‘Sexy yes, obesity no’: the role of vloggers in influencing diet food purchase with self-openness moderation

Abstract

This study aims to evaluate the patterns by which vloggers influence consumers’ purchase intention for sustainable healthy diet food, to prevent obesity. In this process, a model was developed and validated through a collaboration between the lens of similarity-attraction theory (SAT) and the health belief model (HBM). This emphasized the effect of health behaviour similarity and perceived susceptibility-severity, which increased para-social interaction (PSI) and consumer purchase intention on TikTok social media viewers. Furthermore, the moderating effect of self-openness was examined on health behaviour similarity with PSI and consumers’ purchase intention. The role of gender and marital status was also assessed in supporting the relationship between health behaviour similarity (HBS) and PSI. In this case, a total of 438 validated samples were tested using the Covariance-based SEM approach. Based on the results, HBS and perceived susceptibility severity (PSS) were positively related to PSI and consumers’ purchase intention. PSI was also positively related to consumers’ purchase intention. Additionally, self-openness played a role in moderating the relationship between health behaviour similarity to PSI and purchase intention. These results subsequently considered the effectiveness of digital celebrity support for young people in sustainable healthy diet behaviour, especially married and unmarried women.

Keywords:

• Health behaviour similarity
• perceived susceptibility-severity
• PSI
• self-openness
• consumers’ purchase intention

REVIEWING EDITOR:

• Kaouther Kooli, Bournemouth University, United Kingdom

SUBJECTS:

• Advertising
• Consumer Behaviour
• Internet/Digital Marketing/e-Marketing
• Marketing Communications
• Marketing Management
• Marketing Research

1. Introduction

A person’s health is often influenced by diet and the surrounding environment (Hoek et al., 2017a). Based on the scaling up nutrition in 2021, a total of 68 million (35.54%) Indonesian adults were obese due to diet and inadequate activity (Republika.Co.Id, 2021). In previous decades, FAO and WHO have reportedly campaigned intensively for a healthy lifestyle through a sustainable diet, to contribute to food security and nutrition for present and future generations (F.A.O, 2010; Hoek et al., 2017b). In this case, food selection and healthy lifestyle maintenance are also trivially considered. However, most agricultural production processes often use excessive (inorganic) insecticides and food processing methods, which negatively affects health and the environment (Diaconeasa et al., 2019). Technological developments and social media also play important roles in influencing consumer decision-making related to a healthy and sustainable diet (Vermeir et al., 2020), although the most effective functions remain unknown (Choudhary et al., 2020; Zafar et al., 2021b). This important role is often realized through content creativity by vloggers, which are concerned with a healthy lifestyle through a sustainable diet (Sheikh et al., 2019). According to some previous literature, the content and credibility of digital celebrities played an important role in influencing the behaviours of followers (Garg & Bakshi, 2024; Hamilton et al., 2021; Xu et al., 2021). Using digital social interaction space, these celebrities are also interesting to be approached in real terms (Zafar et al., 2021a; Zhang et al., 2018). This has attracted a more in-depth analysis of the effectiveness of vloggers’ roles in influencing consumers, regarding a sustainable healthy diet. Therefore, this study aims to analyze the interpersonal interaction on sustainable healthy diet selections, through the consumption behaviour of vloggers on social media (Añaña & Barbosa, 2023; Ladhari et al., 2020; Meza & Yamanaka, 2020).

YouTube and TikTok are a medium for these digital users, to showcase their lifestyles and share product choices, such as the diet to maintain their slimness to attract followers. This ultimately produces the effect of advertising and selling the brand of a specific product (Babutsidze, 2018; Purwanto, 2021). Besides traditional celebrities, digital public figures are also presently considered more familiar and easy to be approached in real terms (Chapple & Cownie, 2017). This subsequently attracts more followers, which highly consider them friends (Gannon & Prothero, 2016). These conditions increase the interest of study experts in a sustainable healthy diet, regarding the determination of effective interventions on streaming media. Friendship is an important variable in social interaction, especially in the digital development of public influence. Based on previous literature, some followers believed that online friendship highly influenced individual behaviour or attitude (Hamilton et al., 2021). In specific situations, several vloggers and followers have reportedly been involved in decision-making processes, especially in the sustainable maintenance of slimness and fitness (Xu et al., 2021). These attitudes and behaviours are often exhibited by individuals, according to much previous literature (Kandel, 1978). This is in line with Hamilton et al. (2021), where the concept of a ‘journey partner (JP)’ was proposed. In this concept, direct or indirect interactions were often carried out in the decision-making processes on various purchasing procedures and preferences, especially in the healthy food type. Hamilton et al. (2021) also stated that the JP concept had two forms, namely proximal and distal partners. The proximal partners show that vloggers are physically close to their followers and have a strong bond in decision-making. For example, these digital celebrities commonly recommend new culinary delights that should be visited by customers. Meanwhile, distal partners emphasize virtual relationships, where unfamiliar people are always present in the decision-making process, through social media channels such as bloggers, TikTok vloggers, or anonymous author reviews (Xu et al., 2021). Social interaction aims to maintain relationships and motivate followers to remain loyal to vloggers (Argo & Dahl, 2020). For example, followers are often promoted to affiliate with vloggers, regarding similar food selections (McFerran et al., 2010). Based on this description, the presence of social channels facilitates the interactions of the new types of digital celebrities (distal partners), through streaming media. It also helps in the development of stronger bonds as real close friends. These vloggers subsequently have a strong social influence, by establishing an intimate relationship with followers in influencing consumer decision-making (Benito et al., 2020; Jin & Ryu, 2020; Sokolova & Kefi, 2020). Many studies have also reportedly evaluated the influence of traditional bloggers and celebrities on follower decision-making, regarding consumption (Keogh & Chadwick, 2019). However, the influence of vloggers on followers’ behaviour has only recently been assessed, especially the issue of a sustainable healthy diet on social media (Hill et al., 2020; Xu et al., 2021).

