Introduction
WeightWatchers (WW) offers consumers an array of products and services to assist in weight loss, fitness, and the attainment of “more active and more fulfilling lives” (WeightWatchers [WW], n.d.). On February 7th, 2018, WW announced that teenagers between 13 and 17 years old could obtain a free membership of WW’s diet program for 6 weeks (Scritchfield, 2018). WW promotes its diet programs by providing information about the potential for significant reductions in waist circumference and fat mass, along with improvements in blood sugar, cholesterol levels, sleeping, and mental health. However, following the WW teen program announcement, the American Academy of Pediatrics and the U.S. National Eating Disorders Association (NEDA) warned that WW promoted an “unhealthy relationship with food” (Miller, 2018).
In 2019, WW launched a new Kurba app, explicitly designed to assist 8-17 year old children obtain a “healthier” weight by tracking everything they eat and inviting them to embrace the goal of weight loss. Accompanying the launch of the Kurba app was a rebranding effort in which the words “diet” and “dieting” were de-emphasized, replacing them with words associated with “wellness” and “health.” While likely an attempt at neutralizing criticism of WW, as NEDA has noted, WW’s primary business purpose remains supporting weight loss pursuits. In response to the rebranding effort, they made the public pronouncement that “health is more than weight,” and a WW program that “emphasizes and celebrates weight loss is risky for this vulnerable population of children and adolescents at a time when their bodies are undergoing significant changes and are especially susceptible to harm” (National Eating Disorders Association [NEDA], n.d., para. 2). Nevertheless, as the rebranding effort suggests, WW has long sought to promote its unhealthy weight loss programs to youth in the face of public health concerns by prevailing health advocates.
For decades, the use of celebrity endorsements to increase product sales and brand awareness has held a preeminent place within the realm of strategic communication (Garthwaite, 2014). An estimated 10% of U.S. advertisements in magazines use celebrity endorsements (Belch & Belch, 2013). Advertisers routinely idealize and capitalize on brand images associated with celebrities (Russell & Rasolofoarison, 2017) as a persuasion mechanism for influencing attitudes and behaviors toward brands (Bergkvist et al., 2016). Despite five decades of market dominance, in October 2015, WW reported a 45% quarterly revenue decline, a fall of nearly 75% over the preceding three years. Yet, in the midst of this dismal quarterly earnings report, WW’s fortune was abruptly reversed following the announcement that Oprah Winfrey was to take a 10% stake in the company. The company’s ensuing 300% rise in stock value personifies what has been labeled the “Oprah Effect,” denoting the cultural authority and influence of Winfrey’s endorsement (e.g., Loroz & Braig, 2015; Max, 1999; Mourdoukoutas, 2015; Peck, 2002).
The types of influencers, such as journalists, experts, celebrities, and government officials, exert different influences on individuals. For example, a celebrity on television or within a magazine advertisement has a more significant effect on an individual’s attitudes and behaviors than a friend. Also, celebrities’ direct and indirect influence over information/message distribution and its effects on others connected to them within social media networks has surpassed their influence within the traditional media environment (Bakshy et al., 2011). The uniqueness of influencer strategies on social media networks lies in the way influencers’ position is leveraged in the community (network) and its influence enabled by the connections built in it. However, the characteristics of social media, such as social media as uncontrolled media, require further examination as a strategic communication tool. The brand community on social media can be monitored and targeted using influencer marketing strategies, but there remains no way to control the brand/influencer followers’ reactions to social media posts published by the brand or influencers.
Since 2008, there has been an absence of research examining the significance of celebrity endorsements (e.g., Bergkvist et al., 2016; Bergkvist & Zhou, 2016). However, no research remains to address the different effects celebrity and non-celebrity endorsements have within social networks, nor of celebrity endorsements’ potential role in propagating unhealthy narratives associated with commercial “health” and “wellness” campaigns. Accordingly, drawing upon agenda setting and the so-called “Oprah effect” as a case study, this research examines how celebrities and individuals affect the level of engagement in the brand community on social media, and such engagement’s potential to propagate and exacerbate problematic health narratives and behaviors among the public. In doing so, this research also investigates how WW and health advocates, as network influencers, frame each other’s narratives in the brand community.
