Sage Journals: Discover world-class research

Abstract

Objective

The impact of augmented reality (AR) exercise apps on health-related behaviors and outcomes remains underexplored. Grounded in affordance theory, this study investigates how the utilitarian, hedonic, and social affordances of these apps influence user engagement, ultimately affecting continuance intention and subjective well-being in China.

Methods

Data were collected via an online survey of 399 users of AR fitness apps in China, recruited through the professional research platform Credamo.com. All constructs were measured with validated scales adapted to the AR exercise context. The hypothesized relationships were tested using structural equation modeling (SEM).

Results

The results indicate that utilitarian (environmental embedding, visibility), hedonic (playfulness), and social (competition) affordances are positively associated with user engagement. This engagement, in turn, positively influences users’ subjective well-being and their intention to continue using AR exercise apps.

Conclusion

This study extends affordance theory by identifying the affordances in AR fitness apps and validating user engagement as the critical underlying mechanism. The findings suggest that designers should prioritize these affordances and health policymakers should leverage them to develop effective, AR-driven digital sports initiatives.

Keywords

AR exercise apps health-related behavior subjective well-being affordance theory user engagement

1. Introduction

Augmented Reality (AR) integrates digital content into the physical environment, transforming how users perceive and interact with their surroundings.^1,2 Various fields leverage AR for different applications, notably in healthcare, retail, education, and tourism.³ According to Statista,⁴ the mobile AR market was valued at USD 11.9 billion in 2024, with projections indicating a rise to USD 13.8 billion in 2025. In the fitness sector, AR exercise apps have emerged as a promising frontier in digital health.⁵ These apps enable users to interact with virtual elements, such as workout instructions, virtual coaches, and immersive scenarios, while engaging in physical activities and exercises. A notable example is Tiantian Rope Skipping, an app that offers AR-enhanced workouts (e.g., rope jumping and fitness programs) to over 4.9 million predominantly young urban Chinese users by superimposing virtual coaches, interactive scenarios, and real-time feedback onto physical exercise via smartphones or tablets.⁶

The primary goal of AR-enabled fitness technologies is to improve users’ health-related behavior and outcomes.⁷ Aligning with the digital health literatures that emphasize sustainable behavior change and holistic health enhancement,^5,8,9 this study elucidates how AR exercise apps lead to positive health behaviors (e.g., continued use of AR exercise apps) and derive health outcomes (e.g., subjective well-being). Continued use of AR exercise apps reflects health maintenance behavior¹⁰ and serves as a critical factor in achieving long-term health benefits.¹¹ However, fitness apps often face the challenge of user abandonment, as individuals often fail to engage with them consistently over time.^12,13 Meanwhile, subjective well-being captures the psychological benefits of AR technology-mediated exercise, assessing the extent to which immersive experiences contribute to users’ emotional fulfillment and perceived quality of life. Given that enhancing well-being is a central goal of many public policies and considering scholarly calls for research on the sustained use of fitness technologies, this study investigates how AR exercise apps can effectively promote both continuance intention and subjective well-being.

While existing AR studies have primarily focused on domains such as retailing, tourism, and education,^14,15 scholarly understanding of its application in fitness remains limited. Although a growing body of work has begun to examine AR technology in exercise settings, these studies describe that AR can be effective but lack the framework to explain how it works. It remains unclear how and why AR exercise apps are effective in generating wellness outcomes, as well as the psychological mechanisms through which they influence user perceptions and health-related behaviors.

Affordance theory, rooted in ecological psychology¹⁶ and adapted to technology research, highlights users’ perceptions of a technology’s potential to enable specific actions.¹⁷ This theoretical perspective provides the necessary framework for AR fitness, as it moves beyond a feature-centric view to explain how the relationship between the user, the AR app, and the exercise environment facilitates specific outcomes. Although existing studies have explored the utilization of AR-enabled fitness technologies and the effects of their use,^7,13,18,19 research through an affordance lens remains scarce. Therefore, our study applies and extends the affordance perspective to investigate the mechanisms through which AR exercise apps foster health behaviors and outcomes. Furthermore, the features of affordances are not universal but are multidimensional and context-dependent.²⁰ Therefore, identifying the multidimensional nature of the affordances in the AR exercise apps context is crucial. Thus, this leads to our first research question:

RQ1

What are the dimensions of AR exercise apps’ affordances, and how do these affordances influence psychological perception and health-related behavior?

Furthermore, user engagement has been widely recognized as a key psychological mechanism through which technology affordances lead to positive outcomes.^21,22 Evidence from mobile app research indicates that affordances can stimulate user engagement, which subsequently shapes behavioral intentions.^23,24 User engagement emphasizes the deep involvement and interactions with AR exercise apps, its specific mediating role between AR affordances and deeper outcomes like subjective well-being and continuance intention remains inadequately explored. By treating engagement as a multidimensional construct comprising cognitive, emotional, and behavioral dimensions,¹⁴ this research clarifies the underlying psychological mechanism between technology affordances and outcomes. This leads to our second research question:

RQ2

How does multidimensional user engagement mediate the relationship between AR exercise app affordances and outcomes, including subjective well-being and continuance intention?

To address these questions, we construct a comprehensive research model by integrating affordance theory to investigate how AR exercise app affordances foster continuance intention and subjective well-being through the mediation of user engagement. This study makes the following theoretical contributions. First, this study extends digital health research by delineating the mechanisms through which AR exercise apps drive health-related behaviors and outcomes, thereby providing empirical evidence for their effectiveness. Second, the research advances affordance theory by contextualizing it within AR fitness, identifying three salient affordance dimensions: utilitarian, social, and hedonic, that align with user exercise motivations and technological features. By contextualizing these three dimensions, our study enriches the existing literature by explaining how AR-specific actionable possibilities shape key outcomes such as well-being and sustained usage. Third, this study explores the mediating role of user engagement between affordance and AR exercise app use outcomes, providing a deeper understanding of how cognitive, affective, and behavioral engagement translate technological features into tangible benefits. Additionally, this study also offers valuable guidance for stakeholders in the fitness and technology industries, paving the way for the development of more effective and engaging AR-based fitness solutions.

2. Theoretical background

2.1. AR exercise apps

AR exercise apps are digital platforms that integrate AR technology to enhance physical exercise experiences by overlaying digital information, visuals, or interactive elements onto the real-world environment.¹³ Unlike conventional fitness technologies such as step counters or instructional apps, these apps transform the physical space into an interactive exercise environment, enabling users to engage with virtual elements in real time and in context.^13,18

Recent research on AR in sports and fitness indicates that AR technology effectively fosters behavioral changes by enhancing motivation and situational presence.¹⁹ For example, Jeon and Kim¹⁸ discovered that AR-based exercise programs enhanced the sustainability of physical activity and mitigated muscle atrophy in older adults. Sun and Yuan¹³ characterized AR exercise apps as “virtual gyms” that positively affect users’ psychological perception and experiential value. Together, these previous studies demonstrate that by shifting from passive monitoring and screen-based instruction to situated and interactive coaching, AR exercise apps have the potential to enhance user engagement and improve health outcomes.

2.2. Affordance theory

The concept of affordance, originating from ecological psychology, highlights the opportunities or specific actions that the environment offers to individuals.¹⁶ This foundational notion has been expanded into the field of technology and defined as “the possibilities for goal-oriented actions afforded to specified user groups by the technical functionalities of technical objects”²⁵ (p. 622). Affordance theory emphasizes the dynamic interplay between technological capabilities and user goals within specific contexts. Crucially, affordances are not inherent in the technology itself but are actualized through users’ perceptions and interactions with those contexts.²⁶

Recent research has applied affordance theory to diverse digital contexts, including social media platforms,²⁷ brand apps,¹⁷ live streaming commerce,²⁸ and fitness apps,²² as shown in Table 1. These studies have largely examined how technological affordances shape user experiences and behavioral outcomes.^17,21 Notably, the dimensions of affordances exhibit significant variation across different technological contexts. Scholars have called for further empirical investigation into how affordances manifest and function in understudied domains.^17,29 Therefore, this study extends the affordance lens to the domain of AR exercise apps. We identify the context-specific dimensions of affordances in the AR exercise apps context and examine how they promote health-related behaviors and outcomes through the mediating role of user engagement.

Table 1.

Summary of Affordances in digital contexts.

Sources	Affordances	Mediating variables	Outcomes	Research context	Key finding
22	Interactivity, navigability, customization	Relatedness, competence, Autonomy	Engagement with fitness trackers, use of fitness trackers	Fitness trackers technology	Three affordances significantly increased individuals’ sense of autonomy, competence, and relatedness, which in turn led to increased engagement with and use offitness trackers.
30	Health technician affordance and health consultant affordance	Performance expectancy, effort expectancy and exercise self-efficacy	Health-promoting behavior	Wearable fitness devices	The relationship between both health technician and health consultant affordances and health-promoting behavior is mediated by performance expectancy and exercise self-efficacy
17	Visibility, persistence, interactivity, association, selectivity	Value-in-use	Continuance intention, brand loyalty	Branded app	Five affordances are linked to value-in-use, which contributes both directly and indirectly to the two loyalty-related outcomes, through either brand competence or brand warmth.
31	Immersion, presence	Utility, empathy, embodiment	playability	augmented reality games	Technological properties of the ARG system affect opportunities for action available in the environment.
27	Anonymity, Editability, Visibility, Ephemerality	Perceived autonomy, skepticism self-efficacy, self-indulgence	Fact-authenticating, counterarguing, instant sharing	Social media platforms	Four affordances enhance user autonomy, which further leads to cognitive-affective consequences, impacting multiple behavioral responses.
28	Utilitarian (visibility), social (social presence, interactivity, self-presentation), and hedonic affordances (entertainment)	Perceived diagnosticity, psychological distance	Purchase intention	Live commerce	Utilitarian affordances positively contribute to perceived diagnosticity, whereas entertainment and the three social affordances reduce psychological distance.
This study	Utilitarian (environmental embedding, visibility), hedonic (playfulness), and social (competition) affordances	User engagement	subjective well-being and continuance intention	AR exercise apps	Utilitarian, hedonic and social affordances are positively associated with user engagement, and in turn positively influences users’ subjective well-being and their intention to continue using AR exercise apps.

