Electronic Commerce Research and Applications 17 (2016) 62–73

Contents lists available at ScienceDirect

Electronic Commerce Research and Applications

journal homepage: www.elsevier.com/locate/ecra

E-WOM from e-commerce websites and social media: Which will

consumers adopt?
Qiang Yan ⇑, Shuang Wu, Lingli Wang, Pengfei Wu, Hejie Chen, Guohong Wei
School of Economics and Management, Beijing University of Posts and Telecommunications, Beijing, PR China

a r t i c l e i n f o a b s t r a c t

Article history: The influence of user generated content on e-commerce websites and social media has been addressed in
Received 16 January 2015 both practical and theoretical fields. Since most previous studies focus on either electronic word of mouth
Received in revised form 26 March 2016 (eWOM) from e-commerce websites (EC-eWOM) or social media (SM-eWOM), little is known about the
Accepted 26 March 2016
adoption process when consumers are presented EC-eWOM and SM-eWOM simultaneously. We focus on
Available online 28 March 2016
this problem by considering their adoption as an interactive process. It clarifies the mechanism of con-
sumer’s adoption for those from the perspective of cognitive cost theory. A conceptual model is proposed
Keywords:
about the relationship between the adoptions of the two types of eWOM. The empirical analysis shows
Adoption
Credibility
that EC-eWOM’s usefulness and credibility positively influence the adoption of EC-eWOM, but negatively
E-commerce influence that of SM-eWOM. EC-eWOM adoption negatively impacts SM-eWOM adoption, and mediates
Electronic word of mouth the relationship between usefulness, credibility and SM-eWOM adoption. The moderating effects of
Social media consumers’ cognitive level and degree of involvement are also discussed. This paper further explains
Social commerce the adoption of the two types of eWOM based on the cognitive cost theory and enriches the theoretical
Word of mouth research about eWOM in the context of social commerce. Implications for practice, as well as suggestions
for future research, are also discussed.
Ó 2016 Elsevier B.V. All rights reserved.

1. Introduction
Over the past decade, Internet technology has considerably
altered people’s lifestyle. In terms of online shopping,
e-commerce has experienced rapid development, and online shop-
ping has become a popular method for purchasing goods. In 2013,
Amazon’s sales reached US$74.4 billion, while Taobao.com, the
biggest C2C website in China, exceeded US$179.2 billion. In terms
of sharing content online, social media use has increased in recent
years and become one of the most important platforms of commu-
nication. For example, the average number of active users per
month on Twitter and Facebook has reached 0.27 and 1.28 billion,
respectively. In China, Weixin, the most popular social media
platform, has attracted around 0.5 billion users since its launch
in 2011.
With the emergence of Web 2.0, user generated content (UGC) on
networks are increasing. As an important form of electronic word-
of-mouth (eWOM), online reviews provided by e-commerce web-
sites (referred to as EC-eWOM hereafter) have played an important
role in helping consumers make decisions (Cheung and Thadani,
⇑ Corresponding author.
E-mail address: buptyan@gmail.com (Q. Yan).
2012; Davis and Khazanchi, 2008; Duan et al., 2008). According
to a survey, 91% of participants said they use online reviews, blogs
and other forms of UGC before purchasing a new product or
service, and 46% of participants indicated that these comments
influenced their decision (Cheung and Thadani, 2012). Online
reviews can reduce the risks perceived by consumers (Cheung
et al., 2009b; Park and Kim, 2009) and improve their degree of
satisfaction (Liang et al., 2007), as well as their efficiency in making
decisions (Cheung et al., 2009a). It has been proved that the vol-
ume of online reviews is significantly related to sales. The volumes
of both positive and negative reviews impact the consumer’s
decisions (Chatterjee, 2001; Cheung et al., 2009a; Kamins and
Assael, 1987; Berger et al., 2010), while negative reviews have a
greater impact than positive reviews (Park and Lee, 2009a,b;
Herr et al., 1991).
Meanwhile, eWOM in social media (referred to as SM-eWOM
hereafter) has taken on a greater role in promoting online shopping
nowadays, especially with the emergence of social commerce.
These days, shoppers are turning to social media to share their
thoughts about the quality of goods or services they purchased,
or about their shopping experience. This form of SM-eWOM
includes consumer’s posts on Twitter, Facebook, Weibo and so
on, as well as comments on others’ posts. SM-eWOM impacts
consumers from the aspects of impression-management, emotion

