Customer Experience Quality (Affective)

Abstract

The Customer Experience Quality (Affective) scale is a concise, psychometrically validated measurement instrument developed to evaluate an observer's or respondent's assessment of the emotional and affective outcomes derived from a service interaction. Originally operationalized in empirical work on human-computer interaction and service marketing by Liu-Thompkins, Okazaki, and Li (2022) within the Journal of the Academy of Marketing Science, the scale investigates how service encounters—specifically those involving artificial intelligence (AI) and frontline human agents—influence the emotional valence, affective relief, and overall mood state of customers. Comprising three unidimensional items rated on a standard 7-point Likert scale ranging from 1 ("Strongly disagree") to 7 ("Strongly agree"), the instrument measures the degree to which an observer perceives that a service experience left a customer feeling relieved, pleasant, and in an elevated psychological mood. Psychometric evaluations across experimental and pilot investigations demonstrate robust internal consistency (Cronbach's alpha typically exceeding .85), high composite reliability, and clear evidence of convergent, discriminant, and nomological validity within affective computing, customer relationship management (CRM), and services marketing literature. By capturing transient and post-recovery affective states, the scale provides researchers and practitioners with an efficient, minimally burdensome tool to measure emotional shifts in both direct consumer evaluations and third-party experimental designs assessing artificial empathy and service recovery excellence.

Keywords

Affective Customer Experience, Customer Experience Quality, Artificial Empathy, Service Encounter, Emotional Valence, Mood Elevation, Service Recovery, Conversational Agents, Human-AI Interaction, Psychometrics, Consumer Behavior, Affective Computing

Authors

The Customer Experience Quality (Affective) instrument was developed and validated by a multidisciplinary team of marketing and communication scholars examining the frontier of emerging technologies, consumer psychology, and artificial intelligence:

• Yuping Liu-Thompkins, Ph.D.: Professor of Marketing and Director of the CRM and Customer Analytics Center at the Strome College of Business, Old Dominion University, Norfolk, Virginia, USA. Her research focuses on customer loyalty, consumer engagement programs, digital marketing strategies, and consumer-technology interactions.
• Shintaro Okazaki, Ph.D.: Professor of Marketing at King's Business School, King's College London, United Kingdom. An internationally recognized expert in international advertising, digital media, mobile marketing, and consumer psychology.
• Hairong Li, Ph.D.: Professor of Advertising in the Department of Advertising and Public Relations, College of Communication Arts and Sciences at Michigan State University, East Lansing, Michigan, USA. His scholarship investigates consumer behavior in digital environments, interactive advertising, and artificial intelligence applications in marketing communications.

Purpose

Modern customer experience (CX) research conceptualizes the consumer journey as a multidimensional construct encompassing cognitive, affective, behavioral, sensory, and social dimensions. Within service encounters—especially those marked by service failures, inquiries, or automated resolutions—the affective dimension is paramount. The primary purpose of the Customer Experience Quality (Affective) scale is to provide a brief, methodologically sound psychometric index capable of measuring observer-reported or self-reported affective states following a frontline service event.

In contemporary service contexts, frontline service delivery is increasingly mediated by artificial intelligence, including conversational chatbots, virtual assistants, and automated recommendation engines. While automated systems demonstrate high levels of functional and cognitive competence (e.g., speed of query resolution, technical accuracy), they historically struggle to demonstrate authentic emotional intelligence or empathy. Liu-Thompkins et al. (2022) introduced this measure specifically to capture the downstream emotional consequences of "artificial empathy"—defined as the ability of non-human agents to detect, process, and respond adaptively to human emotional cues.

From an applied and experimental standpoint, measuring affective experience quality serves several crucial functions:

• Evaluating AI vs. Human Touchpoints: It allows researchers to quantify whether artificial empathy can bridge the psychological distance between consumers and non-human agents, determining if artificial interactions can elicit affective outcomes comparable to warm human interactions.
• Service Failure and Recovery Analysis: The inclusion of relief-specific items makes this instrument exceptionally suited for assessing service recovery encounters, wherein the primary objective is to alleviate consumer anxiety, frustration, or cognitive distress.
• Observer-Based and Third-Party Experimental Designs: Many experimental paradigms rely on third-party evaluations of recorded transcripts, video encounters, or simulated interactions. This scale was explicitly structured to assess perceived affective outcomes of an observed customer, making it an invaluable tool for scenario-based laboratory experiments and vignette methodologies.
• Minimizing Survey Burden: With only three items, the scale minimizes common method bias driven by respondent fatigue in longitudinal, diary, or multi-stage experimental studies while preserving strict psychometric fidelity.

