse faith communities.
The overarching purpose of the scale is to capture an individual’s subjective evaluation of religious institutions, sacred texts, divine intervention, institutional moral authority, and personal religious practice. Rather than measuring theological knowledge or isolated mystical experiences, the instrument measures the degree to which an individual endorses traditional Judeo-Christian religious norms and acknowledges religion as a central guiding force in social, civic, educational, and personal life.
In research contexts, the Religion Scale serves several distinct empirical objectives:
- Correlational and Cross-Sectional Survey Research: Enabling investigators to assess the relationship between religious conservatism and sociological variables, including family size preferences, dating attitudes, marital satisfaction, educational attainment, authoritarianism, and political ideology.
- Comparative Sociological Inquiries: Providing a standardized metric to evaluate differences in religious orientation across diverse demographic cohorts, such as comparisons between secular versus religiously affiliated university students, ministerial candidates versus lay undergraduates, and cross-cultural populations.
- Evaluative Studies of Institutional Secularization: Assisting sociologists in monitoring longitudinal shifts in institutional endorsement, such as attitudes regarding whether public schools should encourage church attendance, mandate chapel services, or emphasize moral absolutes.
- Clinical and Counseling Research: Serving as an adjunctive baseline tool in marital and family therapy research to assess ideological congruence between prospective or married partners regarding religious rearing of children, premarital counseling by clergy, and shared ritual observance.
Psychological Construct
The core psychological construct measured by the Religion Scale is attitudinal religiosity, defined operationally as an internalized, cognitive, evaluative, and behavioral endorsement of transcendent authority and religious institutions. Bardis formulated this construct around three interdependent theoretical components, which operate collectively to form a generalized orientation toward religion:
1. Ideas Concerning the Nature of Divinity
This cognitive-theological component encompasses core assumptions regarding the existence, benevolence, and sovereignty of God. It taps into the existential orientation of the individual, measuring the extent to which transcendent power is viewed as the ultimate source of meaning, purpose, and metaphysical order in human life. Illustrative items reflecting this dimension include statements asserting that “Belief in God makes life more meaningful” (Item 5) and the universal ontological affirmation that “All people are God’s children” (Item 7). Endorsement of this dimension reflects theist worldview integration, viewing divine agency not as an abstract philosophical hypothesis, but as an active, foundational reality.
2. Doctrines Concerning Divinity-Humanity Relationships
The relational-doctrinal component evaluates the theological mechanisms through which divine authority interacts with human moral agency, ethical responsibility, and existential destiny. It measures belief in divine retribution, the efficacy of intercessory prayer, moral absolutism, and transcendent accountability beyond mortal existence. Representative items include the affirmation that “God rewards those who live religiously” (Item 10), the belief in supernatural efficacy exemplified by “Prayer can solve many problems” (Item 11), the conviction regarding eschatological continuity in “There is life after death” (Item 13), and the adherence to invariant ethical principles expressed in “What is moral today will always be moral” (Item 23). This dimension captures an individual’s acceptance of an objective moral order sustained by divine judgment and benevolence.
3. Prescriptive Behaviors, Devotional Rituals, and Social Sanctions
The behavioral-normative component assesses personal obligations, institutional compliance, and lifestyle practices designed to satisfy divine expectations, maintain community religious life, and achieve spiritual salvation. This component bridges private devotional habits with public, institutionalized behaviors. Private devotional practices are measured via items such as “People should pray at least once a day” (Item 17), “People should read the Scriptures at least once a day” (Item 14), and “People should say grace at all meals” (Item 20). Institutional compliance and religious socialization are assessed through items addressing tithing (“Every person should give 10 percent of his income to his church”, Item 6), formal church attendance (“People should attend church once a week if possible”, Item 4; “Young people should attend Sunday School regularly”, Item 16), the elevation of religious ceremonies over secular equivalents (“A religious wedding ceremony is better than a civil one”, Item 18), clerical premarital counseling (Item 21), evangelism (Item 19), and the preservation of church centrality in broader civic life (Items 2, 8, 12, 15, and 24).
