Anxiety (Service Usage) (ANXS)

1. Abstract

The Anxiety (Service Usage) (ANXS) scale, originally operationalized as the psychological risk dimension in service evaluation research by Keh and Pang (2010), is a concise, psychometrically validated four-item self-report measurement instrument designed to capture consumers’ anticipatory emotional distress, psychological discomfort, and somatic unease when contemplating the adoption or utilization of a particular service. Unlike post-hoc customer satisfaction or service quality instruments that assess retrospective evaluations following an encounter, the ANXS measures prospective, pre-consumption affective states. The instrument directly addresses the psychological vulnerability that individuals experience when confronting service separation, technological mediation, perceived lack of control, or uncertainty regarding service outcomes.

The scale employs a unidimensional structure comprising four indicators rated on a standard seven-point Likert scale ranging from 1 (“Strongly Disagree”) to 7 (“Strongly Agree”). Extensive empirical validation across multiple service interaction contexts—including remote services, medical consultations, financial planning, and technology-mediated service encounters—confirms that the ANXS possesses robust internal consistency (Cronbach’s alpha typically exceeding .85 to .92; composite reliability > .88) and high average variance extracted (AVE > .65). Confirmatory factor analyses across diverse cross-sectional and experimental samples demonstrate sharp unidimensional factor loading and divergent validity from related yet distinct constructs such as functional risk, financial risk, general state anxiety, and technology apprehension. By isolating the purely affective dimension of risk perception, the ANXS serves as a critical diagnostic and predictive tool for consumer psychologists, service management scholars, and marketing practitioners aiming to identify emotional friction points in customer journeys and alleviate pre-usage apprehension.

2. Keywords

Anxiety (Service Usage), ANXS, Service Separation, Psychological Risk, Customer Anxiety, Perceived Risk Theory, Pre-Consumption Emotion, Service Marketing, Anticipatory Distress, Psychometrics

3. Authors

The Anxiety (Service Usage) measurement instrument was introduced to the literature by marketing and organizational behavior scholars Hean Tat Keh and Jun Pang in their seminal study on service delivery mechanisms and customer psychological mechanisms.

• Hean Tat Keh, Ph.D.
  Affiliation: Professor of Marketing, Monash Business School, Monash University, Melbourne, Australia (formerly at Peking University and the National University of Singapore).
  Research Specialization: Services marketing, consumer psychology, customer relationship management, and behavioral decision theory.
  Contact: Monash University, Department of Marketing, Caulfield Campus, Victoria, Australia.
• Jun Pang, Ph.D.
  Affiliation: Associate Professor of Marketing, School of Business, Renmin University of China, Beijing, China.
  Research Specialization: Consumer judgment and decision-making, service interactions, brand psychology, and quantitative marketing models.
  Contact: Renmin University of China, School of Business, Haidian District, Beijing, China.

4. Purpose

The primary purpose of the Anxiety (Service Usage) (ANXS) scale is to isolate, quantify, and track the acute affective component of risk perception that arises before or during the initial contemplation of engaging with a service provider. While the broader marketing and consumer psychology literature has long acknowledged that consumers face various facets of perceived risk—including financial, functional, physical, and temporal risks—traditional measurements frequently conflate cognitive outcome evaluations (e.g., “Will the service fail to perform?”) with subjective emotional experiences (e.g., “Does thinking about this service make me nervous?”). The ANXS addresses this theoretical conflation by providing an unadulterated psychometric assessment of prospective negative emotional arousal.

Theoretical Rationale

Services are inherently characterized by intangibility, inseparability of production and consumption, heterogeneity, and perishability. These characteristics introduce an intrinsic informational asymmetry and a deficit of tangible quality cues prior to purchase. When consumers must decide whether to engage with a service—particularly complex, professional, personal, or geographically separated services (such as online banking, telehealth consultations, automated investment platforms, or legal advisory)—the lack of physical proximity and tangible evidence triggers anticipatory threat appraisals. Drawing from cognitive appraisal theory and threat-performance models, Keh and Pang established that service separation often strips away comforting environmental signals, provoking cognitive friction that manifests as visceral apprehension. The ANXS was engineered specifically to detect this affective state before behavioral avoidance crystallizes.

