Negative WOM Intentions (Data Context) Scale (NWOM-D)

1. Abstract

The Negative WOM Intentions (Data Context) Scale (NWOM-D) is a specialized psychometric instrument designed to assess the extent to which consumers intend to engage in disparaging, retaliatory interpersonal communication regarding a firm following perceived violations of their data privacy. Originally adapted by Martin, Borah, and Palmatier (2017) from the foundational customer retaliation framework of Grégoire and Fisher (2006), the instrument operationalizes consumer retaliation specifically within the contemporary digital ecosystem of customer data vulnerability, corporate data breaches, and non-transparent information harvesting. The scale comprises three concise, high-impact items evaluated via a 7-point Likert response format ranging from 1 (“Strongly disagree”) to 7 (“Strongly agree”).

Psychometrically, the NWOM-D demonstrates exceptional unidimensionality, marked by strong internal consistency reliability (Cronbach's $\alpha ge .90$; Composite Reliability $\mathrm{CR} ge .91$), substantial factor loadings (standardized $lambda > .85$), and high average variance extracted ($\mathrm{AVE} > .75$). Within structural equation models of consumer privacy governance, the scale functions as a critical behavioral outcome. It captures how feelings of emotional violation and declines in cognitive trust translate into market-level consumer punishment. By isolating negative word-of-mouth (NWOM) as an aggressive punitive mechanism rather than mere passive dissatisfaction, the NWOM-D provides marketing scholars, behavioral economists, and corporate risk officers with an empirically validated diagnostic tool to quantify downstream reputation damage and evaluate the protective efficacy of privacy transparency and customer control mechanisms.

2. Keywords

Negative Word-of-Mouth, Data Vulnerability, Consumer Retaliation, Data Privacy, Customer Trust, Psychological Contract Breach, Online Privacy, Consumer Behavior, Psychometrics, Structural Equation Modeling, Marketing Ethics, Firm Performance.

3. Authors

The adaptation of the NWOM construct to the corporate data privacy context was formulated and empirically validated by:

• Kelly D. Martin, Ph.D. — Professor of Marketing and University Board of Governors Professor, College of Business, Colorado State University. Her research focuses on marketing strategy, consumer data privacy, ethical issues in business, and customer vulnerability.
• Abhishek Borah, Ph.D. — Associate Professor of Marketing, INSEAD (formerly at the Foster School of Business, University of Washington). His expertise lies in digital marketing, social media analytics, online communication, and the financial consequences of corporate crises.
• Robert W. Palmatier, Ph.D. — Professor of Marketing and John C. Narver Chair in Business Administration, Michael G. Foster School of Business, University of Washington. He is a prominent scholar in relationship marketing, customer loyalty, marketing strategy, and privacy exchange frameworks.

The conceptual foundation of the measurement items stems from earlier retaliation research by:

• Yany Grégoire, Ph.D. — Professor of Marketing, HEC Montréal.
• Robert J. Fisher, Ph.D. — Professor of Marketing, Alberta School of Business, University of Alberta.

4. Purpose

The contemporary digitized economy relies heavily on the acquisition, aggregation, analysis, and monetization of consumer personal data. While digital personalization provides utility, it simultaneously exposes consumers to substantial information privacy risks, unauthorized third-party sharing, predatory algorithmic targeting, and catastrophic security breaches. The purpose of the Negative WOM Intentions (Data Context) Scale (NWOM-D) is to measure and quantify a consumer's deliberate, retaliatory motivation to inflict commercial and reputational damage on a firm through targeted interpersonal communication following perceived data vulnerability.

Theoretical and Practical Rationale

Prior customer service literature historically treated negative word-of-mouth as a venting mechanism driven by general dissatisfaction, poor service delivery, or product failure. However, within the domain of personal data misuse, customer reactions frequently exceed dissatisfaction; they trigger intense moral outrage, feelings of betrayal, and explicit desires for retribution. Martin, Borah, and Palmatier (2017) demonstrated that when corporate behaviors generate “data vulnerability”—defined as a consumer's susceptible state resulting from unauthorized data access, sharing, or breach—consumers experience an acute violation of implicit social contracts.

The NWOM-D fulfills several vital research and diagnostic functions:

• Isolating Retaliatory Behavior from Passive Disengagement: Unlike customer churn or silent defection, NWOM represents an active, outward-facing mechanism of behavioral punishment. The NWOM-D directly evaluates intentions to dissuade prospective buyers, denigrate the brand, and spread cautionary narratives.
• Quantifying Downstream Firm Risk: In an era of viral social networks, negative word-of-mouth travels exponentially. By measuring NWOM intentions, researchers can link psychometric scores to real-world financial harm, such as abnormal stock return dips, heightened customer acquisition costs, and suppressed enterprise value.
• Evaluating Privacy Safeguards: Organizations implementing privacy-by-design frameworks, transparent data-use disclosures, or customer-facing control panels require sensitive instruments to determine whether these interventions effectively suppress customer retaliatory intentions.
• Cross-Industry Risk Benchmarking: The instrument facilitates comparative risk assessments across high-stakes domains handling sensitive consumer data, including digital banking, telehealth platforms, cloud-based software, and e-commerce networks.