http://dx.doi.org/10.1016/j.elerap.2016.03.004
1567-4223/Ó 2016 Elsevier B.V. All rights reserved.
 Q. Yan et al. / Electronic Commerce Research and Applications 17 (2016) 62–73
regulation, information acquisition, social bonding, and persuasion
(Berger, 2014). According to a report from the China Internet Net-
work Information Center (CNNIC) in 2012, 43.1% of users experi-
ence consumption impulses when they receive recommendations
from friends on social media, 38.3% of users will refer to SM-
eWOM before making shopping decisions, and 37.2% of users will
like to share their comments with friends on social media. Social
media has been used to figure out user preferences, and has
attracted attention from enterprises as a marketing tool.
In the research of Cheung and Thadani (2012), information
adoption is defined as a process in which people purposefully
engage in using information. In this paper, eWOM adoption means
that consumers agree with the content of eWOM and intend to
make purchases based on it.
Currently, the quality of EC-eWOM varies greatly. When con-
sumers visit e-commerce websites, they usually know little about
the authors of the EC-eWOM (Cheung and Thadani, 2012). So not
all EC-eWOM have an equal impact on consumers. Consumers
need to decide whether or not to adopt EC-eWOM based on their
own judgment. When consumers cannot make shopping decisions
based on EC-eWOM, they need to refer to alternative types of infor-
mation, such as SM-eWOM. Despite SM-eWOM currently being
less efficient than EC-eWOM, Internet companies have started
noticing its increasing impact and worked towards improving its
usability. For example, Sogou (www.sogou.com provides a function
to search information on social media ‘weixin.qq.com’. Sogou also
provides open API for programmers to develop more applications
to fully utilize social media content. Another example is from
Xiaomi. (www.mi.com). Xiaomi, one of the leading tech firms in
China, manufactures the popular Redmi and Mi smartphones.
The company attracts many young users and they share content
about Xiaomi’s products on social media. In return, Xiaomi
provides links to this content for consumers to easily access more
information about its products (e.g., www.linkedin.com/company/
xiaomi-technology). This makes both sharing and browsing
comments about Xiaomi’s products convenient.
Specifically, this research focuses on the following scenario.
People browse for goods on e-commerce website, and consumer’s
comments about the goods are available directly on the site,
including links to social media, such as Twitter. At first, consumers
use EC-eWOM. If they can’t make shopping decisions based on
EC-eWOM, they turn to SM-eWOM. In reality, this sequence could
be reversed, meaning consumers use SM-eWOM first, and then
EC-eWOM. However, this scenario is not the focus of this paper
and could be studied in another research.
Regarding EC-eWOM and SM-eWOM, there’s lots of research
demonstrating the individual impact each has on consumer deci-
sions. However, little is known about the consumers’ decisions
when they are simultaneously facing both types of eWOM. To
our knowledge, this is one of the first studies focusing on how con-
sumers choose between EC-eWOM and SM-eWOM. In addition,
previous studies have revealed that the cognitive level and involve-
ment of consumers moderates the impact of eWOM (Cheung et al.,
2009b; Park and Kim, 2009; Gupta and Harris, 2010). This research
will also verify the moderating effects of these factors on the adop-
tion of the two kinds of eWOM.
This research addresses the following questions: (1) which kind
of eWOM will be chosen by consumers when they are facing
EC-eWOM and SM-eWOM simultaneously? And (2) how does the
cognitive level and involvement of consumers moderate the
adoption process for the types of eWOM?
Overall, this research clarifies the mechanism of consumer’s
adoption of EC-eWOM and SM-eWOM from the perspective of cog-
nitive cost theory. The contributions include the following three
aspects: Firstly, it introduces a research model that analyzes the
impacts of multiple channels of eWOM on consumers’ decisions.
63
Attitudes towards EC-eWOM and SM-eWOM are not separated
but interact with each other. Second, the findings show that the
EC-eWOM adoption is not the end of consumers’ information gath-
ering, but instead, a mediating variable that may affect SM-eWOM
adoption. Third, the results demonstrate the influence of cognitive
cost on consumers’ utilization of online information from different
channels. The research also provides practical implications
regarding the management of eWOM on different types of online
platforms in the age of social commerce.
The remainder of the paper is organized as follows. Section 2 is
a review about the related literatures. Section 3 proposes the
research model and hypotheses. Empirical analysis is presented
in Section 4. Conclusions and research limitations are outlined in
Section 5.
2. Literature
2.1. Electronic word of mouth
Word of mouth (WOM) has been acknowledged for many years
as a major influence on what people know, feel and do (Buttle,
1998). It has been described as an informal communication direc-
ted at other consumers about the ownership, usage, or characteris-
tics of particular goods and services or their sellers (Berger, 2014).
EWOM is comments about a product or service provided by cus-
tomers via networks (Hennig-Thurau et al., 2003). It has a direct
impact on consumers’ trust and purchasing behavior (Dellarocas
et al., 2007; Duan et al., 2008). It can overcome the limitations of
traditional WOM, so therefore has been widely analyzed in the
field of e-commerce, information systems and marketing.
With the development of Web 2.0, social media has become
more and more popular. According to a report by CNNIC, as of June
2014, there were 275-million microbloggers in China. The contents
on social media have become an important information source to
help consumers make decisions. Consumers increasingly use com-
ments posted on Facebook, Twitter, etc., to valuate products and
services prior to making a purchase (Yan et al., 2014a). Carleen
(2009) pointed out that many patients use social media to discuss
medical services and their doctors in order to optimize treatments.
Social media content can also influence destination attractiveness
in consumers’ early travel decision-making stage (Shu and Scott,
2014). The companies can use social media as an additional
customer service and communication tool to gain insight into
consumers’ needs, wants, concerns and behaviors in order to serve
them better (He et al., 2013).
2.2. Elaboration likelihood model
The elaboration likelihood model (ELM) is an important informa-
tion processing theory (Park and Lee, 2009a,b; Cheung et al., 2008;
Chu and Kamal, 2008; Gupta and Harris, 2010; Lee et al., 2008;
Petty and Cacioppo, 1986), which explains how individuals are
influenced by the received messages in persuasive communication
(Shih et al., 2013). The model defines two routes for processing: the
central route and the peripheral route. The central route is used
when the message receiver is able to understand the message
without being distracted by any other superficial information.
The central route emphasizes high relevance of the message to
the individual. The greater the relevance and the more interest that
the individual shows in the subject of the message, the higher the
chances that they will think or elaborate on the message (Morris
et al., 2005). The peripheral route occurs when the message recei-
ver is unable or unwilling to engage in much thought about the
message. The message receiver decides whether to agree with
64 Q. Yan et al. / Electronic Commerce Research and Applications 17 (2016) 62–73
the message based on other cues besides the strength of the argu-
ments or ideas in the message.
Previous researchers have empirically confirmed the validity of
ELM in eWOM context. Argument quality (Cheung et al., 2009a;
Chu and Kamal, 2008; Park et al., 2007; Zhang and Watts, 2008)
and decisive information (Cheung et al., 2009b, 2012; Kamins
and Assael, 1987) have been considered as the most significant
central route factors; and source credibility (Cheung et al., 2008;
Zhang and Watts, 2008), information consistency (Zhang and
Watts, 2008), information volume and rating (Cheung et al.,
2009b, Duan et al., 2008; Park and Kim, 2009) are usually consid-
ered as the typical peripheral route factors (Luo et al., 2014).
2.3. EWOM adoption
Previous studies about eWOM can be divided into two groups:
population aspect and individual aspect. The former mainly
focuses on the relationship between eWOM and product sales
based on panel data collected from electronic commerce websites
(Chen and Xie, 2005; Cheung et al., 2012; Chevalier and Mayzlin,
2006; Clemons et al., 2006; Dellarocas et al., 2007; Duan et al.,
2008). The latter treats eWOM as an interactive process between
consumers. The communication between the senders and receivers
of eWOM finally influences the decision of receivers (Cheung et al.,
2009b; Park and Kim, 2009; Zhang and Watts, 2008). Our research
focuses on the adoption of eWOM during a consumer’s decision-
making process.
There have been numerous studies about the factors that may
impact the adoption of eWOM. Nonaka (1994) addressed informa-
tion adoption as a process of information transformation and inter-
nalization. Consumers make decisions based on eWOM because of
its informative and persuasive nature (Goh et al., 2013). According
to the theory of reasoned action (TRA) and the technology acceptance
model (TAM), usefulness is one of the key factors that impact the
acceptance of certain technology or behavior. Sussman and Siegal
(2003) extended the research about usefulness to the field of infor-
mation adoption. They addressed usefulness as a key antecedent
factor of the acceptance of information. Cheung et al. (2009a)
and Liu and Zhang (2010) proved that the usefulness significantly
affects the adoption of eWOM. Cheung et al. (2008) also pointed
out the usefulness of eWOM is affected by its relevance, timeliness,
accuracy and integrity. Baek et al. (2012) found that the consis-
tency of ratings, the count of vocabularies, and the proportions of
negative contents also affect the usefulness of eWOM. Sussman
and Siegal (2003) highlights the assessment of information useful-
ness as a mediator of the information adoption process.
Meanwhile, Cheung et al. (2009a) pointed out that the con-
sumers’ trust contributes to the adoption of eWOM. The research
of Kamins and Assael (1987) proved that the integrity of WOM also
impacts the perceived credibility of consumers. Mcknight et al.
(2002) argued that only when consumers believe the received
information is credible, will they adopt the eWOM to help them
make decisions. In contrast, it is less likely for consumers to adopt
eWOM if they consider the information unbelievable, thus the
impacts of the eWOM decrease correspondingly (Cheung et al.,
2009b). Cheung et al. (2009a) also proved that online information
credibility positively impacts the adoption. Further, previous
research showed that the credibility of information is influenced
by its source. Information is more credible when users are more
familiar with its source. Self (1996) proved that a difference exists
between the impact of information from sources with higher or
lower credibility. Regarding online shopping, consumers perceived
credibility of eWOM was more critical for their subsequent shop-
ping intention and behavior, because they could not touch the
physical goods (Fan et al., 2013).
2.4. Cognitive cost
The cognitive cost model plays an important role in under-
standing consumer decision-making (Punj and Moore, 2007). The
cognitive cost perspective in behavioral decision theory indicates
that consumers undertake cost-benefit tradeoffs when going
through a decision-making task (Todd and Benbasat, 1992). The
cognitive cost model theorizes that consumers display limited cog-
nitive abilities, and thus, their actions during decision-making
tasks depend on the associated cognitive costs (Payne, 1982). For
consumers, more information is not always better. If the informa-
tion load exceeds a certain threshold, it will decrease the con-
sumers’ decision quality and negatively impact the consumers’
shopping experience (Malhotra and Lagakos, 1982; Meyer, 1998;
Scammon, 1977). A recent study also pointed out that information
overload is a main cause of SNS fatigue (Lee et al., 2016).
Payne proposed a cost-benefit framework of cognition, which
was subsequently supported by other researchers (Bettman et al.,
1990). As per this framework, consumers continue to search for
additional information only until the benefit from additional infor-
mation is equal to or greater than the cognitive cost of the search.
Maity and Dass (2014) studied the cognitive cost caused by the dif-
ferent media richness in e-commerce and mobile commerce envi-
ronments, and addressed its influences on consumer’s behavior.
Kuksov and Villas-Boas (2010) has shown that information
overload and search costs can discourage decision makers from
evaluating choices, leading to a scenario where they make no
choices at all. Ghose et al. (2014) also proved that providing more
information can actually lead to fewer purchases. Regarding the
problems in our research, the overflow of online information
increases the cognitive costs to consumers, so they will not accept
all eWOM without considering the cognitive costs. The cognitive
cost is one of the underlying consideration of online information
adoption in the age of information explosion.
3. Model and hypotheses
3.1. Conceptual model
Based on the above-mentioned literatures, information useful-
ness is considered to be a key antecedent factor of information
adoption (Cheung et al., 2009a; Sussman and Siegal, 2003), which
has also been proved in the context of eWOM. Meanwhile,
researchers also use credibility theory to study online information
spreading, proving that credibility is another crucial antecedent
factor of eWOM adoption (Cheung et al., 2009b; McKnight and
Kacmar, 2007). Thus, our research deduces that usefulness and
credibility are two factors influencing eWOM adoption.
Partially based on the work of Cheung and Thadani (2012), we
figured out the independent variables that impact the eWOM use-
fulness and credibility, as illustrated in the conceptual model of
our research (Fig. 1). Consumers’ characters (cognitive level and
involvement degree) moderate the effects of independent variables
on eWOM usefulness and credibility (Park and Kim, 2009; Gupta
and Harris, 2010; Lee et al., 2008; Doh and Hwang, 2009). Our
model diverges from previous work by including SM-eWOM adop-
tion as a new dependent variable.
Compared with EC-eWOM, SM-eWOM is usually scattered
among the messages generated by social media users. SM-eWOM
displays a more unstructured data source. The rating and volume
of SM-eWOM are harder to measure than that of EC-eWOM. In
addition, according to the cognitive cost model, when consumers
browse e-commerce websites, they usually refer to SM-eWOM
only when EC-eWOM is not helpful enough to make decisions.
Thus, in this paper, we study the adoption of SM-eWOM from
the aspects of perceived usefulness and credibility of EC-eWOM.
 Q. Yan et al. / Electronic Commerce Research and Applications 17 (2016) 62–73
Cognitive
level
H7a
H8a
EC-eWOM H2
Volume
H3
EC-eWOM
Rate
Extremism
EC-eWOM H5
Integrity
EC-eWOM H6
Source
Credibility
3.2. Hypotheses
3.2.1. Usefulness related hypotheses
Based on the information adoption theory, information influ-
ences users’ behavior according to users’ judgment about the use-
fulness of information itself. As a special online information, the
usefulness of EC-eWOM impacts the adoption of EC-eWOM
(Cheung et al., 2009a; Liu and Zhang, 2010; Cheung et al., 2008).
According to cognitive cost theory, although EC-eWOM and SM-
eWOM exist simultaneously on the Internet, consumers do not
always consider both every time, in order to reduce the cognitive
cost (Salant, 2011; Liu et al., 2011; Brenčič, 2014). When shopping
on e-commerce websites, consumers refer to SM-eWOM only
when they cannot make shopping decisions based on the EC-
eWOM alone. Thus, we propose the following hypotheses:

Hypothesis 1a. Perceived EC-eWOM usefulness positively influ-
ences the EC-eWOM adoption.

