The consumption of medical aesthetic services has become popular in recent years. Many people have purchased medical aesthetic services and treatments in pursuit of self-beauty. In 2017, the market size of global medical aesthetics was nearly US $230 billion. The growth of Taiwan’s medical aesthetic device market was ranked among the top 3 in Asia. Nearly 60% of Taiwan’s working class spent US $2300 to US $6700 on plastic surgery and microsurgery each year. On average, hyaluronic acid injections were used 193 times daily, and 30% of visitors from Mainland China came exclusively for beauty purposes. As such, the medical aesthetics industry contributes around US $1.4 billion to Taiwan’s gross domestic product and continues to rise.

As technology advances and online communities continue to prosper, consumers often search for answers and support from various online communities before making a decision to purchase [1]. When searching for answers and support from online communities, consumers’ trust toward other members and the community is a key factor impacting the decision to purchase [2].

According to the survey results of the Word-of-Mouth Demand of Online Communities conducted by the Market Intelligence and Consulting Institute, in 2014, nearly 81% of consumers searched the internet for word-of-mouth information via social network sites (45.8%), discussion forums (44.7%), and blogs (33.1%). At present, the exchange of word-of-mouth information among consumers in Taiwan remains concentrated on frequently visited social network sites. On the basis of product type, consumers choose channels with different levels of trust regarding word-of-mouth information. Therefore, businesses should understand consumers’ word-of-mouth demands and internet search habits and should treat such data as important reference when introducing new products or planning annual marketing strategies. Medical research institutes can effectively communicate and connect with more people via online communities [3]. The literature also indicates that consumers tend to believe in and are eager to obtain medical information from online communities [4].

On the basis of the trend described above, this study explored whether a sense of trust is generated in consumers of the medical aesthetics industry from the community support they experience before purchasing or sharing and whether this will ultimately increase their intention to share or purchase.

Social support theory explains how resources and help are acquired through interpersonal interaction within a community unit. There are 2 types of social support, namely structural and functional. Structural social support emphasizes community size, operating systems, relationship intensity among community members, and whether a sense of belonging to the community is generated after interpersonal interaction [5]. On the basis of relationship intensity, structural social support can be further classified as formal structural social support from organizations such as hospitals and schools and informal structural social support from intimate relationships such as families, friends, and peers [6].

Functional social support refers to the subjective feeling formed by an individual after interacting with others. The acquired support intensity is evaluated using the relationship, process, and results of interaction [7]. Support can be categorized as tangible support, informational support, emotional support, esteem support, and network support [8]. Furthermore, functional social support can be delineated as esteem support, instrumental support, informational support, and emotional support [9]. As subjective feelings vary between people, the definition of social support has not yet been standardized in academic research. However, consensus has been reached on the 3 major types of social support, namely tangible support, informational support, and emotional support [10] as the bases for further extension and adjustment.

Social support can also be regarded as a reward acquired from interpersonal interaction. In other words, social support can be viewed as resources acquired by an individual who has exchanged something with others based on the purpose or needs of the individual [11]; examples are tangible support such as monetary assistance and gifts [12] and intangible support such as caring, listening, and giving advice [1]. The higher the social support intensity an individual feels after interacting with others in a group, the higher the individual’s intention to establish an intimate interactive relationship in the group [13].

The theory of social support can be expressed in a more rigorous form by nesting it into a theory of social capabilities–based flexicurity in a learning economy. The basic idea has been developed by various theorists [14-24] who drew upon the insights of Adam Smith; they have proposed a theoretically rigorous and elaborate evaluation of well-being [25,26]. Sen is the originator of this capability approach in recent years [27]. The theoretical criticisms of the utilitarian approach by Sen et al—this approach reduces all qualities into the quanta of utilities—are serious ones. Khan et al has pursued a similar line of criticism in a number of recent papers and in his book Technology, Development, and Democracy [28-31]. This approach makes the capabilities explicitly social and asks what concatenation of institutions, both economic (real and financial) and other (eg, political and social), will allow social support and trust to increase steadily and at the same time equalize them among diverse individuals [32]. In effect, as the following discussion makes clear, we are asking how we can increase and equalize real positive freedom for individuals in specific social contexts—in this study, specifically for online communities and individuals. In his book, Technology, Development and Democracy, Khan points out that trust and freedom are interactive arrangements in a society where concrete institutions at many levels and technologies of production and exchange, all play definite roles [33]. The social support theory can be elaborated in a deeper way by building on these foundations and applying them dynamically to (post)modern networked online communities, with social capability–building through causal connections and positive feedback loop mechanisms. In this approach, methodologically, confirmatory factor analysis can support or disqualify the particular causal mechanisms that are proposed in this study and others in this area of research. We now turn to the specificities of social support in an online environment.

As the use of the internet has grown rapidly in recent decades, more people engage in online communications every day. In particular, as social network sites proliferate, easy exchanges among users are facilitated by the online environment, in which social support has gradually transformed into social support communication [14]. Online social support refers to information and emotional exchanges in the virtual space and is provided through informational support and emotional support [15]. Social support and online communities are closely related [16]. Previous studies showed the connectedness between social support and trust [17]. In this study, we believed that informational support and emotional support benefit the relationships of all related parties and are conducive to relationship building and maintaining a good online environment.

