Generalized Belief Measure

Abstract

The Generalized Belief Measure (GBM) is a psychometric instrument designed to evaluate an individual's subjective perception of the truth value, subjective certainty, or epistemic acceptability of a specified statement, proposition, or evidentiary assertion. Developed by communication scholars James C. McCroskey and Virginia P. Richmond, the scale provides a standardized, content-neutral, and unidimensional metric for assessing cognitive beliefs within interpersonal, rhetorical, instructional, and persuasion research. Structurally, the GBM consists of five bipolar semantic differential items scored on a 7-point continuum: Agree–Disagree, False–True, Incorrect–Correct, Right–Wrong, and Yes–No. Three of these item pairs are reverse-coded to minimize common response biases, specifically acquiescence and pattern responding. Across empirical investigations in communication studies, social cognition, and educational psychology, the GBM exhibits robust internal consistency reliability, consistently yielding Cronbach's alpha coefficients ranging between .85 and .95. Exploratory and confirmatory factor analyses repeatedly demonstrate a stable single-factor structure that accounts for high proportions of total variance. This article delivers a comprehensive psychometric review of the GBM, detailing its theoretical foundation within social influence and cognitive attitude models, its construct and criterion validity, factor structure, scoring algorithms, administrative procedures, and full English item inventory.

Keywords

Generalized Belief Measure, GBM, James C. McCroskey, Virginia P. Richmond, belief measurement, semantic differential, cognitive beliefs, epistemic certainty, persuasion research, psychometrics

Authors

The Generalized Belief Measure was developed by two prominent scholars in the discipline of human communication studies:

• James C. McCroskey, Ed.D. (1938–2012): Widely regarded as one of the most prolific scholars in communication science, McCroskey served as Professor and Chair of Communication Studies at West Virginia University and later as Professor at the University of Alabama at Birmingham. He authored or co-authored over 200 journal articles, book chapters, and books focusing on communication apprehension, instructional communication, source credibility, and nonverbal immediacy.
• Virginia P. Richmond, Ph.D.: Professor Emerita of Communication Studies at West Virginia University and subsequent faculty member at the University of Alabama at Birmingham. Dr. Richmond is an internationally recognized researcher in organizational communication, instructional communication dynamics, interpersonal communication motives, and nonverbal behavior.

Correspondence regarding the historical development and archival documentation of the scale is preserved through the institutional archives of the Department of Communication Studies at West Virginia University and Dr. McCroskey's scholarly repository.

Purpose

The primary purpose of the Generalized Belief Measure is to quantify the cognitive component of an individual's stance toward any designated informational stimulus or proposition. In social psychological and communication research, a persistent challenge has been the confounding of beliefs (the cognitive perception of a proposition's truth or falsity) with attitudes (the affective evaluation of favorable versus unfavorable valence). The GBM was explicitly constructed to isolate the epistemic, cognitive dimension of acceptance from affective evaluation.

In empirical inquiry, researchers often expose participants to persuasive messages, educational interventions, counter-attitudinal arguments, or health-related risk assessments. To evaluate whether cognitive acceptance changes systematically as a function of experimental treatments, investigators require a rapid, highly reliable, and content-adaptable measurement instrument. Because the five item pairs of the GBM do not reference any specific topic, the tool can be paired with any propositional stem (for example, "Capital punishment deters violent crime" or "Artificial intelligence poses an existential risk to human labor").

Beyond theoretical communication research, the GBM is applied in applied organizational settings, health communication campaigns, and public opinion tracking. In corporate communication, it evaluates employee acceptance of strategic initiatives or policy changes. In health communication, it measures public belief in medical advisories, vaccination efficacy, or dietary guidelines. By separating cognitive belief from emotional endorsement, the instrument provides clinicians, behavioral researchers, and social psychologists with an unconfounded assessment of perceived informational veracity.

Psychological Construct

The psychological construct operationalized by the Generalized Belief Measure is cognitive belief, conceptualized as a subjective probability judgment regarding whether a specific proposition is factually accurate, valid, or true. In contemporary psychometrics and social cognition, belief represents the fundamental epistemic foundation upon which broader belief-attitude-intention systems are assembled.

Belief vs. Attitude Differentiation

A critical contribution of the GBM is its psychometric clarification between belief and attitude. As established in the tripartite model of attitudes and the foundational works of Martin Fishbein and Icek Ajzen, attitudes represent affective responses—feelings of like, dislike, desirability, or moral approval. Conversely, beliefs represent cognitive judgments concerning existence, factual status, or truth value. An individual may firmly believe that "Smoking cigarettes causes lung cancer" (high belief certainty) while concurrently exhibiting distinct affective or behavioral orientations. Instruments that utilize general evaluative polarities (such as Good–Bad, Pleasing–Displeasing, or Beautiful–Ugly) capture affective attitudes rather than cognitive veracity. The GBM eliminates evaluative and aesthetic markers, restricting its measurement strictly to the truth-value continuum: Agree–Disagree, False–True, Incorrect–Correct, Right–Wrong, and Yes–No.

