Observational and Self-Report Divergences in Virtual Reality Shopping Evaluation: A Multi-Construct Mixed-Methods Study with Comparative AR Preference Analysis

Observational and Self-Report Divergences in Virtual Reality Shopping Evaluation: A Multi-Construct Mixed-Methods Study with Comparative AR Preference Analysis | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Observational and Self-Report Divergences in Virtual Reality Shopping Evaluation: A Multi-Construct Mixed-Methods Study with Comparative AR Preference Analysis Stephen Agada, Georgios Dafoulas This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9418492/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Virtual Reality (VR) and Augmented Reality (AR) represent two of the most commercially promising immersive modalities for digital retail, yet rigorous empirical evaluation of VR-based shopping applications (VRSA) alongside comparative user preference data for AR-based applications (ARSA) remains limited. This paper presents a mixed-methods study in which 64 participants completed a structured interaction session with a commercial VR shopping mall application, evaluated via a purpose-built eight-item behavioural observation rubric and a seven-construct post-experience survey. Participants subsequently completed a comparative preference instrument contrasting VRSA against a mobile ARSA they also experienced. Statistical analysis employed paired Wilcoxon signed-rank tests with Bonferroni correction, Mann-Whitney U moderation tests, Spearman rank-order correlations, and Cronbach alpha reliability assessment. Across seven evaluated constructs, VRSA exhibited high adoption intention (observational M = 4.47, SD = 0.64; survey M = 4.08, SD = 0.99), strong immersion (observational M = 4.31, SD = 0.73; survey M = 3.64, SD = 0.91), and high visual quality (M = 4.19/4.11). Statistically significant observational-to-survey discrepancies were identified for immersion (delta = + 0.67, W = 115.0, p < .001, d = 0.603) and navigational ease (delta = -0.98, W = 132.5, p < .001, d = -0.748), revealing systematic perceptual biases in retrospective self-report grounded in the peak-end rule and contrast effects. The VRSA-over-ARSA preference mean of 3.95 (SD = 1.20) significantly exceeded the neutral midpoint (W = 1128.0, p < .001). Thematic analysis yielded five preference themes. Findings are critically contextualised against comparable studies and yield actionable design recommendations for immersive retail technology. augmented reality virtual reality technology acceptance model e-commerce user experience immersive technology retail HCI adoption intention cybersickness Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 Figure 10 Figure 11 Figure 12 Figure 13 I. INTRODUCTION The rapid maturation of immersive computing hardware and the accelerating pace of digital retail transformation have positioned Virtual Reality (VR) and Augmented Reality (AR) as two of the most actively explored experiential technologies in commerce. VR generates a fully synthetic, computer-generated three-dimensional environment delivered through head-mounted displays (HMDs), offering users a sense of physical presence within a simulated space (Slater and Sanchez-Vives 2016 ). AR overlays digital content onto the physical environment, typically through smartphone cameras or optical see-through displays, preserving real-world spatial context while augmenting it with virtual information (Azuma 1997 ). Both technologies have attracted substantial commercial investment in retail: IKEA has deployed AR mobile product visualisation and VR kitchen design experiences, while Nike, Amazon, and Sephora have integrated AR try-on features into mobile commerce workflows (Hilken et al. 2022 ; Ricci et al. 2023 ). Despite this commercial activity, the academic literature remains fragmented. Most empirical studies examine either VR or AR in isolation, and comparative studies frequently conflate technology type with application domain, product category, or participant population (Kim et al. 2023 ; Xi et al. 2024 ). A study that rigorously characterises the multi-dimensional user experience of VRSA using both direct behavioural observation and self-report instruments, and situates this characterisation within a comparative framework capturing user preference between VRSA and ARSA, would therefore constitute a meaningful empirical contribution. The retail context was chosen for four mutually reinforcing reasons. First, retail is among the most commercially immediate deployment domains for immersive technologies, where experiential quality directly influences purchase intention, product return rates, and brand perception (Sengupta and Cao 2022 ). Second, the product visualisation task is sufficiently multi-dimensional to exercise a broad range of user experience constructs including perceptual fidelity, spatial navigation, task efficiency, hedonic value, and utilitarian usefulness. Third, shopping tasks generate ecologically valid, observable behaviours amenable to structured observational coding. Fourth, the established literature on consumer technology acceptance provides well-validated theoretical frameworks within which to anchor findings. The study employs an asymmetric instrumentation design: the VRSA was evaluated using a full multi-construct observational rubric and post-experience survey, while the ARSA served as a comparative benchmark evaluated through a preference instrument and qualitative protocol. This design is consistent with within-subjects benchmarking paradigms in technology evaluation research (Nielsen 1993 ), and its implications for comparative claims are explicitly bounded throughout. Three principal contributions emerge from this work: (1) a rigorously instrumented multi-construct VRSA evaluation with full inferential statistical analysis across 64 participants, incorporating the first published systematic test of divergence between observational and self-report measurement modalities in VR retail research; (2) identification and theoretically grounded interpretation of systematic observational-to-survey divergences that carry direct methodological implications for the field; and (3) a five-theme qualitative preference taxonomy and evidence that optimal technology modality is product-category-dependent rather than universal. These contributions are deliberately scoped to what the present design can support and are discussed against their boundary conditions throughout. II. LITERATURE REVIEW A. Virtual Reality in Retail: Immersion, Presence, and Adoption Research on VR-based retail applications has developed substantially over the past decade. Ricci et al. ( 2023 ) examined experiential differences between immersive and desktop VR in virtual fashion retail, finding that HMD-delivered VR produced significantly higher presence and richer product examination behaviour. Slater and Sanchez-Vives ( 2016 ) provided the foundational conceptual distinction between immersion, an objective technological property of the hardware, and presence, the subjective psychological experience of being there. This distinction is directly relevant to the divergences between observational and self-report measures documented in the present study. Han et al. ( 2023 ) applied the Stimulus-Organism-Response (SOR) model to VR shopping, finding that environmental stimuli in virtual retail spaces activate cognitive and affective organism states mediating purchase intention, with experiential value and hedonic arousal identified as organism-level mechanisms. Xi et al. ( 2024 ) extended this line of research in a metaverse acceptance laboratory experiment, finding that VR significantly increased perceived enjoyment. Critically, Xi et al. ( 2024 ) also identified HMD hardware usability constraints as the primary adoption barrier for VR, with participants reporting difficulty navigating immersive environments. This finding is relevant to the present navigational ease results and is distinct from the technology-acceptance profile of mobile AR, where such hardware constraints are absent. Cybersickness represents the most significant inhibitory factor for VR retail adoption. Weech et al. ( 2019 ) documented that cybersickness and sense of presence are negatively related in VR environments, with individual susceptibility varying considerably. Rebenitsch and Owen ( 2016 ) provided a comprehensive review of cybersickness in applications and visual displays, categorising symptoms and mitigation approaches across hardware, software, and behavioural domains. Stanney et al. ( 2003 ) examined exposure durations of 15, 30, 45, and 60 minutes, identifying sustained increases in cybersickness beyond the 15-minute mark and recommending time-limited exposure protocols. The extended TAM for VR proposed by Sagnier et al. ( 2020 ) augmented the foundational TAM (Davis 1989 ) by incorporating cybersickness, hedonic quality, and personal innovativeness, directly relevant to the present VRSA evaluation. B. Augmented Reality in Retail AR retail research emphasises practical usability, contextual product visualisation, and mobile consumer behaviour. Javornik ( 2016 ) identified interactivity, realness, and augmentation quality as primary drivers of consumer response to AR, with interactivity as the strongest predictor of engagement. Flavian et al. ( 2019 ) provided a comprehensive overview of AR in marketing, situating the technology's value generation as reduction of perceived purchase risk and enhancement of consumer decision quality. Lee et al. ( 2022 ) applied the SOR framework to AR virtual try-on, finding that AR augmentation features stimulate a form of mobile telepresence amplifying both utilitarian and hedonic consumer value. Hoffmann et al. ( 2022 ) concluded that AR's real-world spatial contextualisation represents a qualitatively different mode of product comprehension than VR's substitutive virtual environment. Cipresso et al. ( 2018 ) conducted a network and cluster analysis of VR and AR research, identifying distinct research trajectories and application domains for each modality. Oyman et al. ( 2022 ) further demonstrated that perceived AR quality positively influences perceived enjoyment and usefulness within extended TAM frameworks, reinforcing the role of hedonic and utilitarian pathways in mobile AR acceptance. C. Comparative Studies and Research Gap Kim et al. ( 2023 ) contrasted consumer perceptions between AR and VR using IKEA's commercial offerings, finding that VR conditions generated higher presence ratings than AR, while AR was associated with greater perceived interactivity in use. Hilken et al. ( 2022 ) argued that context and product category should determine modality selection. Xi et al. ( 2024 ) found technology-by-context interactions precluding universal modality superiority claims. No published study to date has employed a structured observational instrument alongside a self-report survey to evaluate VRSA across multiple constructs and statistically tested divergences between these measurement modalities in immersive retail contexts. The present study addresses this gap directly. D. Theoretical Framework This study adopts TAM (Davis 1989 ) as its primary theoretical scaffold, extended by the VR-specific additions of Sagnier et al. ( 2020 ) and the experiential value dimensions from the retail-specific literature (Han et al. 2023 ). The seven evaluated constructs collectively span TAM perceived usefulness and ease of use dimensions, VR-specific cybersickness and hedonic quality dimensions, and hedonic and affective dimensions from the SOR and experiential value literatures (Han et al. 2023 ; Lee et al. 2022 ). III. METHODOLOGY A. Research Design This study employed a mixed-methods design combining structured observational coding, post-experience self-report survey, and qualitative preference response analysis. The design is intentionally asymmetric: the VRSA was evaluated as the primary artefact using the full multi-construct instrument battery, while the ARSA served as a comparative benchmark evaluated through a preference instrument and qualitative protocol. This is consistent with within-subjects benchmarking paradigms (Nielsen 1993 ). B. Participants A total of 64 participants were recruited through purposive sampling from Middlesex University London. Age distribution: 60 (93.75%) aged 18 to 27, 2 (3.12%) aged 28 to 43, and 2 (3.12%) above 43 years. This demographic profile is consistent with the established early adopter demographic for immersive commerce technologies (Stecuła et al. 2024 ). Prior general VR experience was reported by 37 participants (57.81%), while 27 (42.19%) had no prior VR exposure. Among experienced users, the dominant usage frequency was less than once a month (n = 32; 86.49%). Prior VRSA experience was rare: only 7 of 37 VR-experienced participants (18.92%) had previously used a VR shopping application. All participants provided written informed consent; ethical approval was granted by the institutional review board. C. Apparatus and Stimuli The VRSA comprised a commercial VR shopping mall application delivered via a Meta Quest 2 HMD, affording navigable three-dimensional retail space, product shelf browsing, spatial product inspection, and a virtual checkout workflow. The ARSA comprised a mobile AR shopping application deployed on an iOS device. Mobile AR was selected because it represents the dominant channel for consumer AR retail experiences (Hoffmann et al. 2022 ; Flavian et al. 2019 ). Product categories spanned consumer clothing, and home furniture to exercise the full range of evaluated constructs. D. Instrumentation Observational Rubric. An eight-item structured observational instrument captured behavioural indicators during each participant's VRSA session, rated on a five-point scale by two trained independent observers. Items assessed: (1) ease of navigation; (2) engagement with interactive elements; (3) immersion inferred from body language; (4) reaction to visual quality; (5) system speed and responsiveness; (6) discomfort indicators; (7) observed likelihood of adopting VRSA; and (8) task completion time in minutes. Inter-rater reliability assessed via Cohen's kappa returned values exceeding 0.75 for all items, indicating substantial to near-perfect agreement (Landis and Koch 1977 ). System Performance (item 5) returned SD = 0.000 across all 64 sessions, reflecting technically uniform application performance consistent with the Meta Quest 2's hardware reliability under controlled conditions. Because the observational System Performance measure has zero variance, it cannot be used in a paired statistical comparison; survey-observed divergence for this construct is therefore reported descriptively only and its statistical row in Table V is marked accordingly. Post-Experience Survey. A 19-item instrument assessed demographic characteristics, prior technology experience, and seven evaluation constructs on five-point Likert scales immediately following the VRSA session, developed through expert validation (content validity index review by three domain experts). Internal consistency analysis returned Cronbach alpha = 0.783 for the full scale (acceptable; Nunnally 1978 ) and alpha = 0.776 for the usability sub-scale. The hedonic sub-scale returned alpha = 0.397, falling below acceptable reliability thresholds. This result indicates that immersion and visual quality function as distinct subjective dimensions rather than a unified hedonic construct in this context; accordingly, no composite hedonic score is computed or interpreted anywhere in this paper, and the two constructs are treated independently throughout. Items were drawn from TAM-derived measures (Davis 1989 ), presence-related scales (Slater and Sanchez-Vives 2016 ), and cybersickness-adjacent discomfort items informed by the Simulator Sickness Questionnaire (Kennedy et al. 1993 ). Comparative Preference Instrument. Two Likert-scale items assessed comparative preferences following both application sessions. Item 16 measured perceived ease comparison: "How easy did you find the AR shopping app compared to the VR shopping app?" (1 = AR much easier, 5 = VR much easier). Item 17 measured preference likelihood: "If you had the option to use either the AR or VR shopping Apps for future purchases, how likely are you to choose the AR shopping app over the VR shopping App?" (1 = much more likely to choose AR, 5 = much more likely to choose VR). A qualitative response item (Q18) asked participants to provide written reasons for their preference: "Would you mind providing a reason for your previous answer?" These written responses were transcribed for thematic analysis. Item 19 captured open-ended feature suggestions for future iterations. E. Procedure Sessions were conducted in a controlled laboratory with three concurrent VR stations, processed in triads across 45-minute sessions. Participants provided informed consent and completed demographic sections before the session. VR-naive participants underwent a supervised tutorial before measured tasks. Standardised shopping tasks included product searching, multi-angle examination, and cart management. Application sequence was counterbalanced across sessions to control for order effects. VRSA interactions were capped at 20 minutes to prevent cybersickness accumulation consistent with established exposure duration recommendations (Stanney et al. 2003 ; Weech et al. 2019 ); ARSA sessions were uncapped, typically concluding within 10 to 15 minutes. Sessions were recorded by a 360-degree camera enabling post-hoc observational coding. The post-experience survey was administered immediately following each application session. F. Data Analysis Shapiro-Wilk tests confirmed that all construct distributions deviated significantly from normality (all p < .001), justifying non-parametric procedures. Descriptive statistics were computed for all constructs. Internal consistency was assessed via Cronbach alpha. Paired Wilcoxon signed-rank tests were conducted to assess the significance of observational-to-survey divergences, with effect sizes reported as Cohen's d and rank-biserial r. All paired comparison p-values were Bonferroni-corrected for seven simultaneous tests (adjusted alpha = .007). System Performance, which returned SD = 0 in the observational instrument, was retained in the descriptive table for completeness but is excluded from inferential interpretation as a zero-variance distribution renders rank-based testing uninformative; this exclusion is noted explicitly in Table V. Spearman correlations characterised inter-construct relationships. Mann-Whitney U tests with Bonferroni correction examined VR experience moderation. Task completion time was analysed descriptively with outliers identified at M + 2SD, and Spearman correlations computed with navigation ease measures. Qualitative preference data were analysed using Braun and Clarke's (2006) six-phase thematic analysis framework, achieving intercoder agreement exceeding 82% prior to consensus resolution. IV. FINDINGS A. Participant Profile Table I summarises sample characteristics. The high rate of VRSA-naive participants among the VR-experienced sub-group (81.08%) confirms that domain-specific novelty effects are present and must be considered when interpreting adoption intention scores. TABLE I Participant Demographics and Prior Technology Experience Variable Category n % Age Range 18 to 27 60 93.75 28 to 43 2 3.12 Above 43 2 3.12 Prior VR Experience No VR Experience 27 42.19 Has VR Experience 37 57.81 VR Frequency (n = 37) Less than once per month 32 86.49 A few times per month 2 5.41 About once per week or more 3 8.10 Prior VRSA Experience (n = 37) No prior VRSA use 30 81.08 Prior VRSA use 7 18.92 Note. VR Frequency percentages are of the VR-experienced sub-group (n = 37). B. VRSA Observational Findings Table II and Fig. 5 present the observational instrument means and standard deviations. Discomfort returned the highest score (M = 4.75, SD = 0.44), indicating that the large majority of participants completed the session without observable severe discomfort. Adoption intention was high (M = 4.47, SD = 0.64) and immersion also strong (M = 4.31, SD = 0.73), consistent with the capacity of HMD-delivered VR to generate high levels of objective immersion (Slater and Sanchez-Vives 2016 ). System Performance returned M = 4.00, SD = 0.000 (see Section III-D for explanation of zero variance). Ease of navigation returned the lowest observational score (M = 3.22, SD = 1.16), with the highest inter-participant variability of all constructs. Task completion time returned M = 2.51 min (SD = 1.06), right-skewed (skewness = 1.14), with two outliers exceeding M + 2SD (4.63 min): TU34 (6.00 min) and TU62 (6.06 min), both VR-naive. TABLE II VRSA Observational Instrument: Means and Standard Deviations (N = 64) Construct Obs. Mean SD Scale Anchors (5 = most positive) Adoption Intention 4.47 0.64 1 = Very Unlikely ... 