Attitude Toward the Tweet (Surprising) (ATTWS)

Abstract

The Attitude Toward the Tweet (Surprising) (ATTWS) is a specialized, unidimensional psychometric instrument developed to evaluate consumer and audience perceptions of epistemic unexpectedness within microblogging and social media marketing communication. Originally introduced and validated by Borah, Banerjee, Lin, Jain, and Eisingerich (2020) in their seminal investigation into improvised marketing interventions (IMIs), the scale quantifies the degree to which an individual tweet defies pre-existing cognitive expectations and catches the reader off guard. Comprising three concise semantic or Likert-type items reflecting core markers of surprise—specifically whether the message is perceived as unexpected, surprising, or unanticipated—the ATTWS functions both as a critical manipulation check in experimental consumer psychology and as an explanatory variable predicting message virality, electronic word-of-mouth (eWOM), brand engagement, and sentiment diffusion.

Psychometrically, the ATTWS demonstrates robust internal consistency, with reported Cronbach’s alpha (α) coefficients routinely exceeding .85 to .90 across diverse experimental cohorts and field samples. Confirmatory factor analysis confirms a unidimensional structure with strong factor loadings (> .80) and high convergent and discriminant validity with respect to adjacent communicative dimensions such as humor, timeliness, relevance, and overall message valence. By isolating surprise from general novelty or valenced affect, the scale provides researchers with a granular, reliable measurement tool for testing cognitive schema-incongruity theories, social transmission dynamics, and real-time digital advertising efficacy.

Keywords

Attitude Toward the Tweet, surprise, schema incongruity, improvised marketing interventions, social media marketing, virality, psychometrics, consumer attitude, microblogging, digital communication, expectancy violation

Authors

The scale was developed and operationalized by an international team of marketing and management scholars:

• Abhishek Borah – INSEAD, Department of Marketing, Boulevard de Constance, 77305 Fontainebleau, France. Expertise in social media analytics, virality, and quantitative marketing models.
• Sourindra Banerjee – University of Leeds, Leeds University Business School, Leeds, United Kingdom. Specializing in marketing strategy, firm capabilities, and marketing agility.
• Yu-Ting Lin – National Taiwan University, College of Management, Taipei, Taiwan. Research interests span social media communication, digital consumer psychology, and algorithmic engagement.
• Apurv Jain – MacroPolo / Independent Researcher, formerly at Microsoft and Harvard Business School. Focuses on macroeconomic data, digital platforms, and behavioral economics.
• Andreas B. Eisingerich – Imperial College Business School, Imperial College London, South Kensington Campus, London, United Kingdom. Specializing in brand attachment, consumer engagement, and digital service innovation.

Purpose

The primary purpose of the Attitude Toward the Tweet (Surprising) (ATTWS) scale is to provide a reliable, psychometrically validated metric that captures the cognitive and emotional shock value embedded within short-form digital marketing communications. In contemporary digital ecosystems—characterized by hyper-saturation, high information velocity, and shortened human attention spans—brands constantly compete for consumer notice. To penetrate cognitive filters, organizations increasingly deploy improvised marketing interventions (IMIs)—unplanned, real-time communicative executions composed and executed in response to exogenous cultural, social, or sporting events. A foundational determinant of whether an IMI succeeds in triggering virality is its capacity to invoke epistemic surprise.

From a research and experimental methodology standpoint, the ATTWS serves several distinct functions:

• Experimental Manipulation Check: In laboratory and online factorial experiments, researchers manipulating message unexpectedness require a clean, unconfounded instrument to confirm that experimental stimuli intended to be surprising are statistically perceived as such by participants, without conflating surprise with positive valence, humor, or confusion.
• Mediational and Mechanistic Analysis: The scale enables scholars to test sophisticated mediation models, investigating whether the effect of real-time brand wit or schema-incongruent messaging on behavioral outcomes (retweets, brand equity, purchase intent) is psychologically mediated by the reader’s subjective experience of surprise.
• Social Transmission and Virality Modeling: Rooted in information transmission theory and emotional contagion models, the ATTWS helps quantify why certain brand communications achieve exponential virality while structurally similar messages languish. Surprise induces an orienting response that drives behavioral re-transmission.
• Managerial Diagnostic Tool: In practical marketing analytics, the ATTWS framework can be deployed during copy pre-testing or automated natural language sentiment evaluation to forecast how radical or unexpected a proposed social media campaign will be perceived across disparate consumer demographic segments.

