Understanding the Potential Influence of WeChat Engagement on Bonding Capital, Bridging Capital, and Electronic Word-of-Mouth Intention

Abstract

As the most prevalent social media platform in mainland China, WeChat enables interpersonal communication among users and serves as an innovative marketing platform for enterprises to interact with consumers. Although numerous studies have investigated the antecedents of electronic word-of-mouth (e-WOM), WeChat users’ specific behaviors still receive limited academic attention. Drawing from social capital theory and social exchange theory, this article develops a model to systematically explore three differentiated types of WeChat behaviors and their association with users’ social capital and e-WOM intention. The conceptual model is explicitly evaluated by utilizing web-based data gathered from 271 young people. Obtained results demonstrate the path effects indicating that: (1) WeChat use behaviors such as seeking, sharing, and liking can positively influence bonding social capital, while only the impacts of sharing and liking on bridging social capital are significant; (2) bonding and bridging social capital are both significant predictors of e-WOM intention, and bonding social capital is the more influential of the two; (3) bonding social capital partially mediates the effect of seeking on e-WOM intention. These findings are eloquent for researchers and operators to further grasp the increasing importance of WeChat adoption and social capital on young generations’ e-WOM intention in the evolving digital age.

1. Introduction

The dramatic advance of social media, including microblogs, video-sharing websites, and instant messaging applications, shows great potential for influencing users’ behavior, which has induced enterprises to implement digital transformations [1,2]. Among these diverse platforms, WeChat is presently recognized as the most preponderant social networking site (SNS) and has grabbed increasing attention from researchers [3,4,5]. Launched by Tencent on 21 January 2011, WeChat is a free application to provide instant messaging services for users and supports all mobile device platforms embracing Apple iOS, Android, and Windows Phone operating systems [6,7]. According to the financial report of WeChat, the number of monthly active accounts of WeChat (both domestic and international versions) reached roughly 1.206 billion in June 2020, with a year-on-year increase of 6.5% [8]. Recently, electronic commerce has developed rapidly in China, and information shared on WeChat can help consumers to make better shopping decisions [9,10,11]. In 2019, the average daily trading volume on Mini-Programs more than doubled compared with that of the previous year, with a total transaction volume exceeding RMB 800 billion [12]. As a substitute for SNSs like Facebook and Twitter in mainland China, WeChat has currently grown into the most widely used social media platform for marketers by virtue of its ease of use, large user base, and direct communication channels [6,13,14].

Owing to the prevalence of SNS, the existing literature on electronic word-of-mouth (e-WOM) has gradually recognized the significance of e-WOM in influencing customers’ decision-making or purchase behavior [15,16]. A growing number of researchers have identified several individual level antecedents of e-WOM, such as SNS use [17] and social capital [18,19]. Despite research progress in the fields of commerce and marketing, this paper identified three key research gaps as follows. Firstly, although a number of recent articles have discussed the relevant issues of e-WOM communication through open SNS platforms, especially Facebook and Twitter [20,21], the reasons behind individuals’ intention of e-WOM engagement in the environment of closed platforms like WeChat has seldom been investigated [15]. The difference between these two types of platforms is that personal relations may play a more prominent role on the close platforms because users exchange messages with other trusted members who belong to the same defined group [3]. Therefore, the present comprehension of e-WOM intention mechanisms may not be easily applied to these innovative SNS services, making it necessary to further study in the new media context. Secondly, while scholars have acknowledged that social interaction is an integral part of WeChat and contributes to the accumulation of social capital [22], the specific mechanism of how different types of WeChat use behaviors could influence individuals’ bonding and bridging social capital remains underexplored. The capability to develop and preserve interpersonal relationships is an essential prerequisite for the build-up of social capital [18]. Existing studies related to WeChat use solely investigated the effect of single behavior [3,4] and seldom systematically delved into different types of behaviors. Thus, this study examines three distinct types of WeChat use behaviors: seeking behavior, liking behavior, and sharing behavior. This present research focuses on these three WeChat use behaviors because they represent the largest proportion of behaviors on social media platforms and have become the most popular ones that have increasingly received researchers’ interest in the field of social media use and e-WOM communication [3,13]. Thirdly, there is a dearth of research on the mediating role of social capital on the relationship between WeChat users’ behavior and e-WOM intention. Some scholars have proved that e-WOM communication strategies in SNS may enjoy a higher financial incentive than traditional e-WOM does [23,24]. It’s obvious that the sociality and interactivity of SNS can influence customers’ decision-making by promoting their interpersonal and social connections [15]. Nevertheless, previous literature only focused on revealing the direct relationship between social capital and consumers’ e-WOM [11,25], and seldom explored the indirect effects of social capital.

