Dual-Pathway Effects of Product and Technological Attributes on Consumer Engagement in Augmented Reality Advertising

Abstract

As augmented reality (AR) advertising becomes increasingly prevalent across digital platforms, understanding how its unique features influence consumer responses is critical for both theory and practice. Based on the elaboration likelihood model (ELM), this study develops and validates a dual-dimension content–dual-route processing model to investigate how different features of AR advertising influence consumer engagement. Specifically, it examines how product-related attributes (attractiveness, informativeness) and technology-related attributes (interactivity, augmentation) shape attitudes toward the ad and purchase intentions through cognitive (information credibility) and affective (enjoyment) pathways. Using data from an online survey (N = 299), the study applies partial least squares structural equation modeling (PLS-SEM) to test the proposed model. The results show that informativeness and augmentation significantly enhance information credibility, while attractiveness primarily influences emotional responses. Interactivity and augmentation positively influence cognitive and affective responses. Mediation analysis confirms the simultaneous activation of central and peripheral processing routes, with flow experience emerging as a significant moderator in selected pathways. By introducing a structured framework for AR advertising content, this study extends the applicability of the ELM in immersive media contexts. It underscores the combined impact of rational evaluation and emotional engagement in shaping consumer behavior and offers practical insights for designing effective AR advertising strategies.

1. Introduction

As augmented reality (AR) technology rapidly advances, more brands integrate AR into their marketing communications to deliver immersive media experiences and enhance consumers’ brand perceptions and interaction quality. In today’s multi-channel media environment, marketers increasingly view AR as a powerful tool for enhancing consumer experience value due to its ability to merge virtual and physical worlds through real-time interaction and layered information [1]. According to ARtillery Intelligence [2,3], the global mobile AR market will grow from USD 10.5 billion in 2023 to USD 21.5 billion by 2028, while the head-mounted AR segment will expand from USD 1.86 billion to USD 5.34 billion during the same period, highlighting AR’s rising commercial and communicative potential.

Compared to traditional image and video advertising, AR advertising delivers a stronger visual impact and gives users greater perceived control and interactive engagement [4]. For instance, Burger King’s “Burn That Ad” campaign leveraged AR technology to encourage consumers to scan competitors’ advertisements, triggering a virtual “burning” effect. This campaign boosted brand recall and increased user participation [5]. Such cases demonstrate the unique advantages of AR advertising in capturing attention, stimulating interest, and influencing consumer behavior [6]. However, despite these benefits and AR’s entertainment appeal [7], most marketers have yet to adopt AR advertising widely. One reason lies in the early-stage development of AR technology [8], which still faces limitations in device and system accessibility. Another stem is that firms have yet to fully understand how consumers derive value from AR advertising experiences. As a result, many firms remain cautious in their implementation. This slow adoption highlights a pressing need to explore how specific AR content features influence consumer engagement and perceived effectiveness.

Prior literature has predominantly focused on technological attributes such as augmentation—the overlay of virtual elements onto physical environments—and interactivity, which refers to the user’s ability to control or manipulate AR content in real time [8,9,10]. However, relatively few studies have examined how product-level informational cues and technology-driven interactive features work together to influence consumers’ information processing. These two content dimensions—product-related and technology-related—may activate distinct psychological mechanisms, yet have rarely been integrated within a unified conceptual framework. Moreover, although many scholars recognize consumer engagement as a key mechanism linking advertising features to brand relationships and behavioral outcomes [11,12], most treat engagement as a static outcome rather than a dynamic psychological process. Few studies investigate the mediating roles of cognitive and emotional engagement during ad processing, and researchers have not systematically examined how cognitive, emotional, and behavioral dimensions interact within AR advertising frameworks.

This study addresses these theoretical gaps by adopting the elaboration likelihood model (ELM) as its conceptual foundation. We propose a dual-route framework in which product-related content features—such as attractiveness (the visual and stylistic appeal of the advertisement) and informativeness (the extent to which it provides useful product information)—are more likely to activate the central route of information processing. In contrast, technology-related features (e.g., augmentation and interactivity) are expected to stimulate peripheral-route responses via sensory and emotional cues. To further refine this framework, we introduce flow as a moderating variable, examining how immersive psychological states may amplify or attenuate the influence of different AR ad content features on consumer engagement. Defined by high concentration and deep involvement, flow has demonstrated positive effects across various digital media contexts. For instance, Pelet et al. [13] found that flow enhances users’ attitudes toward mobile ads and sustains attention. Similarly, Chen and Lin [14] showed that flow improves enjoyment and strengthens purchase intentions in AR-based retail experiences. These findings suggest that flow serves as an experiential response and functions as a critical psychological mechanism that moderates how advertising content influences consumer engagement.

Integrating advertising content, information processing routes, and consumer engagement, this study proposes a unified structural model to investigate how AR advertising triggers multidimensional engagement, including cognitive, emotional, and behavioral dimensions. Specifically, we address the following research questions:

RQ1: How do content features of augmented reality advertising—specifically product-related and technology-related elements—influence the cognitive, emotional, and behavioral dimensions of consumer engagement through distinct processing routes?

RQ2: Does flow play a moderating role in the information processing of augmented reality advertising?

Overall, this study proposes a theoretically grounded and practically relevant framework for understanding how different dimensions of AR advertising content shape consumer engagement through dual processing routes.

The remainder of this paper is structured as follows: The subsequent section outlines the theoretical foundations and presents a review of relevant literature. This is followed by the development of research hypotheses grounded in the elaboration likelihood model. The next section describes the methodology, including data collection and analysis procedures. The empirical results are then presented and interpreted. Subsequently, we discuss the theoretical and practical implications of the findings. Lastly, we conclude the study by addressing its limitations and suggesting directions for future research.

2. Literature Review

As augmented reality (AR) technologies become increasingly accessible, marketers have increasingly adopted AR advertising as a novel and strategically significant medium in brand communication. AR advertising distinguishes itself by embedding digital content such as product models, informational overlays, and brand symbols directly into the consumer’s physical environment in real time, creating a seamless blend of virtual and real experiences [15]. Unlike traditional media that deliver information passively, this immersive and interactive format captures consumer attention, encourages engagement, and reshapes how individuals process brand information and form emotional responses [9,16]. Researchers have found that AR advertising adds value by delivering media richness and enhancing the sense of presence, which strengthens perceived realism and contextual relevance of brand messages [16], ultimately fostering more favorable attitudes and behavioral outcomes [1]. As a result, scholars increasingly focus on understanding how specific AR content features influence consumers’ psychological and behavioral responses.