Hypothesis 1b. Perceived EC-eWOM usefulness negatively
influences the SM-eWOM adoption.
With ELM, when the message receiver is unable or unwilling to
engage in much thought about the message, he decides whether to
agree with the message based on other cues, besides the strength
of the arguments or ideas in the message. Information volume
(Cheung et al., 2009b, Duan et al., 2008; Park and Kim, 2009) is
usually considered to be one of the typical peripheral route cues.
Past studies have proved the positive relationship between online
review volume and product sales (Liu, 2006; Dellarocas et al.,
2007; Duan et al., 2008; Park et al., 2012). Duan et al. (2008) argued
that the volume of EC-eWOM reflects a sense of presence for prod-
ucts, which arouses the awareness effect among consumers.
Reading numerous EC-eWOM by others can reduce consumers’
anxiety when making a purchase decision because consumers rea-
son that many others have also purchased the product. A higher
volume of EC-eWOM means the more sales of products (Duan
et al., 2008; Davis and Khazanchi, 2008; Liu, 2006), which in turn
attracts even more consumers. So from the perspective of
consumers, the volume is a predictor of product quality and
impacts the judgment of EC-eWOM usefulness for consumers
(Berger et al., 2010; Duan et al., 2008). In addition, Berger et al.,
65
Involvement Consumers
degree Characteristics
H7b
H8b
EC-eWOM H1a EC-eWOM
Usefulness Adoption
H1b
H9a
H4a
EC-eWOM SM-eWOM
Credibility H4b Adoption
Fig. 1. Conceptual model.
(2010) also empirically illustrated that even negative publicity
can increase purchase likelihood and sales by increasing product
awareness. Here we have the following hypothesis:

Hypothesis 2. EC-eWOM volume positively influences the per-
ceived EC-eWOM usefulness.
Most e-commerce websites provide rating mechanism. The EC-
eWOM rating means the overall scores given by other consumers
about on a product. Previous research has shown that the EC-
eWOM rating (Cheung et al., 2009b; Duan et al., 2008; Park and
Kim, 2009) has effect on the decision making process of consumers
as a typical peripheral route clue (Luo et al., 2014). The impacts of
EC-eWOM rate on consumers’ shopping decisions and products
sales are also discussed in many literatures (Dellarocas et al.,
2007; Ye et al., 2009; Lee and Lee, 2009; Öğüt and Onur Tasß,
2012; Yan et al., 2014b). Higher rates encourage consumption
while lower rates discourage shopping behavior (Baek et al.,
2012). Compared with medium rates, higher rates and lower rates
may impress consumers more strongly (Cheung et al., 2009b;
Mudambi and Schuff, 2010). While medium rate usually means
an ambiguous attitude towards a product, it may cause consumers
feeling confused and increase the difficulty to make purchase deci-
sions. In this research, we use rate extremism to represent that the
overall rate given by other consumers about a product is biased.
We propose:

Hypothesis 3. EC-eWOM rate extremism positively influences
the perceived EC-eWOM usefulness.
3.2.2. Credibility related hypotheses
Regarding online shopping, consumer’s perceived credibility of
eWOM was more critical for their subsequent shopping intention
and behavior, because they could not touch the physical goods
(Fan et al., 2013). When consumers consider the EC-eWOM to be
credible, they have the confidence to adopt it for assisting decision
making and do not spend any more energy on other sources of
information, to reduce the cognitive cost (Salant, 2011; Liu et al.,
2011; Brenčič, 2014). Otherwise, consumers will not adopt it for
the purpose of risk-reduction (Cheung et al., 2009b). When con-
sumers consider the EC-eWOM to be unreliable, they refer to other
66 Q. Yan et al. / Electronic Commerce Research and Applications 17 (2016) 62–73
sources of information, such as SM-eWOM, to assist decision mak-
ing. Thus, we propose the following hypotheses:

Hypothesis 4a. Perceived EC-eWOM credibility positively influ-
ences the EC-eWOM adoption.

Hypothesis 4b. Perceived EC-eWOM credibility negatively
influences the SM-eWOM adoption.
Integrity means that the EC-eWOM should include both positive
and negative information, covering the pros and cons of a product
or service. It impacts the credibility of EC-eWOM. Although posi-
tive EC-eWOM should be helpful in promoting positive attitudes
toward the products, a few negative messages within the majority
of positive messages are not critically harmful (Doh and Hwang,
2009). This is likely due to consumers’ suspicion that the set of
pure positive EC-eWOM stems from corporate unethical behavior
such as stealth marketing activities. Because some companies
reportedly try to manipulate and manage the voice of general con-
sumers with intentional interruptions in online user-to-user com-
munication, some participants did not trust the sincerity of perfect
sets with all positive EC-eWOM. This suspicion might influence the
results regarding the attitude toward the credibility of eWOM. In
other words, Integrity enhances the persuasion power and makes
EC-eWOM more credible (Kamins and Assael, 1987; Doh and
Hwang, 2009). The following hypothesis is proposed:

Hypothesis 5. EC-eWOM integrity positively influences the per-
ceived EC-eWOM credibility.
Source credibility is defined as the perceived ability and motiva-
tion of the message writer to produce accurate and truthful infor-
mation (Li and Zhan, 2011). It means the extent to which one
perceives a recommendation or review as believable, true, or fac-
tual (Cheung et al., 2009b). According to ELM, source credibility
significantly impacts individual’s attitude through the peripheral
route (Petty and Cacioppo, 1986), and it is a key factor in assisting
a consumer to judge online information (Wathen and Burkell,
2002). Fan et al. (2013) pointed out that source credibility is fairly
consistent with trustworthiness and expertise. In an online envi-
ronment, consumers rarely touch the product or recognize the
eWOM sender. Source credibility is a vital predictor in the early
stage when a consumer is choosing a website, and it contributes
to the perceived credibility of messages on the site (Dabholkar,
2006; Fan et al., 2013). Awad and Ragowsky (2008) suggested that
perceived credibility is a prime determinant in a consumer’s
decision-making process, and reduces uncertainty in both social
and business interactions. Messages from sources with high credi-
bility have greater influence on recipients (Hovland and Weiss,
1951). Source credibility plays an important role in consumer trust
in eWOM and positively influences the perceived eWOM credibil-
ity (Chih et al., 2013). Thus, we propose this hypothesis:

Hypothesis 6. EC-eWOM source credibility positively influences
the perceived EC-eWOM credibility.
3.2.3. Moderating effects
The effects of eWOM vary for different individuals. Regarding
the same eWOM, consumers may have different responses because
of their different cognitive level, experience and involvement
degree (Chaiken, 1980). Researchers widely agree that consumer-
related factors, such as consumer’s cognitive level and involvement
degree, greatly influence eWOM effects (Doh and Hwang, 2009;
Park and Kim, 2009; Fan et al., 2013; Cheung et al., 2009b). Accord-
ing to ELM, when the involvement degree of consumers is low,
eWOM mainly works through the peripheral route, whereas the
central route. Park and Kim (2009) pointed out that cognitive level
has obvious influences on consumers’ use of online reviews, mean-
while, customers with high expertise are able to evaluate informa-
tion according to their own experience and knowledge. Doh and
Hwang (2009) pointed out that consumers’ involvement and prior
knowledge partially moderate the relationship between the ratio
of messages and the eWOM effect.
Consumers tend to expand their information gathering as their
involvement with the product is increased. A higher involvement
may cause consumers to become suspicious about the positive
messages, and the effect of negative messages will thus be
increased. Fan et al., (2013) argued that consumers with higher
involvement have more motivation to understand products and
are more likely to read others’ opinions to gather more informa-
tion. Research of Cheung et al., (2009b) showed that involvement
significantly moderates the relationship between recommendation
consistency and perceived review credibility. They additionally
pointed out that different types of prior knowledge (on the review
topic, on the discussion forum) moderate the impact of different
determinants on eWOM credibility. For example, the higher the
consumer’s prior knowledge of the review topic, the less the source
credibility will affect perceived review credibility. Doh and Hwang
(2009) also proved that consumers with higher prior knowledge
can be more sensitive to negative messages than consumers with-
out prior knowledge.
In this article, we summarize the moderating factors in prior
research as the degree of involvement and the cognitive level. For
consumers with higher involvement degree and cognitive level,
they will have stronger intentions to have a comprehensive
decision-making process, and they will have a stronger ability to
make the purchase decision based on their own expertise, and
not only on the EC-eWOM. Thus, the following two groups of
hypotheses are proposed:

Hypothesis 7a. The cognitive level of consumers negatively
moderates the effects of EC-eWOM volume and rate extremism
on the perceived EC-eWOM usefulness.

Hypothesis 7b. The degree of involvement of consumers nega-
tively moderates the effects of EC-eWOM volume and rate
extremism on the perceived EC-eWOM usefulness.
and:

Hypothesis 8a. The cognitive level of consumers negatively
moderates the effects of EC-eWOM integrity and source credi-
bility on the perceived EC-eWOM credibility.