Literature Review
WeightWatchers vs. Health Advocates
On February 7th, 2018, WW announced a free membership of WW’s diet program for 6 weeks for teenagers (Scritchfield, 2018). WW claims that its program fosters the development of healthy dietary habits among teens (Wiener-Bronner, 2018). According to Oprah Winfrey, “Weight Watchers positively impacts the lives of millions, including my own. I am inspired to be part of this purpose-driven mission as we deepen and expand our connection to communities, making wellness accessible to everyone” (PR Newswire, 2018, para. 5). The NEDA denounced the launch of WW's free teen-targeted program, stating that dieting is a risk factor for eating disorders (Miller, 2018). Echoing the NEDA pronouncement, the American Academy of Pediatrics explained that a healthy lifestyle, not related to weight, helps teens prevent obesity and eating disorders. The American Academy of Pediatrics stated its concern that the WW teens’ program held the potential of triggering eating disorders (Miller, 2018). Leading eating disorder specialists also publicly expressed their concern that the program could lead teenagers to believe that higher weights are unhealthy, rather than unhealthy eating habits. They emphasized that health outcomes derived from weight biases can cause anxiety, stress, depression, low self-esteem, and unhealthy body-ideal issues (Scritchfield, 2018).
In recent years, the Internet has become the dominant platform through which individuals seek information about eating behaviors related to health (Lapinski, 2006). Research has shown that online health information often encourages the development of eating disorders through the provision of body images and graphical representations (Lapinski, 2006). Online information also stresses ways to avoid the threat of weight gain rather than emphasizing the importance of developing healthy behaviors. The information presented displays the severity and susceptibility of weight gain and efficacy behaviors to prevent weight gain. Such efficacy behavior messages have been found to promote such unhealthy behaviors as fasting, purging, and the consumption of diet pills (Lapinski, 2006). Mass media affect adolescent girls and young women directly and indirectly through the cultivation of unattainable beauty ideals and the construction of social comparisons on the basis of thinness. Researchers have consistently found that a leading factor in the development of body dissatisfaction with weight, eating disorder behaviors (Lopez-Guimera et al., 2010), and body ideal images (Park et al., 2009) is mass media.
The National Association of Anorexia Nervosa and Associated Disorders (n.d.) stated that at least 28.8 million American people suffer from eating disorders. Within this context, 0.9% of American women suffer from anorexia, 1.5% of American women suffer from bulimia nervosa, 2.8% of American adults suffer from binge eating disorder, and 1 in 5 anorexia people die by suicide. In the 1970s, the average age that girls started dieting was 14. In the 1990s, the average age dropped to 8 (Iles, 2012). Based on these statistics, eating disorders are not merely individual problems but societal problems.
Agenda Setting & Celebrity Endorsements
Celebrities can play a role in the agenda-setting process (Nownes, 2021). Celebrities can increase attention to an issue and emphasize the importance of the issue. For example, Taylor Swift posted the importance of voting and asked people to visit the registration website on Instagram. Vote.org reported a significant increase in voter registration after Taylor’s Instagram post (France, 2018). According to McCombs and Shaw (1972), audiences not only learn about a particular issue, but the amount of importance to ascribe to that issue based on both the amount of information dedicated within a news story and the story’s position relative to other stories. In this manner, the authors argue that mass media “set the agenda,” determining what issues are important (p. 176). However, as McCombs (2005) notes, “the Internet dramatically changed the communication landscape with the introduction of myriad of new channels” (p. 544). Many agendas now exist in society, disseminated among large portions of the public. This has, in turn, provoked some observers to forecast the end of agenda-setting, wherein audiences are increasingly fragmented and individuals’ access to an abundance of online news and information has resulted in highly individualized media agendas.
McCombs (2005) contends that “in addition to the economic and organizational influences on agendas of online sites, the norms of professional journalism are also a powerful influence on content” (p. 545). He argues that Internet-based sites present agendas that are mostly congruent with the agendas of traditional forms of news media, and that “online sites show considerable resemblance to each other” (p.545). McCombs in turn notes how the Internet has arisen as “the new frontier” for research pertaining to agenda-setting effects. Within the “myriad of new channels” provided by the ascension of the Internet in contemporary society, celebrity endorsements have arisen as notable media actors. Celebrity endorsements can thus be seen as embodying agenda-setting processes.