2.1. Theoretical Framework

This section outlines the theoretical foundations underpinning the present study, specifically the elaboration likelihood model (ELM) and flow theory. Together, these frameworks provide a comprehensive explanation for the dual-pathway information processing and the psychological mechanisms that drive consumer engagement in augmented reality (AR) advertising.

2.1.1. Elaboration Likelihood Model

Petty and Cacioppo [17] developed the elaboration likelihood model (ELM), as a dual-process theory of information processing, which researchers have widely applied to studies on attitude change, persuasion mechanisms, and consumer decision-making. The model proposes that persuasive occurs through two distinct processing routes: central and peripheral [18,19]. An individual’s choice of route depends on their elaboration likelihood, i.e., their willingness and ability to engage in effortful, analytical thinking in a given context [20]. When elaboration likelihood is high, individuals process information through the central route, carefully evaluating the message’s logic, credibility, and relevance to form stable attitudes and well-grounded judgments [14]. When elaboration likelihood is low, individuals rely on the peripheral route, forming quick and shallow attitudes based on cues such as emotional appeal, source attractiveness, or interactive features [21]. Researchers have applied the ELM in traditional advertising and emerging media environments, such as social media and augmented reality, to explain how consumers respond to diverse informational cues in immersive and interactive contexts [22].

This study applies the ELM to explain how product-related and technology-related content in AR advertising trigger different dimensions of consumer engagement through distinct processing routes. Product-related cues, such as product descriptions, functional features, and technical specifications, stimulate the central route by prompting consumers to evaluate the message’s credibility and engage in cognitive processing [23]. In contrast, technological experience cues, including augmentation, interactivity, novelty, and entertainment value, activate the peripheral route by eliciting emotional responses and hedonic experiences that require minimal cognitive effort, thereby increasing emotional engagement and influencing behavioral intentions.

Thus, the ELM offers the theoretical foundation for the proposed dual-route model and a psychological explanation for how different content features drive distinct forms of engagement. Importantly, this study introduces flow experience as a moderating variable. Given that a consumer’s level of immersion may influence their tendency to elaborate, the model incorporates flow to examine whether the persuasive impact of central and peripheral cues varies depending on the intensity of the immersive experience.

2.1.2. Flow Theory

Flow theory suggests that individuals enter a highly immersive, enjoyable, and self-transcendent psychological state, referred to as flow, when they fully engage in an activity [24]. Researchers have widely applied this concept, defined by deep concentration, intrinsic enjoyment, and total absorption, to research on digital media and interactive advertising [25]. In AR advertising, technological content features such as interactivity and augmentation can enhance users’ sense of control, realism, and responsiveness, thereby facilitating flow experiences [10].

Flow functions as an outcome of media use and as a critical antecedent that shapes how consumers process advertising content. When individuals experience high levels of flow, they tend to maintain sustained attention and engage in deeper message elaboration, which strengthens their evaluations of the ad’s credibility. At the same time, flow amplifies affective responses such as enjoyment, novelty, and immersion, thereby increasing the persuasive power of peripheral cues [14].

Therefore, flow is more than a subjective manifestation in AR advertising; it serves as a key moderating variable that influences the strength of both processing routes. Under high-flow conditions, content features are more likely to enhance information credibility (central route) and hedonic enjoyment (peripheral route), whereas low-flow conditions may weaken these effects. Based on this reasoning, the present study incorporates flow as a boundary condition to examine its moderating role in the AR ad information processing model.

2.2. AR Ad Content Characteristics

Unlike traditional advertisements that rely on static text or images to convey information, AR advertising integrates product content with media technology to deliver immersive and highly interactive user experiences. Prior research (see Table 1) identifies two core content dimensions in AR advertisements: (1) product-related informational content, which includes rational cues such as product functions, usage, and technical specifications, and (2) technology-enabled experiential content, which encompasses interactive formats, visual augmentation, and immersive effects [6]. To explore how AR advertising influences consumer engagement, this study categorizes AR ad content into two dimensions—product content and technological functionality—and examines how each shapes consumers’ processing pathways and engagement responses.

Within the product content dimension, this study investigates two key variables: informativeness and attractiveness. Informativeness refers to the extent and completeness of information about a product’s functions, uses, and attributes, reflecting the advertisement’s capacity to satisfy consumers’ informational needs [26]. In AR advertising, highly informative product descriptions help consumers better understand features, functionalities, and usage contexts, thereby facilitating cognitive engagement. In contrast, attractiveness focuses on the ad’s aesthetic and sensory appeal—its visual design, language style, and overall presentation—which draws consumer attention and stimulates initial interest [27], potentially influencing attitudinal responses.

Table 1. AR characteristics.

Ref.	Studies	Augmented Reality Characteristics
[1]	Hilken et al. (2017)	Augmentation
[4]	Uribe et al. (2022)	Informativeness, Entertaining
[7]	Feng & Xie (2018)	Informativeness, Novelty, Entertainment, Complexity
[8]	Barhorst et al. (2021)	Interactivity, Vividness, and Novelty
[9]	Poushneh & Vasquez-Parraga (2017)	Interactivity
[10]	Kumar & Srivastava (2022)	Interactivity, Augmentation
[28]	Yim et al. (2017)	Interactivity, Vividness
[29]	Javornik (2016)	Interactivity, Media Richness
[30]	Chen et al. (2022)	Vividness, Spatial Accuracy
[31]	Rauschnabel et al. (2019)	Augmentation Quality
[32]	Sung et al. (2022)	AR App Control, Design
[33]	Saleem et al. (2022)	Informativeness, Entertainment, Irritation
[34]	Cowan et al. (2024)	Immersive
[35]	Pozharliev et al. (2022)	processing fluency
[36]	Ahn et al. (2023)	Interactivity

Table 1. AR characteristics.

