Jin Et Al., 2021

Information Processing and Management 58 (2021) 102413
Contents lists available at ScienceDirect
Information Processing and Management

journal homepage: www.elsevier.com/locate/infoproman
The differential effects of trusting beliefs on social media users’

willingness to adopt and share health knowledge
Xiao-Ling Jin a, Mengjie Yin a, Zhongyun Zhou b, *, Xiaoyu Yu a
a
Information Management Department, Shanghai University, Shanghai, China
b
Department of Management Science and Engineering, School of Economics and Management, Tongji University, Shanghai, China
A R T I C L E I N F O A B S T R A C T
Keywords: The outline for the “Healthy China 2030” emphasizes social media use in health knowledge
Health knowledge communication communication. Health knowledge communication enabled by social media relies heavily on
Social media recipients’ willingness to adopt and share health knowledge, which is consequently determined
Knowledge adoption willingness
by their trust in content, source, and the platform. However, little is known about the relative
Knowledge sharing willingness
Trusting beliefs
importance and differential effects of these trusting beliefs (content credibility, source credibility,
Differential effects and institution-based trust) on health knowledge adoption versus sharing willingness in social
media. This study examines this critical issue in the context of WeChat Official Accounts (WOAs)
in China. Drawing upon the trust and information communication literature, we develop a
research model and test the proposed hypotheses with data collected from an online survey of
2086 WeChat users. The results reveal that health knowledge adoption (sharing) willingness is
mainly related to source credibility (institution-based trust). Moreover, content credibility has a
stronger relationship with adoption willingness than with sharing willingness, while institution-
based trust shows a stronger relationship with sharing willingness than with adoption willingness.
These findings extend the literature on trust, information communication, and communication
process, and provide rich insights for healthcare institutions to effectively utilize social media
(such as WOAs) on health knowledge communication.
1. Introduction
Leveraging modern information technologies to enhance public health has become a national strategy in China, as specified in the
“Healthy China 2030” initiative released in 2016. “Healthy China 2030” notes that it is important to use various media, especially new
media, to facilitate health education and health science knowledge popularization for improving the public’s health literacy (Gong,
2017). Health knowledge refers to popular science information, including medical knowledge and technical terms, for daily treatment
or disease prevention. Health knowledge communication has been proven to be essential for health promotion and disease prevention
education (Heo et al., 2018). For example, more than 37% of events in the published hospital pages were related to the dissemination
of widespread health science knowledge to respond to COVID-19 on social media in China (e.g., how to wear a mask, wash hands, and
quarantine at home) (Yan et al., 2020). This kind of health knowledge can help improve the public’s health literacy, reduce infection
risks, change their attitudes and behaviors (Park et al., 2013; So et al., 2016), and eventually enhance their health conditions.
However, China is facing severe health knowledge communication challenges, especially in terms of the tension between the
* Corresponding author at: 1410 Tongji Building (Block A), 1500 Siping Road, Tongji University, Shanghai, China.
E-mail address: philzhou@tongji.edu.cn (Z. Zhou).
https://doi.org/10.1016/j.ipm.2020.102413
Received 2 May 2020; Received in revised form 4 October 2020; Accepted 12 October 2020
Available online 23 October 2020
0306-4573/© 2020 Elsevier Ltd. All rights reserved.
X.-L. Jin et al. Information Processing and Management 58 (2021) 102413
public’s strong need (demand) and the lack of sufficient sources (supply) for professional health knowledge. On the demand side, the
proportion of Chinese residents with an adequate level of health literacy has reached 19.2 percent. However, there is still a consid
erable gap to the overall target (30 percent) specified in the “Healthy China 2030” initiative. Moreover, the Chinese population is aging
rapidly (Chang, 2020). By the end of 2019, over three-quarters of this elderly population suffered from one or more chronic diseases
(Wang, 2019). Also, diseases (chronic and acute) are not unique to the elderly, and relatively young people are also in increasing need
of scientific knowledge for health care, especially given that the sub-health state is getting more common (Health News & DXY, 2019).
This phenomenon brings a strong need of the public for health knowledge to guide their daily care. However, there is a general lack of
useful channels for the public to access professional and scientific knowledge about healthcare. On the supply side, the health
knowledge services provided by healthcare professionals, hospitals, and communities are much insufficient. Healthcare institutions
such as hospitals usually lack official channels and incentive mechanisms for doctors to take part in health knowledge communica
tion1. Moreover, the insufficiency and uneven distribution of healthcare resources in China have made doctors overwhelmed with
providing necessary medical services to patients poured into these hospitals (Deloitte, 2017). This makes it difficult for healthcare
professionals such as doctors to engage in health knowledge popularization.
In recent years, the development and widespread use of social media suggest a possible solution for this demand-supply mismatch.
Social media can help healthcare institutions (e.g., hospitals) and professionals (e.g., doctors) to engage in a direct dialogue with the
target public easily and cost-effectively (Gong, 2017). Moreover, the public in China has increasingly relied on social media such as
WeChat Official Accounts as credible and useful sources for acquiring all kinds of information, including health knowledge (Health
News & DXY, 2019). As a result, social media has become an important and novel channel for health knowledge communication (Hu
et al., 2017), not only in China but also in many other countries wherein individuals use social media to acquire health knowledge
(WHO, 2016).
In sum, health knowledge communication via social media is proposed as a remedy to issues such as low public health literacy and
insufficient supply of health knowledge. However, the influence of health knowledge cannot be realized unless it is effectively diffused
from experts to the population at large. Health knowledge communication enabled by social media relies heavily on recipients’
willingness of knowledge adoption and sharing, representing two important but different behaviors involved in the process of health
knowledge communication. Only with a deep understanding of the factors for recipients’ willingness to adopt and share, social media
operators and content creators can effectively manage and create content to facilitate health knowledge communication, thus striking
to influence recipients’ health conditions and behaviors.
Despite this practical pertinence, this issue has received little attention in the health literature. Most of the existing research focused
on discussing how social media are used by doctors or patients for enhancing professional-to-user or user-to-user communications,
especially surrounding professional healthcare services or a specific disease (e.g., cancer) (Bansal et al., 2010; Guo et al., 2017; Liu
et al., 2020; Mano et al., 2014; Wang et al., 2019; Zhang et al., 2020). However, little research has examined health knowledge
communication across a broad spectrum of health-related topics. Shang et al., and Zuo (2020) examined the effect of health belief,
argument quality, and source credibility on health information-sharing intention; Zhao et al., and Chen (2020) investigated the effect
of eHealth literacy (high or low) and content valence (positive or negative) on the intention to share health articles. However, these
previous studies did not directly tap into technological cues, such as users’ trust in the medium platform (i.e., institution-based trust),
which has been suggested to be an important predictor of social media usage (Salehan et al., 2016). In general, the relative importance
of these trusting factors to information communication was not explored in the literature. Moreover, information (knowledge)
adoption (Tseng & Wang, 2016; Zhang et al., 2016) and sharing (Oostervink et al., 2016; Wang et al., 2017) have been investigated
separately in prior research. Although a few studies focused on the difference in sharing and adoption from levels of online activeness
(Muntinga et al., 2011), little research has been devoted to understanding how these two behaviors differ in concepts and antecedent
mechanisms. Motivated by these research gaps, this study aims to study the difference in the factors for content recipients’ willingness
of general health knowledge adoption and sharing in social media.
Trust plays a critical role in content recipients’ willingness to share and adopt health knowledge propagated in social media. The
content recipients who usually do not have professional health knowledge may feel uncertain about the content’s scientific truth
fulness (Li & Wang, 2018). Prior research suggests trusting beliefs can reduce their perceptions of uncertainty and risk, thereby
enhancing their behavioral willingness (McKnight et al., 2002). These important elements in professional communication processes
are the source, content, medium, and recipients (Hovland, 1948; Shannon, 1948; Zhang et al., 2014). Trusting beliefs (e.g., trust in
information quality, source expertise, and social media) have been found to be determinants of recipients’ response to health
knowledge communication (Lin et al., 2016; Yi et al., 2013). Furthermore, prior evidence shows that online information users’ trusting
beliefs can be explored from three different areas: source credibility, message credibility, and medium credibility (Borah, 2014).
Investigating the differential effects of trusting beliefs on behavioral willingness provides useful insights for both researchers and
practitioners. However, few research endeavors have been devoted to understanding the relative importance of these trusting beliefs in
triggering knowledge sharing willingness and adoption willingness.
Therefore, our research objectives are to conceptualize two important, yet distinct, the behaviors involved in the process of health
knowledge communication (e.g., sharing and adoption), and examine the relative importance of various trusting beliefs in triggering the be
haviors involved in the process of health knowledge communication. Specifically, this study intended to examine the differential effects of
three trusting beliefs (content credibility, source credibility, and institution-based trust) on social media users’ willingness to share and
1
https://www.drc.gov.cn/DocView.aspx?chnid=1&leafid=224&docid=2899377.
2
adopt health knowledge.
2. Research background and theoretical foundation
2.1. Health knowledge communication in social media: adoption versus sharing
As the most popular social media platform in China, WeChat creates a convenient, economical, and effective way to promote health
knowledge communication. Considering its powerful ability to incite participation, many healthcare organizations have registered
WeChat Official Accounts (WOAs) to propagate health knowledge. These healthcare organizations use WOAs to broadcast content in
various formats (e.g., text, audio, image, video, and rich media) to all or an intentionally selected subset of their followers (i.e., WeChat
users subscribing to the WOAs) (Wikipedia 2019). There are over ten million WOAs in WeChat, attracting about 80% of WeChat users.
WOAs have become the main channel for the public to access health information, with the proportion of public use exceeding 80%
(Health News & DXY, 2019). WeChat users who receive and read the content from the subscribed WOAs can share it with friends, in
WeChat groups, and WeChat Moments. WeChat Moments allow users to share information (mostly from public accounts) and engage
with (e.g., like, comment on, adopt, or re-share) friends’ posts (Yogesh Khetani, 2019). With the widespread use of WOAs in health
knowledge communication, it is important to understand health knowledge communication (adoption and sharing) among WeChat
users.
We next define knowledge adoption and sharing willingness as reflecting distinct behaviors. In the context of this research,
knowledge adoption willingness refers to the extent to which a recipient agrees with the action suggested in the received content, after
assessing its validity (Wei &Watts, 2008). In comparison, knowledge sharing willingness refers to the recipient’s subjective likelihood to
voluntarily share the content with specific friends or relatives, or within his/her moments in WeChat. To understand the differences
between health knowledge adoption and sharing in this context, we draw upon the communication literature and propose that the key
elements in the communication process include the content (i.e., health knowledge), source (i.e., health professionals), medium (i.e.,
social media platform), and recipient, among others (Hovland, 1948; Milkman & Berger, 2014; Shannon, 1948; Zhang et al., 2014).
This theoretical perspective concerns which elements of the communication process (e.g., content, source, recipient, and medium) are
primarily involved in the focal behavior (i.e., knowledge sharing versus adoption).
In the context of this research, health knowledge adoption consists of the process of accepting knowledge, wherein the content is
internalized by the recipient (Wei &Watts, 2008), reflecting the recipient-content interaction primarily. While it may also involve
elements other than content and recipients, these elements’ importance in determining knowledge adoption is likely low. In com
parison, knowledge sharing is a critical mechanism of knowledge communication in social media (Shi et al., 2014). It is primarily a
social exchange process involving three parties, including the content sharer, source (knowledge creator), and recipients (Shi et al.,
2014). In this sense, it regards recipients’ interactions with not only the content (human-content interaction) but also, more impor
tantly, with the other subjects including the sharer and other potential recipients (interpersonal interaction) in the context. Prior
research suggests that, in addition to and beyond the characteristics of the content and source, the characteristics of recipients and the
