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

Abstract

Objective

Social media fitness influencers are evolving into a new digital form of health communicators whom consumers might turn to for assistance with more physical activity and exercise at home, especially in the current COVID-19 crisis. Drawing from source credibility theory, social cognitive theory, protection motivation theory and literature on physical activity, physical fitness and gender, we analyse how male and female users’ evaluations of social media fitness influencers and user health-related variables impact intentions to exercise with the social media fitness influencer.

Methods

Two consecutive studies using male and female YouTube fitness influencers were carried out. Structural equation modelling was conducted to test the proposed models and estimate the path coefficients.

Results

Study 1 (N = 507) shows that the respective influencer's perceived trustworthiness, expertise and attractiveness are important for the influencer's evaluation, and that it is the perceived motivating power rather than the attitude toward the influencer that increases intentions to exercise for male and female users. Study 2 (N = 445) extends Study 1 and shows that physical fitness, training involvement with YouTube fitness videos and lower health increase behavioural intentions. A more negative body image also raises intentions to exercise, but only among female users.

Conclusion

Social media fitness influencers who are perceived as trustworthy, as experts and as attractive, can be effective for increasing men's and women's physical activity. Perceived motivating power of the influencer emerged as a key variable that predicts intentions to exercise. User health-related variables have different effects on intentions to exercise for men and women.

Keywords

Influencer social media physical activity exercise gender health communications quantitative

Introduction

Social media has developed into a vital source of health information. It provides a communication channel in which users and health communicators can connect regarding various health-related aspects. One popular, health-related topic in many social media channels revolves around physical activity and fitness.¹ On YouTube, one of the most popular social media sites, users have access to videos in which social media fitness influencers demonstrate physical activities and try to motivate the users to mimic their workouts.

In times of globally decreasing levels of physical activity, with significantly negative implications for people's health,^2–4 it is highly important to find ways to encourage people to become more active.^5,6 Given their ability to reach a large audience, social media fitness influencers might be an important digital type of health communicator who could influence health behaviours.^7–9 This is even more apparent with COVID-19 spreading worldwide, which has provided a barrier for people's physical activity routines, as gyms or parks were closed.¹⁰ During the pandemic, interest in online home workouts provided by digital and social media has increased,^11,12 with social media fitness influencers offering fitness workouts on their channels.¹¹ On YouTube, home workouts and instructional videos by social media fitness influencers are free, easy to access online, and often do not require any equipment.¹¹ Thus, exercising at home guided by social media fitness influencers allows individuals to maintain or improve their physical fitness in safety from home, which represents a great advantage in uncertain pandemic times.

While research interest in health and fitness communication and information on social media platforms^5,8,9,13 and on fitness apps^14–17 has increased in recent years, little is known about how social media fitness influencers impact users’ health-related behaviour. Our aim is to fill this research gap. Drawing from source credibility theory, social cognitive theory, protection motivation theory (PMT) and literature on physical activity, physical fitness and gender, we carry out two studies that shed light on the question of what makes a social media fitness influencer successful from the users’ perspectives. Study 1 investigates the impact of the influencers’ source credibility variables (perceived trustworthiness, expertise, and attractiveness) on the attitude toward the influencer, the influencer's perceived motivating power, and, in turn, on intentions to perform the advocated fitness behaviour. Study 2 extends the research and includes important user-related health and fitness variables (self-reported health, physical fitness, body image, physical fitness involvement, and training involvement with YouTube fitness videos). Since gender influences health information behaviour in many different ways,¹⁸ the role of users’ gender is also investigated.

The research contributes to an improved understanding of how users might be motivated by social media fitness influencers to engage in physical activity. Hence, our research is relevant to users and social media fitness influencers as well as to health professionals.

Theoretical and empirical background

Social media fitness influencers as health communicators

Social media influencers are individuals who communicate information through their social media accounts to their followers. Given their large reach, successful influencers are even considered online celebrity endorsers.¹⁹ By sharing content about their lifestyle and engaging with their followers, social media influencers can influence the attitudes and behavioural intentions of their audiences, also regarding exercise.^9,20,21 If they specialize in physical activity and fitness content, they are labelled as social media fitness influencers.

According to source credibility theory, endorsers’ effectiveness in delivering messages to their target audience depends on several factors. The current research differentiates between the endorser's trustworthiness, expertise, and attractiveness as perceived by the users.²² Trustworthiness describes the extent to which a respondent believes the endorser and the endorser's message²² and has been identified as a key success factor of sources in early source credibility studies²³ and also more recently in influencer effectiveness studies.^19,24,25 Expertise describes the respective source's perceived competence.²² For the acceptance of the endorser's information, it is not relevant whether the endorser is indeed an actual expert, but rather whether they are perceived as an expert.^26–28 Attractiveness refers to the physical appearance of the source.²⁷ The endorser's attractiveness is typically judged at the initial contact and influences the evaluation of the person. Considerable empirical research on the effects of celebrity and athlete endorsements has confirmed the importance of these three factors for successful communicators^26,29,30 and shown that higher levels of source trustworthiness, expertise, and attractiveness are typically beneficial for the source and lead to improved attitudes toward the source. Recently, studies have applied these source credibility constructs in social media influencer research,^19,25,31,32 albeit none of them were related to social media fitness influencers.

Since motivation plays a vital role in engaging individuals in physical activity¹⁷ and a lack of motivation has been identified as a major reason for physical inactivity,⁴ the motivating power attributed to the influencer is an important variable to consider. Perceived motivating power refers to how strongly users rate the influencer's capacity to encourage individuals to engage in physical activity. Hence, in the context of social media fitness influencers, we expect that trustworthiness, expertise, and attractiveness improve attitudes toward the social media fitness influencer and also increase the motivating power ascribed to the social media fitness influencer. Therefore, higher trustworthiness, expertise, and attractiveness are expected to be beneficial for the social media fitness influencer, and we hypothesize:

H1: The higher the perceived (a) trustworthiness, (b) expertise and (c) attractiveness of the social media fitness influencer, the more favourable the attitude toward the social media fitness influencer.

H2: The higher the perceived (a) trustworthiness, (b) expertise and (c) attractiveness of the social media fitness influencer, the higher the motivating power of the social media fitness influencer.

Motivation in particular has been identified as a key influence on exercise behaviours in non-social media contexts^33,34 and for fitness apps.¹⁷ Exercise refers to planned, structured and repeated physical activities and is thus a subset of physical activity.³⁵ The extrinsic motivation to exercise can originate from friends and family or from role models. According to social cognitive theory, people learn by observing a model performing a certain behaviour, and the model behaviour can exert motivational influence on the individual.^36,37 By watching the behaviour's consequences, individuals memorize the model's behaviour as a behaviour option that might motivate future behaviour. Influencers can affect their followers’ decision-making and elicit positive effects on intentions to follow their advice³⁸ and motivate certain behaviours. Hence, if a social media fitness influencer elicits favourable attitudes and users attribute high motivating power to the influencer, the users’ intentions to exercise with the social media fitness influencer are expected to be higher. Thus, we propose:

H3: The more favourable the attitude toward the social media fitness influencer, the higher the intention to exercise.

H4: The greater the perceived motivating power of the social media fitness influencer, the higher the intention to exercise.