Ref.	Studies	Augmented Reality Characteristics
[1]	Hilken et al. (2017)	Augmentation
[4]	Uribe et al. (2022)	Informativeness, Entertaining
[7]	Feng & Xie (2018)	Informativeness, Novelty, Entertainment, Complexity
[8]	Barhorst et al. (2021)	Interactivity, Vividness, and Novelty
[9]	Poushneh & Vasquez-Parraga (2017)	Interactivity
[10]	Kumar & Srivastava (2022)	Interactivity, Augmentation
[28]	Yim et al. (2017)	Interactivity, Vividness
[29]	Javornik (2016)	Interactivity, Media Richness
[30]	Chen et al. (2022)	Vividness, Spatial Accuracy
[31]	Rauschnabel et al. (2019)	Augmentation Quality
[32]	Sung et al. (2022)	AR App Control, Design
[33]	Saleem et al. (2022)	Informativeness, Entertainment, Irritation
[34]	Cowan et al. (2024)	Immersive
[35]	Pozharliev et al. (2022)	processing fluency
[36]	Ahn et al. (2023)	Interactivity

The technological functionality dimension emphasizes AR’s interactive and immersive qualities. This study focuses on two characteristics: interactivity and augmentation. Interactivity reflects the extent to which users actively engage with the ad by clicking, rotating, or virtually trying the product, thus highlighting their active role in the communication process [28]. Augmentation refers to the immersive effect of overlaying virtual elements onto the physical environment. It enhances spatial presence and environmental integration [29], increasing perceived realism and immersion.

Together, the four content characteristics—informativeness and attractiveness (product dimension) and interactivity and augmentation (technological functionality)—capture the dual emphasis of AR advertising on rational information and sensory experience. These features vary in how they deliver content and activate psychological processing mechanisms. Product-related features primarily elicit cognitive engagement, drawing attention and facilitating message comprehension, which aligns with the central route of persuasion in the ELM [23]. In contrast, technology-related features rely on sensory stimulation and emotional arousal, triggering peripheral-route processing driven by affective responses and heuristic cues.

By examining how these content types activate cognitive, emotional, and behavioral engagement, this study identifies four independent variables that map onto dual-route persuasion mechanisms. In doing so, it advances understanding of how differentiated AR content shapes consumer responses through distinct information processing pathways.

2.3. Consumer Engagement

Digital marketing and advertising researchers recognize consumer engagement as a core concept. Scholars commonly define it as the psychological investment and behavioral inclination consumers demonstrate during interactions with brand-related content [12]. Unlike traditional unidimensional models that emphasize an attitude–behavior sequence, consumer engagement reflects a more multifaceted and dynamic process, reflecting deeper individual responses within brand communication. As Hollebeek and Macky [12] further argue that in digital content environments, engagement encompasses cognitive, emotional, and behavioral investment, particularly through interaction and co-creation activities with the brand.

Researchers typically conceptualize consumer engagement as a multidimensional construct that includes cognitive, emotional, and behavioral dimensions. Hollebeek et al. [37] define these components as follows: cognitive engagement involves focused attention and mental processing of brand-related information; emotional engagement reflects consumers’ affective responses and interest; and behavioral engagement refers to consumers’ participatory actions. Scholars have widely applied this tripartite framework across various domains, including brand communities, social media, and interactive advertising.

In the context of augmented reality advertising, this three-dimensional structure is particularly relevant. Consumers cognitively engage by actively processing and evaluating product-related content, aligning with the central route of elaboration in the ELM. They experience emotional engagement through immersive and hedonic responses to technological features such as interactivity and augmentation, which activate the peripheral route via sensory cues. Behavioral engagement in this study reflects consumers’ attitudes toward the advertisement and their purchase intentions, capturing the ultimate psychological and behavioral outcomes.

In sum, this study defines consumer engagement as a composite cognitive–emotional–behavioral response process activated by AR advertising features and shaped through message elaboration and immersive experience. This definition aligns with the dual-route perspective of the elaboration likelihood model and underpins the study’s proposed path model and moderating mechanisms.

3. Hypothesis Development

This study draws on ELM, the multidimensional framework of consumer engagement, and the moderating role of flow experience to propose a moderated mediation model that explains how content features in AR advertising influence consumer responses through dual processing routes. This model conceptualizes product-related content (e.g., informativeness and attractiveness) as a central-route cue that fosters cognitive elaboration and enhances perceptions of information credibility. In contrast, it identifies technology-related content (e.g., interactivity and augmentation) as a peripheral-route cue that stimulates hedonic enjoyment through affective and sensory responses. These two mediators—information credibility and enjoyment—shape consumers’ overall attitudes toward the advertisement, which then influences their purchase intentions. To account for the immersive nature of AR media, the model also introduces flow experience as a moderator, assessing whether users’ immersive psychological states strengthen or weaken the influence of content features along each processing path.

Figure 1 illustrates the conceptual model, which guides the development of the hypotheses organized into three thematic sets: (1) how AR content features influence the two mediators, (2) how credibility and enjoyment mediate the effects of content features on attitude and behavioral intention, and (3) how flow moderates the relationships between content features and psychological responses.

3.1. Attractiveness

Advertising attractiveness captures the ad’s ability to draw attention, spark interest, and evoke aesthetic enjoyment through its visual design, stylistic expression, or creative interaction [27]. In AR advertising, advertisers often express this attractiveness through visual appeal, vivid content, and immersive qualities enabled by technological presentation [8]. Prior research suggests that highly attractive advertisements increase consumers’ initial interest and emotional arousal during ad exposure. Within the ELM framework, researchers classify attractiveness as a peripheral cue—one that does not rely on rational elaboration but can still significantly influence information perception [8]. In AR environments, high advertising attractiveness may improve content salience and processing fluency, thereby strengthening perceptions of information credibility. Aesthetically pleasing and creatively designed ads also tend to elicit positive emotional responses, such as enjoyment and satisfaction [6]. Based on this reasoning, we propose the following hypotheses:

H1a: 

Advertising attractiveness has a positive effect on information credibility.

H1b: 

Advertising attractiveness has a positive effect on enjoyment.

3.2. Informativeness

We define informativeness as the degree to which an advertisement delivers valuable, accurate, and detailed information about a product or service, such as its functions, usage methods, key benefits, and applicable scenarios [26]. Prior research shows that when ad content delivers relevant and decision-facilitating information, consumers are more likely to perceive it as a credible and useful source [38]. ELM conceputalizes informativeness as a central-route cue, requiring thoughtful cognitive processing to influence attitude formation and judgment [14]. In AR advertising, structured and information-rich content often stimulates analytical thinking and strengthens perceptions of information credibility. In addition, informativeness can enhance hedonic enjoyment by improving cognitive fluency and reducing ambiguity during ad processing, ultimately increasing consumer satisfaction. This study operationalizes informativeness as the perceived amount, clarity, and usefulness of product-related information, reflecting the rational dimension of AR ad content. Based on these considerations, we propose the following hypotheses:

H2a: 

Informativeness has a positive effect on information credibility.

