The demand for health and exercise among individuals is steadily increasing, and the number of users of fitness apps is also growing annually. Therefore, exploring the factors that influence people’s continuous use of fitness apps has become a significant issue for scholars in this field. Existing literature primarily investigates factors affecting usage attitudes, usage motivations, participation intentions, and continuous use intentions. For instance, Song et al. (7) conducted research based on affordance theory, revealing that social presence, immersion, and credibility positively influenced users’ continuous use intentions of fitness apps. Durau et al. (8) discovered that fitness attitudes and motivation impacted participants’ fitness intentions. Fitness apps offer users fitness videos and online communication services, enabling them to obtain fitness guidance, engage in fitness activities, post fitness updates, and like and comment on other users’ fitness updates. The process involves social interaction between users and the app, as well as among users. During this process, individuals receive not only information support but also emotional support. For example, when individuals receive positive comments from others, they will experience a sense of joy, which subsequently enhances their fitness motivation. Furthermore, through the social features of fitness apps, individuals can view their fitness rankings in comparison to others, leading to social comparison, which fosters a competitive environment that motivates people to engage in fitness exercises. Thus, it is evident that the social features of fitness apps play a crucial role in encouraging individuals to use the app more actively and participate in fitness activities. However, there has been limited research conducted from the perspective of social features. Thus, this paper aims to explore whether the two social features of fitness apps, including social support and social comparison, can influence individuals’ continuous use intentions of fitness apps.

Social support refers to the experience of individuals receiving respect, care, and assistance from others within their social networks. The traditional concept of social support primarily encompasses instrumental support, informational support, and emotional support (9). Social support theory is frequently applied in research on online user communities, occupational burnout, entrepreneurship, and other areas. For example, van Brakel et al. (10) studied the impact of social support on users of social VR platforms. Le et al. (11) based on the stimulus-organism-response (SOR) model, elucidated the mechanisms behind privacy disclosure behavior from the perspective of social support. In the context of fitness apps, Sun and Jiang (12) categorized social support into informational support, self-esteem support, and companionship support, with findings indicating that social support motivated users’ physical exercise through fitness apps. Huang et al. (13) also classified social support into three categories and examined its influence on user participation in physical exercise, concluding that social support could not explain participation in physical exercise. In fitness apps, social support is realized through interactions with the app interface and other users. For instance, users continuously receive informational support regarding exercises through fitness apps. Fitness apps also provide companionship and emotional support by displaying friends’ physical activities. Users’ fitness updates receive likes and comments from others, thereby fostering a sense of support. Research has found that users of fitness apps engage in exercise to receive praise from others (14), indicating that social support is an effective motivational function that fitness apps possess for users.

Social comparison theory was proposed by Festinger (15). The theory posits that individuals possess an intrinsic motivation to evaluate their self-concept and abilities. Driven by this motivation, individuals seek information for self-assessment. However, in the absence of objective and scientific evaluation methods, individuals resort to comparing their abilities with others for self-evaluation. Social comparison typically includes upward and downward comparisons, where individuals compare themselves with others who are in better or worse conditions across multiple dimensions such as abilities and achievements (16). The theory is frequently applied in research on social websites and online social games. For instance, Burke and Rains (17) studied the relationship between individuals’ exposure to others’ exercise posts on social websites and their concerns about weight and attitudes toward physical activity. Kashian and Liu (18) found fitness posts sharing successful experience on social media were positively correlated with self-efficacy. Esteves et al. (19) explored the impact of social comparison on players’ continuous gaming intentions. Huang et al. (13) discovered that upward comparison was positively correlated with participation in physical activities, while downward comparison was negatively correlated with it. However, some studies suggest that social comparison with ideal images of others can lower one’s self-evaluation of their appearance, leading to increased dissatisfaction with their body (20, 21). For example, Arroyo and Brunner (22) found that browsing friends’ fitness posts was positively correlated with negative body talk. Fitness apps typically record individuals’ exercise data and performance, ranking their athletic achievements against friends. Features such as performance rankings create a psychological environment for social comparison, aiding users in assessing their fitness levels accurately. As one of the unique social features of fitness apps, social comparison is significant and should be studied as a behavior change mechanism within the fitness environment (23). In the process of using fitness apps, whether upward comparison and downward comparison have a positive or negative impact on users also requires further research.

