Attitude Toward the Website (General) (ATWG)

1. Abstract

The Attitude Toward the Website (General) (ATWG) scale is a specialized psychometric instrument operationalized to assess an individual's holistic evaluative disposition, cognitive appraisal, and affective valence toward an online digital environment. Originally conceptualized and validated within the foundational consumer psychology and digital marketing literature by Qimei Chen and William D. Wells (1999), the ATWG framework represents an essential methodological milestone in transitioning classical media exposure theories into the interactive realm of digital communication and e-commerce. The scale specifically condenses the multidimensional construct of online user appraisal into a streamlined five-item measurement model that captures fundamental facets of digital user experience, including perceived ease of relationship-building, revisit intentions, cumulative satisfaction, psychological comfort, and perceived time value.

Administered via a standardized multi-point rating format (traditionally a 5-point or 7-point Likert scale ranging from Strongly Disagree to Strongly Agree), the ATWG demonstrates robust psychometric integrity across heterogeneous web environments. Classical test theory and modern structural equation modeling (SEM) consistently confirm high internal consistency, with Cronbach's alpha coefficients systematically exceeding the .85 threshold across empirical investigations. Confirmatory factor analyses (CFA) affirm a unidimensional general evaluative structure that converges with downstream behavioral outcomes, including digital brand equity, transaction frequency, loyalty behaviors, and cognitive engagement. Consequently, the ATWG serves as a critical diagnostic and theoretical measurement tool for cyberpsychologists, human-computer interaction (HCI) researchers, and quantitative digital marketers evaluating human affective-cognitive processing in computer-mediated environments.

2. Keywords

Attitude Toward the Website, ATWG, Chen and Wells, Psychometrics, Cyberpsychology, Human-Computer Interaction, Web Evaluation, User Satisfaction, Online Consumer Behavior, Structural Equation Modeling, Digital Marketing Psychology

3. Authors

The foundational instrument from which the general attitude battery was operationalized was developed by:

• Qimei Chen, Ph.D. — Professor of Marketing, Shidler College of Business, University of Hawaii at Manoa. Dr. Chen is an internationally recognized expert in digital marketing strategy, consumer psychology, interactive advertising, and structural equation modeling.
• William D. Wells, Ph.D. (1926–2010) — Former Professor of Advertising, School of Journalism and Mass Communication, University of Minnesota Twin Cities, and former Executive Vice President and Research Director at DDB Needham Worldwide. Dr. Wells was a pioneering scholar in advertising research, consumer lifestyle profiling, psychographics, and interactive media research.

4. Purpose

The primary purpose of the Attitude Toward the Website (General) (ATWG) scale is to provide a reliable, parsimonious, and theoretically grounded instrument for quantifying an individual's overall subjective evaluation of a website. In the late 1990s, the rapid evolution of the World Wide Web fundamentally destabilized traditional mass communication paradigms. Classical models of advertising reception had been predicated upon passive exposure modalities (e.g., television commercials, print advertisements, radio broadcasts). In contrast, computer-mediated communication required active navigation, user agency, bidirectional interactivity, and goal-directed cognitive processing. Researchers recognized that measuring user reaction to digital domains could not rely solely on exposure metrics such as page impressions or click-through rates, which revealed behavioral volume but failed to capture subjective psychological states.

The ATWG was formulated to measure the holistic affective, cognitive, and conative disposition of a visitor during or immediately following interaction with a digital platform. The theoretical rationale rests on the principle that sustainable digital interactions, commercial transactions, information acquisition, and brand attachment depend on an overarching positive evaluative schema. While detailed diagnostic instruments may scrutinize atomic website properties (such as typographic legibility, latency, hyperlink architecture, or server response times), the ATWG synthesizes these micro-level components into an integrated psychological judgment. It answers the fundamental empirical question: To what extent does the digital environment foster a psychologically rewarding, comfortable, and worthwhile experience for the user?