H2b: 

Informativeness has a positive effect on enjoyment.

3.3. Interactivity

Researchers widely recognize interactivity as a key characteristic of digital and intelligent technologies that shapes immersive consumer experiences [39]. It encompasses a technological feature and a user-perceived experience [40]. This study adopts the technological perspective and defines interactivity as the extent to which users can manipulate and explore product representations in AR advertising, for example, by rotating the product, viewing it in 360 degrees, or aligning it with real-world environments [28]. These interactive functions give users a sense of control and participation, thereby reinforcing the immersive nature of the ad experience [8]. They create more enjoyable user experiences and deepen users’ understanding of product features, which strengthens the credibility of the information presented. Within the ELM, interactivity primarily functions as a peripheral cue that evokes emotional and cognitive responses depending on the user’s level of involvement. Accordingly, this study proposes the following hypotheses:

H3a: 

Interactivity has a positive effect on information credibility.

H3b: 

Interactivity has a positive effect on enjoyment.

3.4. Augmentation

Augmentation is one of the core attributes of AR technology, enabling systems to overlay virtual elements onto the physical environment in real time [29]. This function allows users to perceive and interact with digital content within real-world settings, creating a seamless integration between virtual and physical elements. In AR advertising, augmentation features introduce novelty and interactivity and intensify consumers’ emotional responses and perceived experiential value [41]. For example, virtual try-ons, product visualization, or simulated scenarios allow consumers to preview how a product might look or function in realistic contexts [29,42]. These features support functional evaluation and emotional engagement by delivering personalized and contextually relevant experiences that elicit feelings of novelty, enjoyment, and connection. Furthermore, by presenting spatially rich and visually anchored product representations, augmented elements improve contextual relevance and help consumers more intuitively assess product suitability and performance, ultimately enhancing perceived information credibility [43,44]. Based on this logic, we propose the following hypotheses:

H4a: 

Augmentation has a positive effect on information credibility.

H4b: 

Augmentation has a positive effect on enjoyment.

3.5. Information Credibility

Scholars commonly define information credibility as the degree to which consumers perceive the information conveyed in an advertisement as truthful and trustworthy [45,46]. It reflects consumers’ confidence in the accuracy, completeness, and reliability of the ad content [47]. When consumers perceive an advertisement as untrustworthy or unreliable, they are less likely to believe the advertiser and less inclined to purchase the promoted product [48,49]. Thus, perceived credibility is a critical determinant of advertising effectiveness [50]. When consumers find the information credible, they are more likely to form favorable attitudes toward the advertisement [51], which positively influences their purchase decisions and behavioral intentions [52]. In AR advertising, information credibility becomes even more crucial, as blending virtual content with real-world settings may raise concerns about message authenticity. Delivering clear, relevant, and trustworthy information can alleviate such doubts, increase consumers’ acceptance of the ad, and strengthen their emotional responses. Based on this rationale, we propose the following hypothesis:

H5: 

Information credibility has a positive effect on advertising attitude.

3.6. Enjoyment

Enjoyment constitutes a key perceptual outcome of advertising content and directly shapes consumers’ attitudes toward advertisements [53]. Consumers derive enjoyment when they feel pleasure, entertainment, or aesthetic satisfaction while viewing an ad [54]. This emotional response enhances the positivity of the advertising experience and elicits favorable emotional states such as delight and amusement [55]. These feelings motivate consumers to pay attention to advertising messages and accept them more readily [49]. According to ELM, individuals who lack strong cognitive motivation rely more heavily on peripheral cues, including stylistic elements, background music, or color design [56]. These sensory elements evoke positive emotional responses, which increase message acceptance and improve ad evaluations [7]. Thus, when advertisers design ads that generate enjoyment, they increase the likelihood that consumers will develop positive attitudes toward the advertisement. Based on this logic, we propose the following hypothesis:

H6: 

Enjoyment has a positive effect on advertising attitude.

3.7. Attitude to Ad

Consumers form an advertising attitude as an overall evaluative disposition, favorable or unfavorable, in response to a specific advertising stimulus [57]. This attitude reflects how they assess the ad during message processing, incorporating cognitive elements (e.g., perceived credibility, relevance of the message) and affective elements (e.g., aesthetics, format, emotional appeal) [58]. Consumers’ evaluations of advertising value often shape attitude to the ad, typically determined by perceived informativeness and entertainment [50]. Advertising research widely regards advertising attitude as a core mediating variable that links information processing to consumer behavioral outcomes [59]. Across traditional and digital contexts, studies consistently show that advertising attitude predicts consumers’ acceptance of brands and products, as well as their purchase intentions [60,61].

In AR advertising, consumers interpret informativeness more specifically as information credibility, highlighting its central role in cognitive processing. Based on this, advertising attitude results from central and peripheral-route processing, reflecting the joint influence of credibility judgments and emotional responses such as enjoyment. Prior studies suggest that rational evaluations and sensory impressions shape consumers’ overall attitudes toward an ad, which directly influence their behavioral intentions [62]. In other words, a positive advertising attitude formed during AR ad exposure increases the likelihood of purchase behavior. Accordingly, we propose the following hypothesis:

H7: 

Advertising attitude has a positive effect on purchase intention.

3.8. Flow

Previous research suggests that AR, with its unique features of interactivity, augmentation, and novelty, enables users to enter a flow state, which is a heightened sense of focus, immersion, and enjoyment during task engagement [8]. When users experience flow, they tend to elaborate more cognitively, enjoy the interaction more, perceive lower product risk, and show a greater tendency toward impulse buying [10]. Researchers have also demonstrated that the immersive realism and mobility of AR devices effectively induce flow, which triggers cognitive and emotional reactions toward the brand and advertising interface, ultimately strengthening purchase intention [30]. Building on this evidence, we argue that consumers who experience flow while interacting with AR ads process content features more deeply and respond more positively to the ad. Specifically, flow heightens their sensitivity to product- and technology-related features, thereby amplifying the effects of those features on information credibility and enjoyment. In other words, when consumers are in a state of flow, AR content becomes more persuasive. Accordingly, we propose the following hypothesis:

H8a: 

Flow moderates the relationship between attractiveness and information credibility.