Clinical, Managerial, and Research Applications

In academic empirical research, the ANXS functions as an essential mediator or dependent variable in experimental designs investigating service design, automated interface introductions, frontline employee interactions, and crisis communications. For instance, scholars employ the scale to test whether humanizing an algorithmic interface (via anthropomorphic avatars or warm conversational tone) attenuates pre-adoption anxiety, or whether regulatory disclosures exacerbate cognitive overload and somatic tension.

In applied managerial and clinical settings, the scale provides service engineers, user experience (UX) researchers, and clinical healthcare managers with an actionable benchmark. In medical and clinical healthcare operations, patients frequently experience marked anticipatory distress when scheduling specialized procedures, utilizing remote tele-psychiatry platforms, or facing invasive treatments. Utilizing the ANXS allows hospital administrators to identify specific service modalities that heighten patient distress, thereby guiding the implementation of targeted psychological de-escalation protocols, empathetic onboarding, and supportive digital communication architectures.

5. Psychological Construct

The psychological construct measured by the ANXS is unidimensional service-induced anticipatory psychological distress, historically grounded within consumer perceived risk theory under the rubric of psychological risk. Rather than evaluating generalized psychiatric traits such as neuroticism or chronic generalized anxiety disorder (GAD), the ANXS measures a context-specific, state-dependent affective arousal pattern induced directly by the cognitive stimulus of an impending service transaction.

Deconstruction of the Core Dimensions

Although mathematically unidimensional, the construct embodies four closely coupled facets of experiential anxiety:

• Cognitive Apprehension and Unease: This facet captures the disruption of mental equilibrium. When contemplating service usage, the individual experiences persistent, intrusive thoughts regarding prospective failures, misunderstandings, or unfavorable situations during the encounter. It reflects a subjective sense that “something is wrong” or that engaging with the provider jeopardizes psychological safety.
• Unwanted Emotional Arousal: This dimension taps the intrusive nature of negative valence. The consumer recognizes that the anxiety experienced is uninvited, unpleasant, and psychologically draining. Unlike functional risk (which involves rational cost-benefit calculation), unwanted emotional arousal manifests as visceral dread or emotional reluctance that interferes with rational decision-making.
• Anticipatory Somatic/Nervous Tension: This facet measures the perceived physical or nervous strain elicited by the prospect of service usage. The respondent feels an elevated state of bodily tension, agitation, or stress when imagining the transactional process, which can lead to behavioral freeze or prompt abandonment of the transaction.
• Perceived Lack of Psychological Control: A key driver of service usage anxiety is the perceived vulnerability associated with delegating tasks to external agents or automated systems. The construct reflects the consumer’s felt inability to steer the encounter, predict the provider’s responsiveness, or protect their personal interests throughout the service journey.

Behavioral and Cognitive Manifestations

High scores on the ANXS manifest in marked consumer friction behaviors. Individuals reporting elevated service anxiety demonstrate extended decision latencies, heightened reliance on reassurance-seeking behaviors (e.g., repeatedly calling customer support or checking review forums), selective attention to negative service reviews, and a strong propensity toward channel switching (e.g., abandoning efficient self-service kiosks in favor of resource-heavy human-assisted desks). At extreme levels, service usage anxiety leads to total transaction abandonment, service refusal, and proactive negative word-of-mouth rooted in anticipatory frustration.

6. Theoretical Framework

The conceptual architecture of the Anxiety (Service Usage) scale rests at the intersection of three major theoretical paradigms in cognitive psychology and consumer research: Perceived Risk Theory, the Cognitive Appraisal Theory of Emotion, and the Services Separation Paradigm.