Hypothesis 8b. The degree of involvement of consumers nega-
tively moderates the effects of EC-eWOM integrity and source
credibility on the perceived EC-eWOM credibility.
3.2.4. Mediating effects
EC-eWOM and SM-eWOM are not processed separately. Infor-
mation from different channels impacts consumers together
(Munar and Jacobsen, 2013; Xu et al., 2014). According to studies
about cognitive cost, consumers undertake cost-benefit tradeoffs
when going through a decision-making task (Todd and Benbasat,
1992). Because of limited cognitive abilities, consumers continue
to search for additional information only until the benefit from
additional information is equal to or greater than the cognitive cost
of the search. They will not accept all eWOM without considering
the cognitive costs. Mitchell and McGoldrick (1996) pointed out
that when facing the uncertainty of information, individuals may
refer to information from social network sources for the purpose
of risk reduction. Social media has now become a platform for
eWOM communications (Chu and Kim, 2011).
 Q. Yan et al. / Electronic Commerce Research and Applications 17 (2016) 62–73 67

In our research scenario explained in the introduction section,
consumers access EC-eWOM first when they browse e-commerce
websites. When they cannot make shopping decisions based on
EC-eWOM, they refer to SM-eWOM to get more information and/
or to reduce risks. In more detail, when the perceived usefulness
and credibility of EC-eWOM decrease, the intention to adopt EC-
eWOM will also decrease. That means consumers may find diffi-
culty making decisions merely based on EC-eWOM. Under this cir-
cumstance, consumers may refer to SM-eWOM. If EC-eWOM is
enough for consumers’ to make shopping decisions, they will not
refer to SM-eWOM for the purpose of cost-reduction. Thus, the
adoptions of EC-eWOM and SM-eWOM are not isolated processes.
The consumers’ attitude towards EC-eWOM may influence their
attitude towards SM-eWOM. We propose:

Hypothesis 9a. EC-eWOM adoption negatively influences
SM-eWOM adoption.

Hypothesis 9b. EC-eWOM adoption mediates the effect of
perceived usefulness on SM-eWOM adoption.

Hypothesis 9c. EC-eWOM adoption mediates the effect of
perceived credibility on SM-eWOM adoption.
Our model differs from previous studies because it focuses on
the adoption of SM-eWOM. Although the independent variables
and mediating variable are all about the EC-eWOM, we emphasize
how the changes of these variables may finally influence the adop-
tion of SM-eWOM, through the mediating effects of EC-eWOM
adoption.
4. Empirical analysis
4.1. Questionnaire and data collection
The empirical data were collected through a questionnaire sur-
vey. The questionnaire includes three parts. The first describes the
survey so the respondents understand the questions. The second is
about the demographic information of the respondents. The third
is the main body of the survey which includes 30 items. To ensure
the content validity of the data, all the items for the theoretical
constructs are adopted from previous studies, with elaborate
adjustment, as illustrated in Table 1. The questionnaire uses a 5
point Likert scale, with 1 for ‘‘strongly disagree” and 5 for ‘‘strongly
agree.”
According to a survey conducted by CNNIC in March 2013,
63.4% of Internet users in China are of 18–30 years old and 41%
have a bachelor’s degree or above. Students account for more than
a quarter of Internet users. Based on the overall characteristics of
Internet users in China, our survey mainly aimed at college
students. It was conducted through an online survey website
(www.sojump.com). A total of 300 questionnaires were collected,
and after eliminating those with incomplete information, 289
samples were used for this empirical analysis.
4.2. Demographic and descriptive statistical analysis
The demographic characteristics provided in Table 2. According
to the results, 140 respondents were male and 149 were female
(48.4% vs 51.6%). Most of the people samples were young people,
with 77.9% of the respondents between 21 and 25 years of age.
We used ANOVA to verify if the demographic characteristics of
the samples impact the usefulness and credibility of EC-eWOM.
According to Tables 3 and 4, the values of p are all bigger than
0.05. There is no difference about the usefulness and credibility
of EC-eWOM among samples with different demographic
characteristics.
We utilized Cronbach’s alpha, average variance extracted (AVE)
and composite reliability (CR) for each construct to assess the con-
vergent validity (Fornell and Larcker, 1981). As shown in Table 5,
the values of Cronbach’s alpha for all of the constructs are above

Table 1
Constructs and instruments.

Constructs (sources) Definition Instruments

EC-eWOM volume (Park and Lee, 2009)
EC-eWOM rate extremism (Duan et al., 2008)
EC-eWOM integrity (Luo et al., 2014)
EC-eWOM source credibility (Sussman and Siegal, 2003)
Cognitive level (Park and Kim, 2009)
Degree of Involvement (Wang et al., 2012)
EC-eWOM usefulness (Sussman and Siegal, 2003)
EC-eWOM credibility (Cheung et al., 2009b)
EC-eWOM adoption (Sussman and Siegal, 2003)
SM-eWOM adoption (Sussman and Siegal, 2003)
Total number of posted reviews (Gauri et al.,
2008)
The arithmetic average of all the previous user
review ratings (Duan et al., 2008)
Comprehensiveness of messages refers to their
completeness (Cheung and Thadani, 2012)
Message source’s perceived ability (expertise)
or motivation to provide accurate and truthful
information (trustworthiness) (Cheung et al.,
2008)
The knowledge level (high or low) that
consumers have on products they tend to buy
(Park and Kim, 2009)
The degree of psychological identification and
affective, emotional ties the consumer has with
a stimulus or stimuli (Doh and Hwang, 2009)
The perception of an eWOM message being
useful (Cheung et al., 2008)
The perceived degree to which an eWOM
review provides accurate and truthful
information. (Doh and Hwang, 2009)
A process in which people purposefully engage
in using EC-eWOM (Cheung et al., 2008)
A process in which people purposefully engage
in using SM-eWOM (Cheung et al., 2008)
The quantity of EC-eWOM information is great
The number of EC-eWOM is large
The overall rate given by other consumers is high
The overall rate given by other consumers is low
The overall rate given by other consumers is biased
EC-eWOM only has positive or negative comments
EC-eWOM is biased towards one side or the other
EC-eWOM includes both positive and negative comments
The writer of EC-eWOM is an expert at the product
The writer of EC-eWOM is knowledgeable about the
product
The writer of EC-eWOM is trustworthy
The writer of EC-eWOM is reliable
I am knowledgeable about the product
I have rich purchasing experience on the product
I have the ability of judgement regarding the eWOM
The product is of concern to me
The product is important for me
The product is relevant to me
EC-eWOM is valuable
EC-eWOM is informative.
EC-eWOM is helpful
EC-eWOM is factual
EC-eWOM is accurate
EC-eWOM is credible
I agree with the content of EC-eWOM
I will follow the suggestion of EC-eWOM information
EC-eWOM motivates me to make purchase decision
I agree with the content of SM-eWOM
I will follow the suggestion of SM-eWOM information
SM-eWOM motivates me to make purchase decision
68 Q. Yan et al. / Electronic Commerce Research and Applications 17 (2016) 62–73

Table 2 Table 6
Demographic characteristics of the samples. Correlation Analysis of EC-eWOM Usefulness (Pearson).

Frequency Percentage Construct 1 2 3

(%)
1 Volume 1
Gender Male 140 48.4 2 Rate Extremism 0.502** 1
Female 149 51.6 3 Usefulness 0.877** 0.571** 1
Age Less than 20 19 6.6 **
p < 0.01.
21–25 225 77.9
26–30 40 13.8
More than 30 5 1.7
Online shopping Less than 1 year 19 6.6
regression to test the effects of volume and rate extremism on
experience 1-3 years 103 35.6
4–5 years 118 40.8 the usefulness of EC-eWOM. As shown in Table 7, the model is sig-
6–7 years 36 12.5 nificant, with F = 544.6 and p < 0.01, the adjusted R2 is 0.791. The
More than 7 years 13 4.5 volume and rate extremism positively impact the usefulness of
Education High school and 16 5.5 EC-eWOM, thus H2 and H3 are supported, and the impact of vol-
below ume is stronger than that of rate extremism.
Bachelor degree 91 31.5
Masters 162 56.1
Ph.D. 20 6.9
4.3.2. Analysis of EC-eWOM Credibility
We also performed correlation analysis to test the relation
between EC-eWOM credibility and its antecedent factors. As
Table 3
shown in Table 8, there are significant positive correlations
ANOVA Analysis of EC-eWOM Usefulness. between integrity, source credibility and the EC-eWOM credibility,
with Pearson coefficients of 0.871 and 0.910. The results in Table 8
F p
confirm H5 and H6. We performed multiple regression to test the
Gender 0.660 0.776 effects of integrity and source credibility on the EC-eWOM credi-
Age 1.009 0.438
Online shopping experience 1.082 0.376
bility. As shown in Table 9, the model is significant, with
Education 1.002 0.445 F = 750.9 and p < 0.01, the adjusted R2 is 0.839. The integrity and
source credibility positively impact EC-eWOM credibility, thus
H5 and H6 are supported.
Table 4
ANOVA Analysis of EC-eWOM Credibility. 4.3.3. Analysis of eWOM Adoption
F p
We also performed regression analysis between the EC-eWOM
usefulness, credibility and the adoption of EC-eWOM and SM-
Gender 0.915 0.526
Age 1.166 0.311
eWOM, as shown in Table 10.
Online shopping experience 0.802 0.638 The two models are both significant with p < 0.01, and
Education 0.441 0.936 F = 1266.2 and 946.8. The adjusted R2 values are 0.898 and 0.868,
indicating the good fit of the two models. Usefulness and credibil-
ity of EC-eWOM could be used to predict consumers’ behavior
0.7, all of the AVE values are above 0.6, and the composite reliabil- towards eWOM on both e-commerce websites and social media.
ity values are above 0.8. These results confirmed the convergent According to the coefficients in Table 10, the usefulness and cred-
validity of the data. ibility positively impact the adoption of EC-eWOM (b = 0.371,
p < 0.01; b = 0.628, p < 0.01), and negatively impact the adoption
of SM-eWOM (b = 0.274, p < 0.01; b = 0.700, p < 0.01). When
4.3. Model analysis consumers’ perceived usefulness and credibility of EC-eWOM are
high, they would prefer to make shopping decisions based on the
4.3.1. Analysis of EC-eWOM usefulness comments on e-commerce websites. However, when the perceived
We performed correlation and regression analysis with SPSS usefulness and credibility of EC-eWOM are low, consumers would
20.0. As shown in Table 6, there are significant positive correlations prefer to make decisions based on the comments from friends on
between volume, rate extremism and the usefulness of EC-eWOM, social media. Thus, H1a, H1b, H4a and H4b are supported. In addi-
with Pearson coefficients of 0.877 and 0.571, respectively. The tion, despite the adoption of EC-eWOM, or SM-eWOM, the impact
results in Table 6 confirm H2 and H3. We performed multiple of credibility is stronger than that of usefulness.

