An Empirical Investigation into the Impact of Social Media Fitness Videos on Users’ Exercise Intentions

He Yin, Conceptualization , Methodology , Software , Validation , Investigation , Data curation , Writing – original draft , Writing – review & editing , Xin Huang, Conceptualization , Methodology , Software , Validation , Investigation , Writing – review & editing , and Guangming Zhou, Validation , Investigation , Supervision *

School of Journalism and Communication, Wuhan University, Wuhan 430072, China; nc.ude.uhw@hy_ehniy (H.Y.); nc.ude.uhw@nix-gnauh (X.H.)

Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).

Associated Data

Data are available upon special request from the corresponding author.

Abstract

Social media fitness influencers are driving the maturation of online fitness, which is especially significant in the current era of globally decreasing levels of physical activity. However, there is a paucity of research on online fitness videos, and the mechanism of influence of fitness videos on exercise intention is not well understood. Therefore, based on the stimulus–organism–response (S-O-R) theoretical framework, this study extends the source credibility theory to the field of fitness and adds an examination of the content quality and electronic word-of-mouth of fitness videos to explore how fitness videos motivate users to participate in physical exercise. Through an online survey, 367 valid samples were collected and validated using a structural equation model. The results showed that the three elements contained in source credibility theory have inconsistent importance in the fitness field, with trustworthiness being the most important, followed by attractiveness, and the influence of expertise is not significant. In summary, the attributes of social media fitness influencers, including trustworthiness and physical attractiveness, as well as the content quality and electronic word-of-mouth of their fitness videos, may lead to viewers’ trust and perception of the usefulness of the videos and, furthermore, lead to exercise intention.

Keywords: social media fitness video, social media fitness influencer, exercise intentions, stimulus–organism–response (S-O-R) model, source credibility theory

1. Introduction

Active exercise can help people to prevent and manage non-communicable diseases, with significant benefits for the heart, body, and mind. However, lack of exercise has a negative impact on people’s physical health, community wellbeing, and quality of life, on a global scale [1]. Therefore, in addition to calling for more exercise, it is equally important to find ways to motivate exercise and maintain it [2]. Today, internet technology is more mature and can provide people with the latest and useful information. Social media, in particular, has become a common means for people to seek and communicate health information [3]. At the same time, previous research has found that social media may induce changes in exercise behavior through the content posted [4,5,6].

“The internet is ushering in an era where the role of personal influence will attain unprecedented prominence” [7]. The internet and social media platforms give every user an equal opportunity to express and share, but they also lead to information overload and an uneven quality of information. As a result, social media influencers are increasingly being relied upon. They are often highly followed, have expertise in a specific field, and have the ability to create valuable content and provide useful advice. Thus, similar to traditional media figures, social media influencers are also held up as models of observation and learning who can influence their audience’s intentions and beliefs [8]. In the field of health, a combination of health communicator and influencer strategies are more suitable for the media consumption habits in the mobile internet era, and social media fitness influencers also participate in the construction of health communication patterns in the digital era, as health communicators [2,9]. The phenomenon of following an influencer to exercise is prevalent in today’s social media context, where followers have access to a wide selection of free fitness videos, especially since the COVID-19 pandemic has changed people’s lifestyles. At present, there is not much research on fitness videos and social media fitness influencers.

In fitness-related research, there has been a growing interest in fitness applications and related health behaviors in recent years [10,11], with the focus primarily on the performance expectations, technical characteristics, and user stickiness of applications; however, limited attention has been given to exploring the influence of social media fitness influencers and social media fitness videos on exercise intentions. In the existing studies, the stimulus–organism–response (S-O-R) theoretical framework (this theoretical framework assumes that aspects of the environment work together as a stimulus to influence people’s internal states, thereby inducing individual perceptions that then influence their psychological and behavioral responses) has often been applied to research in the fields of commercial marketing and tourism to explore how stimulus factors affect consumer behavior through the actions of organisms. Few studies have used the S-O-R framework as a basis for health- and fitness-related research. In fact, this framework is also suitable for research in the field of online fitness. Some elements in fitness videos can be regarded as stimulating factors to motivate viewers to exercise. According to the source credibility theory, the trustworthiness (TRU), expertise (EXP), and attractiveness (ATT) of the spokesperson are very important to the persuasiveness of the message [12], and a large number of studies have applied the source credibility structure to the study of social media influencers, confirming the positive effects of these three factors on perceived trust and usefulness [13,14,15]. However, does this influence mechanism also exist in the field of fitness? There is not enough research on this question. In addition to the trainers in fitness videos, the quality of the information and content may also change people’s attitudes and beliefs [16]. Moreover, reviewing previous studies on fitness videos, we found that these studies have neglected the impact of electronic word-of-mouth (eWOM), which is closely related to consumer behavior in marketing practices and has a strong impact on consumers’ purchasing decisions [17]. Hence, one of our main research objectives was to determine whether eWOM also has a significant impact in the field of fitness media information consumption behavior. Perceived trust and perceived usefulness are important determinants of behavioral intention [18,19]. When users trust a fitness video and think that it is useful for their body shape, they may consider using it for exercise. Therefore, this study considered these two factors as organism factors and examined whether they play a mediating effect between fitness videos and exercise intention.

