Source Credibility Measures

1. Abstract

The Source Credibility Measures developed by James C. McCroskey and Jason J. Teven (1999) represent a landmark psychometric operationalization of perceived communicator credibility. Built upon classical Aristotelian rhetorical theory and decades of empirical revisionism in communication studies, the instrument measures the multidimensional construct of perceived source ethos across three distinct dimensions: Competence, Trustworthiness, and Goodwill/Caring. The modern standardized instrument comprises 18 semantic differential items distributed equally across these three 6-item factors. Each item utilizes a 7-point bipolar response scale anchored by diametrically opposed evaluative adjectives (e.g., Intelligent–Unintelligent, Honest–Dishonest, Cares about me–Doesn’t care about me). Psychometric evaluations consistently demonstrate outstanding internal consistency, with Cronbach’s alpha coefficients regularly exceeding .85 to .92 across each subscale in diverse organizational, instructional, political, and mediated communication environments. Confirmatory factor analyses corroborate the oblique three-factor structural model over rival unidimensional or two-factor formulations, showing robust convergent, discriminant, and predictive validity in relation to crucial outcome variables such as subordinate job satisfaction, affective learning, organizational citizenship behaviors, cognitive engagement, and compliance. This comprehensive psychometric profile establishes the instrument as the gold standard assessment tool for measuring interpersonal perceptions of authority figures, supervisors, instructors, and public communicators.

2. Keywords

Source Credibility Measures, James C. McCroskey, Jason J. Teven, Ethos, Competence, Trustworthiness, Goodwill, Caring, Semantic Differential, Organizational Communication, Instructional Communication, Psychometrics

3. Authors

The Source Credibility Measures were developed and refined through a sustained program of psychometric research led by two preeminent communication scholars:

• James C. McCroskey, Ed.D. (1938–2012): Formerly Professor Emeritus and Chair of the Department of Communication Studies at West Virginia University. A foundational figure in communication education, instructional communication, and organizational communication, Dr. McCroskey authored over 30 books and more than 200 journal articles. His decades-long empirical refinement of ethos began with his landmark 1966 scales for the measurement of ethos in Speech Monographs, culminating in the 1999 tripartite reexamination.
• Jason J. Teven, Ed.D.: Professor of Human Communication Studies at California State University, Fullerton. Dr. Teven’s research program centers on instructional and supervisor-subordinate communication processes, with special focus on nonverbal immediacy, perceived caring, instructor socio-communicative style, and source credibility dynamics in workplace and classroom contexts.

4. Purpose

The primary purpose of the Source Credibility Measures is to quantitatively assess the subjective perceptions that receivers form regarding the ethos, believability, and character of a communication source. Historically, the evaluation of source credibility was confounded by divergent operationalizations that either reduced the construct to cognitive capability and honesty alone or conflated it with dynamic extroversion and sociability. McCroskey and Teven designed this 18-item semantic differential instrument to provide an empirically pure, theoretically grounded, and structurally invariant measurement tool that captures the complete classical formulation of ethos without capturing orthogonal personality traits.

In applied organizational environments, the instrument is widely deployed to evaluate managerial leadership effectiveness, supervisory communication quality, and upward employee perceptions. Modern leadership theories emphasize that managerial influence is fundamentally communicative: subordinates do not evaluate directives in a social vacuum; rather, they filter institutional communications through their subjective perceptions of their manager’s competence, character, and genuine benevolence. When applied in corporate audits or diagnostic leadership training, the scale identifies specific deficiencies in relational communication, revealing whether an executive is perceived as technically capable yet interpersonally callous, or morally upright yet functionally inept.

In instructional and academic settings, the measure functions as a primary evaluative metric for classroom instructors, teaching assistants, and institutional administrators. Instructional communication research has demonstrated that student evaluations of teacher credibility mediate the relationship between instructional behaviors (such as nonverbal immediacy, humor, and clarity) and vital educational outcomes, including affective learning, state motivation, student empowerment, and cognitive retention. By providing subscale-specific scores, the scale assists educators in isolating whether their classroom challenges stem from perceived lack of mastery over instructional content, perceived partiality or unfair grading systems, or an apparent emotional detachment from student academic welfare.