5 = Very Likely Immersion 4.31 0.73 1 = Not Immersed ... 5 = Fully Immersed Visual Quality 4.19 0.81 1 = Very Dissatisfied ... 5 = Very Satisfied Interactivity 3.44 0.81 1 = No Engagement ... 5 = Highly Engaged System Performance 4.00 0.00* 1 = Very Slow ... 5 = Very Fast Discomfort (no-discomfort +ve) 4.75 0.44 1 = Severe Discomfort ... 5 = No Discomfort Ease of Navigation 3.22 1.16 1 = Frequent Hesitation ... 5 = Seamless Navigation Task Completion Time 2.51 min 1.06 Recorded in minutes Note. *SD = 0.000: all 64 observers assigned score 4, reflecting uniform application performance. Paired inferential testing is not computed for this construct. C. VRSA Post-Experience Survey Findings Table III presents survey means, SDs, observational means, and divergence scores. Visual quality (M = 4.11, SD = 0.72) and system performance (M = 4.28, SD = 0.81) returned strong survey scores consistent with observational counterparts. Adoption intention remained high (M = 4.08, SD = 0.99). Two constructs showed notable divergences: immersion registered M = 3.64 (SD = 0.91) in the survey versus 4.31 in observation (delta = + 0.67), and ease of navigation registered M = 4.20 (SD = 0.86) in the survey versus 3.22 in observation (delta = -0.98). Statistical testing of these divergences is presented in Section IV-E. The VRSA-over-ARSA preference item returned M = 3.95 (SD = 1.20). TABLE III VRSA Post-Experience Survey and Observational Means with Divergence Scores (N = 64) Construct Survey M SD Obs. M Delta (Obs. minus Survey) Adoption Intention 4.08 0.99 4.47 + 0.39 Immersion 3.64 0.91 4.31 + 0.67*** Visual Quality 4.11 0.72 4.19 + 0.08 Interactivity 3.91 0.88 3.44 -0.47* System Performance 4.28 0.81 4.00 -0.28* Discomfort (no-discomfort +ve) 4.48 0.89 4.75 + 0.27 Ease of Navigation 4.20 0.86 3.22 -0.98*** VRSA-over-ARSA Preference 3.95 1.20 N/A N/A Note. *** Bonferroni-corrected significant (p < .001); * Bonferroni-corrected significant (p < .05). Full statistical details in Table V. D. Qualitative Preference Themes Thematic analysis of the comparative preference protocol yielded five primary themes (Table IV). Theme 1, Immersion and Realism, was the most frequently coded theme among VRSA-preferring participants, framing VR as qualitatively closer to a physical retail visit. Theme 2, Ease of Use, was bidirectional: VRSA-preferring participants cited the VR interaction model as ultimately natural and embodied; ARSA-preferring participants cited the lower cognitive overhead of the smartphone interface. Theme 3, Entertainment and Novelty, was exclusively associated with VRSA preference, foregrounding hedonic motivation. Theme 4, Product Display Fidelity, captured responses explicitly comparing three-dimensional VR product presentation against the flat AR overlay. Theme 5, Contextual Practicality, dominated among ARSA-preferring and neutral participants, emphasising product-category-dependent optimal modality selection. TABLE IV Thematic Analysis of VRSA-vs-ARSA Comparative Preference Responses Theme VRSA/ARSA orientation Representative extract Theoretical relevance 1. Immersion and Realism Pro-VRSA "It felt like I was actually in a store." Presence as qualitative retail analogue 2. Ease of Use Bidirectional "VR felt natural once I got used to it." / "AR is just simpler." Technology familiarity moderates ease perception 3. Entertainment and Novelty Pro-VRSA "It was fun to explore the store." Hedonic motivation as adoption pathway 4. Product Display Fidelity Pro-VRSA "You could see the product from all angles." Spatial product comprehension advantage of VR 5. Contextual Practicality Pro-ARSA / Neutral "AR for large items; VR for visualising clothes." / "Discomfort with VR headsets." Context and product category moderate optimal technology selection Note. Themes ordered by prevalence. Extracts are verbatim participant statements. E. Statistical Testing of Observational-to-Survey Divergences Table V presents the full statistical results for paired Wilcoxon signed-rank tests comparing observational and survey means for each construct, with Bonferroni-corrected significance (adjusted alpha = .007). Two constructs showed Bonferroni-corrected significant divergences: Ease of Navigation (W = 132.5, p_adj < .001, d = -0.748, r_rb = 0.936) and Immersion (W = 115.0, p_adj < .001, d = 0.603, r_rb = 0.945). Two constructs showed uncorrected-significant but correction-failing divergences: Interactivity (p_adj = .025) and System Performance (p_adj = .021). Visual Quality, Discomfort, and Adoption Intention showed no significant divergence after correction. System Performance is retained descriptively in Table V, with the statistical result marked as non-inferential given the zero observational variance. TABLE V Paired Wilcoxon Signed-Rank Tests: Observational vs. Survey Construct Means (N = 64) Construct Obs M Obs SD Sv M Sv SD Delta W p raw p adj d r_rb Sig Ease of Navigation 3.22 1.15 4.20 0.85 -0.98 132.5 < .001 < .001 -0.748 0.936 *** Immersion 4.31 0.73 3.64 0.91 + 0.67 115.0 < .001 < .001 0.603 0.945 *** Visual Quality 4.19 0.81 4.11 0.71 + 0.08 339.0 .630 1.000 0.073 0.837 ns Interactivity 3.44 0.81 3.91 0.86 -0.47 285.5 .004 .025 -0.380 0.863 * System Performance 4.00 0.00 4.28 0.80 -0.28 190.0 .003 .021 -0.349 0.909 * (a) Discomfort 4.75 0.43 4.48 0.88 + 0.27 136.5 .040 .281 0.251 0.934 ns Adoption Intention 4.47 0.64 4.08 0.99 + 0.39 220.0 .014 .096 0.318 0.894 ns Note. Delta = Obs M minus Survey M. p adj = Bonferroni-corrected (adjusted alpha = .007). d = Cohen's d; r_rb = rank-biserial correlation (non-parametric effect size). *** p < .001; * p < .05 (Bonferroni-corrected); ns = not significant after correction. (a) System Performance: SD = 0 in observation; test is retained for descriptive reference only and results should not be interpreted inferentially. F. AR versus VR Preference and Ease The VRSA-over-ARSA preference item returned M = 3.95 (SD = 1.20, Mdn = 4.0). A one-sample Wilcoxon test confirmed that preference significantly exceeded the neutral midpoint of 3 (W = 1128.0, p < .001), indicating a statistically supported moderate lean toward VR. Perceived ease (VR relative to AR) returned M = 3.75 (SD = 1.15). Figure 10 presents the full distributions. G. VR Experience Moderation Table VI presents Mann-Whitney U results comparing survey construct scores by VR experience group. No construct showed a statistically significant difference after Bonferroni correction, indicating that prior VR experience did not systematically moderate survey responses. The largest effect size was for adoption intention (r = 0.223), with VR-naive participants reporting marginally higher adoption intention (M = 4.26) than experienced participants (M = 3.95). Figure 11 presents the boxplots. TABLE VI Mann-Whitney U Tests: Survey Construct Scores by Prior VR Experience Group Construct Exp M Exp SD Naive M Naive SD U p raw p adj r Sig Ease of Navigation 4.27 0.90 4.11 0.80 569.5 .309 1.000 -0.140 ns Immersion 3.65 0.75 3.63 1.11 480.0 .783 1.000 0.039 ns Visual Quality 4.08 0.68 4.15 0.77 474.5 .713 1.000 0.050 ns Interactivity 3.84 0.87 4.00 0.88 464.0 .613 1.000 0.071 ns System Performance 4.22 0.85 4.37 0.74 452.5 .490 1.000 0.094 ns Discomfort 4.46 0.93 4.52 0.85 492.5 .913 1.000 0.014 ns Adoption Intention 3.95 0.97 4.26 1.02 388.0 .108 .756 0.223 ns Note. Exp = VR-experienced (n = 37); Naive = VR-naive (n = 27). p adj = Bonferroni-corrected (adjusted alpha = .007). r = rank-biserial. ns = not significant after correction. H. Inter-Construct Spearman Correlations Spearman correlations revealed a cluster of strong associations among the perceptual quality dimensions. The strongest correlation was Visual Quality with Interactivity (rho = 0.70, p < .001), followed by Interactivity with System Performance (rho = 0.61, p < .001) and Visual Quality with System Performance (rho = 0.47, p < .001). Ease of Navigation correlated significantly with Visual Quality (rho = 0.41, p < .001) and System Performance (rho = 0.40, p < .001). Discomfort showed no significant correlations with other constructs, consistent with cybersickness susceptibility being an individual physiological characteristic largely independent of perceived experiential quality. Adoption Intention showed significant correlations with Ease of Navigation (rho = 0.26, p = .038), Visual Quality (rho = 0.25, p = .050), Interactivity (rho = 0.30, p = .015), and Discomfort (rho = 0.31, p = .012). Figure 12 presents the full matrix. I. Task Completion Time Analysis Task completion time returned M = 2.51 min (SD = 1.06, Mdn = 2.36, skewness = 1.14). Two participants (TU34: 6.00 min; TU62: 6.06 min) exceeded M + 2SD (4.63 min); both had no prior VR experience. Spearman correlation between observed navigation ease and task time was rho = -0.287 (p = .021), confirming that participants rated as navigating more easily by observers completed the task faster. Survey-reported navigation ease to task time correlation was weaker and non-significant (rho = -0.172, p = .173). Figure 13 presents the distribution and scatter plots. V. DISCUSSION A. Contextualising VRSA Findings Within the Extant Literature The VRSA adoption intention means of 4.08 (survey) and 4.47 (observation) are among the highest reported in comparable single-session VR retail evaluations. Kim et al. ( 2023 ) reported strong VR adoption intentions in their comparative IKEA-based study, attributing the effect primarily to hedonic quality rather than utilitarian efficiency. The present findings are directionally consistent: qualitative Theme 3 (Entertainment and Novelty) identifies hedonic motivation as a primary preference driver, and the significant Spearman correlation between interactivity and adoption intention (rho = 0.30, p = .015) provides quantitative support for hedonic engagement as a pathway to adoption in this context. However, the single-session design means novelty inflation cannot be excluded as a contributor. Xi et al. ( 2024 ) acknowledge the external validity constraints of laboratory-based XR shopping studies, noting that single-session conditions may not fully capture longitudinal adoption dynamics. The observed immersion mean of 4.31 is consistent with the high objective immersion documented by Ricci et al. ( 2023 ) for HMD-delivered VR fashion retail. The significant Bonferroni-corrected divergence between observed immersion (M = 4.31) and survey immersion (M = 3.64, W = 115.0, p < .001, d = 0.603) reflects the conceptual distinction between immersion as a hardware property and presence as a subjective psychological state (Slater and Sanchez-Vives 2016 ). The large effect size indicates that this is not a measurement artefact: the Meta Quest 2 hardware delivered high objective immersion, but this was not fully translated into equivalent subjective presence. Likely disruptors include cybersickness episodes, navigational friction (reflected in the low navigation ease scores), and the unfamiliarity of VR interaction for first-time users. The strong intercorrelations among visual quality, interactivity, and system performance (rho = 0.47 to 0.70, all p < .001) are theoretically interpretable as a coherent perceptual quality factor, consistent with the SOR framework of Han et al. ( 2023 ) in which stimulus richness functions as an integrated experiential input. That these three constructs converge while immersion and discomfort remain relatively uncorrelated with the quality cluster suggests that the seven-construct battery captures at least two distinct experiential dimensions: a perceptual quality dimension (visual quality, interactivity, performance, navigation ease) and an experiential depth dimension (immersion, discomfort). The low hedonic sub-scale alpha (0.397) corroborates this bifurcation: immersion and visual quality, though both hedonic in character, are not measuring the same subjective construct in VR retail contexts. This finding has direct implications for the design of future VRSA evaluation instruments and is discussed further in the Limitations section. B. The Navigational Ease Reversal: Mechanisms and Theoretical Grounding The most methodologically significant finding is the Bonferroni-corrected divergence of -0.98 between the observational navigation ease mean (M = 3.22, SD = 1.16) and the survey mean (M = 4.20, SD = 0.86; W = 132.5, p < .001, d = -0.748, r_rb = 0.936). The large rank-biserial r of 0.936 indicates that in nearly every matched participant pair, the observer rated navigation as harder than the participant self-reported. Three non-mutually exclusive mechanisms are proposed, each grounded in established cognitive psychology. First, retrospective competence anchoring (Nielsen 1993 ): participants evaluated their navigation at the skill level attained by session end rather than averaging across the full session trajectory, producing an overly optimistic assessment. This endpoint-weighting is consistent with the peak-end rule (Kahneman et al. 1993 ), which demonstrates that retrospective evaluations are disproportionately influenced by the end-state and peak moment rather than the integrated temporal experience. Applied to VR navigation, a participant who struggled for the first ten minutes but navigated fluently in the final five would retrospectively report high ease, while an observer coding the full session would record a lower mean. This mechanism is directly supported by the observational SD of 1.16 being substantially larger than the survey SD of 0.86: observer ratings captured the full heterogeneous trajectory while self-report compressed it toward the endpoint. Second, contrast effects from counterbalanced ordering: participants who experienced the ARSA before the VRSA may have implicitly contrasted VR navigation against the initially unfamiliar mobile AR interface, inflating their retrospective evaluation of VR navigation relative to an unstated AR baseline. Contrast effects of this type are well documented in comparative judgement paradigms (Schwarz and Bless 1992 ). Third, social desirability bias: participants may have rated navigation favourably in the survey context to appear competent in a structured academic environment with observers present. This mechanism would specifically affect ego-threatening constructs such as navigation competence. The practical implication for VRSA research and design is direct: post-experience surveys alone will systematically overestimate navigational ease in VRSA contexts. The Spearman correlation between observed navigation ease and task time (rho = -0.287, p = .021), compared to the non-significant correlation for survey-reported ease (rho = -0.172, p = .173), provides direct empirical support: the observed measure predicted objective task performance, while the survey measure did not. Observational or task-timing data are therefore essential complements for evaluating navigational usability in VR retail contexts. C. Interactivity and System Performance: Upward Survey Modulation Interactivity showed a Bonferroni-corrected significant divergence (W = 285.5, p_adj = .025, d = -0.380), with the survey mean (M = 3.91) substantially exceeding the observational mean (M = 3.44). System Performance showed a similarly directed divergence (W = 190.0, p_adj = .021, d = -0.349); however, as noted in Section III-D, the zero observational variance for System Performance renders this result descriptively informative rather than inferentially valid, and no causal interpretation is advanced. The interactivity divergence suggests that participants valued the interactive affordances more highly in retrospect than their in-session engagement patterns indicated. This pattern is consistent with the broader finding in technology acceptance research that retrospective evaluations tend to be more positive than concurrent behavioural measures, as users consolidate their understanding of a technology's potential during reflection. The implication for VRSA design is that interactive feature discoverability may be limiting in-session engagement even when those features are retrospectively valued, pointing to the need for clearer onboarding and affordance signalling within VR retail environments. The absence of Bonferroni-corrected significance for discomfort (W = 136.5, p_adj = .281) and adoption intention (W = 220.0, p_adj = .096) indicates that these constructs are relatively stably perceived across measurement modalities. Discomfort stability is particularly noteworthy: cybersickness is an acute physiological experience that is difficult to misremember or retrospectively reframe, which accounts for its measurement consistency. Adoption intention stability suggests that participant enthusiasm for future VRSA use was genuine and not subject to the same retrospective inflation mechanisms affecting navigation ease. D. Contextualising Comparative Preference Findings The overall VRSA-over-ARSA preference mean of 3.95 (SD = 1.20) significantly exceeded the neutral midpoint (W = 1128.0, p < .001), confirming a statistically supported moderate lean toward VR. This is directionally consistent with Kim et al. ( 2023 ), who found VR conditions generated higher presence ratings than AR in a within-subjects IKEA comparison, an effect those authors attributed primarily to the immersive hardware advantage of HMD-delivered environments. The SD of 1.20 reflects genuine individual heterogeneity, echoing Hilken et al. ( 2022 ) strategic argument that neither technology dominates universally and that context, product category, and consumer disposition interact to shape technology preference. The qualitative preference data add critical nuance to the quantitative score. Theme 5 (Contextual Practicality) documents that participants explicitly nominated different optimal technologies for different product categories: AR for large furniture items due to real-world spatial placement, VR for clothing due to immersive try-on experience. This contextual moderation is consistent with the Hilken et al. ( 2022 ) framework and with Lee et al. ( 2022 ) finding that AR telepresence generates distinct forms of product comprehension from VR spatial presence. The optimal immersive modality is product-dependent, not universal, constituting a theoretically grounded product-category-by-modality interaction with direct implications for retailer technology strategy. The bidirectionality of Theme 2 (Ease of Use) is theoretically informative. Among VRSA-preferring participants, the VR interaction model was characterised as natural and embodied once the learning curve was surmounted. Among ARSA-preferring participants, the smartphone interface was valued for its lower cognitive overhead and familiar interaction paradigm. This bidirectionality corroborates Sagnier et al. ( 2020 ) finding that prior technology familiarity moderates ease of use perception in VR contexts, and suggests that the perceived ease advantage of VR may be contingent on sufficient onboarding experience to overcome the initial navigational friction documented in the observational data. E. Limitations and Boundary Conditions Seven limitations bound the interpretation of these findings and should be carefully considered before generalising the results. First, the sample was drawn entirely from a single university in the United Kingdom and is demographically homogeneous: 93.75% of participants were aged 18 to 27. This profile is characteristic of early-adopter convenience samples in immersive technology research, but it substantially constrains generalisability to broader consumer populations. Older adults, in particular, may exhibit greater technology anxiety (Sagnier et al. 2020 ), distinct spatial navigation profiles, and different hedonic motivations than the present sample. All quantitative findings and preference patterns reported here apply to this demographic cohort and should not be extended to general consumer populations without replication across age groups and cultural contexts. Second, the sample size of 64, while adequate for the non-parametric procedures employed, is modest relative to the range of moderation hypotheses that could plausibly be investigated. The VR experience moderation analysis in particular is underpowered to detect small-to-moderate interaction effects; the absence of significant moderation should be treated as inconclusive rather than confirmatory for that analysis. Third, the asymmetric instrumentation design means that construct-level quantitative comparisons between VRSA and ARSA are not possible. The ARSA was not evaluated using the full seven-construct rubric and survey battery. Accordingly, all claims framed as comparative are limited to user-stated preferences and qualitative themes, and no inferential claim about ARSA construct means is advanced anywhere in this paper. Fourth, the single-session design cannot distinguish novelty-driven from utility-driven adoption intention, a critical distinction for predicting real-world deployment behaviour (Xi et al. 2024 ). The high adoption intention scores may be partially attributable to first-exposure enthusiasm rather than stable preference, particularly given that 81.08% of the VR-experienced sub-group had never used a VR shopping application before. Fifth, while counterbalancing controls for order effects, it does not eliminate them. The asymmetric session durations (VRSA capped at 20 minutes; ARSA uncapped at approximately 10 to 15 minutes) introduce a structural inequality in exposure time that may have influenced comparative preference ratings in ways not fully captured by the counterbalancing procedure. Sixth, the hedonic sub-scale reliability of alpha = 0.397 is well below acceptable thresholds. This finding is interpreted in this paper as evidence that immersion and visual quality are empirically distinct constructs in VR retail contexts, and no composite hedonic score is reported. Future instrument development should treat these as separate subscales from the outset. Seventh, the social desirability mechanism proposed for the navigational ease reversal is theoretically plausible but was not directly tested. A follow-up study incorporating implicit navigation performance measures alongside self-report would be needed to isolate this mechanism from the peak-end and contrast-effects accounts. VI. PRACTITIONER IMPLICATIONS The findings yield several actionable recommendations for VRSA designers, ARSA designers, and retail technology strategists. VRSA Navigation and Onboarding Design. The observational navigation ease mean of 3.22 and the significant negative correlation with task completion time (rho = -0.287, p = .021) confirm that VR spatial navigation represents the most significant usability barrier in current VRSA implementations. Designers should implement progressive onboarding with explicit spatial navigation affordances, teleportation as a fallback locomotion option, and visual wayfinding cues. The two task time outliers (TU34 and TU62, both VR-naive, both exceeding 6 minutes) suggest that VR-naive users require more substantial onboarding than a supervised tutorial alone provides. Methodological Triangulation for VRSA Evaluation. The systematic divergence between observed and self-reported navigation ease (d = -0.748) demonstrates that post-experience surveys alone are insufficient for assessing navigational usability in VRSA contexts. Practitioners should complement survey instruments with observational coding or task-timing data. Discomfort and adoption intention are relatively stable across modalities and can be assessed reliably through survey instruments alone. Interactive Feature Discoverability. The upward modulation of interactivity in the survey relative to observation (d = -0.380) suggests that participants valued interactive VR affordances more highly in reflection than they utilised them in session. VRSA designers should invest in interactive feature discoverability: clear visual and haptic cues indicating available interactions, tutorial prompts for key interactive features, and progressive feature introduction. Context-Sensitive Technology Deployment. The quantitative preference lean toward VR (W = 1128.0, p < .001) combined with qualitative evidence for product-category-specific modality preferences (Theme 5) supports a context-sensitive deployment model. Retailers with spatially large product categories such as furniture and appliances may find AR's real-world contextualisation to be the stronger modality. Retailers with wearable or highly visual product categories such as clothing and footwear may derive greater consumer value from VR's immersive spatial presentation. A hybrid deployment strategy exploiting both modalities for their respective product-category advantages is recommended (Hilken et al. 2022 ). Cybersickness Mitigation. Despite generally positive discomfort scores, the non-zero variance indicates a minority experiencing clinically meaningful distress. Established cybersickness mitigation strategies including vignetting during locomotion, field-of-view restriction during rapid movement, and explicit session time limits should be implemented (Rebenitsch and Owen 2016 ; Weech et al. 2019 ). The 20-minute session cap applied in this study appears appropriate as a deployment guideline (Stanney et al. 2003 ). VII. CONCLUSION This paper has presented a rigorously instrumented mixed-methods evaluation of VRSA user experience, drawing on triangulated observational and self-report data from 64 participants supplemented by comparative preference analysis against an ARSA. Three contributions, scoped to what the present design permits, have been delivered. The first and primary contribution is the identification and statistical testing of systematic divergences between observational and self-report measurement modalities in VR retail research. This is, to the authors' knowledge, the first published study to apply paired non-parametric testing with effect size reporting to observational-versus-survey divergences across multiple constructs in a VRSA context. The immersion divergence (d = 0.603, W = 115.0, p < .001) is theoretically interpreted through the immersion-presence distinction (Slater and Sanchez-Vives 2016 ), while the navigational ease reversal (d = -0.748, W = 132.5, p < .001) is grounded in the peak-end rule (Kahneman et al. 1993 ), contrast effects (Schwarz and Bless 1992 ), and social desirability bias. The demonstration that observed navigation ease significantly predicted task completion time (rho = -0.287, p = .021) while survey-reported ease did not (rho = -0.172, p = .173) provides direct empirical support for the non-redundancy of observational and self-report data in this domain. The second contribution is a multi-construct characterisation of VRSA user experience across seven dimensions with full inferential statistical analysis. Adoption intention, immersion, visual quality, discomfort, and system performance registered strongly; navigational ease and task efficiency emerged as areas of significant challenge. This characterisation is bounded to a young adult university sample and a single commercial application and should be treated accordingly. The third contribution is a five-theme qualitative preference taxonomy demonstrating that optimal technology modality is product-category-dependent rather than universal. AR was preferred for large spatially contextualised products, VR for wearable and visually rich categories, a finding that directly informs practitioner deployment strategy. Future work should prioritise longitudinal designs to separate novelty-driven from utility-driven adoption intention, and should recruit demographically broader samples spanning older adult populations to address the generalisability constraints of the present study. Symmetric full-construct instrumentation of both VRSA and ARSA in a single study would enable the construct-level comparative analysis that the present asymmetric design precludes. Neurophysiological measures including galvanic skin response and electroencephalography would provide objective proxies for immersion and cybersickness, eliminating the retrospective bias demonstrated here. Finally, the product-category-by-modality interaction identified qualitatively warrants direct experimental examination. Declarations Funding Statement This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors. Informed consent statement : Informed consent was obtained from all individual participants included in the study. All participants provided written informed consent prior to taking part in the experimental session. The study was approved by the Research Ethics Committee of the Department of Computer Science, Middlesex University London, and was conducted in accordance with the committee's ethical standards and guidelines. Funding Statement This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors. Author Contribution S.A. conceived and designed the study, developed the observational rubric and survey instruments, collected and analysed the data, performed all statistical analyses, conducted the thematic analysis, and wrote the main manuscript text including all sections, tables, and figures. G.D. contributed to the theoretical framework, provided critical supervision throughout the study design and analytical phases, and reviewed and revised the manuscript. Both authors approved the final submitted version. ACKNOWLEDGEMENTS The authors gratefully acknowledge all study participants for their time and engagement, and the Department of Computer Science, Middlesex University London, for supporting the experimental infrastructure used in this research. Data Availability The data that support the findings of this study consist of structured observational ratings, and post-experience survey responses collected from human participants under institutional ethical approval. Full open sharing is not possible as participants consented to data use within the bounds of the approved study protocol, and individual-level responses could potentially identify participants given the controlled laboratory setting and small group sizes. The data are available from the corresponding author upon reasonable request, subject to confirmation that the request is consistent with the original ethical approval. The Python scripts used for statistical analysis, including the non-parametric procedures, correlation analyses, and Bonferroni correction routines, are also available from the corresponding author upon reasonable request. References Azuma RT (1997) A survey of augmented reality. Presence Teleoper Virtual Environ 6(4):355–385. https://doi.org/10.1162/pres.1997.6.4.355 Braun V, Clarke V (2006) Using thematic analysis in psychology. Qual Res Psychol 3(2):77–101. https://doi.org/10.1191/1478088706qp063oa Cipresso P, Giglioli IAC, Raya MA, Riva G (2018) The past, present, and future of virtual and augmented reality research: a network and cluster analysis of the literature. Front Psychol 9:2086. https://doi.org/10.3389/fpsyg.2018.02086 Davis FD (1989) Perceived usefulness, perceived ease of use, and user acceptance of information technology. MIS Q 13(3):319–340. https://doi.org/10.2307/249008 Flavian C, Ibanez-Sanchez S, Orus C (2019) Augmented reality (AR): an overview and future research agenda. Psychol Mark 36(11):1043–1057. https://doi.org/10.1002/mar.21227 Han SL, Kim J, An M (2023) The role of VR shopping in digitalization of SCM for sustainable management: application of SOR model and experience economy. Sustainability 15(2):1277. https://doi.org/10.3390/su15021277 Hilken T, Chylinski M, Keeling DI, Heller J, de Ruyter K, Mahr D (2022) How to strategically choose or combine augmented and virtual reality for improved online experiential retailing. Psychol Mark 39(3):495–507. https://doi.org/10.1002/mar.21600 Hoffmann A, Joerss T, Kolbe C (2022) Consumer behavior in augmented shopping reality: a review, synthesis, and research agenda. Front Virtual Real 3:961236. https://doi.org/10.3389/frvir.2022.961236 Javornik A (2016) It's an illusion, but it looks real! Consumer affective, cognitive and behavioural responses to augmented reality applications. J Mark Manag 32(9–10):987–1011. https://doi.org/10.1080/0267257X.2016.1174726 Kahneman D, Fredrickson BL, Schreiber CA, Redelmeier DA (1993) When more pain is preferred to less: adding a better end. Psychol Sci 4(6):401–405. https://doi.org/10.1111/j.1467-9280.1993.tb00589.x Kennedy RS, Lane NE, Berbaum KS, Lilienthal MG (1993) Simulator sickness questionnaire: an enhanced method for quantifying simulator sickness. Int J Aviat Psychol 3(3):203–220. https://doi.org/10.1207/s15327108ijap0303_3 Kim JH, Kim M, Park M, Yoo J (2023) Immersive interactive technologies and virtual shopping experiences: differences in consumer perceptions between augmented reality (AR) and virtual reality (VR). Telematics Inf 77:101936. https://doi.org/10.1016/j.tele.2022.101936 Landis JR, Koch GG (1977) The measurement of observer agreement for categorical data. Biometrics 33(1):159–174. https://doi.org/10.2307/2529310 Lee H, Xu Y, Porterfield A (2022) Antecedents and moderators of consumer adoption toward AR-enhanced virtual try-on technology: a stimulus-organism-response approach. Int J Consum Stud 46(4):1319–1338. https://doi.org/10.1111/ijcs.12753 Nielsen J (1993) Usability engineering. Academic, Boston Nunnally JC (1978) Psychometric theory, 2nd edn. McGraw-Hill, New York Oyman M, Bal D, Ozer S (2022) Extending the technology acceptance model to explain how perceived augmented reality affects consumers' perceptions. Comput Hum Behav 128:107127. https://doi.org/10.1016/j.chb.2021.107127 Rebenitsch L, Owen C (2016) Review on cybersickness in applications and visual displays. Virtual Real 20(2):101–125. https://doi.org/10.1007/s10055-016-0285-9 Ricci M, Evangelista A, Di Roma A, Fiorentino M (2023) Immersive and desktop virtual reality in virtual fashion stores: a comparison between shopping experiences. Virtual Real 27:2281–2296. https://doi.org/10.1007/s10055-023-00806-y Sagnier C, Loup-Escande E, Lourdeaux D, Thouvenin I, Vallery G (2020) User acceptance of virtual reality: an extended technology acceptance model. Int J Hum Comput Interact 36(11):993–1007. https://doi.org/10.1080/10447318.2019.1708612 Schwarz N, Bless H (1992) Constructing reality and its alternatives: an inclusion/exclusion model of assimilation and contrast effects in social judgment. In: Martin LL, Tesser A (eds) The construction of social judgments. Lawrence Erlbaum, Hillsdale, pp 217–245 Sengupta A, Cao L (2022) Augmented reality's perceived immersion effect on the customer shopping process: decision-making quality and privacy concerns. Int J Retail Distrib Manag 50(8/9):1039–1061. https://doi.org/10.1108/IJRDM-10-2021-0522 Slater M, Sanchez-Vives MV (2016) Enhancing our lives with immersive virtual reality. Front Robot AI 3:74. https://doi.org/10.3389/frobt.2016.00074 Stanney KM, Hale KS, Nahmens I, Kennedy RS (2003) What to expect from immersive virtual environment exposure: influences of gender, body mass index, and past experience. Hum Factors 45(3):504–520. https://doi.org/10.1518/hfes.45.3.504.27254 Stecuła K, Wolniak R, Aydin B (2024) Technology development in online grocery shopping from shopping services to virtual reality, metaverse, and smart devices: a review. Foods 13(23):3959. https://doi.org/10.3390/foods13233959 Weech S, Kenny S, Barnett-Cowan M (2019) Presence and cybersickness in virtual reality are negatively related: a review. Front Psychol 10:158. https://doi.org/10.3389/fpsyg.2019.00158 Xi N, Chen J, Gama F, Korkeila H, Hamari J (2024) Acceptance of the metaverse: a laboratory experiment on augmented and virtual reality shopping. Internet Res 34(7):82–117. https://doi.org/10.1108/INTR-05-2022-0334 Additional Declarations No competing interests reported. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-9418492","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":630279475,"identity":"b79a3f01-39a5-4fbb-b39c-aed6e2ffb9a6","order_by":0,"name":"Stephen Agada","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAz0lEQVRIiWNgGAWjYBAC+QYQycYgx8CQABGRIKTF4ABEizEJWhggWhIbSNDC/PjFjzK79LXtOQYMP2oYEmc2ENAi38BmZtlzLjl325lnCYw9xxgSZxOyheEAg5kBbxtz7rYbyQcYeBsYEucR1sL+zfBvW3262Y3EBsa/xGnhMX7M23Y4wQxoCzPIFoIOMzjMU8Ysc+64Icgvh2WOSRgT9n57++aPb8qq5c2O5xg+fFNjIzvjACFrmBnY4DFxgIhYgWj6QJSyUTAKRsEoGLkAAAmpQYazuFCnAAAAAElFTkSuQmCC","orcid":"","institution":"Middlesex University London","correspondingAuthor":true,"prefix":"","firstName":"Stephen","middleName":"","lastName":"Agada","suffix":""},{"id":630279476,"identity":"50935919-1d73-4071-9adc-bd2cb9f4da7a","order_by":1,"name":"Georgios Dafoulas","email":"","orcid":"","institution":"Middlesex University London","correspondingAuthor":false,"prefix":"","firstName":"Georgios","middleName":"","lastName":"Dafoulas","suffix":""}],"badges":[],"createdAt":"2026-04-14 17:38:44","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9418492/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9418492/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":108385247,"identity":"afe9f49c-2da8-417b-9b71-64a801ca2035","added_by":"auto","created_at":"2026-05-04 06:02:40","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":626913,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eExperimental session setup\u003c/em\u003e\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-9418492/v1/231fc55dfffccad5ede68714.png"},{"id":108385248,"identity":"6319e7f8-7187-4a78-8b59-66dc6f4dc42d","added_by":"auto","created_at":"2026-05-04 06:02:40","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":615628,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003ePrimary environments\u003c/em\u003e\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-9418492/v1/e76759931c49f5b37da92089.png"},{"id":108385249,"identity":"5f06a87d-2520-4e1d-890d-b74cf8c00a8b","added_by":"auto","created_at":"2026-05-04 06:02:40","extension":"jpg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":99640,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eVRSA tutorial scenario during onboarding phase.