Given that users typically seek instrumental performance, intrinsic enjoyment, and social connection when exercising with fitness apps,³² and considering the immersive and interactive characteristics of AR exercise apps,¹³ we adopt the classification framework proposed by Wang, Li²⁸ to categorize affordances into three dimensions: utilitarian, hedonic and social affordance. These dimensions closely correspond to users’ primary goals during exercise and illustrate the mutuality between the user, the features of AR fitness apps, and the exercise context. Specifically, utilitarian affordance serves to fulfill instrumental goals by assisting users in achieving their exercise and fitness goals more effectively. Hedonic affordance refers to the pleasure-oriented capabilities offered by AR exercise apps. Social affordance encompasses those capabilities that foster and sustain social connections among users.

Within these dimensions, we further identify five specific affordances: environmental embedding, anytime-anyplace connectivity, visibility, playfulness, and competition. Firstly, environmental embedding, anytime-anyplace connectivity, and visibility constitute the pivotal utilitarian affordances that facilitate users in attaining fitness goals through AR exercise apps. Specifically, environmental embedding pertains to the extent to which users can obtain integrated and authentic experiences by visually integrating virtual content into their real-world environment.³³ This affordance enhances the user’s sense of spatial presence by making digital objects appear naturally situated in their surroundings. Anytime-anyplace connectivity illustrates “a physical association between users and technology artifacts that afford the users access to services anytime and anywhere”³⁴ (p. 3). Rooted in mobile ubiquity, this affordance supports instrumental goals such as consistency and accessibility.¹³ Visibility indicates the potential for users to access the clear, intuitive presentation of fitness information, such as instructions, performance data, and progress metrics within the AR exercise interface.²⁰ This affordance ensures that workout feedback, guidance, and statistical cues are overlaid in an intuitive and legible manner, enabling users to monitor their exercise performance. Collectively, these affordances enhance the functional value of AR exercise apps: environmental embedding establishes an immersive context, visibility supports cognitive monitoring of performance, and connectivity provides ongoing accessibility.

Furthermore, playfulness represents a core hedonic affordance, pertaining to the capacity of AR apps to elicit intrinsic enjoyment.^31,35 Playfulness encompasses a variety of enjoyable interactions facilitated by AR technology, such as game-like interactions, imaginative scenarios, and intrinsically motivating challenges, emphasizing its potential to make the exercise experience more captivating and pleasurable.³⁶ Additionally, competition embodies the social affordance dimension, reflecting users’ perceptions that AR exercise apps enable them to compare their performance with others.³⁷ Through features such as leaderboards, real-time rankings, and asynchronous challenges, this affordance stimulates social comparison and strengthens interpersonal connections among users.

2.3. User engagement

O’Brien, Cairns³⁸ defined user engagement as “a quality of user experience characterized by the depth of an actor’s cognitive, temporal, affective, and behavioral investment when interacting with a digital system” (p. 29). This concept transcends superficial interaction, representing a state of deep connection and involvement that arises from meaningful immersion in technology-mediated activities. It is widely regarded as a key driver of positive behavioral outcomes, such as user retention,^21,23 WOM intention,²⁴ and co-creation.³⁹ These outcomes underscore the importance of fostering deep and sustained engagement, particularly in interactive digital platforms such as fitness apps.

By overlaying interactive virtual content onto the user’s physical environment, AR exercise apps create a uniquely immersive platform that fosters user engagement. Unlike traditional fitness apps, which often depend on static screens and pre-recorded content, AR exercise apps integrate physical movement, spatial awareness, and real-time feedback into the workout experience, enabling embodied and contextually embedded interactions.³³ This integration of digital elements with real-world environments has the potential to intensify users’ cognitive absorption, emotional excitement, and behavioral persistence, thereby transforming routine physical activity into a compelling and engaging experience. Previous studies have revealed that AR technology significantly promotes user engagement.^40,41 For instance, Jessen, Hilken⁴² showed that AR enables creative customer engagement by visually displaying the relations between products and services. However, most studies on AR and engagement have centered on retail, education, and tourism contexts.^1,14,15,43 Despite the rising popularity of AR in fitness, little is known about its effect on multidimensional user engagement in this domain. As Zhou, Krishnan²² highlighted, understanding engagement with fitness technology is vital for promoting sustained use and lasting health benefits. Therefore, understanding how AR exercise apps affordances shape user engagement is critical to unlocking their full potential in promoting long-term fitness adherence and subjective well-being.

This study considers user engagement as a multidimensional construct, encompassing cognitive, emotional, and behavioral dimensions.^14,21,44 Cognitive engagement reflects the extent of cognitive effort and focused attention that users invest during their exercise process.²³ High cognitive engagement means users are mentally absorbed in the AR-enhanced workout, actively processing real-time feedback, and maintaining sustained concentration to overcome virtual challenges. Emotional engagement pertains to the feelings and emotional connections that users develop with AR exercise apps.^14,45 Emotional engagement involves experiencing positive affective states such as excitement, enjoyment, and passion during AR exercise sessions. Behavioral engagement reflects the observable actions and effort users exert in interacting with the AR exercise app.⁴⁵ It includes the time, energy, and persistence invested in using the app, such as completing full workout sessions, exploring new AR features and exercise programs, or consistently tracking fitness progress.

3. Research model and hypothesis developments

Building on the existing literature, our study constructs a research model to examine how technology affordance influences subjective well-being and continuance intention through user engagement. Specifically, by classifying affordance into three dimensions, we specify the effects of utilitarian, social, and hedonic affordances on user engagement (H1-H5). Subsequently, we propose hypotheses delineating the relationships between user engagement and subjective well-being (H6) as well as continuance intention (H7). In addition, user engagement is treated as a second-order factor consisting of cognitive, affective and behavioral engagement (Figure 1).

Figure 1.

The research model.

3.1. Utilitarian affordance and user engagement

Environmental embedding, regarded as an empowerment factor of AR,^2,43 enables users to align their exercises and postures with real-time virtual scenarios. This affordance transforms the exercise experience by creating a unified reality where digital objects coexist interactively with the user’s environment.¹³ The perception of environmental embedding evokes users’ sense of spatial presence and realism, which in turn is considered to enhance user engagement.⁴⁶ Furthermore, the seamless integration of virtual elements with the user’s bodily awareness fosters a state of sensory immersion,⁴⁷ characterized by focused attention and reduced distraction from the external environment. This immersive experience makes the exercise more captivating and enjoyable,³³ ultimately generating emotional involvement and positive arousal.^2,46 Therefore, we propose the following:

Hypothesis 1

Environmental embedding exerts a positive influence on user engagement.

As a foundational feature of mobile and AR technologies, anytime-anyplace connectivity affordance enables users to integrate physical activity seamlessly into their daily routine, whether at home, in the office, or while traveling.^13,40 These seamless accessibility offers users greater flexibility in scheduling and location choice, potentially lowering barriers to exercise initiation and encouraging more frequent physical activity engagement. As Alalwan, Algharabat⁴⁸ argued in the context of mobile shopping, ubiquitous connectivity is likely to cognitively and emotionally contribute to customer engagement. Although ubiquitous connectivity has become increasingly normalized in today’s digital ecosystem, its value may persist when combined with AR’s immersive and interactive features, as it transforms ordinary spaces into engaging workout environments. In the context of AR exercise apps, the ability to exercise in any setting without sacrificing interactivity or immersion may increase perceived convenience, reduce barriers to exercise, and foster greater enjoyment and satisfaction.⁴⁹ Thus, we postulate that:

Hypothesis 2

Anytime-anyplace connectivity exerts a positive influence on user engagement.

By enhancing information transparency and cognitive legibility, visibility enables users to effortlessly perceive and interpret their performance status.²⁰ For instance, when user see a real-time feedback cue such as “95% accuracy”, they can instantly evaluate the alignment between their movements and target performance standards. This clarity enables users to allocate greater attentional resources to comprehending and refining their exercise execution. This visibility affordance further amplifies goal-gradient motivation, leading users to intensify their effort as they perceive themselves moving closer to desired outcomes.¹⁹ Meanwhile, the presentation of positive performance feedback during training stimulates feelings of satisfaction and triggers positive emotional reactions.⁵⁰ In addition, consistent and positive performance feedback could reinforce perceived competence, a core psychological need in self-determination theory.⁷ As users encounter clear and confirming evidence of their progress, they cultivate stronger self-assurance in their capabilities, which in turn enhances intrinsic motivation and promotes deeper behavioral engagement.²¹ Therefore, the following hypothesis is proposed.