Psychological Construct

The psychological construct captured by this instrument is Affective Customer Experience Quality. While general customer experience encompasses the holistic cognitive and physical journey across touchpoints, affective experience refers specifically to the subjective, valenced emotional states, feelings, and mood alterations induced by a focal service interaction.

Affective customer experience quality within this operationalization consists of three interrelated psychological facets:

1. Affective Relief and Tension Reduction

The first item of the instrument measures the perception that the customer felt relieved. Relief is a unique, post-goal achievement emotion that emerges when an expected or experienced negative outcome is successfully averted, resolved, or mitigated. In psychological literature, relief involves a dual process: the cessation of an unpleasant psychological state (such as anxiety, confusion, or anger) accompanied by a homeostatic return to positive equilibrium. Within service interactions—particularly customer support, technical assistance, or complaint handling—relief is a critical affective milestone. It indicates that the frontline system or agent removed an imminent threat to the customer's resources (e.g., time, money, psychological well-being).

2. Hedonic Pleasantness (Valence)

The second item captures whether the interaction left the customer feeling pleasant. Pleasantness represents the fundamental core dimension of affective valence within Russell's circumplex model of affect. Unlike relief, which is often an outcome of negative tension dissipation, pleasantness reflects intrinsic positive hedonic tone. When a service encounter is perceived as pleasant, it signals that the communicative style, conversational tone, and interpersonal dynamics generated positive sensory or emotional stimulation. In human-AI interactions, high pleasantness indicates that the interaction felt harmonious, socially acceptable, and psychologically comforting rather than cold, robotic, or abrasive.

3. Mood Elevation and Affective Spillover

The third item addresses whether the service experience put the customer in a better mood. Mood differs from discrete emotions (such as momentary anger or joy) in that moods are more diffuse, lower in immediate intensity, longer-lasting, and less tied to a single cognitive target. Improving an individual's general mood indicates successful affective spillover: the positive consequences of the interaction extend beyond the immediate problem-solving context into the customer's ongoing emotional state. Mood elevation is a robust predictor of subsequent pro-social behaviors, brand forgiveness, positive word-of-mouth (WOM), and long-term customer retention.

Theoretical Framework

The design and operationalization of the Customer Experience Quality (Affective) scale are grounded in several prominent psychological and marketing frameworks, notably Cognitive Appraisal Theory, the Computers Are Social Actors (CASA) Paradigm, and the Broaden-and-Build Theory of Positive Emotions.

Cognitive Appraisal Theory

According to Cognitive Appraisal Theory, emotions are generated through subjective cognitive evaluations of an environmental event relative to personal well-being, goals, and coping potential. In a service encounter, customers continuously appraise the agent's behavior across dimensions of goal congruence, agency, and coping potential:

• Goal Congruence: When a service representative (human or AI) successfully solves a problem, the event is appraised as goal-congruent, initiating positive emotional transitions from distress to relief.
• Attribution of Agency: When an agent exhibits empathetic communication, respondents perceive high partner responsiveness, attributing intentional benevolence to the service provider.

The scale directly captures the downstream emotional readouts of these appraisal processes: positive valence (pleasantness), problem resolution (relief), and overall psychological restoration (improved mood).

The Computers Are Social Actors (CASA) Paradigm

Formulated by Clifford Nass and colleagues, the CASA paradigm posits that humans mindlessly apply human social rules and emotional heuristics to technological interfaces when those interfaces display human-like social cues (such as language, turn-taking, and emotional expressions). Liu-Thompkins et al. (2022) leveraged this paradigm to argue that when conversational AI exhibits artificial empathy—such as recognizing emotional distress and responding with comforting language—users react psychologically as if interacting with an empathetic human. The affective customer experience scale functions as an empirical gauge of CASA effects, testing whether emotional responses to synthetic social actors match those elicited by real human interlocutors.