Furthermore, the instrument serves critical functions in clinical communication research, health communication, political polling, marketing, and media studies. In provider-patient relationships, high provider credibility directly correlates with patient adherence to medical regimens and treatment satisfaction. In mediated and digital communication contexts, researchers utilize the scale to evaluate user trust in algorithmic information agents, corporate crisis spokespersons, journalistic outlets, and public health officials during epidemiological crises. The theoretical rationale rests on the principle that credibility is not an objective, static property possessed by a speaker; instead, it is an attributional perception constructed within the mind of the receiver.

5. Psychological Construct

The theoretical construct operationalized by McCroskey and Teven (1999) conceptualizes Source Credibility as a receiver-based perceptual attribution composed of three distinct yet mutually reinforcing psychological dimensions: Competence, Trustworthiness, and Goodwill/Caring. Each dimension captures an indispensable facet of human social evaluation:

Competence (Perceived Expertise and Knowledge)

The Competence dimension reflects the degree to which a receiver perceives a communicator to possess superior knowledge, procedural skill, intellectual capacity, professional training, and domain expertise. Within social exchange and authority structures, competence provides the rational basis for deference and intellectual reliance. If an individual lacks perceived expertise, their directives are met with skepticism, second-guessing, and task resistance. In McCroskey and Teven’s 18-item instrument, Competence is operationalized through six bipolar adjective pairings: Intelligent–Unintelligent, Untrained–Trained, Inexpert–Expert, Informed–Uninformed, Incompetent–Competent, and Bright–Stupid. An individual scoring exceptionally high on this subscale is perceived as an authoritative, technically proficient master of their discipline whose judgment and situational command can be relied upon under conditions of uncertainty.

Trustworthiness (Perceived Character and Integrity)

The Trustworthiness dimension assesses the receiver’s evaluation of the communicator’s moral character, personal integrity, honesty, and ethical uprightness. While competence addresses whether a source can speak the truth or provide correct guidance, trustworthiness assesses whether the source will be candid and principled. In social-cognitive processing, receivers continuously monitor authority figures for cues of deception, manipulation, or ideological hypocrisy. McCroskey and Teven operationalized this dimension via six evaluative pairs: Honest–Dishonest, Untrustworthy–Trustworthy, Honorable–Dishonorable, Moral–Immoral, Unethical–Ethical, and Phony–Genuine. Communicators who register low on trustworthiness evoke psychological reactance and defensive listening strategies, regardless of how intellectually accomplished or technically capable they may be.

Goodwill / Caring (Perceived Intent toward the Receiver)

The Goodwill dimension—often designated interchangeably as Perceived Caring—represents the most profound theoretical contribution of McCroskey and Teven’s 1999 psychometric reexamination. Historical factor-analytic investigations between the 1960s and 1980s had frequently omitted or obscured this dimension, mistakenly subsuming it under broad constructs such as ‘sociability’ or collapsing it into trustworthiness. McCroskey and Teven demonstrated that goodwill is psychologically distinct: a receiver may acknowledge that a source is brilliant (competent) and impeccably honest (trustworthy), yet conclude that the source remains indifferent, hostile, or fundamentally unconcerned regarding the receiver’s personal welfare. Goodwill assesses perceived empathy, responsiveness, and altruistic intent. It is measured through six items: Cares about me–Doesn’t care about me, Has my interests at heart–Doesn’t have my interests at heart, Self-centered–Not self-centered, Concerned with me–Not concerned with me, Insensitive–Sensitive, and Not understanding–Understanding. Empirical findings indicate that perceived goodwill is the single most powerful predictor of interpersonal affection, relational satisfaction, and perceived communicative warmth in hierarchical relationships.

6. Theoretical Framework

The theoretical lineage of the Source Credibility Measures originates in classical Aristotelian rhetoric. In his seminal work Rhetoric (circa 350 BCE), Aristotle argued that persuasive discourse relies on three primary artistic modes of proof: logos (logical argument and rational evidence), pathos (emotional resonance with the audience), and ethos (the perceived personal character of the speaker). Critically, Aristotle posited that ethos is composed of three constituent elements:

• Phronesis: Practical wisdom, intelligence, competence, and broad domain knowledge.
• Arete: Moral virtue, noble character, integrity, and ethical rectitude.
• Eunoia: Benevolence, genuine goodwill, empathy, and active concern toward the audience.