Moderating role of user gender on the effectiveness of social media fitness influencers

Gender has been determined as an important variable in the health and fitness area.³⁹ For instance, men are more motivated and more likely to engage in sports and are more physically active during their leisure time than women.^4,40 Women and men also differ in their motives for engaging in physical activity. While women are primarily interested in weight management or improving their health or physique,^33,41 men are more motivated to increase strength or endurance.⁴² User gender might therefore moderate the proposed model relationships. To explore the role of users’ gender on the hypothesized relationships (see Figure 1), we analyse the following research question:

RQ1: How does user gender moderate the hypothesized relationships?

Study 1

The purpose of Study 1 was to investigate the impact of the influencer source credibility variables (perceived trustworthiness, expertise and attractiveness) on the attitude toward the influencer and the influencer's perceived motivating power and, consequently, on the users’ intentions to engage in physical activity, moderated by user gender, while controlling for influencer gender and user age (Figure 1). Study 1 employed two existing YouTube fitness video clips featuring a German-speaking male and female social media fitness influencer. Relationships were tested simultaneously with multi-group structural equation modelling.

Methods

Stimuli selection and pre-tests

Pre-test 1

The first pre-test aimed at identifying the fitness video's workout topic. It should be popular among users and suitable for all people. The authors created a list of popular YouTube workout topics (such as strength training, cardio, abs or legs exercises). One hundred thirty-four respondents participated in a survey (non-student online sample, M_age = 28.28 years; 55.2% female) and rated their preferred workout topics. Ab exercises emerged as the most popular exercise topic for both groups (see Table 1 for details).

Table 1.

Pre-Test 1: variables with means and standard deviations per group.

Fitness video topics	Women		Men		Total sample
Fitness video topics	M	SD	M	SD	M	SD
Ab workout	3.82	2.154	3.70	2.374	3.77	2.248
Strength training	3.58	2.215	3.66	2.418	3.47	2.454
Legs workout	3.31	2.488	3.58	2.486	3.45	2.213
Cardio workout	3.24	2.172	3.47	2.205	3.39	2.264
Fitness tips	3.22	2.064	3.45	2.158	3.33	2.160
Fitness motivation	3.11	2.117	3.28	2.218	3.26	2.152
Upper body workout	3.08	2.107	3.13	2.213	3.12	2.152
Yoga	2.82	2.091	2.35	2.073	2.61	2.088
Dancing, aerobics, zumba	2.79	2.398	2.25	1.954	2.55	2.217

Pre-test 2

The second pre-test aimed at selecting the video clips for the main study. Based on the results of Pre-test 1, the authors researched YouTube videos featuring ab exercises by popular German social media fitness influencers and selected relatively similar sets of videos for the female and male influencers. Twenty-seven undergraduate students (M_age = 21.24 years; 59.3% female) watched sets of male and female social media fitness influencers’ YouTube video clips and discussed each influencer's trustworthiness, expertise and attractiveness afterward. The final sets of the influencers’ videos showed similar ab workouts, and both influencers were perceived as trustworthy, experts and attractive. Both videos were cut to have the same length (47 s).

Pre-test 3

The third pre-test examined the suitability of the two videos (male version, female version) and the comprehensibility of the questionnaire. Twenty-five undergraduate students participated in the third pre-test (M_age = 22.56 years; 72% female). Participants evaluated the two videos as suitable and reported no difficulties in answering the questionnaire.

Participants

Main Study 1 was a survey study where only de-identified data was collected for which ethical approval is not required under European law. It was a study with human subjects that presents no greater than minimal risk to subjects. Before completing the questionnaire, the subjects were assured in written form of anonymous data collection (i.e. no identifiers can be connected to the data, either directly or through a coding system), of the entirely voluntary nature of participation and of the possibility of terminating the study at any time without consequences. The study meets the criteria of the European General Data Protection Regulation (GDPR).

Study 1 had a non-student sample with 507 participants (M_age = 30.75, from 18–68 years; 50.7% female). Data was collected in the German-speaking area with a structured online questionnaire. Participants were recruited by distributing the questionnaire online via the snowball principle and through the online consumer panel ClickWorker (similar to MTurk in the USA). This sampling procedure has been successfully applied by previous research.⁴³ Respondents received a small compensation (1 EURO) for participating in the survey.

Procedure

The study had a cross-sectional, quasi-experimental 2 (male/female influencer) × 2 (male/female user) between-subject design. Participants were randomly assigned to one of the two fitness video clips (either version). They watched the video clip and then answered questions about the clip. Sociodemographic data on gender, age and nationality were collected at the end of the questionnaire.

Measures

Items were derived from existing multi-item scales when available and adapted to the research context. Table 2 reports the scale references, the items for all variables, their means, standard deviations, Cronbach α values for multi-item measures and Spearman-Brown coefficients for two-item measures.⁴⁴ Factor analyses were performed to investigate the one-dimensionality of the multi-item scales. Reliability analyses were conducted for all variables. Perceived trustworthiness, expertise and attractiveness were each measured with three items from the celebrity endorsement scale by Ohanian.²² The scales achieved good reliability (trustworthiness: α = 0.930; expertise α = 0.949; attractiveness α = 0.946). The motivating power ascribed to the influencer was assessed with the item “The influencer is motivating” (M = 4.03; SD = 1.625), adapted from Döring.⁷ Attitude towards the influencer was measured with two items adapted from Pollay and Mittal.⁴⁵ (Spearman-Brown coefficient = 0.916) Two items adapted from Ajzen⁴⁶ assessed intention to exercise (Spearman-Brown coefficient = 0.953).

Table 2.

Study 1: variables with means and standard deviations per group.

Variables and items	Women		Men		Total sample
Variables and items	M	SD	M	SD	M	SD
Trustworthiness^a (α = 0.930)	4.58	1.464	4.52	1.428	4.55	1.446
The influencer is … trustworthy sincere reliable
Expertise^a (α = 0.949)	4.61	1.521	4.47	1.415	4.54	1.470
The influencer is …experiencedqualifiedskilled
Attractiveness^a (α = 0.946)	3.88	1.618	3.99	1.703	3.93	1.660
The influencer is …beautifulattractivesexy
Motivating power^b	4.08	1.645	3.97	1.607	4.03	1.625
The influencer is motivating.
Attitude toward the influencer^c (Spearman-Brown coefficient = 0.916)	4.47	1.468	4.51	1.430	4.49	1.448
The influencer is …goodpositive
Intention to exercise^d (Spearman-Brown coefficient = 0.953)	2.45	1.750	2.31	1.690	2.38	1.720
I intend to exercise with this YouTube fitness video within the next month.It is very likely that I will exercise with this YouTube fitness video within the next month.

Note. All items were measured on a 7-point Likert scale from 1 (“do not agree at all”) to 7 (“totally agree”).

Based on Ohanian.²²

Adapted from Döring.⁷

Adapted from Pollay and Mittal.⁴⁵

Adapted from Ajzen.⁴⁶

Measurement model

Prior to testing the structural equation model (SEM), we conducted a confirmatory factor analysis (CFA) to assess the measurement model for the whole data set using IBM SPSS AMOS 25. Results yielded an acceptable overall model fit (x²/df = 2.537, TLI = 0.977 CFI = 0.985, RMSEA = 0.055) and acceptable local fit measures.⁴⁷ Table 3 reports the average variance extracted (AVE), composite reliability and indicator reliability for all constructs and the test for discriminant validity,⁴⁸ which was supported as well.

Table 3.

Study 1: reliability and validity measures for variables (CFA), discriminant validity and squared construct correlations.