Hypothesis 3

Visibility exerts a positive influence on user engagement.

3.2. Hedonic affordance and user engagement

Playfulness, a core hedonic affordance of AR, may arise when users participate in activities such as virtual scavenger hunts in their living rooms, hitting rhythm-based targets that appear in sync with music, or receiving playful visual rewards for completing exercise sets.^51,52 Playfulness plays a crucial role in fostering users’ intrinsic motivation for autonomy,⁵³ which in turn has been demonstrated to enhance user engagement.^22,24 Furthermore, playfulness is often associated with positive emotions, such as joy, excitement, and amusement, and can facilitate immersive experience or flow states.^35,54 By positively influencing cognitive and emotional states, playfulness affordance encourages deeper engagement with AR exercise apps, promotes functional exploration, and prolongs interaction time. Consistent with this, Zhou, Lin³⁵ found that playfulness affordance strengthens the desire to engage and enriches emotional experiences in virtual co-creation contexts. Therefore, the following hypothesis is proposed:

Hypothesis 4

Playfulness exerts a positive influence on user engagement.

3.3. Social affordance and user engagement

Social affordance encompasses the social capabilities facilitated by AR fitness apps. While AR platforms support various social affordances, such as collaboration and social support, this study focuses on competition due to its prevalence in gamified fitness apps and its capacity to drive performance-oriented engagement through direct feedback and status differentiation.⁵² By incorporating gamified elements such as leaderboards, real-time rankings, and head-to-head challenges (e.g., “PK mode”), competition can elicit feelings of challenge, exhilaration, and enjoyment, thereby enhancing affective engagement and extrinsic motivation.⁵⁰ Drawing upon social comparison theory, competition motivates users to improve their performance through upward comparisons with others, thereby stimulating greater cognitive investment and participation in fitness activities.²¹ As such, competition can be an effective mechanism to capture users’ attention, evoke positive emotions, and deepen their involvement in exercise activities. Therefore, we propose that:

Hypothesis 5

Competition exerts a positive influence on user engagement.

3.4. User engagement and subjective well-being

Subjective well-being refers to an individual’s self-assessment of their quality of life and emotional states.⁵⁵ It encompasses both hedonic aspects, such as positive emotions and life satisfaction, and eudaimonic dimensions, including the sense of meaning and fulfillment. The ultimate objective of exercise apps is to assist users in establishing a healthy lifestyle and improving their overall well-being.^7,56 Research has shown that physical activity facilitated by digital platforms can enhance subjective well-being. For example, Iii, Seo⁵⁷ demonstrated the positive influence of participation in physical activity through social networking sites on individuals’ subjective well-being. Similarly, Hu, He⁵⁸ found that engagement with fitness apps reduces emotional exhaustion and enhances life satisfaction. In AR contexts, regular physical activity, stimulated by sustained engagement, promotes the release of endorphins and other neurochemicals associated with stress reduction and mood enhancement. In addition, highly engaged users experience satisfaction of their core psychological needs for autonomy, competence, and relatedness, thereby enhancing self-evaluation and perceived life quality.⁷ Therefore, we propose that:

Hypothesis 6

User engagement exerts a positive influence on subjective well-being.

3.5. User engagement and continuance intention

Continued intention, commonly termed post-adoptive intention, reflects the degree to which users intend to continue using AR exercise apps.^13,59 According to the Theory of Planned Behavior,⁶⁰ intention serves as a primary predictor of actual behavior, with stronger intention increasing the likelihood of consistent action. Previous research has revealed a positive relationship between user engagement and continuance intention, suggesting that individuals with high levels of cognitive, emotional, and behavioral engagement are more inclined to sustain their usage of digital services.^21,23,61,62 For example, Zhang, Shao²¹ found that user engagement positively influences user retention in mobile payment contexts. Cheung, Pires⁴⁵ confirmed the role of multidimensional engagement in fostering repurchase intention and ongoing usage behaviors. In the context of AR exercise apps, when users invest cognitive attention, develop emotional attachment, and maintain active behavioral involvement, they are more likely to intend to continue using these apps. Therefore, we propose that:

Hypothesis 7

User engagement exerts a positive influence on continuance intention.

4. Methods

4.1. Measurement of constructs

The measurement items for each latent construct were adapted from the previously validated scales in technology-mediated and health-related research, particularly studies involving AR environments, gamified systems, and mobile health apps, and were rephrased to fit the context of this study (see Table 2). All items were evaluated using a 7-point Likert scale (1 = strongly disagree to 7 = strongly agree). Specifically, items for environmental embedding were adapted from Sun and Yuan.¹³ Items for anytime-anyplace connectivity were adapted from Lee and Li.³⁴ Items for visibility were adapted from Sun, Shao.²⁰ Items for playfulness and competition were adapted from Zhou, Lin.³⁵ User engagement was measured as a reflective second-order construct, exhibited by three first-order dimensions: cognitive, emotional, and behavioral engagement.¹⁴ Items for cognitive engagement and behavioral engagement were adapted from Dessart, Veloutsou,⁶³ while those for emotional engagement were drawn from Zhang, Shao.²¹ Regarding the consequences of user engagement, two items measuring continuance intention were adapted from Bellary, Bala,¹¹ and three items measuring subjective well-being were adapted from Liang, Xin.⁶⁴

Table 2.

Construct, items and sources.

Construct	Items	Sources
Environmental embedding (ENV)	This AR exercise app allows me to integrate my various limbs in various virtual motion scenes.	Adapted from Sun and Yuan¹³
	This AR exercise app allows me to intuitively see how I feel in any virtual motion scenes.
	This AR exercise app allows me to intuitively observe my posture in any virtual motion scenes.
Anytime-anyplace connectivity (ANC)	I would be able to use the AR exercise app at anytime, anywhere.	Adapted from Lee and Li³⁴
	I have access to the AR exercise app whenever and wherever I need it.
	I would find the AR exercise app to be easily accessible, regardless of time and place.
Visibility (VIS)	The AR exercise app provide me with detailed instructions related to fitness.	Adapted from Sun, Shao²⁰
	The AR exercise app makes information about how to achieve fitness goals visible to me.
	The AR exercise app helps me to visualize my workout progress and performance in real time.
Playfulness (PLA)	Using the AR exercise app is a joy.	Adapted from Zhou, Lin³⁵
	Using the AR exercise app is enjoyable.
	Using the AR exercise app is a good experience and not boring.
Competition (COM)	The AR exercise app offers me the possibility to compare my performance with that of others.	Adapted from Zhou, Lin³⁵
	The AR exercise app offers me the possibility to compete with others.
	The AR exercise app offers me the possibility to threaten the status of others by my active participation.
Cognitive engagement (COE)	I like to learn more about the AR exercise activities on this AR exercise app.	Adapted from Dessart, Veloutsou⁶³
	When I’m doing the AR exercises on the APP, I forget everything around me.
	When I’m doing the AR exercises on the APP, I get carried away.
Emotional engagement (EME)	While using this AR exercise app, I feel excited about the exercise activities on this app.	Adapted from Zhang, Shao²¹
	While using this AR exercise app, I am passionate about the exercise activities on this app.
	The use of this AR exercise app makes me happy.
Behavioral engagement (BEE)	I spend a lot of time on the exercise activities available in the AR exercise app.	Adapted from Dessart, Veloutsou⁶³
	I consistently explore new exercise activities and programs available in the AR exercise app.
	I keep track of my personal fitness goals and progress within the AR exercise app.
Continuance intention (CON)	I am highly inclined to continue using the AR exercise app going forward.	Adapted from Bellary, Bala¹¹
Continuance intention (CON)	My intention is to keep using the AR exercise app in the future.	Adapted from Bellary, Bala¹¹
Subjective well-being (SUW)	My experience with the AR exercise app was memorable and enriched my quality of life.	Adapted from Liang, Xin⁶⁴
	I felt that my life was meaningful and fulfilling after using the AR exercise app.
	In general, I felt good about my life shortly after using the AR exercise app.

4.2. Survey design and data collection

This study employed an online survey to collect data for testing the theoretical model. The questionnaire comprised three sections. The initial section included a brief introduction to the study and a consent form. The form clarified that participation was voluntary and anonymous, described data protection measures, and assured participants that their responses would be used solely for academic purposes. These measures were taken to encourage honest responses and minimize social desirability bias. Participants indicated their consent by clicking an “I Agree” button before proceeding and were informed of their right to withdraw at any time. The second section began with two screening questions: “Have you ever used AR exercise apps in your life?” and “How often do you use the AR exercise apps?” These questions were used to filter out respondents who did not meet our requirements. The third section mainly collected the socio-demographic information of the participants. The questionnaire was developed in Chinese following a back-translation procedure to ensure conceptual equivalence.⁶⁵ The questionnaire was initially tested in a pilot study involving fifteen college students and two faculty members, and minor revisions were made based on their feedback.

After obtaining approval from the Institutional Review Board (IRB), we deployed the questionnaire on Credamo.com, a leading Chinese platform that provides services similar to MTurk. With over three million registered users, this platform has become a favored choice among scholars for distributing questionnaires.⁶⁶ This study relied on the platform’s default randomized stratified recruitment protocol for participant sampling. Each participant who completed the questionnaire got a financial reward.