Based on a literature review, this current study utilizes Chinese young people, the most active group on WeChat [3,5], as a research sample and adopts WeChat behaviors and social capital to gauge individuals’ e-WOM intention. Moreover, in this paper, a research model is developed to determine the factors that may significantly influence individuals’ social capital and e-WOM intention in the WeChat context, with a particular focus on the mediation roles of bonding and bridging social capital. Overall, the primary objectives of the research are as follows: (1) to establish the relationships between WeChat use behaviors (seeking, sharing, and liking), social capital (bonding and bridging), and users’ e-WOM intention; (2) to discover whether social capital mediates the effects of WeChat use behaviors on e-WOM intention. To the best of our knowledge, this current research makes an early effort to elucidate the joint influence of differentiated types of WeChat use and social capital on e-WOM intention and comprehensively uncover the underlying mechanisms of these associations among WeChat users in mainland China. By doing so, the present study both bridges the aforementioned research gaps in the existing e-WOM literature on WeChat and provides novel directions for further studies in this field [26,27]. Accordingly, the results will provide meaningful theoretical and managerial implications. Theoretically, this present article contributes to the comprehensive understanding of the application of social capital theory and social exchange theory in e-WOM communication within the WeChat context. Furthermore, the theoretical framework offers fresh insights into the dissimilar impacts of seeking, sharing, and liking behaviors on social capital and e-WOM intention, as well as the mediating role of bonding social capital between seeking and e-WOM intention. Managerially, this study offers several strategic implications and directions for different groups involved in the process of e-WOM communication. The results of this study will allow marketers and managers to better comprehend the diverse functions of WeChat and develop attractive content or activities that are compatible with consumers’ behavior incentive and social capital.

The general framework of this paper is structured as follows: review the relevant streams of literature to establish the theoretical framework and hypotheses; then introduce the research methodology and data analysis strategy, as well as report the results; finally illustrate the discussion, limitations, theoretical, and managerial implications for future lines of research and practice.

2. Theoretical Foundation and Hypotheses Development

2.1. Linking WeChat Use Behaviors to Social Capital

As one of the largest and most universal SNSs in mainland China, WeChat serves as a multifunctional platform allowing users to acquire interesting and useful information and utilize the function of liking and sharing to express their attitudes. Some research has categorized SNS users into two distinct types; quiet members and communicative members [28,29]. Quiet people would like to browse web pages to obtain information and rarely communicate with each other, while communicative people may show more enthusiasm in interacting, posting, commenting, and retweeting. To comprehensively outline WeChat users’ behaviors, this paper identifies three typical behavior patterns of communicative people: seeking, liking, and sharing, which are explicitly documented in prior literature [3,4,13]. Seeking behavior is defined as the process in which users participate in the activities running on SNS to satisfy their desire for information [30]. Liking behavior is a convenient way for users to convey approval or interest in the messages and has gained extensive recognition and acceptance [31]. Sharing behavior reflects when users add some interesting content to the social media page or send it to others as private information [20].

Notably, plenty of scholars have proven that SNS use behaviors may contribute to the depiction of people’s complicated interpersonal relationships on SNS [32,33,34] and can prompt the formation and development of individuals’ social capital [10]. Social capital is the core concept of social capital theory, which typically refers to the total of the realistic and potential resources embedded within and stemming from the interpersonal networks of an individual or social unit [35]. Given the ambiguity and multidimensionality of social capital, previous studies have divided social capital into two components: bonding social capital and bridging social capital [10,11]. The former one is characterized as strong social connections, while the latter one derives from weak social connections. As per pro-social motivation theory and organizational motivation theory, social capital may stimulate the occurrence of seeking and sharing behaviors [36].

Empirically, previous studies have confirmed that general SNS usage is positively associated with individuals’ social capital in the online environment [37,38]. For instance, Aubrey and Rill [38] demonstrated that using Facebook habitually was connected to an increase in online bridging social capital. With the increasing diversification of SNS functions, scholars have discovered that general SNS usage cannot describe users’ behaviors completely [3,4], leading to a need for additional studies on specific SNS behaviors. A quantitative study based on 970 Facebook users discovered that two forms of online behaviors, browsing and participating, can exert extraordinarily positive effects on both bonding and bridging social capitals [10]. Recent research has also made a conspicuous breakthrough in the relationship between users’ different behavior patterns and social capital. As explicated by Sumner, Ruge-Jones, and Alcorn [21], the “like” button’s vital function was embodied in relational facilitation, including building and maintaining social ties. Similarly, Chen, Ma, Wei, and Yang [3] carried out an empirical test on 518 effective samples of Chinese college students who participated in youth league campus activities and observed the positive relation between SNS behavior (seeking information and sharing comments) and social capital. Considering WeChat has become one of the most prevalent communication channels due to its ubiquity, mobility, and interactivity [9,18], it is reasonable to argue that WeChat users’ behaviors may contribute to the accumulating social capital among them. In light of the above arguments, this paper postulates the following hypotheses:

Hypothesis 1 (H1a).

Seeking has a positive effect on WeChat users’ bonding social capital.

Hypothesis 1 (H1b).

Seeking will positively relate to WeChat users’ bridging social capital.

Hypothesis 2 (H2a).