Emotional support refers to demonstrating to others empathy, care, love, understanding, or encouragement [18], making them feel they are being given attention. In contrast to informational support, emotional support emphasizes the emotional side of social support, which may indirectly help overcome problems [34]. Emotional support enables community members to obtain help from other members. Specifically, love is the foundation for developing trust [18]. Therefore, people develop community exchanges and trust through emotional exchange and contact with other members in the community. In addition, if a member obtains love and warmth from other members in the group who have similar experiences of pain and challenges, he/she will often feel more comfortable remaining in the focus group. This is because emotional responses such as love and care satisfy members’ needs for respect and social attention, enabling them to identify as members of the group with a sense of belonging [35].

Informational support provides individuals with advice, guidance, or useful information that helps them solve problems, generate new ideas, or make good decisions [36]. If people continue receiving positive help such as valuable recommendations or instant help from network friends or online groups, they are more likely to show benevolence, integrity, and capability to the other party, further enhancing their trust toward members who provided the related information [37].

As mentioned, interest in a relationship is necessary in the development of cooperative relationships [38]. Therefore, if group members can benefit from others’ opinions, they are likely to agree on the values of the focus group and maintain a long-term relationship with it.

Norms of reciprocity are behavior norms in a social community that obligate people who have received help from others in the community to offer similar help in return in the future [39,40]. When community norms are gradually transformed into individual values in a community, members consider exchange behavior based on a friendly interactive relationship as natural, which impacts their sharing intention in the future [41]. Thus, the higher the intensity of norms of reciprocity, the higher will the frequency of exchange activities be [42]. Furthermore, members’ knowledge collection behavior is enhanced if the knowledge provided on knowledge sharing platforms is reliable and comprehensive [2]. As such, the more the knowledge is aligned to members’ needs, the more willing they will be to offer useful knowledge to other members to help them resolve their problems. Previous studies indicated that norms of reciprocity promote trust in a community [43]. Thus, we hypothesized that norms of reciprocity have a significant positive impact on trust toward members.

Trust is considered a basic factor in the formation of successful relationships [38]. In particular, recent studies focused on the relationship between trust toward providers of products and services and customers’ online purchase intention [44]. As retrieving information from the internet becomes more popular, consumers ask for recommendations from social communities and people they trust. On the contrary, privacy issues mean that consumers are more likely to share their own information with someone they trust [45]. Past studies also confirmed the relationship between brand community and consumer participation [46]. Therefore, the impact of trust on social sharing intention and social buying intention—as well as trust toward members and the community—is discussed in this study.

In this study, trust toward members is defined as an individual’s intention to rely on and believe in other community members, such as believing the information they provide and taking action and making decisions based on other members’ recommendations. Previous studies found that trust toward members impacts online participation behavior such as seeking and providing information in focus groups [47], because in an environment with mutual trust, people often help each other, which develops into common community activities. In particular, information obtained from credible sources is often considered more useful and treated as the basis for decision making [48]. In addition, people are more likely to share their consumption experience of products or services afterward.

When other members display reliable characteristics of the individuals seeking information, such as benevolence, integrity, capableness, or have had a similar experience, information-seeking members feel comfortable talking to one another as they share a common knowledge background. As such, any suspicion or doubts in each other are alleviated.

Trust toward the community refers to the extent to which people rely on and trust a community. Generally accepted standards are enforced by online communities to instill norms of reciprocity among their members. In addition, benevolence and integrity in a community can lower doubts among members. The relationship between trust toward the community and customer loyalty has been confirmed [49]. Essentially, the stronger the trust an individual has for a community, the higher the tendency that individuals will seek advice from that community on products or services they are interested in buying and share their consumption experience within the community afterward.

Relationship commitment reflects many aspects of the long-term relationship [38]. Previous studies confirm the relationship between commitment and users [36], indicating that when users are committed to an online community, they contribute and participate more in the online community [50]. Previous studies also found that online brand community commitment significantly impacts consumers’ decision making [51]. In an online business community environment, the network platform provides a shared space within which users can exchange information and communicate. Users can share their consumption experiences with friends in online communities or seek product recommendations from their network of friends. If users are committed to maintaining a sustainable commitment relationship with an online community, they will strive to maintain this commitment and are more likely to participate in various group activities to help grow the community.

According to the trust transfer theory, interpersonal trust develops into trust toward the community for 2 reasons. First, trust among members makes users believe that information provided in the community is more credible, which reassures members that the community will continue to improve its service quality and provide effective management in building a credible communication environment [52]. Second, previous studies pointed out that trust among members increases members’ trust toward the community [53]. Previous studies reported that when people trust sellers in a Business to Consumer online community, their trust could be transferred to the related applications [54]; to give an example, Asia’s famous and successful retail drugstore brand, Watsons, has its own online community and app. Other studies confirmed the relationship of trust transfer between brand community and consumer participation [46]. When members trust one another, a bigger and stronger community is created, enabling members to consider the community a suitable venue for communication.

