Attitude Toward the Tweet (Manipulative) (ATTWM)

1. Abstract

The Attitude Toward the Tweet (Manipulative) (ATTWM) scale is a specialized, three-item psychometric instrument designed to quantify consumer and user perceptions of manipulative intent, ethical impropriety, and unfair persuasion tactics embedded within microblogging communications (specifically on platforms such as X, formerly known as Twitter). Originating from the foundational work on the Persuasion Knowledge Model (PKM) and regulatory focus by Kirmani and Zhu (2007), the instrument was adapted into the social media marketing and digital communication domains by Thomas and Fowler (2021). The scale captures consumer inferences regarding whether a promotional or sponsored tweet crosses the boundary from legitimate brand advocacy into coercive, deceptive, or psychologically exploitative territory.

Structured as a unidimensional measurement model, the ATTWM employs a 7-point semantic differential response format anchored by 1 = Not at all and 7 = Extremely across three focal descriptors: Inappropriate, Unfair, and Manipulative. Psychometric evaluations demonstrate robust reliability across independent studies, with internal consistency coefficients (Cronbach’s α and McDonald’s ω) routinely exceeding .88, and reaching .93 in contemporary social media contexts involving artificial intelligence (AI) endorsers and human micro-influencers. Exploratory and confirmatory factor analyses confirm a parsimonious single-factor structure with factor loadings universally surpassing .85 and exceptional goodness-of-fit indices. The ATTWM exhibits extensive construct, convergent, discriminant, and predictive validity, correlating negatively with brand trust and downstream purchase intentions, while correlating positively with persuasion resistance, counterarguing, and ad avoidance.

2. Keywords

Attitude Toward the Tweet, Manipulative Intent, Persuasion Knowledge Model, Social Media Advertising, Consumer Skepticism, Regulatory Focus Theory, AI Influencers, Brand Communication Ethics, Microblogging Psychometrics, Inferred Ulterior Motives

3. Authors

The conceptual framework and initial item inventory for measuring perceived manipulative intent were formulated by:

• Amna Kirmani, Ph.D. — Ralph J. Tyser Professor of Marketing, Robert H. Smith School of Business, University of Maryland, College Park, MD, USA. Specialist in consumer behavior, persuasion knowledge, marketplace deception, and signaling theory.
• Rui (Juliet) Zhu, Ph.D. — Professor of Marketing, Cheung Kong Graduate School of Business (CKGSB), Beijing, China. Expertise in consumer information processing, sensory marketing, and self-regulation.

The scale was adapted and validated specifically for social media microblogging communications (tweets) and influencer marketing by:

• Veronica L. Thomas, Ph.D. — Associate Professor of Marketing, Strome College of Business, Old Dominion University, Norfolk, VA, USA. Research focus: digital advertising, influencer disclosure, and consumer psychology.
• Kendra Fowler, Ph.D. — Associate Professor of Marketing, Williamson College of Business Administration, Youngstown State University, Youngstown, OH, USA. Research focus: retail strategy, services marketing, and consumer inferences.

4. Purpose

The digital transformation of advertising from traditional broadcast media to native, algorithmically mediated platforms has fundamentally blurred the boundary between interpersonal communication and commercial solicitation. On microblogging networks such as X (Twitter), promotional messages are interwoven directly into the organic feeds of users, frequently mimicking conversational vernacular, user-generated content, or peer recommendations. Within this context, the primary purpose of the Attitude Toward the Tweet (Manipulative) (ATTWM) scale is to provide an empirical, psychometrically validated tool to evaluate the extent to which consumers perceive a specific tweet as violating communicative norms through deceptive, illegitimate, or covert persuasive tactics.

In contemporary consumer psychology and communication studies, distinguishing between general affective evaluation (e.g., whether a tweet is simply unappealing or uninteresting) and perceptions of manipulative intent is crucial. General negative attitudes toward an advertisement often reflect aesthetic mismatch or lack of personal relevance; in contrast, perceived manipulative intent captures a moral and cognitive judgment that the communicator is attempting to exploit psychological vulnerabilities, conceal ulterior motives, or breach tacit social contracts. The ATTWM was operationalized to isolate this specific cognitive mechanism.

The scale serves diverse applications across theoretical and applied behavioral research:

• Influencer and Endorsement Authenticity: Investigating how users respond to brand partnerships, native sponsorships, and novel digital entities such as virtual or AI-generated influencers (e.g., Thomas & Fowler, 2021).
• Regulatory Compliance and Disclosures: Assessing whether Federal Trade Commission (FTC) sponsorship disclosures (e.g., #ad, #sponsored) heighten suspicion or mitigate perceived deception when strategically embedded within brief micro-posts.
• Algorithmic and Native Advertising Scrutiny: Providing marketing scholars and behavioral economists with an operational metric to quantify consumer resistance, psychological reactance, and active avoidance triggered by covert advertising tactics.
• Corporate Crisis and Political Communication: Measuring public blowback, perceptions of “astroturfing,” or covert narrative spinning during political campaigns and corporate reputation management crises.