medium through which the recipients are reached have essential impacts on knowledge sharing willingness (Barasch & Berger, 2014;
Berger & Iyengar, 2013; Berger & Milkman, 2012; Milkman & Berger, 2014). In sum, knowledge adoption primarily involves the
adopter’s interaction with the content (and possibly source), whereas knowledge sharing regards interpersonal interactions between
the sharer and recipients in specific environments surrounding a media platform.
2.2. Trusting beliefs toward content, source, and platform
Trusting beliefs refer to perceptions that the trustee has attributes beneficial to the recipients (McKnight et al., 2002). People’s
trusting beliefs are suggested to be important for reducing their perceptions of uncertainty in content evaluation, thereby enhancing
their behavioral willingness (Cappella, 2017). While trusting beliefs are often concentrated on person-to-person and
organization-to-organization relationships, prior research suggests that people can also ascribe their trust in non-human entities such
as technologies or media (Lankton et al., 2015; Vance et al., 2008). Typically, key elements of professional communication processes
include the source, content, medium, and recipients (Hovland, 1948; Shannon, 1948; Zhang et al., 2014). Prior research on science
communication has revealed the effects of content-, source-, or audience-related factors such as content characteristics, creator
characteristics, and health literacy (Crook et al., 2016; Milkman & Berger, 2014). Moreover, health knowledge communication in
volves the recipient’s trust in not only health knowledge and professionals but also the social media platform (Lin et al., 2016; Yi et al.,
2013). In the context of this research, recipients’ trusting beliefs can be directed at the content (content credibility), its source (source
credibility), or the medium platform (institution-based trust).
These three trusting beliefs differ in terms of their target objects, i.e., whether they are content-specific versus context-related.
Content credibility refers to the believability of the content itself (Metzger, 2007). It is a recipient-based judgment that involves
objective judgments of content quality or content element accuracy (Freeman & Spyridakis, 2004). It is based on content charac
teristics (e.g., accuracy, objectivity, currency, and argument quality) but reflects an overall assessment of content by taking all these
characteristics together. Thus, it is likely a better representation of content quality, as a whole, than factors used in prior research that
focuses on specific content characteristics such as argument quality (Sussman & Siegal, 2003). Although the assessment of content
credibility may be enhanced by source-related or contextual factors (Metzger, 2007), content credibility per se is a content-specific
construct.
Source credibility refers to recipients’ perception of the credibility of knowledge sources (Sussman & Siegal, 2003). A source may be
3
Fig. 1. Research model and hypotheses.
perceived as credible because of his/her competence (ability) and trustworthiness (morality). It has been found to play an important
role in recipients’ judgments of cognitive authority (Rieh & Belkin, 1998) and content usefulness (Sussman & Siegal, 2003), eventually
leading to content sharing and adoption. In the context of this research, source credibility is a representation of the relationship
between knowledge transmitter and knowledge source in the process of knowledge communication. Source credibility represents a
content-related (rather than context-related) factor, as the source is related to a specific piece of content rather than the situational
context in general.
In this study, recipients’ trusting belief toward the medium platform is mainly reflected as insitution-based trust. According to the
literature (McKnight et al., 2002), institution-based trust can be defined as the belief that the needed structural conditions are present to
enhance the probability of achieving a successful outcome in an endeavor like health knowledge communication. It is a higher-order
factor reflective of two general dimensions: structural assurance and situational normality (McKnight et al., 2002). Structural assurance
refers to the belief that necessary structures (e.g., guarantees, regulations, promises, legal recourse, or other procedures) are in place to
promote success (McKnight et al., 2002; Shapiro, 1987). Situational normality refers to the belief that “the environment is in proper
order and success is likely because the situation is normal or favorable” (McKnight et al., 2002). Situational normality can be further
decomposed into four sub-dimensions, capturing (1) whether one feels good and comfortable about online activities in general (Vance
et al., 2008); and whether he/she perceives that most knowledge providers (i.e., WeChat WOAs) (2) are experts who can do their job
well (competence), (3) are honest, truthful, sincere, reliable, dependable and can keep commitments (integrity), and (4) will act in the
customer’s best interest, try to help, and be genuinely concerned (benevolence) (Ellonen et al., 2008). In the context of this research,
institution-based trust regards the perceptions about knowledge providers in the platform in general rather than specific pieces of
content or sources. Thus, it is essentially a context-related (rather than content-specific) belief.
In sum, three trusting beliefs differ in their targets (content, source, or medium). These were focused on content-specific (content
credibility and source credibility) vs. context-related. Thus, it is likely that they would exert differential effects on health knowledge
adoption versus sharing willingness, which involves different elements of the communication process including the content, source,
recipient, and medium. Despite some piecemeal evidence (Cheung et al., 2008; Wang et al., 2019; Zhang et al., 2014), this study is one
of the first studies to systematically investigate the differential effects of these trusting beliefs on content-related behaviors.
3. Research model and hypotheses
The research model proposed in this study is shown in Fig. 1. Our basic model proposes that these trusting beliefs are positively
related to health knowledge sharing and adoption willingness. This is primarily because recipients usually do not have sufficient
expertise to evaluate the scientific truthfulness of the health knowledge per se and tend to base their behavioral willingness (of
adoption and sharing) on trusting beliefs. Although all three trusting beliefs are related to willingness toward these two behaviors,
their relative importance on the same behavioral willingness could vary (H1A, H1B, H2A, and H2B). Moreover, for each trusting belief,
we propose that its relationships with adoption willingness versus sharing willingness are not identical (H3A and H3B). Below we
justify each hypothesis in detail.
3.1. The relative importance of the three trusting beliefs relating to health knowledge sharing willingness
Source credibility and institution-based trust are expected to have stronger relationships than content credibility with health
knowledge sharing willingness. Prior research has shown that knowledge sharing can strengthen perspective exchanges and inter
personal connections (Milkman & Berger, 2014). According to social interaction theory, the interaction between individuals (e.g.,
knowledge sharing behavior) is motivated by cost and benefit analysis (Blau, 1964). Individuals might maximize their benefits in the
interaction, and minimize the time cost of exchanging knowledge (Razak et al., 2016). Thus, knowledge sharing willingness is
associated with relatively low cognitive effort. The elaboration likelihood model (an information-processing theory) posits that when
the cognitive effort is low, recipients rely more on the peripheral rules (e.g., source credibility) to process information, and pay less
4
attention to the content (Petty & Cacioppo, 1986). Extending this logic, when considering whether to share a relevant article con
taining health knowledge with friends or relatives, WeChat users do not necessarily ascribe much importance to the believability of the
content in this study. Instead, as long as they feel that the source is credible and the platform has provided the necessary institutional
mechanisms for guaranteeing knowledge quality, they tend to believe it is “safe” to share the received knowledge with other people
who may be interested.
Further, the content-independent, context-related trusting belief (institution-based trust) is expected to have a stronger relationship
with health knowledge sharing than content-specific trusting beliefs (content credibility and source credibility). This is because
knowledge sharing is more about social (interpersonal) interactions between the content sharer and the potential recipient than about
the sharer’s interactions with the content or its source (Crook et al., 2016). In prior research, context-related trusting beliefs are more
important for the behavior that emphasizes social interaction (Zhang et al., 2014). Prior research has shown that trust in blog service
providers plays a vital role in promoting users’ sharing behavior (Chai et al., 2011). A significant consideration for the sharer is
whether the platform (i.e., WeChat) provides necessary institutional mechanisms to support sharing behaviors and whether most
WOAs are competent, honest, and benevolent. If so, the sharer tends to share the content promoted by most (if not all) WOAs;
otherwise, the sharer would hesitate to share anything promoted by WOAs in WeChat. Hence, we propose the following hypotheses.
H1A: Source credibility and institution-based trust have a stronger relationship with health knowledge sharing willingness than content
credibility.
H1B: Institution-based trust has a stronger relationship with health knowledge sharing willingness than source credibility.
3.2. The relative importance of the three trusting beliefs relating to health knowledge adoption willingness
Content credibility is expected to have a stronger relationship with “health knowledge adoption willingness” than source credibility
and institution-based trust. In the context of this research, knowledge adoption primarily refers to recipients following the advice
contained in the received content (Crook et al., 2016). Further, there is evidence that health knowledge adoption requires a high level
of recipients’ cognitive effort (Kim & Yang, 2017; Muntinga et al., 2015). According to the elaboration likelihood model, the high
cognitive effort represents in the central route (wherein recipients carefully consider the issues presented in the message) (Petty &
Cacioppo, 1986). Following this logic, in the context of this research, when considering whether to adopt specific health knowledge
contained in a post, the recipients put in efforts to cognitively elaborate the specific content by reflecting deeply on the real-life pros
and cons of following the advice specified (Sussman & Siegal, 2003). Even though they feel the source is credible and the social
platform has provided necessary institutional mechanisms to secure knowledge adoption, these beliefs are insufficient to motivate
them to content’s advice.
Content-specific trusting beliefs (content credibility and source credibility) are expected to have stronger relationships with health
knowledge adoption willingness than the content-independent, context-related trusting beliefs (institution-based trust). Further,
knowledge adoption is primarily a process of identifying with knowledge, which involves recipient-content interaction. Two primary
considerations for the recipients in deciding to follow the advice are content credibility and source credibility (Wang et al., 2008).
Studies have revealed corroborative evidence that content-related factors are important predictors of knowledge adoption (Cheung
et al., 2008; Hussain et al., 2017; Shen et al., 2014). If so, recipients are likely to follow the advice, even though the medium platform
wherein the content is diffused does not provide necessary institutional mechanisms for secure adoption. Overall, even those who do
not trust the informational environment of WOAS may be willing to adopt certain health knowledge as long as they believe in the
content and its source. Hence, we propose the following hypotheses.
H2A: Content credibility has a stronger relationship with health knowledge adoption willingness than source credibility and institution-based
trust.
H2B: Source credibility has a stronger relationship with health knowledge adoption willingness than institution-based trust.
3.3. The differential relationships of trusting beliefs with sharing willingness versus adoption willingness
As a content-specific trusting belief, content credibility is likely to have a stronger relationship with knowledge adoption that
involves human-content interactions than with knowledge sharing that primarily involves social (interpersonal) interactions (Crook
et al., 2016). Prior research has argued that context-related factors are more critical in the behavior that primarily involves social
interactions (Wang et al., 2019; Zhang et al., 2014). Previous studies also found that content-related factors play an essential role in
promoting users’ knowledge adoption (Cheung et al., 2008; Hussain et al., 2017; Shen et al., 2014). In sum, this study argues that
content credibility has a stronger relationship with adoption willingness than sharing willingness. Hence, we propose the following
hypotheses.
H3A: Content credibility has a relationship with health knowledge adoption willingness than with health knowledge sharing willingness.
We propose that institution-based trust has a stronger relationship with sharing willingness than with adoption willingness. The
study finds that institution-based trust is an important factor affecting recipients’ attitudes toward knowledge sharing behavior
(Salehan et al., 2016). Moreover, as a content-independent, context-related trusting belief, institution-based trust should have a
stronger relationship with knowledge sharing that primarily involves social interactions than with knowledge adoption that primarily
implies human interaction with certain content. Hence, we propose the following hypotheses.
H3B: Institution-based trust has a stronger relationship with health knowledge sharing willingness than with health knowledge adoption
willingness.
However, it is difficult to predict the relative importance of source credibility on sharing willingness versus adoption willingness.
5
Table 1
Profile of the respondents (N=2086).
Characteristic Count Percent Characteristic Count Percent
Gender Female 1062 50.9 Tenure < 3 months 38 1.8

