Information & Management 50 (2013) 609–620

Contents lists available at ScienceDirect

Information & Management

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

Intention to purchase on social commerce websites across cultures:

A cross-regional study
Celeste See-Pui Ng *
Department of Information Management, Yuan Ze University, Taiwan

A R T I C L E I N F O A B S T R A C T

Article history: This study investigates the moderating effect of culture and the mediating effect of trust in a social
Received 7 September 2012 network community on the relationship between social interactions and purchase intention. Trust
Received in revised form 5 August 2013 transference theory, social interactions, and Hofstede’s cultural dimensions are considered. The findings
Accepted 10 August 2013
from the research support the moderating effect of culture and the mediating role of trust in a social
Available online 26 August 2013
network community on the relationship between social interactions (in terms of closeness and
familiarity) and intention to purchase in social commerce environments. In addition, the mediating
Keywords:
effect of trust in a social network community is conditional on culture. The findings also suggest that
Social commerce
Cross-cultural study
trust in a social network community may be attributed to the closeness and familiarity developed among
Social interactions its members resulting from social interactions. The results presented herein are in line with the trust
Trust transference transference theory. The empirical study results also suggest that social commerce market penetration in
Social network site East Asia may have relatively fewer barriers to overcome and be more likely to succeed than in the Latin
Subgroup analysis America region.
ß 2013 Elsevier B.V. All rights reserved.

1. Introduction and purpose brand influence, and facilitating advice and assistance sharing
through discussion groups [38].
The popularity and growth of social commerce in recent years Specifically, social commerce involves the use of social media that
have been driven by consumer-product information sharing. supports social interaction, communications, and user-generated
This development in social commerce is due to two main content such as ratings and reviews, blogs, and microblogging to
influential factors: consumer bargaining power and rapid assist in the online buying, selling, and marketing of products and
progress in information technologies, in particular the introduc- services. Its unique characteristics are the provision of an opportunity
tion of Web 2.0. The improved bargaining power among for customers to interact to make better-informed buying decisions
consumers has caused a noticeable shift in the relationship [82] and allowing the customer to connect with, listen to, understand,
between businesses and current and potential consumers. Even and engage other people regarding their purchasing experiences to
more interesting is the fact that these consumers have a very improve their future shopping experience [9].
close relationship, as in the case of social networking sites. The A forecast made by the business consulting firm Booz &
introduction of Web 2.0 allows the creation of user-controlled Company (2012), predicted that the social commerce market will
content and Internet interactivity. be worth $30 billion in annual sales by 2015. A report from the IDC
In this study, we define social commerce as the online buying consulting firm suggested that 10 to 15 percent of consumer
and selling activities initiated via social media, which entails spending in developed countries in the next three to five years
business transactions through either social media (e.g., on a (2011–2015) may go to social networking sites such as Facebook
Facebook store) or other e-commerce sites. Kaplan and Haenlein [49]. The most powerful drivers possibly underpinning the
[30] define social media as ‘‘a group of Internet-based applications explosion of the social commerce market are the word-of-mouth
that build on the ideological and technological foundations of Web marketing and viral marketing capabilities inherent on a social
2.0, and that allow the creation and exchange of user-generated commerce site or fan store. The possible incentive-driven advocacy
content [UGC]’’ (p. 61). UGC is beneficial in providing useful provided by such online stores is also a major factor.
information for product marketing and development, increasing The rapid growth of social networks (e.g., Facebook and Twitter)
over the past decade or so may be attributed to social interaction
[28]. It is social interaction that results in word-of-mouth marketing,
* Corresponding author. Tel.: +886 3 4638800x2614; fax: +886 3 4352077. product advocacy for a consumer brand, and social capital building
E-mail addresses: [email protected], [email protected] [26]. Social commerce sites may be grouped into two types: (1) firms

0378-7206/$ – see front matter ß 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.im.2013.08.002
610 C.-S.P.P. Ng / Information & Management 50 (2013) 609–620

such as Amazon, Levis, TripAdvisor, Mazda, and Gap, which link to goals [73], and enable consumers to browse and view products as
Facebook and have Facebook fan pages and apps that take potential well as add them to a shopping cart [51]. In contrast to traditional
customers to their online-stores; and (2) companies such as brick and mortar stores, social commerce sites allow consumers to
Hallmark, Groupon USA, Pantene North America, and Watson exchange product feedback and provide information that may
Malaysia that connect to Facebook and have Facebook fan pages and influence others’ choices by expressing their own views and
apps that allow potential customers to buy directly from their experiences. This process can create word of mouth (WOM)
Facebook stores. For these consumer brands, the main purposes of marketing effects, as it provides a means for consumers to discuss
utilizing Facebook are to engage with consumers, promote special products and circulate relevant information. This can build active,
offers, and/or provide an opportunity for consumers to have social mutually beneficial consumer-to-consumer and consumer-to-
interactions with other consumers and friends. marketer communications [78].
Social networks such as the Chinese website Taobao that provide In social commerce, the main aim of the seller is to convert
a forum for consumer ratings, evaluations, and advice on products customers into brand advocates, whereas the aim of the buyer is to
have been found to influence other customers’ trust in retailers [16]. make better informed purchasing decisions [30,49]. Either or both
Business consultants [3] report that ‘‘individuals on a social of these aims are achieved when a consumer shares purchasing
networking site are usually connected to other individuals whom experiences. Some social commercial sites such as Groupon and
they trust and respect.’’ (p. 1). This phenomenon occurs because LivingSocial offer the service of group buying to encourage the
social networking environments typically involve an atmosphere of spread of deal information.
trust between friends. These studies, to some extent, indicate that One survey conducted with 1787 respondents on social media
social network users trust the people with whom they communicate showed that one third of shoppers would prefer to make a
or interact in a social network community. purchase directly from Facebook or Twitter [66]. This statistic
However, according to another marketing report by Booz & indicates a promising future for social commerce. For example,
Company in 2010, 73% of online shoppers would not purchase Facebook provides a marketplace, advertising means, and social
goods on Facebook or through other social networking sites. Some network services. Facebook’s strengths as a powerful social
academic researchers [44] have found little evidence of peer commerce site are rooted in its Social Plugins (e.g., the ‘Like,’
influence among college students using Facebook. This lack of ‘Send,’ ‘Comment,’ ‘Recommendation’ buttons) and Open Graph
evidence suggests that social network users do not equally trust or Protocol. Facebook allows a business to conveniently and
are not equally influenced by their connections on a social professionally set-up a Facebook fan store [12].
network. This may be because some friends’ opinions have a
disproportionate influence on one’s behavior [24]. 2.2. Theoretical background
In addition, these inconsistent findings may be due to the
differences in cultural and national contexts [32]. In particular, the 2.2.1. Social interactions
behavior of consumers in the West (e.g., the United States) may not A social network site provides various types of social interac-
be applied to all cultures [53]. As different countries have varying tions and networking features such as message posting, informa-
inherent values, symbols, and cultures, the behavior of social tion exchange and contact management. In addition to basic social
network users and their intention to purchase on social commerce interaction features, other forms of social interaction such as
sites may not be equivalent. Following this possibility, the purpose consumers’ numerical ratings of products (favorability) and the
of this study is to investigate the role of culture in influencing number of online reviews (visibility) have been found to correlate
peers’ online shopping intentions in social commerce conducted with sales [23] and social influence [81]. Wu and Wang [79] also
on Facebook in a cross-regional setting involving Latin American discovered that the social interaction enabled by social network
and East Asian countries – two relatively understudied regions. In sites, which allows users to cultivate, foster, and maintain online
particular, we are interested in studying (1) the moderating effect relationships, is a significant predictor of behavioral intentions.
of the cultural factor on the relationships between social These studies suggest that interactions within a social network
interaction and purchase intention in social commerce, and the community boost reliance and trust in a user’s personal relation-
effect of the cultural factor on the mediating role of trust in a social ship.
network community; and (2) the mediating effect of trust in a According to Kelman [33], there are three categories of social
social network community on the relationship between social influence: compliance, identification, and internalization. Among
interaction and purchase intention. This study attempts to extend these three, the internalization category is the most relevant to
the trust transference theory by incorporating the concepts of online social networking and commerce. Internalization is defined
social interaction, trust in a social network community, and as a situation in which a person accepts and agrees with another’s
cultural issues to better understand how these concepts influence beliefs and behavior both publicly and privately because the
purchase intentions in the context of social commerce. influential content is intrinsically rewarding [33]. Online social
The subsequent organization of this paper is as follows. Section networking and social commerce fall into this category because
2 provides a review of the literature on social commerce and the friends’ interactions, participation in online activities, and product
theoretical background of social interactions, trust transference, recommendations and reviews are written, sent, and read freely,
and culture. Section 3 proposes the research model and hypothe- and voluntarily among members of the site’s community. A user’s
ses, and Section 4 describes the research methods adopted in this selected types of interaction, participation in online activities, and
study. Section 5 presents the research findings, and section 6 offers buy-in for product recommendations are based on her own choices
the contributions and implications of this study for use in research and preferences. Therefore, the social influence of a friend’s
and practice. product recommendation may have an important role in the
product’s acceptance and recommendations among other friends.
2. Literature review One of the salient ways to measure the social influence of a
sender on a receiver is by quantifying the social impact the sender
2.1. Social commerce has on the receiver [40]. The social impact theory states that the
impact of any information source may be accounted for as a
Similar to e-commerce, social commerce sites consist of function of three factors: the number of people who make up that
communities of customers with similar interests, passions or source, the source’s immediacy or closeness with the receiver, and
 C.-S.P.P. Ng / Information & Management 50 (2013) 609–620 611