In general, in order to further understand the relationship between fitness videos and users’ exercise intentions, this study empirically investigated which factors in fitness videos drive users’ cognitive changes that lead to their perception of trust and usefulness of the information and, subsequently, to their intention to exercise. Specifically, we constructed a relationship model of social media fitness influencer attributes, content quality (CQ), electronic word-of-mouth (eWOM), information credibility (INC), perceived usefulness (PU), and exercise intention (EXI), based on the S-O-R theoretical framework, to explore the effects of online fitness videos on users’ exercise intentions.

2. Theoretical Background

2.1. The S-O-R Model

The stimulus–organism–response model (S-O-R model), proposed by Mehrabian and Russell [20] in 1974, posits that aspects of the environment work together as a stimulus (S) to influence people’s internal states (O), thereby inducing individual perceptions that then influence their psychological and behavioral responses (R) [21]. The S-O-R model has been widely used in the field of consumer behavior research. For example, a study on the impact of live broadcast marketing on consumers’ purchase intentions considered live streaming features as a stimulating factor that could positively affect consumers’ perceived value and trust and indirectly facilitate consumers’ purchase intentions [22]. Kim and Johnson [23] used the S-O-R model as a theoretical framework to investigate the impact of brand-related UGC on consumer behavior, and they proposed that brand-related UGC as a stimulus would cause emotional and cognitive responses in consumers and then drive a series of behavioral responses, such as information pass-along, impulse buying, future purchase intention, and brand engagement. To clarify how social media users respond to fitness videos, the stimulus–organism–response model (S-O-R model) paradigm was used as the basis for this research model. We regarded the social media fitness influencer attributes, electronic word-of-mouth (eWOM), and content quality in fitness videos as stimuli that affect viewers’ perceptions, took the perceived credibility and usefulness of the information as organic factors, and used exercise intention to measure viewers’ behavioral responses.

2.2. Source Credibility Theory

Source credibility is defined as the extent to which a recipient of information believes that a source is trustworthy, competent, and reliable [24]. Ohanian [12] proposed that source credibility implies positive characteristics of the communicator, which will affect the acceptability of the information to the recipient. Hovland and colleagues [25] constructed the source credibility model and pointed out that two key determinants of source credibility are perceived expertise and trustworthiness. Source credibility theory is widely used in marketing research and is often used to examine the effectiveness of celebrity endorsements [26]. Ohanian [12] constructed scales of expertise, trustworthiness, and physical attractiveness in his research on celebrity endorsements, which established a scientific and effective way to apply this theory. Smith [27] applied source credibility theory to investigate celebrity endorsements in political elections and found that the trustworthiness, expertise, and attractiveness of endorsers were positively correlated with the image of their political parties. Djafarova and Trofimenko [28] focused on the credibility of micro-celebrity endorsements and proposed new dimensions of source credibility applicable to the online context, based on the existing source credibility model’s dimensions. Based on these, we took “expertise, trustworthiness and physical attractiveness” as the dimensions of social media fitness influencer attributes to be examined in this study.

3. Hypothesis Development

3.1. Social Media Fitness Influencers’ Attributes, Information Credibility, and Perceived Usefulness

3.1.1. Social Media Fitness Influencers’ Attributes and Information Credibility

Information credibility is defined as the extent to which information is perceived to be believable [29]. This study considers information credibility as the credibility of the information provided in videos posted by social media fitness influencers.