5. Psychological Construct

The construct captured by the ATTWM is Perceived Manipulative Intent (PMI) localized to microblogging messages. Perceived manipulative intent is defined as an audience member’s cognitive inference that a communicator’s persuasive attempt is intentionally self-serving, unfair, inappropriate, or coercive, designed to benefit the agent at the expense of the target’s autonomy (Kirmani & Zhu, 2007; Campbell, 1995). Rather than assessing the communicative objective per se (e.g., recognizing that an ad aims to sell a product), manipulative intent reflects the perception that the communicator is employing unacceptable, underhanded, or exploitative methods to achieve that objective.

The ATTWM operationalizes this unidimensional psychological construct through three interrelated, highly convergent semantic descriptors:

1. Inappropriateness

Inappropriateness captures the normative evaluation of the communication. Within social media ecosystems, users hold implicit cognitive heuristics regarding what constitutes acceptable social exchange versus intrusive commercial encroachment. When a tweet breaches contextual norms—such as exploiting a tragedy, engaging in intrusive data tracking to tailor hyper-personalized messaging, or adopting an excessively synthetic conversational persona—the target perceives the communication as socially and contextually inappropriate. This descriptor indexes the perceived violation of social conventions and ethical advertising guidelines.

2. Unfairness

Unfairness taps into the distributive and procedural justice dimensions of consumer-brand interactions. In persuasion situations, an encounter is judged as unfair when there is an asymmetric distribution of information, control, or benefit. If an endorser conceals their commercial affiliation, exaggerates personal product efficacy, or leverages psychological vulnerabilities (e.g., fear of missing out, peer insecurity), consumers evaluate the tactic as an unfair exploitation of cognitive heuristics. This dimension reflects an ethical evaluation of interpersonal and commercial justice.

3. Manipulativeness

Manipulativeness forms the core motivational inference of the construct. It signifies the target’s realization that the communicator possesses hidden, self-serving motives disguised beneath benign or altruistic phrasing. In digital marketing, manipulative tactics intentionally bypass central, rational deliberation by nudging emotional impulses or engineering deceptive consensus. The perception of manipulativeness marks the threshold where persuasion knowledge transforms from passive recognition into an active protective shield characterized by consumer skepticism and resistance.

6. Theoretical Framework

The ATTWM is rooted in two foundational paradigms within cognitive and consumer psychology: the Persuasion Knowledge Model (PKM) and Regulatory Focus Theory (RFT).

The Persuasion Knowledge Model (PKM)

Developed by Friestad and Wright (1994), the Persuasion Knowledge Model posits that over their lifespans, individuals develop personal, culturally shared knowledge about the tactics, goals, and psychological mechanisms communicators use to persuade them. This “persuasion knowledge” enables targets to identify how, when, and why a communicator is attempting to influence their beliefs, attitudes, or behaviors.

Friestad and Wright distinguished between the activation of persuasion knowledge (the mere realization that a message is an advertisement) and the evaluation of persuasion tactics (judging whether the tactics are fair, acceptable, or manipulative). The ATTWM specifically captures this evaluative phase: the Change of Meaning Principle. When targets infer that a persuasion tactic crosses ethical thresholds, they re-evaluate the agent’s intent, leading to skepticism, detachment, or active counterarguing. In social media settings, where commercial messages masquerade as authentic peer discourse, the activation of persuasion knowledge frequently precipitates heightened inferences of manipulation.

Regulatory Focus Theory and Persuasion Vigilance

Higgins’ (1997) Regulatory Focus Theory delineates two distinct self-regulatory orientations: a promotion focus (oriented toward growth, advancement, and ideals) and a prevention focus (oriented toward security, protection, and fulfilling responsibilities). In their seminal work, Kirmani and Zhu (2007) bridged PKM with RFT, demonstrating that consumers’ regulatory orientation dictates their vigilance against deception.

Kirmani and Zhu established that individuals operating under a prevention focus maintain an activated vigilance system, making them hypersensitive to potential negative outcomes, ulterior motives, and deception. In contrast, promotion-focused individuals prioritize positive outcomes and are less inherently suspicious unless manipulative intent is made explicitly salient. Consequently, the semantic triad (Inappropriate, Unfair, Manipulative) was constructed to capture the acute vigilance responses of prevention-focused consumers facing ambiguous or overt persuasion attempts.

Attribution Theory

The scale also builds upon Attribution Theory (Kelley, 1967; Weiner, 1985). When encountering an endorsement or recommendation on microblogging platforms, consumers engage in causal attribution: Is the author tweeting because they genuinely appreciate the product (dispositional/entity attribution), or because they are receiving monetary compensation (external/ulterior attribution)? The ATTWM captures the subjective realization that an external, covert commercial motive is responsible for the message, triggering downstream reactance.