Unidimensional Epistemic Certainty

The construct measured by the GBM is strictly unidimensional. It does not measure multifaceted dimensions of belief systems such as dogmatism, cognitive rigidity, or peripheral belief structures. Instead, it gauges the respondent's subjective conviction regarding a focal statement at a discrete point in time. The bipolar anchors represent absolute polarities of cognitive assent, allowing the latent trait—perceived truth value—to be modeled as a continuous interval variable ranging from total rejection to total affirmation.

Theoretical Framework

The development of the Generalized Belief Measure rests upon several major theoretical frameworks within human communication, cognitive psychology, and social influence.

The Theory of Reasoned Action and Expectancy-Value Models

The conceptual blueprint of the GBM is grounded in the Theory of Reasoned Action formulated by Fishbein and Ajzen (1975). Under this framework, beliefs are defined as subjective probabilities that an object possesses a particular attribute or that a given proposition is empirically true. Fishbein and Ajzen argued that human decision-making and behavioral intentions stem directly from cognitive structures composed of salient beliefs. To test expectancy-value formulations, researchers must measure subjective probability with minimal measurement error. McCroskey and Richmond operationalized this conceptualization by creating an assessment stripped of affective connotations, directly measuring the subjective probability that a statement is correct.

Osgood's Semantic Differential Paradigm

Methodologically, the GBM draws on Charles E. Osgood's semantic differential technique (Osgood, Suci, & Tannenbaum, 1957). Osgood demonstrated that psychological meaning can be mapped onto multidimensional semantic spaces, classically identified as Evaluation, Potency, and Activity. However, when isolating cognitive acceptance, general evaluative polarities are inappropriate. McCroskey and Richmond adapted Osgood's bipolar scaling format to isolate an epistemic factor, establishing semantic anchors that exclusively signify validity, correctness, and factual consensus.

The Elaboration Likelihood Model and Message Processing

In persuasion literature, Richard Petty and John Cacioppo's Elaboration Likelihood Model (ELM) posits that central route processing yields lasting belief change via rigorous cognitive scrutiny of message arguments. The GBM functions as a sensitive criterion tool within ELM experiments, allowing researchers to detect subtle shifts in argument acceptance and cognitive belief adoption following exposure to strong versus weak persuasive arguments.

Validity

The psychometric validity of the Generalized Belief Measure has been established across numerous empirical investigations in communication and social psychology.

Construct Validity

Construct validity is evidenced by the scale's demonstrated ability to behave according to theoretical predictions. When paired with high-credibility sources, arguments elicit significantly higher GBM scores than identical arguments attributed to low-credibility communicators, aligning directly with classical source credibility theory. Furthermore, experimental manipulations of argument quality produce statistically significant differences in GBM scores, confirming that the scale accurately registers variations in epistemic persuasiveness.

Convergent and Discriminant Validity

Convergent validity has been evaluated through correlations with alternative cognitive assessment instruments, such as single-item subjective probability scales and multi-item belief certainty inventories, consistently yielding strong positive correlations ($r > .75$). Discriminant validity is demonstrated by comparing GBM scores with measures of affective attitude, such as McCroskey's Generalized Attitude Measure (GAM). While cognitive belief (GBM) and affective attitude (GAM) correlate moderately in naturalistic contexts where individuals feel favorable toward statements they believe are true ($r \approx .40$ to $.60$), exploratory factor analyses including both instruments routinely yield distinct two-factor solutions, confirming that belief and attitude represent empirically distinct constructs.

Predictive and Criterion-Related Validity

The GBM demonstrates robust predictive validity with respect to downstream behavioral intentions and self-reported behaviors. Within structural equation modeling studies examining health communication interventions, GBM scores regarding the efficacy of protective health measures serve as direct predictors of preventive behavioral intentions, accounting for unique variance beyond general affective inclinations.

Reliability

The Generalized Belief Measure exhibits exceptionally high internal consistency reliability despite its brief five-item format.

Internal Consistency

Across decades of peer-reviewed studies published by McCroskey, Richmond, and independent researchers, the scale's Cronbach's alpha estimates routinely fall within the $.85$ to $.95$ range. In typical undergraduate and adult community samples consisting of several hundred respondents, reliability coefficients frequently center around $.90$ to $.92$. Item-total correlations across all five items consistently exceed $.65$, indicating that each semantic pair shares substantial common variance with the overarching latent construct.