Table 5
Descriptive statistics and convergent validity test.

Construct items No. of Items Max Min Means Standard deviation Cronbach’s a AVE CR
Usefulness 3 5.00 1.33 3.64 0.885 0.845 0.804 0.907
Credibility 3 5.00 1.33 3.66 0.871 0.844 0.806 0.907
Volume 2 5.00 1.33 3.61 0.860 0.789 0.756 0.879
Rate extremism 3 4.67 1.33 3.72 0.823 0.915 0.886 0.941
Integrity 3 5.00 1.33 3.67 1.019 0.875 0.844 0.908
Source credibility 4 5.00 1.50 3.73 0.987 0.841 0.755 0.886
Involvement degree 3 4.67 1.33 3.54 0.864 0.773 0.702 0.874
Cognitive level 3 5.00 1.33 3.68 0.742 0.704 0.644 0.840
EC-eWOM adoption 3 5.00 1.33 3.71 0.939 0.834 0.764 0.901
SM-eWOM adoption 3 4.67 1.00 2.48 0.830 0.832 0.792 0.900
 Q. Yan et al. / Electronic Commerce Research and Applications 17 (2016) 62–73 69

Table 7
Regression analysis of EC-eWOM usefulness.

Constructs Unstandardized coefficients Standardized coefficients t p Adjusted R2 F

B Beta
Volume 0.812 0.789 25.307 0.000 0.791 544.6
Rate Extremism 0.188 0.175 5.623 0.000
Dependent variable: EC-eWOM usefulness

Table 8 As shown in Table 15, with the moderating effect, the 4R2 are
Correlation analysis of EC-eWOM credibility (Pearson). 0.001 and 0.002, indicating no improvement of model fit. The
results in Table 16 also show that the cognitive level’s moderating
Construct 1 2 3
effect for credibility is not significant (p = 0.179 and 0.012). This
1 Integrity 1
means the cognitive level of consumers doesn’t moderate the
2 Source credibility 0.906** 1
3 EC-eWOM credibility 0.871** 0.910** 1 impacts of the antecedent factors of EC-eWOM credibility. So
**
H8a is not supported.
p < 0.01.
R2 also changes little after including the moderating effect of
involvement degree, according to the results in Table 17. The
4.3.4. Analysis of moderating effects involvement degree’s moderating effect for credibility is not signif-
We discussed the moderating effects of consumer characteris- icant (p = 0.132 and 0.126) either, as shown in Table 18. The
tics (cognition level and degree of involvement). We first analyzed involvement degree of consumes doesn’t moderate the impacts
the moderating effects on the influences of the antecedent factors of the antecedent factors of EC-eWOM credibility. So H8b is not
of EC-eWOM usefulness, as shown in Tables 11–14. The variables supported either.
in the analysis are centered.
According to Table 11, the value of 4R2 indicates that the cog-
4.3.5. Analysis of the mediation effect
nitive level moderates the effects between volume, rate extremism
We tested the mediation effect of EC-eWOM adoption on SM-
and EC-eWOM usefulness. The moderating effects are proportional
eWOM adoption based on the method proposed by Wen et al.
to the impacts of the two independent variables, in other words,
(2005). In Fig. 2, Models 1–3 are used to verify the mediation effect
the stronger the impact of independent variable is, the stronger
on the relationship between usefulness and SM-eWOM adoption,
the moderating effect will be. The results in Table 12 also verify
and Models 4–6 are to verify the mediation effect on the relation-
the moderating effects (p < 0.01), and the cognitive level negatively
ship between credibility and SM-eWOM adoption. As illustrated in
moderates the impacts of independent variables, with b of 0.249
Fig. 2(a), to test the mediation effect of EC-eWOM adoption, we
and 0.137. H7a is supported. That means with the increase of the
first need to test the significance of coefficient c1 in Model 1. If c1
cognitive level of consumers, the impacts of volume and rate
is significant, then we need to further test the significance of coef-
extremism of EC-eWOM will become weaker.
ficient a1 and b1 in Models 2 and 3. If a1 and b1 are significant, then
Similarly, according to Tables 13 and 14, consumers’ involve-
we must verify the mediating effect according to c10 . If at least one
ment degree moderates the relationship between volume, rate
of a1 and b1 is not significant, then we must verify the mediating
extremism and EC-eWOM usefulness. The results in Table 14 verify
effect through the Sobel test. The verification of the mediation
the moderating effects (p < 0.01), and the involvement degree neg-
effect on the relationship between credibility and SM-eWOM
atively moderates the impacts of independent variables, with b
adoption are illustrated in Fig. 2(b). The results are shown in Tables
values of 0.164 and 0.121. H7b is supported. Thus, with increase
19 and 20.
in consumer involvement, the impacts of volume and rate extrem-
According to Fig. 2(a) and Table 19, the coefficient a1 (0.920), b1
ism of EC-eWOM will be weaker.
(0.853) and c1 (0.826) are all significant, but c10 did not pass the
Tables 15–18 show the analysis results about the moderating
significance test (0.042, p = 0.54 > 0.05). Thus, EC-eWOM adop-
effects on the influences of the antecedent factors of EC-eWOM
tion has a complete mediation effect on the relationship between
credibility. The variables in the analysis are centered.
EC-eWOM usefulness and SM-eWOM adoption. The decrease of

Table 9
Regression analysis of EC-eWOM credibility.

Construct Unstandardized coefficients Standardized coefficients t p Adjusted R2 F

B Beta
Integrity 0.221 0.259 4.642 0.000 0.839 750.9
Source credibility 0.596 0.675 12.111 0.000

Dep. var.: EC-eWOM Credibility.

Table 10
Regression analysis of EC-eWOM and SM-eWOM adoption.

Model Construct Unstandardized coefficient Standardized coefficient t p Adjusted R2 F

B Beta
1 Usefulness 0.400 0.371 12.13 0.000 0.898 1,266.2
Credibility 0.666 0.628 20.53 0.000
2 Usefulness 0.257 0.274 7.89 0.000 0.868 946.8
Credibility 0.667 0.700 20.14 0.000

Model 1: Dep. var.: EC-eWOM adoption. Model 2: Dep. var.: SM-eWOM adoption.
70 Q. Yan et al. / Electronic Commerce Research and Applications 17 (2016) 62–73

Table 11 Table 15
Fit analysis of the moderating effect of cognitive level for usefulness. Fit analysis of the moderating effect of cognitive level for credibility.

Model Adjusted R2 DR2 Model Adjusted R2 DR2

Baseline 0.791 Baseline 0.840
Model 1 0.819 0.028 Model 1 0.839 0.001
Model 2 0.803 0.012 Model 2 0.842 0.002

Independent variables in each model: Baseline: Cognitive level, Volume, Rate Independent variables in each model:
extremism. Model 1: Cognitive level * Volume, Cognitive level, Volume, Rate extremism. Baseline: Cognitive level, Integrity, Source credibility.
Model 2: Cognitive level * Rate extremism, Cognitive level, Volume, Rate extremism. Model 1: Cognitive level * Integrity, Cognitive level, Integrity, Source credibility.
Model 2: Cognitive level * Source credibility, Cognitive level, Integrity, Source credibility.

Table 12
Regression analysis of the moderating effect of cognitive level for usefulness.

Variable Unstandardized Standardized t p Table 16

coefficient coefficient Regression analysis of the moderating effect of cognitive level for credibility.