the source’s tie strength with the receiver [40]. For simplicity, in 2.2.3. National culture
this study we are only considering one source of information at a A study by [72] emphasized the importance of cultural factors
time. in influencing e-commerce adoption decisions. In studying the
Lee and Kwon [41] used the term ‘‘intimacy’’ to describe the issue of culture, reference is made to the institutional framework
closeness or immediacy factor, saying it is a ‘‘feeling of closeness established through Geert Hofstede’s dimensions of national
and emotional bonding, involving intense liking, moral support, culture. The six dimensions of national culture proposed by
and the ability to tolerate flaws in the service’’ (p. 348). Following Hofstede et al. [19] are power distance, uncertainty avoidance,
this approach, we define intimacy as the feeling of closeness and individualism, masculinity, long-term orientation, and indulgence.
emotional bonding involving intense liking and moral support While the first four dimensions constitute the original Hofstede
from social network friends, and the ability to tolerate social dimensions of national culture, the last two have been suggested
network friends’ mistakes. In this paper, we use ‘intimacy’ and by Michael Bond and Michael Minkov, respectively.
‘closeness’ interchangeably. This research focuses on only two dimensions: individualism
Marsden and Campbell [50] suggest that the best indicators and uncertainty avoidance, which are the most relevant to this
of the strength of a bond with another person are those having study. The definitions of these two dimensions [18] are as follows:
to do with time spent together and the depth of the relationship. uncertainty avoidance is the degree to which the members of a
Consistent with this idea, Brown and Reingen [7] define the culture feel threatened by uncertain or unknown situations, and
strength of ties as the familiarity of one person with another. individualism is the extent to which the ties between individuals
Familiarity is described as the ‘‘feeling of the understanding of are loose.
an entity, often based on previous interactions, experience, and As highlighted in the research [18,19], the national cultural
learning of the what, who, how, and when of what is happening’’ dimensions (with scores ranging from 1 to 120) differ and are
[41]. Based on this description, we associate the strength of ties, clustered quite obviously according to the different regions to
or bonds, with familiarity, and describe familiarity as the feeling which countries belong, such as Latin America, East Asia, Nordic
of understanding of one’s social network friends. This feeling is Europe, Southern Europe, the Middle East, and Africa. This study
often based on previous interactions, experiences, and learning focuses on the regional level, in particular the Latin America
about the what, who, whom, how, when, and why of what is region and East Asia region, because of the (1) research resources
happening. constraint and the infeasibility of collecting data from many
countries; (2) the lack of information on all cultural dimension
2.2.2. Trust transference theory scores for most Latin American countries, such as Nicaragua,
Trust is a critical component in online purchase intentions [35], Honduras and Paraguay; and (3) the proximity of cultural
as a consumer is unable to access and evaluate a real product dimension scores for some countries (see Table 1), such as
online before purchasing it. It has also been studied in the online Taiwan and Thailand scores of 69 and 64 for uncertainty
contexts of product recommendation agents [76], mobile com- avoidance and 17 and 20 for individualism, respectively. In
merce [75], government services [71], and Internet-based interor- addition, the absolute value of the CMSI difference, suggested by
ganizational systems [39]. [68], within the same region, for instance between Taiwan and
Prior studies show that trust may be transferred from one Thailand, is very small (i.e., 8), unlike the absolute value of the
source to another. For example, trust in an organization is CMSI difference between Latin America and East Asia (195–
transferable to trust in its members, trust in an individual may 148 = 47). As a result, the subsequent section builds on the
be transferred to trust in another individual, and trust may be hypotheses from the regional level.
transferred between contexts. Trust transfer occurs when ‘‘the
unknown target [is] being perceived as related to the source of the 3. Research model and hypotheses
transferred trust’’ ([67], p. 6). The perception of relatedness occurs
based on the similarity, proximity, and common fate of the entities Fig. 1 depicts the research model and the associated hypothe-
from/to which trust is transferred [8]. Similarity describes the ses. The definition of the constructs proposed in the hypotheses
entities’ degree of sameness, whereas proximity implies the degree and the references to the constructs (with their measurement
of closeness. Common fate refers to the perception that the entities items) are shown in Table 2.
(such as a social community’s members) are moving in the same
direction. In the context of members of a group, relatedness may 3.1. Social interactions
also be influenced by an individual member’s behavior [77] and the
type of interaction involved such as forming a business partnership Social interactions may breed closeness and familiarity. One
or a social network community [45]. study on continuance intention in the context of web-based
On a social commerce site, prospective consumers make buying services revealed that closeness and familiarity affect users’
decisions based on the advice of friends and family in a social continuance intention [41]. In the context of social promotion
network community that they know or trust [42]. Based on the sites, both cognitive and affective involvements, which are social
idea of relatedness, above, trust in a social network community
may be expected to be transferable to other related sources, i.e.,
Table 1
trust in other firms in the community. The prospective consumers Scores for the two cultural dimensions under examination of the two selected
could believe that the online environment is safe and well regional areas [19].
managed and that everyone there, including the sellers, is
Cultural dimension Score (1–120)
trustworthy. A user will have this belief happen when she
perceives that other firms in the community are related to her Latin America East Asia

close community, in such as way that other firms in the community Guatemala, etc. Taiwan Thailand Average
(1) are using a social network service for the same purpose as any Uncertainty avoidance (UAI) 101 69 64 66.5
other user, (2) closely fit a user’s daily life, such that they become Individualism (IDV) 6 17 20 18.5
an important part of the social network community that she CMSIa 195 152 144 148
interacts with daily, and (3) share the same future goal and a
CMSI–computer-based media support index, computed based on adding UAI to
direction as user. (100-IDV), refer to [67] for more details.
612 C.-S.P.P. Ng / Information & Management 50 (2013) 609–620

H6 communities found that the level of interpersonal trust signifi-

H1
cantly influences members’ intentions to give and receive
Closeness
information [62]. Working toward interpersonal trust may develop
(CLO) H4 one’s reliance on, protection of, and care for others. Trust also
Trust in social H3
network community
Intention to purchase represents the degree of one’s willingness to be vulnerable to
in soc. comm.
H5 (TCO) (INT) others’ actions [52]. We suggest that trust in one’s close social
Familiarity
(FAM) network community (i.e., friends) leads to trust in the whole social
H8
H2
network community, i.e., all members in the website, including
H7 sellers. In other words, if a user trusts a social network community,
H9 she is more likely to make a purchase from a social commerce site
Direct effect
recommended in that community. On the other hand, if she
Culture Moderating effect
believes that a social network community cannot be trusted, she is
Mediating effect
less likely to make a purchase from a social commerce site
recommended by that community. Following this line of argu-
Fig. 1. Research model.
ment, we expect the following:

Table 2 H3. The presence of trust in a social network community will have
Social commerce intention to purchase constructs. a positive influence on the intention to purchase.
Construct Definition Trust may be developed online [11], on social network sites. Prior
Closeness (CLO) The feeling of closeness and emotional bonding, study in the e-commerce environment reveals an association
involving intense liking and moral support between social interactions and trust [14]. In another online context,
from social network friends, and the ability to i.e., in a C2C environment, a similar finding is found. In particular,
tolerate social network friends’ mistakes.
Familiarity (FAM) The feeling of understanding between social
[10] discover that social interactions among members in a C2C
network friends, often based on previous environment increase the level of trust among its members. The
interactions, experiences, and learning of the study further suggests that trust in the community leads to trust in
what, who, whom, how, when, and why of the C2C platform, which includes sellers. We argue that one’s
what is happening.
interaction on a social network site leads to trust in one’s close
Trust in social network The sense of protection, care, and perception
community (TCO) of reliability from the messages, feedback, or community (i.e., friends), and trust in one’s close community leads to
recommendations from other friends on a trust in the entire social network community. Therefore, having
social network site. more social interactions with members in a social network
Culture The degree of national cultural variation in community, leads to greater trust in that community. This
different countries, from the perspective of
the two relevant dimensions of uncertainty
relationship holds because closeness and familiarity help users to
avoidance and individualism; stems from a feel more secure with their peers in an online community. Likewise,
respondent’s region or country of origin. social interactions in a social network community will influence
Intention to The perceptions and intentions to accept a one’s trust in the community itself, as member interactions foster a
purchase (INT) friend’s product recommendation or review
sense of belonging, involvement, and commitment. As such
when purchasing a product on a social
commerce site. reassurances evolve, users tend to believe that other community
members will not deliberately cause harm to a peer because, trust
has been established. As a result, closeness and familiarity will
increase a user’s trust in the social network community. Therefore,
interactions, have been shown to increase social networking users’
purchase intention for deals recommended by online social H4. A stronger the feeling of closeness to a social network commu-
network friends [65]. It is also evident that greater familiarity nity, has a greater influence on the level of trust in that community.
between the information receiver and sender, implies greater
intention to buy the product or service recommended by the H5. A stronger familiarity with a social network community, has a
information sender [47]. These findings suggest that the intention greater its influence on the level of trust in that community.
to purchase is directly motivated by the social interactions Prior studies also have demonstrated that trust plays a mediating
between users and a social network community. A user who has role in the relationship between familiarity and the perception of
more social interactions with others in the social network receiving useful knowledge from others [43]. That is, trust affects the
community, has a stronger feeling of closeness and familiarity perception of the usefulness of information received from a social
with others. Therefore, it is more likely that this user’s intention to network community. Likewise, a higher level of trust in a social
purchase will be influenced by the community. Based on this line of network community fosters a stronger effect on a user’s intention to
argument, we hypothesize that: purchase from a social commerce site. Still, the degree of trust in a
social network community is founded on social interactions. If trust
H1. A stronger feeling of closeness to a social network community,
exists in a social network community, information receivers will
implies a greater influence of the community on the intention to
have a stronger intention to purchase from social commerce sites
purchase.
recommended by fellow community members, with whom they are
H2. A stronger familiarity with a social network community, intimate and familiar. We propose the following:
implies a greater influence of the community on the intention
H6. Trust in a social network community plays a mediating role in
to purchase.
the relationship between closeness and the intention to purchase
from a social commerce site.
3.2. Trust in the social network community
H7. Trust in a social network community plays a mediating role in
Trust is one of the key elements in social capital [54] and is an the relationship between familiarity and the intention to purchase
asset rooted in social relationships [29]. A study of virtual from a social commerce site.
 C.-S.P.P. Ng / Information & Management 50 (2013) 609–620 613

3.3. Culture-uncertainty avoidance and individualism dimensions

4. Research methodology
Countries in Latin America have higher collectivism scores,
whereas East Asian countries have higher scores for individual- The targeted social networking site is Facebook. Facebook is
ism [19]. Together with a high score on uncertainty avoidance, chosen because it is the ‘‘largest social network in the world with 500
this difference implies that Latin American people are more million active users of whom 50 percent log in on a daily basis’’ [37].
likely to listen to the opinions of others in their society rather One global study on Facebook advertising found that companies in
than act individually based on their intrinsic needs alone. Al- many countries choose Facebook for advertising and use the site to
Gahtani et al. [1] argue that countries with low individualism promote their campaigns or brands and to communicate with their
scores might display a strong relationship between subjective customers [70]. To test the research model, we develop an
norm (the influence of people in one’s social environment) experimental social commerce site on Facebook.
and behavioral intention. In addition, people from countries The products sold on the site were accessories and clothing,
with low individualism scores (i.e., with relatively high such as bracelets and scarves. This product category ‘‘accessories
collectivism scores) such as Korea more heavily value support and clothing’’ was chosen because we observed that it was one of
from existing social network relationships than do people from the top three types of items advertised on Facebook [31]. This
high individualism score countries such as the United States experimental social commerce site was launched using a conve-
[36]. Similarly, we believe that the impact of social interactions nience sampling method; it originally had three close friends in
(from a social network site) on one’s intention to purchase total from the two different regions under examination – Latin
from a recommended social commerce site will be stronger in America and East Asia. For ease of reference, we called them the
countries with a higher collective score (Latin America region) 1st-level friends.
than in nations with a higher individual score (East Asia region). We posted pictures and descriptions of all the products on the
In other words, the strength of the relationship between social ‘‘wall’’ of the experimental Facebook account to share with the 1st-
interactions and one’s intention to purchase will be dependent level friends. The 1st-level friends then assessed the products and
on the culture, as differentiated by the individualism dimension. put their positive reviews and recommendations for each, one
Following this phenomenon, we propose the following: product at a time, on their own Facebook accounts as a message for
all their friends (identified as the 2nd-level friends) to read. Next,
H8. Culture, as differentiated by the dimension of collectivism vs. we observed the purchase intentions related to this social
individualism, will moderate the relationship between social inter- commerce site over a period of two weeks. We used the snowball
actions (for both closeness and familiarity) and intention to pur- sampling method to send an online questionnaire to all the 2nd-
chase from a social commerce site, such that the effect is stronger level friends through a Facebook message intended to collect data
in the Latin America region than the East Asia region. on social interaction, trust, respondents’ country of origin, and
Latin American countries have a culture of avoiding the unknown social commerce purchase intentions.
(a high uncertainty avoidance score), that is more pronounced than The survey instrument was developed using items validated in
in nations in the East Asia region, which have a medium-high prior research (see Table 2) and some adaptations of the
uncertainty avoidance score [19]. Prior studies have shown that technologies studied herein. The construct measurement items
countries with a high degree of uncertainty avoidance such as Japan used in the survey are provided in Appendix A1. All items are
may limit their activities using online media [67] unless they find measured on a 7-point Likert scale with anchors from ‘‘strongly
that the online activity is safe and that they are able to trust it [21]. disagree’’ to ‘‘strongly agree.’’ Because this survey was intended for
This difference implies that trust is a relatively more important respondents who speak different languages, we translated the
requirement for countries with a higher degree of uncertainty English version into two other languages, Mandarin (for respon-
avoidance than it is for countries with a lower degree of uncertainty dents from Taiwan) and Thai (for those from Thailand). The English
avoidance. version of the survey instrument was used for the Latin American
For the reasons mentioned above, Latin American people may respondents. To assess the content and translation validity of the
tend to be more careful when considering a purchase intention survey in each language, three people such as university professors
on a social commerce site that they do not trust. As such, Latin and graduate students were recruited who understand both
Americans’ intentions to purchase from a social commerce English and the target foreign language used in the version of the
site are more likely rooted in trust in a social network survey. In addition, we used back translation to ensure the quality
community. In addition, because Latin Americans are more of the translated surveys [5,6]. Back translation was performed by
likely to avoid uncertainty compared to East Asians, their trust translating the non-English versions of the survey back into the
in social communities must be developed through social original language (English in this case) to ensure that these
interactions. In short, their purchase intentions are more likely versions communicate the same meaning and content as the
driven by trust that has strengthened over a long period in a original. This procedure was performed to minimize the construct
social community rather than merely through social interactions bias, method bias, and item bias commonly found in cross-cultural
within that community. In sum, culture will affect the degree studies [46,74].
to which trust in a social community mediates the relationship A total of 632 questionnaires were sent on Facebook in January
between social interaction and one’s intention to purchase 2012 via our experimental Facebook fan page. A total of 284
from a social commerce site. Therefore, we hypothesize the responses were received; therefore, the return rate was 45 percent.
following: A total of 82 percent of the East Asian group and 78 percent of the
Latin American respondents had no social commerce purchasing
experience. The largest group of respondents came from East Asia
H9. Culture, as differentiated by the uncertainty avoidance dimen- at 62 percent, whereas the share of Latin Americans was 38
sion, will moderate the mediating role of trust in a social network percent. A total of 64 percent of the East Asian respondents came
community on the relationship between social interaction (i.e., for from Taiwan, and 42 percent of the Latin American respondents
both closeness and familiarity) and intention to purchase from a were from Guatemala. In addition, 53 percent of the respondents
social commerce site, such that the effect is significantly stronger were male. Most of the respondents were between 23 to 32 years
in Latin America region than in East Asia region. old and possessed a bachelor’s degree (46 percent – East Asia, 49
614 C.-S.P.P. Ng / Information & Management 50 (2013) 609–620