Male 1024 49.1 3-6 months 56 2.7
Age: < 18 17 0.81 6-12 months 181 8.7
18 – 25 556 26.65 1-2 years 551 26.4
26-35 1081 51.82 > 2 years 1260 60.4
36-50 389 18.65 Frequency
51-60 44 2.11
>60 12 0.58 Many times a day 1050 50.3
Education Middle school 19 0.9 Once a day 658 31.5
High school 161 7.7 Once a week 226 10.8
Undergraduate 1665 79.8 Once a month 33 1.6
Master’s degree 224 10.7 Seldom 119 5.7
Doctor’s degree 17 0.8
On the one hand, compared with knowledge adoption, knowledge sharing requires less cognitive effort from recipients; according to
the elaboration likelihood model, recipients use less cognitive effort to examine the information through peripheral cues (e.g., source
credibility) (Petty & Cacioppo, 1986). On the other hand, source credibility is a content-specific heuristic cue that is expected to have a
stronger relationship with knowledge adoption (which regards human-content interactions) (Wang et al., 2008) than with knowledge
sharing (which primarily involves interpersonal or social interactions). As such, the view is contradictory in terms of the relative
importance of source credibility on adoption willingness versus sharing willingness. Thus, we cannot make specific predictions on this
issue but choose to address it in post hoc analysis.
3.4. Relationships between the three trusting beliefs
The trust transfer theory suggests that trust can transfer between targets (Stewart, 2003), e.g., from a known to an unknown target.
For instance, consumers’ trust in an e-commerce platform can enhance their trust in sellers on the platform, eventually leading to a
positive attitude and purchase intention toward specific products (Chen et al., 2016). Following this reasoning, institution-based trust
and source credibility may have positive relationships with content credibility in this study. Moreover, these different combinations of
trust can be reflected in the interaction effects among three trust beliefs (source credibility, institution-based trust, and content
credibility). However, these trust transference effects are already established in the literature and are not the main focuses of this
study. Thus, for the sake of parsimony, we decide not to propose these effects as formal hypotheses but address them in post hoc
analysis.
4. Methodology
4.1. Research design
We used an online survey to collect data from WeChat users, which is appropriate for three reasons. First, it is hardly possible to
acquire archival data in WeChat, which accentuates users’ privacy. Second, a survey is the best way of obtaining data regarding re
spondents’ beliefs and behavioral willingness, while also enjoying the merit of reaching a large sample of respondents and enhancing
the generalizability of research findings (Zhou et al., 2015). Finally, science articles served as study stimuli can enhance the realism of
decision-making situations by providing contextual details.
The science articles were obtained from “Ding Xiang Yi Sheng” and “The Good Doctor Online”, which are among the most popular
healthcare WOAs in China. The two WOAS regularly post a rich set of science-based articles to their followers. The authors of these
articles are typically healthcare professionals from regular hospitals and other healthcare organizations. We extracted 133 articles
posted on these two WOAs over three months using the internet crawler. We chronologically chose the articles containing the author’s
information and assigned them randomly using the random function of the survey website. Each article was treated as a scenario and
administered to 10 to 24 (with a mean of around 16) randomly selected respondents. Respondents read the assigned article and then
answered the questions about our main constructs (as described below) and other variables including demographics, usage experi
ences, and user expertise. They also indicated whether they were followers of the two WOAs and whether they had recently paid
attention to the content topic. In addition, we also collected objective data about the content such as the length of articles (i.e., No. of
words).
4.2. Measurement development
As shown in Appendix A, all constructs were measured with multiple items adapted from prior research to fit this research context.
All the English indicators were translated into Chinese by using the double-translation/back-translation skill from McGorry (2000) to
ensure the Chinese-English questionnaire’s consistency. Items for sharing willingness and adoption willingness were adapted from
Milkman and Berger (2014) and Sussman and Siegal (2003), respectively. Those for content credibility and source credibility were
6
Table 2
Descriptive statistics and psychometric properties (N=2086).
Mean s.d. CC IG IB II IC ISA SC SW AW
CC 3.886 0.742 0.782