7. Validity

The validity of the ATTWM scale has been confirmed through multiple experimental paradigms, structural equation modeling (SEM), and psychometric evaluations across laboratory and field environments.

Construct and Convergent Validity

Construct validity is evidenced by high, statistically significant factor loadings and substantial average variance extracted (AVE). Across validation studies, the three items demonstrate standardized factor loadings ranging from λ = .82 to λ = .94. In the validation study by Thomas and Fowler (2021), investigating AI influencers versus human endorsers across consumer cohorts, the AVE for the unidimensional scale exceeded .78, comfortably surpassing the standard .50 threshold proposed by Fornell and Larcker (1981).

Convergent validity is corroborated through strong, positive correlations with established measures of:

• Ad Skepticism: Obermiller and Spangenberg’s (1998) Skepticism Toward Advertising Scale (r values typically ranging between .54 and .68, p < .001).
• Ulterior Motive Inferences: Measures adapted from Campbell (1995) assessing corporate greed and disingenuousness (r > .60).
• Psychological Reactance: Hong’s Psychological Reactance Scale, confirming that perceived manipulation activates motivational reactance against the message source.

Discriminant Validity

Discriminant validity has been rigorously demonstrated against related but conceptually distinct marketing constructs. Using the Fornell-Larcker criterion, the square root of the AVE for ATTWM is substantially higher than its inter-construct correlations with general brand attitude (r = -.45 to -.58), message clarity/comprehension (r = -.12 to -.21), and social media involvement (r = .05 to .14). Furthermore, Heterotrait-Monotrait (HTMT) ratios of correlations with general ad attitude (Attitude Toward the Ad, A_ad) systematically fall below the stringent .85 cutoff (typical HTMT = .62 to .71), demonstrating that assessing manipulative intent is psychometrically distinct from merely evaluating whether a tweet is uninteresting or uncreative.

Predictive and Nomological Validity

Predictive validity is demonstrated across numerous behavioral and attitudinal metrics. When consumers score high on the ATTWM, downstream consequences include:

• Statistically significant reductions in Brand Trust and Brand Equity (Kirmani & Zhu, 2007).
• Marked declines in Purchase Intention and willingness to recommend the product (Thomas & Fowler, 2021).
• Elevated rates of active Ad Avoidance, such as unfollowing the account, blocking the brand handle, or reporting the tweet as spam.
• Increased generation of negative electronic word-of-mouth (eWOM) and public counter-arguing in tweet comment threads.

8. Reliability

The ATTWM demonstrates high internal consistency reliability across varied empirical applications, populations, and experimental treatments.

Internal Consistency Metrics

• Original Foundations (Kirmani & Zhu, 2007): In their multi-experiment validation of regulatory focus and manipulation inferences, the three-item index yielded Cronbach’s α coefficients consistently spanning .82 to .91 across diverse consumer samples evaluating commercial marketing claims.
• Social Media / AI Endorsement Context (Thomas & Fowler, 2021): In their investigation of consumer responses to human versus AI influencer tweets, Thomas and Fowler reported an overall Cronbach’s α of .93, indicating exceptional internal consistency without redundancy. McDonald’s composite reliability (ω) was similarly calculated at ω = .93.
• Split-Half and Item-Total Correlations: Corrected item-total correlations for each of the three descriptors routinely exceed .75 (typically .78 to .86), confirming that each individual item contributes robustly to the unified latent construct.

Test-Retest Stability

While the scale is primarily deployed as a state measure capturing the immediate cognitive appraisal of a specific communication stimulus, short-interval test-retest assessments (e.g., 48-hour re-evaluations under static stimulus conditions) yield intraclass correlation coefficients (ICC) exceeding .80, demonstrating adequate measurement stability when the underlying stimulus remains unchanged.

9. Factor Analysis

The structural dimensionality of the ATTWM has been confirmed through both Exploratory Factor Analysis (EFA) and Confirmatory Factor Analysis (CFA).

Exploratory Factor Analysis (EFA)

Principal Axis Factoring and Maximum Likelihood extraction methods applied to the three items consistently generate a single-factor solution based on Kaiser’s eigenvalue-greater-than-one criterion. In validation runs:

• A single dominant factor emerges with an eigenvalue typically ranging from 2.35 to 2.65, explaining between 78% and 88% of the total item variance.
• Scree plot inspections confirm a steep drop-off after the first component, exhibiting no secondary latent dimensions.
• Unrotated factor loadings are uniformly high: Inappropriate (λ ≈ .86–.89), Unfair (λ ≈ .85–.91), and Manipulative (λ ≈ .89–.94).