Although vloggers and bloggers have many similarities, their differences are often manifested within more substantial social features in vlogs (video content) than in blogs (text/image content) (Hill et al., 2020). Most of the previous literature also emphasizes the support of traditional celebrities than the digital types, especially in the purchase decision for sustainable healthy food (Schouten et al., 2020). This indicates the presence of a gap between the factors influencing the popularity of vloggers and the purchase decision of followers, regarding a sustainable diet (Hill et al., 2020). According to similarity-attraction theory (SAT), peer interaction in similar attitudes helped to predict interpersonal relationships (Byrne, 1961). Kandel (1978) also stated that similar behaviour led to the strong interaction, attraction, and deep interests. In this case, the followers having similar hobbies, lifestyles, and behaviours with vloggers often possessed stronger interaction and were more positively considered a symbol of their lifestyles (Berscheid & Walster, 1969; Xu et al., 2021). Furthermore, Schouten et al. (2020) proved that the popularity of digital celebrities was determined by the perception of similar behaviour (homophily) with followers, ultimately leading to a positive attitude toward the products purchased by their idols. In this case, ‘homophily’ emphasizes people with similar tendencies in various attributes (Lazarsfeld, 1964). Although perceived similarity behaviour plays an important role in the endorsement of digital celebrities, its mechanisms in consumers’ purchase intentions have still not been widely studied. This similarity mechanism includes the suitability of the product purchased and follower recognition of digital celebrities as endorsed artists on a product with a specific brand (Xu et al., 2021). It also depends on individual behaviours, regarding the self-openness to accept new products (Kumar et al., 2021). In this case, the behaviours of consumers often differ from each other when responding to digital celebrities, depending on their self-openness to accept new things (Harmel & Yeh, 2019). Although followers possess similar behaviour towards digital celebrities, they are still not necessarily receptive to their behaviours. Furthermore, the acceptance of behavioural change prioritizes personal values, which stimulates and promotes the desire to engage in the actions of digital celebrities (Harmel & Yeh, 2019; Kumar et al., 2021). Self-openness to change also motivates consumers to adopt the new products introduced by these celebrities (Schwartz, 1992; Xu et al., 2021), while directly influencing purchase behaviour (Claudy et al., 2015). Regarding some previous literature, the self-openness to accept new consumption products helped to develop a positive self-image, as well as evaluate the perceived benefits and risks (Hansen et al., 2018; Vieira et al., 2013). Therefore, this study aims to explore the similarity-attraction behaviour mechanism felt by followers on the purchase intention of a sustainable healthy diet product, regarding an attitude of self-openness to accept new things. Although they feel they have similar behaviour with vloggers, the understanding of a healthy lifestyle, physical susceptibility, and obesity severity are still low. This leads to the proposition of the health belief model (HBM), which is a theoretical lens for influencing the lifestyle of followers (Ling et al., 2019; Sheeran & Abraham, 1996). According to Ling et al. (2019) and Janz and Becker (1984), HBM described the success and effectiveness of a health program, regarding the perceptions of a person’s severity and susceptibility associated with behavioural responses. This explains that followers are increasingly aware of the importance of healthiness when they are susceptible to various disorders, such as obesity. These awarenesses often promoted their loyalty to the idols having similar healthy behaviour, subsequently leading to the gradual steps in making a purchase decision (Farooq et al., 2020). Additionally, HBM is a follower support mechanism for vloggers, through the self-awareness of health behaviour similarities, beliefs, and risks, especially in obese patients.

The SAT and HBM collaboration is the latest study in the mechanism of vloggers. This influences followers’ decision-making in the realm of a sustainable healthy diet. In this case, SAT proposes that people with similar behaviour are more likely to be attracted to an intimate relationship (Byrne, 1961). This intimacy in social media is unlike real public interaction due to its one-sidedness. It is also a kind of followers’ hallucination, which is often known as para-social interaction (PSI) (Purwanto, 2021; Xiang et al., 2016). In social media marketing, follower interaction commonly occurs when a response to the content created by vloggers is observed. This promotes the awareness of the brand or product used, subsequently influencing consumers’ purchase decisions (Ashley & Tuten, 2015; Kim & Johnson, 2016). Some previous studies also showed that the similarity perceived from social media positively influenced PSI (Haobin Ye et al., 2021; Xu et al., 2021). This is in line with Lee and Watkins (2016) and Sokolova and Kefi (2020), where behaviour similarity emphasized PSI and purchase intention on social media users. However, very few exploratory reports evaluated the effect of HBS (health behaviour similarity) and PSS (perceived susceptibility severity) on PSI and consumers’ purchase intention. In this case, the moderating role of self-openness was also prioritized in the relationship between HBS to PSI and purchase intention. Based on some previous literature, behaviour similarity with vloggers subsequently increased the attitude toward brands or products (Fleck et al., 2012; Yoo & Jin, 2013). From these descriptions, the effect of health behaviour similarity with vloggers is expected on the purchase intention for sustainable healthy food, through the TikTok streaming media platform. This is to enrich the SAT and HBM literature, to help managers select the best online marketing strategy and endorsement, as well as predict consumers’ behaviours. Regarding the final section, explanations and assessments emphasize the following, (1) The development of literature and hypotheses, (2) Methodology, (3) Empirical results and discussion, (4) Conclusions, (5) Theoretical and practical implications, and (6) Limitations and suggestions for subsequent analysis.