B Beta Variable Unstandardized Standardized T P

coefficient coefficient
Volume 0.733 0.712 23.202 0.000
Rate extremism 0.141 0.131 4.245 0.000 B Beta
Cognitive level 0.615 0.319 6.779 0.000
Integrity 0.177 0.207 3.555 0.000
Cognitive 0.372 0.249 5.860 0.000
Source credibility 0.605 0.687 11.844 0.000
level * Volume
Cognitive level 0.135 0.071 1.751 0.081
Cognitive level * Rate 0.217 0.137 2.981 0.003
Cognitive 0.133 0.111 1.348 0.179
extremism
level * Integrity
The significance level is p < 0.01. Cognitive 0.254 0.205 2.524 0.012
level * Source
credibility

Table 13
Fit analysis of the moderating effects of degree of involvement for usefulness.

Model Adjusted R2 4R2

Table 17
Baseline 0.794 Fit analysis of the moderating effect of degree of involvement for credibility.
Model 1 0.808 0.014
Mode 2 0.802 0.008 Model Adjusted R2 4R2
0.838
Independent variables in each model: Baseline: Degree of involvement, Volume, Rate
Model 1 0.838 0
extremism. Model 1: Degree of involvement * Volume, Degree of involvement, Volume,
Model 2 0.838 0
Rate extremism. Model 2: Degree of involvement * Rate extremism, Degree of
involvement, Volume, Rate extremism Independent variables in each model: Baseline: Degree of involvement, Integrity,
Source credibility. Model 1: Degree of involvement * Integrity, Degree of involvement,
Integrity, source credibility. Model 2: Degree of involvement * Source credibility, Degree
the usefulness of EC-eWOM will weaken the adoption of EC- of involvement, Integrity, Source credibility.
eWOM, which in turn will promote the adoption of SM-eWOM.
The results support H9a and H9b.
According to Fig. 2(b) and Table 20, the coefficient a2 (0.865), b2
Table 18
(0.390), c2 (0.917) and c20 (0.579) are all significant. Thus, EC-
Regression analysis of moderating effect of degree of involvement for credibility.
eWOM adoption has a partial mediation effect on the relationship
between EC-eWOM credibility and SM-eWOM adoption. The Variable Unstandardized Standardized t p
coefficient coefficient
decrease of the credibility of EC-eWOM will weaken the adoption
of EC-eWOM, while simultaneously promoting the adoption of B Beta
SM-eWOM. The decrease of the intention to adopt EC-eWOM also Integrity 0.213 0.249 4.40 0.000
promotes the adoption of SM-eWOM. The results support also H9a Source credibility 0.600 0.681 11.96 0.000
Involvement degree 0.010 0.006 0.17 0.869
and H9c.
Involvement 0.128 0.103 1.51 0.132
degree * Integrity
Involvement 0.133 0.102 1.53 0.126
4.3.6. Results of verification of the hypotheses
degree * Source
Table 21 concludes the verification results for the hypotheses in credibility
our research.

Table 14
Regression analysis of the moderating effect of degree of involvement for usefulness.

Variable Unstandardized coefficient Standardized coefficient t p

B Beta
Volume 0.759 0.738 23.62 0.000
Rate extremism 0.140 0.130 4.05 0.000
Degree of Involvement 0.437 0.258 5.87 0.000
Degree of involvement * Volume 0.256 0.164 4.16 0.000
Degree of involvement * Rate extremism 0.180 0.121 2.83 0.005
 Q. Yan et al. / Electronic Commerce Research and Applications 17 (2016) 62–73 71

EC-eWOM SM-eWOM EC-eWOM SM-eWOM

¦1 ¦4
Usefulness Adoption Credibility Adoption

EC-eWOM EC-eWOM
¦2 ¦5
Adoption Adoption

EC-eWOM SM-eWOM EC-eWOM SM-eWOM

¦3 ¦6
Usefulness Adoption Credibility Adoption

SM-eWOM_Adoption=c1*EC-eWOM_Usefulness+¦1 (1) SM-eWOM_Adoption=c2*EC-eWOM_Credibility+¦4 (4)

EC-eWOM_Adoption=a1*EC-eWOM_Usefulness+¦2 (2) EC-eWOM_Adoption=a2*EC-eWOM_Credibility+¦5 (5)

SM-eWOM_Adoption=c’1*EC-eWOM_Usefulness SM-eWOM_Adoption=c’2*EC-eWOM_Credibility
+b1*EC-eWOM_Adoption+¦3 (3) +b2*EC-eWOM_Adoption+¦6 (6)

(a) (b)
Fig. 2. Analysis of the mediation effect of EC-eWOM adoption.