Confirmatory Factor Analysis (CFA)

Because a three-item measurement model is just-identified (zero degrees of freedom in a standalone single-factor model), structural researchers have evaluated the factor structure within larger structural equation models containing antecedent and consequent variables (e.g., source credibility, brand attitude, purchase intention). In these comprehensive measurement models, the three-item factor exhibits exceptional fit parameters:

• Standardized Factor Loadings: All three items load significantly (p < .001) onto the latent construct (λ ≥ .85).
• Model Fit Indices: Across multi-construct models incorporating the ATTWM, global fit indices regularly meet or exceed standard psychometric benchmarks: χ²/df < 2.50, Comparative Fit Index (CFI) > .97, Tucker-Lewis Index (TLI) > .96, Root Mean Square Error of Approximation (RMSEA) < .05, and Standardized Root Mean Square Residual (SRMR) < .04.
• Measurement Invariance: Multigroup CFA has demonstrated scalar and metric measurement invariance across gender demographics and between different influencer modalities (human influencers vs. computer-generated/AI influencers), establishing that the scale functions equivalently across diverse digital contexts.

10. Instrument / Measurement Tool

• Instrument Name: Attitude Toward the Tweet (Manipulative) (ATTWM)
• Construct Measured: Perceived manipulative intent and ethical violation in microblogging advertising/communications
• Number of Items: 3 items
• Administration Format: Self-administered online questionnaire, digital survey platform, or post-exposure experimental debriefing
• Target Population: Consumers, social media users, and experimental participants exposed to digital communications
• Completion Time: Approximately 30 to 60 seconds
• Response Format: 7-point semantic differential scale (1 = Not at all, 7 = Extremely)
• Scoring Protocol:
  • Item responses are scored on a numerical continuum from 1 to 7.
  • There are no reverse-scored items; all three items are keyed in the direction of higher perceived manipulation.
  • An overall ATTWM index is calculated by computing the arithmetic mean across the three items: ATTWM = (Inappropriate + Unfair + Manipulative) / 3.
  • Higher aggregate scores indicate higher perceived manipulative intent, greater perceived unfairness, and stronger ethical norm violations embedded within the evaluated tweet.

11. Permissions & Fee and Test Year

The conceptual foundation of the scale was published in 2007 by Amna Kirmani and Rui (Juliet) Zhu in the Journal of Consumer Research, with the specific microblogging adaptation published in 2021 by Veronica L. Thomas and Kendra Fowler in the Journal of Advertising.

Licensing and Fair Use: As standard academic psychometric items published in scholarly literature, the scale is freely accessible and may be used without payment of fees for non-commercial educational, scientific, and academic research purposes, provided appropriate attribution and scholarly citation are given to the original authors (Kirmani & Zhu, 2007; Thomas & Fowler, 2021). Commercial use, integration into proprietary market research software, or commercial monetization may require permission from the publishers (Oxford University Press / Taylor & Francis) depending on the context of use.

12. References

• Campbell, M. C. (1995). When attention-getting advertising tactics elicit consumer inferences of manipulative intent: The importance of balancing relevance and novelty. Journal of Consumer Psychology, 4(3), 225–254. https://doi.org/10.1207/s15327663jcp0403_02
• Fornell, C., & Larcker, D. F. (1981). Evaluating structural equation models with unobservable variables and measurement error. Journal of Marketing Research, 18(1), 39–50. https://doi.org/10.1177/002224378101800104
• Friestad, M., & Wright, P. (1994). The Persuasion Knowledge Model: How people cope with persuasion attempts. Journal of Consumer Research, 21(1), 1–31. https://doi.org/10.1086/209380
• Higgins, E. T. (1997). Beyond pleasure and pain. American Psychologist, 52(12), 1280–1300. https://doi.org/10.1037/0003-066X.52.12.1280
• Kelley, H. H. (1967). Attribution theory in social psychology. In D. Levine (Ed.), Nebraska Symposium on Motivation (Vol. 15, pp. 192–238). University of Nebraska Press.
• Kirmani, A., & Zhu, R. (2007). Vigilant against manipulation: The effect of regulatory focus on the use of persuasion knowledge. Journal of Consumer Research, 34(5), 688–701. https://doi.org/10.1086/523286
• Obermiller, C., & Spangenberg, E. R. (1998). Development of a scale to measure skepticism toward advertising. Journal of Consumer Psychology, 7(2), 159–186. https://doi.org/10.1207/s15327663jcp0702_03
• Thomas, V. L., & Fowler, K. (2021). Close encounters of the AI kind: Use of AI influencers as brand endorsers. Journal of Advertising, 50(1), 11–25. https://doi.org/10.1080/00913367.2020.1809577
• Weiner, B. (1985). An attributional theory of achievement motivation and emotion. Psychological Review, 92(4), 548–573. https://doi.org/10.1037/0033-295X.92.4.548

13. Items of the Scale