In the past, much research focused on core self-evaluation (CSE) when studying different kinds of human behavior, for example, the relationship between CSE and social support [55]. CSE is defined as the basic evaluation of one’s own values, potential, capabilities, and talents [56]. This study hopes to understand the mediating role of CSE on social sharing and buying intentions. Therefore, we hypothesized that CSE is a moderator variable between trust toward the community and social sharing and buying intentions.

In sum, this study explored whether emotional support, informational support, and norms of reciprocity have an impact on social sharing and social buying intentions through trust toward members and trust toward the community using CSE as the moderator variable. The framework of the study is provided in Figure 1.

The questionnaires were released on January 11 to 31, 2018, and a total of 565 valid questionnaires were collected. Before we started the survey, we first contacted 800 online medical aesthetics community users and asked if they were willing to participate in the survey. A total of 780 online users indicated a willingness to take part in the survey. We then sent questionnaires to those users. The returned surveys yielded a response rate of 78.2% (610/780). After deleting 45 incomplete responses (eg, respondents answered Strongly Disagree or Strongly Agree to all questions), a total of 565 valid observations were collected (a response rate of 70.6%). Table 1 shows the demographic characteristics of the participants. Most of the samples were female, aged 18 years older, and most of them were university graduates. Most of the samples had less than 2 years of community experience, and most of them were willing to spend less than US $10,000 on medical aesthetics.

We analyzed data using SPSS 12 and AMOS 22. The measurement model fit indices (v2/df=2.39; root mean square error of approximation, RMSEA=0.05; confirmatory fit index, CFI=0.982; adjusted goodness of fit index, AGFI=0.906; and Bollen fit index, IFI=0.982). The most widely used incremental fit index, CFI, has a value of 0.982, which exceeds the guideline value of 0.90 for a model of this complexity and size. This model has not been compared with other models; however, the value of parsimony index AGFI (0.90) reflects a good model fit. All these absolute, incremental, and parsimony fit indices suggest an acceptable fit for the measurement model.

Table 2 shows the results of reliability and validity testing. Regarding the reliability of the question items, the factor loading of all the items ranged between 0.7 and 0.95. With their factor loading values being greater than 0.5, these items exhibited high reliability. Average variance extracted (AVE) testing was used to assess the convergent validity of the items. It demonstrated that the set of indicators had an AVE greater than the 0.5 threshold. The composite reliability (CR) of measured variables was estimated to test their internal consistency in the measurement model. CR indicates how well these variables represent latent variables; a CR value of greater than 0.7 generally shows high internal consistency. In this study, the CR values of all constructs in the corrected model ranged between 0.89 and 0.96, indicating that the question items of the constructs had sufficient reliability.

Table 3 tabulates the Pearson’s correlation coefficients of the variables. Significant positive correlations were observed among emotional support, informational support, norms of reciprocity, trust toward members, trust toward the community, social sharing intention, social buying intention, and CSEs. In addition, all constructs achieved discriminant validity, because the square root of each construct’s AVE was higher than its correlation between any of the constructs [57].

This study performed a structural equation model analysis using Amos 22. The model fit indices (v2/df=2.39; RMSEA = 0.05; CFI = 0.98; AGFI = 0.906; and IFI = 0.982) suggested that the model represents a satisfactory fit to the data [58]. The model fit indices supported a good model fit. Therefore, further analyses of the relationships among the modeled constructs were conducted.

The results showed that emotional support and norms of reciprocity positively impact trust toward members, and trust toward members positively impact trust toward the community. This generates trust transfer, which positively impacts social buying and sharing intentions. However, the informational support did not have a significant effect on trust toward members, possibly demonstrating consumer distrust toward medical advertising. The path coefficients of hypothesis testing are summarized in Figure 2 and Table 4.

After all the hypotheses were validated, a Sobel test was conducted to assess mediating effects among variables. A value lower than 0.05 in AMOS and a z score of greater than 1.96 in the Sobel test results indicate significant mediating effects [59]. The Sobel test results showed that emotional support affected trust toward the community through trust toward members (Sobel z=5.99); norms of reciprocity affected trust toward the community through trust toward members (Sobel z=9.55); trust toward members affected social sharing intention through trust toward the community (Sobel z=16.03); trust toward members affected social buying intention on the basis of trust toward the community (Sobel z=15.45). In sum, trust toward members mediated the effects of emotional support and norms of reciprocity on trust toward the community, whereas trust toward the community mediated the effects of trust toward members on social sharing intention and social buying intention.

To understand whether CSE is a moderator on the impact of social sharing and buying intentions, this study averaged the point of CSEs in the sample after statistical analysis and, referring to the past research methods [60], this study divided the sample into 2 groups: high CSE (Average point of CSE>4.85) and low CSE (Average point of CSE<4.85), and compared the group differences based on the past research [61]. Table 5 shows that CSE is a moderator variable between trust toward the community and social buying intention, but CSE is not a moderator variable between trust toward the community and social sharing intention. The path coefficients of hypothesis testing are summarized in Table 6.