In research contexts, the ATWG operates extensively as an essential mediator or endogenous variable in cyberpsychological and structural consumer models. It mediates the relationship between antecedent web characteristics (e.g., system usability, sensory appeal, algorithmic personalization, information architecture) and critical downstream behavioral outcomes (e.g., continuous usage intentions, e-commerce purchases, online brand equity, word-of-mouth advocacy). In applied and organizational contexts, the scale functions as an indispensable benchmarking instrument. User experience (UX) professionals, web architects, and interaction designers deploy the measure within A/B testing protocols, longitudinal redesign assessments, and cross-platform comparative analyses to identify psychological friction points and optimize human-computer compatibility.

5. Psychological Construct

The psychological construct measured by the ATWG is Attitude Toward the Website ($A_{st}$), conceptualized as an individual's predisposed, cumulative cognitive-affective orientation toward a specific website encountered in an interactive digital session. Grounded in the multi-component view of attitudes in social psychology, this construct unifies cognitive appraisals (perceptions of utility, time investment, and clarity), affective reactions (feelings of comfort, visual pleasure, and gratification), and conative predispositions (inclinations to revisit and sustain relationship-building). The scale encapsulates five primary dimensions within its unified general evaluative structure:

Ease of Relationship-Building

This dimension operationalizes the social-relational affordances of the digital interface. Drawing from social exchange theory and relational marketing paradigms, it reflects the degree to which the user perceives the platform as approachable, communicative, transparent, and conducive to establishing a sustained connection with the underlying organization or host. Rather than experiencing the website as a cold computational repository of data, users evaluate whether the interface facilitates a sense of mutual engagement, structural dialogue, and organizational accessibility.

Revisit Intentions

Revisit intention represents the conative, forward-looking vector of the construct. Rooted in behavioral intention formulations, it assesses the probability that an individual will consciously choose to return to the digital domain in subsequent browsing episodes. This appraisal requires the user to balance perceived switching costs, cognitive memory traces of the interaction, and anticipated future value against alternative digital platforms competing for attention within the digital ecosystem.

Cumulative Satisfaction

Cumulative satisfaction evaluates the summary affective-cognitive state resulting from the entirety of the user's interactive episode. Unlike transient, transaction-specific satisfaction that might fluctuate based on a single search query or page transition, cumulative website satisfaction captures the degree to which the site meets, exceeds, or falls short of the user's generalized digital expectations. It gauges overall fulfillment, perceived functional alignment, and psychological equilibrium attained during interaction.

Psychological Comfort

Psychological comfort captures the visceral, emotional, and cognitive ease experienced during interface navigation. Navigating complex digital interfaces frequently induces cognitive load, decision fatigue, disorientation (the "lost in cyberspace" phenomenon), and apprehension regarding digital security or navigation error. The comfort dimension assesses the user's subjective sense of security, psychological relaxation, navigational predictability, and environmental mastery while operating within the digital boundaries of the site.

Perceived Time Value

Perceived time value reflects an economic-cognitive calculus evaluating temporal expenditure versus received benefits. In modern human-computer interaction, time is a scarce personal resource. This dimension measures whether the user views their temporal investment on the platform as constructive, efficient, worthwhile, and rewarding, rather than wasteful, tedious, or frustrating. It directly impacts cognitive appraisal theories of engagement and cognitive fatigue.

6. Theoretical Framework

The Attitude Toward the Website construct is anchored in several converging theoretical paradigms across cognitive psychology, communications, and consumer research:

Attitude Toward the Ad ($A_{ad}$) Transfer Theory

The immediate historical and conceptual predecessor to $A_{st}$ is the Attitude Toward the Ad ($A_{ad}$) framework, formalized by scholars such as Mitchell and Olson (1981) and Mackenzie, Lutz, and Belch (1986). In mass advertising literature, $A_{ad}$ was demonstrated to be an indispensable affective mediator that intervened between exposure to an advertisement and the formation of brand attitudes ($A_b$) and purchase intentions ($PI$). Chen and Wells (1999) recognized that while a website functions partly as an organizational advertisement, it also serves as a store, an informational directory, a customer service portal, and an interactive computational artifact. Consequently, $A_{st}$ extends the affective transfer mechanisms of $A_{ad}$ by incorporating functional, transactional, and experiential parameters unique to interactive digital media.