To ensure data quality, several procedures were adopted to identify and exclude invalid responses. First, participants who completed the survey in less than 3 minutes were excluded, as the average completion time was approximately 5 minutes. Second, an attention-check item was embedded in the questionnaire, which instructed respondents to “select ‘strongly dissatisfied’.” Those who failed to correctly follow this instruction were removed. Third, respondents who provided identical answers to all questions were excluded. After applying these criteria, 399 valid questionnaires were retained for analysis. This sample size meets the recommended minimum of 10 times the largest number of structural paths directed at any latent construct in the model. Table 3 presents the demographic details of the respondents. Of the sample, 45.9% were male and 54.1% were female. Additionally, 72.2% of respondents were between 18 and 34 years old.

Table 3.

Demographics of the participants (N=399).

Demographic variables		Frequency	Percentage (%)
Gender	Male	183	45.9
Gender	Female	216	54.1
Age	18-24	81	20.3
	25-34	207	51.9
	35-44	83	20.8
	45	28	7.0
Education	High school or less	16	4.0
	Junior college/Undergraduate	307	76.9
	Master or above	76	19.0
Prior experience	Every day	99	24.8
	Several times a month	48	12.0
	Several times a month	224	56.1
	Occasionally used	28	7.0

4.3. Control variables

According to previous studies on the usage behavior of fitness apps and AR technology,^67–69 socio-demographic factors such as age, gender, education, and prior experience may influence empirical results. For instance, Zhu, Wang⁷⁰ demonstrated the role of gender in affecting users’ adoption of fitness apps. Therefore, this study considers age, gender, education, and prior experience with AR exercise apps as control variables. These control variables were measured categorically and incorporated into the structural equation model as exogenous variables to account for their potential influence on subjective well-being and continuance intention.

5. Empirical analyses and results

We employed partial least squares structural equation modeling (PLS-SEM) to test the research model. PLS-SEM is deemed appropriate for studies with sample sizes less than 500 and demonstrates flexibility in handling non-normally distributed data.⁷¹ In addition, it is suitable for exploratory studies and the examination of complex relationships.⁷² Therefore, PLS-SEM is particularly appropriate for the context of our research. Using SmartPLS 4.0, we first evaluated the measurement model to assess reliability and validity, followed by estimation of the structural model to test the hypothesized relationships.⁷²

5.1. Common method bias

Given that the data were solely self-reported and obtained from a single source, the potential for common method bias (CMB) exists in our study. To assess CMB issues, we first conducted Harman’s one-factor test.⁷³ The results indicated that the maximum covariance explained by a single factor was 38.83%, below the 50% threshold.⁷⁴ Second, we assessed the variance inflation factor (VIF) values, which ranged from 1.390 to 1.985, all below the threshold of 3.3,⁷⁵ indicating low multicollinearity and a negligible CMB treat. Third, we conducted the marker variable technique.⁷⁶ We used age as a marker variable because it had the lowest degree of correlation with the study variables. The findings revealed that the marker variable had no significant path coefficients with the primary variables, and the significance of all predicted paths did not change. Therefore, CMB is not a serious problem in our study.

5.2. Reliability and validity analysis

Table 4 reports the composite reliabilities (CR), Cronbach’s alpha, average variance extracted (AVE), and intercorrelations. As shown in Table 4, after removing two items from the continuance intention construct (CON3 and CON4) due to loadings below the 0.70 threshold, all remaining constructs show CR and Cronbach’s α values between 0.742 and 0.825, exceeding the recommended level of 0.7. Additionally, the AVE values range from 0.659 to 0.793, all exceeding the threshold of 0.5. Table 5 reports all items’ loadings and crossing-loadings. All item loadings for the first-order constructs surpass the recommended threshold of 0.7, demonstrating good convergent validity.

Table 4.

Reliability and Fornell-Larcker criterion.

	CR	Alpha	AVE	1	2	3	4	5	6	7	8	9	10
1.ANC	0.818	0.820	0.733	0.86
2.BEE	0.758	0.763	0.674	0.44	0.82
3.COE	0.747	0.753	0.664	0.36	0.59	0.82
4.COM	0.784	0.790	0.698	0.41	0.51	0.50	0.84
5.CON	0.739	0.745	0.793	0.41	0.50	0.55	0.40	0.89
6.EME	0.768	0.770	0.683	0.36	0.50	0.63	0.44	0.49	0.83
7.ENV	0.756	0.756	0.672	0.43	0.56	0.57	0.52	0.53	0.55	0.82
8.PLA	0.818	0.824	0.733	0.37	0.56	0.65	0.48	0.55	0.67	0.50	0.86
9.SUW	0.789	0.789	0.703	0.44	0.57	0.59	0.58	0.51	0.60	0.59	0.60	0.84
10.VIS	0.742	0.746	0.659	0.37	0.46	0.47	0.48	0.50	0.45	0.50	0.44	0.51	0.81

Note. The bold diagonal elements represent the square root of the AVEs for each construct.

Table 5.

Factor loading and cross-loading.

	ANC	BEE	COE	COM	CON	EME	ENV	PLA	SUW	VIS
ANC1	0.871	0.395	0.320	0.395	0.335	0.306	0.360	0.329	0.389	0.342
ANC2	0.856	0.378	0.318	0.347	0.387	0.332	0.372	0.322	0.386	0.327
ANC3	0.841	0.347	0.294	0.314	0.338	0.293	0.375	0.297	0.350	0.279
BEE1	0.396	0.859	0.520	0.497	0.409	0.457	0.479	0.489	0.544	0.431
BEE2	0.294	0.793	0.468	0.382	0.384	0.368	0.458	0.425	0.407	0.330
BEE3	0.382	0.810	0.471	0.371	0.427	0.400	0.430	0.451	0.447	0.375
COE1	0.385	0.541	0.838	0.462	0.547	0.577	0.505	0.585	0.563	0.495
COE2	0.250	0.453	0.771	0.350	0.338	0.439	0.448	0.466	0.411	0.319
COE3	0.242	0.450	0.833	0.405	0.440	0.511	0.424	0.527	0.456	0.327
COM1	0.353	0.413	0.436	0.834	0.358	0.361	0.397	0.353	0.497	0.442
COM2	0.338	0.417	0.361	0.813	0.311	0.332	0.425	0.407	0.448	0.365
COM3	0.343	0.449	0.453	0.861	0.341	0.405	0.489	0.447	0.511	0.388
CON1	0.370	0.463	0.519	0.386	0.904	0.460	0.509	0.500	0.451	0.446
CON2	0.367	0.416	0.455	0.329	0.876	0.419	0.430	0.472	0.451	0.439
EME1	0.263	0.419	0.486	0.383	0.350	0.807	0.451	0.512	0.448	0.337
EME3	0.316	0.452	0.555	0.393	0.451	0.850	0.462	0.567	0.522	0.448
EME4	0.320	0.363	0.515	0.312	0.422	0.821	0.438	0.583	0.514	0.321
ENV1	0.363	0.437	0.464	0.432	0.473	0.453	0.816	0.435	0.481	0.407
ENV2	0.349	0.469	0.461	0.456	0.423	0.461	0.821	0.418	0.511	0.430
ENV3	0.347	0.458	0.463	0.400	0.406	0.427	0.822	0.368	0.461	0.387
PLA1	0.330	0.492	0.581	0.464	0.457	0.635	0.449	0.882	0.520	0.345
PLA3	0.319	0.495	0.569	0.417	0.498	0.581	0.450	0.855	0.563	0.409
PLA4	0.299	0.436	0.512	0.349	0.448	0.496	0.370	0.831	0.443	0.380
SUW1	0.357	0.473	0.475	0.462	0.355	0.485	0.470	0.466	0.840	0.404
SUW2	0.384	0.493	0.493	0.482	0.470	0.530	0.492	0.522	0.839	0.399
SUW3	0.362	0.469	0.514	0.518	0.445	0.491	0.524	0.511	0.835	0.488
VIS1	0.296	0.365	0.376	0.357	0.397	0.359	0.402	0.392	0.406	0.797
VIS3	0.297	0.410	0.409	0.427	0.410	0.411	0.416	0.386	0.464	0.829
VIS4	0.310	0.350	0.363	0.374	0.403	0.315	0.394	0.288	0.374	0.811

Note. The bold values represent the within-construct item loadings.

For discriminant validity testing, the square root of the AVE for each construct (shown in bold on the diagonal in Table 4) was found to be greater than its correlations with other constructs (off-diagonal entries). Furthermore, the cross-loadings results in Table 5 indicate that each item loads higher on its intended construct than on any other construct. In addition, as presented in Table 6, all HTMT values were below the recommended threshold of 0.85.⁷⁷ Together, these results indicate an acceptable discriminant validity.

Table 6.

The HTMT results.