IG 3.737 0.871 0.485 0.827
IB 3.596 0.806 0.460 0.574 0.807
II 3.585 0.833 0.428 0.543 0.746 0.786
IC 3.795 0.714 0.472 0.571 0.656 0.652 0.757
ISA 3.666 0.745 0.463 0.587 0.679 0.703 0.691 0.756
SC 3.978 0.659 0.658 0.527 0.459 0.441 0.475 0.487 0.760
SW 3.731 1.030 0.472 0.590 0.541 0.511 0.469 0.502 0.555 0.852
AW 4.014 0.716 0.622 0.521 0.472 0.449 0.473 0.462 0.677 0.649 0.744
Cronbach’s Alpha 0.859 0.812 0.786 0.764 0.781 0.842 0.846 0.887 0.787
Composite Reliability 0.862 0.812 0.848 0.764 0.781 0.842 0.845 0.888 0.788
Average Variance Extracted 0.611 0.684 0.651 0.619 0.573 0.572 0.579 0.725 0.553
Notes. CC: content credibility; IG: situational normality – general; IB: situational normality – benevolence; II: situational normality – integrity; IC:
situational normality – competence; ISA: structural assurance; SC: source credibility; SW: sharing willingness; AW: adoption willingness.
Table 3
Item loadings.
Construct Item loadings Construct Item Loadings
Content Credibility CC1 0.775 Situational Normality - General IG1 0.830

CC2 0.697 IG2 0.824
CC3 0.826 Situational Normality - Benevolence IB1 0.705
CC4 0.822 IB2 0.780
Source Credibility SC1 0.681 IB3 0.750
SC2 0.724 Situational Normality - Integrity II1 0.788
SC3 0.815 II1 0.784
SC4 0.812 Situational Normality - Competence IC1 0.732
Sharing Willingness SW1 0.830 IC2 0.743
SW2 0.846 IC3 0.736
SW3 0.878 Structural Assurance ISA1 0.754
Adoption Willingness AW1 0.740 ISA2 0.745
AW2 0.759 ISA3 0.771
AW3 0.732 ISA4 0.755
adapted from Metzger (2007) and Wu and Shaffer (1987), respectively. Items for the institution-based trust were adapted from
McKnight et al. (2002). Specifically, we modeled institution-based trust as a second-order construct reflectively containing five
first-order dimensions. We used the reflective treatment of the dimensions not only because the trust theory suggests that these di
mensions vary together and are reflective of the overall trust construct (Lankton et al., 2015), but also because we did not seek to
explain variance in these dimensions. All items were measured on 5-point Likert and semantic differential scales.
4.3. Data collection
We first conducted a pilot study among 30 undergraduate students at a large university to refine the survey. The finalized survey
was posted on a popular survey website in China. We distributed the invitation to members belonging to a nationwide panel of users of
an MTurk-like online crowdsourcing market (zbj.com). During a four-week time window, the website server logs recorded 2,181
responses to our survey. We excluded 95 invalid respondents who failed the attention check (2+2=?), resulting in a sample of 2,086
valid responses. Table 1 shows the respondents’ profile in detail.
Our sample was well balanced across genders. Most respondents were relatively young and well-educated, which was consistent
with the general characteristics of WeChat users (WALKTHECAT, 2017). In terms of user experiences, most respondents were expe
rienced (with six-month tenure or longer) and frequent users of WeChat (who used WeChat every day). A wave analysis comparing the
first and last quartile of respondents suggests no significant differences between these two groups of respondents in demographics and
usage experiences with WeChat, suggesting no serious concern about non-response bias (Armstrong & Overton, 1977).
5. Data analysis and results
We used a covariance-based structure equation modeling to analyze the data. In particular, we adopted the analytical software of
AMOS 24 software.
7
Table 4
Assessing the hierarchical model of institution-based trust.
First-order constructs Institution-based trust
Situational Normality - General 0.790***

Situational Normality - Benevolence 0.940***
Situational Normality - Integrity 0.954***
Situational Normality - Competence 0.899***
Structural Assurance 0.904***
Composite Reliability (CR) 0.955
Average Variance Extracted (AVE) 0.809
Notes. This table reports the loadings of the first-order latent variables on the
second-order factor. two-tailed test. * p < 0.05, ** p < 0.01, *** p < 0.001.
∑ ∑ ∑ ∑ ∑ ∑
CR = ( λ)2 /(( λ)2 + ( θ)) AVE = ( λ2 )/(( λ2 ) + ( θ))
Table 5
Structural model assessment (Unstandardized path Coefficients, standardized path coefficients).
Sharing Willingness (SW) Adoption Willingness (AW)
Unstandardized path Standardized path Unstandardized path Standardized path

coefficients coefficients coefficients coefficients
Age 0.004 (1.738) 0.032 (1.738) -0.003 (-1.933) -0.034 (-1.933)

Education -0.035 (-1.114) -0.020 (-1.114) -0.022 (-1.068) -0.018 (-1.068)
Gender 0.055 (0.745) 0.036 (0.745) 0.048 (1.436) 0.044 (1.436)
Tenure 0.019 (1.036) 0.019 (1.036) 0.012 (1.034) 0.018 (1.034)
No. of Words -0.003 (-0.195) -0.003 (-0.195) -0.036 (-3.595) *** -0.060 (-3.595) ***
Content Credibility -0.013 (-0.278) -0.010 (-0.278) 0.201 (6.711) *** 0.227 (6.711) ***
Source Credibility 0.557 (9.113) *** 0.342 (9.113) *** 0.583 (14.040) *** 0.541 (14.040) ***
Institution-based 0.657 (15.911) *** 0.463 (15.911) *** 0.170 (6.983) *** 0.178 (6.983) ***
Trust
R2 (%) 52.2 73.6
Two-tailed t tests: * p < 0.05, ** p < 0.01, *** p < 0.001.
5.1. Measurement model
Harman’s single-factor test and a latent common method factor were used to test for common method bias. First, the result of
Harman’s single-factor test indicated that six factors with eigenvalues greater than one were produced with no single factor accounting
for more than 40% of the variance. Second, after including a latent common method factor in the measurement model, the factor
loadings remained stable across the original measurement model and the measurement model with a common method variance factor.
Thus, the results suggest that the common method bias is unlikely to be a serious concern for this study (Podsakoff et al., 2003).
As our model contained a hierarchical (second-order) construct, namely, institution-based trust, we followed Wetzels et al., and van
Oppen (2009) to assess hierarchical construct models. First, we constructed all first-order latent variables in the research model as
reflectively related to their respective items. As shown in Table 2, composite reliability (CR), Cronbach’s alpha and average variance
extracted (AVE) values exceeded their thresholds of 0.70, 0.70, and 0.50, respectively (Fornell & Larcker, 1981), suggesting adequate
reliability for all first-order constructs. As shown in Table 3, all item loadings were above 0.70, providing evidence of convergent
validity (Fornell & Larcker, 1981). Moreover, the AVE’s square root exceeded the inter-construct correlations (Table 2), indicating
adequate discriminant validity (Fornell & Larcker, 1981).
Second, having established the first-order constructs’ reliability and validity, we then constructed the second-order latent variable
(institution-based trust) by relating it to the blocks of the underlying first-order latent variables in a reflective fashion. As shown in
Table 4, the CR and AVE values were greater than 0.90 and 0.70, providing evidence of reliability. The loadings of all first-order latent
variables on the second-order factor exceeded 0.75, suggesting adequate convergent validity.
Hence, the second-order specification for the construct of institution-based trust was confirmed to be appropriate for this study. For
the sake of parsimony, the results reported are based on this specification and are robust across the alternative first-order specification.
5.2. Structural model
Before testing our hypotheses, we first assessed our basic structural model and obtained adequate fit indices (GFI = 0.949, AGFI =
0.938, NFI = 0.956, TFI = 0.961, CFI = 0.966, and RMSEA= 0.041). As shown in Table 5, the three trust beliefs related positively and
significantly to sharing and adoption willingness, except for the path from content credibility to sharing willingness. In addition,
institution-based trust (b = 0.189, p < 0.001) and source credibility (b = 0.642, p < 0.001) have positive relationships with content
credibility.
8
Table 6
Results of hypotheses testing for H1A-B and H2A-B.
Hypothesis Path coefficient @ Results $ Conclusion
Sharing willingness (SW)

H1A βCC→ SW vs. βIBT→ SW = -0.013 vs.0.657*** P<0.001*** (√) βCC→ SW < βIBT→ SW
βCC→ SW vs. βSC→ SW = -0.013 vs. 0.557*** P<0.001*** (√) βCC→ SW < βSC→ SW
H1B βSC→ SW vs. βIBT→ SW = 0.557*** vs. 0.657*** P<0.05* (√) βSC→ SW< βIBT→ SW
Adoption willingness (AW)
H2A βCC→ AW vs. βIBT→ AW =0.201*** vs. 0.170 *** n.s. (x) βCC→ AW > βIBT→ AW
βCC→ AW vs. βSC→ AW =0.201*** vs. 0.583*** P<0.001*** (x) βCC→ AW > βSC→ AW
H2B βSC→ AW vs. βIBT→ AW =0.583*** vs. 0.170 *** P<0.001*** (√) βSC→ AW > βIBT→ AW
Notes. βX→Y indicates the path coefficient from X to Y.