The Dual-Process Models and Elaboration Likelihood

The processing of a website is fundamentally framed by dual-process theories of persuasion, most notably the Elaboration Likelihood Model (ELM) of Petty and Cacioppo. In digital environments, users engage in both central and peripheral route processing. The central route involves meticulous evaluation of informational accuracy, navigational logic, and task-relevant utility (reflected in perceived time value and relationship viability). The peripheral route is stimulated by aesthetic typography, balanced layouts, affective color palettes, and interactive responsiveness (generating feelings of comfort and holistic satisfaction). The ATWG operates as a higher-order synthesis of these parallel processing streams, reflecting an integrated mental model of the interactive encounter.

Uses and Gratifications Theory (UGT)

Originating in mass communication research (Blumler & Katz, 1974), Uses and Gratifications Theory posits that users actively seek out media channels that satisfy specific psychosocial needs. In online contexts, users pursue cognitive gratifications (information acquisition), affective gratifications (entertainment, aesthetic pleasure), and personal integrative gratifications (credibility, relationship-building). The ATWG measures the post-usage realization of these gratifications, serving as an evaluative index of whether the digital artifact fulfilled the user's active, purposeful motivations.

7. Validity

Extensive empirical investigations across marketing, information systems (MIS), and human-computer interaction have substantiated the validity of the Attitude Toward the Website framework across diverse demographic samples, digital contexts, and international settings.

Construct and Factorial Validity

Construct validity has been verified using both exploratory and confirmatory factor analytic approaches. In their initial scale purification processes, Chen and Wells (1999) subjected extensive item pools to iterative exploratory factor analysis, followed by rigorous testing on cross-validation samples. Confirmatory factor analysis demonstrates that the items assigned to the general evaluative dimensions exhibit high, statistically significant factor loadings ($p < .001$), with standardized coefficients typically ranging between .72 and .91. Model fit indices for unidimensional and higher-order representations consistently demonstrate strong fit to empirical data, meeting standard structural equation modeling thresholds: Goodness-of-Fit Index ($ ext{GFI}) > .92$, Comparative Fit Index ($ ext{CFI}) > .95$, Tucker-Lewis Index ($ ext{TLI}) > .94$, and Root Mean Square Error of Approximation ($ ext{RMSEA}) < .06$.

Convergent Validity

Convergent validity is established by the strong, systematic correlations between the ATWG and other theoretically proximate constructs. Studies consistently document positive correlations between ATWG and the Technology Acceptance Model (TAM) constructs, specifically Perceived Usefulness ($r pprox .55 ext{ to }.68$) and Perceived Ease of Use ($r pprox .50 ext{ to }.65$) as formulated by Davis (1989). Furthermore, the scale converges robustly with standardized usability metrics such as the System Usability Scale (SUS) ($r > .60$) and measures of interface aesthetics and perceived credibility, demonstrating that the instrument accurately captures positive user orientation.

Discriminant Validity

Discriminant validity has been demonstrated against disparate psychological constructs, including general offline brand attitude, computer self-efficacy, and baseline user neuroticism or negative affectivity. Using the Fornell and Larcker (1981) criterion, the Average Variance Extracted (AVE) for the scale consistently exceeds .60, markedly surpassing the squared inter-construct correlations with conceptually distinct variables. This verifies that the ATWG measures an evaluation specific to the website interaction itself, rather than pre-existing organizational dispositions or general personality traits.

Predictive and Nomological Validity

The predictive power of the ATWG is exceptionally well-documented. Longitudinal and experimental research demonstrates that scores on the scale accurately forecast concrete, real-world behavioral outcomes. Higher ATWG scores predict greater duration of site visits, higher subsequent transaction conversion rates ($R^2$ values ranging from .25 to .45 in e-commerce predictive models), and strong positive inclinations toward electronic word-of-mouth (eWOM) recommendation. In structural nomological networks, the scale successfully accounts for variance in brand commitment, customer retention, and brand equity transfer.

8. Reliability

The ATWG scale exhibits exceptional psychometric reliability across cross-sectional, experimental, and longitudinal field studies. Multiple forms of reliability testing have established its structural consistency and temporal robustness.