Table 19 Table 21
Mediation effect of EC-eWOM adoption on relation between usefulness and SM- Results of hypotheses verification.
eWOM adoption.
No. Hypothesis Result
Model Coefficient T p Adjusted R2
H1a Perceived EC-eWOM usefulness positively influences the Supported
(1) c1 0.826 24.8 0.00 0.68 EC-eWOM adoption
(2) a1 0.920 39.8 0.00 0.85 H1b Perceived EC-eWOM usefulness negatively influences the Supported
(3) b1 0.853 12.5 0.00 0.79 SM-eWOM adoption
c10 0.042 0.61 0.54 H2 EC-eWOM volume positively influences the perceived EC- Supported
eWOM usefulness
H3 EC-eWOM rate extremism positively influences the Supported
perceived EC-eWOM usefulness
Table 20 H4a Perceived EC-eWOM credibility positively influences the Supported
Mediation effect of EC-eWOM adoption on relation between credibility and SM- EC-eWOM adoption
eWOM adoption. H4b Perceived EC-eWOM credibility negatively influences the Supported
SM-eWOM adoption
Model Coefficient t p Adjusted R2
H5 EC-eWOM integrity positively influences the perceived Supported
(4) c2 0.917 38.8 0.00 0.84 EC-eWOM credibility
(5) a2 0.865 29.3 0.00 0.75 H6 EC-eWOM source credibility positively influences the Supported
(6) b2 0.390 -9.47 0.00 0.88 perceived EC-eWOM credibility
c20 0.579 -14.1 0.00 H7a The Cognitive level of consumers negatively moderates Supported
the effects of EC-eWOM volume and rate extremism on
the perceived EC-eWOM usefulness
H7b The involvement degree of consumers negatively Supported
5. Conclusion moderates the effects of EC-eWOM volume and rate
extremism on the perceived EC-eWOM usefulness
5.1. Academic contributions H8a The cognitive level of consumers negatively moderates Not
the effects of EC-eWOM integrity and source credibility supported
on the perceived EC-eWOM credibility
UGC has become an important information source for con-
H8b The involvement degree of consumers negatively Not
sumers to make shopping decisions. Studies have proven that the moderates the effects of EC-eWOM integrity and source supported
perceived usefulness and credibility has significant influences on credibility on the perceived EC-eWOM credibility
the adoption of EC-eWOM. The volume and rating of EC-eWOM H9a EC-eWOM adoption negatively influences SM-eWOM Supported
also impacts the sales of products. Meanwhile, with the develop- adoption
H9b EC-eWOM adoption mediates the effect of perceived Supported
ment of social media, the influences of SM-eWOM are increasing.
usefulness on SM-eWOM adoption
Social media based marketing has also attracted the attentions of H9c EC-eWOM adoption mediates the effect of perceived Supported
enterprises. However, there is currently little research studying credibility on SM-eWOM adoption
the relationship between EC-eWOM and SM-eWOM. To compen-
sate this gap, our research focuses on the adoption process when
consumers face these two kinds of eWOM simultaneously.
The main content of our research includes the following three consumers’ decision-making, the high volume of online informa-
aspects. First, it is one of the first to study the inter-relationship tion increases the cognitive cost for consumers. To reduce this cost,
between the adoption of EC-eWOM and SM-eWOM. It enriches consumers do not adopt all information without selection. For cus-
the theoretical research about eWOM under the circumstances of tomers of e-commerce, if EC-eWOM can help them make their
social commerce. EC-eWOM’s usefulness and credibility positively shopping decisions, they prefer not to refer to SM-eWOM anymore,
impact the adoption of EC-eWOM, but negatively impact the adop- to reduce the cognitive cost. Only when EC-eWOM is not enough
tion of SM-eWOM. The adoption of EC-eWOM negatively impacts during the decision making process do customers of e-commerce
that of SM-eWOM, and mediates the effects of usefulness and cred- prefer to refer to SM-eWOM.
ibility on SM-eWOM adoption. Finally, consistent with previous researches, this research veri-
Second, our research explains the adoption process of the two fies the positive relation between volume, rate extremism and use-
kinds of eWOM based on the cognitive cost theory. Although EC- fulness of EC-eWOM, as well as the positive relation between
eWOM and SM-eWOM both provide supportive information for integrity, source credibility and the EC-eWOM credibility.
72 Q. Yan et al. / Electronic Commerce Research and Applications 17 (2016) 62–73
This research also shows that the influence of credibility on the
adoption of both kinds of eWOM is stronger than that of useful-
ness. This may because under the network environment, con-
sumers may make more efforts to figure out if the eWOM is
believable or not based on the credibility related clues. The integ-
rity and source credibility are both factors that may improve the
perceived credibility of eWOM. The research further proves that
the cognitive level and involvement degree moderate the per-
ceived usefulness of EC-eWOM. But the moderating effect for the
perceived credibility is not significant. According to the results,
for consumers with high cognitive and involvement level, the
impacts of volume and rate extremism of EC-eWOM on perceived
usefulness will decrease. That means during the information pro-
cess, consumers with high cognitive and involvement level mainly
make decisions based on their own knowledge, experience and
judgment. They are more likely to determine the usefulness of
eWOM reasonably. Thus the impacts of volume and rate on the
usefulness of eWOM will decrease correspondingly.
In conclusion, our study makes the following theoretical contri-
butions. First, it introduces a research model that analyzes the
impacts of multiple channels of eWOM on consumers’ decisions.
The attitudes towards EC-eWOM and SM-eWOM are not separated,
and interact with each other. Second, the findings show that the
EC-eWOM adoption is not the end of consumers’ information pro-
cessing, but is a mediating variable that may affect the SM-eWOM
adoption. Third, the results demonstrate the influences of cognitive
cost on consumers’ utilization of online information from different
channels.
5.2. Practical implications
Our research has several practical implications. First, for the
e-commerce operators, they should further enhance the SNS
elements of the websites, bridge the e-commerce websites with
social media, and proactively present the SM-eWOM to their
customers. Currently, consumers need to click the hyperlinks for
social media by themselves to find out the information they want.
In the future, if the e-commerce websites can provide SM-eWOM
proactively, according to the identities of consumers and the
products they are interested in, it would help consumers to make
purchase decisions more efficiently. Second, for social media
operators, it is worthwhile to improve the organization of the
SM-eWOM to make them easier to be accessed and retrieved by
consumers. The ability of socialized marketing can also be
enhanced based on the improved management of SM-eWOM.
5.3. Limitations
There are also some limitations of our research. First, our
research did not focus on the factors directly influencing the SM-
eWOM, but instead studied the relation between the adoption of
EC-eWOM and SM-eWOM, from the perspective of e-commerce
customers. This is because the organization of SM-eWOM is not
as good as that of EC-eWOM. On the other hand, for e-commerce
customers, they refer to eWOM during browsing on e-commerce
websites in the specific research scenario. So it is reasonable to
research the adoption of SM-eWOM based on its relationship with
EC-eWOM. However, with the improvement of SM-eWOM man-
agement, as well as with the increase of SM-eWOM impacts,
there’s a need to study the direct influencing factors for the adop-
tion of SM-eWOM.
Second, this research did not consider the impact of different
product categories (experience products versus search products).
Previous studies have proved the moderating effects of product
categories on the adoption of eWOM (Baek et al., 2012; Bae and
Lee, 2011). So the question about the impact of product categories
when consumers are facing different kinds of eWOM simultane-
ously needs to be further considered. Third, our research focuses
on a specific scenario where consumers browse eWOM on e-
commerce websites at the beginning and only refer to SM-
eWOM when necessary. However, the sequence could be reversed,
meaning consumers may refer to SM-eWOM first, and then EC-
eWOM. Future research could include more practical scenarios.
Finally, our research doesn’t cover all possible antecedent factors
for usefulness and credibility in order to be more focused on the
relationship between EC-eWOM and SM-eWOM. More factors
could be included in the future research of this topic.
Acknowledgments
The research was supported by Program for New Century Excel-
lent Talents in University (NCET-11-0597), Beijing Philosophy and
Social Science Planning Project (13KDB011).
References
Awad, N.F., Ragowsky, A., 2008. Establishing trust in electronic commerce through
online word of mouth: an examination across genders. J. Manage. Inf. Syst. 24
(4), 101–121.
Bae, S., Lee, T., 2011. Product type and consumers’ perception of online consumer
reviews. Electron. Markets 21 (4), 255–266.
Baek, H., Ahn, J., Choi, Y., 2012. Helpfulness of online consumer reviews: readers’
objectives and review cues. Int. J. Electron. Commerce 17 (2), 99–126.
Berger, J., 2014. Word of mouth and interpersonal communication: a review and
directions for future research. J. Consum. Psychol. 24 (4), 586–607.
Berger, J., Sorensen, A.T., Rasmussen, S.J., 2010. Positive effects of negative publicity:
when negative reviews increase sales. Marketing Sci. 29 (5), 815–827.
Bettman, J.R., Johnson, E.J., Payne, J.W., 1990. A componential analysis of cognitive
effort in choice. Organ. Behav. Hum. Decis. Process. 45 (1), 111–139.
Brenčič, V., 2014. Search online: evidence from acquisition of information on online
job boards and resume banks. J. Econ. Psychol. 42 (C), 112–125.
Buttle, F.A., 1998. Word of mouth: understanding and managing referral marketing.
J. Strategic Marketing 6 (3), 241–254.
Carleen, H., 2009. Take two aspirin and tweet me in the morning: how Twitter,
Facebook, and other social media are reshaping health care. Health Aff. 28 (2),
361–368.
Chaiken, S., 1980. Heuristic versus systematic information processing and the use of
source versus message cues in persuasion. J. Pers. Soc. Psychol. 39 (5), 752.
Chatterjee, P., 2001. Online reviews: do consumers use them? Adv. Consum. Res. 28
(1), 129–133.
Chen, Y., Xie, J., 2005. Third-party product review and firm marketing strategy.
Marketing Sci. 24 (2), 218–240.
Cheung, C., Xiao, B., Liu, I.L., 2012. The impact of observational learning and
electronic word of mouth on consumer purchase decisions: the moderating role
of consumer expertise and consumer involvement. In: Sprague, R. (Ed.),
Proceedings of the 45th Hawaii International Conference on System Science.
IEEE Computer Society Press, Washington, DC, pp. 3228–3237.
Cheung, C.M., Lee, M.K., Rabjohn, N., 2008. The impact of electronic word-of-mouth:
the adoption of online opinions in online customer communities. Internet Res.
18 (3), 229–247.