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-9418492/v1/118fd44ce76ca8c375d4d77e.jpg"},{"id":108385250,"identity":"97eaf59d-5c40-4598-8b5f-7aab50fe5f90","added_by":"auto","created_at":"2026-05-04 06:02:40","extension":"jpg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":17936,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eParticipant demographics: age distribution (left), prior VR experience (centre), and VR usage frequency among experienced users (right; n = 37).\u003c/em\u003e\u003c/p\u003e","description":"","filename":"4.jpg","url":"https://assets-eu.researchsquare.com/files/rs-9418492/v1/73d1e03908616eadaa290d0c.jpg"},{"id":108492321,"identity":"8c860546-d3dd-421e-89da-7df97c92462a","added_by":"auto","created_at":"2026-05-05 09:57:26","extension":"jpg","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":25580,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eObservational rubric construct means: radar chart (left) and bar chart with SD (right; N = 64). Dashed line = neutral midpoint (3).\u003c/em\u003e\u003c/p\u003e","description":"","filename":"5.jpg","url":"https://assets-eu.researchsquare.com/files/rs-9418492/v1/cbaaad90feecbdc57d882e20.jpg"},{"id":108492779,"identity":"f6e0448b-0da6-4baf-acbc-72bccb6da5d2","added_by":"auto","created_at":"2026-05-05 09:58:37","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":316328,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eVRSA application screenshots showing product selection and navigation interface.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"6.png","url":"https://assets-eu.researchsquare.com/files/rs-9418492/v1/02168d1bc74bd7a4ecee3c3b.png"},{"id":108385252,"identity":"56a304f4-71fa-4b17-af4e-2f170d2248d3","added_by":"auto","created_at":"2026-05-04 06:02:40","extension":"jpg","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":35873,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eFrequency distributions of the seven VRSA post-experience survey constructs (N = 64). Red dashed line indicates mean.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"7.jpg","url":"https://assets-eu.researchsquare.com/files/rs-9418492/v1/e49ae77f3e376ec7252a61bc.jpg"},{"id":108492352,"identity":"0abf7763-612f-4f53-892d-48df7aa5ceab","added_by":"auto","created_at":"2026-05-05 09:57:33","extension":"png","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":506850,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eARSA application screenshots\u003c/em\u003e\u003c/p\u003e","description":"","filename":"8.png","url":"https://assets-eu.researchsquare.com/files/rs-9418492/v1/ff0a9c1aff73e7b09868c382.png"},{"id":108385255,"identity":"00cd840d-0a38-47f0-8cf1-6a637da27142","added_by":"auto","created_at":"2026-05-04 06:02:40","extension":"jpg","order_by":9,"title":"Figure 9","display":"","copyAsset":false,"role":"figure","size":26578,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eObservation vs. survey means (left) and delta divergence with Cohen's d and Bonferroni-adjusted significance (right). *** p \u0026lt; .001; * p \u0026lt; .05 (corrected); ns = not significant after correction.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"9.jpg","url":"https://assets-eu.researchsquare.com/files/rs-9418492/v1/56e4c3331f5b7730382bfde9.jpg"},{"id":108803932,"identity":"ec6f1af0-e18b-46c3-a7ac-2686dcba1ede","added_by":"auto","created_at":"2026-05-08 15:11:52","extension":"jpg","order_by":10,"title":"Figure 10","display":"","copyAsset":false,"role":"figure","size":23311,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eVRSA-over-ARSA preference distribution (left) and perceived ease comparison (right; N = 64). W = 1128.0, p \u0026lt; .001 vs. neutral midpoint.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"10.jpg","url":"https://assets-eu.researchsquare.com/files/rs-9418492/v1/9ad865bd1b0c1b9f824dd770.jpg"},{"id":108385256,"identity":"b34e0d9a-5942-4fb2-8681-7b0056ec7b46","added_by":"auto","created_at":"2026-05-04 06:02:40","extension":"jpg","order_by":11,"title":"Figure 11","display":"","copyAsset":false,"role":"figure","size":25251,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eSurvey construct scores by prior VR experience group (N = 64; Bonferroni-corrected Mann-Whitney U). No construct reached adjusted significance (alpha = .007).\u003c/em\u003e\u003c/p\u003e","description":"","filename":"11.jpg","url":"https://assets-eu.researchsquare.com/files/rs-9418492/v1/31f2df465471893dfd810a9d.jpg"},{"id":108385257,"identity":"28115ef6-5dbf-4cea-a982-feb14c8f8ff6","added_by":"auto","created_at":"2026-05-04 06:02:40","extension":"jpg","order_by":12,"title":"Figure 12","display":"","copyAsset":false,"role":"figure","size":34473,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eSpearman correlation matrix across seven VRSA survey constructs (lower triangle; N = 64). * p \u0026lt; .05, ** p \u0026lt; .01, *** p \u0026lt; .001.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"12.jpg","url":"https://assets-eu.researchsquare.com/files/rs-9418492/v1/fb2423d3508a5c376893caaf.jpg"},{"id":108492747,"identity":"b757be5c-7bcb-4923-9fe8-c928aff152b7","added_by":"auto","created_at":"2026-05-05 09:58:31","extension":"jpg","order_by":13,"title":"Figure 13","display":"","copyAsset":false,"role":"figure","size":22790,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eTask completion time distribution (left); scatter plots against observed navigation ease (centre; rho = -0.287, p = .021) and survey-reported navigation ease (right; rho = -0.172, p = .173, ns).\u003c/em\u003e\u003c/p\u003e","description":"","filename":"13.jpg","url":"https://assets-eu.researchsquare.com/files/rs-9418492/v1/d416c298b9e77517702d1bde.jpg"},{"id":108808968,"identity":"01105864-1f7a-473e-8efc-048b3f3954dc","added_by":"auto","created_at":"2026-05-08 15:48:25","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":3278894,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9418492/v1/69f10d5f-2871-4b63-b57a-1622822d1f37.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Observational and Self-Report Divergences in Virtual Reality Shopping Evaluation: A Multi-Construct Mixed-Methods Study with Comparative AR Preference Analysis","fulltext":[{"header":"I. INTRODUCTION","content":"\u003cp\u003eThe rapid maturation of immersive computing hardware and the accelerating pace of digital retail transformation have positioned Virtual Reality (VR) and Augmented Reality (AR) as two of the most actively explored experiential technologies in commerce. VR generates a fully synthetic, computer-generated three-dimensional environment delivered through head-mounted displays (HMDs), offering users a sense of physical presence within a simulated space (Slater and Sanchez-Vives \u003cspan class=\"CitationRef\"\u003e2016\u003c/span\u003e). AR overlays digital content onto the physical environment, typically through smartphone cameras or optical see-through displays, preserving real-world spatial context while augmenting it with virtual information (Azuma \u003cspan class=\"CitationRef\"\u003e1997\u003c/span\u003e). Both technologies have attracted substantial commercial investment in retail: IKEA has deployed AR mobile product visualisation and VR kitchen design experiences, while Nike, Amazon, and Sephora have integrated AR try-on features into mobile commerce workflows (Hilken et al. \u003cspan class=\"CitationRef\"\u003e2022\u003c/span\u003e; Ricci et al. \u003cspan class=\"CitationRef\"\u003e2023\u003c/span\u003e).\u003c/p\u003e\n\u003cp\u003eDespite this commercial activity, the academic literature remains fragmented. Most empirical studies examine either VR or AR in isolation, and comparative studies frequently conflate technology type with application domain, product category, or participant population (Kim et al. \u003cspan class=\"CitationRef\"\u003e2023\u003c/span\u003e; Xi et al. \u003cspan class=\"CitationRef\"\u003e2024\u003c/span\u003e). A study that rigorously characterises the multi-dimensional user experience of VRSA using both direct behavioural observation and self-report instruments, and situates this characterisation within a comparative framework capturing user preference between VRSA and ARSA, would therefore constitute a meaningful empirical contribution.\u003c/p\u003e\n\u003cp\u003eThe retail context was chosen for four mutually reinforcing reasons. First, retail is among the most commercially immediate deployment domains for immersive technologies, where experiential quality directly influences purchase intention, product return rates, and brand perception (Sengupta and Cao \u003cspan class=\"CitationRef\"\u003e2022\u003c/span\u003e). Second, the product visualisation task is sufficiently multi-dimensional to exercise a broad range of user experience constructs including perceptual fidelity, spatial navigation, task efficiency, hedonic value, and utilitarian usefulness. Third, shopping tasks generate ecologically valid, observable behaviours amenable to structured observational coding. Fourth, the established literature on consumer technology acceptance provides well-validated theoretical frameworks within which to anchor findings.\u003c/p\u003e\n\u003cp\u003eThe study employs an asymmetric instrumentation design: the VRSA was evaluated using a full multi-construct observational rubric and post-experience survey, while the ARSA served as a comparative benchmark evaluated through a preference instrument and qualitative protocol. This design is consistent with within-subjects benchmarking paradigms in technology evaluation research (Nielsen \u003cspan class=\"CitationRef\"\u003e1993\u003c/span\u003e), and its implications for comparative claims are explicitly bounded throughout. Three principal contributions emerge from this work: (1) a rigorously instrumented multi-construct VRSA evaluation with full inferential statistical analysis across 64 participants, incorporating the first published systematic test of divergence between observational and self-report measurement modalities in VR retail research; (2) identification and theoretically grounded interpretation of systematic observational-to-survey divergences that carry direct methodological implications for the field; and (3) a five-theme qualitative preference taxonomy and evidence that optimal technology modality is product-category-dependent rather than universal. These contributions are deliberately scoped to what the present design can support and are discussed against their boundary conditions throughout.\u003c/p\u003e\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n"},{"header":"II. LITERATURE REVIEW","content":"\u003cp\u003e\u003cstrong\u003eA. Virtual Reality in Retail: Immersion, Presence, and Adoption\u003c/strong\u003e\u003c/p\u003e\u003cp\u003eResearch on VR-based retail applications has developed substantially over the past decade. Ricci et al. (\u003cspan class=\"CitationRef\"\u003e2023\u003c/span\u003e) examined experiential differences between immersive and desktop VR in virtual fashion retail, finding that HMD-delivered VR produced significantly higher presence and richer product examination behaviour. Slater and Sanchez-Vives (\u003cspan class=\"CitationRef\"\u003e2016\u003c/span\u003e) provided the foundational conceptual distinction between immersion, an objective technological property of the hardware, and presence, the subjective psychological experience of being there. This distinction is directly relevant to the divergences between observational and self-report measures documented in the present study.\u003c/p\u003e\u003cp\u003eHan et al. (\u003cspan class=\"CitationRef\"\u003e2023\u003c/span\u003e) applied the Stimulus-Organism-Response (SOR) model to VR shopping, finding that environmental stimuli in virtual retail spaces activate cognitive and affective organism states mediating purchase intention, with experiential value and hedonic arousal identified as organism-level mechanisms. Xi et al. (\u003cspan class=\"CitationRef\"\u003e2024\u003c/span\u003e) extended this line of research in a metaverse acceptance laboratory experiment, finding that VR significantly increased perceived enjoyment. Critically, Xi et al. (\u003cspan class=\"CitationRef\"\u003e2024\u003c/span\u003e) also identified HMD hardware usability constraints as the primary adoption barrier for VR, with participants reporting difficulty navigating immersive environments. This finding is relevant to the present navigational ease results and is distinct from the technology-acceptance profile of mobile AR, where such hardware constraints are absent.\u003c/p\u003e\u003cp\u003eCybersickness represents the most significant inhibitory factor for VR retail adoption. Weech et al. (\u003cspan class=\"CitationRef\"\u003e2019\u003c/span\u003e) documented that cybersickness and sense of presence are negatively related in VR environments, with individual susceptibility varying considerably. Rebenitsch and Owen (\u003cspan class=\"CitationRef\"\u003e2016\u003c/span\u003e) provided a comprehensive review of cybersickness in applications and visual displays, categorising symptoms and mitigation approaches across hardware, software, and behavioural domains. Stanney et al. (\u003cspan class=\"CitationRef\"\u003e2003\u003c/span\u003e) examined exposure durations of 15, 30, 45, and 60 minutes, identifying sustained increases in cybersickness beyond the 15-minute mark and recommending time-limited exposure protocols. The extended TAM for VR proposed by Sagnier et al. (\u003cspan class=\"CitationRef\"\u003e2020\u003c/span\u003e) augmented the foundational TAM (Davis \u003cspan class=\"CitationRef\"\u003e1989\u003c/span\u003e) by incorporating cybersickness, hedonic quality, and personal innovativeness, directly relevant to the present VRSA evaluation.\u003c/p\u003e\u003cp\u003e\u003cstrong\u003eB. Augmented Reality in Retail\u003c/strong\u003e\u003c/p\u003e\u003cp\u003eAR retail research emphasises practical usability, contextual product visualisation, and mobile consumer behaviour. Javornik (\u003cspan class=\"CitationRef\"\u003e2016\u003c/span\u003e) identified interactivity, realness, and augmentation quality as primary drivers of consumer response to AR, with interactivity as the strongest predictor of engagement. Flavian et al. (\u003cspan class=\"CitationRef\"\u003e2019\u003c/span\u003e) provided a comprehensive overview of AR in marketing, situating the technology's value generation as reduction of perceived purchase risk and enhancement of consumer decision quality. Lee et al. (\u003cspan class=\"CitationRef\"\u003e2022\u003c/span\u003e) applied the SOR framework to AR virtual try-on, finding that AR augmentation features stimulate a form of mobile telepresence amplifying both utilitarian and hedonic consumer value. Hoffmann et al. (\u003cspan class=\"CitationRef\"\u003e2022\u003c/span\u003e) concluded that AR's real-world spatial contextualisation represents a qualitatively different mode of product comprehension than VR's substitutive virtual environment. Cipresso et al. (\u003cspan class=\"CitationRef\"\u003e2018\u003c/span\u003e) conducted a network and cluster analysis of VR and AR research, identifying distinct research trajectories and application domains for each modality. Oyman et al. (\u003cspan class=\"CitationRef\"\u003e2022\u003c/span\u003e) further demonstrated that perceived AR quality positively influences perceived enjoyment and usefulness within extended TAM frameworks, reinforcing the role of hedonic and utilitarian pathways in mobile AR acceptance.\u003c/p\u003e\u003cp\u003e\u003cstrong\u003eC. Comparative Studies and Research Gap\u003c/strong\u003e\u003c/p\u003e\u003cp\u003eKim et al. (\u003cspan class=\"CitationRef\"\u003e2023\u003c/span\u003e) contrasted consumer perceptions between AR and VR using IKEA's commercial offerings, finding that VR conditions generated higher presence ratings than AR, while AR was associated with greater perceived interactivity in use. Hilken et al. (\u003cspan class=\"CitationRef\"\u003e2022\u003c/span\u003e) argued that context and product category should determine modality selection. Xi et al. (\u003cspan class=\"CitationRef\"\u003e2024\u003c/span\u003e) found technology-by-context interactions precluding universal modality superiority claims. No published study to date has employed a structured observational instrument alongside a self-report survey to evaluate VRSA across multiple constructs and statistically tested divergences between these measurement modalities in immersive retail contexts. The present study addresses this gap directly.\u003c/p\u003e\u003cp\u003e\u003cstrong\u003eD. Theoretical Framework\u003c/strong\u003e\u003c/p\u003e\u003cp\u003eThis study adopts TAM (Davis \u003cspan class=\"CitationRef\"\u003e1989\u003c/span\u003e) as its primary theoretical scaffold, extended by the VR-specific additions of Sagnier et al. (\u003cspan class=\"CitationRef\"\u003e2020\u003c/span\u003e) and the experiential value dimensions from the retail-specific literature (Han et al. \u003cspan class=\"CitationRef\"\u003e2023\u003c/span\u003e). The seven evaluated constructs collectively span TAM perceived usefulness and ease of use dimensions, VR-specific cybersickness and hedonic quality dimensions, and hedonic and affective dimensions from the SOR and experiential value literatures (Han et al. \u003cspan class=\"CitationRef\"\u003e2023\u003c/span\u003e; Lee et al. \u003cspan class=\"CitationRef\"\u003e2022\u003c/span\u003e).\u003c/p\u003e"},{"header":"III. METHODOLOGY","content":"\u003cp\u003e\u003cstrong\u003eA. Research Design\u003c/strong\u003e\u003c/p\u003e\u003cp\u003eThis study employed a mixed-methods design combining structured observational coding, post-experience self-report survey, and qualitative preference response analysis. The design is intentionally asymmetric: the VRSA was evaluated as the primary artefact using the full multi-construct instrument battery, while the ARSA served as a comparative benchmark evaluated through a preference instrument and qualitative protocol. This is consistent with within-subjects benchmarking paradigms (Nielsen \u003cspan class=\"CitationRef\"\u003e1993\u003c/span\u003e).\u003c/p\u003e\u003cp\u003e\u003cstrong\u003eB. Participants\u003c/strong\u003e\u003c/p\u003e\u003cp\u003eA total of 64 participants were recruited through purposive sampling from Middlesex University London. Age distribution: 60 (93.75%) aged 18 to 27, 2 (3.12%) aged 28 to 43, and 2 (3.12%) above 43 years. This demographic profile is consistent with the established early adopter demographic for immersive commerce technologies (Stecuła et al. \u003cspan class=\"CitationRef\"\u003e2024\u003c/span\u003e). Prior general VR experience was reported by 37 participants (57.81%), while 27 (42.19%) had no prior VR exposure. Among experienced users, the dominant usage frequency was less than once a month (n = 32; 86.49%). Prior VRSA experience was rare: only 7 of 37 VR-experienced participants (18.92%) had previously used a VR shopping application. All participants provided written informed consent; ethical approval was granted by the institutional review board.