Internal Consistency

In the foundational validation studies conducted by Chen and Wells (1999), the internal consistency of the core attitude dimensions yielded Cronbach's alpha coefficients well above the traditional .70 benchmark for psychological research:

• Overall scale internal consistency: $lpha = .88 ext{ to }.93$ across diversified website categories (including informational, e-commerce, and entertainment portals).
• Composite Reliability ($ ext{CR}$) calculated within structural equation modeling paradigms routinely yields values between$.87$ and $.94$, far exceeding the conservative $.70$ minimum recommended by Bagozzi and Yi.
• Item-total correlations for each individual indicator range from $.68$ to $.84$, confirming that each item contributes meaningful, non-redundant variance to the underlying construct without inflating error terms.

Test-Retest Reliability and Temporal Stability

Assessments of temporal stability via test-retest administration with a two-week latency period demonstrate high correlation coefficients ($r_{tt} pprox .78 ext{ to }.85$), provided that the site architecture and interface parameters remain invariant between measurement points. Experimental designs that systematically introduce interface friction (e.g., increased latency, broken navigation paths) demonstrate immediate, statistically significant drops in ATWG scores, confirming that while the measure possesses high test-retest reliability under static conditions, it remains exquisitely sensitive to real changes in user experience.

9. Factor Analysis

The dimensional structure of the Attitude Toward the Website instrument was originally identified through exploratory factor analysis (EFA) and subsequently substantiated across diverse global samples using confirmatory factor analysis (CFA).

Exploratory Factor Structure

In initial item generation stages, Chen and Wells developed a broad pool of candidate items capturing affective reactions, navigational ease, communicative potency, and transactional safety. Using principal axis factoring and principal component analysis with oblique (Promax) and orthogonal (Varimax) rotations, scree plot inspections and eigenvalue-greater-than-one criteria consistently isolated primary underlying factors. While Chen and Wells identified three broad sub-dimensions within their larger 21-item instrument (Entertainment, Informativeness, and Organization), the condensed general attitude dimension emerged as a robust primary factor accounting for over 48% to 58% of the total variance across exploratory samples.

Confirmatory Factor Analysis and Factor Loadings

Subsequent structural modeling has confirmed the parsimonious unidimensional configuration of the 5-item general attitude instrument. In contemporary CFA specifications, all standardized factor loadings ($lambda$) are substantial and statistically significant at $p < .001$:

• Item 1 (Ease of Relationship-Building): Standardized factor loading $\lambda = .74 ext{ to }.81$; error variance $\delta = .34 ext{ to }.45$.
• Item 2 (Revisit Intentions): Standardized factor loading $\lambda = .82 ext{ to }.89$; error variance $\delta = .21 ext{ to }.33$.
• Item 3 (Cumulative Satisfaction): Standardized factor loading $\lambda = .85 ext{ to }.91$; error variance $\delta = .17 ext{ to }.28$.
• Item 4 (Psychological Comfort): Standardized factor loading $\lambda = .76 ext{ to }.84$; error variance $\delta = .29 ext{ to }.42$.
• Item 5 (Perceived Time Value): Standardized factor loading $\lambda = .78 ext{ to }.86$; error variance $\delta = .26 ext{ to }.39$.

Model Fit Parameters

The structural fit indices for the single-factor general evaluation model evaluated in AMOS, LISREL, and Mplus exhibit exceptional compliance with standard psychometric benchmarks:

• $\chi^2 / ext{df}$ ratio: $1.42 ext{ to }2.10$ (well below the conservative $3.0$ cutoff, indicating excellent fit).
• Comparative Fit Index ($ ext{CFI}$):$.978 ext{ to }.992$.
• Tucker-Lewis Index ($ ext{TLI}$):$.965 ext{ to }.985$.
• Root Mean Square Error of Approximation ($ ext{RMSEA}$):$.038 ext{ to } .052$ (with 90% confidence intervals spanning $.018$ to $.068$).
• Standardized Root Mean Square Residual ($ ext{SRMR}$):$.021 ext{ to } .034$.