	ANC	BEE	COE	COM	CON	EME	ENV	PLA	SUW
BEE	0.552
COE	0.459	0.785
COM	0.513	0.657	0.647
CON	0.531	0.660	0.728	0.526
EME	0.457	0.650	0.824	0.563	0.653
ENV	0.548	0.733	0.750	0.678	0.705	0.716
PLA	0.451	0.702	0.824	0.596	0.701	0.841	0.628
SUW	0.545	0.734	0.762	0.737	0.661	0.769	0.765	0.739
VIS	0.474	0.612	0.625	0.621	0.670	0.586	0.663	0.562	0.666

5.3. Structural model testing

To assess the significance of the path coefficients, a bootstrapping procedure with 5,000 resamples was employed, applying bias-corrected and accelerated confidence intervals across all model paths. This approach also enabled evaluation of the model’s explanatory power (R²) and effect sizes (f²). The results are presented in Figure 2 and Table 7. Except H2, all hypotheses were supported, as indicated by p-values below 0.05 and t-values exceeding 1.96. Specifically, the affordances of environmental embedding (β = 0.269, p < 0.001), visibility (β = 0.123, p < 0.01), playfulness (β = 0.469, p < 0.001), and competition (β = 0.114, p < 0.01) exerted significant positive effects on user engagement, supporting H1, H3, H4, and H5. However, the anytime-anyplace connectivity affordance did not exert a significant influence on user engagement (β = 0.076, p > 0.05), thus H2 was rejected. Furthermore, user engagement positively impacted subjective well-being (β = 0.694, p < 0.001) and continuance intention (β = 0.607, p < 0.001), supporting H6 and H7. Finally, the first-order factors (cognitive engagement, emotional engagement, and behavioral engagement) were largely and significantly loaded on their second-order factor (user engagement).

Figure 2.

Results of hypothesis testing.

Table 7.

Hypotheses test results.

Hypotheses	Paths	β	S.D.	T-values	P-values	f²	VIF	Remarks
H1	ENV → UEN	0.269^***	0.047	5.681	0.000	0.136	1.723	Supported
H2	ANC → UEN	0.076^ns	0.040	1.875	0.061	0.014	1.347	Unsupported
H3	VIS → UEN	0.123^**	0.040	3.047	0.002	0.032	1.526	Supported
H4	PLA → UEN	0.469^***	0.051	9.271	0.000	0.465	1.530	Supported
H5	COM → UEN	0.114^**	0.036	3.117	0.002	0.025	1.648	Supported
H6	UEN → SUW	0.684^***	0.032	21.330	0.000	0.801	1.145	Supported
H7	UEN → CON	0.595^***	0.035	16.821	0.000	0.497	1.145	Supported

Additionally, as shown in Figure 2, our model explains 69.0%, 49.0%, and 37.7% of the variance (R²) in user engagement, subjective well-being, and continuance intention, respectively. These R² values indicate moderate to substantial explanatory power. Furthermore, following Hair,⁷⁸ f² values of 0.02, 0.15, and 0.35 indicate small, medium, and large effect sizes, respectively. The results in Table 7 show that three links (e.g., playfulness to user engagement, user engagement to subjective well-being, and user engagement to continuance intention) depict a large effect size with f² values greater than the threshold 0.350. The f² values of 0.014 from anytime-anyplace connectivity to user engagement, less than 0.02, demonstrate a negligible effect. The remaining paths depict small effect sizes (0.02 < f² < 0.15). Regarding control variables, only education significantly influenced subjective well-being (β = 0.067, p < 0.05), showing a positive association. Age, gender, and prior experience showed no statistically significant effects on either subjective well-being or continuance intention.

The mediation effect of the model was further examined using a two-tailed test and bias-corrected accelerated bootstrapping with 5,000 subsamples. As shown in Table 8, all indirect effects were statistically significant, with the 95% confidence intervals not encompassing zero. This supports the conclusion that user engagement fully or partially mediates the relationships between affordances and both continuance intention and subjective well-being. Specifically, user engagement fully mediated the effects of visibility on subjective well-being (direct effect: β = 0.094, 95% CI [-0.008, 0.195]), and of competition on continuance intention (direct effect: β = -0.041, 95% CI [-0.150, 0.027]). For all other paths, both direct and indirect effects were significant, indicating partial mediation.

Table 8.

Mediating effect results.

Path	Direct effect			Indirect effect			Results
Path	Effect	LLCI	ULCI	Effect	LLCI	ULCI	Results
ENV→UEN→ CON	0.166	0.064	0.266	0.062	0.026	0.113	Partial
ENV→UEN→ SUW	0.149	0.018	0.296	0.081	0.046	0.128	Partial
VIS→UEN→ CON	0.177	0.077	0.270	0.028	0.008	0.058	Partial
VIS→UEN→ SUW	0.094	-0.008	0.195	0.037	0.010	0.074	Full
PLA→UEN→ CON	0.187	0.080	0.108	0.279	0.047	0.185	Partial
PLA→UEN→ SUW	0.148	0.045	0.247	0.140	0.075	0.210	Partial
COM→UEN→ CON	-0.041	-0.150	0.027	0.027	0.007	0.058	Full
COM→UEN→ SUW	0.183	0.071	0.307	0.034	0.011	0.065	Partial

6. Discussion and implications

6.1. Discussion

This research elucidates how the affordances of AR exercise apps influence user engagement and, in turn, shape users’ subjective well-being and continuance intention. Our empirical results yield several research findings.

First, the results demonstrate that utilitarian affordances, particularly environmental embedding and visibility, positively influence user engagement. Although the effect size of environmental embedding was small (f² < 0.15), its significant positive impact is consistent with previous findings^2,41,42 that AR’s capacity to integrate digital content into physical environments enhances user interaction experience. Visibility also exhibited a significant positive effect on engagement, with small effect sizes (f² < 0.15). This finding aligns with prior research by Chen, Zhu⁷⁹ and Li, Lu,⁸⁰ which emphasizes the importance of perceptual clarity and feedback visibility in shaping users’ mental states and behavioral intention. In the context of AR exercise apps, enhancing the visibility of interactive elements, such as real-time performance metrics or progress indicators, may help sustain user engagement over time.

However, our analysis did not reveal a significant correlation between anytime-anyplace connectivity and user engagement, which seems to contradict previous research findings.⁴⁸ One explanation is that in China’s current digital landscape, ubiquitous connectivity is largely taken for granted, diminishing its novelty and motivational force. In this context, connectivity may function more as a hygiene factor, necessary but insufficient to drive engagement. Another explanation is that our sample’s familiarity with always-on mobile technology likely reduced the perceived salience of connectivity. Future research could employ longitudinal or experimental designs to better disentangle the boundary conditions under which connectivity influences engagement, particularly among less tech-savvy user groups or in contexts with intermittent access.

Second, we found that playfulness emerged as the most powerful driver of user engagement. This finding aligns with the arguments presented by Holdack, Lurie³⁶ and Mclean and Wilson,⁸¹ who suggest that the playfulness affordance derived from the immersive augmentation of AR brings user enjoyment and intrinsic motivation for frequent and prolonged exercise. This finding supports prior research by demonstrating that enjoyment-driven design is not merely supplementary but central to engagement in fitness contexts.

Third, our investigation reveals that competition affordance has a significant positive impact on user engagement. This finding aligns with previous research on online gamification-based exercise,³⁷ which suggests that incorporating competitive elements into mobile fitness apps can significantly enhance user motivation and participation. The greater the competition affordance individuals perceive, the more likely they are to engage actively with AR-enabled exercise activities.

Finally, our findings suggest a positive association between user engagement and subjective well-being. This finding aligns with previous literature indicating that higher levels of engagement in various activities, such as exercise and digital interactions, are associated with improved emotional states and overall life satisfaction.^7,8 Furthermore, user engagement positively correlated with continuance intention, aligning with existing studies^21,23,26 that user engagement serves as a critical mechanism that translates technological capabilities into sustained usage intentions.

6.2. Theoretical implications

This study makes several contributions to the existing literature. First, this study broadens the scope of AR research and advances the fields of digital health and sports. While previous studies have primarily focused on the influence of AR technology in areas such as tourism, shopping, and education,^14,15 this research demonstrates how AR exercise apps promote health-related outcomes through immersive, technology-mediated physical activity. By establishing a direct link between AR exercise app affordances and user engagement, subjective well-being, and long-term exercise adherence, this study extends the current understanding of AR technology within digital health ecosystems.

Second, this study extends affordance theory by applying it to the emerging context of AR exercise apps. Affordance theory is particularly well-suited to capturing the complex relational structure between a fitness app and its users, which shapes potential outcomes in specific contexts.^17,31 While previous research has applied the theory to digital platforms such as social media and live commerce,^28,82 this study advances the theory by identifying and validating a set of context-specific affordances that are salient in AR exercise apps. Based on the core characteristics of AR exercise apps and users’ exercise motivation, we identify five affordances: environmental embedding, anytime-anyplace connectivity, visibility, playfulness, and competition, which reflect the utilitarian, social, and hedonic dimensions of affordances. By integrating affordance theory into the domain of AR exercise apps, this study establishes a theoretical link between context-specific affordance and health-related outcomes, showing how AR-enabled possibilities for action can promote subjective well-being and behavior change in digital health contexts.

