The Similarity of Endorsers Scale<\/strong> is a concise, unidimensional psychometric instrument developed by Veronica L. Thomas<\/a> and Kendra Fowler<\/a> (2021) to quantify the degree to which an ensemble or portfolio of brand endorsers is perceived as homogeneous, mutually substitutable, and cognitively interchangeable. Originating within consumer psychology and advertising research published in the Journal of Advertising<\/em>, the scale was created to address the expanding paradigm of synthetic media, virtual influencers, and artificial intelligence (AI<\/a>) brand ambassadors relative to traditional human celebrity and peer endorsers. The instrument comprises three self-report items administered on a 7-point Likert scale<\/a> ranging from 1 (Strongly disagree<\/em>) to 7 (Strongly agree<\/em>), yielding a composite mean score where higher values reflect heightened perceived homogeneity and cognitive overlap among endorsers.<\/p>\n Psychometrically, the instrument demonstrates robust internal consistency reliability across experimental cohorts, routinely reporting Cronbach\u2019s alpha<\/a> ($\u0007lpha$<\/span>) coefficients exceeding .85 and composite reliability values well above benchmark thresholds. Factor analytic evaluations confirm a parsimonious single-factor architecture with high factor loadings ($lambda > .80$<\/span>) and strong convergent, discriminant, and predictive validity. Functionally, the construct mediates and moderates critical advertising outcomes, including brand portfolio coherence, endorser redundancy, brand dilution, consumer skepticism, and overall persuasion effectiveness. By providing an empirical metric for evaluating how audiences cognitively cluster human and non-human endorsers, the scale serves as a foundational assessment tool across advertising strategy, human-computer interaction (HCI<\/a>), and social perception research.<\/p>\n Similarity of Endorsers, Endorser Interchangeability, Perceived Homogeneity, Artificial Intelligence Influencers, Virtual Endorsers, Source Credibility, Social Categorization Theory, Advertising Psychology, Entitativity, Psychometrics<\/p>\n The scale was developed and validated by:<\/p>\n Correspondence regarding the original empirical validation study can be directed via academic channels to the Department of Marketing at Towson University or through the Journal of Advertising<\/em> editorial office.<\/p>\n The primary purpose of the Similarity of Endorsers Scale is to systematically assess consumer evaluations of uniformity versus differentiation across multiple brand representatives. In contemporary marketing campaigns, firms increasingly deploy multi-endorser strategies\u2014contracting multiple celebrities, micro-influencers, or computer-generated virtual personalities\u2014to appeal to diverse market segments. Thomas and Fowler (2021) introduced this measurement tool to investigate how consumers psychologically categorize arrays of endorsers, particularly when comparing emergent AI-generated influencers (e.g., CGI avatars such as Lil Miquela or Shudu) against conventional human spokespersons.<\/p>\n From an applied research perspective, the scale fulfills several vital diagnostic functions:<\/p>\n In theoretical and experimental contexts, the instrument enables scholars to isolate perceived endorser homogeneity as a focal independent, mediating, or moderating variable. It allows researchers to clarify how cognitive categorization mechanisms shape fundamental communication outcomes, such as source trustworthiness, perceived authenticity, attitude toward the ad ($A_{ad}$<\/span>), attitude toward the brand ($A_b$<\/span>), and behavioral purchase intentions.<\/p>\n The psychological construct captured by the scale is Perceived Endorser Similarity and Interchangeability<\/strong>. This construct reflects the cognitive degree to which an observer perceives a set of social or pseudo-social actors as sharing common morphological, behavioral, symbolic, or functional properties such that any single member can substitute for another without a perceptible loss of identity or communicative utility.<\/p>\n This construct encompasses three interrelated psychological facets:<\/p>\n In the context of artificial intelligence and synthetic media, the construct acquires heightened psychological significance. Human beings naturally evaluate other humans using rich evolutionary mechanisms dedicated to facial individuation, personal history attribution, and perceived free will. Conversely, artificial entities often evoke cognitive categorization schemas associated with manufactured products. Consequently, consumers may reflexively infer high interchangeability among AI endorsers, treating them as modular components of a software algorithm rather than unique social communicators.