When the modern social-scientific discipline of communication emerged in the mid-twentieth century, early empirical researchers—most notably Carl Hovland and his colleagues at the Yale Communication and Attitude Change Program (Hovland, Janis, & Kelley, 1953)—operationalized source credibility primarily through a two-dimensional framework consisting of ‘Expertise’ (corresponding to phronesis) and ‘Trustworthiness’ (corresponding to arete). In their psychometric operationalizations, Hovland and contemporary social psychologists largely neglected eunoia, assuming either that goodwill was an artifactual subset of trustworthiness or that it exerted negligible empirical influence over persuasion.

Throughout the 1960s and 1970s, James C. McCroskey developed several empirical ethos instruments, culminating in the landmark 1981 critique by McCroskey and Young published in the Central States Speech Journal. McCroskey and Young reviewed three decades of credibility research and warned that psychometric scales had proliferated into theoretical anarchy: researchers were dumping dozens of personality-trait scales into atheoretical exploratory factor analyses, generating artificial factors such as ‘dynamism’, ‘extroversion’, ‘sociability’, ‘composure’, and ‘size’. McCroskey and Young proved that these extra dimensions were orthogonal personality descriptors rather than core components of receiver-perceived source ethos. They urged scholars to return strictly to Aristotelian foundations.

In 1999, McCroskey and Teven executed their definitive reexamination to resolve the longstanding absence of eunoia. Drawing upon cognitive appraisal theory, social attribution models, and relational schema theories, they hypothesized that communicators who fail to communicate relational empathy cannot achieve optimal credibility. By developing psychometrically robust items specifically crafted to reflect eunoia (operationalized as perceived caring/goodwill), McCroskey and Teven proved through exploratory and confirmatory modeling that receiver cognitive appraisals cleanly bifurcate into three distinct perceptual channels: ‘What does this person know?’ (Competence), ‘Is this person ethical and honest?’ (Trustworthiness), and ‘Does this person care about my well-being or are they purely self-interested?’ (Goodwill). This tripartite theoretical framework successfully integrated classical philosophical rhetoric with rigorous twentieth-century psychometric construct validation.

7. Validity

The Source Credibility Measures have undergone extensive empirical validation across varied populations, professional sectors, and communication modalities, demonstrating superior construct, convergent, discriminant, and predictive validity.

Construct and Factorial Validity

In their initial validation study comprising 792 participants evaluating both liked and disliked targets across diverse social, instructional, and managerial roles, McCroskey and Teven (1999) established strong construct validity. Confirmatory factor analysis verified that the hypothesized three-factor oblique model fit empirical data significantly better than alternative rival models, including a unidimensional general credibility model and a two-factor (competence and character) model. All 18 items exhibited robust standardized factor loadings (ranging from .68 to .89) on their specified theoretical latent dimensions without problematic cross-loadings above .30.

Convergent Validity

Convergent validity has been repeatedly established through robust positive correlations with external measures of constructive communicative behavior. The Competence subscale correlates strongly with measures of cognitive clarity, communicative effectiveness, and task performance evaluations ($r = .55$ to $.72, p < .001$). The Trustworthiness subscale displays high convergent correlations with interpersonal trust inventories, such as the Workplace Trust Scale ($r = .65$ to $.80, p < .001$). The Goodwill/Caring subscale demonstrates exceptionally strong convergent validity with measures of supervisor socio-communicative style, perceived nonverbal immediacy, and instructional responsiveness ($r = .60$ to $.78, p < .001$).

Discriminant Validity

Discriminant validity is supported by both statistical and theoretical benchmarks. Average Variance Extracted (AVE) analyses conducted across multiple structural equation modeling studies confirm that the AVE for each latent factor exceeds the squared inter-factor correlations ($\phi^2$), fulfilling the rigorous Fornell-Larcker criterion. Although the three dimensions are intercorrelated—typically exhibiting moderate to strong positive correlations ($r = .40$ to $.65$)—they do not share sufficient variance to justify collapsing into a single construct. Furthermore, the scale demonstrates clear discriminant divergence from constructs such as source attractiveness, physical appearance, and communicative extroversion, confirming that it measures ethos rather than superficial personality or aesthetic appeal.

Predictive and Criterion-Related Validity

Criterion-related validity is evidenced by the scale’s profound predictive power across organizational and pedagogical outcomes:

• In instructional settings, Teven and McCroskey (1997, 1999) demonstrated that perceived instructor caring accounts for substantial unique variance in students’ cognitive learning, affective attitudes toward course material, and willingness to recommend instructors, over and above instructor competence alone.
• In workplace leadership contexts, managerial credibility measured via this scale accounts for over 45% of the variance in subordinate job satisfaction, 30% of the variance in organizational citizenship behaviors (OCBs), and significantly reduces employee turnover intentions.
• In crisis communication investigations, post-crisis organizational trust is overwhelmingly predicted by corporate spokespersons’ scores on the Trustworthiness and Goodwill subscales.