McCracken (1989) defined a celebrity endorser as “any individual who enjoys public recognition and who uses this recognition on behalf of a consumer good by appearing with it in an advertisement” (p.310). Bergkvist and Zhou (2016) defined a celebrity endorsement as “an agreement between an individual who enjoys public recognition (a celebrity) and an entity (e.g., a brand) to use the celebrity for the purpose of promoting the entity” (p.644). Research has found that celebrity endorsements positively affect attitudes and behaviors towards products and services (e.g., Bergkvit et al., 2016; Elberse & Verleun, 2012; Knittel & Stango, 2014). Source and endorsement factors have also been found to affect attitudes toward a brand. The perceived motivation of the celebrity has a significant effect on attitudes toward a brand. In other words, when individuals understand a celebrity endorser’s motivation for endorsing the brand, they positively evaluate products and/or brands (Bergkvit et al., 2016). Celebrity attractiveness has also been shown to have positive effects on brand evaluations (Eisend & Langner, 2010). All in all, celebrity endorsements can affect teens’ health and weight ideas. This study thus investigated the role of celebrities in the WW Twitter community.
Social Network Theory
Social Influence on Networks
The concept of a social network can be formally defined as a composite of a set of objects (nodes) and relations (edges) between the objects (Kadushin, 2012). The simplest network contains two objects, 1 and 2, and one relationship that links them. In the case of Twitter, each Twitter account (i.e., nodes) is located within a shared topic network as an individual object, and they communicate with one another through various relationships (i.e., edges) such as tweets, replies, and mentions. Individuals in the network are influenced when they interact with each other (Watts, 2004), and such influences spreading through the network can be seen in any kind of social association, e.g., a shared interest, information, and/or experiences, linking individuals within the social network (Christakis & Fowler, 2013) such as the Twitter network.
Network analysis enables the visualization of key influencers and relationships on a given network using different indicators and measurements, which reflect the structural characteristics of that network and the nodes’ strategic potential within it (Yun et al., 2016). Additionally, because message distributions and interactions built between nodes on social media can now be tracked down easily, in recent years this method has received attention from researchers exploring the complex relationships between nodes and interpersonal influence spreading through social media networks. When the flow of a new idea, product, or practice is the research subject, network-based methods are considered one of the optimal approaches to assess the role of personal influence in the chain of events (Kadushin, 2012).
Knowing who the influencers are has become the first step in maximizing the diffusion of information, including in the manner in which they tailor their messages or in the creation of promotional information for new products (Keller & Berry, 2003; Weimann, 1994). Through the posts, messages, and information published and distributed by social media influencers, companies can directly reach not only each influencer’s network, but also the individual followers’ networks indirectly. Information about the relationships between influencers and followers, built through tweets, replies, and mentions, and how the relationships were repeated can also be used to measure each object’s influence within the network (Kadushin, 2012), including its influence on the attitudes and behavioral intentions of individuals and/or groups located on the network (Yun et al., 2016). As such, knowledge about influencers in the network can be leveraged to achieve organizational goals in persuading individuals within their social networks.
Different Types of Influencers
Identifying different types of social media influencers and measuring their influence over individuals/groups in the brand community remains increasingly relevant but undetermined. Most commonly, the number of followers on Twitter, which reflects network size and serves as an indicator of popularity, has been used to identify the influential nodes in a network (De Veirman et al., 2017). However, the number of followers is more akin to one of the node attributes explaining the node characteristics itself, rather than reflecting the node as a part of the network. Thus, looking at the mere number of followers cannot provide an overall understanding of each node’s influence and reasoning behind it, such as the node location and different types of connections with varying values in the network. With consideration of such limitations associated with viewing an individual’s number of followers as a popularity indicator, centrality approaches have been adopted to measure influencers in social networks. Popularity can be broken down into several different ideas under the general umbrella of “centrality” (Freeman, 1979). Centrality refers to the number of links that a specific node has within the network and captures how “important” (central) a node is within the network (Hansen et al., 2011).
Several centrality measures can be calculated through social network analysis, including degree centrality, betweenness centrality, and eigenvector centrality. Degree centrality, which is similar to the number of followers, has been the most commonly applied (Liu et al., 2015). The rest of the centrality measures can be thought of, respectively, as a bridge score (betweenness centrality) and an influence score (eigenvector centrality; Hansen et al., 2011).