Broaden-and-Build Theory of Positive Emotions

Barbara Fredrickson's Broaden-and-Build Theory asserts that positive affective states (pleasantness, contentment, relief) broaden an individual's momentary thought-action repertoires and build enduring physical, intellectual, and social resources. In commercial and interpersonal environments, eliciting positive affect widens consumer cognitive flexibility, lowers skepticism, enhances openness to recommendations, and strengthens relational bonding with the provider organization.

Validity

The validity of the Customer Experience Quality (Affective) scale has been established through rigorous construct, convergent, discriminant, and criterion-related empirical testing across diverse experimental settings.

Construct and Convergent Validity

Construct validity is evidenced by the scale's ability to systematically capture the theoretical spectrum of affective customer experience. In confirmatory factor models, all three items load uniformly high on a single underlying latent factor, with standardized factor loadings consistently exceeding .80. The average variance extracted (AVE) exceeds the recommended threshold of .50 (typically exceeding .70), demonstrating that the majority of variance in the observed items is directly attributable to the latent affective construct rather than measurement error.

Discriminant Validity

In empirical testing conducted by Liu-Thompkins et al. (2022), affective customer experience quality demonstrated clear empirical divergence from related but conceptually distinct constructs, such as:

• Social Customer Experience Quality: Pertaining to perceived social connection, warmth, and interpersonal communion.
• Functional Service Quality: Pertaining to promptness, accuracy, and operational competence.
• Perceived Artificial Empathy: Pertaining to the perceived ability of the AI to recognize and express empathy.

Using the Fornell-Larcker criterion and the Heterotrait-Monotrait (HTMT) ratio of correlations, the square root of the AVE for affective customer experience systematically exceeded its cross-construct correlations with other service evaluation dimensions, and HTMT ratios remained well below the conservative .85 threshold.

Predictive and Nomological Validity

Nomological validity is verified by the instrument's documented relationships with critical antecedent and outcome variables. The scale functions as an influential mediator between front-end service characteristics (e.g., empathetic language, responsiveness, conversational human-likeness) and macro-level behavioral outcomes, including customer satisfaction, brand loyalty, repurchase intention, and willingness to recommend. Furthermore, in experimental vignette studies involving artificial intelligence versus human service agents, the instrument showed high sensitivity in detecting subtle variations in emotional outcomes across different levels of interaction complexity.

Reliability

The reliability of the Customer Experience Quality (Affective) instrument has been evaluated through multiple indices of internal consistency and scale precision:

Internal Consistency Metrics

• Cronbach's Alpha (α): Across empirical deployments, the three-item instrument exhibits excellent internal consistency reliability, consistently generating Cronbach's α coefficients between .87 and .94. These values well exceed the standard academic benchmark of .70 recommended by Nunnally and Bernstein for established scales.
• Composite Reliability (CR): Structural equation modeling (SEM) evaluations show composite reliability coefficients ranging from .88 to .95, confirming that the indicators are cohesive and accurately reflect the underlying affective customer experience construct.
• Average Variance Extracted (AVE): The AVE scores routinely hover above .72, indicating that more than 72% of the variance captured by the indicators is direct construct variance.

Item Homogeneity and Stability

Inter-item correlations among the three items (relieved, pleasant, better mood) typically range between .65 and .82, demonstrating an optimal balance between item cohesion and non-redundancy. Corrected item-total correlations uniformly exceed .70, confirming that each statement contributes significantly to the measurement domain. Given that experimental studies repeatedly reproduce these reliability statistics across varying participant samples (e.g., MTurk, Prolific, undergraduate panels, and commercial customer panels), the scale demonstrates robust external stability.

Factor Analysis

Empirical evaluations using exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) confirm that the Customer Experience Quality (Affective) scale is strictly unidimensional.

Exploratory Factor Analysis (EFA)

When subjected to principal component analysis or maximum likelihood factor extraction, the three items consistently load onto a single dominant factor:

• Eigenvalue Criterion: The primary factor yields an initial eigenvalue substantially exceeding 2.0 (often between 2.25 and 2.60), accounting for 75% to 86% of the total variance across the indicators.
• Scree Plot: The scree plot displays a steep downward break after the first factor, with second and third components yielding eigenvalues far below the Kaiser-Guttman threshold of 1.0.
• Factor Loadings: Unrotated factor loadings for all three items consistently exceed .85, indicating strong shared variance.