The Stimulus-organism-response model originates from the field of environmental psychology, where S represents various stimuli from the external environment. External stimuli (S) affect the internal states (O) of the organism, which in turn drives the subsequent behavioral response (R) (24). The SOR model is primarily used to study the internal state responses of individuals to various external environmental stimuli. And the behavioral responses resulting from changes in internal state responses. The organism refers to cognitive and emotional responses. Cognitive responses indicate the psychological processes involved in acquiring, retaining, retrieving, and processing information, while emotional responses reflect the arousal, resonance, and pleasure generated by environmental stimuli (25). Individuals transform environmental stimuli into meaningful information and influence their subsequent specific behavioral responses.

Scholars have utilized the SOR model to investigate individuals’ behavioral responses across various research domains, including impulse buying, online learning, and online social networks (26, 27). In the field of fitness, recent studies have also applied the SOR theory to investigate the impact of short video features and platform characteristics as environmental stimuli on individuals’ internal states. As well as the resulting behavioral responses related to fitness apps. For example, Chen et al. (28) found a positive correlation between the visibility of fitness short videos and the willingness to exercise with fitness influencers. The results from Elsotouhy et al. (29) indicated that the quality of information, system quality, and service quality on fitness platforms significantly impacted user engagement with fitness apps. Teng and Bao (30) discovered that interactions between individuals and information, as well as interactions among individuals, served as environmental stimuli that affected individuals’ internal states, thereby influencing their stickiness to fitness apps. This study suggests that when individuals use fitness apps, they often have certain goal-oriented needs. Users interact with others in the fitness app to obtain information and emotional supports from the app or other users. Thus, the social support can be regarded as an external stimulus. When individuals engage in upward and downward comparisons with other users, social comparison features of fitness apps help users perceive the “real presence” of other users more vividly, facilitating a precise assessment of their own fitness levels. Simultaneously, the heightened sense of competition enhances the enjoyment of fitness, increasing users’ interest in participating in fitness activities, thereby strengthening their continuous use intention of fitness apps (31). Consequently, this study employs the SOR model to explore how the two external stimuli of social support and social comparison within the fitness environment influence individuals’ organism, subsequently affecting their continuous use intention.

Social support theory has been utilized in the study of online and brand communities. For instance, Trepte et al. (32) discovered that online communities provided social support and a sense of competence to their users. Zhao et al. (33) classified social support into information support and emotional support, finding that social support positively affected users’ perceived benefits during online information searches. In this study, information support refers to the recommendations and suggestions provided by fitness apps that assist individuals in engaging in fitness activities more efficiently and safely. When individuals encounter challenges, they can quickly receive informational assistance from fitness apps or other users. Fitness apps integrate a variety of functions including AI coaching, intelligent pacing, dietary analysis, and personalized exercise plans. AI coaching delivers professional exercise demonstrations, interactive Q&A, image recognition, and voice-guided instructions, enhancing training safety and efficiency of users. Dietary analysis function automatically identifies food calories and nutritional components, aiding users in recording and analyzing their diets to achieve weight loss goals. Fitness apps can also generate personalized fitness plans based on user information, daily exercise records, and fitness goals, and monitor the implementation of users. These integrated functions provide users with information support, enhancing their fitness knowledge, improving fitness outcomes, and facilitating the achievement of fitness goals.

Literature indicates that in the context of online shopping, information support can moderate the effect of social presence on trust in online merchants (34). When using fitness apps, users genuinely perceive communication with others when they receive social support from other users, leading to feelings of warmth and closeness. On one hand, fitness apps provide individuals with more professional fitness information and guidance, helping them manage health issues and improve their fitness management levels. On the other hand, through comment interactions, users can view comments from other users, and these suggestions assist them in making better decisions. Particularly when users encounter difficulties and confusion, the advice provided by others helps them overcome challenges, making them feel that they are not alone and thus providing a sense of psychological companionship. It is evident that the information support provided by fitness apps enhances the social presence of users.