@ Standardized path coefficients are reported here. However, unstandardized path coefficients are required inputs for calculating t values to test
differential effects.
$ t values (* p < 0.05, ** p < 0.01, *** p < 0.001, n.s.: nonsignificant) were calculated based on the following formula (Li et al., 2013).
βi − βj
t = √̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
( )
√ 2
√ 1 − RY 2
sdy ii sdy 2 jj sdy 2 rij
√ ∗ ∗ r + ∗ r − 2 ∗ √ ̅̅̅̅̅̅̅̅̅̅̅̅
̅
n− k− 1 sdi 2 sdj 2 sdi ∗sdj rii ∗ rjj
where βi is the unstandardized path coefficient of the independent variable I (see the unstandardized path coefficients in Appendix B); sdi is the
standard deviation of I, rij are the elements of the inverted correlation metrics, R2Y is the explained variance of the dependent variable Y, n is a sample
size, and k is the number of total independent variables.
Table 7
Results of hypotheses testing for H3A-B.
Hypothesis Path coefficient @ Results $ Conclusion
H3A βCC→SW vs. βCC→AW = -0.013 vs. 0.201*** P<0.001*** (√) ВCC→SW < βCC→AW
H3B βIBT→SW vs. βIBT→AW =0.657*** vs. 0.170 *** P<0.001*** (√) ВIBT→SW > βIBT→AW
– βSC→SW vs. βSC→AW =0.557*** vs. 0.583*** n.s. No significant difference
Notes: βX→Y indicates the path coefficient from X to Y.

@ Standardized path coefficients are reported here. However, unstandardized path coefficients are required inputs for calculating t values to test
differential effects.
$ t values (* p < 0.05, ** p < 0.01, *** p < 0.001, n.s.: non-significant) were calculated following these procedures (Li et al., 2013): We first obtained
unstandardized path coefficients (Table 5), and then generated the estimated value of one dependent variable, say ÂW. We subtracted ÂW from
another dependent variable, say, SW (i.e., SW - ÂW), and then regressed the new variable, namely, SW - ÂW, on the original set of independent
variables. The resulting path coefficient of an independent variable and its significance level, respectively, indicate the magnitude and significance of
the difference in the independent variable’s impacts on SW vs. AW.
5.3. Hypotheses tests: path comparisons
To test the proposed hypotheses, we adopted the path comparison method proposed by Cohen et al., and Aiken (2003) and applied
widely in information systems research (e.g., Li et al., 2013). The results are summarized in Tables 6 and 7.
As for the relative effects of the three trusting beliefs on sharing willingness, path comparisons between each pair suggest sig
nificant differences. Specifically, source credibility (p < 0.001) and institution-based trust (p < 0.001) had a strong relationship than
content credibility, supporting H1A. Further, the relationship of institution-based trust was stronger than that of source credibility (p <
0.05), supporting H1B.
Source credibility showed a significantly stronger relationship (p < 0.001) than content credibility with adoption willingness
(which is opposite to our expectation). Content credibility had a stronger relationship than institution-based trust with adoption
willingness, but the difference in path coefficients was found to be insignificant. Thus, H2A was not supported. Further, source
credibility showed a significantly stronger relationship (p < 0.001) than institution-based with adoption willingness, supporting H2B.
Content credibility had a stronger relationship with adoption willingness than with sharing willingness (p < 0.001), supporting
H3A. Also, the institution-based trust had a stronger relationship with sharing willingness than with adoption willingness (p < 0.001),
supporting H3B. Further, we did not find a significant difference in the relationship of source credibility with adoption willingness and
that with sharing willingness.
5.4. Post-hoc analysis results
To investigate the interaction effects of source credibility, institution-based trust, and content credibility on users’ sharing and
adoption willingness, we adopted a stepwise procedure (See the detailed stepwise test in Appendix B). The results suggested that all the
two-way interaction effects (e.g., CC*SC, SC*IBT) are significant on knowledge adoption willingness except for CC*IBT. However, only
CC*SC showed a significant effect on knowledge sharing willingness whereas the other two-way interaction effects (e.g., SC*IBT,
9
Table 8
The moderation of user expertise.
Path tspooled Path Coefficients
High UE (N=1092) Low UE (N=994)
SC →SW 1.615 0.742*** 0.348***