Third, this study uncovers the mediating mechanism of user engagement within the context of AR exercise apps, demonstrating that technological affordances generate a deeper, more immersive form of engagement, thereby enhancing subjective well-being and continuance intention. Unlike traditional fitness apps that might support motivation through instruction or tracking,²² AR exercise apps' unique value lies in their ability to seamlessly merge digital challenges with the physical environment, thereby creating a heightened state of focused attention and enjoyment. These findings reveal that the advantage of AR in exercise contexts stems not merely from superior visual presentation, but from a fundamentally more effective psychological pathway for sustaining meaningful engagement and promoting lasting behavior change. By investigating the antecedents and consequences of user engagement in the context of AR exercise apps, we extend the existing literature on technology-mediated engagement.^63,83

6.3. Practical implications

This study offers several practical implications for designers, industry practitioners, and public health stakeholders. First, for AR exercise app designers, it is essential to optimize utilitarian, hedonic, and social affordances. To enhance environmental embedding, designers could incorporate realistic 3D models, develop intuitive interfaces that simplify interaction with AR elements, and enable integration with wearable devices. Visibility can be strengthened via customizable AR overlays that display real-time feedback, such as heart rate, repetition count, or form guidance, while allowing users to adjust the size, position, and transparency of these elements to reduce visual clutter and cognitive load. To strengthen hedonic affordances, designers are encouraged to integrate playful features that increase enjoyment and motivation. Examples include AR mini-games that align with physical movements, themed workouts based on engaging narratives such as adventure or fantasy scenarios, and unique AR-specific challenges that leverage spatial interaction. For competition affordance, it is recommended that designers offer diverse competition modes to cultivate a more dynamic and interactive user experience. Designers may introduce reward-based mini-games tailored for beginners, skill-driven challenges for advanced users, and asynchronous competitions that allow users to compete with friends over time.

Second, for fitness industry practitioners, they could embed AR features into comprehensive fitness ecosystems. For example, fitness equipment manufacturers can collaborate with app developers to enable seamless AR integration on smart treadmills or mirrors, allowing users to visualize workout data or virtual coaches directly in their training environment. Similarly, gym operators can deploy AR-enhanced group classes, such as immersive dance or martial arts sessions to differentiate their offerings and attract younger, tech-savvy members. In addition, from a business perspective, introducing AR features as pilot add-ons to existing apps can help assess user engagement and retention benefits before committing to large-scale development, providing a clearer basis for evaluating return on investment.

Third, from a public health standpoint, AR technologies offer a promising avenue to increase population-level engagement in physical activity. Policymakers are encouraged to collaborate with industry stakeholders in integrating AR into digital health and wellness strategies. For example, policymakers can establish safety standards and data privacy regulations to ensure user safety and foster trust in AR-based digital health solutions. Policymakers can also partner with app developers and local governments to co-design public health initiatives, such as location-based AR fitness trails in parks and urban recreational areas, and launch AR scavenger hunts that encourage walking or cycling.

6.4. Limitations and future research

Our study also has some limitations. First, the study sample is limited to a single cultural background, specifically the Chinese context. Cultural factors, such as attitudes towards competition and playfulness, may significantly influence how AR affordances are perceived and valued. Future research could enhance the generalizability of the findings by including samples from different cultural backgrounds.

Second, the questionnaire data used in this study are cross-sectional, which constrains causal inference and precludes examination of long-term engagement dynamics. The cross-sectional data cannot establish temporal precedence, leaving open the possibility of reverse causality (e.g., whether higher well-being leads to more positive affordance perceptions) or the influence of unobserved variables. More critically, this static data cannot address the crucial question of sustained engagement. A key theoretical challenge is understanding how playfulness can be maintained over time and how habituation effects can be mitigated. It is plausible that the hedonic affordances found to be powerful drivers of initial engagement may see their effect diminish with repeated exposure as the novelty wears off. Future research could adopt longitudinal or experimental designs to verify the proposed causal pathways and to investigate how dynamic design elements, such as adaptive challenges, evolving content, or user-generated features, might be necessary to combat habituation and sustain the psychological appeal of AR fitness apps in the long run.

Third, our study relied on self-reported survey data. Some bias (e.g., social desirability bias and acquiescence bias) might exist. Additionally, the use of continuance intention as a proxy for long-term actual behavior inherently encounters the intention-behavior gap. Therefore, future research could incorporate multi-method approaches, such as integrating in-depth interviews to explore subjective experiences, or tracking objective behavioral data through longitudinal field studies or application usage logs, to provide a more comprehensive and robust understanding of user behavior in AR exercise contexts.

Fourth, this study did not theorize about or empirically examine how individual differences might systematically moderate the relationships in our model. These differences include demographic factors (e.g., age, gender) and personality traits (e.g., competitive vs. collaborative orientation), and users' prior experience with AR technology. For instance, it is plausible that the effect of hedonic affordances on engagement is stronger for younger users, or that the importance of utilitarian affordances varies by gender. Similarly, competitive affordances may enhance engagement for users with a competitive orientation, while potentially alienating those with a more collaborative or avoidant motivation. Furthermore, the lack of a defined minimum user experience level means we could not precisely isolate the effects of the affordances themselves from the influence of prior familiarity. Further research could employ methods such as multi-group analysis to investigate how user characteristics shape the psychological mechanisms of AR exercise apps, thereby enhancing the personalization and targeted design of digital health strategies.

Finally, this study primarily examined the independent effects of the five affordances on user engagement, leaving their potential interactions and broader influences unexplored. In practice, these affordances may operate synergistically (e.g., where competitive features enhance playfulness) or even exhibit trade-offs (e.g., when overemphasis on utilitarian functionality diminishes hedonic enjoyment). Moreover, our model does not account for other theoretically relevant factors, such as health consciousness,¹¹ goal setting,⁸⁴ and users’ familiarity with AR.⁸⁵ These variables may act as antecedents, moderators, or mediators in the engagement process. Future research could therefore adopt a more integrative framework to simultaneously investigate the dynamic interplay between affordances and the broader role of individual and contextual variables, ultimately advancing a more nuanced and contextualized understanding of the underlying mechanisms in AR fitness apps.

7. Conclusion

Drawing upon affordance theory, our study delves into the mechanisms underlying how three distinct dimensions of technology affordances influence health-related behaviors and outcomes within the context of AR exercise apps. Our findings validate a research model that positions multidimensional user engagement as the central psychological pathway linking AR exercise app affordances to health outcomes. The empirical results confirm that utilitarian affordances (environmental embedding, visibility), hedonic affordance (playfulness), and social affordance (competition) are significant drivers of user engagement, which in turn influence both continuance intention and subjective well-being. Theoretically, this study extends affordance theory by contextualizing and validating it within the AR fitness domain, providing a nuanced explanation of how AR exercise apps create value through psychological mechanisms. Practically, it calls for a strategic shift in app development from novelty-driven approaches to integrated designs that balance contextual utility, intrinsic enjoyment, and social motivation. For public health policymakers, AR fitness emerges as a viable tool for population health, warranting initiatives such as location-based AR trails in public spaces and clear data privacy standards to harness its societal potential.

Supplemental material

Supplemental material - Enhancing fitness and health outcomes through augmented reality exercise apps: The role of technology affordances and user engagement

Supplemental material for Enhancing fitness and health outcomes through augmented reality exercise apps: The role of technology affordances and user engagement by Yilong Chen, Yanhong Chen and Yaner Yang in Digital Health

Footnotes

ORCID iDs

Yilong Chen

Yaner Yang

Consent to participate

Informed consent was obtained from all subjects involved in the study.

Author contributions

Conceptualization, YL.C. and YH.C.; methodology, YH.C. and YE.Y.; formal analysis, YH.C.; investigation, YL.C. and YH.C.; software, YH.C.; writing—original draft preparation, YL.C. and YH.C.; writing—review and editing, YL.C., YH.C. and YE.Y.; project administration, YL.C.; funding acquisition, YL.C. All authors have read and agreed to the published version of the manuscript.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was funded by the MOE (Ministry of Education in China) Project of Humanities and Social Sciences (24YJC890006).

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data Availability Statement

The raw data supporting the conclusions of this article are available upon reasonable request from the corresponding author.*

Institutional review board statement

The authors obtained informed consent from all participants, and the study did not involve any personal or private information. The study was conducted in accordance with the Declaration of Helsinki, and the protocol was approved by the Ethics Committee of Jinan University on 10 July 2024.

Supplemental material

Supplemental material for this article is available online.

References

Lin

Huang

. Shopping in the digital world: How augmented reality mobile applications trigger customer engagement. Technology in Society 2024; 77: 102540. https://doi.org/10.1016/j.techsoc.2024.102540

Dong

, et al. Shopping through mobile augmented reality: The impacts of AR embedding and embodiment attributes on consumer-based brand equity. Information & Management 2024; 61(6): 103999. https://doi.org/10.1016/j.im.2024.103999

Rauschnabel

. Augmented reality is eating the real-world! The substitution of physical products by holograms. International Journal of Information Management 2021; 57: 102279. https://doi.org/10.1016/j.ijinfomgt.2020.102279

Statista . Augmented reality (AR) - statistics & facts. https://www.statista.com/topics/3286/augmented-reality-ar/#topicOverview2024

Ruth

Willwacher

Korn

. Acceptance of digital sports: a study showing the rising acceptance of digital health activities due to the SARS-CoV-19 pandemic. Int J Environ Res Public Health 2022; 19(1): 596. https://doi.org/10.3390/ijerph19010596

QuestMobile . 2025 Sports and Health Industry Report. https://www.questmobile.com.cn/research/report/19601797946174054412025

Sun

Yuan

. Augmented reality fitness games and subjective well-being of players: a mixed methods study from China. Behaviour & Information Technology 2023; 44: 1–17. https://doi.org/10.1080/0144929x.2023.2279664

Dorian

Calvo

Ryan

. Designing for Motivation, Engagement and Wellbeing in Digital Experience. Frontiers in Psychology 2018; 9: 797.