Sharing has a positive effect on WeChat users’ bonding social capital.

Hypothesis 2 (H2b).

Sharing will positively relate to WeChat users’ bridging social capital.

Hypothesis 3 (H3a).

Liking has a positive effect on WeChat users’ bonding social capital.

Hypothesis 3 (H3b).

Liking will positively relate to WeChat users’ bridging social capital.

2.2. Linking Social Capital to e-WOM Intention

E-WOM refers to the online sharing of useful evaluations of products, services, or brand-related marketing among consumer groups [39], which can have an intense impact on individuals’ purchasing intentions and the growth of product sales [40]. Compared with traditional word-of-mouth, e-WOM is becoming a more powerful marketing tool due to the advantages of its speed, convenience, breadth, and lack of face-to-face interpersonal communication and pressure [18]. The rapid development and unprecedented popularity of SNSs have transformed e-WOM communication, creating more opportunities for consumers to exchange their product reviews beyond the limitation of time and space [41]. A growing number of researchers are predisposed to determine the pre-condition of e-WOM via building research frameworks and using a variety of testing methods [5,15]. Among the existing research findings, social capital has been widely studied [41,42] and has been proven to positively influence SNS users’ attitudes towards e-WOM intention [11].

Prior studies have revealed that individuals’ willingness to engage in e-WOM communication depends on the intensity of social ties: strong connections tend to promote information dissemination and referral behavior at the micro-level, while weak connections act as a bridge at the macro-level [18,43,44]. Fitzgerald Bone [45] highlighted that due to information from strong ties being more credible, WOM is diffused more extensively among people who have strong relation networks than among those who have weak relationships. By contrast, other scholars have stated that weak relation networks can offer a novel source of related messages [46], which would play a more prominent role in the process of decision-making [42]. Although bonding and bridging social capital may have different influence mechanisms, it is undeniable that they are both critical predictors of e-WOM intention. A survey of 238 SNS users determined that tie strength is the key factor encouraging consumers to participate in e-WOM communication [18]. More recently, González-Soriano, Feldman, and Rodríguez-Camacho [25] adopted the perspective of social identity and confirmed the positive effect of social capital on the generation of e-WOM. In short, social capital owned by SNS users can increase their interest in communicating and disseminating product-related information, thus increasing e-WOM. Accordingly, this paper hypothesizes:

Hypothesis 4 (H4).

Bonding social capital has a positive effect on e-WOM intention.

Hypothesis 5 (H5).

Bridging social capital has a positive effect on e-WOM intention.

2.3. Linking WeChat Use Behaviors to e-WOM Intention

The intrinsic social nature of social commerce urges scholars to obtain theoretical insights from relevant social theories [47]. Social exchange theory, as the most influential theory in the field of social interaction, assumes that people are more likely to manage their behaviors in the hopes of gaining enough benefits so as to exceed the costs they paid for these actions [48]. It’s evident that the increasing prevalence of SNS substantially reduces the social expenses of online users, ultimately promoting the frequency and intensity of SNS use behavior. Additionally, network characteristics of convenience, synchronization, and persistence have become a huge driving force for plenty of users to exchange information in the online community [49,50]. Via official accounts and Moments on WeChat, consumers can browse original content posted by other people such as pictures, links, and videos, and then press the “like” and “share” buttons to express personal interest and agreement [13]. In the social media-saturated environment, WeChat is emerging to play an enormous role in improving the effectiveness of communication and interaction, along with pushing the transmission of e-WOM in social networks [4,9].

Prior studies on marketing have actively explored the relevance between users’ behavior and e-WOM intention in order to guide theoretical research and enterprise management [51,52]. Pöyry, Parvinen, and Malmivaara [29] found that communicative members may provide more feedback including liking and sharing behavior after browsing the relevant content, resulting in a higher willingness to buy. There is sufficient evidence to suggest that messages received in the process of seeking substantially affects the purchase decisions of consumers [53,54]. Apart from seeking behavior, scholars have examined other types of behavior, for example, Kudeshia, Sikdar, and Mittal [52] conducted a study on Facebook clients’ liking behavior on fan pages to elaborate the positive link between liking behavior and e-WOM about the brands. Likewise, Liang and Yang [13] demonstrated that company-hosted official WeChat accounts are a key factor for consumers’ e-WOM intention; the more frequent the participation, the higher the e-WOM intention. As social exchange theory explains, consumers active on SNS are encouraged to adopt beneficial activities like sharing brand-related messages for social rewards or respect [10]. Hence, this article formulates the following:

Hypothesis 6 (H6).

Seeking will positively relate to e-WOM intention.

Hypothesis 7 (H7).

Sharing will positively relate to e-WOM intention.

Hypothesis 8 (H8).

Liking will positively relate to e-WOM intention.