CC → SW -1.538 -0.244** -0.025
IBT →SW -1.650 0.27*** 0.443
SC →AW 2.943*** 1.024*** 0.508***
CC → AW -3.835*** -0.117 0.23***
IBT →AW -2.260** 0.047 0.189***
Notes: * p<0.05, ** p<0.01, *** p<0.001.
CC*IBT) were insignificant. The three-way interaction effect was significantly negative on knowledge sharing and adoption willing
ness. Precisely, when users perceived low source credibility, the relationship between high institution-based trust and knowledge
adoption willingness was weaker than high source credibility. Given high institution-based trust, the relationship between content
credibility and knowledge sharing willingness was stronger when users perceived low than high source credibility.
Furthermore, consideration of the moderating role of personal characteristics can help reconcile inconsistent findings and provide a
more comprehensive explanation for behavioral willingness (Venkatesh et al., 2003). User expertise (i.e., whether the user has relevant
professional health knowledge needed to make informed health decisions) is one of the most important personal characteristic var
iables for knowledge communication. A recipient with more professional knowledge is able to accept and process the information
deeply. Thus, user expertise may moderate the relationship of content credibility, source credibility, and institution-based trust with
sharing willingness and adoption willingness. The analysis of these regulatory effects helps deepen our understanding of the mech
anisms involved.
We conducted two between-group analyses, which permitted the comparison of the structural model across groups. First, according
to the median of user professionalism 3.5, the total sample was divided into high-UE (UE ≥ 3.5) and low-UE (UE < 3.5) subsamples.
Then, the validity and equivalence of the measurement model in the two-component samples are confirmed. Finally, the path co
efficients obtained from the analysis of the two subsamples are compared Sia et al., 2009), as shown in formulas ((2) and (3).
√̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
{[ ] } {[ ] }
(N1 − 1)2 (N2 − 1)2
Sspooled = × SE1 2 + × SE2 2 (2)
N1 + N2 − 2 N1 + N2 − 2
(PC1 − PC2 )
tspooled = √̅̅̅̅̅̅̅̅̅̅̅̅̅̅ (3)
1
Spooled × N1
+ N12
Where Spooled is the pooled estimator for the variance, tspooled refers to the t-statistic with (N1+N2− 2) degrees of freedom; Ni is the
sample size of data set for group i (i= 1, 2); SEi is the standard error of path in the structural model of group i (i =1, 2); PCi is the path
coefficient in the structural model of group i (i=1, 2).
As shown in Table 8, we found that user expertise positively moderated the relationship between source credibility and adoption
willingness. User expertise negatively moderated the relationship of content credibility and institution-based trust with adoption
willingness. User expertise did not significantly moderate the relationships of the three trust beliefs with knowledge sharing
willingness.
6. Discussion
6.1. Key findings
This study investigated the relative importance and differential relationships of three trusting beliefs (content credibility, source
credibility, and institution-based trust) in triggering health knowledge sharing and adoption willingness in the context of health
knowledge communication enabled by social media (WeChat in particular) in China. The results of our study reveal a set of interesting
findings. First, sharing willingness is primarily related to peripheral cues (more by institution-based trust and less by source) rather
than the central cue (content credibility); adoption willingness is primarily related to content-specific trusting beliefs (content cred
ibility and source credibility), rather than the context-related trusting belief (institution-based trust). Second, as expected, content-
specific cue (content credibility) exerts a stronger relationship with adoption willingness than with sharing willingness. In compari
son, context-related cue (institution-based trust) has a stronger relationship with sharing willingness than with adoption willingness.
Finally, one unexpected finding is that adoption willingness is also primarily related to peripheral cues (source credibility and
institution-based trust) rather than the central cue (content credibility). Specifically, source credibility shows the strongest relation
ship with adoption willingness, and the relationship of content credibility is not significantly weaker than that of institution-based
10
trust. One plausible explanation lies in the uniqueness of our research context, i.e., health knowledge communication enabled by social
media in China, wherein most recipients lack health expertise. The Chinese residents with a relatively low level of health literacy
reached 19.2 percent (24.8 percent for urban residents and 15.7 percent for rural residents). Furthermore, prior research suggests that
when subjects’ expertise is low, their cognitive effort tends to be low as well (Sussman & Siegal, 2003). Thus, in this research, these
recipients who lack health expertise still rely heavily on peripheral cues to decide whether to follow the advice contained in the
content. Consistent with this argument, prior research has found that source credibility can improve recipients’ likelihood to act on the
advice offered in online health knowledge (Wang et al., 2008).
6.2. Theoretical implications
As one of the first studies investigating the differential factors in health knowledge communication (e.g., adoption and sharing)
enabled by social media (WeChat in particular), this study provides several implications for theory. First, previous research about
health knowledge communication has mainly discussed how social media are used by doctors or patients for enhancing doctor-to-
patient or patient-to-patient communications (Guo et al., 2017; Liu et al., 2020; Zhang et al., 2020; Zhang et al., 2017), especially
surrounding professional healthcare services (Bansal et al., 2010; Mano et al., 2014) or a certain disease (e.g., cancer) (Wang et al.,
2019). This study explores the differential relationships of the credibility of content, source, and platform with users’ willingness of
general health knowledge communication (adoption and sharing), thus extending the literature on online communication of scientific
knowledge.
Second, this study extends the trust literature by showing how trust targets (e.g., content, source, or platform) matter in triggering
knowledge behavior willingness (e.g., adoption and sharing). Prior research on trust has differentiated trusting beliefs in terms of their
targets (i.e., which elements in the decision-making process recipients’ trusting beliefs are directed toward) (Borah, 2014), but rarely
examined their differential relationships with other factors. This study extends this body of research by revealing a clear-cut pattern of
how target-oriented trusting beliefs differ in leading to adoption versus sharing willingness in the context of health knowledge. That is,
trusting beliefs specific to a particular piece of content and its source (content credibility and source credibility) are more important in
driving knowledge adoption, which primarily regards human-content interactions. As a comparison, the generic trusting belief toward
the platform (institution-based trust) is more important in stimulating users’ willingness to share health knowledge. This behavior
involves social or interpersonal interactions between potential sharers and recipients.
Finally, it is critical to understand the specific mechanisms through which companies can engage social media users in knowledge
communication (sharing) and consumption (adoption). While prior research on knowledge behaviors has examined information
(knowledge) adoption (e.g., Tseng & Wang, 2016) and sharing (e.g., Wang et al., 2017) separately, this study contributes to the
knowledge communication literature by integrating these two behavioral willingnesses and exploring how they are related differently
to the same set of trusting beliefs (i.e., content credibility, source credibility, and institution-based trust). By doing so, this research
contributes to the literature on social media enabled knowledge communication by providing rich insights into ways to engage users in
adoption and sharing.
6.3. Practical implications
In China, the public has increasingly relied on WOAs as credible and useful sources to acquire health knowledge. However, a small
portion of health knowledge is adopted and shared by users. Our research findings also provide rich insights for practitioners
(especially healthcare organizations operating WOAs in China) to attract public attention and effectively propagate health knowledge.
First, our results suggest that although source credibility and institution-based trust are common antecedents of health knowledge
sharing and adoption, their relative importance differs depending on the focal behavior. One implication of this finding is that different
strategies should be applied to stimulate recipients’ willingness of knowledge sharing versus knowledge adoption (i.e., following the
given advice). To stimulate users to participate in viral transmissions of health knowledge (i.e., knowledge sharing), WOA operators
should fulfill their duty of providing necessary institutional mechanisms to gain institution-based trust (e.g., by acting in the recipients’
best interests, trying to help, providing a standardized audit mechanism, and being genuinely concerned), which is the strongest
enabler of users’ sharing willingness.
In contrast, for encouraging recipients to adopt knowledge, WOA operators should pay more attention to identifying health pro
fessionals that are perceived to be credible and using design approaches to improve laypeople’s perceptions of source credibility
(which is the strongest predictor of adoption willingness). Possible techniques include (but are not limited to) leveraging recipients’
feedback such as likes and comments to evaluate the credibility of knowledge sources (health professionals), and providing presen
tational cues (e.g., professional title, affiliations, head portraits, etc.) to endorse source authority and credibility. Using these tech
niques, WOAs can effectively disseminate health knowledge from authorities to laypeople (Wang et al., 2017), improve the public’s
health literacy, and trigger changes in health behaviors among recipients. In addition to these social values, successfully stimulating
changes in behaviors among recipients would help health WOAs gain a reputation as authoritative in providing practical suggestions
for health management.
11
Second, many WOA operators in China who seek to motivate individuals to disseminate healthcare knowledge put massive efforts
into phrasing or framing healthcare knowledge in a way that draws greater believability, reflects more objectivity, increases perceived
reliability, and triggers trust. However, our results suggest that this commonly used strategy of content credibility may not be as
effective as previously thought. Beyond the important role of content credibility, social media users’ trusting beliefs about content
sources and platforms exert vital effects on their content-engagement behaviors. Specifically, in the context of health knowledge
communication enabled by social media in China, content credibility can promote both adoption and sharing willingness. However, as
most laypersons with relatively low health literacy cannot judge whether the health knowledge per se is credible and authoritative, the
other trusting beliefs (i.e., source credibility and institution-based trust) can be equally (even more) important. Thus, to effectively
disseminate health knowledge in China, WOA operators should not only put efforts into content credibility but also take actions to
enhance recipients’ trusting beliefs toward the source and platform.
6.4. Limitations and directions for future research
Like any other research, this study is not free of limitations. First, we used an online survey to collect data. Although this is common
in prior research (e.g., Milkman & Berger, 2014) and appropriate for this study, the cross-sectional nature of our data makes it difficult
to establish causal relationships. While we leverage theoretical arguments to make causal inferences, we encourage future research to
adopt longitudinal or experimental designs to establish causality relationships. Second, we take a trusted perspective to investigate the
antecedent differences between knowledge adoption and sharing. We do so intentionally to contribute to the trust literature but
encourage future research to examine other antecedents of, and their differential effects on, knowledge adoption and sharing. Finally,
the uniqueness of the research context (i.e., health knowledge communication in WeChat) may limit the generalizability of our
research findings. We deliberately selected this research context to make novel contributions to both theory and practice. Yet, some of
our research findings (e.g., the triviality of content credibility in predicting knowledge adoption) may be specific to this unique
context, which may not be generalizable to all types of communications. To improve the generalizability of research findings, we urge
future researchers to cross-validate these findings in different contexts (e.g., other types of knowledge) and countries.
There are also interesting directions for future research to extend this study. As mentioned earlier, the comparison of knowledge
adoption and sharing in concepts and antecedents opens a new avenue for research on understanding various user engagement be
haviors in social media. The current study is just a first step in this direction. Future research can go further by integrating other
engagement behaviors (e.g., liking or commenting on a post, following or subscribing to a WOA, and others) and comparing them with
the outcome behaviors in this study. Besides, future research may extend this study by incorporating potential moderators (e.g.,
demographics and user experiences) into the model. Finally, future research can explore whether and how different types of health
knowledge would need to be processed differently by social media users, so that certain topics may be more likely to be shared (or
adopted) than others.
7. Conclusion
In recent years, the development of social media has provided a possible solution to the mismatch between supply and demand for
professional health knowledge in China. Social media plays a vital role in health knowledge communication to improve public health
literacy in China (and many other countries). Recipients’ adoption and sharing willingness are the keys to the effectiveness of health
knowledge communication enabled by social media, and likely result from recipients’ trusting beliefs about the content, source, and
social media platform. Based on the trust and communication literature, we theorize the conceptual differences between knowledge
adoption and sharing willingness in the health knowledge communication enabled by social media, and investigate how they are
influenced differently by three trusting beliefs (content credibility, source credibility, and institution-based trust). The results of our
study conducted in China suggest a clear pattern of differential effects of trusting beliefs on behavioral willingness. This study enriches
the literature of trust and knowledge communication and provides important insights for practitioners (especially healthcare orga
nizations operating WOAs in China).
Funding
This work was supported by the National Natural Science Foundation of China [grant numbers 71871162, 71872112], Shanghai
Philosophy and Social Science Planning Program [grant number 2017BGL022], Humanities and Social Science Fund of Ministry of
Education of China [grant number 17YJC630237], and Program for Professor of Special Appointment (Eastern Scholar) at Shanghai
Institutions of Higher Learning [grant number TP2018016].
Appendix A
Table A.
12
Table A
Measurement items.
Content Credibility (CC): Adapted from (Metzger, 2007)
CC1. Not trustful/Trustful
CC2. Not objective/Objective
CC3. Not reliable/Reliable
CC4. Not believable/Believable
Institution-based Trust (IBT): Adapted from (McKnight et al., 2002)
Situational Normality-General (IG)
IG1. I feel good about how things go when I sharing content or other activities on WeChat.
IG2. I am comfortable sharing content on WeChat.
Situational Normality-Benevolence (IB)
IB1. I feel that most WeChat official accounts would act in a users’ best interest.
IB2. If a user required help, most WeChat official accounts would do their best to help.
IB3. Most WeChat official accounts are interested in user well-being, not just their own-being.
Situational Normality-Integrity (II)
II1. I am comfortable relying on WeChat official accounts to meet their obligations.
II2. I always feel confident that I can rely on WeChat official accounts to do their part when I interact with them.
Situational Normality-Competence (IC)
IC1. In general, most WeChat official accounts are competent at serving their customers.
IC2. Most WeChat official accounts do a capable job of meeting customer needs.
IC3. I feel that most WeChat officials are good at what they do.
Structural Assurance (ISA)
ISA1. WeChat has enough safeguards to make me feel comfortable using it to transmit information.