Fitness apps offer users a variety of fitness courses, personalized fitness plans, and professional online fitness guidance, all of which help alleviate fitness-related stress and enable users to engage in fitness activities more easily and enjoyably. The extensive social interaction features within the app enhance its appeal, encouraging users to continue using it. Therefore, it is evident that the information support function of fitness apps plays a significant role in enhancing users’ fitness interest. Consequently, we propose the following hypotheses:

Emotional support refers to the emotional and spiritual assistance received from fitness apps, such as care and attention. When individuals feel encouraged and supported, their confidence in interacting within the app is enhanced. Users can engage with other users within the fitness apps to receive emotional support, which helps motivate them to participate in fitness activities more enthusiastically. Previous research has indicated a positive correlation between emotional support and social presence. Emotional support helps reduce dissatisfaction and stress, assists individuals in overcoming challenges, thereby alleviating pressure and enhancing positive emotions (35). For instance, when users feel disappointed and frustrated during fitness process, other users can communicate with and comfort them. When users feel lonely during their fitness activities, other users can provide companionship (36). The emotional support helps users minimize emotional adversity, alter their negative perceptions toward fitness issues, and promote a positive understanding of fitness outcomes. Consequently, users can restore their confidence and positive attitude toward fitness and regain emotional stability when dealing with fitness-related issues. With the emotional support received, users experience feelings of being understood, comforted, and motivated, which in turn enhances their fitness interest. Conversely, when users encounter setbacks during their fitness journey and lack relevant emotional support, their enthusiasm and passion for fitness will decrease. Therefore, we propose the following hypothesis:

Upward comparison refers to comparing oneself with individuals who perform better in a certain task, while downward comparison refers to comparing oneself with individuals who perform worse (37). Research indicates that, regarding fitness behaviors, individuals tend to compare themselves with other users in their social networks, and social comparison is associated with a significant increase in fitness activities (38). In particular, health interventions that utilize social comparison to create a competitive social environment can enhance the effectiveness of fitness activities.

Recently, fitness apps generally track users’ exercise records and rank their fitness performance against friends within the app. Features such as performance rankings enhance the importance of social comparison for fitness app users and should be examined as a behavior change mechanism within the fitness environment. While early research connected mobile fitness features (such as rankings) to the social comparison process, few studies have verified whether upward or downward comparisons are activated in this context. Recent studies have primarily focused on fitness posts shared on social platforms, confirming that social comparison can influence individuals’ exercise attitudes and fitness intentions. For example, Burke and Rains (17) found that both upward and downward comparisons increased concerns about weight when observing others’ exercise-related posts, with only downward comparisons showing a negative association with exercise attitudes. In the study of fitness posts on Instagram, Peng et al. (39) discovered that upward comparisons could motivate individuals to engage in physical exercise through self-enhancement, while no such effect was found for downward comparisons. Based on it, this study argues that the mechanism by which the social comparison features of fitness apps influence individuals’ organism states, thereby affecting fitness behaviors, requires further investigation. And it is essential to distinguish between different types of social comparisons.

In fitness apps, when individuals compare themselves to those who have better performance, they are more likely to regard these individuals as goals and believe that they can attain better fitness achievements in the future, thereby engaging in fitness activities more efficiently. Users are easily motivated, which can lead to positive emotions and an increase in fitness interest. Conversely, when individuals compare themselves to those who perform worse, users may experience feelings of shame and frustration. Engaging in downward comparison can lead to a decrease in fitness interest, as users may worry about deteriorating their performance to the level of the downward target, resulting in negative emotions (13). Therefore, we propose the following hypothesis:

Social presence refers to the degree to which individuals feel the presence of others during communication, reflecting the sense of intimacy or direct experience that individuals develop toward the communication medium in interpersonal interactions (40). Social presence plays a crucial role in users’ psychology and behavior. When using fitness apps, users communicate with others, allowing them to compare their own fitness situations with other users. Through interaction, users can help bridge the social distance between them, enhancing their perception of others’ presence in fitness apps. As users experience a sense of authenticity and warmth from other users, they are more likely to engage in interpersonal interactions, perceive support and trust, thereby leading to continuous use of the fitness apps and an increased fitness intention. Existing literature has confirmed that social presence positively affects consumers’ online purchase intentions in social commerce (41). In related studies on online platforms, scholars have found that social presence not only affects users’ initial trust in the website but also influences their level of affection for the site and perceived usefulness (42). In online live streaming, social presence enhances consumers’ sense of identity, thereby increasing their purchase intentions (40). Another study indicates that social presence has a positive and significant impact on consumers’ trust and flow state, thereby triggering impulsive buying behavior (26). It is evident that social presence significantly influences users’ online purchase intentions, affection, trust, and sense of identity. This study applies the concept of social presence in the fitness context. When users perceive the presence of others, for instance, by interacting with other users through comments, obtaining necessary information, and receiving encouragement from others, their affection and trust for fitness apps will increase, which helps enhance their continuous use intention. Therefore, this study posits that the stronger the social presence perceived by users while using fitness apps, the more likely they are to increase their continuous use intention of fitness apps.

The higher the level of social presence that users experience while using fitness apps, the greater their motivation and intention to engage in fitness activities, such as exercise, diet, and monitoring. Previous studies have confirmed that social presence significantly impacts students’ satisfaction with online learning (43). In research on live streaming, scholars found that social presence positively influenced users’ willingness to watch live broadcasts (44). In this study, when users feel the presence and companionship of other users and receive social support from them, the sense of distance between users diminishes, and users’ intention to continue exercising increases. Therefore, we propose the following hypothesis:

Fitness interest refers to the degree of attention or interest that users have toward the outcomes of fitness behaviors (45). Interest is the first step in guiding users to engage in activities. Therefore, cultivating users’ interest in fitness behaviors is crucial, as this may lead them to adopt a positive attitude toward health behaviors and make better health-related choices. For instance, in this study, fitness apps can provide users with useful information support and positive feedback regarding fitness outcomes, thereby increasing their expectations of behavioral results and maintaining their interest in future fitness activities. Existing research indicates that fitness interest influences fitness behaviors (46, 47), and enhancing exercise interest will increase physical activity among middle school students, fostering a regular exercise lifestyle (48). In studies involving smartwatches, exercise interest is a significant criterion for segmenting smartwatch users’ exercise habits (49). In the fitness context, users’ fitness interest will evoke positive impressions of fitness apps and trigger users’ continuous use intention. Specifically, users’ fitness interest during the use of fitness apps helps to establish a deeper connection between the app and the user, enhancing user satisfaction and stickiness toward the fitness apps, thereby increasing the adoption rate.

From a hedonic perspective, as individuals cultivate an interest in using fitness apps, their enthusiasm and enjoyment for fitness activities also increase, thereby encouraging further participation in fitness (50). Since interest is a strong predictor of behavioral intention (51), users who develop interest, appreciation, and enjoyment for fitness activities while using fitness apps are more likely to engage in actual fitness activities. Because they believe they will derive greater utilitarian achievements, social value, and hedonic value from these activities (52). Therefore, we propose the following hypothesis:

This study employs the questionnaire survey method to collect data. The questionnaire is divided into three parts: the first part includes basic descriptive information about the questionnaire, providing relevant explanations about fitness apps. It guides respondents with the following prompt: “The mobile fitness application referred to in this questionnaire is an application that allows users to engage in fitness training through mobile devices, such as Keep, Joyrun, Mint Health, Xiaomi Sports, and Codoon. Please fill out this questionnaire based on your most recent experience using a fitness app.” The second part surveys the basic information of the respondents, including gender, age, education level and usage of fitness apps. The third part measures eight latent variables related to the social features of fitness apps, users’ organism and continuous use intention. To ensure the reliability and validity of the scales, the measurement items for the variables in this study are referenced from mature scales. Specifically, information support is based on the studies by Oh et al. (53) and Lin et al. (54), emotional support is based on the studies by Lin et al. (54) and Liang et al. (55). Upward and downward comparisons reference Kim’s (56) research and are adapted to fit the context of fitness apps. Social presence is based on the studies by Gao et al. (57) and Chen et al. (40), fitness interest is based on research of Jang et al. (45). Continuous use intention of fitness apps is based on the studies by Hsiao and Chiou (58) and Esteves et al. (19). And fitness intention is based on Bhattacherjee’s (59) research. Specific measurement items are shown in Table 1. Except for the statistical characteristics of the sample, all measurement items are assessed using a 7-point Likert scale, with 1 to 7 representing levers from “strongly disagree” to “strongly agree.” The questionnaire was sent to experts in the field of e-commerce for review. After multiple revisions based on expert feedback, the initial questionnaire was formed. The initial questionnaire was randomly sent online to 50 respondents for a pre-survey, and based on the results of the pre-survey, the content and wording of the items were revised to form the final questionnaire.