8. Reliability

The psychometric reliability of the Source Credibility Measures is exceptionally high and remarkably stable across cultural groups, age cohorts, and administrative contexts. Both internal consistency and temporal stability metrics satisfy the most stringent psychometric standards for applied diagnostic and academic research applications.

Internal Consistency

In McCroskey and Teven’s (1999) foundational investigation, the instrument demonstrated outstanding internal consistency across all three subscales:

• Competence: Cronbach’s $\alpha = .89$ to $.92$
• Trustworthiness: Cronbach’s $\alpha = .91$ to $.94$
• Goodwill/Caring: Cronbach’s $\alpha = .89$ to $.92$

Subsequent investigations across diverse contexts have continuously replicated these high reliability estimates. In instructional contexts, studies routinely report alphas between .88 and .95 for all three dimensions. In corporate organizational surveys evaluating senior leadership, alpha values consistently range from .87 to .93. Composite reliability ($
ho_c$) and McDonald’s \omega ($omega$) coefficients calculated in modern structural equation modeling studies mirror these values, consistently registering above .90, thereby confirming that the high internal consistency is not an artifact of tau-equivalence violations.

Test-Retest Stability

When evaluated in test-retest reliability designs with stable communication targets (e.g., assessing tenured professors or long-term supervisors over a two- to four-week period without intervening communicative shocks), stability coefficients range from $r_{tt} = .78$ to $.86$. Because source credibility is inherently a dynamic state attribution sensitive to immediate interactional changes, slight temporal shifts are theoretically expected following negative disclosures, managerial policy reversals, or communication breakdowns, confirming the scale’s sensitivity to situational communicative phenomena.

9. Factor Analysis

The structural morphology of the 18-item Source Credibility Measures has been rigorously analyzed via both Exploratory Factor Analysis (EFA) and Confirmatory Factor Analysis (CFA).

Exploratory Factor Structure

In the scale’s derivation studies, Principal Components Analysis (PCA) and Principal Axis Factoring (PAF) utilizing oblique rotations (Promax and Direct Oblimin) consistently extracted three primary eigenvalues exceeding 1.0 (Kaiser-Guttman criterion), together accounting for approximately 65% to 73% of the total variance across the 18 items. Scree plot analyses consistently revealed an unambiguous break after the third factor. The six items designated for Competence load cleanly onto Factor 1, the six items for Trustworthiness load onto Factor 2, and the six items for Goodwill/Caring load onto Factor 3. In cleanly executed datasets, primary factor pattern coefficients range from .65 to .88, with secondary cross-loadings uniformly suppressed below .25.

Confirmatory Factor Analysis and Model Fit

Contemporary psychometric evaluations employing CFA with maximum likelihood estimation (often using robust standard errors, MLM/MLR) confirm the superior fit of the hypothesized three-factor oblique structure over alternative nested models. Standard goodness-of-fit benchmarks demonstrate strong alignment:

• Model Chi-Square / Degrees of Freedom: $\chi^2/df < 2.50$
• Comparative Fit Index (CFI): $ge .95$
• Tucker-Lewis Index (TLI): $ge .94$
• Root Mean Square Error of Approximation (RMSEA): $le .055$ (with 90% confidence interval ranging from .042 to .066)
• Standardized Root Mean Square Residual (SRMR): $le .045$

When researchers model a single-factor unidimensional solution, model fit deteriorates dramatically (CFI drops below .70, RMSEA spikes above .14), demonstrating that source credibility cannot be measured validly as an undifferentiated global score without sacrificing conceptual and statistical precision. The three-factor solution remains invariant across gender, organizational hierarchy, and cultural settings.