Fei

Zhou

, et al. Continuance intention and digital health resources from the perspective of elaboration likelihood model and DART model: a structural equation modeling analysis. Frontiers in Public Health 2024; 12: 1416750. https://doi.org/10.3389/fpubh.2024.1416750

10.

James

Belanger

Lowry

, et al. The Mediating Role of Fitness Technology Enablement of Psychological Need Satisfaction and Frustration on the Relationship between Goals for Fitness Technology Use and Use Outcomes. Journal of the Association of Information Systems 2022; 23(4): 913–965. https://doi.org/10.17705/1jais.00745

11.

Bellary

Bala

Chakraborty

. Exploring cognitive-behavioral drivers impacting consumer continuance intention of fitness apps using a hybrid approach of text mining, SEM, and ANN. Journal of Retailing and Consumer Services 2024; 81: 104045. https://doi.org/10.1016/j.jretconser.2024.104045

12.

Veronika

Lee

Lang

, et al. A systematic review and future agenda on continuance intentions in mobile apps. International Journal of Information Management Data Insights 2025; 5(2): 100352. https://doi.org/10.1016/j.jjimei.2025.100352

13.

Sun

Yuan

. A virtual gym in your pocket: the influence of augmented reality exercise app characteristics on user’s continuance intention. Virtual Reality 2024; 28(1): 54. https://doi.org/10.1007/s10055-024-00959-4

14.

Recalde

Jai

Jones

. I can find the right product with AR! The mediation effects of shopper engagement on intent to purchase beauty products. Journal of Retailing and Consumer Services 2024; 78: 103764. https://doi.org/10.1016/j.jretconser.2024.103764

15.

Drljević

Botički

Wong

L-H

. Investigating the different facets of student engagement during augmented reality use in primary school. British Journal of Educational Technology 2022; 53(5): 1361–1388. https://doi.org/10.1111/bjet.13197

16.

Gibson

. The Ecological Approach to Visual Perception. Houghton Mifflin, 1979, pp. 98–99.

17.

Fang

Y-H

. An app a day keeps a customer connected: Explicating loyalty to brands and branded applications through the lens of affordance and service-dominant logic. Information & Management 2019; 56(3): 377–391. https://doi.org/10.1016/j.im.2018.07.011

18.

Jeon

Kim

. Effects of Augmented-Reality-Based Exercise on Muscle Parameters, Physical Performance, and Exercise Self-Efficacy for Older Adults. International Journal of Environmental Research and Public Health 2020; 17(9): 3260. https://doi.org/10.3390/ijerph17093260

19.

Y-L

, et al. Effectiveness of virtual and augmented reality-enhanced exercise on physical activity, psychological outcomes, and physical performance: A systematic review and meta-analysis of randomized controlled trials. Computers in Human Behavior 2019; 99: 278–291. https://doi.org/10.1016/j.chb.2019.05.026

20.

Sun

Shao

, et al. How live streaming influences purchase intentions in social commerce: An IT affordance perspective. Electronic Commerce Research Applications 2019; 37: 100886. https://doi.org/10.1016/j.elerap.2019.100886

21.

Zhang

Shao

Benitez

, et al. How to improve user engagement and retention in mobile payment: a gamification affordance perspective. Decision Support Systems 2023; 168: 113941. https://doi.org/10.1016/j.dss.2023.113941

22.

Zhou

Krishnan

Dincelli

. Examining user engagement and use of fitness tracking technology through the lens of technology affordances. Behaviour & Information Technology 2022; 41(9): 2018–2033. https://doi.org/10.1080/0144929x.2021.1915383

23.

Yang

. The affordances of AI‐enabled automatic scoring applications on learners’ continuous learning intention: An empirical study in China. British Journal of Educational Technology 2020; 51(5): 1674–1692. https://doi.org/10.1111/bjet.12995

24.

Bitrián

Buil

Catalán

. Enhancing user engagement: The role of gamification in mobile apps. Journal of Business Research 2021; 132: 170–185. https://doi.org/10.1016/j.jbusres.2021.04.028

25.

Lynne

Silver

. A Foundation for the Study of IT Effects: A New Look at DeSanctis and Poole's Concepts of Structural Features and Spirit. Journal of the Association for Information Systems 2008; 9(10): 1–25.

26.

Yan

Chen

Shao

, et al. How IT affordances influence customer engagement in live streaming commerce? A dual-stage analysis of PLS-SEM and fsQCA. Journal of Retailing and Consumer Services 2023; 74: 103390. https://doi.org/10.1016/j.jretconser.2023.103390

27.

Zhou

Scheibenzuber

, et al. Fake news land? Exploring the impact of social media affordances on user behavioral responses: A mixed-methods research. Computers in Human Behavior 2023; 148: 107889.

28.

Wang

Yan

, et al. How to enhance consumers’ purchase intention in live commerce? An affordance perspective and the moderating role of age. Electronic Commerce Research Applications 2024; 67: 101438. https://doi.org/10.1016/j.elerap.2024.101438

29.

Evans

Pearce

Vitak

, et al. Explicating Affordances: A Conceptual Framework for Understanding Affordances in Communication Research. Journal of Computer‐mediated Communication 2017; 22(1): 35–52. https://doi.org/10.1111/jcc4.12180

30.

Yoo

Park

. How can I trust you if you’re fake? Understanding human-like virtual influencer credibility and the role of textual social cues. Journal of Research in Interactive Marketing 2025; 19(4): 730–748. https://doi.org/10.1108/jrim-02-2024-0092

31.

Shin

. Does augmented reality augment user affordance? The effect of technological characteristics on game behaviour. Behaviour & Information Technology 2022; 41(11): 2373–2389. https://doi.org/10.1080/0144929x.2021.1928286

32.

Soulé

Marchant

Verchère

. Sport and fitness app uses: a review of humanities and social science perspectives. European Journal for Sport Society 2022; 19(2): 170–189. https://doi.org/10.1080/16138171.2021.1918896

33.

Hilken

De Ruyter

Chylinski

, et al. Augmenting the eye of the beholder: exploring the strategic potential of augmented reality to enhance online service experiences. Journal of the Academy of Marketing Science 2017; 45: 884–905. https://doi.org/10.1007/s11747-017-0541-x

34.

Lee

K-W

C-Y

. It is not merely a chat: Transforming chatbot affordances into dual identification and loyalty. Journal of Retailing and Consumer Services 2023; 74: 103447.

35.

Zhou

Lin

Mou

. Unpacking the effect of gamified virtual CSR cocreated on users’ pro-environmental behavior: A holistic view of gamification affordance. Technology in Society 2023; 73: 102228. https://doi.org/10.1016/j.techsoc.2023.102228

36.

Holdack

Lurie-Stoyanov

Fromme

. The role of perceived enjoyment and perceived informativeness in assessing the acceptance of AR wearables. Journal of Retailing and Consumer Services 2020; 65: 102259.

37.

Yang

Goh

Teo

, et al. Compete with Me? The Impact of Online Gamified Competition on Exercise Behavior. Journal of the Association of Information Systems 2023; 24(3): 912–935. https://doi.org/10.17705/1jais.00806

38.

O’Brien

Cairns

Hall

. A practical approach to measuring user engagement with the refined user engagement scale (UES) and new UES short form. International Journal of Human-Computer Studies 2018; 112: 28–39. https://doi.org/10.1016/j.ijhcs.2018.01.004

39.

Khan

. Customer engagement outcomes in mobile applications: Self-congruence as a moderator. Journal of Retailing and Consumer Services 2023; 75: 103497. https://doi.org/10.1016/j.jretconser.2023.103497

40.

Rauschnabel

Felix

Heller

, et al. The 4C framework: Towards a holistic understanding of consumer engagement with augmented reality. Computers in Human Behavior 2024; 154: 108105. https://doi.org/10.1016/j.chb.2023.108105

41.

Scholz

Smith

. Augmented reality: Designing immersive experiences that maximize consumer engagement. Business Horizons 2016; 59: 149–161.

42.

Jessen

Hilken

Chylinski

, et al. The playground effect: How augmented reality drives creative customer engagement. Journal of Business Research 2020; 116: 85–98. https://doi.org/10.1016/j.jbusres.2020.05.002

43.

Huang

T-L

. Restorative experiences and online tourists’ willingness to pay a price premium in an augmented reality environment. Journal of Retailing and Consumer Services 2021; 58: 102256. https://doi.org/10.1016/j.jretconser.2020.102256

44.

Brodie

Ilic

Juric

, et al. Consumer engagement in a virtual brand community: An exploratory analysis. Journal of Business Research 2013; 66(1): 105–114. https://doi.org/10.1016/j.jbusres.2011.07.029

45.

Cheung

Pires

Rosenberger

, et al. The role of consumer-consumer interaction and consumer-brand interaction in driving consumer-brand engagement and behavioral intentions. Journal of Retailing and Consumer Services 2021; 61: 102574. https://doi.org/10.1016/j.jretconser.2021.102574

46.