H8b: 

Flow moderates the relationship between informativeness and information credibility.

H8c: 

Flow moderates the relationship between attractiveness and enjoyment.

H8d: 

Flow moderates the relationship between informativeness and enjoyment.

H8e: 

Flow moderates the relationship between interactivity and information credibility.

H8f: 

Flow moderates the relationship between augmentation and information credibility.

H8g: 

Flow moderates the relationship between interactivity and enjoyment.

H8h: 

Flow moderates the relationship between augmentation and enjoyment.

4. Methodology

This section outlines the research methodology, including data collection procedures, measurement instruments, and analytical techniques employed in the study.

4.1. Data Collection

We collected data for this study using the online survey platform Credamo and obtained 340 responses. After rigorous data screening and removing cases with excessively short response times and highly homogeneous answer patterns, we retained 299 valid responses for analysis, resulting in an effective response rate of 87.94%. This final sample size meets the minimum threshold for structural equation modeling (SEM) as recommended by Hair et al. [63] ensuring sufficient statistical power for model estimation and hypothesis testing. To further ensure the adequacy of our sample size, we applied the inverse square root method [64], a recent and practical method for estimating sample adequacy in PLS-SEM. Assuming a small effect size of β = 0.10—aligned with Cohen’s [65] benchmark for small effects—and adopting a constant of c = 0.7, which is appropriate for models of moderate complexity, the estimated minimum required sample size is approximately 188. Our final sample of 299 thus exceeds this threshold. Furthermore, with this sample size, the minimum detectable path coefficient is approximately β = 0.08, indicating that the model has sufficient statistical power to detect small-to-moderate effects commonly observed in PLS-SEM research.

All participants reviewed an informed consent statement before beginning the study and voluntarily agreed to participate, with the option to withdraw at any time. No personally identifiable information was collected, and all responses were treated with strict confidentiality. Anonymity and data privacy were fully ensured, and all data were used exclusively for academic research purposes.

4.2. Measures Instrument

This study adapted measurement items from established and validated scales in prior literature. To ensure contextual relevance, we made minor wording adjustments to fit the augmented reality (AR) advertising context and align with the linguistic environment of the target population, while preserving the original constructs and measurement dimensions. A bilingual expert translated all original English items into Chinese, and then a second independent expert backtranslated them into English. The backtranslated version closely matched the original, demonstrating strong semantic equivalence.

We measured all constructs using previously validated scales, with items rated on a seven-point Likert scale (1 = strongly disagree, 7 = strongly agree), unless otherwise noted. Specifically, we assessed the following factors (we provide information about the scales used in brackets):

• Attractiveness (two-item scale adapted from Baum et al. [66])
• Informativeness (three items based on Ducoffe [26] and Liu et al. [67])
• Interactivity (a three-item scale adapted from Yim et al. [28])
• Augmentation (three items from Kumar and Srivastava [10])
• Information Credibility (three items from Brackett and Carr [68])
• Enjoyment (a three-item scale from Yim et al. [28])
• Flow (a two-item semantic differential scale developed by Yim et al. [28])
• Attitude to ad (two items adapted from Holbrook and Batra [69] and Uribe et al. [4])
• Purchase Intention (a two-item scale based on Spears and Singh [70])

4.3. Data Analysis

The data were analyzed using a combination of statistical software packages. Initially, SPSS 27 was employed for data entry, screening, and cleaning, as well as for conducting descriptive statistics and assessing common method bias. Subsequently, SmartPLS 4.0 was used to perform advanced statistical analyses based on partial least squares structural equation modeling (PLS-SEM). This study utilized structural equation modeling (SEM) to test the proposed hypotheses and evaluate the relationships between independent and dependent variables, including both mediating and moderating effects. Unlike simple linear regression, SEM allows for the simultaneous testing of multiple mediation pathways and is particularly suitable for evaluating complex models. Given that PLS algorithms do not produce traditional parametric significance tests for individual parameters, we applied the non-parametric bootstrapping method to estimate standard errors and t-values. Specifically, 10,000 bootstrap resamples were used to assess the significance of both direct and indirect effects within the model.

5. Results

The empirical findings are presented in the following subsections, including participant demographics, measurement model, and structural model assessment.

5.1. Participants’ Characteristics

Among the 299 respondents, 67.6% were female (n = 202) and 32.4% were male (n = 97). Most participants (48.8%, n = 146 were between 26 and 35 years old, followed by those aged 18–25 (27.1%, n = 81), 36–50 (22.1%, n = 66), and 51 and older (2.0%, n = 6). Regarding education, most held a bachelor’s degree (71.2%, n = 213), while others reported having a master’s degree (19.4%, n = 58), an associate degree or lower (8.7%, n = 26), or a doctoral degree or higher (0.7%, n = 2).

Table 2. Sample profile.

Variable	Category	Absolute Number	Percentage
Gender	Male	97	32.4
	Female	202	67.6
Age	18–25 years	81	27.1
	26–35 years	146	48.8
	36–50 years	66	22.1
	51+	6	2.0
Education	Junior college or less	26	8.7
	Undergraduate	213	71.2
	Postgraduate	58	19.4
	Doctorate or above	2	0.7

summarizes the demographic characteristics of the sample.

5.2. Common Method Bias

To assess potential common method bias (CMB), this study employed Harman’s single-factor test using SPSS 27. We conducted an exploratory factor analysis (EFA) on all measurement items using unrotated principal component analysis. The results indicated that the first factor accounted for 28.9% of the total variance, which is well below the commonly accepted threshold of 40%. Therefore, common method bias is not a serious concern in this study.

5.3. Measurement Model

To evaluate the reliability and validity of the reflective measurement model, this study followed the guidelines recommended by Hair et al. [63], assessing standardized factor loadings, composite reliability (CR), average variance extracted (AVE), and Cronbach’s alpha. The measurement model was analyzed using SmartPLS 4.0, which provided the relevant reliability and validity statistics. As shown in

Table 3. Measurement properties.