\u003c/p\u003e\u003cp\u003e\u003cstrong\u003eC. Apparatus and Stimuli\u003c/strong\u003e\u003c/p\u003e\u003cp\u003eThe VRSA comprised a commercial VR shopping mall application delivered via a Meta Quest 2 HMD, affording navigable three-dimensional retail space, product shelf browsing, spatial product inspection, and a virtual checkout workflow. The ARSA comprised a mobile AR shopping application deployed on an iOS device. Mobile AR was selected because it represents the dominant channel for consumer AR retail experiences (Hoffmann et al. \u003cspan class=\"CitationRef\"\u003e2022\u003c/span\u003e; Flavian et al. \u003cspan class=\"CitationRef\"\u003e2019\u003c/span\u003e). Product categories spanned consumer clothing, and home furniture to exercise the full range of evaluated constructs.\u003c/p\u003e\u003cp\u003e\u003cstrong\u003eD. Instrumentation\u003c/strong\u003e\u003c/p\u003e\u003cp\u003e\u003cstrong\u003eObservational Rubric.\u003c/strong\u003e An eight-item structured observational instrument captured behavioural indicators during each participant's VRSA session, rated on a five-point scale by two trained independent observers. Items assessed: (1) ease of navigation; (2) engagement with interactive elements; (3) immersion inferred from body language; (4) reaction to visual quality; (5) system speed and responsiveness; (6) discomfort indicators; (7) observed likelihood of adopting VRSA; and (8) task completion time in minutes. Inter-rater reliability assessed via Cohen's kappa returned values exceeding 0.75 for all items, indicating substantial to near-perfect agreement (Landis and Koch \u003cspan class=\"CitationRef\"\u003e1977\u003c/span\u003e). System Performance (item 5) returned SD = 0.000 across all 64 sessions, reflecting technically uniform application performance consistent with the Meta Quest 2's hardware reliability under controlled conditions. Because the observational System Performance measure has zero variance, it cannot be used in a paired statistical comparison; survey-observed divergence for this construct is therefore reported descriptively only and its statistical row in Table V is marked accordingly.\u003c/p\u003e\u003cp\u003e\u003cstrong\u003ePost-Experience Survey.\u003c/strong\u003e A 19-item instrument assessed demographic characteristics, prior technology experience, and seven evaluation constructs on five-point Likert scales immediately following the VRSA session, developed through expert validation (content validity index review by three domain experts). Internal consistency analysis returned Cronbach alpha = 0.783 for the full scale (acceptable; Nunnally \u003cspan class=\"CitationRef\"\u003e1978\u003c/span\u003e) and alpha = 0.776 for the usability sub-scale. The hedonic sub-scale returned alpha = 0.397, falling below acceptable reliability thresholds. This result indicates that immersion and visual quality function as distinct subjective dimensions rather than a unified hedonic construct in this context; accordingly, no composite hedonic score is computed or interpreted anywhere in this paper, and the two constructs are treated independently throughout. Items were drawn from TAM-derived measures (Davis \u003cspan class=\"CitationRef\"\u003e1989\u003c/span\u003e), presence-related scales (Slater and Sanchez-Vives \u003cspan class=\"CitationRef\"\u003e2016\u003c/span\u003e), and cybersickness-adjacent discomfort items informed by the Simulator Sickness Questionnaire (Kennedy et al. \u003cspan class=\"CitationRef\"\u003e1993\u003c/span\u003e).\u003c/p\u003e\u003cp\u003e\u003cstrong\u003eComparative Preference Instrument.\u003c/strong\u003e Two Likert-scale items assessed comparative preferences following both application sessions. Item 16 measured perceived ease comparison: \"How easy did you find the AR shopping app compared to the VR shopping app?\" (1 = AR much easier, 5 = VR much easier). Item 17 measured preference likelihood: \"If you had the option to use either the AR or VR shopping Apps for future purchases, how likely are you to choose the AR shopping app over the VR shopping App?\" (1 = much more likely to choose AR, 5 = much more likely to choose VR). A qualitative response item (Q18) asked participants to provide written reasons for their preference: \"Would you mind providing a reason for your previous answer?\" These written responses were transcribed for thematic analysis. Item 19 captured open-ended feature suggestions for future iterations.\u003c/p\u003e\u003cp\u003e\u003cstrong\u003eE. Procedure\u003c/strong\u003e\u003c/p\u003e\u003cp\u003eSessions were conducted in a controlled laboratory with three concurrent VR stations, processed in triads across 45-minute sessions. Participants provided informed consent and completed demographic sections before the session. VR-naive participants underwent a supervised tutorial before measured tasks. Standardised shopping tasks included product searching, multi-angle examination, and cart management. Application sequence was counterbalanced across sessions to control for order effects. VRSA interactions were capped at 20 minutes to prevent cybersickness accumulation consistent with established exposure duration recommendations (Stanney et al. \u003cspan class=\"CitationRef\"\u003e2003\u003c/span\u003e; Weech et al. \u003cspan class=\"CitationRef\"\u003e2019\u003c/span\u003e); ARSA sessions were uncapped, typically concluding within 10 to 15 minutes. Sessions were recorded by a 360-degree camera enabling post-hoc observational coding. The post-experience survey was administered immediately following each application session.\u003c/p\u003e\u003cp\u003e\u003cstrong\u003eF. Data Analysis\u003c/strong\u003e\u003c/p\u003e\u003cp\u003eShapiro-Wilk tests confirmed that all construct distributions deviated significantly from normality (all p \u0026lt; .001), justifying non-parametric procedures. Descriptive statistics were computed for all constructs. Internal consistency was assessed via Cronbach alpha. Paired Wilcoxon signed-rank tests were conducted to assess the significance of observational-to-survey divergences, with effect sizes reported as Cohen's d and rank-biserial r. All paired comparison p-values were Bonferroni-corrected for seven simultaneous tests (adjusted alpha = .007). System Performance, which returned SD = 0 in the observational instrument, was retained in the descriptive table for completeness but is excluded from inferential interpretation as a zero-variance distribution renders rank-based testing uninformative; this exclusion is noted explicitly in Table V. Spearman correlations characterised inter-construct relationships. Mann-Whitney U tests with Bonferroni correction examined VR experience moderation. Task completion time was analysed descriptively with outliers identified at M + 2SD, and Spearman correlations computed with navigation ease measures. Qualitative preference data were analysed using Braun and Clarke's (2006) six-phase thematic analysis framework, achieving intercoder agreement exceeding 82% prior to consensus resolution.\u003c/p\u003e"},{"header":"IV. FINDINGS","content":"\u003cp\u003e\u003cstrong\u003eA. Participant Profile\u003c/strong\u003e\u003c/p\u003e\u003cp\u003eTable I summarises sample characteristics. The high rate of VRSA-naive participants among the VR-experienced sub-group (81.08%) confirms that domain-specific novelty effects are present and must be considered when interpreting adoption intention scores.\u003c/p\u003e\u003cp\u003e\u003cstrong\u003eTABLE I\u003c/strong\u003e\u003c/p\u003e\u003cp\u003e\u003cem\u003eParticipant Demographics and Prior Technology Experience\u003c/em\u003e\u003c/p\u003e\u003cdiv class=\"gridtable\"\u003e\n\u003cdiv class=\"colspec\" align=\"char\"\u003e\u0026nbsp;\u003c/div\u003e\n\u003ctable id=\"Tabb\" border=\"1\"\u003e\n\u003cthead\u003e\n\u003ctr\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eVariable\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eCategory\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003en\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003e%\u003c/p\u003e\n\u003c/th\u003e\n\u003c/tr\u003e\n\u003c/thead\u003e\n\u003ctbody\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eAge Range\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e18 to 27\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e60\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e93.75\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e28 to 43\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e2\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e3.12\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eAbove 43\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e2\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e3.12\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003ePrior VR Experience\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eNo VR Experience\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e27\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e42.19\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eHas VR Experience\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e37\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e57.81\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eVR Frequency (n = 37)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eLess than once per month\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e32\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e86.49\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eA few times per month\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e2\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e5.41\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eAbout once per week or more\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e3\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e8.10\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003ePrior VRSA Experience (n = 37)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eNo prior VRSA use\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e30\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e81.08\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003ePrior VRSA use\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e7\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e18.92\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/tbody\u003e\n\u003ctfoot\u003e\n\u003ctr\u003e\n\u003ctd colspan=\"4\"\u003e\u003cem\u003eNote. VR Frequency percentages are of the VR-experienced sub-group (n = 37).\u003c/em\u003e\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/tfoot\u003e\n\u003c/table\u003e\n\u003c/div\u003e\u003cp\u003e\u003cstrong\u003eB. VRSA Observational Findings\u003c/strong\u003e\u003c/p\u003e\u003cp\u003eTable II and Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e5\u003c/span\u003e present the observational instrument means and standard deviations. Discomfort returned the highest score (M = 4.75, SD = 0.44), indicating that the large majority of participants completed the session without observable severe discomfort. Adoption intention was high (M = 4.47, SD = 0.64) and immersion also strong (M = 4.31, SD = 0.73), consistent with the capacity of HMD-delivered VR to generate high levels of objective immersion (Slater and Sanchez-Vives \u003cspan class=\"CitationRef\"\u003e2016\u003c/span\u003e). System Performance returned M = 4.00, SD = 0.000 (see Section III-D for explanation of zero variance). Ease of navigation returned the lowest observational score (M = 3.22, SD = 1.16), with the highest inter-participant variability of all constructs. Task completion time returned M = 2.51 min (SD = 1.06), right-skewed (skewness = 1.14), with two outliers exceeding M + 2SD (4.63 min): TU34 (6.00 min) and TU62 (6.06 min), both VR-naive.\u003c/p\u003e\u003cp\u003e\u003cstrong\u003eTABLE II\u003c/strong\u003e\u003c/p\u003e\u003cp\u003e\u003cem\u003eVRSA Observational Instrument: Means and Standard Deviations (N = 64)\u003c/em\u003e\u003c/p\u003e\u003cdiv class=\"gridtable\"\u003e\n\u003ctable id=\"Tabc\" border=\"1\"\u003e\n\u003cthead\u003e\n\u003ctr\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eConstruct\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eObs. Mean\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eSD\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eScale Anchors (5 = most positive)\u003c/p\u003e\n\u003c/th\u003e\n\u003c/tr\u003e\n\u003c/thead\u003e\n\u003ctbody\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eAdoption Intention\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e4.47\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e0.64\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e1 = Very Unlikely ... 5 = Very Likely\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eImmersion\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e4.31\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e0.73\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e1 = Not Immersed ... 5 = Fully Immersed\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eVisual Quality\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e4.19\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e0.81\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e1 = Very Dissatisfied ... 5 = Very Satisfied\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eInteractivity\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e3.44\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e0.81\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e1 = No Engagement ... 5 = Highly Engaged\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eSystem Performance\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e4.00\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e0.00*\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e1 = Very Slow ... 5 = Very Fast\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eDiscomfort (no-discomfort +ve)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e4.75\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e0.44\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e1 = Severe Discomfort ... 5 = No Discomfort\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eEase of Navigation\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e3.22\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e1.16\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e1 = Frequent Hesitation ... 5 = Seamless Navigation\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eTask Completion Time\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e2.51 min\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e1.06\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eRecorded in minutes\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/tbody\u003e\n\u003ctfoot\u003e\n\u003ctr\u003e\n\u003ctd colspan=\"4\"\u003e\u003cem\u003eNote. *SD = 0.000: all 64 observers assigned score 4, reflecting uniform application performance. Paired inferential testing is not computed for this construct.\u003c/em\u003e\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/tfoot\u003e\n\u003c/table\u003e\n\u003c/div\u003e\u003cp\u003e\u0026nbsp;\u003c/p\u003e\u003cp\u003e\u003cstrong\u003eC. VRSA Post-Experience Survey Findings\u003c/strong\u003e\u003c/p\u003e\u003cp\u003eTable III presents survey means, SDs, observational means, and divergence scores. Visual quality (M = 4.11, SD = 0.72) and system performance (M = 4.28, SD = 0.81) returned strong survey scores consistent with observational counterparts. Adoption intention remained high (M = 4.08, SD = 0.99). Two constructs showed notable divergences: immersion registered M = 3.64 (SD = 0.91) in the survey versus 4.31 in observation (delta = + 0.67), and ease of navigation registered M = 4.20 (SD = 0.86) in the survey versus 3.22 in observation (delta = -0.98). Statistical testing of these divergences is presented in Section IV-E. The VRSA-over-ARSA preference item returned M = 3.95 (SD = 1.20).\u003c/p\u003e\u003cp\u003e\u003cstrong\u003eTABLE III\u003c/strong\u003e\u003c/p\u003e\u003cp\u003e\u003cem\u003eVRSA Post-Experience Survey and Observational Means with Divergence Scores (N = 64)\u003c/em\u003e\u003c/p\u003e\u003cdiv class=\"gridtable\"\u003e\n\u003cdiv class=\"colspec\" align=\"left\"\u003e\u0026nbsp;\u003c/div\u003e\n\u003cdiv class=\"colspec\" align=\"left\"\u003e\u0026nbsp;\u003c/div\u003e\n\u003ctable id=\"Tabd\" border=\"1\"\u003e\n\u003cthead\u003e\n\u003ctr\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eConstruct\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eSurvey M\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eSD\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eObs. M\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eDelta (Obs. minus Survey)\u003c/p\u003e\n\u003c/th\u003e\n\u003c/tr\u003e\n\u003c/thead\u003e\n\u003ctbody\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eAdoption Intention\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e4.08\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e0.99\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e4.47\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e+ 0.39\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eImmersion\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e3.64\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e0.91\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e4.31\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e+ 0.67***\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eVisual Quality\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e4.11\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e0.72\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e4.19\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e+ 0.08\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eInteractivity\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e3.91\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e0.88\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e3.44\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e-0.47*\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eSystem Performance\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e4.28\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e0.81\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e4.00\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e-0.28*\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eDiscomfort (no-discomfort +ve)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e4.48\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e0.89\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e4.75\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e+ 0.27\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eEase of Navigation\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e4.20\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e0.86\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e3.22\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e-0.98***\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eVRSA-over-ARSA Preference\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e3.95\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e1.20\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eN/A\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eN/A\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/tbody\u003e\n\u003ctfoot\u003e\n\u003ctr\u003e\n\u003ctd colspan=\"5\"\u003e\u003cem\u003eNote. *** Bonferroni-corrected significant (p \u0026lt; .001); * Bonferroni-corrected significant (p \u0026lt; .05). Full statistical details in Table V.