<\/p>\n The construct must be carefully distinguished from related yet distinct psychological phenomena:<\/p>\n The Similarity of Endorsers Scale is anchored at the intersection of several core paradigms in social cognition and consumer psychology:<\/p>\n Rooted in Social Identity Theory<\/a> and Self-Categorization Theory (Tajfel & Turner, 1986; Turner et al., 1987), humans naturally organize social stimuli into cognitive categories to reduce processing demands. A robust manifestation of this process is the Outgroup Homogeneity Effect<\/strong> (Park & Rothbart, 1982; Linville et al., 1989), which posits that individuals perceive members of an outgroup as “all alike” and interchangeable, while recognizing nuanced variance and individuality among ingroup members.<\/p>\n When Thomas and Fowler (2021) examined AI versus human influencers, this mechanism proved central. Because human consumers inherently classify AI avatars as non-human or machine outgroups, they exhibit a cognitive bias toward outgroup homogeneity. Unless virtual influencers are imbued with distinct, multidimensional narratives, consumers generalize across synthetic agents, viewing them as interchangeable algorithmic representations.<\/p>\n Donald Campbell (1958) introduced entitativity<\/a> to describe the degree to which a collection of individuals is perceived as having the nature of an entity. Key Gestalt cues driving entitativity include common fate, similarity, and proximity. In multi-endorser advertising, high endorser similarity prompts observers to perceive the spokespersons not as independent corroborating witnesses of a product\u2019s quality, but as a singular collective unit. Consequently, information provided by interchangeable endorsers is discounted because subsequent endorsers offer no independent validation (Harkins & Petty, 1981).<\/p>\n The CASA paradigm<\/a>, formulated by Clifford Nass and Byron Reeves (Nass & Reeves, 1996), demonstrates that humans mindlessly apply social rules and expectations to technological artifacts when those artifacts exhibit anthropomorphic visual or auditory cues. However, when consumers discover an agent is non-human, systematic cognitive recalibration occurs. Perceptions of agency, emotional authenticity, and moral responsibility diminish (Gray et al., 2007). In advertising, this recalibration accentuates perceived interchangeability: lacking an authentic inner life or personal biographical trajectory, AI influencers risk being reduced to decorative, fungible marketing props.<\/p>\n Empirical assessment of the Similarity of Endorsers Scale has established extensive construct, convergent, discriminant, and predictive validity across varied experimental manipulations of endorser type (human vs. AI), product category, and portfolio size.<\/p>\n Construct validity is substantiated by the scale\u2019s ability to faithfully capture the cognitive clustering of endorser characteristics. Confirmatory factor analyses (CFA<\/a>) consistently exhibit high standardized factor loadings exceeding the conventional .70 benchmark (individual loadings across validation studies: $\\lambda_1 = .84$<\/span>, $\\lambda_2 = .88$<\/span>, $\\lambda_3 = .86$<\/span>). The Average Variance Extracted<\/a> (AVE) regularly surpasses .70, comfortably exceeding the .50 cutoff proposed by Fornell and Larcker (1981) and confirming that the variance captured by the latent construct is substantially greater than variance attributable to measurement error.<\/p>\n Discriminant validity has been rigorously demonstrated against adjacent consumer evaluation metrics:<\/p>\n The instrument exhibits strong predictive validity in advertising response paradigms. Thomas and Fowler (2021) demonstrated that high endorser similarity scores systematically predicted:<\/p>\n The Similarity of Endorsers Scale demonstrates exceptionally strong reliability across diverse empirical investigations. Despite its brief 3-item length, the instrument maintains high internal consistency without exhibiting redundant item collinearity.<\/p>\n Corrected item-total correlations for each of the three items consistently exceed $.72$<\/span>, indicating that all items contribute robustly to the core latent dimension. Deletion of any individual item results in a reduction of the overall alpha coefficient, confirming that each item captures a critical facet of the similarity-interchangeability spectrum. Split-half reliability estimates ($r_{sb} > .85$<\/span>) further support the internal stability of the instrument across varied demographic samples.<\/p>\n Empirical validation of the scale involves both exploratory (EFA<\/a>) and confirmatory factor analyses (CFA<\/a>) to establish its dimensional structure and measurement properties.<\/p>\n Principal Axis Factoring and Maximum Likelihood extraction methods with oblimin and varimax rotations consistently reveal an unambiguous unidimensional solution:<\/p>\n Structural equation modeling has confirmed excellent model fit for the unidimensional structure. When testing the 3-item single-factor model across diverse experimental datasets, the measurement model demonstrates exemplary fit indices conforming to rigorous structural criteria (Hu & Bentler, 1999):<\/p>\n Multi-group CFA testing has evaluated measurement invariance across experimental cohorts evaluating human versus AI influencers. The instrument demonstrates configural invariance (identical factor structure across groups), metric invariance (equal factor loadings across agent types), and scalar invariance (invariant item intercepts). This ensures that mean score differences observed between AI and human endorsers reflect true differences in perceived similarity rather than measurement artifacts.