2. Literature development and hypotheses

2.1. Similarity-attraction theory (SAT)

The perception of interpersonal attraction prioritizes the desire of people to meet and know each other, as well as always be together (Byrne, 1961; Montoya & Horton, 2004). According to psychologists, this relationship is reportedly associated with similarity/homophily (Eyal & Rubin, 2003), which shows similar interaction in social status, beliefs, education, and hobbies. Moreover, similarity and homophily are often used interchangeably by several study experts (Ladhari et al., 2020). Irrespective of these conditions, homophily is more popularly introduced by virtual environment analysis, regarding the modern reports on social media marketing communication (Bu et al., 2022; Ladhari et al., 2020). Based on the impact of this phenomenon on purchase intention, some existing literature mainly focused on the relationship between influencers and followers, para-social interaction, as well as brand value and awareness (Ladhari et al., 2020; Lou & Yuan, 2019). In the similarity-attraction theory, relationships often prioritized similar attitudes, values, and beliefs, regarding the level of reward or benefit to be obtained (Byrne, 1961; Newcomb, 1956). When a person interacts with a public figure, appearance similarity often leads to the development of pride, by strengthening the consistency of perceptions and opinions. In this case, the possession of similar values, cognitive processes, and emotional commitments also complements the need for social identity, leading to the development of a useful interaction process (Byrne, 1961; Zhang & Bloemer, 2008). For Byrne (1971), several individuals are more likely to be attracted to one another due to similarities. Byrne and Griffitt (1973) also stated that interpersonal attraction was a person’s attitude towards others, which commonly leads to friendship and romantic relationships. This attraction theory is frequently used to explore the effect of similarity on human interaction. For example, the similarity of a customer toward other people directly or indirectly promotes brand recognition through attraction, as the antecedent to a deep relationship with the product (Elbedweihy et al., 2016).

According to Kwon et al. (2016) and Xu et al. (2021), the perceived similarity to other buyers positively influenced satisfaction and subsequent behaviour. Benbasat et al. (2020) also explained the importance of similarity in humans and recommended the interaction as a repair agent. Furthermore, SAT explains that people with similar attitudes are more likely to associate and bond with themselves, leading to an assessment of the relationship’s closeness (Schlager et al., 2018). In this case, one’s attraction is likely to change the attitude of others with a persuasive influence (Cohen et al., 2018; Ruijten, 2021). The perception of social attraction is also an important basis for decision-making. For example, the similarity between consumers and their idols (characters, artists, celebrities) often predicts diverse affective (engagement), cognitive (evaluation of people and products), relational (friendship), and behavioural (purchase) aspects (Brocato et al., 2012; Ilmarinen et al., 2016). In Rogers and Bhowmik (1970), the basic principle of human communication focused on the exchange of messages. The interactions between the source and the receiver are also observed due to the possession of similar behaviours. This is in line with McCroskey et al. (1974), where homophily or similarity was used to enhance the effect of communication. In this case, similarities emphasized easier conveyance of messages to the potentially influenced audience, leading to more attraction toward the content delivered.

2.2. Health belief model (HBM)

HBM shows that people’s beliefs about health problems, self-efficacy, as well as perceived action benefits and barriers, explain adequate or inadequate involvement in health-promoting behaviours (Janz & Becker, 1984; Sheeran & Abraham, 1996). This behavioural theory emphasizes the aspects of belief and perception, where the existence of a good or bad opinion originates from the knowledge, experience, and information obtained by the individual concerned (Glanz et al., 2008). The theory is also developed from a behavioural concept, which assumes that a person’s behaviour depends on the following, (1) The value assigned by the individual on a goal, and (2) An individual’s attitudinal estimate of the likelihood to achieve a goal. According to some previous studies, this construct was a significant predictor of health motivation (Bish & Michie, 2010; Farooq et al., 2020). Perceived severity is defined as an individual’s assessment of a situation’s complexity, regarding health consequences (Ling et al., 2019). Meanwhile, perceived susceptibility is an assessment of the likelihood of being vulnerable in a specific situation (Ling et al., 2019). The similarity to other people also considered the development of motivation by public figures, to enable behavioural adoption regarding a healthy lifestyle. This condition emphasizes the avoidance of health severity and susceptibility (Xu et al., 2021). Besides its use in healthy behavioural change interventions, HBM has also been recently used in social media-influenced consumption reports (Farooq et al., 2020; Xu et al., 2021; Zeng et al., 2010). Therefore, this study uses the concept of perceived susceptibility-severity to conceptualize threat perception (health belief) as a significant predictor of health motivation, to primarily maintain slimness through a sustainable hygienic diet (Bish & Michie, 2010; Xu et al., 2021).

2.3. The relationship between healthy behaviour and para-social interaction

According to McCroskey et al. (1974), similarity had two dimensions, namely behaviour and motive. Behaviour similarity reflects the extent to which a person perceives that other people possess similar attitudes. Meanwhile, motive similarity shows the extent to which a person assumes that other people have identical social features. A previous study also stated that behaviour and motive similarities significantly and insignificantly affected emotional closeness and purchase, respectively (Ladhari et al., 2020). Moreover, Li and Du (2011) stated that behaviour similarity played an important role in predicting social influence support, as well as consumers’ attitudes and purchase intention. This construct subsequently increases interpersonal interaction and strengthens emotional bonds between digital celebrities and followers (Zafar et al., 2021a; Zhang et al., 2018). When individuals think that members of a community are similar to themselves, a more positive attitude is observed toward participatory willingness (Hook et al., 2020). The followers with the perception of having similarity to digital celebrities are also more likely to positively comply and participate. Irrespective of these conditions, behavioural interaction and participation are still not as significantly related to vloggers (the illusion of facial interaction) as the communication between the audience and actors, as in TV interviews, which only form a one-sided relationship (Horton & Wohl, 1956). Subsequently, it is applied to social media platforms to express the one-sided closeness between followers and vloggers as public influencers (Xu et al., 2021).