\u003c/em\u003e\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/tfoot\u003e\n\u003c/table\u003e\n\u003c/div\u003e\u003cp\u003e\u0026nbsp;\u003c/p\u003e\u003cp\u003e\u003cstrong\u003eD. Qualitative Preference Themes\u003c/strong\u003e\u003c/p\u003e\u003cp\u003eThematic analysis of the comparative preference protocol yielded five primary themes (Table IV). Theme 1, Immersion and Realism, was the most frequently coded theme among VRSA-preferring participants, framing VR as qualitatively closer to a physical retail visit. Theme 2, Ease of Use, was bidirectional: VRSA-preferring participants cited the VR interaction model as ultimately natural and embodied; ARSA-preferring participants cited the lower cognitive overhead of the smartphone interface. Theme 3, Entertainment and Novelty, was exclusively associated with VRSA preference, foregrounding hedonic motivation. Theme 4, Product Display Fidelity, captured responses explicitly comparing three-dimensional VR product presentation against the flat AR overlay. Theme 5, Contextual Practicality, dominated among ARSA-preferring and neutral participants, emphasising product-category-dependent optimal modality selection.\u003c/p\u003e\u003cp\u003e\u003cstrong\u003eTABLE IV\u003c/strong\u003e\u003c/p\u003e\u003cp\u003e\u003cem\u003eThematic Analysis of VRSA-vs-ARSA Comparative Preference Responses\u003c/em\u003e\u003c/p\u003e\u003cdiv class=\"gridtable\"\u003e\n\u003cdiv class=\"colspec\" align=\"left\"\u003e\u0026nbsp;\u003c/div\u003e\n\u003cdiv class=\"colspec\" align=\"left\"\u003e\u0026nbsp;\u003c/div\u003e\n\u003ctable id=\"Tabe\" border=\"1\"\u003e\n\u003cthead\u003e\n\u003ctr\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eTheme\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eVRSA/ARSA orientation\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eRepresentative extract\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eTheoretical relevance\u003c/p\u003e\n\u003c/th\u003e\n\u003c/tr\u003e\n\u003c/thead\u003e\n\u003ctbody\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e1. Immersion and Realism\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003ePro-VRSA\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e\"It felt like I was actually in a store.\"\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003ePresence as qualitative retail analogue\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e2. Ease of Use\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eBidirectional\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e\"VR felt natural once I got used to it.\" / \"AR is just simpler.\"\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eTechnology familiarity moderates ease perception\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e3. Entertainment and Novelty\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003ePro-VRSA\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e\"It was fun to explore the store.\"\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eHedonic motivation as adoption pathway\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e4. Product Display Fidelity\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003ePro-VRSA\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e\"You could see the product from all angles.\"\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eSpatial product comprehension advantage of VR\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e5. Contextual Practicality\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003ePro-ARSA / Neutral\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e\"AR for large items; VR for visualising clothes.\" / \"Discomfort with VR headsets.\"\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eContext and product category moderate optimal technology selection\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/tbody\u003e\n\u003ctfoot\u003e\n\u003ctr\u003e\n\u003ctd colspan=\"4\"\u003e\u003cem\u003eNote. Themes ordered by prevalence. Extracts are verbatim participant statements.\u003c/em\u003e\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/tfoot\u003e\n\u003c/table\u003e\n\u003c/div\u003e\u003cp\u003e\u003cstrong\u003eE. Statistical Testing of Observational-to-Survey Divergences\u003c/strong\u003e\u003c/p\u003e\u003cp\u003eTable V presents the full statistical results for paired Wilcoxon signed-rank tests comparing observational and survey means for each construct, with Bonferroni-corrected significance (adjusted alpha = .007). Two constructs showed Bonferroni-corrected significant divergences: Ease of Navigation (W = 132.5, p_adj \u0026lt; .001, d = -0.748, r_rb = 0.936) and Immersion (W = 115.0, p_adj \u0026lt; .001, d = 0.603, r_rb = 0.945). Two constructs showed uncorrected-significant but correction-failing divergences: Interactivity (p_adj = .025) and System Performance (p_adj = .021). Visual Quality, Discomfort, and Adoption Intention showed no significant divergence after correction. System Performance is retained descriptively in Table V, with the statistical result marked as non-inferential given the zero observational variance.\u003c/p\u003e\u003cp\u003e\u003cstrong\u003eTABLE V\u003c/strong\u003e\u003c/p\u003e\u003cp\u003e\u003cem\u003ePaired Wilcoxon Signed-Rank Tests: Observational vs. Survey Construct Means (N = 64)\u003c/em\u003e\u0026nbsp;\u003c/p\u003e\u003cdiv class=\"gridtable\"\u003e\n\u003cdiv class=\"colspec\" align=\"left\"\u003e\u0026nbsp;\u003c/div\u003e\n\u003ctable id=\"Tabf\" border=\"1\"\u003e\n\u003cthead\u003e\n\u003ctr\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eConstruct\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eObs M\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eObs SD\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eSv M\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eSv SD\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eDelta\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eW\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003ep raw\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003ep adj\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003ed\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003er_rb\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eSig\u003c/p\u003e\n\u003c/th\u003e\n\u003c/tr\u003e\n\u003c/thead\u003e\n\u003ctbody\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eEase of Navigation\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e3.22\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e1.15\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e4.20\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e0.85\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e-0.98\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e132.5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e\u0026lt; .001\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e\u0026lt; .001\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e-0.748\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e0.936\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e***\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eImmersion\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e4.31\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e0.73\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e3.64\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e0.91\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e+ 0.67\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e115.0\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e\u0026lt; .001\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e\u0026lt; .001\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e0.603\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e0.945\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e***\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eVisual Quality\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e4.19\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e0.81\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e4.11\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e0.71\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e+ 0.08\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e339.0\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e.630\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e1.000\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e0.073\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e0.837\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003ens\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eInteractivity\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e3.44\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e0.81\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e3.91\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e0.86\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e-0.47\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e285.5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e.004\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e.025\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e-0.380\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e0.863\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e*\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eSystem Performance\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e4.00\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e0.00\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e4.28\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e0.80\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e-0.28\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e190.0\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e.003\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e.021\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e-0.349\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e0.909\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e* (a)\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eDiscomfort\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e4.75\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e0.43\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e4.48\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e0.88\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e+ 0.27\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e136.5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e.040\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e.281\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e0.251\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e0.934\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003ens\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eAdoption Intention\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e4.47\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e0.64\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e4.08\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e0.99\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e+ 0.39\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e220.0\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e.014\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e.096\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e0.318\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e0.894\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003ens\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/tbody\u003e\n\u003ctfoot\u003e\n\u003ctr\u003e\n\u003ctd colspan=\"12\"\u003e\u003cem\u003eNote. Delta = Obs M minus Survey M. p adj = Bonferroni-corrected (adjusted alpha = .007). d = Cohen's d; r_rb = rank-biserial correlation (non-parametric effect size). *** p \u0026lt; .001; * p \u0026lt; .05 (Bonferroni-corrected); ns = not significant after correction. (a) System Performance: SD = 0 in observation; test is retained for descriptive reference only and results should not be interpreted inferentially.\u003c/em\u003e\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/tfoot\u003e\n\u003c/table\u003e\n\u003c/div\u003e\u003cp\u003e\u003cstrong\u003eF. AR versus VR Preference and Ease\u003c/strong\u003e\u003c/p\u003e\u003cp\u003eThe VRSA-over-ARSA preference item returned M = 3.95 (SD = 1.20, Mdn = 4.0). A one-sample Wilcoxon test confirmed that preference significantly exceeded the neutral midpoint of 3 (W = 1128.0, p \u0026lt; .001), indicating a statistically supported moderate lean toward VR. Perceived ease (VR relative to AR) returned M = 3.75 (SD = 1.15). Figure\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e10\u003c/span\u003e presents the full distributions.\u003c/p\u003e\u003cp\u003e\u003cstrong\u003eG. VR Experience Moderation\u003c/strong\u003e\u003c/p\u003e\u003cp\u003eTable VI presents Mann-Whitney U results comparing survey construct scores by VR experience group. No construct showed a statistically significant difference after Bonferroni correction, indicating that prior VR experience did not systematically moderate survey responses. The largest effect size was for adoption intention (r = 0.223), with VR-naive participants reporting marginally higher adoption intention (M = 4.26) than experienced participants (M = 3.95). Figure\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e11\u003c/span\u003e presents the boxplots.\u003c/p\u003e\u003cp\u003e\u003cstrong\u003eTABLE VI\u003c/strong\u003e\u003c/p\u003e\u003cp\u003e\u003cem\u003eMann-Whitney U Tests: Survey Construct Scores by Prior VR Experience Group\u003c/em\u003e\u003c/p\u003e\u003cdiv class=\"gridtable\"\u003e\n\u003ctable id=\"Tabg\" border=\"1\"\u003e\n\u003cthead\u003e\n\u003ctr\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eConstruct\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eExp M\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eExp SD\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eNaive M\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eNaive SD\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eU\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003ep raw\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003ep adj\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003er\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eSig\u003c/p\u003e\n\u003c/th\u003e\n\u003c/tr\u003e\n\u003c/thead\u003e\n\u003ctbody\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eEase of Navigation\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e4.27\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e0.90\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e4.11\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e0.80\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e569.5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e.309\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e1.000\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e-0.140\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003ens\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eImmersion\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e3.65\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e0.75\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e3.63\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e1.11\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e480.0\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e.783\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e1.000\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e0.039\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003ens\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eVisual Quality\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e4.08\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e0.68\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e4.15\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e0.77\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e474.5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e.713\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e1.000\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e0.050\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003ens\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eInteractivity\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e3.84\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e0.87\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e4.00\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e0.88\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e464.0\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e.613\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e1.000\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e0.071\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003ens\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eSystem Performance\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e4.22\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e0.85\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e4.37\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e0.74\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e452.5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e.490\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e1.000\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e0.094\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003ens\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eDiscomfort\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e4.46\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e0.93\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e4.52\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e0.85\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e492.5\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e.913\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e1.000\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e0.014\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003ens\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eAdoption Intention\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e3.95\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e0.97\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e4.26\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e1.02\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e388.0\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e.108\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e.756\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"char\"\u003e\n\u003cp\u003e0.223\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003ens\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/tbody\u003e\n\u003ctfoot\u003e\n\u003ctr\u003e\n\u003ctd colspan=\"10\"\u003e\u003cem\u003eNote. Exp = VR-experienced (n = 37); Naive = VR-naive (n = 27). p adj = Bonferroni-corrected (adjusted alpha = .007). r = rank-biserial. ns = not significant after correction.