5. Psychological Construct

The psychological construct measured by the NWOM-D is Retaliatory Negative Word-of-Mouth Intentions in response to corporate data practices. In psychometric and consumer behavior theory, this construct reflects an individual's conscious, planned behavioral disposition to disseminate adverse evaluative information, warn social ties, and denigrate a business entity with the explicit or implicit objective of damaging its reputation or economic welfare.

Dimensional Decomposition of the Construct

Although the NWOM-D is modeled as a parsimonious, unidimensional scale, it encompasses three interrelated psychological and behavioral facets:

• Broad Dissemination (Item 1: “I will spread negative word-of-mouth about the company to other people.”): This facet captures the diffusion velocity and public orientation of retaliatory communication. It encompasses both online mechanisms (e.g., social media broadsides, review platform denunciation) and offline discussions. The underlying cognitive driver is justice restoration: publicizing the firm's privacy malpractice alerts the broader community while punishing the offending institution.
• Interpersonal Denigration within Close Social Networks (Item 2: “I will denigrate the company to my friends.”): Moving beyond objective warnings, this component reflects emotionally charged degradation of the company's standing within primary reference groups. Denigration involves socially devaluing the firm, characterized by hostile appraisals of its ethics, integrity, and operational legitimacy. The motivation stems from psychological reactance and the need to align social circles against the transgressive firm.
• Active Commercial Deterrence and Persuasive Intervention (Item 3: “When my friends look for a similar service/product, I will tell them not to buy from the company.”): This dimension represents an intentional effort to inflict direct commercial harm on the firm by diverting sales to competitors. Rather than merely expressing displeasure, the consumer actively acts as a barrier in the purchasing funnel of peers, neutralizing the firm's market conversion efforts.

Distinction from Related Constructs

The NWOM-D construct differs substantially from related consumer responses:

• NWOM vs. Customer Churn / Defection: Churn is an avoidance coping strategy aimed at terminating a commercial relationship to mitigate personal risk. NWOM is an approach coping strategy focused on outward retribution and social deterrence.
• NWOM vs. Direct Third-Party Complaining: Direct complaining involves seeking formal remediation from consumer protection agencies or legal entities. Retaliatory NWOM operates through informal interpersonal networks, often aiming to inflict social and economic penalties rather than seeking personal restitution.
• NWOM vs. Private Emotional Venting: While venting serves an affective catharsis function, retaliatory NWOM in the data context is explicitly instrumental, designed to alter peer purchasing behavior and weaken the offending firm.

6. Theoretical Framework

The NWOM-D is anchored in a synthesis of Psychological Contract Theory, Cognitive Appraisal Theory of Emotion, and the Social Exchange Retaliation Model.

Psychological Contract Breach and Betrayal

Psychological Contract Theory posits that relationships are governed by unwritten, reciprocal expectations of fairness, integrity, and respect. When consumers entrust sensitive data—such as financial records, location trails, browsing histories, and personal identifiers—to an enterprise, they operate under an implicit contract that the firm will safeguard this asset. When a firm surreptitiously shares data, suffers an unforced breach due to inadequate oversight, or monetizes user profiles without explicit consent, a profound psychological contract breach occurs. This breach is perceived not merely as an operational failure, but as corporate betrayal.

The Dual Mediation Mechanism: Emotional Violation vs. Cognitive Trust

Martin, Borah, and Palmatier (2017) formalized how data vulnerability translates into retaliatory outcomes through two parallel psychological mechanisms:

1. The Affective Pathway (Feeling of Violation): In accordance with the Cognitive Appraisal Theory of Emotion (Lazarus & Folkman, 1984), events appraised as personally threatening, unjust, and intentionally imposed generate primary negative emotions: moral outrage, anger, and feelings of vulnerability. Emotional violation represents this affective realization of betrayal. This heightened emotional arousal demands behavioral discharge, which manifests directly as retaliatory negative word-of-mouth.
2. The Cognitive Pathway (Cognitive Trust): Cognitive trust reflects the customer's rational confidence in a firm's competence, dependability, and systemic integrity. Data vulnerabilities dismantle these rational expectations. The collapse of cognitive trust eliminates the firm's psychological buffer, removing any perceived social obligation to protect its commercial interests and facilitating active negative communication to social peers.