The questionnaire was distributed and collected both online and offline using Questionnaire Star. Online distribution was conducted through WeChat, QQ, and email to relatives and friends. As for offline, it was distributed in classrooms by scanning QR codes for university students, and student survey teams were organized in densely populated areas such as shopping malls and cinemas to distribute the questionnaire to passersby via QR code scanning. A total of 500 questionnaires were distributed, with 378 returned. After excluding those that did not use any fitness apps, those with contradictory trap questions, those with completely identical answer options, and those with response times less than 1 min, 343 valid questionnaires remained. The descriptive statistics is presented in Table 2. Among the participants in the survey, 58.60% were male and 41.40% were female; the education level was primarily concentrated in Bachelor’s and Master’s degrees, accounting for 82.22% of the overall sample; the age group was mainly between 18 and 45, comprising 89.50% of the total sample; and 67.93% of the sample had over 1 year of experience using fitness apps.

SPSS 26.0 was utilized to assess the internal reliability of the research variable scales using Cronbach’s α. A Cronbach’s α value greater than 0.7 indicates good reliability, with higher values reflecting greater reliability (60). Typically, a CA coefficient value exceeding 0.6 suggests that the reliability level of the observed variables is acceptable (60). The Cronbach’s α coefficients for each research variable are presented in Table 3, where the coefficients for information support, emotional support, upward comparison, downward comparison, social presence, fitness interest, continuous use intention of fitness apps, and fitness intention all exceed 0.8, indicating that the scale possesses high internal consistency and can reliably measure the latent variables.

To validate the reliability and validity of the measurement model, confirmatory factor analysis was conducted using AMOS 26.0, with the results presented in Table 3. The findings indicate that the standardized factor loadings for each observed variable exceed 0.7. Calculations reveal that the composite reliability (CR) for each factor is greater than 0.8, indicating strong reliability for each factor (61). The average variance extracted (AVE) for latent variables is above 0.6, demonstrating good convergent validity (61). And it is observed that the square roots of AVE are always greater than the absolute correlations between the constructs. AVE ranges from 0.778 to 0.892 (see Table 4). It indicates that all the constructs report acceptable validity (61).

The path coefficients and validation results of the model are presented in Table 5, where hypothesis H4b was not supported, while the other hypotheses were supported.

This paper employs the Bootstrap method to test the mediating effects of social presence and fitness interest. The Bootstrap method is a commonly used resampling technique in mediation analysis, which estimates the indirect effects of mediating variables and the characteristics of the sampling distribution through repeated random sampling, providing confidence intervals for the indirect effects. When the confidence interval does not include 0, the effect is considered significant. Using AMOS 26.0, the Bootstrap random sampling was set to 5,000 iterations, and the mediation effects were tested at a 95% confidence interval. A parallel mediation model was employed for testing, and the results are presented in Table 6. Except for the mediating effects of fitness interest in the downward comparison with continuous use intention of fitness apps and downward comparison with fitness intention, which were not significant, all other mediating effects were significant and indicated partial mediation.

This study employs fitness apps as the research context and constructs a research model based on SOR theory to investigate the impact of social features of fitness apps on users’ continuous use intention and fitness intention. Data were collected through a questionnaire. The empirical findings indicate that the social support and social comparison features of fitness apps exert varying degrees of stimulating effects on users’ social presence and fitness interest, thereby positively influencing users’ continuous use intention and fitness intention.

Due to objective constraints, this study has certain limitations: (1) This paper primarily examines two prominent social features in fitness apps, which presents a limitation. There is a plenty of social features of fitness apps, and future research can further incorporate additional factors to conduct a more systematic investigation into users’ behavioral intentions. (2) The social features of fitness apps can influence users’ organism and responses. The mechanisms underlying these influences, as well as the potential existence of other mediating and moderating variables, can be further explored in future research.