Construct	Indicator reliability	Composite reliability	AVE	Highest squared construct correlation
1 Trustworthiness	0.810–0.858	0.941	0.841	0.728
2 Expertise	0.858–0.911	0.957	0.882	0.596
3 Attractiveness	0.835–0.914	0.955	0.877	0.341
4 Attitude	0.814–0.877	0.916	0.845	0.494
5 Motivating power	-	-	-	-
6 Intention to exercise	0.898–0.923	0.953	0.910	0.321

Note. AVE: average variance extracted; -: not applicable.

SEM results

The multi-group SEM resulted in an acceptable model fit (x²/df = 2.091, CFI = 0.977, TLI = 0.968, RMSEA = 0.046) and was valid for female and male respondents. Trustworthiness had the strongest impact on attitude toward the social media fitness influencer for men and women (γ_women = 0.457, p < 0.01; γ_men = 0.485, p < 0.01). Expertise (γ_women = 0.261, p < 0.05; γ_men = 0.265, p < 0.05) and attractiveness (γ_women = 0.270, p < 0.01; γ_men = 0.226, p < 0.01) also yielded significant influence on the attitude toward the influencer for both groups. The results confirmed H1a-c. Trustworthiness was also the strongest predictor of motivating power for women (γ_women= 0.379, p < 0.01; γ_men = 0.209, p < 0.05), while the influencers’ expertise (γ_women= 0.194, p < 0.05; γ_men = 0.413, p < 0.01) was the strongest predictor of motivating power for men. Attractiveness also influenced motivating power significantly for both genders (γ_women= 0.294, p < 0.01; γ_men = 0.270, p < 0.01). H2a-c was also confirmed by the data.

Interestingly, attitude toward the influencer had no effect on the participants’ intentions to perform physical activity (γ_women = 0.078, not significant (n.s.); γ_men = 0.023, n.s.), disproving H3, while motivating power led to higher intentions to perform physical activity for both genders (γ_women = 0.541, p < 0.01; γ_men = 0.518, p < 0.01), confirming H4. Regarding the control variables, the influencer's gender had no significant effect on behavioural intentions, either for male or female users (γ_women = −0.008, n.s.; γ_men = −0.007, n.s.). User age had a significant positive effect on intentions to exercise for men (γ_men = 0.129, p < 0.05), but no significant effect for women (γ_women = −0.016, n.s.).

We conducted pairwise comparisons to test for significant differences between female and male users in the model relationships. The impact of expertise on motivating power was slightly stronger among males than females at the 10% level. The path from user age to intention to exercise differed on the 10% level as well (z-value >|1.65|). Overall, the model and its relationships were very similar for both genders. Table 4 displays the path coefficients, significance, and pairwise comparison results.

Table 4.

Study 1: SEM results.

Path	H	Path coefficient				C.R.
		Women		Men
Trustworthiness → Attitude	H1a	0.457***		0.485***		0.214
Expertise → Attitude	H1b	0.261**		0.265**		0.252
Attractiveness → Attitude	H1c	0.270***		0.226***		−0.655
Trustworthiness → Motivating power	H2a	0.379***		0.209**		−1.352
Expertise → Motivating power	H2b	0.194**		0.413***		1.888*
Attractiveness → Motivating power	H2c	0.294***		0.270***		−0.426
Attitude → Intention to exercise	H3	0.078		0.023		−0.501
Motivating power → Intention to exercise	H4	0.541***		0.518***		−0.295
Influencer gender → Intention to exercise		−0.008		−0.007		0.012
User age → Intention to exercise		−0.016		0.129**		1.828*
	x²/df	RMSEA	CFI		TLI
Structural model Study 1	2.091	0.046	0.977		0.968

Note. *** p < 0.01; ** p < 0.05; * p < 0.10.

Discussion of study 1

The results of Study 1 imply that social media fitness influencers’ trustworthiness is more important for users’ attitude toward the influencer than social media fitness influencers’ expertise and attractiveness. With regard to perceived motivating power, trustworthiness is the most relevant variable for women. Among men, expertise is more important than the other source credibility variables. Attractiveness is less important for both genders. Another important result of Study 1 is that the attitude toward the influencer is largely irrelevant for users’ intentions to exercise, while the motivating power attributed to the influencer is the key to positively influencing users’ intentions to exercise for both groups.

Impact of user health and fitness variables on the intention to engage in physical activity

Study 1 explored the roles that social media fitness influencers’ trustworthiness, expertise and attractiveness play in female and male users’ intentions to follow the advocated physical activity and investigated whether the users’ gender matters. In the following, we extend the model with the important health and fitness variables: self-reported health, physical fitness, body image, physical fitness involvement and training involvement with YouTube fitness videos. We expect that these variables might impact users’ intentions to engage in the advocated physical activity. We are again interested in the question of whether the user's gender moderates the model relationships. Figure 2 displays the final conceptual social media fitness influencer model. The variables and relationships will be elaborated in the following paragraphs.

A variable that is likely to play a role in users’ intentions to exercise with a social media fitness influencer is users’ health status.⁴⁹ While a person's health status is undoubtedly complex and multi-faceted,^50,51 individuals typically judge their perceived health as a single assessment (e.g. “good” or “poor”), that is, instead of evaluating different facets of their health separately, individuals often assess their overall health status. This summary assessment is referred to as self-reported health and is commonly used to measure one's perceived health status.^52,53 In a social media fitness influencer setting, we expect lower self-reported health to lead to increased intentions to perform physical activity with the social media fitness influencer's video. We draw our rationale from PMT⁵⁴ which is a major health psychology theory that aims at gaining a better understanding of the cognitive processes inherent to adopting health-related behaviours, for example physical activity. PMT suggests that when individuals perceive a threat, they appraise possible coping mechanisms that help them to remove the threat.^54–56 We assume that individuals likely perceive a low health status as threatening and therefore try to engage in coping mechanisms. One way to cope is by engaging in physical activity that contributes positively to one's health⁵⁵ and can easily be carried out. Two reviews^55,56 and several empirical studies in different contexts on PMT and physical activity^57,58 suggest that the PMT can predict health-related behaviours. Emulating exercises performed by social media fitness influencers promises relatively easy improvements of one's health status. Social media fitness influencers usually demonstrate exercise behaviours that are not too difficult to carry out and exercises can be performed at home; hence, they are fairly easy to realize. Thus, we hypothesize:

H5: The lower the individual‘s self-reported health, the higher the intention to exercise.

Physical fitness can be considered as a set of attributes that are either health- or skill-related and that people have or achieve.³⁵ Health-related components of physical fitness include cardiorespiratory endurance, muscular endurance, muscular strength, body composition and flexibility, while skill-related attributes include agility, balance, coordination, power, speed and reaction time, which can be assessed by physical tests.^35,59 Usually, the health-related components are the focus of empirical studies on fitness.^35,59

We also focus on the health-related components and follow the notion that physical fitness is a state of well-being with low risk of premature health problems and the ability to participate in a variety of physical activities,^59,60 which is a desirable state for individuals. Physical fitness can be achieved through (sustained) physical activity. Ortega et al.⁶¹ showed that individuals’ self-reported overall fitness was strongly associated with their actual overall fitness (as measured by objective laboratory tests). We focus on the perceived overall fitness of individuals. It is also an individual characteristic that can influence exercise intentions.⁶² Lower perceived physical fitness probably leads to higher intentions to exercise with the fitness influencers in order to improve one's own physical fitness, similar to the effect of the health status on the intention to exercise. Following the PMT, individuals may again perceive their low physical fitness as a threat to their overall health and again develop intentions for counteractive behaviours, such as the intention to exercise to reduce such threats.