Measure	Item	Loadings	CR	AVE	Cronbach’s α
Attractiveness (ATTR)	ATTR 1	0.873	0.843	0.728	0.628
	ATTR 2	0.833
Informativeness (INF)	INF 1	0.741	0.798	0.569	0.621
	INF 2	0.810
	INF 3	0.707
Interactivity (INT)	INT 1	0.793	0.825	0.611	0.682
	INT 2	0.785
	INT 3	0.766
Augmentation (AUG)	AUG 1	0.808	0.836	0.630	0.705
	AUG 2	0.842
	AUG 3	0.727
Flow (FL)	FL 1	0.828	0.842	0.728	0.628
	FL 2	0.878
Information Credibility (IC)	IC 1	0.800	0.836	0.629	0.706
	IC 2	0.790
	IC 3	0.790
Enjoyment (ENJ)	ENJ 1	0.717	0.774	0.533	0.563
	ENJ 2	0.718
	ENJ 3	0.755
Attitude to ad (ATTI)	ATTI 1	0.836	0.786	0.648	0.459
	ATTI 2	0.772
Purchase Intention (PI)	PI 1	0.898	0.896	0.812	0.769
	PI 2	0.904

Note: CR: Composite Reliability; AVE: Average Variance Extracted.

, all items exhibited factor loadings above the commonly accepted threshold of 0.70, indicating strong indicator reliability.

The composite reliability (CR) values for all constructs ranged from 0.774 to 0.896, exceeding the recommended threshold of 0.70, indicating satisfactory internal consistency. The average variance extracted (AVE) values ranged from 0.533 to 0.812, indicating satisfactory convergent validity. Notably, Cronbach’s alpha, which assumes equal indicator loadings, tends to underestimate reliability compared to CR. In this study, Cronbach’s alpha for some constructs (e.g., ATTR, INF, ATTI) fell below 0.70. However, given that their CR values met the recommended standard, the overall reliability of these constructs remains acceptable [63].

To further assess discriminant validity, this study applied the method proposed by Fornell and Larcker [71], which compares the square root of the AVE for each construct with its correlations with other constructs (see

Table 4. Discriminant validity evaluation.

Construct	ATTR	INF	INT	AUG	IC	ENJ	ATTI	PI	FL
ATTR	0.853
INF	0.251	0.754
INT	0.408	0.197	0.782
AUG	0.465	0.285	0.281	0.794
IC	0.378	0.332	0.429	0.427	0.793
ENJ	0.543	0.375	0.475	0.504	0.420	0.730
ATTI	0.499	0.383	0.390	0.439	0.437	0.512	0.805
PI	0.425	0.298	0.358	0.378	0.512	0.472	0.349	0.901
FL	0.377	0.214	0.314	0.374	0.325	0.503	0.482	0.311	0.853

Note: The diagonal items are the square root of AVE values, and the non-diagonal ones are inter-construct correlations.

). The results show that for all constructs, the square root of the AVE (diagonal elements) exceeded the corresponding inter-construct correlations (off-diagonal elements), indicating that the measurement model demonstrates satisfactory discriminant validity. In addition, we assessed discriminant validity using the Heterotrait–Monotrait ratio (HTMT). As shown in

Table 5. HTMT analysis.

Construct	ATTR	INF	INT	AUG	IC	ENJ	ATTI	PI	FL
ATTR
INF	0.390
INT	0.619	0.314
AUG	0.701	0.420	0.404
IC	0.565	0.501	0.615	0.600
ENJ	0.906	0.621	0.765	0.797	0.655
ATTI	0.914	0.710	0.696	0.778	0.761	1.008
PI	0.612	0.427	0.495	0.513	0.694	0.713	0.578
FL	0.597	0.330	0.477	0.542	0.485	0.845	0.885	0.443

, most HTMT values were below the liberal threshold of 0.90 [72], indicating satisfactory discriminant validity among the majority of constructs. Although some HTMT values slightly exceed the conventional threshold of 0.90, simulation-based bias analysis by Roemer et al. [73] suggests that even under conditions of high construct correlations (φ = 0.90 or φ = 1.0) and a sample size of n = 250, the HTMT method maintains a low level of estimation bias—particularly when the heterogeneity of indicator loadings is minimal. Therefore, the observed HTMT values in this study can still be regarded as statistically acceptable.

5.4. Structural Model

After evaluating the measurement model, we analyzed the structural model using SmartPLS 4.0 to test the proposed research hypotheses. We examined the significance and strength of the hypothesized relationships among constructs. Figure 2 depicts the results of our structural model.

Table 6. Path analysis results.

Path	Coefficient	p-Value	Hypothesis	Results
ATTR→IC	0.071	0.239	H1a	Not Supported
ATTR→ENJ	0.248	0.000	H1b	Supported
INF→IC	0.179	0.001	H2a	Supported
INF→ENJ	0.183	0.000	H2b	Supported
INT→IC	0.273	0.000	H3a	Supported
INT→ENJ	0.181	0.000	H3b	Supported
AUG→IC	0.219	0.000	H4a	Supported
AUG→ENJ	0.170	0.001	H4b	Supported
IC→ATTI	0.270	0.000	H5	Supported
ENJ→ATTI	0.399	0.000	H6	Supported
ATTI→PI	0.349	0.000	H7	Supported

presents the path coefficients, significance levels, and hypothesis testing results. The results show that most hypotheses received empirical support. Specifically, informativeness (INF) and interactivity (INT) significantly and positively influenced information credibility (IC) and enjoyment (ENJ) (p < 0.01), supporting H2a, H2b, H3a, and H3b. Augmentation (AUG) exerted a significant positive effect on IC (β = 0.219, p < 0.001) and ENJ (β = 0.170, p = 0.001), thereby supporting H4a and H4b. Regarding advertising attitude (ATTI), IC (β = 0.270, p < 0.001) and ENJ (β = 0.399, p < 0.001) emerged as significant predictors, supporting H5 and H6. In addition, ATTI significantly and positively affected purchase intention (PI) (β = 0.349, p < 0.001), supporting H7.

Notably, while attractiveness (ATTR) significantly increased enjoyment (ENJ) (β = 0.248, p < 0.001), supporting H1b, it did not significantly influence information credibility (IC) (β = 0.071, p = 0.2329), rejecting H1a. This finding suggests that attractiveness is more likely to elicit emotional experiences rather than enhance the perceived credibility of the information. Overall, the structural model results confirm the majority of the hypothesized relationships and highlight the positive influence of various AR advertising content features on users’ perceptions, attitudes, and behavioral intentions.