Table 3
Mean, standard deviation and the reliability tests for the constructs.

Mean Std. Dev. Factor loading SMCa Composite reliability AVE

Closeness or Intimacy (CLO) 0.856 0.749

CLO 1 4.68 1.446 0.85 0.723
CLO 2 4.28 1.552 0.88 0.774
Familiarity (FAM) 0.878 0.707
FAM 1 4.88 1.384 0.88 0.774
FAM 2 5.14 1.392 0.84 0.706
FAM 3 4.52 1.312 0.80 0.640 0.708 0.548
Trust in social network community (TCO)
TCO2 4.88 1.324 0.74 0.548 0.913 0.842
TCO3 4.15 1.571 0.74 0.548
Purchase intention (INT)
INT1 4.36 1.861 0.85 0.723
INT2 3.84 1.737 0.98 0.960
a
Squared multiple regression.

percent - Latin America). Half of the East Asian region’s model fit with normalized x2 = 59.87, df = 44, RMSEA = .036,
respondents were employed, whereas slightly more than half GFI = .97, AGFI = .94, NFI = .98, CFI = .99 and P-value = .056.
(57 percent) of the Latin American respondents were college Table 3 shows that the factor loading of all construct items
students. A more detailed summary of the respondents’ char- ranges from 0.74 to 0.98. The composite reliability of the survey
acteristics is provided in Appendix A2. items intended to measure the same construct are higher than 0.7
for all constructs (see Table 3). The average variance extracted
(AVE) for all constructs are higher than 0.50 (Table 3), and the
5. Results and analysis
square roots of the AVEs are greater in all cases than the off-
diagonal elements in their corresponding row and columns
5.1. Validity and reliability assessment
(Table 4); these facts confirm the convergent and discriminant
validity of the constructs and the measurement items used in the
The Covariance Based Structural Equation Modeling (CBSEM)
survey instrument [13]. In addition, the correlation between latent
approach is adopted in this study as it serves the purpose of this
constructs does not exceed 0.70 (Table 4), which further signals
study, i.e., theory testing and development [17]. For this purpose,
measure distinctness [58].
the Lisrel 9.1 software was utilized in this study.
This study, consisting of 284 responses, meets the minimum
5.2. The structural model
sample size requirement, also known as the rule of ten [17]. That
is, the sample size is larger than ten times the largest number of
The beta path coefficients for closeness and familiarity to
structural paths (three) directed at a particular latent construct
intention to purchase (in Fig. 2A) are positive (in the expected
in the structural model. To reduce the potential for survey
direction) and statistically significant. Therefore, hypotheses H1
evaluation apprehension and to increase the validity of survey
and H2 are supported. Similarly, the beta path coefficient for trust
responses, procedural remedies to control for common method
in the social network community to the intention to purchase (in
bias as suggested by Podsakoff et al. [59] are considered in this
Fig. 2B) is also positive and statistically significant, supporting
study. These remedies include protecting respondent anonymity,
hypothesis H3.
ensuring data confidentiality, providing concise and clear
Both familiarity and closeness in Fig. 2B are significant in
questions, and avoiding double-barreled questions. To test for
predicting the level of trust in a social network community and are
the problem of common method bias in the collected data,
able to explain 56 percent of the variance in trust in the social
Harman’s single factor test [59] is conducted. The test results
network community examined. These findings support hypotheses
confirm that the common method bias is unlikely to be a serious
H4 and H5. In overall, the entire model in Fig. 2B explains 33
problem in this study.
percent of the variance exhibited in the intention to purchase
The two-step approach suggested by [2] is applied here to test through social commerce in this sample.
the validity of the measurement model and then to assess the In testing the mediation effect of trust in a social community,
model fits of the structural models. To do so, confirmatory factor the three conditions suggested by [34] and colleagues [4,27] have
analysis (CFA) is first performed to test the unidimensionality, been examined. The results in Fig. 2A and B show that the three
convergent validity, internal consistency reliability, and the conditions for testing the mediation effect of trust in a social
discriminant validity of the measurement items used herein. This network community on the relationship between closeness (and
analysis is conducted by creating a measurement model based on familiarity) and the intention to purchase from a social commerce
the latent variables used in the research model. Prior to CFA, item site are all met. Closeness, familiarity, and trust in a social network
CLO3 is dropped to satisfy the requirement for face validity. After
CFA, each item’s unidimensionality is assessed by evaluating its
shared residual variance and checking the modification index for Table 4
theta-epsilon with other items. The pairs of items with standard- Discriminant validity (intercorrelations) of variable constructs.

ized residuals far above 2.58 and modification indexes (for 1 2 3 4

theta-delta and theta-epsilon) above 5, thus threatening the 1. Closeness 0.89
unidimensionality, are dropped one at a time until there are no 2. Familiarity 0.57 0.84
problematic measurement items [15]. Following these rules, TCO1, 3. Purchase intention 0.40 0.37 0.74
TCO4 and TCO5 are dropped to prevent the violation of unidimen- 4. Trust in a social network community 0.62 0.67 0.55 0.92
sionality. Following these procedures, the CFA achieved satisfactory Note: Diagonal is the square-root of the construct’s AVE.
 C.-S.P.P. Ng / Information & Management 50 (2013) 609–620 615

(2A) Without ‘trust in social network community’ (2B) With ‘trust in social network community’
0.30*** 0.17 n.s.

Closeness Closeness 0.44

Intention to ** * Intention to
purchase in soc. Trust in social
purchase in soc.
comm. network community 0.48*** comm.
(R2=.21) *** (R2=.56)
0.39 (R2=.33)
Familiarity
Familiarity

0.14* -0.04 n.s.

Note:
2A: 2 =30.74, df=11, RMSEA=.079, GFI=.973, AGFI=.931, 2B: 2 =53.93, df=21, RMSEA=.074, GFI=.973, AGFI=.931, NFI=.978
NFI=.978 and CFI=.986 and CFI=.986
* Significant at p < 0.05 , ** Significant at p < 0.01, ***Significant at p < 0.001, n.s.- not significant

Fig. 2. The initial structural model results for the whole sample.