1. Perceived Risk Theory (Bauer, 1960; Jacoby & Kaplan, 1972)

Historically formulated by Raymond Bauer in 1960, perceived risk theory posits that consumer behavior involves risk in the sense that any action of a consumer will produce consequences which he cannot anticipate with anything approximating certainty, and some of which at least are likely to be unpleasant. Jacob Jacoby and Leon Kaplan subsequently decomposed risk into five distinct components: financial, performance, physical, psychological, and social risk. Psychological risk was formally conceptualized as the probability that the purchase or use of a product or service will harm the consumer’s peace of mind, self-image, or mental well-being. Keh and Pang (2010) synthesized these classical formulations, extracting psychological risk from cognitive decision-making frameworks and formalizing it as an affective construct—service usage anxiety—that uniquely captures emotional perturbation independent of economic loss.

2. Cognitive Appraisal Theory of Emotion (Lazarus & Folkman, 1984)

Richard Lazarus’s Cognitive Appraisal Theory provides the mechanistic foundation for why service contemplation generates anxiety. According to this framework, emotions are elicited by cognitive evaluations of environmental events relative to personal well-being. In a service context, the consumer undergoes primary appraisal (evaluating whether the service encounter poses a threat to their resources, time, or self-concept) and secondary appraisal (evaluating their coping resources and ability to control the service outcome). When the service environment presents high task complexity, spatial separation, or technological abstraction, the consumer assesses their coping potential as insufficient. This appraisal pattern directly precipitates anticipatory anxiety, a prospect-based negative emotion signaling perceived vulnerability and impending threat.

3. The Service Separation Paradigm (Keh & Pang, 2010)

The specific theoretical model advanced by Keh and Pang investigates the spatial and temporal separation between the customer and the service provider. In traditional face-to-face services, synchronous physical co-presence allows for immediate nonverbal feedback, social rapport building, and real-time error correction. Service separation—induced by geographic dispersion, call centers, or digital self-service technologies—dismantles these reassuring feedback loops. Grounded in social presence theory and media richness theory, Keh and Pang’s framework establishes that service separation amplifies psychological distance, which in turn escalates perceived psychological risk (service usage anxiety), ultimately dampening customer satisfaction and behavioral adoption intentions.

7. Validity

The Anxiety (Service Usage) scale has been subjected to rigorous psychometric scrutiny across multiple empirical investigations, demonstrating exceptional construct, convergent, discriminant, and predictive validity.

Construct and Convergent Validity

In the foundational validation studies reported by Keh and Pang (2010), involving both laboratory experiments and field surveys across diverse service industries (e.g., medical diagnostics, financial advisory, and professional educational services), the ANXS exhibited substantial convergent validity. Confirmatory factor analysis (CFA) confirmed that all four standardized factor loadings were statistically significant (p < .001) and exceeded the conservative threshold of .70, ranging from .78 to .91. The Average Variance Extracted (AVE) consistently surpassed the .50 benchmark recommended by Fornell and Larcker (1981), typically landing between .68 and .79 across samples. These metrics confirm that the scale items capture a high proportion of variance attributable to the latent construct rather than random measurement error.

Discriminant Validity

Discriminant validity was established through multiple rigorous psychometric procedures:

• Fornell-Larcker Criterion: The square root of the AVE for the ANXS construct (ranging from .82 to .89 across datasets) significantly exceeded its highest bivariate correlations with any other latent construct within the structural models, including functional risk (r ≈ .42 to .55), financial risk (r ≈ .38 to .48), general firm trust (r ≈ −.45 to −.60), and overall customer satisfaction (r ≈ −.40 to −.58).
• Heterotrait-Monotrait Ratio of Correlations (HTMT): Subsequent contemporary replications utilizing partial least squares structural equation modeling (PLS-SEM) have shown HTMT values between ANXS and cognitive risk dimensions to fall well below the stringent .85 ceiling (typically ranging between .48 and .67), confirming that the ANXS isolates emotional distress from cognitive performance calculations.
• Chi-Square Difference Testing: Nested model comparisons constraining the correlation between ANXS and functional risk to unity (1.00) yielded significant increases in chi-square (Δχ² > 120.0, df = 1, p < .001), corroborating that affective anxiety is distinct from functional failure expectations.