Social media fitness influencers promise a relatively convenient way to improve one's physical fitness by emulating the influencer's workout from home, which might be particularly attractive to individuals with lower fitness capabilities, because they can exercise at their own pace without being observed by others. We hypothesize:

H6: The lower the individual‘s physical fitness, the higher the intention to exercise.

Body image is a multidimensional construct that refers to perceptions, feelings, thoughts and behaviours about one's body and can range from positive to negative.⁶³ Body image mirrors how satisfied or dissatisfied a person feels about their shape and weight. Exercise has been shown to improve body image.⁶⁴ In fitness-related image-based social media, the ideal female body is typically portrayed as thin, fit and athletic,^65,66 while the ideal male body is presented as muscular, toned and lean.⁶⁷ Social media fitness influencers with such an ideal body are likely to influence body-related perceptions and behaviours, such as physical activity, with potential positive and negative outcomes.^37,66 Negative outcomes of social media use in relation to body image (e.g. body image concerns, body dissatisfaction, disordered eating) have been investigated by a large body of research.^66,68,69 However, exercise might have positive effects on one's body image as well,⁶⁴ since individuals might develop positive expectations in relation to their health (e.g. getting fit, having a well-trained healthy body), particularly if the goal seems achievable,²¹ which, in turn, encourages an engagement in physical activity and exercise. Hence, lower body image might increase exercise intentions with the social media fitness influencer. Influencers are often considered as peers and their positive fitness results might seem achievable and motivating,^21,70 because individuals are likely to expect an improvement of their bodies’ appearance in relation to their health through exercise. Therefore, we assume – similar to the proposed effects of health status and physical fitness on intention to exercise – that a lower body image increases exercise intentions, since social media fitness influencers encourage users to exercise and create a positive body image by developing a healthy and fit body. We postulate:

H7: The lower the individual‘s body image, the higher the intention to exercise.

Involvement is a central construct for predicting and understanding behaviour.⁷¹ Physical fitness involvement refers to a person's general concern about and interest in physical fitness and how important achieving physical fitness is to the individual.^72,73 As stated previously, physical fitness is considered a desired health outcome of physical activity.^35,62 Physical fitness involvement has been shown to successfully predict exercise and leisure behaviours.^73,74 For instance, individuals with higher physical fitness involvement were found to be more likely to engage in related behaviours, such as physical activity.^72,75 An elevated physical fitness involvement can be seen as a prerequisite for exercise intention, as it indicates a general interest in sport exercises. We hypothesize:

H8: The higher the individual‘s physical fitness involvement, the higher the intention to exercise.

In the context of YouTube fitness videos, training involvement with YouTube fitness videos – a more specific form of physical fitness involvement – might influence intentions to exercise as well. If YouTube fitness video training plays a significant role in an individual's life, this involvement might then positively influence the user's intention to perform a fitness workout.⁶² Thus, we hypothesize:

H9: The higher the individual‘s training involvement with YouTube fitness videos, the higher the intention to exercise.

Study 2

The purpose of Study 2 was to test the extended model relationships. As in Study 1, relationships were tested simultaneously with multi-group SEM. Study 2 used different videos, but the same influencers as in Study 1, and again featured ab exercises. Two pre-tests were conducted to select the stimuli.

Methods

Stimuli selection and pre-test

Pre-Test 1

The first pre-test was a guided interview with the aim of selecting a different set of video clips featuring the same influencers as in Study 1. The authors researched YouTube fitness videos with a focus on abs from the two social media fitness influencers. Eleven subjects (M_age = 24.18 years) watched 3 × 2 (male vs female) YouTube social media fitness influencer video clips performing the same ab exercise. Then, they evaluated their attitude toward the influencers with a paper-and-pencil questionnaire. Afterwards, influencers’ trustworthiness, expertise, attractiveness and motivation were discussed in the group. The final two videos (male version, female version) had the most favourable and similar results. They were cut to the same length of 22 s.

Pre-Test 2

The second pre-test (student sample, N = 25, M_age = 20.92 years) aimed at confirming the suitability of the stimuli and the comprehensibility of the online questionnaire. Participants reported no difficulties in answering the questionnaire and evaluated the clips as suitable.

Participants

Consistent with Study 1, Study 2 presented no greater than minimal risk to subjects and participants were assured in written form of anonymous data collection, of the entirely voluntary nature of participation and of the option to end the study at any time without consequences. Again, the criteria of the European GDPR were met.

Study 2 had a non-student sample with 445 participants (M_age = 29.6, from 16–63 years; 55.3% female). Data was collected with a structured online questionnaire in the German-speaking area. As in Study 1, participants were recruited online via the snowball principle and through the online consumer panel ClickWorker. Again, respondents received a small compensation (1 EURO) for their participation.

Procedure

The study had a cross-sectional, quasi-experimental 2 (male/female influencer) × 2 (male/female user) within-subject design. Each participant randomly watched one fitness video clip (the female or male version first), answered questions about it, then watched the other video clip and completed questions about it. Respondents reported sociodemographic data (gender, age, nationality) at the end of the questionnaire.

Measures

Study 2 included the same constructs as Study 1 (trustworthiness, expertise, attractiveness, attitude towards the influencer, motivating power, behavioural intention, age and gender) and the additional health and fitness variables. As in Study 1, items were measured on a 7-point-Likert scale from 1 (“do not agree at all”) to 7 (“totally agree”). Factor analyses were conducted to ensure the one-dimensionality of the multi-item scales. Reliability analyses were performed for all variables. All scales had acceptable Cronbach-Alpha values and Spearman-Brown coefficients, respectively. Items, mean values, standard deviations and reliability values of Study 2 are displayed in Table 5.

Table 5.

Study 2: variables with means and standard deviations per group.

Variables and items
	Women		Men		Total sample
	M	SD	M	SD	M	SD
Trustworthiness^a (α = 0.938)	4.52	1.457	4.30	1.510	4.42	1.485
The influencer is …trustworthysincerereliable
Expertise^a (α = 0.953)	4.46	1.518	4.25	1.533	4.37	1.528
The influencer is …experiencedqualifiedskilled
Attractiveness^a (α = 0.936)	3.54	1.601	3.74	1.695	3.63	1.646
The influencer is …beautifulattractivesexy
Motivating power^b (α = 0.899)	3.73	1.556	3.56	1.488	3.65	1.527
The influencer …is motivatingencourages to keep goingencourages to carry on
Attitude toward the influencer^c (Spearman-Brown coefficient = 0.940)	4.55	1.549	4.31	1.614	4.44	1.582
The influencer is …goodpositive
Intention to exercise^d (Spearman-Brown coefficient = 0.956)	2.15	1.480	2.24	1.684	2.19	1.574
I intend to exercise with this YouTube fitness video within the next month.It is very likely that I will exercise with this YouTube fitness video within the next month.
Self-reported health^e	5.37	1.406	5.61	1.319	5.48	1.372
My overall health is good.
Physical fitness^f (Spearman-Brown coefficient = 0.895)	4.47	1.338	5.11	1.374	4.75	1.390
My general physical fitness is …
My satisfaction with my general physical fitness is …
Body image^g (Spearman-Brown coefficient = 0.911)	5.05	1.484	5.63	1.242	5.31	1.410
I have a positive attitude toward my body.
My feelings toward my body are positive, for the most part.
Physical fitness involvement^h (Spearman-Brown coefficient = 0.954)	5.47	1.503	5.63	1.365	5.54	1.444
I am interested in physical fitness.
Physical fitness is important to me.
Training involvement with YouTube fitness videos^h (Spearman-Brown coefficient = 0.944)	2.79	1.696	2.61	1.751	2.71	1.722
I am interested in doing my physical fitness training with YouTube fitness videos.
Doing my physical fitness training with YouTube fitness videos is important to me.