This study assessed the explanatory power of the structural model using the adjusted coefficient of determination (adjusted R²) for each endogenous construct. The model accounted for 32.0% of the variance in information credibility (IC), 52.5% in enjoyment (ENJ), and 31.8% in advertising attitude (ATTI). For the final behavioral outcome—purchase intention (PI)—the adjusted R² was 11.9%. These results suggest that while the model demonstrates moderate explanatory power for consumers’ cognitive and emotional responses to AR advertising, its ability to predict behavioral intentions is relatively limited. This underscores the need for future models to incorporate additional predictors or contextual variables to enhance behavioral prediction.

To assess the model’s predictive relevance, we applied the blindfolding procedure and calculated Q² values for all endogenous constructs. All Q² values exceeded zero, indicating acceptable predictive relevance [74]. Specifically, enjoyment (ENJ) yielded a Q² value of 0.268, information credibility (IC) and advertising attitude (ATTI) had Q² values of 0.198 and 0.200, respectively, all indicating moderate predictive relevance. Although purchase intention (PI) showed a lower Q² value of 0.094, it still exceeded the minimum threshold, confirming some degree of predictive relevance. These results suggest that the model provides explanatory strength and predictive utility across the endogenous variables.

To examine potential multicollinearity, we calculated variance inflation factor (VIF) values. All values ranged from 1.097 to 1.638, well below the recommended threshold of 3.0 [75], indicating no serious multicollinearity and confirming the stability and interpretability of the model estimates.

To explore the mediation mechanisms in greater depth, we conducted bias-corrected bootstrapping with 10,000 resamples using SmartPLS 4.0. We assessed whether AR advertising content features influenced advertising attitude and purchase intention through the mediators of information credibility and enjoyment.

Table 7. Indirect and mediating effects.

	Bootstrap 10,000 Times			Percentile 95%
	β	SE	T Statistics	Lower	Upper
ATTR→IC→ATTI	0.019	0.018	1.078	−0.008	0.050
INF→IC→ATTI	0.048	0.017	2.803 **	0.023	0.080
INT→IC→ATTI	0.074	0.020	3.619 ***	0.044	0.110
AUG→IC→ATTI	0.059	0.020	3.030 **	0.030	0.094
ATTR→ENJ→ATTI	0.099	0.027	3.673 ***	0.057	0.146
INF→ENJ→ATTI	0.073	0.020	3.614 ***	0.044	0.111
INT→ENJ→ATTI	0.072	0.021	3.415 ***	0.040	0.110
AUG→ENJ→ATTI	0.068	0.023	2.906 **	0.030	0.106
ATTR→IC→ATTI→PI	0.007	0.007	1.003	−0.003	0.019
INF→IC→ATTI→PI	0.017	0.007	2.335 *	0.007	0.031
INT→IC→ATTI→PI	0.026	0.009	2.794 **	0.013	0.043
AUG→IC→ATTI→PI	0.021	0.008	2.478 *	0.009	0.036
ATTR→ENJ→ATTI→PI	0.034	0.012	2.868 **	0.017	0.056
INF→ENJ→ATTI→PI	0.025	0.009	2.826 **	0.013	0.043
INT→ENJ→ATTI→PI	0.025	0.009	2.755 **	0.012	0.042
AUG→ENJ→ATTI→PI	0.024	0.010	2.382 *	0.009	0.041

Note(s): * p < 0.05; ** p < 0.01; *** p < 0.001; β: path coefficients; SE: standard errors.

presents several statistically significant indirect effects, showing that product- and technology-related features partially exert their influence through cognitive evaluations and emotional experiences.

More specifically, enjoyment (ENJ) significantly mediated the effects of informativeness (INF), interactivity (INT), and augmentation (AUG) on advertising attitude (ATTI) (p < 0.01 for all paths). The indirect effects included: INF→ENJ→ATTI (β = 0.073, p < 0.001), INT→ENJ→ATTI (β = 0.072, p < 0.001), and AUG→ENJ→ATTI (β = 0.068, p < 0.01), supporting enjoyment’s role as a key mechanism along the peripheral route. Simultaneously, information credibility (IC) also mediated the effects of INT and AUG on ATTI, including a significant indirect path from AUG→IC→ATTI (β = 0.099, p < 0.001), highlighting the central route through cognitive evaluation.

The analysis also identified several serial mediation paths, such as INF→ENJ→ATTI→PI and INT→ENJ→ATTI→PI, which further demonstrated how emotional experience and advertising attitude bridge the path to purchase intention. For instance, the indirect effect of INT→ENJ→ATTI→PI was β = 0.025 (p < 0.01), while AUG→IC→ATTI→PI yielded an indirect effect of β = 0.021 (p < 0.05).

In summary, the results confirm that enjoyment (ENJ) and information credibility (IC) function as key affective and cognitive mediators, respectively, in the relationship between perceived ad content and consumer responses. These findings align with the dual-process assumptions of the elaboration likelihood model (ELM), which predicts that consumers process information through emotional and rational pathways.

Finally, this study also examined whether flow experience moderates the relationships between advertising content features and the mediating variables. Per

Table 8. Moderation analysis.

Interaction Term	Path Coefficient (β)	Hypothesis	Results
FL×ATTR→IC	0.017	H8a	Not Supported
FL×INF→IC	−0.059	H8b	Not Supported
FL×ATTR→ENJ	0.127 *	H8c	Supported
FL×INF→ENJ	−0.061	H8d	Not Supported
FL×INT→IC	0.048	H8e	Not Supported
FL×AUG→IC	−0.071	H8f	Not Supported
FL×INT→ENJ	−0.037	H8g	Not Supported
FL×AUG→ENJ	−0.150 *	H8h	Supported

Note: * p < 0.05.

, two of the eight interaction effects reached statistical significance (p < 0.05). Specifically, flow significantly strengthened the positive effect of attractiveness (ATTR) on enjoyment (ENJ) (β = 0.127, p < 0.05), supporting H8c. This result suggests that visual attractiveness becomes more effective at eliciting emotional responses under conditions of high immersion. Conversely, flow significantly weakened the positive relationship between augmentation (AUG) and enjoyment (β = –0.150, p < 0.05), supporting H8h. This finding implies that in highly immersive states, the added value of augmentation features may diminish, potentially due to sensory saturation or a reduced sense of novelty.

6. General Discussion

Grounded in the elaboration likelihood model [23], this study investigated how different content features in augmented reality (AR) advertising, specifically, product-related attributes (attractiveness and informativeness) and technology-related attributes (interactivity and augmentation), influence consumers’ attitudes toward the advertisement and purchase intentions through the central route (information credibility) and peripheral route (enjoyment). The study developed and empirically tested a moderated mediation model using PLS-SEM to reveal the psychological mechanisms and structural pathways through which AR ad content affects consumer responses.