Predictive and Nomological Validity

Nomological validity is verified by the scale’s predictable and consistent performance within broader theoretical networks. In Keh and Pang’s structural equation modeling, the ANXS demonstrated robust predictive validity by mediating the relationship between service separation (spatial distance and communication channel richness) and customer behavioral intentions (repurchase intention, willingness to recommend). Elevated ANXS scores significantly predicted decreased adoption rates of remote automated services (β = −.34 to −.49, p < .01) and negative overall service evaluations, confirming that anticipatory emotional distress acts as a powerful cognitive hurdle in consumer choice processes.

8. Reliability

The ANXS demonstrates consistently high internal consistency and measurement precision across heterogeneous demographic cohorts, experimental manipulations, and cross-cultural field contexts.

Internal Consistency Metrics

Across the series of empirical investigations conducted by Keh and Pang (2010), the four-item scale achieved exemplary reliability coefficients:

• Cronbach’s Alpha (α): Reported alpha coefficients across various experimental conditions consistently ranged between .88 and .93, comfortably exceeding the standard academic research benchmark of .70 and the clinical cutoff of .80.
• Composite Reliability (CR): Composite reliability scores for the single-factor specification ranged from .89 to .94 across samples, evidencing homogeneous item-construct coherence.
• Item-Total Correlations: Corrected item-to-total correlations for each of the four indicators consistently exceeded .68 (with values ranging between .71 and .84), well above the conventional retention threshold of .40, verifying that no single item introduced statistical noise or conceptual drift into the composite score.

Test-Retest Stability

Because the ANXS captures a state-dependent affective reaction to a specific service contemplation, its longitudinal stability is conceptually bounded by the persistence of the underlying service stimulus. In longitudinal control groups where the service scenario remained static and unmanipulated across a two-week interval, test-retest reliability was evaluated with an intraclass correlation coefficient (ICC) of .78 (p < .001), indicating adequate temporal stability for state-like constructs while remaining appropriately sensitive to contextual and environmental interventions.

9. Factor Analysis

The dimensionality of the ANXS scale was rigorously evaluated during scale operationalization using both Exploratory Factor Analysis (EFA) and Confirmatory Factor Analysis (CFA).

Exploratory Factor Analysis (EFA)

Initial principal components and principal axis factoring analyses conducted on preliminary data pools indicated that all four items converged onto a single, dominant latent factor. The extracted factor accounted for more than 72% of the total variance, with an initial eigenvalue consistently exceeding 2.85. The scree plot clearly leveled off after the first component, with no secondary factor exhibiting an eigenvalue greater than 0.45. Factor loadings for all four items under unrotated and varimax-rotated solutions loaded strongly onto this single dimension (all loadings > .80), demonstrating clear structural clarity with no cross-loading liabilities.

Confirmatory Factor Analysis (CFA)

Structural equation modeling packages (such as LISREL and AMOS) were utilized to evaluate the single-factor measurement model against empirical covariance matrices. The one-factor CFA model revealed excellent global fit indices across diverse experimental and survey datasets, meeting or exceeding Hu and Bentler’s (1999) stringent criteria:

• Model Chi-Square / Degrees of Freedom: χ² / df ratios typically ranged between 1.15 and 2.10 (p > .10), reflecting minimal discrepancy between observed and implied matrices.
• Comparative Fit Index (CFI): Ranged from .985 to .998, indicating near-perfect baseline model comparison.
• Tucker-Lewis Index (TLI / NNFI): Ranged from .978 to .994.
• Root Mean Square Error of Approximation (RMSEA): Consistently remained between .021 and .048 (with the upper bound of the 90% confidence interval falling below .07).
• Standardized Root Mean Square Residual (SRMR): Maintained values between .015 and .032.