3. Research Methodology

3.1. Research Model

Drawing on social capital theory and previous studies on WeChat use behavior [7,14,27], this paper constructs a research model (presented in Figure 1) to systematically explore the factors that influence individuals’ e-WOM intention. WeChat use behavior is further decomposed into seeking, sharing, and liking, and is expected to exert direct effects on social capital and e-WOM intention. Social capital is comprised of bonding and bridging social capital and the model also illustrates the direct correlation between social capital and e-WOM intention.

3.2. Sample and Data Collection

Prior literature has argued that adolescents and young adults are valued as the significant customer segment that could play a vital role in the early adoption of products or services via their corresponding online social networks [55]. Mishra et al. [56] also pointed out that adolescents can be perceived as the crucial group in driving the internet economy. Furthermore, younger generations make up the majority of WeChat users and engage in various WeChat activities more frequently [5,57]. According to a previous report [6], 73% of SNS users in China are between 13 and 34 years old. Consequently, Chinese young people aged 14–33 were identified as the target population for this survey. Based on earlier literature in the context of WeChat [5,6], this paper required that the qualified respondents should be monthly active WeChat subscribers who sent out at least one message last month.

The empirical data needed for further research was gathered from the professional electronic questionnaire survey platform (www.wjx.cn) (accessed on 12 June 2021). Meanwhile, all participants were clearly informed about the general objective of this proposed study in advance and were assured that their responses would be entirely voluntary and anonymous. A small number of monetary rewards were set up to encourage more participants to fill out the online-based questionnaire actively and earnestly. In the data collection from 22 July 2020 to 11 August 2020, a total of 271 valid responses were obtained after eliminating the incomplete responses or the answers with insufficient completion time.

Table 1. Demographic characteristics of research sample (N = 271).

	Frequency	%
Gender
Males	96	35.4
Females	175	64.6
Age
14–18	12	4.4
19–23	166	61.3
24–28	83	30.6
29–33	10	3.7
Education
Middle school	7	2.6
High school	19	7.0
University	115	42.4
Master	127	46.9
Doctor	3	1.1
How long have you used WeChat?
Below 1 year	3	1.1
1–2 years	15	5.5
2–3 years	23	8.5
Above 3 years	230	84.9
Daily time spent on WeChat
Below 1 h	46	17.0
1–2 h	73	26.9
2–3 h	63	23.3
Above 3 h	89	32.8
How many friends do you have on WeChat?
Below 100	60	22.1
100–200	86	31.7
201–300	54	19.9
301–400	21	7.8
Above 400	50	18.5

represents the demographic statistics of the research sample in detail. The respondents were 35.4% male and 64.6% female, and the age of the majority was between 19 and 23 (61.3%). The participants are inclined to have a good educational background (42.4% of them are undergraduate students, and 46.9% are postgraduate students). A total of 84.9% of the respondents had been using WeChat for more than three years, 56.1% of them spent more than two hours on daily activities on WeChat, and 77.9% have over 100 friends on WeChat.

3.3. Measurement

The questionnaire designed to collect data consists of two sections. The first section intends to gather basic information of respondents, like age, gender, education, etc. The second section comprises six constructs, comprising seeking, sharing, liking, bonding social capital, bridging social capital, and e-WOM intention. This paper used validated measures from the previous literature and made some minor modifications to better suit the research background. As the language of the original questionnaire is in English, this paper has invited three professional scholars to translate it into Chinese to ensure its validity. Moreover, five users with rich shopping experiences on WeChat were invited to participate in a pilot survey during the initial questionnaire compilation. Based on their feedback, some items were revised to guarantee the clarity and comprehensibility of the questionnaire.

3.3.1. Seeking

Three items were taken from the relevant study of Chu and Kim [58] and carefully modified to assess WeChat users’ seeking behavior. The paper selected this measure because it is the most commonly utilized scale of specific media usage, and it has been validated with different samples [3,30]. The example items included “I get opinions from other users on WeChat before I purchase new goods (SE1),” “When I think over new goods, I search for other users’ advice on WeChat (SE2),” and “I will be more comfortable in selecting goods when I have obtained other users’ suggestions on WeChat (SE3)”. A 5-point Likert scale (1 = strongly agree, 5 = stringly disagree) was used to rate the abovementioned three items. This scale revealed a high level of reliability (Cronbach’s alpha = 0.912, over 0.7).

3.3.2. Sharing

This study employed extant valuable scale items from Ridings et al. [59] to measure WeChat users’ sharing behavior. The specific items are “I want to share my comments with other users on WeChat (SH1),” “I intend to share my opinions with other users on WeChat (SH2),” “I try to share my viewpoints with other users on WeChat (SH3),” and “I like to advise other users on WeChat (SH4)”. The work used such measures because it can offer conceptual clarification regarding sharing behavior and its motivation in an online setting [20,44,60]. This paper used a five-point Likert scale to estimate these four items with 1 and 5 signifying strongly agree and strongly disagree, respectively. Cronbach’s alpha for this four-item scale was 0.885, over 0.7, which demonstrated good reliability.