Based on the similarity-attraction theory, identical people are more likely to build emotional bonds, which leads to favourable evaluations and intimate relationships (PSI) (Schlager et al., 2018). This was in line with Lee and Watkins (2016), where PSI affected various factors such as brand attitude and purchase. Behaviour similarity (BS) with vloggers also influenced PSI, which subsequently affected the brand perception and purchase intention. Xiang et al. (2016) then proved that the attitude of BS with other social media users positively influenced PSI and then affected impulse purchases. In Hung (2014), followers with similarities to the behaviours of digital celebrities also impacted PSI and positive behavioural intention toward the endorsed brand. Therefore, this study proposes that followers feeling more similar to their idols (digital celebrities) regarding healthy lifestyle behaviour, and are more likely to build social interaction and have purchase intention towards the products or brands bought. Based on these conditions, the following hypotheses are proposed,

H_1: Health behaviour similarity is positively related to para-social interaction.

H_3: Health behaviour similarity is positively related to purchase intention.

2.4. Relationship of perceived susceptibility-severity to PSI

Understanding individuals’ health beliefs emphasize an awareness of the susceptibility and severity of obesity to sustainable wellness. This leads to the adoption of the health belief model (HBM), which motivates individuals toward living a healthy life through the right diet. The model also relates demographic variables and psychological characteristics to effective and cognitive states, such as perceived susceptibility severity, health motivation, and behavioural response benefits (Sheeran & Abraham, 1996). Additionally, health professionals use HBM to investigate the design and intervention of medical behavioural change (Eldredge et al., 2016; Orji et al., 2012), including the process of healthy-living attitudes due to the compliance with traditional and digital celebrities (Xu et al., 2021). To conceptualize the perception of the health threats related to obesity, this study uses the perceived susceptibility-severity variable. This construct is subsequently determined as a significant predictor, to motivate people toward sustainable healthy diets through suitable methods (Farooq et al., 2020; Xu et al., 2021). Perceived severity is defined as an individual’s assessment of a situation’s complexity, regarding health consequences (Ling et al., 2019). Meanwhile, perceived susceptibility is an assessment of the likelihood of being vulnerable in a specific situation, such as obesity (Ling et al., 2019; Xu et al., 2021). In this case, the people feeling the susceptibility and severity of obesity are more motivated to follow the behaviours of their idols, concerning the consumption of sustainable healthy food products (Gong & Li, 2017; Pradhan et al., 2016).

For example, a recent study showed that the severity-susceptibility perception of obesity was a significant predictor of a sustainable healthy diet, regarding the products purchased by celebrities (Farooq et al., 2020; Xu et al., 2021). Huang et al. (2015) also found that emotional and physical closeness increased the likelihood of sharing information, as well as the desire to interact with many people or public figures. The mechanism responsible for this observation is known as the perceived relevance of the situation, for example, vloggers are more likely to digitally share information when the content is relevant to them (Chua & Banerjee, 2018). This explains that when people perceive a higher degree of severity and susceptibility, they are more fanatical about the behaviour of the vloggers purchasing sustainable healthy diet products. These were in line with Xu et al. (2021) and Kahle and Homer (1985), where the perceived alignment of the products used by vloggers increased the followers’ desire to build a strong interaction relationship. In this case, the perception of celebrity attraction increased with the elevation of the product-vlogger alignment with followers or vice versa (Yoo & Jin, 2015). Another study also showed that low product-vlogger alignment negatively impacted digital celebrity image, with followers increasingly reluctant to engage in interaction (Park & Lin, 2020). This proved that people feeling susceptible and severely over an unwanted situation often attempted to align the use of the product with public figures, to defuse disturbing conditions (Keogh & Chadwick, 2019; Ladhari et al., 2020; Xu et al., 2021). Additionally, the similarity of people’s product use with their idols (vloggers) commonly increases the establishment of a personal relationship. This was in line with Elbedweihy et al. (2016), where the similarity of digital celebrities with followers had a stronger interpersonal attraction. In this process, the perceived attraction of the vloggers increased PSI when the product-vlogger alignment with followers was higher. Based on these descriptions, the following hypothesis is proposed,

H₂: When individuals understand the susceptibility and severity of obesity, they are more motivated to adapt to products-vloggers and establish para-social interaction with vloggers.

In Goodwin et al. (2009), the food purchase by consumers increased due to being influenced by personal circumstances and the environment. Joshi and Rahman (2015) also identified that emerging concerns such as obesity majorly affected consumers’ behaviour. This indicated that perceived severity and susceptibility influenced other diseases in obese people, with consumers often reacting by buying the products frequently consumed by their idols. Based on these descriptions, the following hypothesis is presented,

H₄: When individuals understand the susceptibility and severity of obesity, they are more motivated to purchase healthy diet food.