<\/p>\n The operational specifications of the Similarity of Endorsers Scale are summarized below:<\/p>\n The Similarity of Endorsers Scale was published in 2021<\/strong> in the Journal of Advertising<\/em> by Veronica L. Thomas and Kendra Fowler. The scale items and psychometric properties are detailed in the published academic record:<\/p>\n Thomas, V. L., & Fowler, K. (2021). Close Encounters of the AI Kind: Use of AI Influencers as Brand Endorsers. Journal of Advertising, 50(1), 11\u201325.<\/em><\/p>\n Licensing and Usage:<\/strong> The scale is available for non-commercial academic research, pedagogical use, and scientific inquiry without royalties or licensing fees, under standard academic fair-use conventions. Researchers utilizing the instrument should properly cite the original publication in any resulting scholarly papers, conference presentations, or dissertations. Commercial entities seeking to integrate the scale into proprietary market testing platforms or syndicated brand analytics tools should consult the copyright policies of the American Academy of Advertising and the publisher (Taylor & Francis<\/a>) to ensure compliance.<\/p>\n Campbell, D. T. (1958). Common fate, similarity, and other indices of the status of aggregates of persons as social entities. Behavioral Science<\/em>, 3(1), 14\u201325. https:\/\/doi.org\/10.1002\/bs.3830030103<\/a><\/p>\n Fiske, S. T., Cuddy, A. J., Glick, P., & Xu, J. (2002). A model of (often mixed) stereotype content: Competence and warmth respectively follow from perceived status and competition. Journal of Personality and Social Psychology<\/em>, 82(6), 878\u2013902. https:\/\/doi.org\/10.1037\/0022-3514.82.6.878<\/a><\/p>\n Fornell, C., & Larcker, D. F. (1981). Evaluating structural equation models with unobservable variables and measurement error. Journal of Marketing Research<\/em>, 18(1), 39\u201350. https:\/\/doi.org\/10.1177\/002224378101800104<\/a><\/p>\n Gray, H. M., Gray, K., & Wegner, D. M. (2007). Dimensions of mind perception. Science<\/em>, 315(5812), 619. https:\/\/doi.org\/10.1126\/science.1134475<\/a><\/p>\n Harkins, S. G., & Petty, R. E. (1981). The multiple source effect in persuasion: The effects of commonality of source expertise on cognitive effort and attitude change. Journal of Personality and Social Psychology<\/em>, 41(4), 625\u2013635. https:\/\/doi.org\/10.1037\/0022-3514.41.4.625<\/a><\/p>\n Hu, L. T., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal<\/em>, 6(1), 1\u201355. https:\/\/doi.org\/10.1080\/10705519909540118<\/a><\/p>\n Linville, P. W., Fischer, G. W., & Salovey, P. (1989). Perceived distributions of the characteristics of in-group and out-group members: Empirical evidence and a computer simulation. Journal of Personality and Social Psychology<\/em>, 57(2), 165\u2013188. https:\/\/doi.org\/10.1037\/0022-3514.57.2.165<\/a><\/p>\n McCroskey, J. C., Richmond, V. P., & Daly, J. A. (1975). The development of a measure of perceived homophily in interpersonal communication. Human Communication Research<\/em>, 1(4), 323\u2013332. https:\/\/doi.org\/10.1111\/j.1468-2958.1975.tb00281.x<\/a><\/p>\n Nass, C., & Reeves, B. (1996). The media equation: How people treat computers, television, and new media like real people and places<\/em>. Cambridge University Press.<\/p>\n Ohanian, R. (1990). Construction and validation of a scale to measure celebrity endorsers’ perceived expertise, trustworthiness, and attractiveness. Journal of Advertising<\/em>, 19(3), 39\u201352. https:\/\/doi.org\/10.1080\/00913367.1990.10673191<\/a><\/p>\n Park, B., & Rothbart, M. (1982). Perception of out-group homogeneity and levels of social categorization: Memory for the subordinate attributes of in-group and out-group members. Journal of Personality and Social Psychology<\/em>, 42(6), 1051\u20131068. https:\/\/doi.org\/10.1037\/0022-3514.42.6.1051<\/a><\/p>\n Tajfel, H., & Turner, J. C. (1986). The social identity theory of intergroup behavior. In S. Worchel & W. G. Austin (Eds.), Psychology of intergroup relations<\/em> (pp. 7\u201324). Nelson-Hall.<\/p>\n2. Keywords<\/h2>\n
3. Authors<\/h2>\n
\n
4. Purpose<\/h2>\n
\n
5. Psychological Construct<\/h2>\n
\n
\n
6. Theoretical Framework<\/h2>\n
Social Categorization Theory and the Outgroup Homogeneity Effect<\/h3>\n
Perceived Entitativity<\/h3>\n
The Computers Are Social Actors (CASA) Paradigm<\/h3>\n
7. Validity<\/h2>\n
Construct and Convergent Validity<\/h3>\n
Discriminant Validity<\/h3>\n
\n
Predictive and Nomological Validity<\/h3>\n
\n
8. Reliability<\/h2>\n
Internal Consistency Metrics<\/h3>\n
\n
Item-Total Correlations and Stability<\/h3>\n
9. Factor Analysis<\/h2>\n
Exploratory Factor Analysis (EFA)<\/h3>\n
\n
Confirmatory Factor Analysis (CFA)<\/h3>\n
\n
Measurement Invariance<\/h3>\n
10. Instrument \/ Measurement Tool<\/h2>\n
\n
\n
\n
11. Permissions & Fee and Test Year<\/h2>\n
12. References<\/h2>\n