The Retaliation Model

Drawing directly from Grégoire and Fisher (2006), customer retaliation is understood as an effort to “punish and cause harm to a firm for the damages it has caused.” In social exchange terms, an individual subjected to an unfair transaction will seek to restore equilibrium by imposing costs on the transgressor. When traditional legal or regulatory remedies are inaccessible or ineffective for individual consumers, interpersonal denigration and negative word-of-mouth become their primary available punitive instruments.

7. Validity

The validity of the NWOM-D scale has been demonstrated through rigorous measurement testing, including construct, convergent, discriminant, and criterion-related validity assessments within advanced marketing analytics environments.

Construct and Convergent Validity

In structural equation modeling (SEM) evaluations conducted by Martin et al. (2017), the three-item instrument exhibited high convergent validity. The standardized factor loadings ($lambda$) for each of the three items substantially exceeded the accepted psychometric threshold of $.707$, typically loading between $.86$ and $.95$ ($p < .001$). The Average Variance Extracted ($\mathrm{AVE}$) consistently surpasses $.75$, far exceeding the conservative $.50$ benchmark established by Fornell and Larcker (1981). These metrics confirm that the variance captured by the underlying latent construct is substantially greater than the variance attributable to measurement error.

Discriminant Validity

To establish that retaliatory NWOM is empirically distinct from other post-vulnerability consumer reactions, Martin et al. (2017) tested the scale against related behavioral intentions and psychological states:

• Switching Intentions: While switching focuses on severing the personal purchasing relationship, NWOM focuses on deterring others. Fornell-Larcker tests demonstrate that the shared variance (squared correlation, $\phi^2$) between NWOM and switching intentions is significantly lower than their individual AVE estimates.
• Direct Complaining: The correlation between private complaining to firm management and public-facing NWOM is moderate ($r \approx .35 – .48$), confirming that external disparagement functions as a distinct behavioral construct.
• Feelings of Violation and Cognitive Trust: Factor-analytic comparisons of alternative nested models (collapsing NWOM items into antecedent affective and cognitive constructs) resulted in significant deterioration of model fit ($\Delta \chi^2$ with $p < .001$), supporting the empirical independence of the behavioral intention from its cognitive and emotional drivers.

Criterion-Related and Predictive Validity

The predictive validity of the NWOM-D was confirmed through longitudinal and multi-method empirical testing:

• Behavioral Intentions to Longitudinal Action: Laboratory and field experiments demonstrated that elevated NWOM-D scores reliably predict downstream communication behaviors, including posting negative reviews online and participating in coordinated digital boycotts.
• Firm Financial Performance: Martin et al. (2017) combined consumer survey data with econometric event studies analyzing public announcements of corporate data leaks. Elevated customer retaliation and NWOM intentions were directly linked to negative abnormal stock returns and sustained declines in quarterly sales revenue growth.

8. Reliability

The reliability of the NWOM-D scale has been corroborated through both classical test theory metrics and modern structural equation modeling parameters across multiple independent samples.

Internal Consistency Reliability

The scale consistently exhibits exceptional internal consistency across diverse empirical studies:

• Cronbach’s Alpha ($\alpha$): Reported values for the scale routinely exceed $.90$. In the validation studies by Martin et al. (2017), the Cronbach's alpha for the three-item instrument was $.92$, reflecting strong inter-item correlations without excessive redundancy.
• Composite Reliability ($\mathrm{CR}$): Because Cronbach’s alpha can underestimate reliability under departures from tau-equivalence, Composite Reliability (Raykov's $rho$) was evaluated. The scale regularly yields $\mathrm{CR}$ estimates ranging between $.91$ and $.94$, exceeding the $.80$ threshold required for advanced structural analysis.
• Average Variance Extracted ($\mathrm{AVE}$): Values routinely exceed $.75$, indicating that over three-quarters of the variance in the indicator variables is directly accounted for by the underlying NWOM construct.

Measurement Stability Across Contexts

The reliability of the NWOM-D remains stable across varied operational contexts:

• Industry Sector Invariance: Reliability benchmarks remain stable across diverse industries, such as financial services ($\alpha = .93$), retail e-commerce ($\alpha = .91$), and digital social platforms ($\alpha = .90$).
• Experimental Manipulations: The scale maintains high reliability across varied experimental conditions, ranging from transparent opt-in data policies to severe, unannounced third-party data breaches.

9. Factor Analysis

The dimensionality of the NWOM-D scale has been evaluated using both Exploratory Factor Analysis (EFA) and Confirmatory Factor Analysis (CFA).