The empirical findings demonstrate that rational and sensory cues in AR advertising affect consumer responses in distinct ways. Informativeness and augmentation significantly enhanced information credibility, suggesting that ad content characterized by high realism and rich informational depth increases consumers’ perceived trust in the message. This result aligns with Tsai et al. [16], who emphasized the role of realism and media richness in boosting credibility in AR contexts. In contrast, attractiveness significantly influenced enjoyment but did not affect information credibility. This result suggests that while visual expressiveness enhances emotional experiences [28,66], it does not strengthen consumers’ cognitive evaluation of message reliability, functioning more as an affective trigger than a rational cue.

Interactivity and augmentation positively influenced information credibility and enjoyment, reinforcing their dual roles in activating central and peripheral processing routes [76]. However, these effects may depend on the quality of ad design and the effective integration of technological elements. Overly simplistic interactivity or poorly contextualized augmentation may weaken emotional impact. Moreover, since participants engaged with the AR ads through video demonstrations rather than direct interaction, the passive viewing mode likely limited the emotional intensity that interactive AR typically elicits. Future research should incorporate fully immersive environments or greater user agency to garner a better understanding of how technological content interacts with emotional responses.

Regarding the structural model, information credibility and enjoyment significantly predicted advertising attitude, which in turn influenced purchase intention. These findings confirm that central and peripheral routes jointly shape behavioral intentions in AR advertising. Mediation analysis further validated this dual-route process, reinforcing the relevance of the ELM in immersive media environments. Unlike prior research focused on static, two-dimensional stimuli, this study extends ELM’s theoretical boundaries by applying it to interactive and immersive advertising formats such as AR.

Although prior literature emphasizes the importance of flow states in immersive media [14], the current study did not observe a significant moderating effect of flow across all AR advertising paths. One possible explanation lies in the nature of the stimuli, i.e., video demonstrations without real-time interactivity may have failed to induce a genuine flow state. Moreover, as a subjective psychological experience, flow varies across individuals and contexts, making it difficult to measure reliably. Variations in participants’ situational awareness or engagement likely affected the accuracy of flow assessments in this study.

6.1. Theoretical Implications

The findings offer several insights for theory. First, by integrating ELM with AR advertising, the study introduces and validates a “dual-feature–dual-route” framework that explains how content characteristics influence engagement through cognitive and affective mechanisms, and extends the applicability of the ELM to immersive media environments. Second, by distinguishing between product-related and technology-related features, the study expands the conceptualization of AR advertising beyond traditional perceptual dimensions and offers a structured measurement model for future research. Third, by mapping dual processing routes onto cognitive, emotional, and behavioral engagement outcomes, the study identifies multiple mediated pathways that explain how AR ad content shapes consumer attitudes and purchase intentions, highlighting the joint role of rational evaluation and emotional experience in shaping consumer responses. This insight contributes to the theoretical understanding of how rational evaluation and emotional experience operate concurrently within immersive advertising.

6.2. Practical Implications

From a managerial perspective, the findings suggest that advertisers should balance rational and emotional appeals in AR ad design. Providing detailed and specific product information enhances perceived credibility, while strengthening visual appeal, interactivity, and immersion can increase enjoyment and engagement motivation. Rather than treating AR as a mere “technological showcase,” firms should adopt a consumer-centric approach that constructs perceptually valuable, contextually immersive, and operationally coherent user journeys. While flow did not moderate the model significantly in this study, it remains a valuable psychological indicator of immersive engagement. Future AR ad design can still benefit from using flow as a reference point for optimizing interactivity and user experience.

6.3. Limitations and Future Research

Despite making theoretical and empirical contributions, this study has several areas that future research should address. First, the study employed an online survey design and presented AR advertising content through video demonstrations. While this approach helped control for confounding variables, it lacked the interactivity and immersion of real-world AR experiences. As a result, participants may have evaluated technological features like augmentation and interactivity differently than they would in hands-on scenarios, potentially affecting the depth of cognitive processing and emotional responses. Future studies should consider immersive experimental platforms, field-based experiences, or real-time mobile AR interactions to capture more effectively consumers’ perceptual pathways and behavioral responses.

Second, flow, as a psychological moderator, is highly individualized and context-dependent. This study measured flow using self-reported scales without enabling actual physical interaction, relying solely on video exposure. This method may have limited the activation of immersive states and contributed to the non-significant moderation results. Future studies could enhance experimental realism by incorporating gamification elements, AR navigation tasks, or customized interaction scenarios to elicit flow more effectively. Combining these setups with process-based measures (e.g., eye-tracking, response times) could also improve measurement precision.

Finally, the sample primarily comprised young Chinese consumers recruited through the Credamo survey platform. While this tech-savvy demographic provided relevant insights, the limited diversity restricts the generalizability and cross-cultural applicability of the findings. Given that perceptions and acceptance of AR advertising may vary based on cultural values, media literacy, and technological familiarity, future research should conduct comparisons across countries, age groups, or media usage patterns to test the robustness of the proposed model in different contexts.

7. Conclusions

This study investigated how content features of augmented reality advertising—specifically product-related (attractiveness, informativeness) and technology-related (interactivity, augmentation) elements—influence consumer engagement through different psychological pathways, and whether flow moderates these effects. Drawing on the elaboration likelihood model (ELM), the findings confirm that AR advertising content activates both central and peripheral routes of information processing. Informativeness and augmentation enhance information credibility, leading to stronger cognitive engagement, while attractiveness and interactivity primarily stimulate affective engagement through increased enjoyment. These psychological responses further shape consumers’ attitudes toward the ad and their purchase intentions, validating the dual-route processing model in immersive media contexts. The moderating role of flow was only partially supported, suggesting that its influence may depend on the level of immersion and the nature of user interaction. These results extend the theoretical applicability of the ELM by demonstrating its relevance in dynamic, multi-sensory advertising environments like AR. Practically, the study provides insights for marketers and designers, emphasizing the need to integrate informative, emotionally resonant, and technologically engaging content to foster multidimensional consumer engagement. Future research could explore fully interactive AR experiences, examine individual differences in flow susceptibility, or investigate long-term behavioral outcomes to further advance understanding of AR advertising effectiveness.