Erdogan [30], in the study “Celebrity endorsement: a literature review”, described trustworthiness as the honesty, integrity, and believability of an endorser. Previous studies have shown that, on social media platforms, influencers’ trustworthiness is one of the most critical influencing factors on users’ attitudes and behavior [31,32]. Source credibility significantly affects consumers’ perceptions of the credibility of information [13,33,34].

Hovland and his associates [25] defined expertise as “the extent to which a communicator is perceived to be a source of valid assertions”; Ohanian [12] added that expertise is also referred to as “authoritativeness”, “competence”, “expertness”, or “qualification”. Crisci and Kassinove [35] found that having a more credible title, such as “Dr.”, will strengthen the impact of the communicator and their message. A previous study found that when a product has a perceived risk, consumers are more likely to be influenced by information from influencers who are market experts before making a purchase decision [36].

Patzer [37] defined physical attractiveness as the degree to which a communicator’s face and body image are pleasing to observe, and experimentally demonstrated that communicators with higher levels of physical attractiveness were more persuasive and perceived with more trust and as having greater expertise than communicators with lower levels of attractiveness. Information recipients develop positive beliefs toward attractive communicators, and attractiveness will enhance the effectiveness of persuasive information [38,39]. Physical attractiveness has been shown to be positively correlated with information credibility [13,40].

Based on previous studies, we established the following hypotheses to verify the influence of fitness influencers’ attributes on information credibility:

The trustworthiness of social media fitness influencers is positively correlated with information credibility.

The expertise of social media fitness influencers is positively correlated with information credibility.

The physical attractiveness of social media fitness influencers is positively correlated with information credibility.

3.1.2. Social Media Fitness Influencers’ Attributes and Perceived Usefulness

Perceived usefulness refers to the degree to which prospective users expect to increase their performance using a particular system, and it strongly influences people’s intentions and is a major factor in determining whether people use a particular system [19]. Perceived usefulness in this study refers to the extent to which users perceive the content of a fitness video on social media to be useful for their exercise.

Trust in the source of information will affect the recipient’s assessment of the usefulness of the information. A study on how travel vloggers’ YouTube videos influence the future behavior of tourists showed that source credibility has a significant impact on the perceived usefulness of information [15]. From the perspective of trust transfer, Hu et al. [41] showed that if social media users have enough trust in influencers, they will consider applications approved by those influencers to be highly useful.

Expertise consists of accumulated knowledge, skills, and competencies, and experts often have the ability to provide reliable advice that reduces costs, such as time and effort spent searching for information and evaluating products [42]. Tien et al. [43] found that source expertise greatly predicted cosmetic users’ evaluation of the usefulness of eWOM on social media, in that information providers with a high level of expertise were more able to convince other consumers.

Physically attractive people proved to be more likely to be effective communicators and, thus, more persuasive than less physically attractive people [44]. For social media users, the best proof of the usefulness of fitness videos’ content is the visual appearance of the social media fitness influencer, such as their body shape, skin, and energy. A physically attractive influencer will inspire the viewer’s desire to “want to be like him or her”, thus generating the willingness to exercise, as Durau et al. [2] defined the concept of “the motivating power”.

Based on previous studies, we established the following hypotheses to verify the influence of fitness influencers’ attributes on the perceived usefulness of their videos’ content:

The trustworthiness of social media fitness influencers is positively correlated with perceived usefulness.

The expertise of social media fitness influencers is positively correlated with perceived usefulness.

The physical attractiveness of social media fitness influencers is positively correlated with perceived usefulness.

3.2. Content Quality of Fitness Videos, Information Credibility, and Perceived Usefulness

Content quality includes the accuracy, completeness, relevance, and timeliness of information [45]. High-quality information can help consumers to reduce risk and make better decisions, and consumers are more likely to have confidence that the information and the information provider are reliable [46]. Vila and Kuster [47] demonstrated that the information quality of a website can enable consumers to build more perceived trust. Saima and Khan [48] proposed that, in social media marketing interactions, the quality of information provided by influencers significantly affects their credibility and consumers’ purchase intentions.

Lin and Lu [49] explored the factors that affect users’ acceptance or rejection of websites, and their results showed that websites that provide higher-quality information will be perceived as more useful by the user. The consumer’s decision-making process is energy-consuming, and if a website does not provide clear, complete, and current information, consumers will abandon it [50]. Effective information will lead to the internalization of behavior, and the more specific the information, the easier it is to provoke changes in individual behavior [51].

