Expertise (Source) (EXPSRC)

1. Abstract

The Expertise (Source) (EXPSRC) scale is an eight-item semantic differential psychometric instrument designed to assess audience perceptions of an information source’s knowledge, competence, training, intellect, and practical experience. Developed originally within the psychometric lineage of James C. McCroskey’s pioneering 1966 investigation into the empirical dimensions of rhetorical ethos and source credibility, this measure captures the epistemic authority attributed to communicators, including human spokespersons, institutional actors, academic experts, organizational entities, and mediated communicators. Structured as an eight-item, seven-point bipolar semantic differential scale, the instrument operationalizes perceived competence through established adjectival polar opposites: Reliable – Unreliable, Informed – Uninformed, Qualified – Unqualified, Intelligent – Unintelligent, Valuable – Worthless, Expert – Inexpert, Experienced – Inexperienced, and Competent – Incompetent. Across dozens of empirical replications in speech communication, marketing research, health communication, and social psychology, the scale demonstrates exceptional psychometric properties, consistently exhibiting high internal consistency reliability (Cronbach’s alpha typically ranging between .85 and .94) and robust unidimensionality under exploratory and confirmatory factor analytic evaluations. Convergent and predictive validity studies establish that high scores on the EXPSRC significantly predict message acceptance, attitude change, source trustworthiness ratings, and behavioral compliance within dual-process frameworks such as the Elaboration Likelihood Model (ELM) and the Heuristic-Systematic Model (HSM). This article provides an exhaustive review of the scale’s historical foundations, theoretical and cognitive mechanisms, psychometric architecture, structural validity, and empirical administration rules.

2. Keywords

source expertise, source credibility, ethos, James C. McCroskey, semantic differential, communication competence, persuasion, elaboration likelihood model, epistemic trust, message endorsement

3. Authors

The foundational psychometric lineage of the scale stems from the empirical research of James C. McCroskey, Ed.D. (1934–2012), one of the most prolific scholars in the history of the communication discipline. At the time of the publication of his seminal monograph, “Scales for the measurement of ethos” (1966), McCroskey was affiliated with the Department of Speech at the Pennsylvania State University, subsequently serving as Professor and Chair of the Department of Communication Studies at West Virginia University and later at the University of Alabama at Birmingham. McCroskey was recognized as a National Communication Association (NCA) Distinguished Scholar and authored over 200 journal articles, focusing comprehensively on communication apprehension, rhetorical ethos, nonverbal immediacy, and instructional communication.

4. Purpose

The primary objective of the Expertise (Source) (EXPSRC) scale is to furnish an efficient, reliable, and construct-valid standardized instrument capable of quantifying the perceived intellectual and technical capability of an information source. In psychological and communication investigations, the efficacy of persuasive discourse, instructional messaging, risk communication, and marketing appeals is heavily governed by receiver assessments of the communicator’s background and skill. While historical formulations of persuasion dating back to classical antiquity recognized the paramount importance of the speaker’s personal character, experimental psychometrics requires precise instrumentation to disentangle the communicator’s cognitive mastery from other affective dimensions of credibility, such as perceived warmth, moral character, benevolence, or dynamism.

In applied research contexts, the EXPSRC scale is deployed across diverse experimental and naturalistic environments:

• Health and Risk Communication: Quantifying patient and public trust in public health spokespersons, medical practitioners, scientific advisory panels, and algorithmic diagnostic engines during epidemiological crises or vaccination campaigns.
• Persuasion and Social Influence Experiments: Serving as either a critical manipulation check or an independent predictor in laboratory studies testing dual-process cognitive models (Petty & Cacioppo, 1986). Researchers assess whether high-expertise cues bias argument elaboration or function as peripheral heuristic shortcuts under low personal involvement.
• Consumer Psychology and Advertising: Evaluating corporate credibility, celebrity endorser fit, brand-influencer alignment, and the epistemic authority of organizational spokespersons in corporate social responsibility (CSR) initiatives and product recalls.
• Political Communication and Civic Engagement: Measuring voter perceptions of candidates’ policy preparedness, intellectual capacity, legislative grasp, and administrative competence relative to debate outcomes and policy adoption.
• Human-Computer Interaction (HCI) and Artificial Intelligence: Examining user acceptance of automated recommendations, conversational agents, large language models (LLMs), and algorithmic decision-support systems, specifically analyzing how technical competence vs. perceived warmth influences system adoption.