Nguyen

Cheng

JMS

, et al. Customer engagement in the context of retail mobile apps: A contingency model integrating spatial presence experience and its drivers. Journal of Retailing and Consumer Services 2022; 66: 102950. https://doi.org/10.1016/j.jretconser.2022.102950

47.

Song

Baek

Choo

. Try-on experience with augmented reality comforts your decision: Focusing on the roles of immersion and psychological ownership. Information Technology & People 2019; 33(4): 1214–1234. https://doi.org/10.1108/itp-02-2019-0092

48.

Alalwan

Algharabat

Baabdullah

, et al. Examining the impact of mobile interactivity on customer engagement in the context of mobile shopping. Journal of Enterprise Information Management 2020; 33(3): 627–653. https://doi.org/10.1108/jeim-07-2019-0194

49.

C-Y

Fang

Y-H

Chiang

Y-H

. Can AI chatbots help retain customers? An integrative perspective using affordance theory and service-domain logic. Technological Forecasting and Social Change 2023; 197: 122921. https://doi.org/10.1016/j.techfore.2023.122921

50.

Suh

Wagner

. How gamification of an enterprise collaboration system increases knowledge contribution: an affordance approach. Journal of Knowledge Management 2017; 21(2): 416–431. https://doi.org/10.1108/jkm-10-2016-0429

51.

Rogers

Strudler

Decker

, et al. Can augmented-reality technology augment the fan experience? A model of enjoyment for sports spectators. Journal of Sports Media 2017; 12(2): 25–44. https://doi.org/10.1353/jsm.2017.0009

52.

Rehman

Abbasi

Ting

, et al. Exploring the Impact of Gamified Experiences on User Engagement in Fitness Apps: A GAMEFULQUEST Perspective. IEEE Transactions on Engineering Management 2023; 71: 3613–3628.

53.

Koivisto

Hamari

. The rise of motivational information systems: A review of gamification research. International Journal of Information Management 2019; 45: 191–210. https://doi.org/10.1016/j.ijinfomgt.2018.10.013

54.

Dacko

. Enabling smart retail settings via mobile augmented reality shopping apps. Technological Forecasting and Social Change 2017; 124: 243–256.

55.

Ryan

Deci

. On Happiness and Human Potentials: A Review of Research on Hedonic and Eudaimonic Well-Being. Annual Review of Psychology 2001; 52(1): 141–166. https://doi.org/10.1146/annurev.psych.52.1.141

56.

Cai

. Exercise or lie down? The impact of fitness app use on users' wellbeing. Frontiers in Public Health 2024; 11: 1281323. https://doi.org/10.3389/fpubh.2023.1281323

57.

Iii

Seo

Jordan

, et al. Exploring the impact of social networking sites on running involvement, running behavior, and social life satisfaction. Sport Management Review 2015; 18(2): 182–192. https://doi.org/10.1016/j.smr.2014.02.006

58.

Zhang

, et al. Examining the impacts of fitness app features on user well-being. Information & Management 2023; 60(5): 103796. https://doi.org/10.1016/j.im.2023.103796

59.

Pang

Ruan

. Disentangling composite influences of social connectivity and system interactivity on continuance intention in mobile short video applications: The pivotal moderation of user-perceived benefits. Journal of Retailing and Consumer Services 2024; 80: 103923. https://doi.org/10.1016/j.jretconser.2024.103923

60.

Ajzen

. The theory of planned behavior. Organizational behavior and human decision processes 1991; 50(2): 179–211. https://doi.org/10.1016/0749-5978(91)90020-t

61.

Shao

Chen

. Understanding individuals’ engagement and continuance intention of MOOCs: the effect of interactivity and the role of gender. Internet Research 2021; 31(4): 1262–1289. https://doi.org/10.1108/intr-10-2019-0416

62.

Elsotouhy

Ghonim

Alasker

, et al. Investigating Health and Fitness App Users' Stickiness, WOM, and Continuance Intention Using S-O-R Model: The Moderating Role of Health Consciousness. International journal of human-computer interaction 2024; 40: 1–16.

63.

Dessart

Veloutsou

Morgan-Thomas

. Capturing consumer engagement: duality, dimensionality and measurement. Journal of Marketing Management 2016; 32(5–6): 399–426. https://doi.org/10.1080/0267257x.2015.1130738

64.

Liang

Xin

Xiaoyan

, et al. Bike sharing and users’ subjective well-being: An empirical study in China. Transportation Research Part A: Policy Practice 2018; 118: 14–24.

65.

Brislin

. Back-translation for cross-cultural research. Journal of cross-cultural psychology 1970; 1(3): 185–216. https://doi.org/10.1177/135910457000100301

66.

Zhang

, et al. Fostering well-being: Exploring the influence of user-AI assistant relationship types on subjective well-being. International Journal of Information Management 2024; 79: 102822. https://doi.org/10.1016/j.ijinfomgt.2024.102822

67.

Durau

Diehl

Terlutter

. Motivate me to exercise with you: The effects of social media fitness influencers on users’ intentions to engage in physical activity and the role of user gender. Digital Health 2022; 8: 20552076221102769. https://doi.org/10.1177/20552076221102769

68.

Angosto

Garcia-Fernandez

Grimaldi-Puyana

. A systematic review of intention to use fitness apps (2020–2023). Humanities and Social Sciences Communications 2023; 10(1): 512. https://doi.org/10.1057/s41599-023-02011-3

69.

Chen

Carl

Hinz

. Continued usage of mobile fitness applications: a systematic literature review. Management Review Quarterly 2025; 1–52. https://doi.org/10.1007/s11301-025-00537-1

70.

Zhu

Wang

Zeng

, et al. Does gender really matter? Exploring determinants behind consumers' intention to use contactless fitness services during the COVID-19 pandemic: a focus on health and fitness apps. Internet Research 2023; 33(1): 280–307. https://doi.org/10.1108/intr-07-2021-0454

71.

Chin

Marcolin

Newsted

. A Partial Least Squares Latent Variable Modeling Approach for Measuring Interaction Effects: Results from a Monte Carlo Simulation Study and an Electronic-Mail Emotion/Adoption Study. Information Systems Research 2003; 14(2): 189–217. https://doi.org/10.1287/isre.14.2.189.16018

72.

Hair

Ringle

Sarstedt

. PLS-SEM: Indeed a Silver Bullet. Journal of Marketing Theory and Practice 2011; 19(2): 139–152. https://doi.org/10.2753/mtp1069-6679190202

73.

Podsakoff

MacKenzie

Podsakoff

. Sources of Method Bias in Social Science Research and Recommendations on how to Control it. Annual Review of Psychology 2012; 63: 539–569. https://doi.org/10.1146/annurev-psych-120710-100452

74.

Podsakoff

Mackenzie

Lee

, et al. Common method biases in behavioral research: a critical review of the literature and recommended remedies. Journal of Applied Psychology 2003; 88(5): 879–903. https://doi.org/10.1037/0021-9010.88.5.879

75.

Kock

. Common Method Bias in PLS-SEM: A Full Collinearity Assessment Approach. International Journal of e-Collaboration (IJeC) 2015; 11: 1–10.

76.

Lindell

Whitney

. Accounting for common method variance in cross-sectional research designs. J Appl Psychol 2001; 86(1): 114–121. https://doi.org/10.1037/0021-9010.86.1.114

77.

Henseler

Ringle

Sarstedt

. A new criterion for assessing discriminant validity in variance-based structural equation modeling. Journal of the Academy of Marketing Science 2015; 43(1): 115–135. https://doi.org/10.1007/s11747-014-0403-8

78.

Hair

. A primer on partial least squares structural equation modeling (PLS-SEM). Sage, 2014.

79.

Chen

Zhu

. Being there and being with them: the effects of visibility affordance of online short fitness video on users’ intention to cloud fitness. Frontiers in Psychology 2024; 15: 1267502. https://doi.org/10.3389/fpsyg.2024.1267502

80.

Akram

, et al. Understanding how customer social capital accumulation in brand communities: A gamification affordance perspective. Journal of Retailing and Consumer Services 2024; 78: 103761. https://doi.org/10.1016/j.jretconser.2024.103761

81.

Mclean

Wilson

. Shopping in the digital world: Examining customer engagement through augmented reality mobile applications. Computers in Human Behavior 2019; 101: 210–224. https://doi.org/10.1016/j.chb.2019.07.002

82.

Jia

Zhao

Song

, et al. How vicarious learning increases users’ knowledge adoption in live streaming: The roles of parasocial interaction, social media affordances, and knowledge consensus. Information Processing & Management 2024; 61(2): 103599. https://doi.org/10.1016/j.ipm.2023.103599

83.

Liu

Yang

Nie

. Research on the influence of online fitness user engagement on value co-creation behavior. Asia Pacific Journal of Marketing and Logistics 2023; 35(12): 3138–3155. https://doi.org/10.1108/apjml-12-2022-1061

84.

Lee

Lin

. The continuous usage of artificial intelligence (AI)-powered mobile fitness applications: the goal-setting theory perspective. Industrial management & data systems 2023; 123: 1840–1860. https://doi.org/10.1108/imds-10-2022-0602

85.

Bonnin

. The roles of perceived risk, attractiveness of the online store and familiarity with AR in the influence of AR on patronage intention. Journal of Retailing and Consumer Services 2020; 52(Jan.): 101938. https://doi.org/10.1016/j.jretconser.2019.101938