Note. All items were measured on a 7-point Likert scale from 1 (“do not agree at all”) to 7 (“totally agree”).

Based on Ohanian.²²

Adapted from Döring⁷ and extended.

Adapted from Pollay and Mittal.⁴⁵

Adapted from Ajzen.⁴⁶

Adapted from Stokols, Shumaker and Martinez.⁷⁶

Adapted from the German version of the International Fitness Scale (IFIS).⁶¹

Adapted from Avalos et al.⁶³

Adapted from Zaichkowsky.⁷¹

Study 1 identified motivating power as a key variable that predicts intentions to exercise, measured with one item. In Study 2, the authors decided to measure motivating power with three items for a more solid measurement (α = 0.899). Self-reported health was measured with one item adapted from Stokols et al.⁷⁶ Subjects were asked for their level of agreement to the statement “My overall health is good” (M = 5.48, SD = 1.372). Physical fitness was measured with two items adapted from the German version of the International Fitness Scale.⁶¹ (Spearman-Brown coefficient = 0.895) Body image was measured with two items from the Body Appreciation Scale.⁶³ (Spearman-Brown coefficient = 0.911) Physical fitness involvement (Spearman-Brown coefficient = 0.954) and training involvement with YouTube fitness videos (Spearman-Brown coefficient = 0.944) were each measured with two items adapted from Zaichkowsky.⁷¹

Measurement model

In order to assess the measurement model, we carried out a CFA using IBM SPSS AMOS 25. Results yielded an acceptable model fit.⁴⁷ (x²/df = 3.807, CFI = 0.970, TLI = 0.957, RMSEA = 0.056). The local fit measures AVE, indicator reliability, composite reliability and discriminant validity fulfilled the psychometric requirements (see Table 6).

Table 6.

Study 2: reliability and validity measures for variables (CFA), discriminant validity and squared construct correlations.

Construct	Indicator reliability	Composite reliability	AVE	Highest squared construct correlation
1 Trustworthiness	0.843–0.869	0.946	0.854	0.723
2 Expertise	0.847–0.916	0.960	0.888	0.605
3 Attractiveness	0.807–0.886	0.946	0.855	0.321
4 Attitude	0.867–0.906	0.939	0.886	0.416
5 Motivating power	0.534–0.911	0.911	0.775	0.257
6 Intention to exercise	0.898–0.932	0.956	0.916	0.262
7 Self-reported health	-	-	-	-
8 Physical fitness	0.717–0.913	0.890	0.802	0.469
9 Body image	0.730–0.960	0.919	0.850	0.404
10 Physical fitness involvement	0.880–0.944	0.954	0.912	0.052
11 YT fitness training involvement	0.884–0.905	0.944	0.894	0.043

Note. AVE: average variance extracted; -: not applicable.

SEM results

The hypotheses were tested simultaneously with multi-group SEM. We analysed the conceptual model with the extended measurement of motivating power and the control variables influencer gender and user age (Figure 2). Again, two groups were analysed (male vs female respondents) to investigate the moderating effects of user gender. The results of the multi-group SEM yielded an acceptable model fit (x²/df = 2.222, CFI = 0.969, TLI = 0.963, RMSEA = 0.037).

Trustworthiness, expertise and attractiveness had significant effects on attitude toward the influencer and motivating power, which subsequently had significant positive impacts on intentions to exercise for men and women. As in Study 1, H1a-c and H2a-c were confirmed. Trustworthiness had the strongest effect on attitude toward the influencer for both groups (γ_women = 0.577, p < 0.01; γ_men = 0.656, p < 0.01). Expertise had the strongest influence on motivating power for women (γ_women = 0.347, p < 0.01), while trustworthiness had the highest impact on motivating power for men (γ_men = 0.428, p < 0.01). Attitude toward the influencer had significant positive effects on intentions to exercise for women (γ_women = 0.198, p < 0.01), but not for men (γ_men = 0.063, n.s.), leading to partial support for H3. Motivating power (γ_women = 0.271, p < 0.01; γ_men = 0.431, p < 0.01) yielded significant positive effects on intentions to exercise for both genders, confirming H4. Self-reported health had a significant negative effect on behavioural intentions for men and women (for women only at a 10% level) (γ_women = −0.094, p < 0.10; γ_men = −0.190, p < 0.05), confirming H5. Physical fitness yielded a significant positive effect on intentions to exercise for both groups (γ_women = 0.137, p < 0.05; γ_men = 0.217, p < 0.05), rejecting H6. Body image had a significant negative effect on intentions to exercise for women, but not for men (γ_women = −0.180, p < 0.05; γ_men = −0.068, n.s.). Thus, H7 was confirmed for women, but not for men. Involvement with physical fitness had a significant, yet negative, effect on intentions to exercise for women (γ_women = −0.089, p < 0.05), but suggested no significant effect for men (γ_men = 0.055, n.s.), rejecting H8 for both groups. Training involvement with YouTube fitness videos positively influenced intentions to exercise for male and female users (γ_women = 0.431, p < 0.01; γ_men = 0.426, p < 0.01), confirming H9. The control variable, influencer gender, again had no significant effect on the behavioural intentions of women and men (γ_women = −0.056, n.s.; γ_men = 0.006, n.s.). User age had a significant negative effect on intentions to exercise for women but had no significant effect for men (γ_women = −0.079, p < 0.05; γ_men = 0.007, n.s.).

We conducted pairwise comparisons to test for significant differences between the path coefficients of male and female users. The paths from trustworthiness, expertise and attractiveness to motivating power; from motivating power to intentions to exercise; and from physical fitness involvement to intentions to exercise implied significant group differences (z-value >|1.96|). The path from attitude to intention to exercise differed on a 10% level (z-value >|1.65|). Table 7 exhibits the path coefficients, significances and critical ratios.

Table 7.

Study 2: SEM results.

Path	H
		Women	Men	C.R.
Trustworthiness → Attitude	H1a	0.577***	0.656***	0.863
Expertise → Attitude	H1b	0.187***	0.163**	−0.175
Attractiveness → Attitude	H1c	0.199***	0.169***	−1.134
Trustworthiness → Motivating power	H2a	0.180***	0.428***	2.364***
Expertise → Motivating power	H2b	0.347***	0.223***	−2.441***
Attractiveness → Motivating power	H2c	0.163***	0.105***	−2.713***
Attitude → Intention to exercise	H3	0.198***	0.063	−1.695*
Motivating power → Intention to exercise	H4	0.271***	0.431***	3.139***
Self-reported health → Intention to exercise	H5	−0.094*	−0.190**	−1.389
Physical fitness → Intention to exercise	H6	0.137**	0.217**	0.739
Body image → Intention to exercise	H7	−0.180**	−0.068	1.148
Physical fitness involvement → Intention to exercise	H8	−0.089**	0.055	2.354***
Training involvement with YouTube fitness videos → Intention to exercise	H9	0.431***	0.426***	−0.258
Influencer gender → Intention to exercise		−0.056	0.006	1.226
User age → Intention to exercise		−0.079**	0.007	1.107
	x²/df	RMSEA	CFI	TLI
Structural model	2.222	0.037	0.969	0.963

Note. ***p < 0.01; **p < 0.05; *p < 0.10.