10. Instrument / Measurement Tool

The operational specifications of the Source Credibility Measures are structured as follows:

• Instrument Name: Source Credibility Measures (also cited as the McCroskey-Teven Source Credibility Scale or Ethos Scale).
• Authors: James C. McCroskey and Jason J. Teven (1999).
• Assessment Type: Receiver-report psychometric rating instrument utilizing a 7-point semantic differential format.
• Target Population: Adults, working professionals, and students assessing any specified communicative source (e.g., manager, executive, teacher, political candidate, healthcare professional).
• Item Count: 18 bipolar adjective pairs (6 items per dimension).
• Response Scale: 7-point semantic differential continuum (1 to 7) anchored between opposing adjectives. Instructions specify:
  • Numbers 1 and 7 indicate a very strong feeling.
  • Numbers 2 and 6 indicate a strong feeling.
  • Numbers 3 and 5 indicate a fairly weak feeling.
  • Number 4 indicates you are undecided.
• Subscale Item Distribution:
  • Competence Subscale: Items 1, 2, 7, 11, 13, 16
  • Trustworthiness Subscale: Items 4, 6, 10, 12, 14, 17
  • Goodwill/Caring Subscale: Items 3, 5, 8, 9, 15, 18
• Scoring Rules and Transformation:
  • Items marked with an asterisk (*) in the scoring key are reverse-coded to ensure that higher numeric scores uniformly reflect higher perceived credibility. Specifically, reverse-code items: 1, 3, 4, 5, 9, 10, 11, 12, and 16 using the transformation formula:
    
    Recoded Score = 8 - Original Raw Score (i.e., $1=7, 2=6, 3=5, 4=4, 5=3, 6=2, 7=1$).
  • Sum the six recoded/raw items belonging to each subscale to obtain three separate dimensional scores ranging from 6 to 42 each.
  • Alternatively, calculate the dimensional mean score (range 1.0 to 7.0) by dividing the dimensional sum by 6 for direct semantic interpretation.
  • Scores from 6 to 18 (means 1.0–3.0) represent low credibility; scores from 19 to 29 (means 3.1–4.9) represent moderate/ambivalent credibility; scores from 30 to 42 (means 5.0–7.0) represent high perceived credibility.

11. Permissions & Fee and Test Year

The definitive tripartite version of the Source Credibility Measures was published in 1999 by James C. McCroskey and Jason J. Teven in Communication Monographs. Prior developmental antecedents were published in 1966, 1969, 1974, and 1981.

In alignment with Dr. James C. McCroskey’s lifetime dedication to open academic inquiry and communication scholarship, the Source Credibility Measures instrument is placed in the public domain for scholarly research and educational purposes. No licensing fees or royalty payments are required for academic scholars, graduate researchers, or classroom teachers utilizing the scale in non-commercial investigations. However, formal bibliographic citation of McCroskey and Teven (1999) is mandatory in all research publications, theses, and technical reports. Organizations or commercial consulting firms utilizing the scale for commercial diagnostics, fee-based leadership training, or proprietary assessment platforms should review institutional copyright policies and provide proper attribution.

12. References

The following foundational publications document the theoretical evolution, psychometric validation, and empirical application of the Source Credibility Measures:

• Aristotle. (350 BCE). Rhetoric (W. Rhys Roberts, Trans.). Modern Library.
  https://en.wikipedia.org/wiki/Rhetoric_(Aristotle)
• Hovland, C. I., Janis, I. L., & Kelley, H. H. (1953). Communication and persuasion: Psychological studies of opinion change. Yale University Press.
• McCroskey, J. C. (1966). Scales for the measurement of ethos. Speech Monographs, 33(1), 65–72.
  https://doi.org/10.1080/03637756609375482
• McCroskey, J. C., & Mehrley, R. S. (1969). The effects of disorganization and nonfluency on attitude change and source credibility. Speech Monographs, 36(1), 13–21.
  https://doi.org/10.1080/03637756909375604
• McCroskey, J. C., Holdridge, W., & Toomb, J. K. (1974). An instrument for measuring the source credibility of basic speech communication instructors. The Speech Teacher, 23(1), 26–33.
  https://doi.org/10.1080/03634527409378053
• McCroskey, J. C., & Teven, J. J. (1999). Goodwill: A reexamination of the construct and its measurement. Communication Monographs, 66(1), 90–103.
  https://doi.org/10.1080/03637759909376464
• McCroskey, J. C., & Young, T. J. (1981). Ethos and credibility: The construct and its measurement after three decades. Central States Speech Journal, 32(1), 24–34.
  https://doi.org/10.1080/10510978109368075
• Teven, J. J., & McCroskey, J. C. (1997). The development of a relationship between perceived teacher caring and student learning in higher education. Communication Education, 46(1), 1–9.
  https://doi.org/10.1080/03634529709379069

13. Items of the Scale (Questionnaire)