3.3.3. Liking

Three five-point Likert-type items of Gan [4] have been revised to assess participants’ agreement with statements about liking behavior, from very likely (1) to very unlikely (5). All scale items have been used extensively and reported satisfactory validity and reliability [29,31,32]. Items were composed such as “I often click ‘like’ on WeChat (LI1),” “I always click ‘like’ on WeChat (LI2),” and “In general, I like to click ‘like’ on WeChat (LI3)”. Cronbach’s alpha for this three-item scale was 0.914, over the recommended minimum critical value of 0.7, which manifested a high degree of reliability.

3.3.4. Bonding Social Capital

Five items from the existing validated scale of Horng and Wu [10] were adapted to measure WeChat users’ perceived bonding social capital. The example statements consist of “I have some trusted friends on WeChat to assist me in solving problems (BO1),” “I can ask someone on WeChat for advice when making important decisions (BO2),” “When I feel lonely, I can communicate with several friends on WeChat (BO3),” “Individuals whom I contact on WeChat may put their reputation on the line for me (BO4),” and “Individuals whom I contact on WeChat would recommend jobs to me (BO5)”. These five items were anchored on a five-point Likert scale, ranging from “1 = stongly agree” to “5 = strongly disagree”. This scale exhibited good internal consistency (Cronbach’s alpha = 0.893, over 0.7).

3.3.5. Bridging Social Capital

WeChat users’ bridging social capital was measured on seven five-point Likert-type items adapted from the previously developed scales of Zhang, Liang, and Qi [15] (1 = strongly agree, 5 = strongly disagree). The study used this measure because it could clearly show the extent of the heterogeneous and external ties among people [11,18]. Users are asked to rate the following statements: (1) “Make me willing to try new things” (BR1), (2) “Inform me that every person in the world is connected” (BR2), (3) “Make me interested in other members’ ideas” (BR3), (4) “Make me interested in what occurs outside my hometown” (BR4), (5) “Make me believe that I belong to a larger group” (BR5), (6) “Make me feel connected to the bigger picture” (BR6), (7) “Make me want to spend time supporting the routine on WeChat community” (BR7). Cronbach’s alpha for this scale was 0.922, over 0.7, which showed a high degree of reliability.

3.3.6. E-WOM Intention

E-WOM intention was evaluated by three items that indicated high statistical validity and were borrowed from Lee et al. [61]. The research employed the scale because it has been widely applied in the context of investigating the nature of implicit e-WOM communication in social media settings [17,56]. The example items consist of, for example, “I will say positive things about specific products or services to other people in future (EW1),” “I will recommend specific products or services to anyone who seeks advice about these items in future (EW2),” and “I will encourage other people to get specific products or services in future (EW3).” This study used a five-point Likert scale moving from “1 = very likely” to “5 = very unlikely”, to rate these three items. Cronbach’s alpha for this three-item scale was 0.879, over 0.7, which demonstrated good reliability.

4. Data Analysis Strategy

Firstly, cleaning and pretreatment of data were implemented in Microsoft Excel to delete substandard answers. Secondly, this paper uses SPSS 22 to carry out the descriptive statistical analysis and the test of common method variance. Lastly, the structural equation modeling (SEM) approach is utilized to examine the proposed hypotheses using AMOS 22. The two-step method can make the outcomes more meaningful and trustworthy [62]: A confirmatory factor analysis (CFA) is adopted to test the research model (including overall model fit, construct reliability, and validity) and then SEM is employed to verify the structural relationships. This structural equation model is appropriate in the current research since this method could take a confirmatory approach to analyzing survey data by stating certain associations among main variables [14,63]. In addition, utilizing SEM also permits the scholars to gauge the factorial validity of the questions that constitute selected constructs by clarifying the extent to which it is likely to assess identical concepts and variables [3,64].

5. Results

CFA was applied to measure the model. As conferred in

Table 2. Fit indices for the measurement model.

Model Fit Measures	Model Fit Criterion	Index Value	Good Model Fit (Y/N)
Absolute Fit Indices
RMSEA	<0.08	0.067	Y
RMR	<0.05	0.037	Y
${\mathsf{\chi }}^2$/d.f. (${\mathsf{\chi }}^2=$577.817, d.f. = 260)	<3	2.222	Y
Incremental Fit indices
CFI	>0.9	0.938	Y
AGFI	>0.8	0.808	Y
IFI	>0.9	0.938	Y
TLI	>0.9	0.928	Y

, a reasonable model is assessed by absolute fit indices (${\mathsf{\chi }}^2$/d.f. = 2.222; RMSEA = 0.067; RMR = 0.037) and incremental fit indices (CFI = 0.938; AGFI = 0.808; IFL = 0.938; TIL = 0.928). All those indices have met the standard suggesting the acceptable model fit [65]. As for the reliability of the measurement model, Cronbach’s alpha and composite reliability (CR) are good indices to test internal consistency. The Cronbach alpha and CR values have all outstripped the critical value of 0.7, revealing good reliability [66]. This paper inspects the validity through convergent validity and discriminant validity. Convergent validity was evaluated by standardized factor loadings, average of variance extracted (AVE), and squared multiple correlations (SMC) [66,67]. In

Table 3. Summary of confirmatory factor analysis.