From a diagnostic and clinical perspective, variants of the scale have been employed in psychotherapy and counseling psychology to assess therapeutic alliance, specifically investigating how a client’s perception of the therapist’s clinical competence and seasoned experience interacts with treatment adherence, drop-out rates, and symptomatic recovery.

5. Psychological Construct

Perceived source expertise constitutes a cognitive evaluation formed by an audience member or message recipient concerning the extent to which a communicator possesses valid information, formal training, intellectual skill, and authoritative insight regarding a designated domain of knowledge. Psychometrically, source expertise is differentiated from objective expertise (such as accredited degrees, objective test scores, or formal credentials). While objective expertise resides in the communicator’s actual training and history, perceived expertise exists strictly as a psychological attribution constructed in the mind of the receiver. As such, an objective novice can be perceived as highly expert if communicative cues signal competence, whereas a decorated scientist may be perceived as inexpert if their communicative behavior signals disorientation or incompetence.

The eight semantic differential pairs within the EXPSRC instrument comprehensively capture the core facets of this psychological construct:

• Formal Competence and Intellectual Capacity: Reflected by the pairs Intelligent – Unintelligent and Competent – Incompetent. These items capture the general cognitive endowment and baseline functional capacity attributed to the communicator. They measure whether the source has the cognitive infrastructure required to analyze complex domain-specific phenomena accurately.
• Epistemic Familiarity and Factual Preparation: Operationalized through Informed – Uninformed. This dimension targets the currency, depth, and informational accuracy of the communicator’s knowledge base concerning the immediate discourse context.
• Credentialing and Domain Authority: Captured by Qualified – Unqualified and Expert – Inexpert. These items evaluate institutional or social legitimacy, assessing whether the communicator meets the normative criteria established for subject-matter authority in their specified arena.
• Practical Tenure and Applied Practice: Indexed by Experienced – Inexperienced. Recognizing that intellectual capacity alone is insufficient for authoritative credibility, this facet captures perceived pragmatic application, situational familiarity, and experiential seasoning.
• Epistemic Utility and Soundness: Manifested in Reliable – Unreliable and Valuable – Worthless. These items reflect the functional dependability of the information furnished. In epistemic cognition, an expert source is judged by the pragmatic utility and predictive stability of their insights, distinguishing substantive competence from superficial articulateness.

Together, these eight indicators form a coherent psychological gestalt. When a receiver confronts an information source, fast-acting cognitive heuristics synthesize perceptual inputs—such as vocabulary density, nonverbal fluency, institutional affiliations, and logical coherence—into a unified semantic evaluation of perceived expertise.

6. Theoretical Framework

The theoretical architecture of the Expertise (Source) scale is situated at the intersection of classical rhetorical theory, mid-century social psychological attitudes research, and modern cognitive psychology. The conceptual lineage traces back directly to Aristotle’s treatise Rhetoric, wherein persuasion achieved through the personal character of the speaker is termed ethos (Book I, Chapter 2). Aristotle posited that ethos is partitioned into three indispensable components: phronesis (practical wisdom, intelligence, competence), arete (moral virtue, goodness), and eunoia (goodwill toward the audience). The EXPSRC instrument serves as the direct operationalization of phronesis.