Table 5 measurement parameters are the same across all groups and
Sobel significance test results for indirect effects.
whether the same construct is measured in all groups. A Chi-
Path coefficient Standard Error Z-values square difference test was performed for this purpose. The results
Closeness for the measurement model are shown in Appendix A3. The large
Direct effect (c1) 0.171 0.0992 1.721 P-value suggests that there is sufficient evidence to accept the null
Indirect effect (Sobel Test) hypothesis that the factor loadings of the measurement model are
CLO- > TCO (a1) 0.437 0.0901 4.853 invariant across the two cultural groups.
TCO- > INT (b1) 0.475 0.119 3.999
Fig. 3 illustrates the structural model results for both the East
Sobel Z-value 3.082 (significant at P < 0.01)
Familiarity Asia and Latin America regions. In comparing the model results
Direct effect (c2) 0.0422 0.0875 -0.482 between these two groups, it was found that only the relationship
Indirect effect (Sobel Test) between closeness and the intention to purchase was significant
FAM->TCO (a2) 0.387 0.0858 4.514
within the East Asia group, whereas only familiarity was significant
TCO->INT (b2) 0.475 0.119 3.999
Sobel Z-value 2.963 (significant at P < 0.01)
in predicting the intention to purchase within the Latin America
group. A Chi-square difference test was conducted to test whether
these two groups have equal parameter estimates in the path
model. The results in Table 6 show that the Chi-square difference
community (the mediator) significantly predict the intention to was Dx2 = 13.41, df = 6, P = .037. This finding indicates that
purchase; see Fig. 2A and B. However, the direct effects of closeness differences in parameter estimates exist between the two model
and familiarity on the intention to purchase are no longer analyses. However, H8 was supported only for the relationship
significant after controlling for the mediator. The results of the between familiarity and intention to purchase but not for the
Sobel significance test, referring to [4], for the indirect effects of relationship between closeness and intention to purchase because
the two independent variables on the dependent variable via the this relationship was insignificant within the Latin America group.
mediator are all significant at P < 0.01, as shown in Table 5. In The moderated mediation tested in this study was conducted by
addition, complete mediation occurs in the direct effects of examining the mediation effects and then the moderating effect.
closeness and familiarity on the intention to purchase after Similarly, to test the mediation effect of trust in a social community
controlling for the mediator. Therefore, H6 and H7 are supported. on the two different regions, the three required conditions were
Similar to the study by [67], cultural factors were neither met as shown in Fig. 4A and B. The mediation test of trust in a social
collected nor measured in our survey instrument. Therefore, the network community, using the Sobel significance test, on the
interaction terms between culture and other independent vari- relationships (1) between closeness and the intention to purchase
ables could not be computed and included in the measurement is significant at P < 0.05, with a Z-value of 2.342, for the East Asia
model. In this situation, subgroup analysis is appropriate for group, and (2) the relationship between familiarity and the
testing the moderation strength and is analyzed using Lisrel [4]. intention to purchase is significant at P < 0.05, with a Z-value of
To justify subgroup analysis and analysis of moderating effects 2.077, for the Latin America group. Partial mediation was observed
in different groups, testing the measurement model invariance is as the predictive utility of closeness was reduced from beta = 0.41,
required [63]. This procedure is followed to determine whether the P < 0.001 to beta = 0.28, P < 0.01, and there was a drop in the

(3A) East Asia (3B) Latin America

0.41*** -0.10 n.s.

Closeness Closeness
Intention to Intention to
purchase in soc. purchase in soc.
comm. comm.
(R2=.23) (R2=.10)
Familiarity Familiarity

-0.10 n.s. 0.34**

Note:
3A: 2 =45.41, df=11, RMSEA=.133, GFI=.942, AGFI=.852, 3B: 2 =18.36, df=11, RMSEA=.079, GFI=.954, AGFI=.883,
NFI=.959 and CFI=.968 NFI=.949 and CFI=.978
* Significant at p < 0.05 , ** Significant at p < 0.01, ***Significant at p < 0.001, n.s.- not significant

Fig. 3. Structural model results of social interactions for the two regions.
616 C.-S.P.P. Ng / Information & Management 50 (2013) 609–620

Table 6 Table 7
Parameter estimates for the two groups – Chi-square difference test results. Mediation test of TCO for East Asia and Latin America.

Hypothesis Chi-square DF P-value East Asia (N = 176) Latin America (N = 108)

Equal parameter estimates (H0) 148.89 39 0.0001 Base Constraint model Base Constraint model
Unequal parameter estimates (H1) 135.48 33 0.0001
CLO->TCO 0.417*** 0.466*** 0.206 0.205
Difference 13.41 6 0.0370
FAM->TCO 0.350*** 0.302** 0.634*** 0.635***
CLO->INT 0.284** (Constrained) NA NA
FAM->INT NA NA 0.006 (Constrained)
significance of the direct effects of closeness on the intention to TCO->INT 0.313** 0.603*** 0.464* 0.470***
x2 66.26 73.94 37.81 37.81
purchase after controlling for the mediator. To assess the statistical
DF 22 23 22 23
conclusion validity of the type of mediation, the significance of CFI 0.973 0.968 0.972 0.974
indirect effects is further examined [48]. This assessment was RMSEA 0.107 0.112 0.082 0.077
accomplished by comparing the constraint model (where the Dx2 – 7.68 – 0
coefficient of the direct effect path is constrained to ‘zero’) with the P-value – 0.0056a – 1b

base model (which includes all the direct and indirect paths). The NA – not applicable for testing the mediating role of TCO as the paths (FAM->TCO in
addition of this constraint (CLO->TCO set to ‘zero’) in the East Asia the East Asia and CLO->TCO in the Latin America) are not significant as indicated in
Fig. 3A and B respectively.
group significantly deteriorates the model (Dx2 = 7.68, df = 1,
(Constrained) – the path is set to ‘zero’.
P = 0.0056) compared to the base model, see Table 7. This finding a
Reject H0: the path coefficient of CLO- > INT in the base model and the
indicates that the indirect effects are significant and that the direct constraint model is the same.
b
effect remained significant even in the presence of the mediator. Accept H0: the path coefficient of FAM- > INT in the base model and the
This supports the partial mediation model. constraint model is the same.
*
Significant at P < 0.05
In contrast, the mediation test effect of trust in a social **
Significant at P < 0.01.
community between familiarity and the intention to purchase, ***
Significant at P < 0.001.
shown in Fig. 4B, indicates that complete mediation occurred. This
finding exists because the previously significant relationship
between familiarity and the intention to purchase becomes Table 8
Mediation test–Between the two group comparisons.
insignificant after controlling for trust in a social network
community. The model comparison between the constraint model Modela Constraint x2 DF Dx2 P-value RMSEA
and base model in the Latin America group showed that the Default (No structural 64.82 50 0.046
addition of this constraint (FAM->TCO sets to ‘zero’) did not constraint)
significantly deteriorate the model (Dx2 = 0, df = 1, P = 1) com- Model A CLO->INT 99.08 51 34.26 0.0001 0.084
pared to the base model as shown in Table 7. This finding proves Model B FAM->INT 112.64 51 47.82 0.0001 0.093
a
that adding the FAM->TCO path does not provide any additional Measurement model was constrained to be equal across the two cultural groups
information. Therefore, trust in a social network community fully (East Asia and Latin America) in all the models
mediates the relationship between familiarity and the intention to
purchase in the Latin America group.
To compare these results across subgroups in a more rigorous imposed [64]. On the other hand, the constraint model has an
way, between-group structural equation model analyses; see [60], additional equality constraint of setting the path coefficient (CLO-
were conducted in Table 8. These analyses were accomplished by >INT or FAM->INT) to be equal across the two groups in models A
comparing the constraint model with the default model. In the and B, respectively. The Dx2 values in models A and B in Table 8 were
default model (as presented in Fig. 2B), all the structural parameters significant (model A: Dx2 = 34.26, df = 1, P = .0001; model B:
were allowed to vary freely across the East Asia and Latin America Dx2 = 47.82, df = 1, P = .0001). These showed deterioration in model
groups, but the measurement model equivalence constraints were fit, indicating that the paths (CLO->INT in model A and FAM->INT in

Fig. 4. Structural model results of the mediating effect of TCO in the two regions.
 C.-S.P.P. Ng / Information & Management 50 (2013) 609–620 617

Table 9
Hypotheses conclusions.