\u003c/em\u003e\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/tfoot\u003e\n\u003c/table\u003e\n\u003c/div\u003e\u003cp\u003e\u003cstrong\u003eH. Inter-Construct Spearman Correlations\u003c/strong\u003e\u003c/p\u003e\u003cp\u003eSpearman correlations revealed a cluster of strong associations among the perceptual quality dimensions. The strongest correlation was Visual Quality with Interactivity (rho = 0.70, p \u0026lt; .001), followed by Interactivity with System Performance (rho = 0.61, p \u0026lt; .001) and Visual Quality with System Performance (rho = 0.47, p \u0026lt; .001). Ease of Navigation correlated significantly with Visual Quality (rho = 0.41, p \u0026lt; .001) and System Performance (rho = 0.40, p \u0026lt; .001). Discomfort showed no significant correlations with other constructs, consistent with cybersickness susceptibility being an individual physiological characteristic largely independent of perceived experiential quality. Adoption Intention showed significant correlations with Ease of Navigation (rho = 0.26, p = .038), Visual Quality (rho = 0.25, p = .050), Interactivity (rho = 0.30, p = .015), and Discomfort (rho = 0.31, p = .012). Figure\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e12\u003c/span\u003e presents the full matrix.\u0026nbsp;\u003c/p\u003e\u003ch3\u003eI. Task Completion Time Analysis\u003c/h3\u003e\u003cp\u003eTask completion time returned M = 2.51 min (SD = 1.06, Mdn = 2.36, skewness = 1.14). Two participants (TU34: 6.00 min; TU62: 6.06 min) exceeded M + 2SD (4.63 min); both had no prior VR experience. Spearman correlation between observed navigation ease and task time was rho = -0.287 (p = .021), confirming that participants rated as navigating more easily by observers completed the task faster. Survey-reported navigation ease to task time correlation was weaker and non-significant (rho = -0.172, p = .173). Figure\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e13\u003c/span\u003e presents the distribution and scatter plots.\u003c/p\u003e"},{"header":"V. DISCUSSION","content":"\u003cp\u003e\u003cstrong\u003eA. Contextualising VRSA Findings Within the Extant Literature\u003c/strong\u003e\u003c/p\u003e\u003cp\u003eThe VRSA adoption intention means of 4.08 (survey) and 4.47 (observation) are among the highest reported in comparable single-session VR retail evaluations. Kim et al. (\u003cspan class=\"CitationRef\"\u003e2023\u003c/span\u003e) reported strong VR adoption intentions in their comparative IKEA-based study, attributing the effect primarily to hedonic quality rather than utilitarian efficiency. The present findings are directionally consistent: qualitative Theme 3 (Entertainment and Novelty) identifies hedonic motivation as a primary preference driver, and the significant Spearman correlation between interactivity and adoption intention (rho = 0.30, p = .015) provides quantitative support for hedonic engagement as a pathway to adoption in this context. However, the single-session design means novelty inflation cannot be excluded as a contributor. Xi et al. (\u003cspan class=\"CitationRef\"\u003e2024\u003c/span\u003e) acknowledge the external validity constraints of laboratory-based XR shopping studies, noting that single-session conditions may not fully capture longitudinal adoption dynamics.\u003c/p\u003e\u003cp\u003eThe observed immersion mean of 4.31 is consistent with the high objective immersion documented by Ricci et al. (\u003cspan class=\"CitationRef\"\u003e2023\u003c/span\u003e) for HMD-delivered VR fashion retail. The significant Bonferroni-corrected divergence between observed immersion (M = 4.31) and survey immersion (M = 3.64, W = 115.0, p \u0026lt; .001, d = 0.603) reflects the conceptual distinction between immersion as a hardware property and presence as a subjective psychological state (Slater and Sanchez-Vives \u003cspan class=\"CitationRef\"\u003e2016\u003c/span\u003e). The large effect size indicates that this is not a measurement artefact: the Meta Quest 2 hardware delivered high objective immersion, but this was not fully translated into equivalent subjective presence. Likely disruptors include cybersickness episodes, navigational friction (reflected in the low navigation ease scores), and the unfamiliarity of VR interaction for first-time users.\u003c/p\u003e\u003cp\u003eThe strong intercorrelations among visual quality, interactivity, and system performance (rho = 0.47 to 0.70, all p \u0026lt; .001) are theoretically interpretable as a coherent perceptual quality factor, consistent with the SOR framework of Han et al. (\u003cspan class=\"CitationRef\"\u003e2023\u003c/span\u003e) in which stimulus richness functions as an integrated experiential input. That these three constructs converge while immersion and discomfort remain relatively uncorrelated with the quality cluster suggests that the seven-construct battery captures at least two distinct experiential dimensions: a perceptual quality dimension (visual quality, interactivity, performance, navigation ease) and an experiential depth dimension (immersion, discomfort). The low hedonic sub-scale alpha (0.397) corroborates this bifurcation: immersion and visual quality, though both hedonic in character, are not measuring the same subjective construct in VR retail contexts. This finding has direct implications for the design of future VRSA evaluation instruments and is discussed further in the Limitations section.\u003c/p\u003e\u003cp\u003e\u003cstrong\u003eB. The Navigational Ease Reversal: Mechanisms and Theoretical Grounding\u003c/strong\u003e\u003c/p\u003e\u003cp\u003eThe most methodologically significant finding is the Bonferroni-corrected divergence of -0.98 between the observational navigation ease mean (M = 3.22, SD = 1.16) and the survey mean (M = 4.20, SD = 0.86; W = 132.5, p \u0026lt; .001, d = -0.748, r_rb = 0.936). The large rank-biserial r of 0.936 indicates that in nearly every matched participant pair, the observer rated navigation as harder than the participant self-reported. Three non-mutually exclusive mechanisms are proposed, each grounded in established cognitive psychology.\u003c/p\u003e\u003cp\u003eFirst, retrospective competence anchoring (Nielsen \u003cspan class=\"CitationRef\"\u003e1993\u003c/span\u003e): participants evaluated their navigation at the skill level attained by session end rather than averaging across the full session trajectory, producing an overly optimistic assessment. This endpoint-weighting is consistent with the peak-end rule (Kahneman et al. \u003cspan class=\"CitationRef\"\u003e1993\u003c/span\u003e), which demonstrates that retrospective evaluations are disproportionately influenced by the end-state and peak moment rather than the integrated temporal experience. Applied to VR navigation, a participant who struggled for the first ten minutes but navigated fluently in the final five would retrospectively report high ease, while an observer coding the full session would record a lower mean. This mechanism is directly supported by the observational SD of 1.16 being substantially larger than the survey SD of 0.86: observer ratings captured the full heterogeneous trajectory while self-report compressed it toward the endpoint.\u003c/p\u003e\u003cp\u003eSecond, contrast effects from counterbalanced ordering: participants who experienced the ARSA before the VRSA may have implicitly contrasted VR navigation against the initially unfamiliar mobile AR interface, inflating their retrospective evaluation of VR navigation relative to an unstated AR baseline. Contrast effects of this type are well documented in comparative judgement paradigms (Schwarz and Bless \u003cspan class=\"CitationRef\"\u003e1992\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eThird, social desirability bias: participants may have rated navigation favourably in the survey context to appear competent in a structured academic environment with observers present. This mechanism would specifically affect ego-threatening constructs such as navigation competence.\u003c/p\u003e\u003cp\u003eThe practical implication for VRSA research and design is direct: post-experience surveys alone will systematically overestimate navigational ease in VRSA contexts. The Spearman correlation between observed navigation ease and task time (rho = -0.287, p = .021), compared to the non-significant correlation for survey-reported ease (rho = -0.172, p = .173), provides direct empirical support: the observed measure predicted objective task performance, while the survey measure did not. Observational or task-timing data are therefore essential complements for evaluating navigational usability in VR retail contexts.\u003c/p\u003e\u003cp\u003e\u003cstrong\u003eC. Interactivity and System Performance: Upward Survey Modulation\u003c/strong\u003e\u003c/p\u003e\u003cp\u003eInteractivity showed a Bonferroni-corrected significant divergence (W = 285.5, p_adj = .025, d = -0.380), with the survey mean (M = 3.91) substantially exceeding the observational mean (M = 3.44). System Performance showed a similarly directed divergence (W = 190.0, p_adj = .021, d = -0.349); however, as noted in Section III-D, the zero observational variance for System Performance renders this result descriptively informative rather than inferentially valid, and no causal interpretation is advanced. The interactivity divergence suggests that participants valued the interactive affordances more highly in retrospect than their in-session engagement patterns indicated. This pattern is consistent with the broader finding in technology acceptance research that retrospective evaluations tend to be more positive than concurrent behavioural measures, as users consolidate their understanding of a technology's potential during reflection. The implication for VRSA design is that interactive feature discoverability may be limiting in-session engagement even when those features are retrospectively valued, pointing to the need for clearer onboarding and affordance signalling within VR retail environments.\u003c/p\u003e\u003cp\u003eThe absence of Bonferroni-corrected significance for discomfort (W = 136.5, p_adj = .281) and adoption intention (W = 220.0, p_adj = .096) indicates that these constructs are relatively stably perceived across measurement modalities. Discomfort stability is particularly noteworthy: cybersickness is an acute physiological experience that is difficult to misremember or retrospectively reframe, which accounts for its measurement consistency. Adoption intention stability suggests that participant enthusiasm for future VRSA use was genuine and not subject to the same retrospective inflation mechanisms affecting navigation ease.\u003c/p\u003e\u003cp\u003e\u003cstrong\u003eD. Contextualising Comparative Preference Findings\u003c/strong\u003e\u003c/p\u003e\u003cp\u003eThe overall VRSA-over-ARSA preference mean of 3.95 (SD = 1.20) significantly exceeded the neutral midpoint (W = 1128.0, p \u0026lt; .001), confirming a statistically supported moderate lean toward VR. This is directionally consistent with Kim et al. (\u003cspan class=\"CitationRef\"\u003e2023\u003c/span\u003e), who found VR conditions generated higher presence ratings than AR in a within-subjects IKEA comparison, an effect those authors attributed primarily to the immersive hardware advantage of HMD-delivered environments. The SD of 1.20 reflects genuine individual heterogeneity, echoing Hilken et al. (\u003cspan class=\"CitationRef\"\u003e2022\u003c/span\u003e) strategic argument that neither technology dominates universally and that context, product category, and consumer disposition interact to shape technology preference.\u003c/p\u003e\u003cp\u003eThe qualitative preference data add critical nuance to the quantitative score. Theme 5 (Contextual Practicality) documents that participants explicitly nominated different optimal technologies for different product categories: AR for large furniture items due to real-world spatial placement, VR for clothing due to immersive try-on experience. This contextual moderation is consistent with the Hilken et al. (\u003cspan class=\"CitationRef\"\u003e2022\u003c/span\u003e) framework and with Lee et al. (\u003cspan class=\"CitationRef\"\u003e2022\u003c/span\u003e) finding that AR telepresence generates distinct forms of product comprehension from VR spatial presence. The optimal immersive modality is product-dependent, not universal, constituting a theoretically grounded product-category-by-modality interaction with direct implications for retailer technology strategy.\u003c/p\u003e\u003cp\u003eThe bidirectionality of Theme 2 (Ease of Use) is theoretically informative. Among VRSA-preferring participants, the VR interaction model was characterised as natural and embodied once the learning curve was surmounted. Among ARSA-preferring participants, the smartphone interface was valued for its lower cognitive overhead and familiar interaction paradigm. This bidirectionality corroborates Sagnier et al. (\u003cspan class=\"CitationRef\"\u003e2020\u003c/span\u003e) finding that prior technology familiarity moderates ease of use perception in VR contexts, and suggests that the perceived ease advantage of VR may be contingent on sufficient onboarding experience to overcome the initial navigational friction documented in the observational data.\u003c/p\u003e\u003cp\u003e\u003cstrong\u003eE. Limitations and Boundary Conditions\u003c/strong\u003e\u003c/p\u003e\u003cp\u003eSeven limitations bound the interpretation of these findings and should be carefully considered before generalising the results.\u003c/p\u003e\u003cp\u003eFirst, the sample was drawn entirely from a single university in the United Kingdom and is demographically homogeneous: 93.75% of participants were aged 18 to 27. This profile is characteristic of early-adopter convenience samples in immersive technology research, but it substantially constrains generalisability to broader consumer populations. Older adults, in particular, may exhibit greater technology anxiety (Sagnier et al. \u003cspan class=\"CitationRef\"\u003e2020\u003c/span\u003e), distinct spatial navigation profiles, and different hedonic motivations than the present sample. All quantitative findings and preference patterns reported here apply to this demographic cohort and should not be extended to general consumer populations without replication across age groups and cultural contexts.\u003c/p\u003e\u003cp\u003eSecond, the sample size of 64, while adequate for the non-parametric procedures employed, is modest relative to the range of moderation hypotheses that could plausibly be investigated. The VR experience moderation analysis in particular is underpowered to detect small-to-moderate interaction effects; the absence of significant moderation should be treated as inconclusive rather than confirmatory for that analysis.\u003c/p\u003e\u003cp\u003eThird, the asymmetric instrumentation design means that construct-level quantitative comparisons between VRSA and ARSA are not possible. The ARSA was not evaluated using the full seven-construct rubric and survey battery. Accordingly, all claims framed as comparative are limited to user-stated preferences and qualitative themes, and no inferential claim about ARSA construct means is advanced anywhere in this paper.\u003c/p\u003e\u003cp\u003eFourth, the single-session design cannot distinguish novelty-driven from utility-driven adoption intention, a critical distinction for predicting real-world deployment behaviour (Xi et al. \u003cspan class=\"CitationRef\"\u003e2024\u003c/span\u003e). The high adoption intention scores may be partially attributable to first-exposure enthusiasm rather than stable preference, particularly given that 81.08% of the VR-experienced sub-group had never used a VR shopping application before.\u003c/p\u003e\u003cp\u003eFifth, while counterbalancing controls for order effects, it does not eliminate them. The asymmetric session durations (VRSA capped at 20 minutes; ARSA uncapped at approximately 10 to 15 minutes) introduce a structural inequality in exposure time that may have influenced comparative preference ratings in ways not fully captured by the counterbalancing procedure.\u003c/p\u003e\u003cp\u003eSixth, the hedonic sub-scale reliability of alpha = 0.397 is well below acceptable thresholds. This finding is interpreted in this paper as evidence that immersion and visual quality are empirically distinct constructs in VR retail contexts, and no composite hedonic score is reported. Future instrument development should treat these as separate subscales from the outset.\u003c/p\u003e\u003cp\u003eSeventh, the social desirability mechanism proposed for the navigational ease reversal is theoretically plausible but was not directly tested. A follow-up study incorporating implicit navigation performance measures alongside self-report would be needed to isolate this mechanism from the peak-end and contrast-effects accounts.\u003c/p\u003e"},{"header":"VI. PRACTITIONER IMPLICATIONS","content":"\u003cp\u003eThe findings yield several actionable recommendations for VRSA designers, ARSA designers, and retail technology strategists.\u003c/p\u003e \u003cp\u003e \u003cb\u003eVRSA Navigation and Onboarding Design.\u003c/b\u003e The observational navigation ease mean of 3.22 and the significant negative correlation with task completion time (rho = -0.287, p = .021) confirm that VR spatial navigation represents the most significant usability barrier in current VRSA implementations. Designers should implement progressive onboarding with explicit spatial navigation affordances, teleportation as a fallback locomotion option, and visual wayfinding cues. The two task time outliers (TU34 and TU62, both VR-naive, both exceeding 6 minutes) suggest that VR-naive users require more substantial onboarding than a supervised tutorial alone provides.\u003c/p\u003e \u003cp\u003e \u003cb\u003eMethodological Triangulation for VRSA Evaluation.\u003c/b\u003e The systematic divergence between observed and self-reported navigation ease (d = -0.748) demonstrates that post-experience surveys alone are insufficient for assessing navigational usability in VRSA contexts. Practitioners should complement survey instruments with observational coding or task-timing data. Discomfort and adoption intention are relatively stable across modalities and can be assessed reliably through survey instruments alone.\u003c/p\u003e \u003cp\u003e \u003cb\u003eInteractive Feature Discoverability.\u003c/b\u003e The upward modulation of interactivity in the survey relative to observation (d = -0.380) suggests that participants valued interactive VR affordances more highly in reflection than they utilised them in session. VRSA designers should invest in interactive feature discoverability: clear visual and haptic cues indicating available interactions, tutorial prompts for key interactive features, and progressive feature introduction.\u003c/p\u003e \u003cp\u003e \u003cb\u003eContext-Sensitive Technology Deployment.\u003c/b\u003e The quantitative preference lean toward VR (W\u0026thinsp;=\u0026thinsp;1128.0, p \u0026lt; .001) combined with qualitative evidence for product-category-specific modality preferences (Theme 5) supports a context-sensitive deployment model. Retailers with spatially large product categories such as furniture and appliances may find AR's real-world contextualisation to be the stronger modality. Retailers with wearable or highly visual product categories such as clothing and footwear may derive greater consumer value from VR's immersive spatial presentation. A hybrid deployment strategy exploiting both modalities for their respective product-category advantages is recommended (Hilken et al. \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2022\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e\u003cb\u003eCybersickness Mitigation.