Based on previous studies, we established the following hypotheses to verify the influence of the content quality of fitness videos on the credibility of the videos’ information and the perceived usefulness of their content:

The content quality of fitness videos is positively correlated with information credibility.

The content quality of fitness videos is positively correlated with perceived usefulness.

3.3. Electronic Word-of-Mouth of Fitness Videos, Information Credibility, and Perceived Usefulness

The increasing maturity of internet technology and social media has facilitated the development of online WOM, that is, electronic word-of-mouth (eWOM); eWOM is not limited by time and space, providing online consumers with a platform to connect the whole world via the internet, where they can obtain a large amount of information on product experiences from experienced consumers as a reference [52]. Furthermore, eWOM can provide consumers with high levels of transparency on market and product information, and it is easier to understand than seller information [53]; so, consumers search and read online recommendations in the pre-purchase phase, mainly to save time and make better decisions [54]. Experienced consumers are considered to have no vested interest in a product; their personal experiences and opinions create vicarious experiences for users, and thus, they are perceived as trustworthy [13,55].

eWOM is becoming the main reference information for people’s purchasing decisions [56]. Huang’s research showed that users tend to rely on social cues as a reference for perceptions of usefulness before experiencing an information system [57]. Positive word-of-mouth from users who have used it will convince prospective users that the system is a useful tool. In addition, people may try to observe others and follow them in making decisions in order to reduce uncertainty and avoid information asymmetry [58].

Based on previous studies, we established the following hypotheses to verify the influence of eWOM about fitness videos on the credibility of the videos’ information and the perceived usefulness of their content:

The eWOM of fitness videos is positively correlated with information credibility.

The eWOM of fitness videos is positively correlated with perceived usefulness.

3.4. Information Credibility and Exercise Intention

Behavioral intention is considered to be the most effective and closest predictor of actual behavior [59]. Many studies have confirmed the positive correlation between credibility and behavioral intention, especially in business transactions. For example, Choi et al. confirmed that believable information and services are factors in the intention to use mobile apps for travel-related purposes [60]. Research by Kim et al. [61] showed that trust has a strong positive impact on consumers’ online transactions. The degree of trust is an important variable in predicting behavioral intention, and the degree of trust that users have for websites is a key antecedent for their online purchases [62]. Based on these research foundations to predict the effect of credibility on behavioral intention, we established the following hypothesis:

Information credibility is positively correlated with exercise intention.

3.5. Perceived Usefulness and Exercise Intention

The influence of perceived usefulness on behavioral intention has been reported in various studies. Scholars found that information quality has an indirect effect on continuous use intention, mediated by perceived usefulness, in a study on the determinants of the continuous use intention of food-delivery applications [63]. In addition, a study on the acceptance and use of mobile learning services by Malaysian university students found that usage intention was facilitated by usefulness [64]. Chen et al. proposed that perceived usefulness can greatly predict the continuous usage intention of mobile service users [65]. Based on previous research on the relationship between perceived usefulness and behavioral intention, we established the following hypothesis:

Perceived usefulness is positively correlated with exercise intention.

Based on the discussion above, we proposed a conceptual model (see Figure 1 ).

The research model.

4. Methodology

4.1. Measures

To test the research model and answer the research questions, we designed a questionnaire based on and adapted from previous studies. The first part of the questionnaire collected the demographic information of the participants, including gender, age, and level of education. The second part measured the variables in the study model using a 5-point Likert scale from 1 (“strongly disagree”) to 5 (“strongly agree”). Table A1 presents a detailed overview of the measurement scales, including all items.

Trustworthiness/expertise/attractiveness: The participants were asked to rate social media fitness influencers’ trustworthiness (Cronbach’s α = 0.960), expertise (Cronbach’s α = 0.943), and attractiveness (Cronbach’s α = 0.926), and we adapted Ohanian’s scale [12] for validating celebrity endorsers. Trustworthiness consisted of four items: trustworthy, sincere, honest, and dependable. Expertise consisted of four items: knowledgeable, experienced, skilled, and expert. Attractiveness consisted of four items: classy, beautiful, sexy, and attractive.

eWOM: To measure the impact of eWOM, we adapted Abubakar and Ilkan’s scale [66], and participants were asked to indicate the extent to which eWOM affected their perceived trust and the perceived usefulness of social media fitness videos; one example item was “I often gather information from other fitness enthusiasts’ online reviews before using fitness videos for workouts” (Cronbach’s α = 0.931).