Psychological Construct

The construct measured by the ATTWS is perceived message surprise within the context of microblogging communications. Psychologically, surprise is classified as an epistemic, neutral affective-cognitive state triggered by an organism’s encounter with schema-discrepant or unexpected stimuli. Unlike pure affective valences such as joy or anger, surprise serves fundamentally as a cognitive interruption mechanism: it alerts the central executive system that existing predictive models of the environment have failed to anticipate an event accurately.

The ATTWS operationalizes this construct along three closely coupled, synonymous facets:

• Unexpectedness: The degree to which the communicative stimulus departs from the recipient’s normative probabilistic forecast. In the context of a tweet, unexpectedness reflects the recipient’s immediate realization that the brand or user was not anticipated to say what was said, or that the association drawn was statistically improbable within ordinary brand discourse. For example, when an airline leverages a viral pop-culture mishap to crack a spontaneous joke, the unexpectedness score rises because consumers do not anticipate corporate accounts deviating from formal customer support registers.
• Surprise: The primary affective-cognitive reaction marked by momentary startle and cognitive re-orientation. This facet captures the immediate psychological “jolt” that dislodges the consumer from passive digital scrolling. It constitutes the experiential core of the construct, demanding attentional allocation to parse the communicative intent.
• Unanticipated Nature: The structural or temporal absence of prior anticipation or mental priming. This dimension captures the lack of preparatory cognitive schema; the tweet enters the respondent’s perceptual field completely unheralded, without antecedent cues or telegraphing.

Crucially, the construct tapped by the ATTWS is inherently valence-neutral. A message can be profoundly surprising yet either delightfully witty or intensely offensive. The ATTWS deliberately restricts its psychometric scope to the degree of schema disruption and cognitive unexpectedness, ensuring that evaluations of message liking, humor appreciation, or moral outrage are treated as theoretically and empirically distinct constructs.

Theoretical Framework

The ATTWS is rooted in several interconnected theories spanning cognitive psychology, information processing, and consumer behavior:

1. Schema Incongruity Theory

Originally formulated by Piaget and expanded into consumer psychology by Mandler (1982), Schema Incongruity Theory posits that human beings interpret reality through cognitive schemas—structured frameworks of knowledge regarding objects, entities, and social roles. When an incoming stimulus matches an activated schema (schema congruence), processing is fluent, effortless, and typically elicits mild, routine cognitive evaluations. When an incoming stimulus deviates moderately from the schema (schema incongruity), it triggers cognitive arousal and epistemic curiosity. Borah et al. (2020) argue that improvised social media marketing succeeds precisely because it introduces deliberate, controlled incongruity: consumers maintain a rigid schema of corporate accounts as sterile, self-promotional, and bureaucratic. A swift, culturally astute, self-deprecating, or contextual tweet violently conflicts with that schema, evoking a sharp elevation on the ATTWS.

2. Expectancy Violations Theory (EVT)

Developed by Judee Burgoon (1978; 1993), Expectancy Violations Theory suggests that individuals develop normative expectations regarding communicative interaction. When an expectation is violated, the recipient’s attention is immediately diverted from the core message to the violator and the nature of the violation. If the violation is evaluated positively (a positive violation valence), it produces heightened communicative effectiveness, interpersonal attraction, and persuasive impact compared to strict norm conformity. The ATTWS captures the cognitive detection of the violation itself—the necessary antecedent before the recipient assigns positive or negative valence to the brand’s audacity.