Cheung, C.M., Lee, M.K., Thadani, D.R., 2009a. The impact of positive electronic
word-of-mouth on consumer online purchasing decision. In: Proceedings of the
2nd World Summit on the Knowledge Society (WSKS 2009), Chania, Greece.
Springer, Berlin Heidelberg, Germany, pp. 501–510.
Cheung, C.M., Thadani, D.R., 2012. The impact of electronic word-of-mouth
communication: a literature analysis and integrative model. Decis. Support
Syst. 54 (1), 461–470.
Cheung, M.Y., Luo, C., Sia, C.L., Chen, H., 2009b. Credibility of electronic word-of-
mouth: informational and normative determinants of on-line consumer
recommendations. Int. J. Electron. Commerce 13 (4), 9–38.
Chevalier, J.A., Mayzlin, D., 2006. The effect of word of mouth on sales: online book
reviews. J. Marketing Res. 43 (3), 345–354.
Chih, W.H., Wang, K.Y., Hsu, L.C., Huang, S.C., 2013. Investigating electronic word-
of-mouth effects on online discussion forums: the role of perceived positive
electronic word-of-mouth review credibility. Cyberpsychol. Behav. Soc. Netw.
16 (9), 658–668.
Chu, S.C., Kamal, S., 2008. The effect of perceived blogger credibility and argument
quality on message elaboration and brand attitudes: an exploratory study. J.
Interact. Advertising 8 (2), 26–37.
Chu, S.C., Kim, Y., 2011. Determinants of consumer engagement in electronic word-
of-mouth (eWOM) in social networking sites. Int. J. Advertising 30 (1), 47–75.
Clemons, E.K., Gao, G.G., Hitt, L.M., 2006. When online reviews meet hyper
differentiation: a study of the craft beer industry. J. Manage. Inf. Syst. 23 (2),
149–171.
 Q. Yan et al. / Electronic Commerce Research and Applications 17 (2016) 62–73
Dabholkar, P.A., 2006. Factors influencing consumer Choice of a ‘‘rating Web site ”:
an experimental investigation of an online interactive decision aid. J. Marketing
Theory & Practice 14 (4), 259–273.
Davis, A., Khazanchi, D., 2008. An empirical study of online word of mouth as a
predictor for multi product category e-commerce sales. Electron. Markets 18
(2), 130–141.
Dellarocas, C., Zhang, X.M., Awad, N.F., 2007. Exploring the value of online product
reviews in forecasting sales: the case of motion pictures. J. Interact. Marketing
21 (4), 23–45.
Doh, S.J., Hwang, J.S., 2009. How consumers evaluate eWOM (electronic word-of-
mouth) messages. Cyber Psychol. Behav. 12 (2), 193–197.
Duan, W., Gu, B., Whinston, A.B., 2008. Do online reviews matter? an empirical
investigation of panel data. Decis. Support Syst. 45 (4), 1007–1016.
Fan, Y.W., Miao, Y.F., Fang, Y.H., Lin, R.Y., 2013. Establishing the adoption of
electronic word-of-mouth through consumers’ perceived credibility. Int. Bus.
Res. 6 (3), 58–65.
Fornell, C., Larcker, D.F., 1981. Evaluating structural equation models with
unobservable variables and measurement error. J. Marketing Res., 39–50
Gauri, D., Bhatnagar, A., Rao, R., 2008. Role of word of mouth in online store loyalty,
association for computing machinery. Commun. ACM 51 (3), 89–91.
Ghose, A., Ipeirotis, P.G., Li, B., 2014. Examining the impact of ranking on consumer
behavior and search engine revenue. Manage. Sci. 60 (7), 1632–1654.
Goh, K.Y., Heng, C.S., Lin, Z., 2013. Social media brand community and consumer
behavior: quantifying the relative impact of user-and marketer-generated
content. Inf. Syst. Res. 24 (1), 88–107.
Gupta, P., Harris, J., 2010. How e-WOM recommendations influence product
consideration and quality of choice: a motivation to process information
perspective. J. Bus. Res. 63 (9), 1041–1049.
He, W., Zha, S., Li, L., 2013. Social media competitive analysis and text mining: a case
study in the pizza industry. Int. J. Inf. Manage. 33 (3), 464–472.
Hennig-Thurau, T., Walsh, G., Walsh, G., 2003. Electronic word-of-mouth: motives
for and consequences of reading customer articulations on the Internet. Int. J.
Electron. Commerce 8 (2), 51–74.
Herr, P.M., Kardes, F.R., Kim, J., 1991. Effects of word-of-mouth and product-
attribute information on persuasion: an accessibility-diagnosticity perspective.
J. Consum. Res. 17 (1), 454–462.
Hovland, C.I., Weiss, W., 1951. The influence of source credibility on communication
effectiveness. Public Opin. Q. 15 (4), 635–650.
Kamins, M.A., Assael, H., 1987. Two-sided versus one-sided appeals: a cognitive
perspective on argumentation, source derogation, and the effect of
disconfirming trial on belief change. J. Marketing Res., 29–39
Kuksov, D., Villas-Boas, J.M., 2010. When more alternatives lead to less choice.
Marketing Sci. 29 (3), 507–524.
Lee, A.R., Son, S.M., Kim, K.K., 2016. Information and communication technology
overload and social networking service fatigue: a stress perspective. Comput.
Human Behav. 55 (A), 51–61.
Lee, J., Lee, J.N., 2009. Understanding the product information inference process in
electronic word-of-mouth: an objectivity: subjectivity dichotomy perspective.
Inf. Manage. 46 (5), 302–311.
Lee, J., Park, D.H., Han, I., 2008. The effect of negative online consumer reviews on
product attitude: an information processing view. Electron. Commerce Res.
Appl. 7 (3), 341–352.
Li, J., Zhan, L., 2011. Online persuasion: how the written word drives WOM evidence
from consumer-generated product reviews. J. Advertising Res. 51 (1), 239–257.
Liang, T.P., Lai, H.J., Ku, Y.C., 2007. Personalized content recommendation and user
satisfaction: theoretical synthesis and empirical findings. J. Manage. Inf. Syst. 23
(3), 45–70.
Liu, Q.B., Karahanna, E., Watson, R.T., 2011. Unveiling user-generated content:
designing websites to best present customer reviews. Bus. Horiz. 54 (3), 231–
240.
Liu, R.R., Zhang, W., 2010. Informational influence of online customer feedback: an
empirical study. J. Database Marketing Customer Strategy Manage. 17 (2), 120–
131.
Liu, Y., 2006. Word of mouth for movies: its dynamics and impact on box office
revenue. J. Marketing 70 (3), 74–89.
Luo, C., Wu, J., Shi, Y., Xu, Y., 2014. The effects of individualism–collectivism cultural
orientation on eWOM information. Int. J. Inf. Manage. 34 (4), 446–456.
Maity, M., Dass, M., 2014. Consumer decision-making across modern and traditional
channels: e-commerce, m-commerce, in-store. Decis. Support Syst. 61, 34–46.
Malhotra, N.K., Lagakos, S.W., 1982. The information overload controversy: an
alternative viewpoint. J. Marketing 46 (2), 27–37.
McKnight, D.H., Choudhury, V., Kacmar, C., 2002. The impact of initial consumer
trust on intentions to transact with a web site: a trust building model. J.
Strategic Inf. Syst. 11 (3), 297–323.
73
McKnight, D.H., Kacmar, C.J., 2007. Factors and effects of information credibility. In:
Proceedings of the Ninth International Conference on Electronic Commerce,
August. ACM Press, New York, NY, pp. 423–432.
Meyer, J.A., 1998. Information overload in marketing management. Marketing
Intell. Plann. 16 (3), 200–209.
Mitchell, V.W., McGoldrick, P.J., 1996. Consumer’s risk-reduction strategies: a
review and synthesis. Int. Rev. Retail, Distrib. Consum. Res. 6 (1), 1–33.
Morris, J.D., Woo, C., Singh, A.J., 2005. Elaboration likelihood model: a missing
intrinsic emotional implication. J. Targeting, Meas. Anal. Marketing 14 (1), 79–
98.
Mudambi, S.M., Schuff, D., 2010. What makes a helpful online review? a study of
customer reviews on Amazon.com. MIS Q. 34 (1), 185–200.
Munar, A.M., Jacobsen, J.K.S., 2013. Trust and involvement in tourism social media
and web-based travel information sources. Scand. J. Hospitality Tourism 13 (1),
1–19.
Nonaka, I., 1994. A dynamic theory of organizational knowledge creation. Organ.
Sci. 5 (1), 14–37.
Öğüt, H., Onur Tasß, B.K., 2012. The influence of Internet customer reviews on the
online sales and prices in hotel industry. Serv. Ind. J. 32 (2), 197–214.
Park, C., Lee, T.M., 2009a. Information direction, website reputation and eWOM
effect: a moderating role of product type. J. Bus. Res. 62 (1), 61–67.
Park, D.H., Kim, S., 2009. The effects of consumer knowledge on message processing
of electronic word-of-mouth via online consumer reviews. Electron. Commerce
Res. Appl. 7 (4), 399–410.
Park, D.H., Lee, J., 2009b. EWOM overload and its effect on consumer behavioral
intention depending on consumer involvement. Electron. Commerce Res. Appl.
7 (4), 386–398.
Park, D.H., Lee, J., Han, I., 2007. The effect of on-line consumer reviews on consumer
purchasing intention: the moderating role of involvement. Int. J. Electron.
Commerce 11 (4), 125–148.
Park, J., Gu, B., Lee, H., 2012. The relationship between retailer-hosted and third-
party hosted WOM sources and their influence on retailer sales. Electron.
Commerce Res. Appl. 11 (3), 253–261.
Payne, J.W., 1982. Contingent decision making. Psychol. Bull. 9 (2), 382–402.
Petty, R.E., Cacioppo, J.T., 1986. Communication and Persuasion: Central and
Peripheral Routes to Attitude Change. Springer-Verlag, New York, NY.
Punj, G.N., Moore, R., 2007. Smart versus knowledgeable online recommendation
agents. J. Interact. Marketing 21 (4), 46–60.
Salant, Y., 2011. Procedural analysis of choice rules with applications to bounded
rationality. Am. Econ. Rev. 101 (2), 724–748.
Scammon, D.L., 1977. Information load and consumers. J. Consumer Res. 4 (3), 148–
155.
Self, C.S., 1996. Credibility. In: Salwen, M., Stacks, D. (Eds.), An Integrated Approach
to Communication Theory and Research. Erlbaum, Mahwah, NJ.
Shih, H.P., Lai, K.H., Cheng, T.C., 2013. Informational and relational influences on
electronic word of mouth: an empirical study of an online consumer discussion
forum. Int. J. Electron. Commerce 17 (4), 137–166.
Shu, M., Scott, N., 2014. Influence of social media on Chinese students’ choice of an
overseas study destination: an information adoption model perspective. J.
Travel Tourism Marketing 31 (2), 286–302.
Sussman, S.W., Siegal, W.S., 2003. Informational influence in organizations: an
integrated approach to knowledge adoption. Inf. Syst. Res. 14 (1), 47–65.
Todd, P., Benbasat, I., 1992. The use of information in decision making: an
experimental investigation of the impact of computer-based decision aids. MIS
Q. 31 (1), 373–393.
Xu, L., Duan, J.A., Whinston, A., 2014. Path to purchase: a mutually exciting point
process model for online advertising and conversion. Manage. Sci. 60 (6), 1392–
1412.
Zhang, W., Watts, S.A., 2008. Capitalizing on content: information adoption in two
online communities. J. Assoc. Inf. Syst. 9 (2), 73–94.
Wang, X., Yu, C., Wei, Y., 2012. Social media peer communication and impacts on
purchase intentions: a consumer socialization framework. J. Interactive
Marketing 26 (4), 198–208.
Wathen, C.N., Burkell, J., 2002. Believe it or not: factors influencing credibility on the
web. J. Am. Soc. Inform. Sci. Technol. 53 (2), 134–144.
Wen, Z.L., Hau, K.T., Chang, L., 2005. A comparison of moderator and mediator and
their applications. Acta Psychol. Sin. 37 (2), 268–274.
Yan, G., He, W., Shen, J., Tang, C., 2014a. A bilingual approach for conducting Chinese
and English social media sentiment analysis. Comput. Netw. 75, 491–503.
Yan, Q., Wang, Q., Liu, X., 2014b. Research on the interactive effects of online scores.
Electron. Commerce Res. Appl. 13 (6), 402–408.
Ye, Q., Law, R., Gu, B., 2009. The impact of online user reviews on hotel room sales.
Int. J. Hospitality Manage. 28 (1), 180–182.