Degree centrality is the simplest centrality measure, counting how many connections each node has, and therefore capable of being conceptualized as a popularity measure most similar in nature to the number of followers an individual has (Easley & Kleinberg, 2010). Betweenness centrality measures how often a given node lies on the shortest path between two other nodes. Nodes with high betweenness may influence information flows between others because they are bridges between different groups within the network. The elimination of such nodes with high betweenness centrality can disrupt communications within the network (Freeman, 1979; Kadushin, 2012). Lastly, eigenvector centrality implies that not all nodes in the network are equivalent, with some being more valuable than others, even if they have the same number of links to other sources. A Twitter account with few connections could have a very high eigenvector centrality if those few connections were much more valuable than other connections. Eigenvector centrality allows connections to have a variable value, so connecting to some nodes has more benefits than connecting to others (Hansen et al., 2011). For instance, a media Twitter account with high eigenvector centrality means that the account is important within the network because it is linked to other important accounts, such as accounts with more resources or greater audience size. Thus, each centrality indicates different types of influencers, and they are structural measurements explaining the values of each influencer’s location and their relationships with others within the network.
Research Questions
Individual influence attributes are related to credibility, expertise, or enthusiasm, whereas network influence attributes are related to connectivity or centrality (Gladwell, 2000). Individuals who have a large number of followers are more likely to be influential in the future (Bakshy et al., 2011). In what Bakshy and his colleagues (2011) emphasized as “who listens to whom” (p. 65), while word-of-mouth type information spreads via connections between ordinary individuals, it is mostly triggered by individuals who have a larger number of influencers (Bakshy et al., 2011).
Oprah Winfrey’s O: The Oprah Magazine, with an online audience of 8 million, represents one of the most successful magazine start-ups in history (Hearst, 2020). Her Book Club has been accountable for 28 consecutive best sellers and the estimated sale of 55 million “Oprah editions” (Minzesheimer, 2011). Coupled with her vast media holdings, the influence of Oprah Winfrey’s “cultural authority” has led her to be recognized as one of the most influential people of the 20th century by Time magazine (Clemeston, 2001). In 2015, Oprah Winfrey purchased a 10% stake in WW and became the WW brand’s leading endorser. However, her ascension into this role has been accompanied by WW efforts to build a commercially-driven program that “celebrates weight loss,” programs that leading health advocacy organizations have consistently labeled as a direct health threat to a “vulnerable population of children and adolescents.” Given her preeminent cultural position and longstanding reputation for influence, this study investigates how social influencers affect the level of mentions on WW Twitter, and the potential implications for the propagation of unhealthy “health and wellness” narratives in the public sphere. Identifying different types of social media influencers are important to anticipate who leads topics and shapes the public sphere on social media. When applying the network perspective, the measure of “centrality” is in regard to the number of connections that a specific account has within the network and captures how “important” (central) the account is within the network (Hansen et al., 2011). Influential Twitter accounts within a network tend to have a greater number of connections (i.e., degree centrality) and tend to be mentioned often by other media accounts, thus bridging different groups of Twitter accounts (i.e., betweenness centrality) or connections to important accounts, such as accounts with greater resources or greater audience size on the network (i.e., eigenvector centrality). Therefore, this research examined different types of influencers and centralities in the WW Twitter community.
RQ1: To what extent were different types of celebrity accounts successful at occupying different influencer positions (degree centrality, betweenness centrality, eigenvector centrality) within the WeightWatchers tweet network?
Some individual accounts can also be more visible than most (i.e., influencers) within the social network (Rasmussen, 2016) when they are successful at engaging in public discussions and other accounts by being replied-to or mentioned by other accounts (Freelon, 2014). For example, Twitter accounts refer to/mention other accounts to support or refute their agendas. Through a series of replies and mentions, they can become allies or opponents. When they stay in the center of public attention and conversations related to WW, they may be replied-to or mentioned by other accounts more frequently. Therefore, in this study, how influencers were being replied-to and mentioned by others in the network was examined, adding to the three types of centrality measures.
RQ2: To what extent were different types of celebrity accounts successful at occupying influencer positions by being mentioned and replied-to by other accounts within the WeightWatchers tweet network?
Foundations for developing eating disorders and an obsession with overweight increasingly emerges in youth. When WW announced its’ free teens programs, the NEDA, pediatrics, and health advocates expressed their concern that participation could lead adolescents to only consider weight in their construction of healthy living rather than developing healthy eating behaviors. Given the social significance of these health concerns, WW’s leading market position in dietary health programs, and the reputation of Oprah Winfrey’s celebrity endorsement, this study investigates the impact of social influencers on discussions related to the free teens’ diet program and how WW and health advocates frame the program and each other’s narratives on Twitter.