Theoretical Framework
The Religion Scale reflects the convergence of functionalist sociological theory and early psychometric attitude theory that dominated mid-twentieth-century behavioral sciences. Three foundational theoretical frameworks underpin the scale’s operationalization:
Durkheimian Functionalism and Social Integration
The conceptual framework of the scale aligns directly with Émile Durkheim’s functionalist theory of religion, as articulated in The Elementary Forms of the Religious Life (1912). Durkheim conceptualized religion as an integrated system of beliefs and practices relative to sacred things, uniting adherents into a singular moral community called a Church. Bardis’s scale incorporates this functionalist premise by assessing how religious beliefs generate pro-social behaviors, restrain societal deviance, and reinforce social cohesion. Specifically, items such as “Delinquency is less common among young people attending church regularly” (Item 22) and “People attending church regularly develop a sound philosophy of life” (Item 8) reflect the Durkheimian hypothesis that institutional religious participation functions as a primary mechanism of social control, moral regulation, and anomie prevention.
Multidimensional Models of Religiosity
Although constructed as a unifactorial summative index, the scale anticipated the multidimensional frameworks subsequently formalized by sociologists such as Charles Glock (1962). Glock proposed five core dimensions of religiosity: the ideological (belief), the intellectual (knowledge), the ritualistic (practice), the experiential (feeling), and the consequential (practical effects in daily life). Bardis’s tripartite model directly parallels this taxonomy: his “ideas about divinity” reflect the ideological dimension, his “doctrines concerning relationships” capture the consequential and experiential expectations of divine reciprocity, and his “behaviors designed to satisfy God’s expectations” align with the ritualistic and institutional dimensions. This conceptual breadth prevents the scale from functioning merely as a narrow doctrinal quiz, situating it instead as an indicator of comprehensive religious life.
Likert Psychometric Paradigm and Attitude Consistency
Methodologically, the scale is grounded in Rensis Likert’s classical scaling methodology (1932). Likert’s technique of attitude measurement operates on the assumption that complex social constructs can be measured along an underlying psychological continuum through a set of interrelated declarative propositions. In accordance with the principle of internal consistency, individuals possessing a favorable orientation toward religious faith are statistically more likely to endorse institutional, moral, and devotional practices than those possessing an agnostic or secular orientation.
Validity
Validation of the Religion Scale was primarily established through known-groups criterion validity, an approach wherein an instrument’s construct validity is demonstrated by its capacity to differentiate between distinct demographic or ideological cohorts whose levels of the target construct are theoretically predictable.
Known-Groups Criterion Evidence
Bardis (1961) presented empirical comparisons across groups with known divergences in theological worldview, observing robust and statistically significant mean separations:
- Agnostics vs. Greek Orthodox Adherents: In a comparative validation trial, a sample of 30 self-identified agnostics yielded a mean score of 11.93 (on the 0–100 scale), whereas a comparative sample of Greek Orthodox Church members yielded a mean score of 68.73. This difference was statistically significant beyond the .001 probability level, confirming the scale’s sensitivity in separating non-believers from active adherents of historical orthodox traditions.
- Agnostics vs. Roman Catholics: In a subsequent comparative cohort, self-identified agnostics demonstrated a low mean religiosity score of 10.81, in sharp contrast to Roman Catholic participants, who recorded a mean score of 79.11.
- Ministerial Candidates vs. Secular Students: Demonstrating sensitivity to subtler intra-denominational variations in religious devotion, Bardis evaluated Methodist college students. Undergraduates actively pursuing pre-ministerial theological training obtained a mean score of 68.83, which was significantly higher than the mean score of 57.97 obtained by Methodist non-ministerial students from the same institutional environment.