2.5. Relationship of PSI to consumers’ purchase intention

A previous empirical study stated that followers often established a community, as well as shared values, beliefs, and interests (Nambisan & Watt, 2011). This showed that the two-way relationship between followers and vloggers was real, although it was only a cyberspace illusion (Labrecque, 2014; Lee & Watkins, 2016). Kim et al. (2015) also found that social media use was positively related to PSI. For example, young women commonly follow all activities of vloggers, ranging from habits and lifestyle to fashion, which generate self-confidence (Djafarova & Rushworth, 2017). Moreover, Lee and Watkins (2016) stated that the PSI between vloggers and followers positively influenced the emotions to buy products and want similarity to their idols. Other literature also stated that PSI was positively related to purchase intention, due to the suitability of healthy living behaviour (Choi & Lee, 2019; Chung & Cho, 2017; Hwang & Zhang, 2018). This indicated that the more users perceive similar health-related behaviour to vloggers, the greater they are motivated toward increasing the frequency of interaction, consuming longer time, and elevating higher PSI. These activities subsequently increased the purchase intention of the recommended products. Based on these descriptions, the following hypothesis is proposed,

H₅: Para-social interaction positively affects consumers’ purchase intention.

2.6. Moderating effect of self-openness

Self-openness is an individual’s value, which promotes the desire to accept change and engage in independent action (Harmel & Yeh, 2019). This personal value consists of stimulation and self-direction (Bilsky & Schwartz, 1994). According to some previous literature, self-openness was an individual behaviour, which easily accepted new products and influences purchase attitude (Claudy et al., 2015; Schwartz, 1992). Concerning the changes in consumption patterns, self-openness also enables the realization of a positive identity and self-image, regarding the sustainable diet obtained from digital celebrity information (Vieira et al., 2013; Xu et al., 2021). Furthermore, Hansen et al. (2018) stated that the openness to change was influenced by the perceived benefits and risks of using a product or service. Although digital celebrities are very strong in influencing the attitude of followers, the acceptance still depends on the individuals’ behaviour toward self-openness to change (Kumar et al., 2021). Irrespective of the inexistence of a solid theoretical basis, some previous studies still emphasized consumers’ benefit and risk perceptions of using sustainable diet products (Kumar et al., 2021; Xu et al., 2021). High self-openness is also favourable to the presence of new brands and positively affects purchase intention (Chaudhuri & Holbrook, 2002; Mauri & Minazzi, 2013). Therefore, this study believes that a moderating effect of high self-openness is present in the relationship between HBS and the desire to build a higher PSI. High self-openness also enables a stronger relationship between HBS and purchase intention, for sustainable healthy diet food. In this condition, a person’s self-openness to something new (product, brand, service, person) prioritizes the positive or negative evaluation (Bhandari & Rodgers, 2018; Reimer & Benkenstein, 2016). Based on these explanations, this study posits that followers with higher self-openness to change are more likely to develop social interaction and purchase sustainable healthy diet products. This focuses on individuals with greater self-openness, which are more likely to assess the obtained benefits and risks. These activities subsequently lead to self-behaviour, such as accepting or rejecting the presence of something new (Kumar et al., 2021) (Figure 1). Regarding these descriptions, the following hypotheses are proposed:

Figure 1. Conceptual model.

H₆: Self-openness to change moderates the relationship between health behavior similarity to a para-social interaction development intention.

H₇: Self-openness to change moderates the relationship between health behavior similarity to consumers’ purchase intention.

3. Methods

3.1. Participants

The analytical samples are TikTok viewers, considering the popularity of the platform among young adults in Indonesia, which is highly estimated at 94% or 190 million users (Katadata.co.id, 2021). The vloggers with sustainable healthy eating content were also emphasized, as the trending content throughout 2021 focused on the feeding habits of digital celebrity, TheSantoso, which has been watched by more than 28 million users (TikTok What’s Next Report, 2022). Furthermore, the millennial generation (ages 18–29) was selected as the sample, due to having the highest number of social media accounts (McCormick, 2016). Data were also obtained through an online survey for users of the Indonesian streaming media platforms, which are interested in sustainable healthy living. In this case, a total of 600 participants were invited through social media and personal communication. These participants were instructed to recall their last viewed content, regarding healthy diets and experiences recommended by digital celebrities. From this context, those without any social experience and interest in answering the questionnaire were excluded, leading to a valid sample of 438 TikTok users (73.00% response rate). Based on gender and age, 306 (69.86%) and 387 (88.35%) participants were female and 18–25 years, respectively. This was accompanied by those with undergraduate and graduate education levels at 72.88% and 27.66%, respectively. Additionally, the participants actively following celebrity content with more than two views were 87.91%.

3.2. Measurement

A questionnaire digitally designed through Google Forms was used for data collection, where screening questions were initially tested on 30 participants to detect possible errors such as ambiguous and offensive items (Cooper & Schindler, 2013). Moreover, the participants were instructed to answer questions related to health behaviour similarity, perceived susceptibility severity, openness to change, para-social interaction, and purchase intention. Based on several previous studies, a total of 18 statement items were adopted in this analysis (Table 1). In this case, the measurements of health behaviour similarity and perceived susceptibility-severity contained 3 and 4 items, which were adopted from Xu et al. (2021) and Ladhari et al. (2020), as well as Bish and Michie (2010), Farooq et al. (2020), and Ling et al. (2019).

Table 1. Instrument, loading factor CRs and AVEs.