Method
Sample
The time frame of the sample is four weeks, starting from WW's announcement of the free teen program on February 2nd, 2018 (Shea, 2018). Before data collection, an initial search on Twitter’s network produced the relevance of two keywords associated with the area of inquiry, WeightWatchers and #WWFreestyle. These keywords were analyzed for two weeks in order to identify the fitness of the keywords. After this monitoring process, considering the volume of search results returned and popularity of the keywords in related tweets within the WW community, the search keyword of WeightWatchers was selected. NodeXL, a Microsoft Excel application add-in network and computer-assisted content analysis program, was used for Twitter data acquisition. Only those Tweets containing the search keyword were included in the sample, and the current social media stream was searched and saved daily. A total of 218,061 edges (tweets, mentions, reply-to relationships between Twitter accounts) and 11,202 nodes (Twitter accounts active in the network graphed using the keyword, WeightWatchers) were collected and included in a database for analysis.
Data Manipulation
To achieve and manage the WW-related Tweet database, Microsoft Access 2016 was used. First, all the tweet data archived in the 30 databases created daily using NodeXL were saved and converted to an Access database. In this study, Access was used to retrieve a choice of information from the massive datasets initially collected by NodeXL, which is regarded as the strength of this tool (MacDonald, 2010). Multiple daily Tweet databases were combined using a text command written in a specialized language called Structured Query Language (SQL), such as a union query, useful in merging results from more than one database and then producing them into a single Access datasheet. After the completion of this data manipulation process, the database of Tweets was ready for further analysis.
Data Analysis
To analyze text-based data and identify influencers, computer-assisted content and network analysis software, NodeXL was primarily used. First, network analysis was used to identify key influencers among various Twitter accounts using three types of network centrality measures-degree centrality, betweenness centrality, and eigenvector centrality. After the preliminary analysis, closeness centrality was dropped due to no evident closeness centrality score distribution among nodes. Secondly, by conducting computer-assisted content analysis, through the series of text-based data treatment processes, meaningful information within the text streams indicating topics (top words) or narratives (top word-pairs) proposed/delivered by influencers were identified, such as the most propagated words and word-pairs in the network.
Network Analysis
To identify influential Twitter accounts, a series of network analyses was conducted. Network elements such as the characteristics of relationships (direction) and the number of relationships (popularity) were used to calculate the network centrality of each Twitter account and their influence within the network. Several centrality measures were calculated through social network analysis, including degree centrality, betweenness centrality, and eigenvector centrality. To identify different types of influencers within the network, the top 10 Twitter accounts were rank-ordered based on their social roles as a popular figure (degree centrality), bridge (betweeness centrality), and influencer connected to important sources within the network (eigenvector centrality). Additionally, each Twitter account’s influence within the WW network was measured using two other types of network influence concepts: Top replied-to & Top mentioned. This denotes the degree to which each Twitter account had become subjects of conversation in the network and had been referenced by others, respectively.
Computer-Assisted Content Analysis
The frequency of each word and word-pairs detected in the tweets were used to identify the top 10 words and word pairs within the WW network. By analyzing popular words and word pairs within the network, the topic agendas of public discussion within the network were identified, as well as the significance of the discussion topics in relation to WW during the data collecting period.
Measurement
In this study, the influence of each Twitter account within the WW network was measured using various types of network influence concepts, including network centrality (measured by centrality measures) and being replied-to and mentioned, and popular topics propagated within the network (popular words and word-pairs).
Network Centrality
In this study, network centrality, calculated based on the number of links that a specific node has within the network, was used to measure each media Twitter account’s popularity. Three types of centrality measures, including degree centrality, betweenness centrality, and eigenvector centrality, were used to identify different types of influencers within the network. By using the degree centrality measure, a Twitter account’s capability to reach other Twitter accounts directly within the network was measured. Betweenness centrality indicates the extent to which an influencer connects to other Twitter accounts within the network. Lastly, eigenvector centrality was used to evaluate the influence of each Twitter account by assessing the value of their links to important accounts within the network.
Being replied-to and mentioned
Through an analysis of how each Twitter account had been replied-to or frequently mentioned in other tweets regarding the topic of Weight Watchers, each Twitter account’s ability to stay in the center of attention/conversation was measured. When it comes to the marketing/advertising power of each Twitter account, this ability to get the brand exposed to the public, enabled by their greater network centrality, has become an important potential of each node, rather than merely being able to distribute information.
Results
The sampling frame of Tweets used for analysis was four weeks beginning on February 2nd, 2018. During this time frame, a total of 218,061 edges and 11,202 nodes were collected and analyzed in order to identify influencers in the WW tweet network.