Constructs and Items	Loading (>0.7)	SMC (>0.5)	Cronbach’s Alpha	CR (>0.7)	AVE (>0.5)
Seeking (SE)			0.912	0.916	0.784
SE1	0.904	0.817
SE2	0.936	0.876
SE3	0.812	0.659
Sharing (SH)			0.885	0.887	0.664
SH1	0.786	0.618
SH2	0.894	0.799
SH3	0.776	0.602
SH4	0.797	0.636
Liking (LI)			0.914	0.915	0.782
LI1	0.870	0.757
LI2	0.904	0.817
LI3	0.878	0.771
Bonding social capital (BO)			0.893	0.896	0.634
BO1	0.788	0.621
BO2	0.838	0.702
BO3	0.767	0.589
BO4	0.836	0.698
BO5	0.749	0.560
Bridging social capital (BR)			0.922	0.923	0.632
BR1	0.808	0.653
BR2	0.823	0.677
BR3	0.844	0.713
BR4	0.791	0.626
BR5	0.801	0.641
BR6	0.756	0.571
BR7	0.737	0.543
e-WOM intention (EW)			0.879	0.882	0.714
EW1	0.885	0.784
EW2	0.894	0.799
EW3	0.748	0.560

Notes: SMC, squared multiple correlations; CR, construct reliability; AVE, average variance extracted.

, the standardized loading estimates ranged from 0.737 to 0.936 above 0.7, which shows that all factors converge on the latent construct. AVE highlights the average variance of item extraction loaded on the construct and the values of all constructs have exceeded 0.5, providing evidence for good convergent validity [66,67]. SMC indicates the variance of a measured variable explained by a latent construct and all values are higher than 0.5, supporting adequate convergence [66].

Table 4. Descriptive statistics and discriminant validity.

	M	SD	SE	SH	LI	BO	BR	EW
SE	2.794	0.968	(0.885)
SH	2.807	0.925	0.578 **	(0.815)
LI	2.950	1.004	0.324 **	0.502 **	(0.884)
BO	2.565	0.884	0.581 **	0.635 **	0.484 **	(0.796)
BR	2.647	0.829	0.416 **	0.556 **	0.463 **	0.679 **	(0.795)
EW	2.731	0.890	0.573 **	0.559 **	0.404 **	0.687 **	0.654 **	(0.845)

Notes: Double asterisk (**) represents p < 0.001; SE, seeking; SH, sharing; LI, liking; BO, bonding social capital; BR, bridging social capital; EW, e-WOM intention; M, mean; SD, standard deviation; () = the square root of each construct’s AVE.

illustrates the analysis consequences of discriminate validity. The square root of each variable’s AVE (diagonal entries) is higher than the other correlation coefficients (off-diagonal entries), which proves good discriminate validity. In conclusion, the measurement model established in this study has an eligible model fitting degree, good reliability, sufficient convergence, and discriminant validity.

Common method variance (CMV) refers to the artificial covariation between prediction variables and criterion variables caused by similar data sources or raters and the same measurement environment. This covariance is a systematic error, which seriously interferes with the research results and potentially misleads the conclusions [68]. To solve such a problem, the questionnaire ensures the anonymity and confidentiality of participants and informs them that there are no right or wrong answers. This method may reduce respondents’ evaluation apprehension and social expectations, which can control the common method bias within an acceptable range. Moreover, all of the samples have experiences of WeChat usage, further reducing the possibility of CMV appearance [69]. Two methods were used in this study to identify CMV. First, Harman’s single-factor test shows that six clearly interpretable factors were found through exploratory factor analysis (EFA) and no single factor has a factor loading over 50%, demonstrating CMV hardly affects the validity [70]. Second, comparing a single factor to multifactor structure, the CFA results ($\Delta {\mathsf{\chi }}^2$ = 1761.346, Δ.f. = 16, p-value = 0; multi-factor: ${\mathsf{\chi }}^2$ = 577.817/d.f. = 260; one-factor: ${\mathsf{\chi }}^2$ = 2339.163/d.f. = 276) indicate that CMV is not an issue [71]. In general, the effect of CMV is acceptable in this paper.