Discussion and implications

This research aimed to examine whether and how social media fitness influencers can be successful health communicators and which factors lead to higher intentions of users to engage in physical activity. We analysed how the social media fitness influencer's trustworthiness, expertise and attractiveness influence the attitude toward the influencer and the motivating power that users ascribe to the influencer, how these variables, in turn, influence users’ intentions to exercise, and whether users’ gender affects these relationships. Furthermore, five user-related health and fitness variables were included in the analysis. Our findings from two empirical studies contribute to our understanding of how users evaluate social media fitness influencers, what makes a social media fitness influencer successful from the users’ perspective, and which variables lead to encouraging respondents’ intention to engage in (more) physical activity. Therefore, this study extends research on source credibility to the context of social media fitness influencers. The findings of the two studies show that trustworthiness, expertise and attractiveness help to understand how social media fitness influencers might contribute to motivating users to engage in physical activity. Social media fitness influencers have established themselves as influential in the health and fitness scene.^9,21 Extending previous research findings on source credibility of social media influencers,^19,25,31,32 this research confirms the high relevance of these three factors especially for social media fitness influencers.

Trustworthy, expert and attractive influencers might be effective health communicators in the digital world and might contribute to increasing physical activity^2,3 by motivating people to exercise at home. This research further highlights the positive effects that social media^8,77 and social media fitness influencers might have on users.^9,21 Increasing physical exercise and thus contributing to individuals’ health via digital and social media seems especially important when gyms and fitness clubs are subject to access restrictions, as in the COVID-19 pandemic.^10,11

Among the source credibility variables, the influencers’ trustworthiness appeared to be particularly important for all participants in relation to attitude formation toward the social media fitness influencers in both studies. This finding corroborates previous research that identified trustworthiness as a central influencer characteristic.^19,25 Notably, in both studies, the influencers’ motivating power had a stronger impact on the intention to exercise than the attitude toward the influencer in both studies, indicating that perceived motivating power is a key variable for intentions to exercise for men and women. Following social cognitive theory,³⁶ users’ observations of the influencer's exercise behaviour might indeed increase motivation to work out, when the influencer's behaviour is evaluated as a promising behavioural option. A favourable attitude toward the influencer played a less important role.

While perceived motivating power emerged as a key variable for exercise intentions, several health and fitness variables showed effects on exercise intentions as well. Users’ lower self-reported health indicated higher intentions to exercise. Apparently, in line with the PMT,^55,56 if subjects consider their health to be rather poor and perceive the risk of possible diseases as a threat to their own health, they seem to be more inclined to engage in appropriate coping behaviours, that is, physical activity, with the goal of improving their health. Therefore, individuals appear to be aware of the fact that engaging in regular physical activity is beneficial for their health.⁷⁸

Contrary to what was predicted, physical fitness positively affected users’ intentions to exercise. It seems that a certain level of physical fitness is regarded as a prerequisite for intention to exercise. Thus, a more positive evaluation of one's own physical fitness is likely to positively affect intentions to exercise with a fitness video. By positively addressing users’ physical fitness, for instance, by complimenting the level of physical fitness already achieved as an outcome of regular physical activity, social media fitness influencers might actually be able to increase users’ motivation to exercise. If users then perceive their own fitness as higher, they are probably more motivated to continue exercising.

Body image exerted a significant negative influence on the motivation to exercise for female users, but not for male users; that is, a lower body image among women indicated higher intentions to exercise. Previous research has suggested that exposure to fit and thin fitness models might have both positive and negative health effects.^79,80 Negative effects can, for instance, result from comparisons with well-trained models that lead to dissatisfaction with one's own appearance. Positive effects can result when the models inspire individuals to pursue healthy lifestyle behaviours by imitating the models.⁶⁶ In our study, women with lower body image reported increased intentions to exercise after viewing the fitness influencer's video. Thus, while watching the influencer and the exercise, positive expectations in relation to their own body shape might have arisen. Apaolaza-Ibáñez et al.⁸¹ were able to demonstrate for cosmetics that women gain a positive feeling from using cosmetics, as they do everything they can for their appearance, and thus experience a feeling of relief from dissatisfaction with their physical appearance. This might also hold true for physical activity and one's own body image. By exercising with the influencer, women might experience less dissatisfaction with their body image. This effect certainly warrants additional research. For men, body image seems to be less relevant for working out with a fitness video, pointing in the same direction as prior research that indicates that men tend to pay less attention to social norms about their physical appearance than women.⁸⁰

Interestingly, physical fitness involvement had no effect among male users and a negative effect among female users on their exercise intentions. While previous research in other contexts found positive effects of physical fitness involvement on exercise intentions,^73,75 our results indicate that physical fitness involvement does not predict motivations to work out with social media fitness influencers. Possibly, physical fitness involvement is too general to predict higher and more positive intentions to exercise with a YouTube fitness influencer. It might be that users see other physical activities as better suited to achieve their physical fitness goals. Training involvement with YouTube fitness videos, a more specific measurement of involvement, yielded the expected positive results for both groups. Men and women who are involved in social media fitness influencer YouTube training sessions and consider such fitness videos to be important for their lifestyle indicated higher intentions to conduct their workouts with such videos.

In summary, our research suggests that social media fitness influencers who are evaluated as trustworthy, as experts, as attractive, and as having a high perceived motivating power might indeed be effective for increasing people's intention to engage in physical activity (i.e. exercising at home with social media fitness influencers) and thus might improve the physical condition and health of users via digital media. The motivating power of the influencer emerged as a central variable for predicting intentions to exercise, rather than the attitude towards the influencer. Physical fitness, (lower) health status, and training involvement with YouTube fitness videos were found to be important for men's and women's exercise intentions. Body image and physical fitness involvement had different effects on intentions to exercise reported by both groups.

Limitations and directions for future research

Beyond our study's important contributions, there are some limitations. One limitation of this research is that we employed German-speaking influencers in both studies. Future research might benefit from examining the developed social media fitness influencer model in an international context, for instance, in other countries with other social media fitness influencers. Additionally, our study analysed one fitness activity for one specific body area. Future research might want to examine other fitness activities and different body areas. Investigating gender differences in relation to other workouts could provide new discoveries.

This research identified perceived motivating power as a central influencing factor on intentions to exercise and analysed the influence of five important health and fitness variables on behavioural intentions. Beyond the source credibility variables trustworthiness, expertise and attractiveness, that we applied in our study, there might be additional source variables that have an influence on respondents’ attitude formations and the social media fitness influencers’ perceived motivating power. Apart from attitude and motivating power, there might be further variables that could influence users’ behavioural intentions.

As some of our results showed gender-related differences, it might be interesting to explore the reasons for these differences in more detail.

We focused on how social media fitness influencers might increase users’ intentions to become physically active. Future research might examine actual physical activity planning and behaviour^10,82,83 with social media fitness influencers’ video content. Future research should also elaborate on the causal relationship between the variables and whether the successful achievement of regular physical activity is also important to increase or maintain an individual's intention to exercise. We measured the overall fitness perception. Though Ortega et al.⁶¹ showed that the overall self-reported fitness is a good indicator for the actual physical fitness measured through objective laboratory tests, future studies might want to use objective tests for measurement and might also analyse different forms of physical fitness such as cardiorespiratory fitness, muscular strength, speed, agility and flexibility.^35,61 Future research may also apply methods such as experiments and conduct longitudinal studies.

Footnotes

Contributorship

JD researched literature and all authors jointly conceived the study. JD collected and analysed the data. SD and RT provided input into the study design and data analysis. JD wrote the first draft of the manuscript. SD and RT reviewed and edited the manuscript. JD integrated the feedback into the manuscript. All authors approved the final version of the manuscript.