Network Influencers
The first research question examined network influencers in the WW Twitter community. Table 1 displays the key Twitter accounts within the network identified using the three types of network centrality measures: degree, betweenness, and eigenvector centrality. As shown in Table 1, the official Twitter account of WW, @weightwatchers, was the account with the greatest degree centrality. Celebrity endorsers of WW, being either officially promoting the brand or providing personal testimonials in public, @anavarro, @oprah, @djkhaled, and @andy, were the top four of the top 10 Twitter accounts with a degree centrality measure. Health advocates/organizations, including @nedastaff, @balanceedtc, @fionabodyposaus, @flynnfluencer (activist), and @fyeahmfabello (researcher) followed.
Within a given period of time, with the greatest betweenness centrality, @weightwatchers was the most influential bridge connecting account groups to each other, along with Weight Watchers’ official WW account, @ww_uk. Some celebrity endorser accounts (@ananavarro, @oprah, and @djkhaled) and health advocates (@laurathomasphd, @flynnfluencer, @gemmar333, and @balanceedtc) remained in the top 10.
As shown in Table 1, @weightwatchers ranked first with the highest eigenvector centrality, followed by @oprah, @nedastaff, @balanceedtc, @chr1styharrison, @joseesovinskyrd, @djkhaled, and @nic85m. Among the Top 10 nodes, except for @weightwatchers and two celebrity endorsers, @oprah and @djkhaled, the rest of the accounts were health advocates and dietitians (RDs). The results show that, with the highest eigenvector centrality, @weightwatchers was aggressive in linking to other influential accounts, but that the others, those in opposition to WW free teens’ program, were also being active at connecting with other nodes with important connections in the network.
Taken together, this study found that @weightwatchers ranked highest among Twitter accounts across all the three lists using the three types of network centrality measures. @oprah was ranked second when the influence was measured using eigenvector centrality. During the data collecting period, the top 10 influencers list was dominated by two groups of accounts, either Weight Watcher’s official accounts and celebrity endorsers or health advocates. This shows that health advocates, researchers, activists, and non-profit organizations entered into the WW brand community as WW announced their new plan on providing their program to teens for free, and consequently led public discussions within the community by setting an agenda about teens’ health and diet.
This The second research question examined another type of network influencer by the degree of being replied-to and mentioned by others in the network, staying in the center of public attention and conversations related to WW. For the top 10 accounts ranked by the degree to which they were replied-to, @weightwatchers ranked first, followed by @ananavarro, @beastieaw, @balanceedtc, @realrazor, @svaniamato, @alisonwwcoach, @oprah, @djkhaled, and @fionabodyposaus (see Table 2). Along with celebrity endorsers (@ananavarro, @oprah, @djkhaled) and health advocates (@balanceedtc and @fionabodyposaus) identified as influencers using centrality measures, @beastieaw, @realrazor (athlete, self-testimonial), @svaniamato (WW coach), and @alisonwwcoach (WW coach) entered into the ranking. This indicates that these newly emerging influencers represent the ones involved in the discussions through re-tweeting, replying and mentioning.
The account with the highest level of being mentioned in others’ tweets was @weightwatchers, followed by @oprah, @nedastaff, @balanceedtc, @fionabodyposaus, @djkhaled, @bedaorg, @moreloveorg, @ bodylove_ryb, and @jenkreatsoulas. There were a considerable number of accounts that appeared in this influencer ranking. Health advocate accounts, including @bedaorg, @moreloveorg, @bodylove_ryb, and @jenkreatsoulas are the accounts never ranked using centrality measures but mentioned to a significant degree by others in the WW network.
Mentioning and replying are hyperlinking practices occurring in social networks oftentimes, but they also indicate strategic ways of building connections, such as allies or opponents (Freelon, 2014). Health advocate accounts newly appeared in the top 10 influencer accounts by the degree of being replied-to and mentioned by others in the network may indicate a type of influencers exerting influence not by directly engaging in the WW network but by being influential outside community support to within-network influencers, setting agendas around the weight loss program effects on eating disorders.
Network Topics and Narratives
The third research question investigated to what extent different types of social influencers lead to different types of network topics related to the free teens’ diet program. Table 3 provides a summary of the most popular words and word-pair frequencies found within the network. The top 10 words on the Twitter accounts’ network shaped around the keyword WeightWatchers, were weightwatchers (n=182188), oprah (n=46467), teens (n=44859), wakeupweightwatchers (n=2897), weight (n=31211), diet (n=26451), dieting (n=23219), eating (n=22664), body (n=21042), and free (n=18107). This study found that social influencers led the topic related to weight management and discussed the effects of the WW program. For instance, WeightWatchers was the most popular keyword from the news of the new free teen WW program launch. At the same time, the hashtag #wakeupweightwatchers was utilized by health advocate accounts as a response to the free teens’ diet program launched by WW. This keyword trended on Twitter together with slogans such as risks of putting teens on diets, dieting is not the answer, body positivity, and dieting and eating disorders.