Hypotheses conclusions Result supported

H1: A stronger feeling of closeness to a social network community increases the Yes
community’s influence on the intention to purchase.
H2: A stronger familiarity with a social network community increases the Yes
community’s influence on the intention to purchase
H3: Presence of trust in a social network community has a positive influence Yes
on the intention to purchase.
H4: A stronger feeling of closeness to a social network community has a stronger Yes
influence on the level of trust in that community.
H5: A stronger familiarity with a social network community has a stronger Yes
influence on the level of trust in that community.
H6: Trust in a social network community plays a mediating role in the Yes – complete mediation
relationship between closeness and the intention to purchase from a
social commerce site.
H7: Trust in a social network community plays a mediating role in the Yes – complete mediation
relationship between familiarity and the intention to purchase from a
social commerce site.
H8: Culture, as differentiated by the dimension of collectivism vs. individualism, Yes – moderation occurs
moderates the relationship between social interactions (for both closeness and (But, only for FAM->INT, the effect is found to be
familiarity) and the intention to purchase from a social commerce site such stronger in the LA region)
that the effect is stronger in the Latin America region than the East Asia region
H9: Culture, as differentiated by the uncertainty avoidance dimension, will Inconclusive – as no comparable relationship can be
moderate the mediating role of trust in a social network community on the made between the two regions from our sample data.
relationship between social interaction (i.e., for both closeness and (However, complete mediation of TCO happens within
familiarity) and the intention to purchase from a social commerce site the LA region for FAM->INT, and partial mediation
such that the effect is significantly stronger in the Latin America region occurs within the EA region for CLO->INT.)
than in the East Asia region.

model B) were not equal in the two regions [64]. These results social commerce site – i.e., a partial mediation of closeness’s direct
indicate that the strengths of the mediator role of trust in a social effect in the East Asia region and a complete mediation of
network community (TCO) in these two relationships (CLO->INT familiarity’s direct effect in the Latin America region. This finding
and FAM->INT) vary significantly across the two regional settings. In implies that the mediating effect of trust in a social network
this case, the effect of the mediator is stronger in the Latin America community is relatively stronger within the LA region.
group as complete mediation occurs, whereas partial mediation
occurs in the East Asia group. 6.2. Theoretical contribution

6. Conclusions and implications This study shows that trust transference theory helps in to
explain social commerce purchase intention. It is clear from the
Table 9 presents the outcomes of the hypotheses tested in this results that one third of the variations in the intention to purchase
study. from a social commerce site observed in our sample are based on
trust in a social network community. This finding shows that trust in
6.1. Conclusions other firms in a social network community may be transferred from
trust in one’s close social network community. This finding is
The results from the sample in this study suggest that closeness consistent with that of [10]; whose study shows that mutual trust
and familiarity (social interactions) are good predictors of trust in a between members boosts their trust in the C2C platform provider.
social network community. A total of 56% of the variation in trust in This finding highlights the importance of building up trust in a social
a social network community may be explained by a user’s network community to boost users’ intentions to purchase from a
closeness and familiarity with a social network community social commerce site. In addition, this study has extended the trust
(Fig. 2B). Furthermore, based on our whole sample, trust in a transference theory by suggesting that culture somehow plays a
social network community is a salient factor influencing the moderating role in the transfer of trust in one’s close social network
intention to purchase on social commerce websites. community to trust other firms in that social network community.
In comparison to closeness, familiarity is a weaker measure of This study has also contributed to research on purchase
social impact (see Fig. 2A). Our conclusion is consistent with that of intentions via social commerce in the following ways: (1) confirming
[41], who also argued that closeness/intimacy is a purer affective that culture plays a moderating role in predicting purchase
concept and has a greater impact on online users’ behavior or intentions through the social commerce context; (2) showing that
compared to familiarity. both the closeness and familiarity of members on a social network
In addition, this study verifies that trust in a social networking site are important in building trust in a social network community,
community plays a mediating role in the relationship between social and (3) showing that trust in a social network community mediates
interactions and the intention to purchase from a social commerce the relationship between social interactions and the intention to
site. Whereas closeness was significant in the East Asia group, purchase, and this mediation is conditional on culture.
familiarity was significant in the Latin America group in predicting
users’ purchase intention on social commerce sites. Our results also 6.3. Implications for practice
provide evidence that (1) the differences in culture between
regions moderate the relationship between social interaction and Social media brings businesses closer to their customers. A
the intention to purchase from a social commerce site and (2) the study conducted by [16] suggests that ‘‘social networks are the
differences in culture between regions moderate the mediating role most prominent feature in predicting how consumers choose their
of trust in a social network community differently in the relationship transaction partners.’’ A better understanding of the characteristics
between social interactions and the intention to purchase from a of social commerce users and the impacts of social interactions on
618 C.-S.P.P. Ng / Information & Management 50 (2013) 609–620

social commerce purchasing decisions, may help online companies 6.4. Implications for future research directions and limitations
to better plan and target their marketing strategies at a predefined
segment of customers on a predetermined social network site. The use of Holfstede’s cultural dimensions in studying cross-
According to a global study [69], if social interactions (which regional cultural issues may face some criticisms [25]. Wu [80]
may create closeness and familiarity) on a social networking site emphasizes that values in a specific culture change and evolve over
and trust in the community are properly managed, a social time, as they may be influenced by political changes, social factors
networking site can also serve as a venue for (1) retaining and and increasingly integrated global economies. As a result,
increasing an online company’s customer loyalty (as such sites purchasers’ behavior will change over time as well. Nevertheless,
facilitate a direct and open point of contact with customers), (2) this study may serve as a foundation for comparative studies in
obtaining more customer feedback and electronic word-of-mouth future research, for example, by comparing the relative relevance
advertisings, and (3) managing enterprise knowledge. Our study and robustness of Holfstede’s cultural dimensions with the
has proven that social interactions with a social network recently developed GLOBE approach [20] for measuring cultural
community may result in a stronger sense of trust in the values. Moreover, future studies may repeat this study’s research
community and may drive the intention to purchase from a social model under different cultural settings such as examining Western
commerce site in that community. countries versus Middle Eastern countries or under different
Online and offline companies must be open to and aware of national economic environments such as developed versus
external changes that are occurring worldwide with regard to developing countries.
managing brand awareness and customer relationship management In addition, this study considers social capital factors only.
to remain competitive [56]. Companies may incorporate and apply Consequently, other factors related to social commerce systems
the right social media strategy by modularizing this innovative such as system quality, information quality and user-friendly
business strategy in their organizational routines [55] to shape interfaces (cf., [61]), after-sales service, and operational environ-
organizational rhetorical practices [22] or by adopting a project ments (e.g., reliability and availability services provided by the
evaluation model suggested in the research [57] to manage their postal system in a country) may merit further investigation. In
organizational commitments for social media strategy. addition, other future studies may consider or control payment
Similar to other popular online companies such as Google and behavior, examine macroeconomic differences between different
Yahoo, social networking sites such as Facebook that are free of world regions (e.g., Internet infrastructure and legal system), and
charge allow companies to attract a high volume of users and study differences at the individual level, such as an individual’s
visitors. Setting up a store to sell products on a social networking purchasing power and habits, Internet usage behavior, prior
site may be accomplished at almost no cost, provides unlimited purchasing experience, and interest in a product as well as a
access to a large potential base of customers, and may increase an product’s price. In addition, trust in a social network community
online company’s brand equity. However, a firm understanding of may change over time; therefore, a longitudinal study may be
the importance of social interactions and of the characteristics of a beneficial for a deeper understanding of the dynamics of trust
particular customer segment is required to guarantee success in transfers occurring in a social commerce environment.
social commerce. For instance, our study has shown that both
closeness and trust in a social network community are significant Acknowledgments
predictors of the intention to purchase on social commerce sites in
the East Asia region (Fig. 4A) but that only trust in a social network This research was supported by the National Science Council of
community is a significant predictor for the Latin America region Taiwan, grant NSC 101-2410-H-155-005 and NSC 99-2410-H-155-
(Fig. 4B). These findings suggest that, if social commerce requires 031. The author is indebted to the guest editor, Yunjie (Calvin) Xu, and
social interactions among the members in its community, social reviewers for their constructive comments and suggestions, which
commerce market entry or penetration in East Asian countries may have greatly improved the quality of this paper. Finally, the author
have relatively fewer barriers and has a greater potential for would like to thank Danny, JJ, Alex and Anurag for the assistance in
success than in Latin American countries. translating and/or sending out the survey to the respondents.