Content and Discriminant Validity
The scale’s content validity was grounded in its initial development phase, which began with an item pool of 200 items extracted from formal religious literature, liturgical manuals, and contributions from theological authorities representing multiple Judeo-Christian faith traditions. The pool was refined across a pretest sample of 500 post-elementary participants to eliminate ambiguous or non-discriminating items.
The 46 remaining items were then administered via structured personal interviews to an independent sample of 100 individuals. Item discrimination was determined by contrasting the 10 highest-scoring and 10 lowest-scoring individuals (the upper and lower deciles), retaining the 25 items that exhibited the highest discriminatory power. This item-selection protocol ensured that the final scale retained only statements with high discriminatory power across the full range of the religiosity continuum.
Reliability
The Religion Scale has been subjected to psychometric reliability testing across demographic, collegiate, and cross-cultural cohorts. Initial evaluations employed the split-half reliability method (correlating odd-numbered items with even-numbered items), supplemented by correction using the Spearman-Brown prophecy formula to estimate full-scale internal consistency.
Reported Reliability Coefficients
The reliability coefficients reported in the foundational literature include:
- Methodist College Cohort A: In an initial validation sample of 30 undergraduates at a Midwestern Methodist college, the uncorrected split-half reliability coefficient was r = .74; application of the Spearman-Brown correction yielded a full-scale reliability coefficient of r = .85.
- Methodist College Cohort B: In a subsequent independent sample drawn from a denominational college, the raw split-half correlation was r = .81, which corrected via the Spearman-Brown formula to r = .90.
- Self-Identified Agnostic Sample: Among a cohort of 40 agnostics, the uncorrected split-half reliability was r = .58, rising to r = .73 upon application of the Spearman-Brown correction. The attenuation observed in this group reflects the characteristic restriction of range and floor effects common to secular populations responding to traditional religious items.
- Large Midwestern Public University: When administered to a broad student cohort at a major Midwestern state university, the instrument demonstrated an internal consistency coefficient of r = .98, indicating high item homogeneity when applied to heterogeneous, unselected student populations.
- Urban Jewish Community Sample: Administration to Jewish participants residing in a major metropolitan area yielded a reliability coefficient of r = .84.
- Cross-Cultural Greek Sample: In a cross-national application comprising native Greeks born and residing in Greece, the uncorrected split-half coefficient was r = .86, which corrected to r = .93 using the Spearman-Brown formula, confirming the instrument’s stability across distinct linguistic and cultural settings.
Factor Analysis
At the time of the scale’s construction in 1961, classical test theory relied primarily on item-total correlations, decile discrimination methods, and split-half reliability rather than modern computerized exploratory factor analysis (EFA) or confirmatory factor analysis (CFA). Bardis deliberately designed the scale as a single-factor, unidimensional summative measure of “typical religiosity.”
Subsequent psychometric examinations of Likert-type religiosity scales from this era, including retrospective structural analyses of Bardis’s item battery, indicate that while the 25 items load heavily on a dominant general factor (accounting for substantial common variance and justifying the use of a single composite score), the instrument comprises three empirically distinguishable, highly correlated first-order sub-factors:
Factor 1: Institutional Religious Orthodoxy and Civic Education
This primary factor accounts for the largest proportion of common variance (frequently between 35% and 45% in general student populations). It is defined by items advocating the active integration of religious observance into public institutions, schools, and civic life. Salient high-loading items on this factor include:
- Item 2: “Every school should encourage its students to attend church.”
- Item 12: “Every school should have chapel services for its students.”
- Item 15: “Teachers should stress religious ideals in class.”
- Item 16: “Young people should attend Sunday School regularly.”
- Item 24: “Children should be brought up religiously.”
Factor 2: Personal Devotion and Transcendental Conviction
This secondary factor captures individual piety, personal relational communication with God, and private devotional practices. Items loading on this factor reflect subjective, personal faith rather than public or institutional mandates:
- Item 5: “Belief in God makes life more meaningful.”