The fourth research question tested how WW and health advocates frame each other’s narratives on social media. For the top 10 word pairs, weightwatchers- oprah (n=33445) ranked first, followed by balanceedtc-weightwatchers (n=8276), eating-disorder (n=8074), new-weightwatchers (n=7231), eating-disorders (n=7155), link-between (n=6900), between-dieting (n=6886), dieting-eatingdisorders (n=6839), eatingdisorders-clear (n=6809), and weightwatchers-promotion (n=6744). When the narratives represented by each popular word-pair are considered, there are two types of narratives identified from the WW network: weight loss as the scientifically proven benefit of the new WW programs vs. the link between weight loss program/dieting and eating disorders. Based on tweets, this study indicated that words and word-pairs related to celebrity endorsers (e.g., Oprah), new programs free for teens (e.g., teens, free, promotion), and health advocates’ agendas (e.g., eating disorders, dieting) appeared as high frequency.
Discussion
This research found that brand/product issues can encourage supporters and opponents to share various perspectives. As the results of the network analysis showed, celebrity endorsers and health advocates were the most prominent influencers within the network. After the WW announced the free teen program, the NEDA, pediatrics, and health advocates criticized the WW program, and then joined the WW brand communities built on Twitter to discuss issues related to the program, such as the effects of weight loss programs on teens’ health and how dieting can cause eating disorders. This shows that a brand community on social media, which can be captured by social media analytics, does not function as a mere support community with static brand loyalty or equity. The community is more like an open community sharing certain characteristics with public spheres.
When Twitter users started to criticize WW, they simultaneously criticized Oprah Winfrey as they automatically connected the WW brand to its well-known brand ambassador. This echoes previous research (e.g., Elberse & Verleun, 2012; Knittel & Stango, 2014) finding that celebrities and brands share each other’s images and reputations. Nevertheless, analysis of the data revealed that Oprah Winfrey was a significant presence within the WW network community on Twitter following the free teens’ program launch. This in turn implies that Winfrey played a notable role in promoting the brand and program and her network position as an influencer within a brand community.
Oprah Winfrey shared the story of her own struggle with weight management within WW advertising and on Twitter, highlighting how much weight she had lost and how much healthier she had become through the WW dietary program. Such content appears to reflect the diffusion of a successful health narrative to audiences as a basis for persuading others to enact healthy eating and lifestyle behaviors. Oprah shared private information about her weight conditions in the public sphere. When individuals are exposed to celebrities’ experiences and stories related to their health status, they are likely to have less fears related to health issues, such as diabetes. Nevertheless, relating to the personal health-related narratives of celebrities comes with the potential ramifications of supporting the celebrity’s opinions and following their behaviors (Beck et al., 2014). Previous research (Beck et al., 2014; Elberse & Verleun, 2012) highlights the rationale for health advocates’ publicly expressed concern over the WW free teens program. The well-documented and unprecedented success of Oprah Winfrey as a cultural influencer has been labeled the “Oprah Effect” (Baum, 2006; Max, 1999; Mourdoukoutas, 2015; Peck, 2002). While this study revealed that the platform of Twitter facilitated the emergence of a brand community with avenues for the discussion of social concerns in conflict with the overarching narrative and frame constructed by WW and Oprah Winfrey, it also illustrates the degree to which the construction of public health frames by health advocates competes with celebrity influencers with greater levels of influence.
Network influencer analysis results show that health advocates are operating as agenda setters in the WW network by responding to the WW new teens’ free program launch and leading public discussions about teens’ health and diet. Top words and word-pair analysis results corresponded to the findings from network analysis results. One of the most trending keywords was wakeupweightwatchers from the #wakeupweightwatchers social media initiative during the time where WW from the new teens’ program launch news was the most popular keyword. The narrative about the weight loss program or dieting as a predictor of teens’ eating disorders, identified from the most popular word-pair analysis, also shows how health advocates set the competing agenda in the WW network. In the WW case, the existence of different types of network influencers may have worked to sustain network agenda diversity around the important public health issue.