\u003c/b\u003e Despite generally positive discomfort scores, the non-zero variance indicates a minority experiencing clinically meaningful distress. Established cybersickness mitigation strategies including vignetting during locomotion, field-of-view restriction during rapid movement, and explicit session time limits should be implemented (Rebenitsch and Owen \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2016\u003c/span\u003e; Weech et al. \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). The 20-minute session cap applied in this study appears appropriate as a deployment guideline (Stanney et al. \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2003\u003c/span\u003e).\u003c/p\u003e"},{"header":"VII. CONCLUSION","content":"\u003cp\u003eThis paper has presented a rigorously instrumented mixed-methods evaluation of VRSA user experience, drawing on triangulated observational and self-report data from 64 participants supplemented by comparative preference analysis against an ARSA. Three contributions, scoped to what the present design permits, have been delivered.\u003c/p\u003e \u003cp\u003eThe first and primary contribution is the identification and statistical testing of systematic divergences between observational and self-report measurement modalities in VR retail research. This is, to the authors' knowledge, the first published study to apply paired non-parametric testing with effect size reporting to observational-versus-survey divergences across multiple constructs in a VRSA context. The immersion divergence (d\u0026thinsp;=\u0026thinsp;0.603, W\u0026thinsp;=\u0026thinsp;115.0, p \u0026lt; .001) is theoretically interpreted through the immersion-presence distinction (Slater and Sanchez-Vives \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2016\u003c/span\u003e), while the navigational ease reversal (d = -0.748, W\u0026thinsp;=\u0026thinsp;132.5, p \u0026lt; .001) is grounded in the peak-end rule (Kahneman et al. \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e1993\u003c/span\u003e), contrast effects (Schwarz and Bless \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e1992\u003c/span\u003e), and social desirability bias. The demonstration that observed navigation ease significantly predicted task completion time (rho = -0.287, p = .021) while survey-reported ease did not (rho = -0.172, p = .173) provides direct empirical support for the non-redundancy of observational and self-report data in this domain.\u003c/p\u003e \u003cp\u003eThe second contribution is a multi-construct characterisation of VRSA user experience across seven dimensions with full inferential statistical analysis. Adoption intention, immersion, visual quality, discomfort, and system performance registered strongly; navigational ease and task efficiency emerged as areas of significant challenge. This characterisation is bounded to a young adult university sample and a single commercial application and should be treated accordingly.\u003c/p\u003e \u003cp\u003eThe third contribution is a five-theme qualitative preference taxonomy demonstrating that optimal technology modality is product-category-dependent rather than universal. AR was preferred for large spatially contextualised products, VR for wearable and visually rich categories, a finding that directly informs practitioner deployment strategy.\u003c/p\u003e \u003cp\u003eFuture work should prioritise longitudinal designs to separate novelty-driven from utility-driven adoption intention, and should recruit demographically broader samples spanning older adult populations to address the generalisability constraints of the present study. Symmetric full-construct instrumentation of both VRSA and ARSA in a single study would enable the construct-level comparative analysis that the present asymmetric design precludes. Neurophysiological measures including galvanic skin response and electroencephalography would provide objective proxies for immersion and cybersickness, eliminating the retrospective bias demonstrated here. Finally, the product-category-by-modality interaction identified qualitatively warrants direct experimental examination.\u003c/p\u003e"},{"header":"Declarations","content":"\u003ch2\u003eFunding Statement\u003c/h2\u003e\n\u003cp\u003eThis research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.\u003c/p\u003e\n\u003ch2\u003eInformed consent\u0026nbsp;\u003cstrong\u003estatement\u003c/strong\u003e:\u003c/h2\u003e\n\u003cp\u003eInformed consent\u0026nbsp;was obtained from all individual participants included in the study. All participants provided written informed consent prior to taking part in the experimental session. The study was approved by the Research Ethics Committee of the Department of Computer Science, Middlesex University London, and was conducted in accordance with the committee's ethical standards and guidelines.\u003c/p\u003e\n\u003ch2\u003eFunding Statement\u003c/h2\u003e\n\u003cp\u003eThis research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.\u003c/p\u003e\n\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\n\u003cp\u003eS.A. conceived and designed the study, developed the observational rubric and survey instruments, collected and analysed the data, performed all statistical analyses, conducted the thematic analysis, and wrote the main manuscript text including all sections, tables, and figures. G.D. contributed to the theoretical framework, provided critical supervision throughout the study design and analytical phases, and reviewed and revised the manuscript. Both authors approved the final submitted version.\u003c/p\u003e\n\u003ch2\u003eACKNOWLEDGEMENTS\u003c/h2\u003e\n\u003cp\u003eThe authors gratefully acknowledge all study participants for their time and engagement, and the Department of Computer Science, Middlesex University London, for supporting the experimental infrastructure used in this research.\u003c/p\u003e\n\u003ch2\u003eData Availability\u003c/h2\u003e\n\u003cp\u003eThe data that support the findings of this study consist of structured observational ratings, and post-experience survey responses collected from human participants under institutional ethical approval. Full open sharing is not possible as participants consented to data use within the bounds of the approved study protocol, and individual-level responses could potentially identify participants given the controlled laboratory setting and small group sizes. The data are available from the corresponding author upon reasonable request, subject to confirmation that the request is consistent with the original ethical approval. The Python scripts used for statistical analysis, including the non-parametric procedures, correlation analyses, and Bonferroni correction routines, are also available from the corresponding author upon reasonable request.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eAzuma RT (1997) A survey of augmented reality. Presence Teleoper Virtual Environ 6(4):355\u0026ndash;385. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1162/pres.1997.6.4.355\u003c/span\u003e\u003cspan address=\"10.1162/pres.1997.6.4.355\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBraun V, Clarke V (2006) Using thematic analysis in psychology. Qual Res Psychol 3(2):77\u0026ndash;101. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1191/1478088706qp063oa\u003c/span\u003e\u003cspan address=\"10.1191/1478088706qp063oa\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCipresso P, Giglioli IAC, Raya MA, Riva G (2018) The past, present, and future of virtual and augmented reality research: a network and cluster analysis of the literature. Front Psychol 9:2086. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3389/fpsyg.2018.02086\u003c/span\u003e\u003cspan address=\"10.3389/fpsyg.2018.02086\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDavis FD (1989) Perceived usefulness, perceived ease of use, and user acceptance of information technology. MIS Q 13(3):319\u0026ndash;340. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.2307/249008\u003c/span\u003e\u003cspan address=\"10.2307/249008\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFlavian C, Ibanez-Sanchez S, Orus C (2019) Augmented reality (AR): an overview and future research agenda. Psychol Mark 36(11):1043\u0026ndash;1057. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/mar.21227\u003c/span\u003e\u003cspan address=\"10.1002/mar.21227\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHan SL, Kim J, An M (2023) The role of VR shopping in digitalization of SCM for sustainable management: application of SOR model and experience economy. Sustainability 15(2):1277. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3390/su15021277\u003c/span\u003e\u003cspan address=\"10.3390/su15021277\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHilken T, Chylinski M, Keeling DI, Heller J, de Ruyter K, Mahr D (2022) How to strategically choose or combine augmented and virtual reality for improved online experiential retailing. Psychol Mark 39(3):495\u0026ndash;507. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/mar.21600\u003c/span\u003e\u003cspan address=\"10.1002/mar.21600\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHoffmann A, Joerss T, Kolbe C (2022) Consumer behavior in augmented shopping reality: a review, synthesis, and research agenda. Front Virtual Real 3:961236. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3389/frvir.2022.961236\u003c/span\u003e\u003cspan address=\"10.3389/frvir.2022.961236\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJavornik A (2016) It's an illusion, but it looks real! Consumer affective, cognitive and behavioural responses to augmented reality applications. J Mark Manag 32(9\u0026ndash;10):987\u0026ndash;1011. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1080/0267257X.2016.1174726\u003c/span\u003e\u003cspan address=\"10.1080/0267257X.2016.1174726\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKahneman D, Fredrickson BL, Schreiber CA, Redelmeier DA (1993) When more pain is preferred to less: adding a better end. Psychol Sci 4(6):401\u0026ndash;405. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/j.1467-9280.1993.tb00589.x\u003c/span\u003e\u003cspan address=\"10.1111/j.1467-9280.1993.tb00589.x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKennedy RS, Lane NE, Berbaum KS, Lilienthal MG (1993) Simulator sickness questionnaire: an enhanced method for quantifying simulator sickness. Int J Aviat Psychol 3(3):203\u0026ndash;220. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1207/s15327108ijap0303_3\u003c/span\u003e\u003cspan address=\"10.1207/s15327108ijap0303_3\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKim JH, Kim M, Park M, Yoo J (2023) Immersive interactive technologies and virtual shopping experiences: differences in consumer perceptions between augmented reality (AR) and virtual reality (VR). Telematics Inf 77:101936. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.tele.2022.101936\u003c/span\u003e\u003cspan address=\"10.1016/j.tele.2022.101936\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLandis JR, Koch GG (1977) The measurement of observer agreement for categorical data. Biometrics 33(1):159\u0026ndash;174. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.2307/2529310\u003c/span\u003e\u003cspan address=\"10.2307/2529310\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLee H, Xu Y, Porterfield A (2022) Antecedents and moderators of consumer adoption toward AR-enhanced virtual try-on technology: a stimulus-organism-response approach. Int J Consum Stud 46(4):1319\u0026ndash;1338. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/ijcs.12753\u003c/span\u003e\u003cspan address=\"10.1111/ijcs.12753\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNielsen J (1993) Usability engineering. Academic, Boston\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNunnally JC (1978) Psychometric theory, 2nd edn. McGraw-Hill, New York\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOyman M, Bal D, Ozer S (2022) Extending the technology acceptance model to explain how perceived augmented reality affects consumers' perceptions. Comput Hum Behav 128:107127. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.chb.2021.107127\u003c/span\u003e\u003cspan address=\"10.1016/j.chb.2021.107127\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRebenitsch L, Owen C (2016) Review on cybersickness in applications and visual displays. Virtual Real 20(2):101\u0026ndash;125. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s10055-016-0285-9\u003c/span\u003e\u003cspan address=\"10.1007/s10055-016-0285-9\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRicci M, Evangelista A, Di Roma A, Fiorentino M (2023) Immersive and desktop virtual reality in virtual fashion stores: a comparison between shopping experiences. Virtual Real 27:2281\u0026ndash;2296. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s10055-023-00806-y\u003c/span\u003e\u003cspan address=\"10.1007/s10055-023-00806-y\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSagnier C, Loup-Escande E, Lourdeaux D, Thouvenin I, Vallery G (2020) User acceptance of virtual reality: an extended technology acceptance model. Int J Hum Comput Interact 36(11):993\u0026ndash;1007. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1080/10447318.2019.1708612\u003c/span\u003e\u003cspan address=\"10.1080/10447318.2019.1708612\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchwarz N, Bless H (1992) Constructing reality and its alternatives: an inclusion/exclusion model of assimilation and contrast effects in social judgment. In: Martin LL, Tesser A (eds) The construction of social judgments. Lawrence Erlbaum, Hillsdale, pp 217\u0026ndash;245\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSengupta A, Cao L (2022) Augmented reality's perceived immersion effect on the customer shopping process: decision-making quality and privacy concerns. Int J Retail Distrib Manag 50(8/9):1039\u0026ndash;1061. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1108/IJRDM-10-2021-0522\u003c/span\u003e\u003cspan address=\"10.1108/IJRDM-10-2021-0522\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSlater M, Sanchez-Vives MV (2016) Enhancing our lives with immersive virtual reality. Front Robot AI 3:74. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3389/frobt.2016.00074\u003c/span\u003e\u003cspan address=\"10.3389/frobt.2016.00074\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eStanney KM, Hale KS, Nahmens I, Kennedy RS (2003) What to expect from immersive virtual environment exposure: influences of gender, body mass index, and past experience. Hum Factors 45(3):504\u0026ndash;520. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1518/hfes.45.3.504.27254\u003c/span\u003e\u003cspan address=\"10.1518/hfes.45.3.504.27254\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eStecuła K, Wolniak R, Aydin B (2024) Technology development in online grocery shopping from shopping services to virtual reality, metaverse, and smart devices: a review. Foods 13(23):3959. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3390/foods13233959\u003c/span\u003e\u003cspan address=\"10.3390/foods13233959\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWeech S, Kenny S, Barnett-Cowan M (2019) Presence and cybersickness in virtual reality are negatively related: a review. Front Psychol 10:158. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3389/fpsyg.2019.00158\u003c/span\u003e\u003cspan address=\"10.3389/fpsyg.2019.00158\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eXi N, Chen J, Gama F, Korkeila H, Hamari J (2024) Acceptance of the metaverse: a laboratory experiment on augmented and virtual reality shopping. Internet Res 34(7):82\u0026ndash;117. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1108/INTR-05-2022-0334\u003c/span\u003e\u003cspan address=\"10.1108/INTR-05-2022-0334\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"augmented reality, virtual reality, technology acceptance model, e-commerce, user experience, immersive technology, retail, HCI, adoption intention, cybersickness","lastPublishedDoi":"10.21203/rs.3.rs-9418492/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9418492/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eVirtual Reality (VR) and Augmented Reality (AR) represent two of the most commercially promising immersive modalities for digital retail, yet rigorous empirical evaluation of VR-based shopping applications (VRSA) alongside comparative user preference data for AR-based applications (ARSA) remains limited. This paper presents a mixed-methods study in which 64 participants completed a structured interaction session with a commercial VR shopping mall application, evaluated via a purpose-built eight-item behavioural observation rubric and a seven-construct post-experience survey. Participants subsequently completed a comparative preference instrument contrasting VRSA against a mobile ARSA they also experienced. Statistical analysis employed paired Wilcoxon signed-rank tests with Bonferroni correction, Mann-Whitney U moderation tests, Spearman rank-order correlations, and Cronbach alpha reliability assessment. Across seven evaluated constructs, VRSA exhibited high adoption intention (observational M\u0026thinsp;=\u0026thinsp;4.47, SD\u0026thinsp;=\u0026thinsp;0.64; survey M\u0026thinsp;=\u0026thinsp;4.08, SD\u0026thinsp;=\u0026thinsp;0.99), strong immersion (observational M\u0026thinsp;=\u0026thinsp;4.31, SD\u0026thinsp;=\u0026thinsp;0.73; survey M\u0026thinsp;=\u0026thinsp;3.64, SD\u0026thinsp;=\u0026thinsp;0.91), and high visual quality (M\u0026thinsp;=\u0026thinsp;4.19/4.11). Statistically significant observational-to-survey discrepancies were identified for immersion (delta\u0026thinsp;=\u0026thinsp;+\u0026thinsp;0.67, W\u0026thinsp;=\u0026thinsp;115.0, p \u0026lt; .001, d\u0026thinsp;=\u0026thinsp;0.603) and navigational ease (delta = -0.98, W\u0026thinsp;=\u0026thinsp;132.5, p \u0026lt; .001, d = -0.748), revealing systematic perceptual biases in retrospective self-report grounded in the peak-end rule and contrast effects. The VRSA-over-ARSA preference mean of 3.95 (SD\u0026thinsp;=\u0026thinsp;1.20) significantly exceeded the neutral midpoint (W\u0026thinsp;=\u0026thinsp;1128.0, p \u0026lt; .001). Thematic analysis yielded five preference themes. Findings are critically contextualised against comparable studies and yield actionable design recommendations for immersive retail technology.\u003c/p\u003e","manuscriptTitle":"Observational and Self-Report Divergences in Virtual Reality Shopping Evaluation: A Multi-Construct Mixed-Methods Study with Comparative AR Preference Analysis","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-05-04 06:02:34","doi":"10.21203/rs.3.rs-9418492/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"878d1c87-818a-4bfa-8ee0-03f4cdadfc21","owner":[],"postedDate":"May 4th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2026-05-05T09:01:41+00:00","versionOfRecord":[],"versionCreatedAt":"2026-05-04 06:02:34","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9418492","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9418492","identity":"rs-9418492","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}