Construct	Item	Loading/weight	Cronbach’s a
Health behaviour similarity CR: 0.95, AVE: 0.79	The vloggers’ healthy diet behaviour is the same as mine.	0.83	0.87
	Vloggers share many of the same healthy diet behaviour as me.	0.81
	Vloggers have thoughts and ideas about a healthy diet similar to mine.	0.78
Perceived susceptibility-severity CR: 0.91, AVE: 0.85	I believe that obesity is another health risk condition.	0.75	0.83
	When my condition is very susceptible to disease, then I need to perform a sustainable healthy diet as performed by vloggers.	0.74
	I feel how severe the impact of obesity is.	0.71
	Since obesity is dangerous to infectious health, I adhere to a sustainable healthy diet.	0.73
Para-social interaction CR: 0.96, AVE: 0.83	I am looking forward to following the vloggers on their TikTok channel.	0.72	0.85
	I think vloggers are like my best friend for real.	0.75
	When the vloggers show up on another platform’s channel, I will watch that TikTok.	0.73
	When I watch the vloggers, I feel I am part of the group.	0.69
Self-openness CR: 0.91, AVE: 0.81	I am very open to new things.	0.81	0.82
	I want to be the first to try new things in the market.	0.77
	As long as my idol thinks it is good then I will follow it	0.79
	I do not want to be left behind on new products in the market.	0.72
Purchase intention CR: 0.93, AVE: 0.77	Because I feel there is a similarity in my behaviour, I am motivated to buy a product consumed.	0.71	0.84
	TikTok videos of the vloggers will help me make a healthy food purchase decision.	0.72
	I am interested in buying healthy diet food products consumed by vloggers.	0.71

The measurements of PSI, purchase intention and openness to change were also measured by 4, 3, and 4 items, which were adopted from Schlager et al. (2018); Lee and Watkins (2016), and Xiang et al. (2016), Choi and Lee (2019), Hwang and Zhang (2018), and Chung and Cho (2017), as well as Claudy et al. (2015) and Kumar et al. (2021), respectively. Additionally, all the items were measured using a 7-point Likert scale of SD (1; Strongly Disagree) and SA (7; Strongly Agree).

3.3. Analysis technique

The AMOS and covariance-based SEM methods were used for data analysis, due to being very popular and widely used by various experts (Dhir et al., 2021; Kaur et al., 2021; Kumar & Yadav, 2021). This analysis was carried out in two steps, namely (1) Testing the reliability and validity of the measurement model, and (2) Testing the structural model to verify the hypotheses. The measurement of the construct validity (CV) model was also evaluated by examining the measurement of model convergence and DV (discriminant validity) (Hair et al., 2011). In this case, various criteria were used for CV, such as the following, (1) The outer loading on the SEM, which should be ≥ 0.7, (2) The Average Variance Extracted (AVE) of each structure, which should be ≥ 0.5, and (3) The composite reliability (CR), which should be ≥ 0.7.

4. Results

4.1. Confirmatory factor analysis

This analysis was used to measure the validity of the structure. Firstly, the adequacy of each scale in the question items was tested, indicating the observation of a significant standard of Average Variance Extracted (AVE) with convergent validity. Based on Table 1, all items showed sufficient convergent validity and an AVE of 0.77–0.85. Meanwhile, construct reliability was tested using Cronbach’s alpha, whose values ranged from 0.82 to 0.87 (Table 1). This indicated that the measurement model had good reliability. Furthermore, the validity of the model was tested by comparing the square root of the AVE and the construct correlation (Table 2). From the results, the square root of the AVE on all constructs was more than 0.70, indicating sufficient discriminant validity (DV) and can be continued for the hypothesis test (Fornell & Larcker, 1981). Secondly, the fit of the model was tested, where all hypotheses were accepted with a GFI (goodness-of-fit-index) and CFI (comparative fit index) of 0.941 and 0.917, respectively. A Chi-square of 616.312 (100 df, p = 0.000) was also obtained with no standard residual of more than 2.0, indicating that the overall model fit was acceptable (Hair et al., 2011).

Table 2. Correlation matrix and discriminant validity.

	Health behaviour similarity	Perceived susceptibility-severity	Para-social interaction	Purchase intention	Self-openness
Health behaviour similarity	0.813
Perceived susceptibility-severity	0.251	0.872
Para-social interaction	−0.031	0.113	0.784
Purchase intention	0.212	0.121	0.021	0.761
Self-openness	0.411	0.211	0.521	−0.327	0.827
Average variance extracted (AVE)	0.328	0.477	0.765	0.124	0.342
Mean	0.453	0.344	0.876	0.432	0.512
Standard deviation (SD)	0.244	0.081	0.043	0.021	−0.241
Model fit: χ2 = 2,145, p < .01, df = 1,409; CFI = 0.913; TLI = 0.904; RMSEA = 0.077; SRMR = 0.07

^aThe square roots of AVE for each construct are presented in bold on the diagonal of the correlation matrix.

^bAVEs of formative indicators are not applicable.

^cNote. N = 438.

Thirdly, a test was carried out using common method variance (CMV) on the data. In this case, two approaches were used to test CMV, namely (1) The commonly used single-factor test, where no single factor explained a substantial portion of the total variance in the data, and (2) The general method-factor test, where the model variance was smaller than the substantive variance (ratio 57:1) (Podsakoff et al., 2003). This small variance indicated that CMV was not the main problem in the data.

4.2. Structural model and hypothesis test

The structural model used to test the hypotheses had a goodness of fit, where health behaviour similarity and perceived susceptibility severity was positively related to PSI (H1: = 0.21, p < 0.001; H₂: = 0.57, p < 0.004), respectively. Health behaviour similarity, perceived susceptibility severity, and PSI were also positively related to purchase intention (H₃: = 0.19, p < 0.001; H4: = 0.23, p < 0.003; H₅: = 0.47, p < 0.001), respectively. Based on these results, all the proposed hypotheses were accepted, as shown in Table 3 and Figure 2.