Conflict of interests

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

Ethical approval

Ethical approval was not required for this study under European law. The study meets the criteria of the European General Data Protection Regulation (GDPR).

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The authors acknowledge the financial support by the University of Klagenfurt.

Guarantor

JD.

ORCID iD

Julia Durau

References

Carrotte

Vella

Lim

MSC

Predictors of “liking” three types of health and fitness-related content on social media: a cross-sectional study. J Med Internet Res 2015; 17: e205.

WHO. Global recommendations on physical activity for health. http://apps.who.int/iris/bitstream/10665/44399/1/9789241599979_eng.pdf (2010, accessed 22 August 2018).

WHO. Physical activity. http://www.who.int/news-room/fact-sheets/detail/physical-activity (2020, accessed 12 November 2021).

European Commission. Special Eurobarometer 472. Report. Sport and physical activity: survey requested by the European commission, directorate-general for education, youth, sport and culture and co-ordinated by the directorate-general for communication. http://ec.europa.eu/commfrontoffice/publicopinion/index.cfm/ResultDoc/download/DocumentKy/82432 (2018, accessed 12 November 2021).

Zhou

Krishnan

. What predicts exercise maintenance and well-being? Examining the influence of health-related psychographic factors and social media communication. Health Commun 2019; 34: 589–597.

Waters

Hawkins

. Awareness of health outcomes associated with insufficient physical activity and associations with physical activity intentions and behavior. J Health Commun 2018; 23: 634–642.

Döring

. Gesundheitskommunikation auf YouTube. Fallstudien zu 25 fitness- und lifestyle-Kanälen. [health communication on YouTube. Case studies on 25 fitness and lifestyle channels]. In: Schäfer

Quiring

Rossmann

, et al. (eds) Gesundheitskommunikation im gesellschaftlichen Wandel. [social change for health communication]. Baden-Baden: Nomos, 2015, pp. 105–118.

Lutkenhaus

Jansz

Bouman

. Tailoring in the digital era: stimulating dialogues on health topics in collaboration with social media influencers. Digit Health 2019; 5: 1–11.

Pilgrim

Bohnet-Joschko

. Selling health and happiness how influencers communicate on Instagram about dieting and exercise: mixed methods research. BMC Public Health 2019; 19: 1054.

10.

Kaushal

Keith

Aguiñaga

, et al. Social cognition and socioecological predictors of home-based physical activity intentions, planning, and habits during the COVID-19 pandemic. Behav Sci (Basel) 2020; 10: 31.

11.

Suciu

. Fitness goes to social media during COVID-19 outbreak. https://www.forbes.com/sites/petersuciu/2020/03/19/fitness-goes-to-social-media-during-covid-19-outbreak/?sh=4d1db8bf38ea (2020, accessed 15 January 2021).

12.

Mutz

Müller

Reimers

. Use of digital media for home-based sports activities during the COVID-19 pandemic: results from the German SPOVID survey. Int J Environ Res Public Health 2021; 18: 4409.

13.

Peng

C-T

T-Y

Chen

, et al. Comparing and modeling via social media: the social influences of fitspiration on male Instagram users’ work out intention. Comput Human Behav 2019; 99: 156–167.

14.

Oyibo

Adaji

Vassileva

. Social cognitive determinants of exercise behavior in the context of behavior modeling: a mixed method approach. Digit Health 2018; 4: 1–19.

15.

Oyibo

Vassileva

. Investigation of persuasive system design predictors of competitive behavior in fitness application: a mixed-method approach. Digit Health 2019; 5: 1–16.

16.

Baretta

Bondaronek

Direito

, et al. Implementation of the goal-setting components in popular physical activity apps: review and content analysis. Digit Health 2019; 5: 1–10.

17.

Hosseinpour

Terlutter

. Your personal motivator is with you: a systematic review of mobile phone applications aiming at increasing physical activity. Sports Med 2019; 49: 1425–1447.

18.

Rowley

Johnson

Sbaffi

. Gender as an influencer of online health information-seeking and evaluation behavior. J Assoc Inf Sci Technol 2017; 68: 36–47.

19.

Lou

Yuan

. Influencer marketing: how message value and credibility affect consumer trust of branded content on social media. J Interact Advert 2019; 19: 58–73.

20.

Freberg

Graham

McGaughey

, et al. Who are the social media influencers?: a study of public perceptions of personality. Public Relat Rev 2011; 37: 90–92.

21.

Sokolova

Perez

. You follow fitness influencers on YouTube. But do you actually exercise? How parasocial relationships, and watching fitness influencers, relate to intentions to exercise. J Retail Consum Serv 2021; 58: 102276.

22.

Ohanian

. Construction and validation of a scale to measure celebrity endorsers’ perceived expertise, trustworthiness, and attractiveness. J Advert 1990; 19: 39–52.

23.

Hovland

Weiss

. The influence of source credibility on communication effectiveness. Public Opin Q 1951; 15: 635–650.

24.

De Veirman

Cauberghe

Hudders

. Marketing through Instagram influencers: the impact of number of followers and product divergence on brand attitude. Int J Advert 2017; 36: 798–828.

25.

Djafarova

Rushworth

. Exploring the credibility of online celebrities’ Instagram profiles in influencing the purchase decisions of young female users. Comput Human Behav 2017; 68: 1–7.

26.

Amos

Holmes

Strutton

. Exploring the relationship between celebrity endorser effects and advertising effectiveness. Int J Advert 2015; 27: 209–234.

27.

Erdogan

. Celebrity endorsement: a literature review. J Mark Manag 1999; 15: 291–314.

28.

Wilson

Sherrel

. Source effects in communication and persuasion research: a meta-analysis of effect size. J Acad Mark Sci 1993; 21: 101–112.

29.

Kahle

Homer

. Physical attractiveness of the celebrity endorser: a social adaptation perspective. J Consum Res 1985; 11: 954.

30.

Priester

Petty

. The influence of spokesperson trustworthiness on message elaboration, attitude strength, and advertising effectiveness. J Consum Psychol 2003; 13: 408–421.

31.

Schouten

Janssen

Verspaget

. Celebrity vs. Influencer endorsements in advertising: the role of identification, credibility, and product-endorser fit. Int J Advert 2020; 39: 258–281.

32.

De Jans

Van de Sompel

De Veirman

, et al. #Sponsored! how the recognition of sponsoring on Instagram posts affects adolescents’ brand evaluations through source evaluations. Comput Human Behav 2020; 109: 106342.

33.

Egli

Bland

Melton

, et al. Influence of age, sex, and race on college students’ exercise motivation of physical activity. J Am Coll Health 2011; 59: 399–406.

34.

Hagger

Chatzisarantis

. Self-determination theory and the psychology of exercise. Int Rev Sport Exerc Psychol 2008; 1: 79–103.

35.

Caspersen

Powell

Christenson

. Physical activity, exercise, and physical fitness: definitions and distinctions for health-related research. Public Health Rep 1985; 100: 126–131.

36.

Bandura

. Self-efficacy: toward a unifying theory of behavioral change. Psychol Rev 1977; 84: 191–215.

37.

Bandura

. Social cognitive theory of mass communication. Media Psychol 2001; 3: 265–299.

38.

Casaló

Flavián

Ibáñez-Sánchez

. Influencers on Instagram: antecedents and consequences of opinion leadership. J Bus Res 2020; 117: 510–519.

39.