3. Cognitive Dissonance and Epistemic Curiosity

Surprise creates a temporary cognitive gap (Loewenstein, 1994). When a tweet generates high scores on the ATTWS, it challenges existing informational assumptions, compelling the reader to process the text more deeply to reconcile the incongruity. This increased depth of processing (Petty & Cacioppo, 1986) strengthens memory encoding, heightens cognitive elaboration, and significantly increases the probability that the consumer will share the tweet with peers to validate or share the surprising experience.

Validity

The validity of the ATTWS has been extensively evaluated across both laboratory experimental designs and observational field datasets in digital marketing and psychometric testing:

Construct and Content Validity

Content validity was established through thorough domain sampling of linguistic descriptors of surprise in cognitive psychology literature (e.g., Meyer, Niepel, Rudolph, & Schützwohl, 1991). By employing three closely synonymous items (unexpected, surprising, unanticipated), the instrument achieves saturated content coverage of the cognitive-affective state without diluting the unidimensional focus or introducing semantic drift into adjacent constructs such as shock, disgust, or confusion.

Convergent and Discriminant Validity

In the empirical testing conducted by Borah et al. (2020), convergent validity was supported by high and statistically significant standardized factor loadings (> .85) and an Average Variance Extracted (AVE) well exceeding the recommended benchmark of .50 (frequently surpassing .75). Discriminant validity was established using the Fornell-Larcker criterion and Heterotrait-Monotrait (HTMT) ratios against closely related social media evaluation constructs:

• Attitude Toward the Tweet (Humorous): HTMT ratios remained well below the conservative .85 threshold, demonstrating that respondents clearly distinguish between finding a tweet funny versus finding it surprising.
• Message Timeliness: Despite both constructs being central to real-time marketing, correlations between timeliness and surprise were moderate (typically r = .30 to .45), indicating that a tweet can be recognized as prompt without being perceived as unexpected.
• Attitude Toward the Brand: Pre-existing brand affinity did not load on or confound the ATTWS, affirming that the measure isolates message-specific perceptual reactions rather than generalized halo effects.

Predictive and Nomological Validity

Predictive validity is demonstrated by the scale’s capacity to forecast consumer engagement behaviors. High scores on the ATTWS significantly predict retweeting intentions, actual retweet counts, reply volume, and positive sentiment in downstream commentary. Mediational analyses confirm that the ATTWS mediates the relationship between improvised communicative strategies and viral reach, satisfying Baron and Kenny’s as well as Preacher and Hayes’ bootstrapping criteria for nomological validity.

Reliability

The ATTWS demonstrates exceptionally high internal consistency and measurement reliability across multiple experimental replications and contextual conditions:

• Internal Consistency (Cronbach’s Alpha): Across the empirical studies presented in Borah et al. (2020), the three-item instrument consistently yielded Cronbach’s alpha (α) coefficients between .88 and .94. These values reflect exemplary item homogeneity, well above the standard .70 cut-off for research instruments.
• Composite Reliability (CR): Structural equation modeling reveals composite reliability values exceeding .90, demonstrating that the measurement error associated with the latent variable is minimal.
• Test-Retest Stability: While the construct of surprise is inherently dynamic and state-dependent—decaying rapidly upon repeated exposure to the exact same stimulus—immediate test-retest reliability across split-half survey sections during pre-testing showed robust correlation coefficients (r > .82), demonstrating that the scale consistently records the immediate cognitive impact of the stimulus.

Factor Analysis

Both Exploratory Factor Analysis (EFA) and Confirmatory Factor Analysis (CFA) have been executed to evaluate the latent dimensional architecture of the ATTWS.

Exploratory Factor Analysis (EFA)

When subjecting the three items alongside items measuring humor, timeliness, and brand attitude to Principal Axis Factoring with Promax rotation, the ATTWS items load cleanly onto a single, independent factor. The first eigenvalue accounts for over 75% of the total variance among the target items, with no cross-loadings onto humor or general attitude exceeding .20. Kaiser-Meyer-Olkin (KMO) measures of sampling adequacy routinely exceed .80, and Bartlett’s Test of Sphericity demonstrates statistical significance (p < .001).