During the mid-twentieth century, this rhetorical taxonomy was integrated into behavioral experimental psychology by Carl I. Hovland and his colleagues within the Yale Communication and Attitude Change Program (Hovland, Janis, & Kelley, 1953). In their landmark volume Communication and Persuasion, Hovland and associates identified source credibility as an influential factor in inducing attitude modification, bifurcating it into two primary dimensions: expertness (the communicator’s perceived ability to make valid assertions) and trustworthiness (the communicator’s perceived motivation to communicate valid assertions without bias). Hovland’s group demonstrated that attributing identical persuasive texts to high-expertise sources (e.g., the New England Journal of Medicine or J. Robert Oppenheimer) produced significantly greater immediate opinion shift than when attributed to low-expertise sources (e.g., the Soviet newspaper Pravda), although the famous sleeper effect revealed that source dissociations occur over temporal delays.

In the 1960s, communication scholars sought to transition from ad-hoc operationalizations of credibility to rigorous psychometric instruments. Utilizing Charles E. Osgood, George J. Suci, and Percy H. Tannenbaum’s (1957) semantic differential methodology, researchers began factor-analyzing expansive batteries of bipolar adjectives. While David K. Berlo, James B. Lemert, and Robert J. Mertz (1969) proposed dimensions of “Safety,” “Qualification,” and “Dynamism,” James C. McCroskey (1966) rigorously established that ethos consistently bifurcated into two primary, highly stable orthogonal factors across diverse communication settings: Authoritativeness (subsequently refined as Competence/Expertise) and Character (subsequently designated as Trustworthiness). The eight items comprising the EXPSRC scale trace their empirical origin to this authoritativeness cluster.

In contemporary cognitive psychology, the scale is interpreted through Petty and Cacioppo’s Elaboration Likelihood Model (ELM) and Shelly Chaiken’s Heuristic-Systematic Model (HSM). Under conditions of low cognitive elaboration—characterized by low personal relevance, cognitive fatigue, or high distraction—perceived source expertise functions as a peripheral heuristic cue (e.g., “Experts can be trusted”), enabling rapid decision-making without detailed argument inspection. Conversely, under high elaboration conditions, high source expertise can bias systematic processing or serve as a persuasive argument in itself if the source’s competence is directly relevant to the issue at hand.

7. Validity

The construct, convergent, predictive, and discriminant validity of the Expertise (Source) scale has been established through extensive psychometric testing across six decades of research.

Construct and Factorial Validity

Construct validity is evidenced by the consistent emergence of an independent competence/expertise factor across heterogeneous populations and varied communicative stimuli. Factorial analyses repeatedly show that the eight items load heavily onto a single latent factor, explaining between 55% and 72% of the common variance among the items, with primary factor loadings routinely exceeding .70. Cross-cultural replications have affirmed that the semantic pairings retain their conceptual coherence across linguistic adaptations, demonstrating strong scalar and metric measurement invariance.

Convergent Validity

Convergent validity has been evaluated by correlating the EXPSRC with alternative scales of communicator credibility, interpersonal attraction, and epistemic authority:

• Berlo’s Qualification Dimension: Correlations between the EXPSRC and the Qualification subscale of Berlo, Lemert, and Mertz (1969) consistently fall between r = .78 and r = .88 (p < .001).
• McCroskey & Teven’s (1999) Revised Source Credibility Scale: Correlations between the EXPSRC and the modernized 6-item Competence subscale approach collinearity (r = .89 to .94), confirming that both measures tap the identical epistemic construct.
• Epistemic Authority Scales: The scale correlates moderately to strongly (r = .62 to .74) with Kruglanski et al.’s (2005) Epistemic Authority Scale, reflecting perceived domain mastery.

Discriminant Validity

Discriminant validity is supported by the measure’s ability to decouple cognitive expertise from affective or moral judgments. In factor analyses incorporating items measuring source character, perceived goodwill (caring), and physical attractiveness, the eight expertise items consistently dissociate from non-epistemic items. While correlations between perceived expertise and perceived trustworthiness are moderately positive (typically ranging between r = .35 and .55, reflecting general halo effects), average variance extracted (AVE) estimates routinely exceed shared squared correlations, meeting the rigorous Fornell-Larcker criterion for discriminant validity.