Bidmon

Terlutter

. Gender differences in searching for health information on the internet and the virtual patient-physician relationship in Germany: exploratory results on how men and women differ and why. J Med Internet Res 2015; 17: e156.

40.

Chalabaev

Sarrazin

Fontayne

, et al. The influence of sex stereotypes and gender roles on participation and performance in sport and exercise: review and future directions. Psychol Sport Exerc 2013; 14: 136–144.

41.

Azevedo

Araújo

CLP

Reichert

, et al. Gender differences in leisure-time physical activity. Int J Public Health 2007; 52: 8–15.

42.

Gonzàlez-Cutre

Sicilia

Aguila

. Interplay of different contextual motivations and their implications for exercise motivation. J Sports Sci Med 2011; 10: 274–282.

43.

Ninaus

Diehl

Terlutter

. Employee perceptions of information and communication technologies in work life, perceived burnout, job satisfaction and the role of work-family balance. J Bus Res 2021; 136: 652–666.

44.

Eisinga

Grotenhuis

Pelzer

. The reliability of a two-item scale: Pearson, Cronbach, or Spearman-Brown? Int J Public Health 2013; 58: 637–642.

45.

Pollay

Mittal

. Here’s the beef: factors, determinants, and segments in consumer criticism of advertising. J Mark 1993; 57: 99–114.

46.

Ajzen

. Constructing a theory of planned behavior questionnaire. http://people.umass.edu/∼aizen/pdf/tpb.measurement.pdf (2006, accessed 13 October 2017).

47.

Hooper

Coughlan

Mullen

. Structural equation modelling: guidelines for determining model fit. Electron J Bus Res Methods 2008; 6: 53–60.

48.

Fornell

Larcker

. Evaluating structural equation models with unobservable variables and measurement error. J Mark Res 1981; 18: 39–50.

49.

Von dem Knesebeck

Verde

Dragano

. Education and health in 22 European countries. Soc Sci Med 2006; 63: 1344–1351.

50.

Bergner

Rothman

. Health status measures: an overview and guide for selection. Annu Rev Public Health 1987; 8: 191–210.

51.

Feeny

Furlong

Boyle

, et al. Multi-attribute health status classification systems. Pharmacoeconomics 1995; 7: 490–502.

52.

Zajacova

Dowd

. Reliability of self-rated health in US adults. Am J Epidemiol 2011; 174: 977–983.

53.

Zajacova

Huzurbazar

Todd

. Gender and the structure of self-rated health across the adult life span. Soc Sci Med 2017; 187: 58–66.

54.

Rogers

. A protection motivation theory of fear appeals and attitude change. J Psychol 1975; 91: 93–114.

55.

Plotnikoff

Trinh

. Protection motivation theory: is this a worthwhile theory for physical activity promotion? Exerc Sport Sci Rev 2010; 38: 91–98.

56.

Bui

Mullan

McCaffery

. Protection motivation theory and physical activity in the general population: a systematic literature review. Psychol Health Med 2013; 18: 522–542.

57.

Plotnikoff

Rhodes

Trinh

. Protection motivation theory and physical activity: a longitudinal test among a representative population sample of Canadian adults. J Health Psychol 2009; 14: 1119–1134.

58.

Tulloch

Reida

Slovinec D’Angeloa

, et al. Predicting short and long-term exercise intentions and behaviour in patients with coronary artery disease: a test of protection motivation theory. Psychol Health 2009; 24: 255–269.

59.

Corbin

Pangrazi

Franks

. Definitions: health, fitness, and physical activity. Pres Counc Phys Fit Sports Res Dig 2000; 3: 1–11.

60.

Howley

Franks

. Health fitness instructor’s handbook. 3rd ed. Champaign, Ill.: Human Kinetics, 1997.

61.

Ortega

Ruiz

España-Romero

, et al. The international fitness scale (IFIS): usefulness of self-reported fitness in youth. Int J Epidemiol 2011; 40: 701–711.

62.

Sibley

Hancock

Bergman

. University students exercise behavioral regulation, motives, and physical fitness. Percept Mot Skills 2013; 116: 322–339.

63.

Avalos

Tylka

Wood-Barcalow

. The body appreciation scale: development and psychometric evaluation. Body Image 2005; 2: 285–297.

64.

Hausenblas

Fallon

. Exercise and body image: a meta-analysis. Psychol Health 2006; 21: 33–47.

65.

Boepple

Ata

Rum

, et al. Strong is the new skinny: a content analysis of fitspiration websites. Body Image 2016; 17: 132–135.

66.

Simpson

Mazzeo

. Skinny is not enough: a content analysis of fitspiration on pinterest Health Commun 2017; 32: 560–567.

67.

Pritchard

Cramblitt

. Media influence on drive for thinness and drive for muscularity. Sex Roles 2014; 71: 208–218.

68.

Fardouly

Vartanian

. Social media and body image concerns: current research and future directions. Curr Opin Psychol 2016; 9: 1–5.

69.

Holland

Tiggemann

. A systematic review of the impact of the use of social networking sites on body image and disordered eating outcomes. Body Image 2016; 17: 100–110.

70.

Kleemans

Daalmans

Carbaat

, et al. Picture perfect: the direct effect of manipulated Instagram photos on body image in adolescent girls. Media Psychol 2018; 21: 93–110.

71.

Zaichkowsky

. Measuring the involvement construct. J Consum Res 1985; 12: 341–352.

72.

Wiley

CGE

Shaw

Havitz

. Men’s and women’s involvement in sports: an examination of the gendered aspects of leisure involvement. Leis Sci 2000; 22: 19–31.

73.

Funk

Jordan

Ridinger

, et al. Capacity of mass participant sport events for the development of activity commitment and future exercise intention. Leis Sci 2011; 33: 250–268.

74.

Park

S-H

. Relationships between involvement and attitudinal loyalty constructs in adult fitness programs. J Leis Res 1996; 28: 233–250.

75.

Mahan

Seo

Jordan

, et al. Exploring the impact of social networking sites on running involvement, running behavior, and social life satisfaction. Sport Manag Rev 2015; 18: 182–192.

76.

Stokols

Shumaker

Martinez

. Residential mobility and personal well-being. J Environ Psychol 1983; 3: 5–19.

77.

Thomas

Chavez

Browne

, et al. Instagram As a tool for study engagement and community building among adolescents: a social media pilot study. Digit Health 2020; 6: 1–13.

78.

Warburton

DER

Nicol

Bredin

SSD

. Health benefits of physical activity: the evidence. CMAJ 2006; 174: 801–809.

79.

Grabe

Ward

Hyde

. The role of the media in body image concerns among women: a meta-analysis of experimental and correlational studies. Psychol Bull 2008; 134: 460–476.

80.

Strahan

Wilson

Cressman

, et al. Comparing to perfection: how cultural norms for appearance affect social comparisons and self-image. Body Image 2006; 3: 211–227.

81.

Apaolaza-Ibáñez

Hartmann

Diehl

, et al. Women satisfaction with cosmetic brands: the role of dissatisfaction and hedonic brand benefits. Afr J Bus Manag 2011; 5: 792–802.

82.

Rhodes

. Chapter five - the evolving understanding of physical activity behavior: a multi-process action control approach. In: Elliot

(ed.) Advances in motivation science. Elsevier, 2017; 4: 171–205.

83.

Gourlan

Bord

Cousson-Gélie

. From intentions formation to their translation into behavior: an extended model of theory of planned behavior in the exercise domain. Sport Exerc Perform Psychol 2019; 8: 317–333.