Content quality: To measure the effect of the content quality of social media fitness videos on viewers, we adapted a scale developed by Magno [67], consisting of four items: reliable, updated, accurate, and high-quality. The participants were asked to evaluate the content quality of fitness videos (Cronbach’s α = 0.923).

Information credibility: We adapted a scale from Lederman et al. [68] and asked the participants to indicate what they thought of the information provided by the fitness videos: “I think fitness video information is truthful” (1 = strongly disagree; 5 = strongly agree) (Cronbach’s α = 0.962).

Perceived usefulness: Measures of perceived usefulness were adapted from Nagy [69] and Davis [70] and consisted of three items. The participants were asked to rate the effectiveness of the fitness videos; one example item was “Using fitness videos helps me improve my fitness efficiency” (Cronbach’s α = 0.950).

Exercise intention: The measure of exercise intention was adapted from Bhattacherjee and Sanford [24]; the participants were asked to indicate whether they intended to exercise after watching a fitness video. One example item was “I intend to use this fitness video to exercise in the near future” (Cronbach’s α = 0.881).

4.2. Data Collection and Sample

The questionnaire was distributed through Wenjuanxing (www.wjx.cn, accessed on 25 October 2023), a professional online questionnaire survey platform. We sent questionnaires via social media to potential participants from China. We set screening questions in the questionnaire to check whether they had watched fitness videos on social media and whether they had experience using fitness videos for exercise. Potential participants would have to answer “yes” to participate in the survey, and unqualified participants were not allowed further access. Finally, we collected a total of 367 valid questionnaires after excluding invalid answers. Table 1 shows the sociodemographic characteristics of the participants, who comprised 240 females (65.40%) and 127 males (34.60%). The largest proportions of participants were 19–35 years old (43.05%) and 36–59 years old (40.87%), while smaller numbers were 18 years and under or 60 years and over. Among all of the participants, 26.70% had a high school education or below, 23.98% had a college education, 38.42% had a bachelor’s degree, and 10.90% had a graduate degree.

Table 1

Sociodemographic characteristics of the participants (n = 367).

5. Data Analysis and Results

The study applied the partial least squares structural equation modeling (PLS-SEM) statistical approach and analyzed the data using SmartPLS 3.2.9 software (SmartPLS GmbH, Oststeinbek, Germany). When applying structural equation modeling (SEM), there are two types of methods to estimate the relationships hypothesized in the model: covariance-based techniques (CB-SEM), and variance-based partial least squares (PLS-SEM) [71]. PLS-SEM operates much like a “multiple regression analysis” [72]. Unlike CB-SEM, which “estimates model parameters so that the discrepancy between the estimated and sample covariance matrices is minimized”, PLS-SEM “maximizes the explained variance of the endogenous latent variables by estimating partial model relationships in an iterative sequence of ordinary least squares (OLS) regressions” [71]. CB-SEM is suitable for testing a theory in the confinement of a concise theoretical model, while PLS is generally more conducive to smaller samples and more complex models. PLS emphasizes prediction and is preferred for research aimed at the prediction and explanation of target constructs [71,73]. The evaluation of the model followed the two-step method of measurement model evaluation and structural model evaluation.

5.1. Measurement Model Evaluation

To assess the validity and reliability of the instrument, we examined whether the instrument met the commonly suggested criteria for measurement model assessment: indicator reliability, construct reliability, convergent validity, and discriminant validity. Indicator reliability was measured by item loadings, and the recommended threshold is 0.7 [74]. As shown in Table 2 , the standardized factor loading of each item in this study was higher than 0.7. Construct reliability was tested by composite reliability (CR) and Cronbach’s alpha (CA), and Table 1 shows that all constructs had CR and CA above the recommended threshold of 0.7 [75], thus confirming construct reliability. The convergent validity was assessed by the average variance extracted (AVE), and the AVE values for all constructs in Table 2 were higher than 0.5, which means that all constructs had satisfactory convergent effectiveness [74]. The discriminant validity, which reflects the degree of uniqueness of the latent variable, was assessed by the square root of the average variance extracted (AVE). According to Table 3 , the square roots of the AVE for all potential variables were higher than their intercorrelations, which indicates that the scales had high discriminant validity [76]. Hence, this study’s constructs can be further used to test the structural model.

Table 2

Factor loading, Cronbach’s α, composite reliability, and AVE.