ISA2. I feel assured that legal and technological structures adequately protect me from problems on WeChat.
ISA3. I feel confident that encryption and other technological advances on WeChat make it safe for me to transmit information there.
ISA4. In general, WeChat is now a robust and safe environment in which to transmit information.
Source Credibility (SC): Adapted from (Wu & Shaffer, 1987)
SC1. How knowledgeable is the author who wrote the content on the topic of the content? (Not knowledgeable / Knowledgeable)
SC2. To what extent is the author who wrote the content an expert on the content topic? (Not expert / Expert)
ST1. How trustworthy is the author who wrote the content on the topic of the content? (Not trustworthy / Trustworthy)
ST2. How reliable is the author who wrote the content on the topic of the content? (Not reliable / Reliable)
User Expertise (UE): Adapted from (Stamm & Dube 1994)
EX1.How informed are you on the subject matter of this issue? Novice/expert
EX2.To what extent are you an expert on the topic of this content? Not at all/To a great extent
Sharing Willingness (SW): Adapted from (Milkman & Berger, 2014a)
How likely will you share this content …? (Very unlikely / Very likely)
SW1. in your WeChat friends circle
SW2. with someone of your WeChat friends directly
SW3. in a WeChat Group
Adoption Willingness (AW): Adapted from (Sussman & Siegal, 2003)
AW1. To what extent do you agree with the suggestions in the content? (Completely disagree / Completely agree)
AW2. To what extent will you follow its suggestions in the content? (Not closely / Very closely)
AW3. To what extent does this content motivate you to take action? (Not motivated / Highly motivated)
Appendix B
We adopted a stepwise procedure: Model F1 included the control variables (Age, Gender, Education, Long, Word), Model F2
included the control variables and main factors (CC, SC, IBT), Model F3 incorporated the factors in Model F2 and the interaction terms
(CC*SC, CC*IBT, SC*IBT), Model F4 incorporated the factors in Model F3 and the interaction terms (CC*SC*IBT).
Figure B1, B2, B3, B4 and B5 show the interaction plots and Table B1 summarizes the results.
Fig. B1. The two-way interaction effect between content credibility (CC) and source credibility (SC) on sharing willingness.
13
Fig. B2. The three-way interaction effect among content credibility (CC), institution-based trust (IBT), and source credibility (SC) on sharing
willingness.
Fig. B3. The two-way interaction effect between content credibility (CC) and source credibility (SC) on adoption willingness.
Fig. B4. The two-way interaction effect between institution-based trust (IBT) and source credibility (SC) on adoption willingness.
Fig. B5. The three-way interaction effect among content credibility (CC), institution-based trust (IBT), and source credibility (SC) on adoption
willingness.
Table B1
Regression analysis results.
Sharing Willingness Adoption Willingness
Model F1 Model F2 Model F3 Model F4 Model F1 Model F2 Model F3 Model F4
Control variables
Age 0.021*** 0.003 0.004* 0.004* 0.017*** -0.004* -0.002 -0.002
Gender -0.004 0.022 0.020 0.025 0.018 0.046 0.039 0.042
Education 0.000 -0.033 -0.031 -0.031 -0.013 -0.038 -0.033 -0.033
Tenure 0.117*** 0.016 0.016 0.016 0.120*** 0.014 0.015 0.014
No. of Word -0.006 -0.006 -0.004 -0.006 -0.049* -0.055*** -0.052*** -0.052***
Main effects
CC 0.058** 0.055* 0.067** 0.255*** 0.246*** 0.252***
SC 0.276*** 0.239*** 0.247*** 0.401*** 0.341*** 0.344***
IBT 0.421*** 0.432*** 0.452*** 0.200*** 0.223*** 0.231***
Two-way interactions
CC*SC -0.057*** -0.071 -.106*** -0.113***
CC*IBT 0.008 -0.001 .001 -0.003
SC*IBT 0.009 -0.017 .039* 0.028
Three-way interaction
CC*SC*IBT -0.029*** -0.013*
R2 0.034 0.442 0.447 0.451 0.028 0.540 0.556 0.557
Note: CC: content credibility; SC: source credibility; IBT: institution-based trust.*p < .05, **p < .01, ***p < .001.
14
References
Armstrong, J. S., & Overton, T. S. (1977). Estimating nonresponse bias in mail surveys. Journal of Marketing Research, 14, 396–402.
Bansal, G., Zahedi, F., & Gefen, D. (2010). The impact of personal dispositions on information sensitivity, privacy concern and trust in disclosing health information
online. Decision Support Systems, 49(2), 138–150.
Barasch, A., & Berger, J. (2014). Broadcasting and narrowcasting: How audience size affects what people share. Journal of Marketing Research, 51(3), 286–299.
Berger, J., & Iyengar, R. (2013). Communication channels and word of mouth: How the medium shapes the message. Journal of Consumer Research, 40(3), 567–579.
Berger, J., & Milkman, K. L. (2012). What makes online content viral? Journal of Marketing Research, 49(2), 192–205.
Blau, P. M. (1964). Exchange and power in social life. Piscataway, NJ: Transaction Publishers.
Borah, P. (2014). The hyperlinked world: A look at how the interactions of news frames and hyperlinks influence news credibility and willingness to seek information.
Journal Of Computer-Mediated Communication, 19(3), 576–590.
Cappella, J. N. (2017). Vectors into the future of mass and interpersonal communication research: Big data, social media, and computational social science. Human
Communication Research, 43(3), 545–558.
Chai, S., Das, S., & Rao, H. R. (2011). Factors affecting bloggers", knowledge sharing: An investigation across gender. Journal of Management Information Systems,
309–341, 28:3.
Chang, Y. Q. (2020). The hot spot and development trend of population aging. China Population Daily. March 20, p. 3.10.28125/n.cnki.ncrcb.2020.000412.
Chen, X., Huang, Q., Davison, R. M., & Hua, Z. (2016). What drives trust transfer? The moderating roles of seller-specific and general institutional mechanisms.
International Journal of Electronic Commerce, 20(2), 261–289.
Cheung, C. M. K., Lee, M. K. O., & Rabjohn, N. (2008). The impact of electronic word-of-mouth. Internet Research, 18(3), 229–247.
Cohen, J., Cohen, P., West, S. G., & Aiken, L. S. (2003). Applied multiple regression / correlation analysis for the behavioral sciences. London: Lawrence Erlbaum Associates.
Crook, B., Stephens, K. K., Pastorek, A. E., Mackert, M., & Donovan, E. E. (2016). Sharing health information and influencing behavioral intentions: The role of health
literacy, information overload, and the internet in the diffusion of healthy heart information. Health Communication, 31(1), 60–71.
Deloitte (2017).,"China medical service industry",pp. 2–3.
Ellonen, R., Blomqvist, K., & Puumalainen, K. (2008). The role of trust in organisational innovativeness. European Journal of Innovation Management, 11(2), 160–181.
Fornell, C., & Larcker, D. F. (1981). Evaluating structural equation models with unobservable variables and measurement error. Journal of Marketing Research, 18(1),
39–50.
Freeman, K. S., & Spyridakis, J. H. (2004). An examination of factors that affect the credibility of online health information. Technical Communication, 51(2), 239–263.
Gong, H. (2017). Research on health communication in the age of social media. Chinese Journal of Social Sciences, 3. November 2.
Guo, S., Guo, X., & Fang, Y. (2017). How doctors gain social and economic returns in online health-care communities: A professional capital perspective. Journal of
Management Information Systems, 34(2), 487–519.
Heo, J., Chun, M., Lee, H. W., & Woo, J. H. (2018). Social media use for cancer education at a community-based cancer center in South Korea. Journal of Cancer
Education, 33(4), 769–773.
Hovland, C. I. (1948). Social communication. Proceedings of the American Philosophical Society, 92, 371–375.
Hu, S., Gu, J., Liu, H., & Huang, Q. (2017). The moderating role of social media usage in the relationship among multicultural experiences, cultural intelligence, and
individual creativity. Information Technology & People, 30(2), 265–280.
Hussain, S., Ahmed, W., Jafar, R. M. S., Rabnawaz, A., & Jianzhou, Y. (2017). eWOM source credibility, perceived risk and food product customer’s information
adoption. Computers in Human Behavior, 66, 96–102.
Kim, C., & Yang, S.-U. (2017). Like, comment, and share on Facebook: How each behavior differs from the other. Public Relations Review, 43(2), 441–449.
Lankton, N. K., McKnight, D. H., & Tripp, J. (2015). Technology, humanness, and trust: Rethinking trust in technology. Journal of the Association for Information
Systems, 16(10), 880–918.
Lankton, N. K., McKnight, D. H., & Tripp, J. (2015). Technology, humanness, and trust: Rethinking trust in technology. Journal of the Association for Information
Systems, 16(10), 880–918.
Li, X., Hsieh, J. J. P.-A., & Rai, A. (2013). Motivational differences across post-acceptance information system usage behaviors: An investigation in the business
intelligence systems context. Information Systems Research, 24(3), 659–682.
Li, Y., & Wang, X. (2018). Seeking health information on social media: A perspective of trust, self-determination, and social support. Journal of Organizational & End
User Computing, 30(1), 1–22.
Lin, W. Y., Zhang, X., & Song, H. (2016). Health information seeking in the Web 2.0 age: Trust in social media, uncertainty reduction, and self-disclosure. Computers in
Human Behavior, 61, 690-690.
Liu, Y., Ren, C., & Shi, D. (2020). Evaluating the social value of online health information for third-party patients: Is uncertainty always bad? Information Processing &
Management, 57(5).
Mano, R. S. (2014). Social media and online health services: A health empowerment perspective to online health information. Computers in Human Behavior, 39,
404–412.
McKnight, D. H., Choudhury, V., & Kacmar, C. (2002). Developing and validating trust measures for e-commerce: An integrative typology. Information Systems
Research, 13(3), 334–359.
Metzger, M. J. (2007). Making sense of credibility on the web: Models for evaluating online information and recommendations for future research. Journal of the
American Society for Information Science and Technology, 58(13), 2078–2091.
Milkman, K. L., & Berger, J. (2014). The science of sharing and the sharing of science. Proceedings of the National Academy of Sciences of the United States of America, 111
(14), 13642–13649.
Muntinga, D. G., Moorman, M., & Smit, E. G. (2015). Introducing COBRAs. International Journal of Advertising, 30(1), 13–46.
Oostervink, N., Agterberg, M., & Huysman, M. (2016). Knowledge sharing on enterprise social media: Practices to cope with institutional complexity. Journal Of
Computer-Mediated Communication, 21(2), 156–176.
Park, H., Rodgers, S., & Stemmle, J. (2013). Analyzing health organizations’ use of Twitter for promoting health literacy. Journal of Health Communication, 18(4),
410–425.
Petty, R. E., & Cacioppo, J. T. (1986). Communication and persuasion: Central and peripheral routes to attitude change. New York: Springer-Verlag.
Podsakoff, P. M., MacKenzie, S. B., Lee, J. Y., & Podsakoff, N. P. (2003). Common method biases in behavioral research: A critical review of the literature and
recommended remedies. Journal of Applied Psychology, 88(5), 879–903.
Razak, N. A., Pangil, F., Zin, M. L. M., Yunus, N. A. M., & Asnawi, N. H. (2016). Theories of knowledge sharing behavior in business strategy. Procedia Economics and
Finance, 37, 545–553.
Rieh, S. Y., & Belkin, N. J. (1998). Understanding judgment of information quality and cognitive authority in the WWW. In , 35. Proceedings of ASIS Annual Meeting
(pp. 279–289).
Salehan, M., Kim, D. J., & Koo, C. (2016). A study of the effect of social trust, trust in social networking services, and sharing attitude, on two dimensions of personal
information sharing behavior. The Journal of Supercomputing, 74(8), 3596–3619.
Shang, L., Zhou, J., & Zuo, M. (2020). Understanding older adults’ intention to share health information on social media: The role of health belief and information
processing. Internet Research.
Shannon, C. E. (1948). A mathematical theory of communication. Bell System Technical Journal, 27(3), 379–423.
Shapiro, S. P. (1987). The social control of impersonal trust. American Journal of Sociology, 93(3), 623–658.
Shen, X. L., Zhang, K., & Zhao, S. J. (2014). Understanding information adoption in online review communities: The role of herd factors. In Hawaii international
conference on system sciences. IEEE Computer Society.
Shi, Z., Rui, H., & Whinston, A. B. (2014). Content sharing in a social broadcasting environment: Evidence from twitter. MIS Quarterly, 38(1), 123–142.
15
Sia, C., Lim, L., K, H., & Leung, K. (2009). Web strategies to promote internet shopping: Is cultural-customization needed? MIS Quarterly, 33(3), 491–512.
So, J., Prestin, A., Lee, L., Wang, Y., Yen, J., & Chou, W. Y. (2016). What do people like to "share" about obesity? A content analysis of frequent retweets about obesity
on Twitter. Health Communication, 31(2), 193–206.
Stamm, K. R., & Dube, R. (1994). The relationship of attitudinal components to trust in media. Communication Research, 21, 105–123.
Stewart, K. J. (2003). Trust transfer on the world wide web. Organization Science, 14(1), 5–17. https://doi.org/10.1287/orsc.14.1.5.12810.
Sussman, S. W., & Siegal, W. S. (2003). Informational influence in organizations: An integrated approach to knowledge adoption. Information Systems Research, 14(1),
47–65.
Tseng, S. Y., & Wang, C. N. (2016). Perceived risk influence on dual-route information adoption processes on travel websites. Journal of Business Research, 69,
2289–2296.
Vance, A., Elie-Dit-Cosaque, C., & Straub, D. W. (2008). Examining trust in information technology artifacts: The effects of system quality and culture. Journal of
Management Information Systems, 24(4), 73–100.
Venkatesh, V., Morris, M. G., & Davis, G. B. (2003). User acceptance of information technology: Toward a unified view. MIS Quarterly, 27(3), 425–478.
WALKTHECAT. (2017)., “CAICT WeChat economic and social impact report”, Retrieved from https://technode.com/2018/05/15/wechat-impact-report-2018/.
Wang, C., Zhou, Z., Jin, X. L., Fang, Y., & Lee, M. K. O. (2017). The influence of affective cues on positive emotion in predicting instant information sharing on
microblogs: Gender as a moderator. Information Processing and Management, 53, 721–734.
Wang, L. M. (2019). A study on the prevalence and burden of chronic diseases among the elderly in China. Chinese Journal of Epidemiology, 40(3), 277–283.
Wang, L., Huang, L., Xu, H., Yang, J., & Shen, J. (2017). The application of WeChat official accounts in health education and health promotion (in Chinese). Chinese
Journal of Health Education, 7, 669–671.
Wang, X., Chen, L., Shi, J., & Peng, T.-Q. (2019). What makes cancer information viral on social media? Computers in Human Behavior, 93, 149–156.
Wang, Z., Walther, J. B., Pingree, S., & Hawkins, R. P. (2008). Health information, credibility, homophily, and influence via the internet: Web sites versus discussion
groups. Health Communication, 23(4), 358–368.
Watts, S. A., & Wei, Z. (2008). Capitalizing on content: Information adoption in two online communities. Journal of the Association for Information Systems, 9(2), 73–94.
Wetzels, M., Odekerken-Schroder, G., & van Oppen, C. (2009). Using PLS path modeling for assessing hierarchical construct models: guidelines and empirical
illustration. MIS Quarterly, 33(1), 177–195.
World Health Organization (2016)., Global diffusion of eHealth: Making universal health coverage achievable, December.
Wu, C., & Shaffer, D. R. (1987). Susceptibility to persuasive appeals as a function of source credibility and prior experience with the attitude object. Journal of
Personality and Social Psychology, 52(4), 677–688.
Yan, A. H., Zou, Y., & Mirchandani, D. A. (2020). How hospitals in mainland China responded to the outbreak of COVID-19 using information technology-enabled
services: An analysis of hospital news webpages. Journal of the American Medical Informatics Association, 27(7), 991–999.
Yi, M. Y., Yoon, J. J., & Davis, J. M. (2013). Untangling the antecedents of initial trust in Web-based health information: The roles of argument quality, source
expertise, and user perceptions of information quality and risk. Decision Support Systems, 55(1), 284–295.
Zhang, L., Peng, T. Q., Zhang, Y.-P., Wang, X. H., & Zhu, J. J. H. (2014). Content or context: Which matters more in information processing on microblogging sites.
Computers in Human Behavior, 31, 242–249.
Zhang, L., Zhao, J., & Xu, K. (2016). Who creates trends in online social media: The crowd or opinion leaders? Journal Of Computer-Mediated Communication, 21(1),
1–16.
Zhang, X. F., Guo, F., Xu, T. X., & Li, Y. L. (2020). What motivates physicians to share free health information on online health platforms? Information Processing &
Management, 57(2).
Zhang, X., Liu, S., & Deng, Z. (2017). Knowledge sharing motivations in online health communities: A comparative study of health professionals and normal users.
Computers in Human Behavior, 75, 797–810.
Zhao, H., Fu, S., & Chen, X. (2020). Promoting users’ intention to share online health articles on social media: The role of confirmation bias. Information Processing &
Management, 57(6).
Zhou, Z., Jin, X. L., Fang, Y., & Vogel, D. (2015). Toward a theory of perceived benefits, affective commitment, and continuance intention in social virtual worlds:
Cultural values (indulgence and individualism) matter. European Journal of Information Systems, 24(3), 247–261.
16