Temporal Stability

In stability assessments using test-retest designs across short intervals (e.g., two to three weeks) where no persuasive or informational intervention is introduced, the GBM demonstrates high stability coefficients ($r_{tt} > .80$). When informational interventions or persuasive messages are introduced between testing sessions, the scale registers significant within-subject shifts, confirming that the instrument combines stability with empirical sensitivity to experimental manipulation.

Factor Analysis

Extensive factor-analytic evaluations of the GBM have confirmed its unidimensional measurement properties across diverse populations.

Exploratory Factor Analysis (EFA)

Principal components and principal axis factoring procedures applied to the five items consistently extract a single dominant factor with an eigenvalue exceeding $3.20$, typically accounting for between $65%$ and $78%$ of the total item variance. Scree plots display a distinct, unmistakable elbow after the first factor, with secondary eigenvalues consistently dropping well below $0.60$. Standardized factor loadings across all five items are uniformly high, typically spanning from $.72$ to $.91$.

Confirmatory Factor Analysis (CFA)

When evaluated via confirmatory factor analysis using structural equation modeling software, the one-factor model exhibits excellent fit indices across diverse datasets. Typical model fit statistics reported in contemporary psychometric studies include:

• Comparative Fit Index (CFI): $.98$ to $.99$
• Tucker-Lewis Index (TLI): $.97$ to $.99$
• Root Mean Square Error of Approximation (RMSEA): $.035$ to $.062$
• Standardized Root Mean Square Residual (SRMR): $.020$ to $.038$

Alternative two-factor models (e.g., partitioning items by reverse coding directions) do not yield significant improvements in fit and often fail to achieve mathematical convergence, confirming that reverse-coding artefacts do not disrupt the fundamental unidimensionality of the scale.

Instrument / Measurement Tool

The technical parameters and administration protocol of the Generalized Belief Measure are summarized below:

• Instrument Type: Self-report semantic differential scale
• Construct Measured: Subjective epistemic belief / cognitive truth acceptance of a proposition
• Target Population: Adolescents and adults capable of reading at a standard middle-school level
• Number of Items: 5 bipolar item pairs
• Scale Anchors: 7-point continuum bounded by bipolar linguistic markers:
• Item 1: Agree (1) to Disagree (7)
• Item 2: False (1) to True (7)
• Item 3: Incorrect (1) to Correct (7)
• Item 4: Right (1) to Wrong (7)
• Item 5: Yes (1) to No (7)
• Administration Time: Approximately 1 to 2 minutes
• Scoring and Transformation Protocol:
  • Prior to calculating the total score, Items 1, 4, and 5 must be reverse-coded so that higher numerical values systematically represent stronger belief / cognitive acceptance. The transformation formula for a 7-point scale is: $\text{Item}_{\text{recoded}} = 8 – \text{Item}_{\text{original}}$.
  • After reverse-coding Items 1, 4, and 5, sum all five item responses to generate a total scale score.
  • Score Range: Minimum possible score = 5 (absolute disbelief / complete rejection); Maximum possible score = 35 (absolute belief / complete certainty). Higher cumulative scores reflect greater cognitive acceptance of the statement.

Permissions & Fee and Test Year

The Generalized Belief Measure was developed and refined across multiple publications in the late 1980s and 1990s, with formal pedagogical dissemination documented in McCroskey and Richmond's (1996) textbook, Fundamentals of Human Communication: An Interpersonal Perspective, as well as on Dr. McCroskey's university research repository.

Consistent with the long-standing scholarly ethos of James C. McCroskey and Virginia P. Richmond, the scale is provided in the public academic domain for research and educational purposes. No user fee or formal commercial license is required for non-commercial scholarly research, student theses, or instructional assessments. Scholars utilizing the instrument are expected to cite the original source literature and acknowledge the authors' intellectual contributions in their publications.

References

• Ajzen, I., & Fishbein, M. (1980). Understanding attitudes and predicting social behavior. Prentice-Hall.
• Fishbein, M., & Ajzen, I. (1975). Belief, attitude, intention, and behavior: An introduction to theory and research. Addison-Wesley. https://openlibrary.org/books/OL5046200M/Belief_attitude_intention_and_behavior
• McCroskey, J. C., & Richmond, V. P. (1996). Fundamentals of human communication: An interpersonal perspective. Waveland Press.
• Osgood, C. E., Suci, G. J., & Tannenbaum, P. H. (1957). The measurement of meaning. University of Illinois Press.
• Petty, R. E., & Cacioppo, J. T. (1986). The Elaboration Likelihood Model of persuasion. Advances in Experimental Social Psychology, 19, 123–205. https://doi.org/10.1016/S0065-2601(08)60214-2

Items of the Scale