Appendix A1. Predictor and predicted latent construct items

Survey items Dropped item

Closeness or Intimacy (CLO)

CLO 1: I feel a sense of closeness with my friends on Facebook.
CLO 2: I feel a sense of intimacy with my friends on Facebook.
CLO 3: I feel my friend’s product recommendations or product reviews are a very important part of my consumption life. Yes – to satisfy face validity
Familiarity or Tie strength (FAM)
FAM 1: I am familiar with my friends on Facebook through message exchanges.
FAM 2: I am familiar with my friends on Facebook through photo sharing.
FAM 3: I have a very high level of interaction with each person on Facebook.
Trust in social network community (TCO)
TCO1: I feel fine interacting with the social network community (e.g., friends, and relatives) because it fulfills my needs Yes – dropped in CFA
of interaction efficiently.
TCO2: I always feel confident that I can rely on the social network community’s (e.g., friends, and relatives) responses
and feedback when I interact with them.
TCO3: I assume my Facebook friends would always look out for my interests.
TCO4: I assume my Facebook friends would make sure that I was not harmed or in danger. Yes – dropped in CFA
TCO5: I feel like my Facebook friends care what happens to me. Yes – dropped in CFA
Purchase intention (INT)
INT1: I have a strong intention to purchase a product online if the product is recommended by my social network friends.
INT2: I choose to accept my friend’s product recommendations or product reviews without any hesitation when
purchasing goods on the Internet.
EXP: Have you bought a product from any companies that promote themselves on Facebook? Yes – dropped in CFA
 C.-S.P.P. Ng / Information & Management 50 (2013) 609–620 619

Appendix A2. Summary of respondent characteristics (n = 284)

Frequency Percent

Country region
Latin American countries Guatemala 46 16
(n = 108) Nicaragua 18 6
Honduras 13 5
Other 31 11
East Asian countries Taiwan 113 46
(n = 176) Thailand 59 24
Other 4 1
Interaction level return rate Low 134/402 (sent), 134/246 (returned) 33, 54
Average 68/132, 68/246 52, 28
High 44/98, 44/246 45, 18
Social Commerce Experience 49/246 20
East Asia 32/176 18
Latin America 24/108 22
Age (for both mode and median) East Asia 23–32 65
Latin America 23–32 75
Gender Male 130/246 53
Female 116/246 47
Education level (mode) East Asia Bachelor Degree 46
Latin America Bachelor Degree 49
Employment (mode) East Asia Employed 50
Latin America College Student 57

Appendix A3. Invariance of factor loadings–Chi-square difference test results

Hypothesis Chi-square DF P-value

Equal factor loadings (H0) 143.82 95 0.0009

Unequal factor loading (H1) 137.88 88 0.0005
Difference 5.94 7 0.5468

References [17] J.F. Hair, C.M. Ringle, M. Sarstedt, PLS-SEM: indeed a silver bullet, Journal of
Marketing Theory and Practice 19 (2), 2011, pp. 139–151.
[1] S.S. Al-Gahtani, G.S. Hubona, J. Wang, Information technology (IT) in Saudi Arabia: [18] Hofstede, G., Hofstede, G. J., 2011. Dimensions of national cultures. http://
culture and the acceptance and use of IT, Information & Management 44 (8), 2007, www.geerthofstede.nl/culture/dimensions-of-national-cultures.aspx.
pp. 681–691. [19] G. Hofstede, G.J. Hofstede, M. Minkov, Cultures and Organizations: Software of the
[2] J.C. Anderson, D.W. Gerbing, Structural equation modeling in practice: a review Mind, McGraw-Hill USA, New York, 2010.
and recommended two-step approach, Psychological Bulletin 103 (3), 1988, pp. [20] R.J. House, P.J. Hanges, M. Javidan, P.W. Dorfman, G. Vipin, Culture, Leadership,
411–423. and Organizations: The GLOBE Study of 62 Societies, Sage, Thousand Oaks,
[3] M. Anderson, J. Sims, J. Price, J. Brusa, Turning Like to Buy Social Media Emerges as 2004 .
a Commerce Channel, Booz & Company Inc., Chicago, 2011 pp. 1–12. [21] M.-H. Hsieh, S.-L. Pan, R. Setiono, Product-, corporate-, and country-image dimen-
[4] R.M. Baron, D.A. Kenny, The moderator-mediator variable distinction in social sions and purchase behavior: a multicountry analysis, Journal of the Academy of
psychological research: conceptual, strategic, and statistical considerations, Jour- Marketing Science 32 (3), 2004, pp. 251–270.
nal of Personality and Social Psychology 51 (6), 1986, pp. 1173–1182. [22] J. Huang, J. Baptistaa, R.D. Galliers, Reconceptualizing rhetorical practices in
[5] R.W. Brislin, Back-translation for cross-cultural research, Journal of Cross-Cultural organizations: the impact of social media on internal communications, Informa-
Psychology 1 (3), 1970, pp. 185–216. tion & Management 50 (2-3), 2013, pp. 112–124.
[6] R.W. Brislin, Cross cultural research in psychology, Annual Review of Psychology [23] J. Huang, W.F. Boh, K.H. Goh, From a social influence perspective: the impact of
34 (February), 1983, p. 363L 400. social media on movie sales, The 15th Pacific Asia Conference on Information
[7] J.J. Brown, P.H. Reinegn, Social ties and word-of-mouth referral behavior, Research systems (PACIS), Brisbane, Australia, Queensland University of Technology, 2011,
in Consumer Behavior 14 (3), 1987, pp. 350–362. pp. 1–10.
[8] D.T. Campbell, Common fate, similarity, and other indices of the status of [24] R. Iyengar, C. Van den Bulte, T.W. Valente, Opinion leadership and social contagion
aggregates of persons as social entities, Behavioral Science 3 (1), 1958, pp. 14–25. in new product diffusion, Marketing Science 30 (2), 2011, pp. 195–212.
[9] Cecere, L., 2010. Pioneers on the Move: Rise of Social Commerce. Altimeter Group. [25] M. Javidan, R.J. House, P.W. Dorfman, P.J. Hanges, M.S. de Luque, Conceptualizing
http://www.altimetergroup.com/2010/11/pioneers-on-the-move-rise-of-social- and measuring cultures and their consequences: a comparative review of GLOBE’s
commerce.html. and Hofstede’s approaches, Journal of International Business Studies 37 (6), 2006,
[10] J. Chen, C. Zhang, Y. Xu, Understanding the role of mutual trust in building up pp. 897–914.
customer loyalty to the C2C platform provider, International Journal of Electronic [26] C. Jin, The perspective of a revised TRAM on social capital building: the case of
Commerce 14 (1), 2009, pp. 147–171. facebook usage, Information & Management 50 (4), 2013, pp. 162–168.
[11] C. Dwyer, S.R. Hiltz, Trust and Privacy Concern within Social Networking Sites: A [27] C.M. Judd, D.A. Kenny, Process analysis: estimating mediation in treatment
Comparison of Facebook and MySpace, The Thirteenth Americas Conference on evaluations, Evaluation Review 1981, pp. 5602–5619.
Information Systems, Colorado, August, 09–12, 2007, CD-ROM. [28] G. Jung, B. Lee, How did facebook outspace myspace with open innovation? An
[12] Facebook, 2011. Core Concepts. Facebook developer. http://developers.facebook.- analysis of network competition with changes of network topology The 15th
com/docs/coreconcepts/. Pacific Asia Conference on Information systems (PACIS), Brisbane, Australia,
[13] C. Fornell, D.F. Larcker, Evaluating structural equation models with unobservable Queensland University of Technology, 2011, pp. 1–14.
variables and measurement error: algebra and statistics, Journal of Marketing [29] A. Kankanhalli, B.C.Y. Tan, K.-K. Wei, Contributing knowledge to electronic
Research 18 (3), 1981, pp. 382–388. knowledge repositories: an empirical investigation, MIS Quarterly 29 (1),
[14] D. Gefen, E-commerce: the role of familiarity and trust, Omega: The International 2005, pp. 113–143.
Journal of Management Science 28 (6), 2000, pp. 725–737. [30] A.M. Kaplan, M. Haenlein, Users of the world, unite! the challenges and oppor-
[15] D. Gefen, Assessing unidimensionality through LISREL: an explanation and ex- tunities of social media Business Horizons 53 (1), 2010, pp. 59–68.
ample, Communications of the Association for Information Systems 12 (1), 2003, [31] Keath, J., 2012. 105 Facebook Advertising Case Studies. http://socialfresh.com/
pp. 23–47. facebook-advertising-examples/.
[16] S. Guo, M. Wang, J. Leskovec, The role of social networks in online shopping: [32] M. Keil, B.C.Y. Tan, K.K. Wei, T. Saarinen, V.K. Tuunainen, A. Wassenaar, A cross-
information passing, price of trust, and consumer choice, ACM Conference on cultural study on escalation of commitment behavior in software projects, MIS
Electronic Commerce 15, 2011, pp. 7–166. Quarterly 24 (2), 2000, 299-226.
620 C.-S.P.P. Ng / Information & Management 50 (2013) 609–620