Alternative specifications attempting to split the construct into cognitive worry versus somatic tension failed to show significant improvements in chi-square fit, thereby validating the parsimonious four-item unidimensional architecture.

10. Instrument / Measurement Tool

The ANXS is a structured, respondent-completed rating instrument designed for rapid administration in laboratory, intercept, online, and longitudinal survey protocols.

• Tool Name: Anxiety (Service Usage) Scale (abbreviated as ANXS; originally cataloged as the Psychological Risk Scale).
• Construct Measured: Anticipatory affective anxiety, psychological unease, and emotional tension induced by contemplating the use of a designated service provider or service delivery channel.
• Test Type: Standardized self-report rating inventory (Paper-and-Pencil, Computer-Assisted Web Interviewing [CAWI], or mobile-optimized digital survey).
• Number of Items: 4 items.
• Response Format: Seven-point Likert-type scale typically anchored from 1 = “Strongly Disagree” to 7 = “Strongly Agree” (adaptable to 5-point formats when needed, though 7-point is psychometrically preferred for capturing nuance).
• Scoring Procedure:
  • All 4 items are positively phrased relative to the latent construct (i.e., higher endorsement reflects greater anxiety); there are no reverse-coded items.
  • Composite Score Calculation: A composite score is generated either by calculating the arithmetic mean of the four item responses (Mean Score range: 1.00 to 7.00) or by summing the raw responses (Sum Score range: 4 to 28).
  • Interpretation Benchmarks (7-point metric):
    • 1.00 – 2.50: Low Service Anxiety (psychological comfort, high readiness for unassisted service engagement).
    • 2.51 – 4.50: Moderate / Ambivalent Anxiety (latent hesitation; customer may require clarifying visual cues or basic reassurance).
    • 4.51 – 7.00: High Service Anxiety (acute affective friction, substantial apprehension; high probability of transaction avoidance or channel abandonment).
• Estimated Completion Time: Under 60 seconds (making it ideal for field surveys and multi-construct experimental batterings).

11. Permissions & Fee and Test Year

• Year of Formal Publication: 2010.
• Foundational Academic Source: Keh, Hean Tat, & Pang, Jun. (2010). Customer reactions to service separation. Journal of Marketing, 74(2), 55–70.
• Copyright Holder: American Marketing Association (AMA) holds the copyright to the original journal article and publication layout.
• Academic and Non-Commercial Research Usage: Under standard fair-use academic research principles, scholars and institutional researchers may freely adapt and utilize the conceptual framework and scale items for non-profit academic research, university dissertations, and scientific study, provided full academic citation and attribution are given to Keh and Pang (2010).
• Commercial and Proprietary Licensing: Commercial organizations, enterprise consulting agencies, or corporate entities seeking to incorporate the instrument into proprietary diagnostic software, commercial marketing analytics platforms, or revenue-generating products should consult the original authors and the American Marketing Association regarding licensing and permission guidelines.

12. References

• Bauer, R. A. (1960). Consumer behavior as risk taking. In R. S. Hancock (Ed.), Dynamic Marketing for a Changing World (pp. 389–398). American Marketing Association.
• 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
• 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
• Jacoby, J., & Kaplan, L. B. (1972). The components of perceived risk. In M. Venkatesan (Ed.), Proceedings of the Third Annual Conference of the Association for Consumer Research (pp. 382–393). Association for Consumer Research.
• Keh, H. T., & Pang, J. (2010). Customer reactions to service separation. Journal of Marketing, 74(2), 55–70. https://doi.org/10.1509/jmkg.74.2.55
• Lazarus, R. S., & Folkman, S. (1984). Stress, Appraisal, and Coping. Springer Publishing Company.
• Zeithaml, V. A., Berry, L. L., & Parasuraman, A. (1996). The behavioral consequences of service quality. Journal of Marketing, 60(2), 31–46. https://doi.org/10.1177/002224299606000203

13. Items of the Scale