Confirmatory Factor Analysis (CFA)

In structural equation modeling frameworks, a single-factor CFA model demonstrates superior model fit indices when evaluating the latent construct:

• Standardized Factor Loadings (λ): Loadings for the three items range from .84 to .95, each statistically significant at p < .001.
• Goodness-of-Fit Indices: Multi-factor measurement models incorporating the ATTWS along with companion scales exhibit excellent fit: Root Mean Square Error of Approximation (RMSEA) ≤ .045; Comparative Fit Index (CFI) ≥ .98; Tucker-Lewis Index (TLI) ≥ .98; and Standardized Root Mean Square Residual (SRMR) ≤ .030.
• Measurement Invariance: The single-factor model exhibits metric and scalar invariance across diverse consumer groups (e.g., differing age cohorts, high vs. low social media usage frequency), confirming that the instrument evaluates the same underlying psychological construct across distinct sub-populations.

Instrument / Measurement Tool

• Instrument Name: Attitude Toward the Tweet (Surprising) (ATTWS)
• Alternative Designations: Perceived Tweet Unexpectedness Scale, Surprising Dimension of Attitude Toward the Tweet
• Target Population: Social media users, consumers, and experimental survey participants evaluating digital brand communications.
• Construct Measured: Perceived unexpectedness, schema incongruity, and cognitive surprise elicited by a specific tweet.
• Item Count: 3 items (unidimensional)
• Response Format: Typically administered as a 7-point Likert scale (ranging from 1 = Strongly Disagree to 7 = Strongly Agree) or a 7-point semantic differential scale anchored by bipolar adjective pairs (e.g., 1 = Not at all surprising to 7 = Very surprising).
• Administration Time: Approximately 30 to 45 seconds.
• Scoring Protocol: Individual item scores are summed or averaged to yield a composite score ranging from 1.0 to 7.0. Higher scores reflect a higher degree of perceived surprise and cognitive unexpectedness. Reverse scoring is not required as all items are positively keyed.

Permissions & Fee and Test Year

• Publication Year: 2020
• Original Publication Outlet: Journal of Marketing, published by the American Marketing Association (AMA) in partnership with SAGE Publications.
• Copyright & Licensing: The theoretical construct, empirical validation, and publication copyright are held by the American Marketing Association / SAGE Publications and the original authors (Borah et al., 2020).
• Research & Educational Use: Academic researchers may utilize the scale items for non-commercial scholarly research, classroom instruction, and educational theses provided appropriate formal academic citation is rendered. Commercial deployment, inclusion in proprietary survey batteries, or commercial market research software may require explicit permission or licensing from the copyright holder.

References

• Borah, A., Banerjee, S., Lin, Y.-T., Jain, A., & Eisingerich, A. B. (2020). Improvised marketing interventions in social media. Journal of Marketing, 84(2), 69–91. https://doi.org/10.1177/0022242919899383
• Burgoon, J. K. (1978). A communication model of personal space violations: Explication and an initial test. Human Communication Research, 4(2), 129–142. https://doi.org/10.1111/j.1468-2958.1978.tb00603.x
• Burgoon, J. K. (1993). Interpersonal expectations, expectancy violations, and emotional communication. Journal of Language and Social Psychology, 12(1-2), 30–48. https://doi.org/10.1177/0261927X93121003
• Loewenstein, G. (1994). The psychology of curiosity: A review and reinterpretation. Psychological Bulletin, 116(1), 75–98. https://doi.org/10.1037/0033-2909.116.1.75
• Mandler, G. (1982). The structure of value: Accounting for taste. In M. S. Clark & S. T. Fiske (Eds.), Affect and Cognition: The 17th Annual Carnegie Symposium on Cognition (pp. 3–36). Lawrence Erlbaum Associates.
• Meyer, W. U., Niepel, M., Rudolph, U., & Schützwohl, A. (1991). An experimental analysis of surprise. Cognition & Emotion, 5(4), 295–311. https://doi.org/10.1080/02699939108411042
• 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