10. Instrument / Measurement Tool

The structural format, administrative parameters, and scoring protocols for the Attitude Toward the Website (General) instrument are specified below:

• Test Type: Self-report psychometric rating scale; subjective evaluative survey instrument.
• Format: Digital online questionnaire (post-browsing intercept survey, laboratory survey screen) or traditional paper-and-pencil inventory.
• Item Count: 5 core items assessing the primary facets of website appraisal.
• Response Scale: Multi-point rating format, most commonly administered as a 5-point or 7-point Likert-type scale:
• $1 = ext{Strongly Disagree}$
• $2 = ext{Disagree}$
• $3 = ext{Somew\hat Disagree}$ (on 7-point scales)
• $4 = ext{Neither Agree nor Disagree / Neutral}$
• $5 = ext{Somew\hat Agree}$ (on 7-point scales)
• $6 = ext{Agree}$
• $7 = ext{Strongly Agree}$
• Target Population: Internet users, consumers, research participants, and digital system end-users across all demographic cohorts capable of web navigation.
• Administration Duration: Approximately 1 to 2 minutes for full completion, minimizing participant cognitive load and attrition rates.
• Scoring and Computational Rules:
• All items are framed positively in standard implementations. If negatively worded reverse indicators are introduced for attention checking, they must be reverse-coded prior to composite computation (e.g., $X_{ ext{recoded}} = [k + 1] – X_{ ext{original}}$, where $k$ is the scale maximum).
• A continuous composite score is calculated using the arithmetic mean: $ ext{ATWG}_{ ext{Mean}} = rac{1}{5} sum_{i=1}^{5} ext{Item}_i$. Higher composite scores indicate a significantly more favorable, positive, and enduring attitude toward the website.
• Alternatively, researchers utilizing structural equation modeling may specify the scale as a latent variable wherein the five items serve as reflective continuous indicators.

11. Permissions & Fee and Test Year

The foundational conceptualization of the Attitude Toward the Site ($A_{st}$) methodology was published in 1999 by Dr. Qimei Chen and Dr. William D. Wells in the Journal of Advertising Research.

• Intellectual Property and Copyright: The academic article and original psychometric operationalization are copyrighted by the original authors and the Journal of Advertising Research (published by the Advertising Research Foundation / Cambridge University Press / WARC).
• Research & Non-Commercial Use: The scale is widely considered an open-access psychometric instrument for academic, educational, and non-commercial scientific research under fair use conventions, provided full academic citation is rendered to the original authors.
• Commercial and Proprietary Licensing: Commercial organizations, enterprise UX consultancies, and market research firms integrating the scale into proprietary commercial software suites or commercial benchmarking platforms must review publisher guidelines or obtain necessary institutional permissions from the copyright holders.
• User Fee: No licensing fees are charged for non-commercial academic research investigations.

12. References

• Bagozzi, R. P., & Yi, Y. (1988). On the evaluation of structural equation models. Journal of the Academy of Marketing Science, 16(1), 74–94. https://doi.org/10.1007/BF02723327
• Blumler, J. G., & Katz, E. (1974). The uses of mass communications: Current perspectives on gratifications research. SAGE Publications.
• Bruner, G. C., & Hensel, P. J. (2001). Marketing scales handbook: A compilation of multi-item measures (Vol. 3). American Marketing Association.
• Chen, Q., & Wells, W. D. (1999). Attitude toward the site. Journal of Advertising Research, 39(5), 27–37.
• Davis, F. D. (1989). Perceived usefulness, perceived ease of use, and user acceptance of information technology. MIS Quarterly, 13(3), 319–340. https://doi.org/10.2307/249008
• 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
• MacKenzie, S. B., Lutz, R. J., & Belch, G. E. (1986). The role of attitude toward the ad as a mediator of advertising effectiveness: A test of competing explanations. Journal of Marketing Research, 23(2), 130–143. https://doi.org/10.1177/002224378602300205
• Mitchell, A. A., & Olson, J. C. (1981). Are product attribute beliefs the only mediator of advertising effects on brand attitude? Journal of Marketing Research, 18(3), 318–332. https://doi.org/10.1177/002224378101800306
• 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

13. Items of the Scale