Exploratory Factor Analysis (EFA)

Initial principal components and maximum likelihood extractions on the three items demonstrate a clear single-factor solution. A single eigenvalue substantially exceeds unity (Eigenvalue $> 2.45$), accounting for over $80%$ of the total explained variance across the indicator items. Scree test scree plots show a sharp bend immediately following the first factor, with no secondary factors showing conceptual or mathematical viability.

Confirmatory Factor Analysis (CFA)

When evaluated within a multi-construct structural model including data vulnerability, transparency, regulatory control, trust, and switching behaviors, the measurement model for NWOM-D demonstrates exceptional fit to empirical data. Standardized factor loadings and fit statistics are summarized below:

Global Model Fit Indices

Within structural measurement models incorporating the full set of relational constructs, the global fit indices consistently satisfy rigorous criteria (Hu & Bentler, 1999):

• Comparative Fit Index (CFI): $ge .98$
• Tucker-Lewis Index (TLI): $ge .97$
• Root Mean Square Error of Approximation (RMSEA): $le .045$ ($90% \text{ CI } [.028, .061]$)
• Standardized Root Mean Square Residual (SRMR): $le .028$
• Chi-Square / Degrees of Freedom ($\chi^2 / df$): $le 2.10$

10. Instrument / Measurement Tool

The NWOM-D scale is structured as follows:

• Instrument Name: Negative WOM Intentions (Data Context) Scale (NWOM-D)
• Construct Measured: Consumer intentions to engage in retaliatory negative word-of-mouth following corporate data vulnerability or privacy breaches
• Theoretical Ancestry: Retaliation Scale (Grégoire & Fisher, 2006); adapted to data privacy by Martin, Borah, & Palmatier (2017)
• Test Format: Self-report psychometric questionnaire administered via paper-and-pencil or computerized survey platforms (e.g., Qualtrics, Confirmit)
• Item Count: 3 items
• Response Scale: 7-point Likert scale (1 = Strongly disagree to 7 = Strongly agree)
• Administration Time: Approximately 1 to 2 minutes
• Target Population: Consumers, commercial clients, digital platform users, and survey panelists subjected to varied corporate data governance practices
• Scoring Protocol: All three items are positively worded (reflecting high retaliatory intent). Items are unweighted and averaged to generate an overall negative word-of-mouth intention score ranging from 1.00 to 7.00:
  $$\text{NWOM-D Score} = \frac{\text{Item}_1 + \text{Item}_2 + \text{Item}_3}{3}$$
• Interpretation Guidelines:
  • 1.00 – 2.49 (Low Intentions): Minimal retaliatory threat; customer perceives acceptable privacy practices or harbors low moral indignation.
  • 2.50 – 4.50 (Moderate Intentions): Emerging dissatisfaction; indicates latent vulnerability where minor subsequent triggers could prompt active disparagement.
  • 4.51 – 7.00 (High / Severe Intentions): Active retaliatory posture; represents severe reputation and commercial deterrence risk for the focal firm.

11. Permissions & Fee and Test Year

The NWOM-D scale was developed and published in 2017 in the Journal of Marketing (Martin, Borah, & Palmatier, 2017). The underlying retaliation items from which it was adapted were published in 2006 in the Journal of the Academy of Marketing Science (Grégoire & Fisher, 2006).

• Permissions & Academic Use: Under international scholarly fair-use conventions, the three survey items may be used freely by non-commercial researchers, academic institutions, students, and educators for scientific investigation, theoretical replication, and educational instruction, provided appropriate formal citation is given to Martin et al. (2017) and Grégoire and Fisher (2006).
• Commercial and Proprietary Licensing: Commercial organizations, enterprise consulting agencies, and for-profit market research firms wishing to integrate the items into commercial diagnostics, proprietary risk algorithms, or fee-based audit systems must consult the copyright holder, the American Marketing Association (AMA), or the authors regarding copyright permissions and licensing terms.
• Fee: $0 for non-profit academic research.

12. References

• 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
• Grégoire, Y., & Fisher, R. J. (2006). The effects of relationship quality on customer retaliation. Journal of the Academy of Marketing Science, 34(1), 31–44. https://doi.org/10.1177/0092070305279338
• 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, 6(1), 1–55. https://doi.org/10.1080/10705519909540118
• Lazarus, R. S., & Folkman, S. (1984). Stress, appraisal, and coping. Springer Publishing Company.
• Martin, K. D., Borah, A., & Palmatier, R. W. (2017). Data privacy: Effects on customer and firm performance. Journal of Marketing, 81(1), 36–58. https://doi.org/10.1509/jm.15.0497
• Rousseau, D. M. (1995). Psychological contracts in organizations: Understanding written and unwritten agreements. SAGE Publications. https://doi.org/10.4135/9781452231594

13. Items of the Scale