Predictive and Criterion Validity

Extensive experimental literature supports the criterion validity of the EXPSRC:

• In persuasive messaging experiments, manipulating objective speaker qualifications (e.g., Nobel laureate vs. high school student) yields statistically robust, large-magnitude differences on the EXPSRC scale (typically Cohen’s d > 1.40).
• In consumer research, higher scores on the EXPSRC reliably predict post-exposure attitude toward the brand (β = .42, p < .001) and purchase intentions (β = .36, p < .01).
• In instructional communication, student ratings of instructor expertise using this scale significantly predict student affective learning (r = .48) and cognitive learning outcomes (r = .38).

8. Reliability

The psychometric reliability of the Expertise (Source) scale has been documented in hundreds of experimental investigations, consistently demonstrating internal consistency coefficients that surpass the standard thresholds recommended for psychometric research (Nunnally & Bernstein, 1994).

Internal Consistency Reliability

Across empirical studies involving varied targets (political candidates, corporate entities, scientific spokespersons, healthcare workers, peer communicators):

• Cronbach’s Alpha (α): The 8-item scale consistently yields Cronbach’s alpha coefficients between .86 and .94. In McCroskey’s foundational studies (1966), authoritativeness scales demonstrated split-half and alpha coefficients routinely exceeding .90.
• McDonald’s Omega (ω): Contemporary psychometric evaluations utilizing composite reliability and McDonald’s hierarchical omega (ω_h) confirm values ranging from .88 to .95, indicating that common factor variance dominates total scale score variance, with minimal item-specific error variance.
• Item-Total Correlations: Corrected item-to-total correlations for all eight bipolar pairings consistently exceed .60, with central items such as Expert – Inexpert and Competent – Incompetent frequently exhibiting item-total correlations greater than .75.

Test-Retest Reliability and Temporal Stability

Because perceived expertise is designed to reflect an evaluative attitude toward a specific target, temporal stability is dependent on the stability of target-related information. In test-retest experimental paradigms where no new information regarding the source is introduced across an intervening 14-day interval, test-retest correlation coefficients range between r_tt = .78 and .85, confirming strong temporal stability. In the presence of new contradictory evidence (such as source performance blunders or credential fraud exposure), the instrument demonstrates appropriate sensitivity to state-level cognitive re-evaluations.

9. Factor Analysis

The structural dimensionality of the items comprising the EXPSRC scale has been evaluated using both exploratory factor analysis (EFA) and confirmatory factor analysis (CFA).

Exploratory Factor Analysis (EFA)

In exploratory factor analytic studies utilizing principal axis factoring or maximum likelihood extraction with oblique rotations (e.g., Promax or Direct Oblimin), the eight items consistently converge on a dominant single factor. Characteristic psychometric findings include:

• Eigenvalues: A single primary eigenvalue typically exceeding 4.50, accounting for 56% to 70% of the total variance, followed by a precipitous drop to the second eigenvalue (consistently < 0.70), clearly establishing an unambiguous scree plot elbow.
• Factor Loadings: Standardized factor loadings across diverse empirical datasets demonstrate high pattern matrix coefficients:
  • Competent – Incompetent: .82 to .89
  • Expert – Inexpert: .80 to .88
  • Qualified – Unqualified: .78 to .86
  • Informed – Uninformed: .74 to .82
  • Experienced – Inexperienced: .71 to .81
  • Intelligent – Unintelligent: .68 to .78
  • Reliable – Unreliable: .65 to .75
  • Valuable – Worthless: .62 to .74

Confirmatory Factor Analysis (CFA)

Confirmatory factor analytic investigations evaluating a unidimensional measurement model have confirmed adequate to excellent goodness-of-fit indices across large sample cohorts (N > 500). Representative model fit statistics reported in contemporary literature include:

• Chi-Square / Degrees of Freedom Ratio: χ²/df typically between 1.85 and 2.90 (indicating acceptable fit below the standard threshold of 3.0).
• Comparative Fit Index (CFI): .96 to .99 (exceeding the .95 benchmark for superior fit).
• Tucker-Lewis Index (TLI): .95 to .98.
• Root Mean Square Error of Approximation (RMSEA): .042 to .061 (90% CI [.030, .072]), with a non-significant probability of close fit (p_close > .50).
• Standardized Root Mean Square Residual (SRMR): .025 to .038 (well below the .08 maximum threshold).