Figure 2. Hypotheses test.

Table 3. Hypotheses test results.

	Path	B	t statistic	p values
H1	HBS → PSI	0.21	7.635	0.001
H2	PSS → PSI	0.57	6.125	0.004
H3	HBS → PI	0.19	9.711	0.001
H4	PSS → PI	0.23	11.496	0.003
H5	PSI → PI	0.47	21.231	0.001

Note: Health behaviour similarity = HBS, purchase intention = PI, Perceived susceptibility-severity = PSS, Para-social interaction = PSI.

4.3. Moderation analysis

The moderating role of openness to change on health behaviour similarity, PSI, and purchase intention was analyzed using PROCESS macro. Regarding the results, self-openness determined the followers’ attitudes to establish a relationship with digital celebrities and continue their purchase intentions. This explained that self-openness positively moderated the relationship between health behaviour similarity, PSI, and purchase intention, subsequently supporting H6 and H7 (Table 4).

Table 4. Moderation analysis of openness to change.

Path	Coeff	Ze	T	P	LLCI	ULCI	Remarks
HBS-PSI	0.311	0.07	1.39	0.02	0.001	1.015	Support
HBS-PI	0.421	0.06	2.57	0.01	0.004	1.321	Support

Note: Health behaviour similarity = HBS, purchase intention = PI.

Based on Figure 3(a), the high and low self-openness groups strengthened and degraded the positive relationship between HBS and PSI, respectively. Similarly, the relationship between HBS and consumers’ intention was strengthened by high self-openness as shown in Figure 3(b).

Figure 3. (a) The moderating effect of self-openness. (b) The moderating effect of self-openness.

4.4. Post hoc analysis

This post hoc test analysis was used to examine the presence or absence of differences, regarding the effect of marital status and gender on HBS and PSI. In this process, marital status and gender were allowed to interact with all predictors of HBS. The results are shown in Table 5, where the PSI predictors, as interaction terms, were all significant. For health behaviour similarity with gender and marital status (p < 0.05), the terms also significantly affected the relationship between HBS and PSI.

Table 5. The interaction effect of marital status and gender on health behaviour similarity.

Predictors	Predicting health behaviour similarity
Gender	0.03*
Male	0.07* ns
Female	0.02*
Marital status	0.00*
R^2	67%

Based on Table 5, marital status and gender supported the relationship between behaviour similarity and PSI. However, one interaction term was not significant for the male gender, indicating that the effect of HBS decreased for men. This proved that men were less concerned with obesity than women, leading to a decreased desire to become loyal followers or build interaction with digital celebrities. From these results, men were less fanatical about vloggers, regarding sustainable healthy diet food. This was in line with Gwyther and Holland (2012), where men had different self-regulation abilities on appearance than women, which opted to maintain their slimness. This indicated that men were more indifferent to obesity as a health risk. Therefore, the effect of gender on PSI (p < 0.05) was strengthened, where women developed a more intimate relationship with vloggers. As a health problem, gender illustrated the confidence level in the risk of obesity, with women specifically more disciplined towards a healthy diet, to maintain body slimness and sustainable hygiene.

5. Discussion and conclusion

The development of social media, especially TikTok, has increased sharply and is used as a public space to interact about various interests (Schouten et al., 2020; Xu et al., 2021). This study aimed to explore the behaviours of media users making healthy-living decisions through a sustainable diet influenced by digital celebrity TikTok content. In this experiment, the results obtained supported all proposed hypotheses. HBS also played an important role in making a healthy life decision through a sustainable diet informed by vloggers, regarding various types of content. Based on Figure 2, HBS was positively related to purchase intention for sustainable healthy diet food. However, it depended on one’s self-openness attitude toward the interaction with digital celebrities and purchase intention. Regarding an important result, self-openness moderated the relationship between HBS, PSI, and purchase intention. It also compared the perceived benefits to the potentially accepted risks. When individuals feel these benefits are greater, they subsequently believe and follow the recommendations provided by digital celebrities (Hansen et al., 2018; Kumar et al., 2021). This result was very helpful in informing consumers, public health organizations, and producers to persuasively convey hygienic messages and promote healthy food for a sustainable diet. It was also helpful in educating the community, regarding the positive impact of food consumption patterns.

Based on these results, followers often spend more time watching TikTok shows when they believe in having similar behaviour with vloggers. They also participate in interactions and develop friendships with celebrities. This was the first-ever result showing that HBS motivated followers to participate in vlogs as a social attraction. In this case, active followers positively participated in driving PSI and influencing purchase intention (PI). Furthermore, a moderating role was found between HBS, PSI, and PI. This was in line with Kumar et al. (2021), Lee and Watkins (2016), and Xu et al. (2021), where the relationship between health behaviour similarity and celebrity social interaction was determined by vlogger credibility and brand trust, which ultimately increased purchase intention. The result was also beyond the knowledge of previous literature, where increasing social interaction was influenced by followers’ health awareness. This was carried out through the perceived susceptibility-severity of the people willing to adopt the health belief model (HBM) theory. Regarding this theory, obese people are often severely susceptible to various disorders due to their respective conditions (Farooq et al., 2020; Ling et al., 2019). This confirmed that the perceived susceptibility severity increased the interaction between followers and vloggers. In this process, the followers aware and worried about their situation were found to increase their PSI. They also hoped to continuously increase interaction with vloggers, as well as adhere to health-related messages and recommended product purchases. This subsequently helped vloggers with the endorsement of brands, to become the advertisements accepted by their followers. These results were consistent with various previous reports, where the individuals with higher perceived susceptibility severity were more compliant with vloggers. They were also willing to build social interaction with digital celebrities, which ultimately purchased and recognized their advertisements (Evans et al., 2017; Gong & Li, 2017; Reijmersdal et al., 2016).