AMOS 22 was applied to further examine the structure model. The model fit indices (${\mathsf{\chi }}^2$ = 1.847 < 3; RMSEA = 0.056 < 0.08; RMR = 0.036 < 0.05; CFI = 0.957 > 0.9; AGFI = 0.845 > 0.8; IFL = 0.957 >0.9; TIL = 0.950 > 0.9) are all above the threshold showing a good model fit. Then, this paper assessed the structural model to testify the proposed hypotheses. Seeking ($\beta$ = 0.297, p < 0.001), sharing ($\beta$ = 0.378, p < 0.001), and liking ($\beta$ = 0.198, p < 0.001) exert significant and positive impacts on bonding social capital. Thus, H1a, H2a, and H3a are supported. Sharing ($\beta$ = 0.379, p < 0.001) and liking ($\beta$ = 0.200, p < 0.001) are the powerful predictors of bridging social capital. Hence, H2b and H2c are statistically supported. Additionally, with regard to the standardized direct effects (STD) of the structural model [50], sharing behavior seems to play a more dominant part than seeking and liking in promoting bonding and bridging social capital. Bonding social capital ($\beta$ = 0.303, p < 0.001), bridging social capital ($\beta$ = 0.273, p < 0.001), and seeking ($\beta$ = 0.187, p < 0.001) also significantly and positively affect e-WOM intention. Thus, H4, H5, and H6 are statistically corroborated. However, seeking is found to have a minute influence on bridging social capital. Therefore, H1b is not supported. The impacts of sharing and liking on e-WOM intention are not significant. Thus, H7 and H8 are rejected. The outcomes of hypotheses testing and path coefficients are displayed in

Table 5. Statistical results of structural model.

Hypotheses	Paths	Path Coefficient	p-Value
H1a	Seeking → Bonding social capital	0.297	0.000 **
H1b	Seeking → Bridging social capital	0.113	0.075
H2a	Sharing → Bonding social capital	0.378	0.000 **
H2b	Sharing → Bridging social capital	0.379	0.000 **
H3a	Liking → Bonding social capital	0.198	0.000 **
H3b	Liking → Bridging social capital	0.200	0.000 **
H4	Bonding social capital → e-WOM intention	0.303	0.000 **
H5	Bridging social capital → e-WOM intention	0.273	0.000 **
H6	Seeking → e-WOM intention	0.187	0.000 **
H7	Sharing → e-WOM intention	0.035	0.619
H8	Liking → e-WOM intention	0.005	0.922

Note: Double asterisk (**) represents p < 0.001.

and Figure 2.

Following H1–H8, this study explores the mediating effects of bonding and bridging social capital between seeking and e-WOM intention. The bootstrapping results verify that the relationship between seeking and e-WOM intention is partially mediated by bonding social capital (β = 0.087, p = 0.015 < 0.05; 95% confidence interval (CI) = [0.020, 0.190]).

6. Discussion

6.1. Summary of the Key Results

This current paper initially carries out empirical research among Chinese young people in ascertaining the underlying mechanism between WeChat use behavior, social capital, and e-WOM intention in mainland China. In addition, building upon social capital and social exchange theoretical frameworks, this paper establishes a conceptual model that examines the moderating roles of bonding social capital and bridging social capital for the intricate relationships among distinct WeChat use behaviors and e-WOM intention. Through the web-based study, this research offers several insights into WeChat users’ e-WOM intention in the new media context.

First, the significant outcomes for Hypotheses 2–3 corroborate the path effects suggesting that sharing and liking can positively influence bonding and bridging social capital. The results are in accordance with previous studies that identified social media use as a key factor contributing to individual’s social capital in the computer-mediated environment [7,21,72]. Regarding seeking behavior, H1a is supported while H1b is rejected, showing that seeking only has effects on bonding social capital. The fundamental explanation is that Chinese users prefer to trust the messages received from their close friends or relatives and have less confidence in commentaries sent by regular friends or strangers [6,9,13]. This is a new finding that offers empirical support to theorize that seeking behavior normally occurs among acquaintances, which can be beneficial for bonding social capital yet makes few contributions to bridging social capital. Additionally, as a convenient and efficient way to express users’ perspectives towards user-generated content, sharing and liking behavior have acquired huge popularity among young groups and promote the accumulation of social capital.

Second, as supported by the significant results for H4 and H5, both bonding and bridging social capital exhibit positive impacts on e-WOM intention, which is consistent with González-Soriano, Feldman, and Rodríguez-Camacho [25]. Therefore, this study further indicates that individuals with higher social capital are more prone to communicate their own e-WOM experience with other individuals, no matter whether they are intimate friends or strangers. Additionally, the outcomes of the standardized direct effects explain that bonding social capital is the more influential factor of e-WOM intention than bridging social capital, which is consistent with the discovery of Chu and Kim [58]. As online access to information becomes more convenient and diverse, users can easily share reviews of specific products or services with close friends, which can aid in their shopping decisions.

Third, H6 is supported in this paper indicating that seeking is a significant predictor of e-WOM intention. This conclusion is congruent with information seeking and social media use literature [3,51]. As the findings show, people with active seeking behavior may have more experience in searching for opinions and comments on concrete products or services and will display a stronger capability in selecting information tallying with their demands [73]. Consequently, they can better realize the benefits of useful recommendations, which may enhance their intention to share related e-WOM experiences [44,58]. However, contrary to our expectation, the results for H7 and H8 are non-significant, implying that sharing and liking cannot affect e-WOM intention. The possible reasons are twofold. First, social image has been proven to affect users’ behavioral intentions in SNS [60,74], which impels Chinese young people to be more cautious about sharing and liking e-WOM experiences visible to other WeChat members, including elders and teachers. Second, since clicking the “like” button is an easy and quick way to interact with others, it’s difficult to determine whether people who enjoy liking behavior will be willing to spend time on sharing or posting e-WOM experiences.