Jin Et Al., 2021

Uploaded by

Copyright:

Available Formats

Jin Et Al., 2021

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Jin Et Al., 2021

Uploaded by

Copyright:

Available Formats

Information Processing and Management 58 (2021) 102413

Contents lists available at ScienceDirect

Information Processing and Management

The differential effects of trusting beliefs on social media users’

adopt health knowledge.

2. Research background and theoretical foundation

2.1. Health knowledge communication in social media: adoption versus sharing

2.2. Trusting beliefs toward content, source, and platform

Fig. 1. Research model and hypotheses.

3. Research model and hypotheses

Gender Female 1062 50.9 Tenure < 3 months 38 1.8

3.4. Relationships between the three trusting beliefs

4.1. Research design

4.2. Measurement development

CC 3.886 0.742 0.782

Content Credibility CC1 0.775 Situational Normality - General IG1 0.830

4.3. Data collection

5. Data analysis and results

Situational Normality - General 0.790***

Unstandardized path Standardized path Unstandardized path Standardized path

Age 0.004 (1.738) 0.032 (1.738) -0.003 (-1.933) -0.034 (-1.933)

Two-tailed t tests: * p < 0.05, ** p < 0.01, *** p < 0.001.

5.1. Measurement model

5.2. Structural model

Sharing willingness (SW)

Notes. βX→Y indicates the path coefficient from X to Y.

Notes: βX→Y indicates the path coefficient from X to Y.

5.3. Hypotheses tests: path comparisons

5.4. Post-hoc analysis results

High UE (N=1092) Low UE (N=994)

SC →SW 1.615 0.742*** 0.348***

Notes: * p<0.05, ** p<0.01, *** p<0.001.

6.1. Key findings

6.2. Theoretical implications

6.3. Practical implications

6.4. Limitations and directions for future research

You might also like

Two-tailed t tests: * p < 0.05, p < 0.01, * p < 0.001.

SC →SW 1.615 0.742* 0.348*

Notes: * p<0.05, p<0.01, * p<0.001.