5.1. Theoretical implications

This study provided important contributions to the existing literature. Firstly, expansions were provided on the present literature, by testing and validating the models collaborating between SAT and HBM. The behaviour similarity attitude was also provided as an important predictor of purchase intention, customer loyalty (Lee & Watkins, 2016; Sakib et al., 2020), and brand image (Liu et al., 2019). Furthermore, interpersonal interaction plays an important role for vloggers and traditional celebrity influencers (bloggers). Irrespective of these results, the analysis emphasizing behaviour similarity with vloggers on social media platforms moderated by self-openness was still rarely conducted. Using the lens of similarity-attraction theory and the health belief model, the empirical gap was filled, with the effect of HBS (and perceived susceptibility-severity increasingly and digitally clarified on consumers’ purchase intention. In this case, the effect of HBS and PSS mediated by PSI on the purchase intention were tested for sustainable healthy food, regarding the prevention of obesity. The effect of HBS was also moderated by self-openness on PSI and consumers’ purchase intention. Moreover, the presence or absence of differences in the HBS effect on PSI and purchase intention was tested, regarding gender and marital status. Therefore, the results obtained contribute to the similarity-attraction literature and perceived susceptibility-severity in the health care influenced by digital celebrities.

Secondly, an understanding of self-openness was provided in the relationship mechanism between HBS to PSI. In this case, the interpersonal attraction was stronger when followers feel they have similar healthy behaviour with vloggers and provide greater benefits. This indicates that the existence of HBS promotes them to interact with their idols and follow the pre-purchase decision phase structuring journey partner activities in social media. Thirdly, perceived susceptibility-severity was the motivation of followers to increase interaction with digital celebrities as a factor in their self-awareness, regarding the dangers of obesity to health and personal appearance. Although several previous studies traditionally evaluated HBS and PSS, some analyses on HBS, healthy food, and sustainable diet were still limited. Therefore, this study contributes to the online purchase literature through an empirical analysis, regarding sustainable healthy food.

5.2. Managerial implications

These results have important implications for healthcare providers/public health organizations (WHO, FAO) and food product marketers, by incorporating vlog content into their marketing communication strategies. Firstly, producers are able to decide the selection patterns of vloggers in their segment, to maintain and develop communication strategies for the most influential and viral types of digital celebrities. This indicates that the vloggers related to healthy lifestyles influence consumers’ purchase intention, through HBS, PSS, and PSI. Therefore, marketers should assess the strength of the vloggers’ social appeal when developing the content related to healthy food products. This is to increase interaction, affective closeness, and emotional bonding with potential followers, as well as generate a positive attitude in attracting consumer buying interest. Secondly, managers are able to understand the role of consumer attitude regarding self-openness in accepting new things, which are likely to provide support for interactions with digital celebrities and purchase behaviour. The HBS with these celebrities should also consider the suitability of the products used by vloggers with followers. Intuitively, the existence of vlogger-follower product suitability increase trust and develop a favourable attitude. This is to build stronger social interactions and severally elevate digital celebrity trust in purchase decision, advertising recognition, and others. Therefore, the target is better and more effective in increasing the sales of sustainable healthy diet products.

5.3. Limitations and future directions

Firstly, a social media platform increasingly popular for the millennial generation was used as the object. This indicates that the results are likely to differ when the model is retested in different environments and cultures, for example, in developing countries with the largest number of millennials using TikTok. Based on this limitation, academics should subsequently and futuristically test and verify these results in different cultural backgrounds.

Secondly, the social interaction between followers and vloggers was only used in influencing consumers’ purchase intention with the antecedents of behaviour similarity and health perception. However, self-openness in purchase decision and recognition has not been examined. This shows that subsequent studies are recommended in determining the presence or absence of followers’ recognition for their idols. These are important for marketers in selecting digital star celebrities for their product advertisements. The development of PSI on social media is also a unique content ensuring more public attraction, leading to the willingness of consumers to continue watching and actively participating in vlog contents. Therefore, future studies need to implement a broader approach to analyze the development and dynamics of PSI’s effect on various consumers.

Thirdly, this study only emphasized a TikTok account, leading to the recommendation of subsequent comparative analysis with other social media presently popular among young people. Fourthly, since this study is limited to health-related vlogs, the consideration of its impact on other types of natural personal care products and consumer motivation is necessary for broader outcomes, regarding the utilitarian involving price and discount as well as hedonic consisting of experience and design motives to be explored.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by Madika Foundation.

Notes on contributors

Purwanto P.

Purwanto P. is a PhD in Marketing Management from the Airlangga University of Surabaya, Indonesia. The research interests focus on Marketing Strategic, Digital Marketing, Consumer Behavior, Business Strategic and Management information System. Additionally, he is skilled in research Small Medium Enterprises and innovation in both theory and practice.

Rahayu S.

Rahayu S. is a dedicated Senior Lecturer bringing over 17 years of experience to the realm of higher education. Her academic journey is anchored by a Ph.D. in Marketing with a specialized focus on consumer Behavior from Tujuh Belas Agustus 1945 University of Surabaya. Her academic interest areas are Marketing Strategic, Digital Marketing, consumer behavior, statistic, business research, social media marketing.