Finally, this paper also discovers that bonding social capital can be regarded as the mediator of the associations between seeking and e-WOM intention, which is similar to an extant study [10]. In the context of online communication, bonding social capital may play a more crucial role in mediating the effect of seeking on WeChat users’ intention to interchange experience and evaluation of specific commodities. Furthermore, users are more likely to communicate with reliable members in closed social media such as WeChat and bonding social capital can make greater contributions by enhancing the credibility of information [15,63].

6.2. Theoretical and Managerial Implications

Based on the obtained results, this article offers several theoretical contributions to the emerging literature on e-WOM engagement and online social commerce consumer behavior [27,58]. First, the majority of existing studies on e-WOM intention has mainly concentrated on the potential benefits of general social media use [5,47,75], but neglected the possible impact of distinct patterns of WeChat use on individual-level social capital and e-WOM intention. Thus, by dividing WeChat use behaviors into three specific categories including seeking, sharing, and liking, this current study represents an initial attempt to systematically explore how these different patterns of WeChat use behaviors could influence social capital and e-WOM intention in the online environment. Through determining various behaviors’ distinct impacts, the discovery of this article supplies interesting insights into the research of WeChat users’ behaviors and provides a fresh way to study e-WOM communication. Second, while numerous researchers have identified the direct association between social media adoption and e-WOM intention [13,17,52], few studies have been done to systematically explore the underlying mechanism between these scaled variables. Therefore, this paper expands our comprehension by demonstrating that seeking exhibits an indirect association via the mediation of bonding social capital in the WeChat context. Moreover, the findings echo the theoretical proposition that e-WOM has become the vital source of social capital [15], and further support the idea that individuals are more likely to engage in e-WOM due to personal desire to forge relationships with others and enhance bonding social capital. Third, this study contributes by establishing a comprehensive conceptual model that probes the antecedents of social capital and e-WOM intention on WeChat from the perspectives of social capital theory and social exchange theory in a SNS environment. Accordingly, the outcomes of the present study may enrich the body of relevant studies and provide a novel theoretical angle for following studies further related to the emergence and development of SNSs.

From a managerial perspective, this article also provides some important guidance for multiple stakeholders, including consumers, WeChat owners, and companies in the spreading process of e-WOM. First, as discussed earlier, for consumers who want to develop and maintain their social capital, sharing behavior tends to be more effective than seeking and liking. Our results suggest that marketers could design attractive social activities for WeChat users that permit them to communicate and engage, further strengthening interpersonal connectedness [13,76]. Second, the empirical examination of the users’ actual WeChat use provides evidence that social capital could play a significant role in users’ continuance intention on social media [74]. For WeChat owners who intend to retain users, this paper thus suggests that they improve the facility of the application so that users can easily manage their behavior to sustain social relationships. Lastly, this study further reveals that seeking behavior exerts a significant influence on users’ e-WOM intention. Thus, this result suggests that it’s necessary for social media companies to set up an interactive area to the product detail page independent of the normal comment area to encourage users to seek useful information and build strong consumer–brand relationships [77]. Further, with respect to e-WOM communication, marketers could also leverage users’ demand for meaningful information by offering value-added product or service information that conveys superior customer value when disseminated.

6.3. Limitations and Recommendations for Future Research

Although the research has shed fresh light on future studies, there are still several limitations remaining to be discussed. Firstly, this paper only focuses on Chinese young people’s WeChat use behavior, social capital, and e-WOM intention, hence these research findings only conform to the characteristics of young groups and cannot apply to other age groups. As potent consumers, the middle-aged population segment’s attitudes and behaviors differ from young people [5], so this article calls for an extension of research on other populations. Secondly, the cross-sectional data collected in this article has limited our further analysis of WeChat use behavior. Thus, it is urgent to launch longitudinal research with a larger time span to precisely evaluate how WeChat users’ long-term behaviors imperceptibly affect the accumulation of social capital. Lastly, this article only concentrated on the effects of WeChat use behavior and social capital on e-WOM intention. However, other key factors such as information richness [78], hedonic or utilitarian motivation [13], and self-enhancement [25] remain worthy of consideration and research. Scholars can cluster the abovementioned variables into the drivers of e-WOM intention and the extended outcomes may offer evidence for enterprises to obtain competitive advantages. Besides, along with the recent interest in fuzzy-set qualitative comparative analysis (fsQCA) in information systems and digital marketing [27,79], future studies could also employ fsQCA to investigate which of the factors from WeChat use behavior and bonding and bridging social capital, or their combinations, are necessary and sufficient to interpret e-WOM intention [80] or how personal demographics could be utilized as the condition that would enable case by case analysis. For these reasons, e-WOM on social media requires further research, especially considering its dynamic and interactive communication processes.