Confirmatory Factor Analysis (CFA)

In structural equation modeling frameworks, a single-factor CFA model demonstrates exceptional fit to empirical data. Because a three-indicator model is mathematically just-identified (zero degrees of freedom), fit statistics are evaluated within multi-factor measurement models containing accompanying constructs (e.g., perceived artificial empathy, functional service quality, overall loyalty). Representative fit indices from these models demonstrate superior structural alignment:

• Comparative Fit Index (CFI): ≥ .98
• Tucker-Lewis Index (TLI): ≥ .97
• Root Mean Square Error of Approximation (RMSEA): ≤ .05 (90% CI [.00, .08])
• Standardized Root Mean Square Residual (SRMR): ≤ .03

Standardized parameter loadings (λ) for the three individual items typically range as follows:

• Item 1 ("relieved"): λ = .82 – .89
• Item 2 ("pleasant"): λ = .88 – .94
• Item 3 ("better mood"): λ = .85 – .92

No error covariances between items are necessary to achieve optimal fit, affirming the clean unidimensional property of the measurement model.

Instrument / Measurement Tool

The administrative profile of the Customer Experience Quality (Affective) scale is detailed below:

• Test Type: Psychometric rating scale; self-report or third-party observer evaluation.
• Construct Assessed: Affective customer experience quality (emotional valence, relief, mood elevation).
• Target Population: Consumers evaluating actual service encounters, or research participants evaluating service scenarios, conversational AI dialogues, or recorded frontline interactions.
• Administration Format: Paper-and-pencil, computer-assisted self-interview (CASI), or online web surveys (e.g., Qualtrics, Gorilla, CloudResearch).
• Number of Items: 3 items.
• Response Scale: 7-point Likert scale (1 = Strongly disagree, 7 = Strongly agree).
• Time Required for Administration: Approximately 30 to 60 seconds, minimizing participant attrition and survey fatigue.
• Scoring and Aggregation Guidelines:
  • All three items are framed in a positive affective direction; there are no reverse-coded items.
  • An overall composite score is calculated by computing the arithmetic mean of all three items.
  • Higher mean values indicate a stronger perceived affective customer experience (range: 1.0 to 7.0).
  • Alternatively, researchers utilizing structural equation modeling may specify the three items as reflective indicators of a single latent variable.

Permissions & Fee and Test Year

The Customer Experience Quality (Affective) scale was published in 2022 by Yuping Liu-Thompkins, Shintaro Okazaki, and Hairong Li in the Journal of the Academy of Marketing Science. The intellectual property is held by the authors and the publisher (Springer Nature). Under standard academic fair-use guidelines, the scale items may be utilized without licensing fees for non-commercial academic research, empirical theses, and educational purposes, provided that appropriate scholarly attribution is given to the original 2022 article. Commercial organizations seeking to integrate the instrument into proprietary diagnostic software or commercial customer tracking platforms should consult Springer Nature's permissions portal or contact the corresponding author.

References

• Fredrickson, B. L. (2001). The role of positive emotions in positive psychology: The broaden-and-build theory of positive emotions. American Psychologist, 56(3), 218–226. https://doi.org/10.1037/0003-066X.56.3.218
• Lazarus, R. S. (1991). Emotion and Adaptation. Oxford University Press.
• Liu-Thompkins, Y., Okazaki, S., & Li, H. (2022). Artificial empathy in marketing interactions: Bridging the human-AI gap in affective and social customer experience. Journal of the Academy of Marketing Science, 50(6), 1198–1218. https://doi.org/10.1007/s11747-022-00858-0
• Nass, C., Steuer, J., & Tauber, E. R. (1994). Computers are social actors. In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems (pp. 72–78). ACM. https://doi.org/10.1145/191666.191703
• Nunnally, J. C., & Bernstein, I. H. (1994). Psychometric Theory (3rd ed.). McGraw-Hill.
• Russell, J. A. (1980). A circumplex model of affect. Journal of Personality and Social Psychology, 39(6), 1161–1178. https://doi.org/10.1037/h0077714

Items of the Scale