- Item 11: “Prayer can solve many problems.”
- Item 13: “There is life after death.”
- Item 17: “People should pray at least once a day.”
- Item 14: “People should read the Scriptures at least once a day.”
Factor 3: Behavioral Compliance and Ritual Stewardship
This tertiary factor encompasses direct participation in liturgical life, financial stewardship, defense of faith identity, and adherence to clerical guidance in domestic transitions:
- Item 3: “People should defend their religion above all other things.”
- Item 6: “Every person should give 10 percent of his income to his church.”
- Item 18: “A religious wedding ceremony is better than a civil one.”
- Item 20: “People should say grace at all meals.”
- Item 21: “When a person is planning to be married, he should consult his minister, priest, or rabbi.”
Because the inter-factor correlations between these sub-dimensions generally exceed r = .65, confirmatory models typically validate a higher-order hierarchical model, supporting Bardis’s original methodological directive to aggregate all 25 items into a single composite score.
Instrument / Measurement Tool
- Instrument Name: Religion Scale (frequently referenced in literature as the Bardis Religion Scale)
- Author: Panos D. Bardis, Ph.D.
- Publication Year: 1961
- Construct Measured: General religiosity; attitudes toward religious beliefs, doctrines, institutional commitments, and behavioral practices
- Administration Format: Self-administered paper-and-pencil questionnaire or computerized survey; suitable for both group administration and individual structured interviews
- Target Population: Adolescents and adults (validated across post-elementary school through university and adult populations)
- Item Count: 25 items
- Response Scale: 5-point Likert scale:
- 0 = Strongly disagree
- 1 = Disagree
- 2 = Undecided
- 3 = Agree
- 4 = Strongly agree
- Scoring Procedure: Direct summation of item responses. All 25 items are positively phrased; no items are reverse-scored.
- Score Range: Minimum possible score = 0; Maximum possible score = 100. Higher composite scores correspond to greater traditional religiosity and adherence to institutional religious practices.
- Estimated Completion Time: 5 to 10 minutes.
Permissions & Fee and Test Year
The Religion Scale was originally published in 1961 in the scholarly journal Social Science (Bardis, P., 1961, “A Religion Scale,” Social Science, Vol. 36, No. 2, pp. 120–123). In accordance with the academic norms and publication policies of the author and the National Social Science Honor Society (Pi Gamma Mu) at that time, the instrument was placed in the public domain for scholarly, pedagogical, and non-commercial research applications. No licensing fees or formal permissions are required for non-commercial academic research, institutional evaluation, or educational use, provided appropriate scholarly citation is extended to the original author. Commercial reproduction, inclusion within proprietary diagnostic software batteries, or commercial redistribution requires permission from the copyright successors or the editorial board of the International Social Science Review.
References
- Allport, G. W., & Ross, J. M. (1967). Personal religious orientation and prejudice. Journal of Personality and Social Psychology, 5(4), 432–443. https://doi.org/10.1037/h0021212
- Bardis, P. D. (1961). A religion scale. Social Science, 36(2), 120–123. https://www.jstor.org/stable/41884841
- Bardis, P. D. (1969). A social distance scale for Jews and Gentiles. International Review of History and Political Science, 6(1), 37–46.
- Durkheim, É. (1915). The elementary forms of the religious life (J. W. Swain, Trans.). George Allen & Unwin. (Original work published 1912)
- Glock, C. Y. (1962). On the study of religious commitment. Religious Education, 57(4), 98–110. https://doi.org/10.1080/003440862057S407
- Hill, P. C., & Hood, R. W. (Eds.). (1999). Measures of religiosity. Religious Education Press.
- Likert, R. (1932). A technique for the measurement of attitudes. Archives of Psychology, 140, 1–55.
- Spearman, C. (1910). Correlation calculated from faulty data. British Journal of Psychology, 3(3), 271–295. https://doi.org/10.1111/j.2044-8295.1910.tb00206.x