When evaluated within multidimensional credibility structures (e.g., alongside Trustworthiness/Character and Goodwill/Caring dimensions), multi-factor CFA models demonstrate distinct latent constructs with inter-factor correlations between .40 and .60, outperforming alternative one-factor or collapsed models across Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC) model comparison metrics.

10. Instrument / Measurement Tool

• Instrument Name: Expertise (Source) Scale (EXPSRC)
• Alternative Designations: McCroskey Authoritativeness Scale (Competence Subscale); Source Expertise Dimension
• Primary Author: James C. McCroskey (1966)
• Target Population: General adolescent and adult populations capable of language-based semantic comprehension (suitable for laboratory, field, and survey research)
• Administration Format: Paper-and-pencil self-report or computer-assisted web interview (CAWI)
• Completion Time: Approximately 1 to 2 minutes
• Number of Items: 8 bipolar adjective pairs
• Response Scale: 7-point semantic differential scale (ranging from 1 to 7 between polar opposite adjectives)
• Scoring Procedure:
  • Response values are assigned integers from 1 to 7 across the continuum between semantic anchors.
  • Items oriented with the positive/competent anchor on the left and the negative anchor on the right are scored directly: 7 = Most Expert, 1 = Least Expert.
  • Items presenting negative anchors on the left are reverse-scored so that higher scores consistently reflect greater perceived source expertise.
  • A composite score is derived by computing either the arithmetic sum (ranging from 8 to 56) or the mean score (ranging from 1.00 to 7.00) across the eight items. Higher aggregate scores denote greater perceived source expertise.

11. Permissions & Fee and Test Year

The conceptual framework and foundational item battery for the scale were first published in 1966 by James C. McCroskey in Speech Monographs (now Communication Monographs), published by the Speech Association of America (now the National Communication Association). Consistent with Dr. McCroskey’s career-long commitment to open educational and academic research access, his psychometric scales were widely placed in the public domain for non-commercial educational and scholarly use without royalty requirements. Researchers may administer the instrument in non-profit academic studies, dissertations, theses, and instructional settings without formal written permissions, provided appropriate academic citation is maintained. For commercial deployments, proprietary consulting, or integration into fee-based software platforms, licensing guidance should be referenced through published institutional agreements or scholarly rights holders.

12. References

Aristotle. (1954). The rhetoric and the poetics of Aristotle (W. R. Roberts & I. Bywater, Trans.). Modern Library.

Berlo, D. K., Lemert, J. B., & Mertz, R. J. (1969). Dimensions of source credibility. Speech Monographs, 36(4), 463–476. https://doi.org/10.1080/03637756909375635

Chaiken, S. (1980). Heuristic versus systematic information processing and the use of source versus content cues in persuasion. Journal of Personality and Social Psychology, 39(5), 752–766. https://doi.org/10.1037/0022-3514.39.5.752

Hovland, C. I., Janis, I. L., & Kelley, H. H. (1953). Communication and persuasion: Psychological studies of opinion change. Yale University Press.

Kruglanski, A. W., Raviv, A., Bar-Tal, D., Raviv, A., Sharvit, K., Ellis, S., Bar, R., Pierro, A., & Mannetti, L. (2005). Says who? Epistemic authority and its effects on knowledge formation. Advances in Experimental Social Psychology, 37, 345–392. https://doi.org/10.1016/S0065-2601(05)37006-7

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., & 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

Nunnally, J. C., & Bernstein, I. H. (1994). Psychometric theory (3rd ed.). McGraw-Hill.

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

Pornpitakpan, C. (2004). The persuasiveness of source credibility: A critical review of five decades’ evidence. Journal of Applied Social Psychology, 34(2), 243–281. https://doi.org/10.1111/j.1559-1816.2004.tb02547.x

13. Items of the Scale