The case of the Tiger Woods scandal and its effects on the reputation of Nike demonstrated that the emergence of negative publicity surrounding a celebrity negatively influences the image of the brand endorsed by that celebrity (Knittel & Stango, 2014). The results of this study re-confirm the relationship between celebrity endorsers and brands in the case of negative publicity, yet in a manner that illustrates how this process is reciprocal in nature. Within the context of WW free teen program and Oprah, the analysis of the brand community on Twitter revealed that the negative publicity incurred by the brand exerted a negative impact on the celebrity endorser. When topics such as teens’ eating disorders and body image entered into the WW brand community’s discussions on Twitter through the entrance of health advocate Twitter accounts, the promotional tweets about the new WW program and celebrity endorsers’ self-testimonial could not dominate the community’s discussion.
The emergence of new health advocates’ accounts as network influencers shows how the brand community has changed into a battlefield over topics and narratives centered on WW as a weight loss program. Twitter became a public sphere in which to discuss a variety of issues with various perspectives in the form of public dialogue. Tweets presented concerns surrounding the potential promotion of eating disorders, but at the same time, these tweets received attention from the mass media and users. This study discovered a unique opportunity for insight into how health issues are promoted on social network sites.
Lastly, if the popularity measures heavily depending on the follower numbers were solely used, the emergence and disappearance of different types of influencers in the WW network could not be captured. The official Twitter account of WW and celebrity endorsers such as @oprah are the accounts with the greatest number of followers, but the influence of health advocate accounts was developed and strategically built through retweeting, replying to other tweets, and also being mentioned by other accounts in the network. They indicate that another type of network influencer is also noteworthy, those who are highly visible within the network by functioning as an influential outside source to support within-community agendas. This demonstrates how the network concepts addressed in this study can further assist interpersonal influence studies on social media and those pertaining to strategic health communication marketing.
Weight management programs can help people lose weight (Baetge et al., 2017; Dansinger et al., 2005). However, according to Spahlholz et al. (2016), in the U.S., 40% of adults have experienced weight stigma in their workplace, healthcare, schools, and interpersonal sources. Negative weight ideas can induce individuals’ devaluation of themselves and lead to greater levels of depression, anxiety, and stress (Himmelstein et al., 2020). As previous research indicated, even adults with negative weight ideas and body images are prone to negative physical and mental health (e.g., eating disorders, self-blame, exercise avoidance, and depression). While in adolescence, teenagers can create their own health perspectives. Once they believe that higher weights are unhealthy, instead of unhealthy eating behaviors and excise avoidance, it is hard to change their beliefs. Also, negative weight stereotypes can cause depression, low self-esteem, low life satisfaction, etc. (Scritchfield, 2018).
In addition, children may be unable to understand the difference between calories and nutrition. Media, specifically advertisements, portray lower calories as healthy, but do not provide nutrition information (Harrison, 2005). For example, food is often advertised as fat-free. However, they commonly do not include information regarding protein, fiber, or any nutritional aspects. Even though WW emphasizes “health habits,” the WW website still highlighted “losing weight easier” and “improv[ing] BMI status” and provided testimony on how much weight members lost. Even eating vegetables and fruits alone cannot fulfill the whole nutritional needs of growing children. Therefore, youth weight programs should educate youths on how to understand nutrition related to healthy behaviors rather than for weight loss. Social influencers should carefully lead agendas related to health issues, specifically when promoting messages aimed at children and teens.
Limitations & Future Study
There are some limitations inherent in this study. First, within this study, a computer program that enables network analysis and computer-assisted content was used to collect tweets and extract meaningful information about the important nodes and relationships among them. Despite the methodological strength, it has a sampling limitation. The search results returned using the API may give incomplete results under certain circumstances (Thelwall, 2014). Thus, the collected dataset should be treated as a sample rather than a comprehensive set.
This study also explored the network concepts that can supplement measurements of social influence in social media networks. During the preliminary analysis, closeness centrality was eliminated from further analysis because the evident closeness centrality score distribution across and among nodes was not found. The way tweets regarding WW were collected in this study may be understood as the reason why all the tweets searched and archived are highly likely connected to each other through the search keyword, Weight Watchers, within short steps. Future research needs to further test such social influence-network concepts across different settings and communities on social media.
This study collected data before and after WW announced the free teens’ program. In doing so, this study found the connections between WW and health advocates who are against the free program. Future research should focus on natural status in order to facilitate a better understanding of the effects of social influencers on contagious opinions and behaviors.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation
Print