[33] H.C. Kelman, Compliance, identification, and internalization: three processes of [61] F. Rahimniaa, J.F. Hassanzadehb, The impact of website content dimension and
attitude change, Journal of Conflict Resolution 2, 1958, pp. 51–60, March. e-trust on e-marketing effectiveness: the case of iranian commercial saffron
[34] Kenny, D.A., 2011. Mediation. http://davidakenny.net/cm/mediate.htm. corporations, Information & Management 50 (5), 2013, pp. 240–247.
[35] H.-W. Kim, Y.C. Xu, S. Gupta, Which is more important in internet shopping, [62] C.M. Ridings, D. Gefen, B. Arinze, Some antecedents and effects of trust in
perceived price or trust? Electronic Commerce Research and Applications 11 (3), virtual communities, Journal of Strategic Information Systems 11 (3–4), 2003,
2012, pp. 241–252. pp. 271–295.
[36] Y. Kim, D. Sohn, S.M. Choi, Cultural difference in motivations for using social [63] M. Sarstedt, C.M. Ringle, Treating unobserved heterogeneity in pls path model-
network sites: a comparative study of American and Korean college students, ling: a comparison of FIMIX-PLS with different data analysis strategies, Journal of
Computers in Human Behavior 27 (1), 2011, pp. 365–372. Applied Statistics 37 (8), 2010, pp. 1299–1318.
[37] KISSmetrics, 2012. Facbook Statistics. http://blog.kissmetrics.com/facebook-sta- [64] R.E. Schumacker, R.G. Lomax, A Beginner’s Guide to Structural Equation Modeling,
tistics/. Routledge, New York, 2010.
[38] D.M. Kroenke, Using MIS, Pearson Prentice Hall, New Jersey, 2011. [65] J.-K. Shin, M.-S. Park, Y. Ju, The effect of the online social network structure
[39] I.K.W. Lai, V.W.L. Tong, D.C.F. Lai, Trust factors influencing the adoption of characteristics on network involvement and consumer purchasing intention:
internet-based interorganizational systems, Electronic Commerce Research and focus on Korean social promotion sites, The 11th International DSI and the
Applications 10 (1), 2011, pp. 85–93. 16th PDSI Joint Meeting, Taipei, Taiwan, 2011.
[40] B. Latané, The psychology of social impact, American Psychologist 36 (4), 1981, [66] shop.org, New Study Evaluates Consumer Behaviors, Attitudes towardSocial
pp. 343–356. Commerce, The Shop.org Community, 2011 http://www.shop.org/press/
[41] Y. Lee, O. Kwon, Intimacy, familiarity and continuance intention: an extended 20110527.
expectation–confirmation model in web-based services, Electronic Commerce [67] K.J. Stewart, Trust transfer on the world wide web, Organization Science 14 (1),
Research and Applications 2011, pp. 10342–10357. 2003, pp. 5–17.
[42] LeeRaito.com, 2007. Social Commerce Chapter. LeeRaito.com. http://leeraito.com/ [68] D. Straub, M. Keil, W. Brenner, Testing the technology acceptance model
social-commerce-chapter/. across cultures: a three country study, Information & Management 33 (1),
[43] D.Z. Levin, R. Cross, The strength of weak ties you can trust: the mediating role of 1997, pp. 1–11.
trust in effective knowledge transfer, Management Science 50 (11), 2004, pp. [69] Syzygy Group, F-Commerce Beyond the F-Store: 20 Notable Examples, Social
1477–1490. Commerce Today, 2012 http://socialcommercetoday.com/f-commerce-beyond-
[44] K. Lewis, M. Gonzalez, J. Kaufman, Social selection and peer influence in an online the-f-store-20-notable-examples/.
social network, Proceedings of the National Academy of Sciences of the United [70] TBG Digital, 2012. Global Facebook Advertising Report 2012 Q2. http://clearsli-
States of America 2012, pp. 68–72. de.com/view/mail?iID=A8GU559Z78B9DX9V4Q8A.
[45] B. Lickel, D.L. Hamilton, G. Wieczorkowska, A. Lewis, S.J. Sherman, A.N. Uhles, [71] T.S.H. Teo, S.C. Srivastava, L. Jiang, Trust and electronic government success: an
Varieties of groups and the perception of group entitativity, Journal of Personality empirical study, Journal of Management Information Systems 25 (3), 2009, pp.
and Social Psychology 78 (2), 2000, pp. 223–246. 99–131.
[46] Y.-H. Lin, C.-Y. Chen, P.-K. Chiu, An overview on issues of cross-cultural research [72] S.M.B. Thatcher, W. Foster, L. Zhu, B2B E-commerce adoption decisions in Taiwan:
and back-translation, The Sport Journal 8 (4.), 2005. the interaction of cultural and other institutional factors, Electronic Commerce
[47] Y. Liu, G. Liu, Study of the social network’s impact on College Students Purchasing Research and Applications 5 (2), 2006, pp. 92–104.
Behavior, The 7th International Conference on Wireless Communications, Net- [73] TribalCafe, Social Commerce—e-Commerce becomes Social, TribalCafe, 2011
working and Mobile Computing (WiCOM), 23–25 September, 2011, pp. 1–4. http://www.tribalcafe.co.uk/2011/01/social-commerce/.
[48] D.P. Mackinnon, C.M. Lockwood, J. Williams, Confidence limits for the indirect [74] F. Van de Vijver, R.K. Hambleton, Translating tests: some practical guidelines,
effect: distribution of the product and resampling methods, Multivariate Behav- European Psychologist 1 (2), 1996, pp. 89–99.
ioral Research 39 (1), 2004, pp. 99–128. [75] A. Vance, C. Elie-Dit-Cosaque, D.W. Straub, Examining trust in technology arti-
[49] P. Marsden, F-commerce selling on facebook: the opportunity for consumer facts: the effects of system quality and culture, Journal of Management Informa-
brands, SYZYGY Group 2011, pp. 1–33. tion Systems 24 (4), 2008, pp. 73–100.
[50] P.V. Marsden, K.E. Campbell, Measuring tie strength, Social Forces 63 (2), 1984, pp. [76] W. Wang, I. Benbasat, Trust in and adoption of online recommendation agents,
482–501. Journal of the Association for Information Systems 6 (3), 2005, pp. 72–101.
[51] Matsumoto, F., 2009. Social Commerce: Strategies for Extending Online Shopping [77] D. Wilder, A.F. Simon, in: C. Sedikides, J. Schopler, C.A. Insko (Eds.), Categorical and
beyond the e-Commerce Site. Direct Marketing News. http://www.dmnews.com/ Dynamic Groups: Implications for Social Perception and Intergroup Behavior.
social-commerce-strategies-for-extending-online-shopping-beyond-the-e-com- Intergroup Cognition and Intergroup Behavior, Lawrence Erlbaum Associates,
merce-site/article/157433/. New Jessey, Mahwah, 1998, pp. 27–44.
[52] D.H. McKnight, L.L. Cummings, N.L. Chervany, Initial trust formation in new [78] WOMMA, An Introduction to WOM Marketing with Definitions, WOMMA, 2010
organizational relationships, Academy of Management Review 1998, pp. http://womma.org/wom101/.
23473–23490. [79] H.-L. Wu, J.-W. Wang, An empirical study of flow experiences in social
[53] J. Moon, D. Chadee, S. Tikoo, Culture, product type, and price influences on network sites, The 15th Pacific Asia Conference on Information systems
consumer purchase intention to buy personalized products online, Journal of (PACIS), Brisbane, Australia, Queensland University of Technology, 2011,
Business Research 61 (1), 2008, pp. 31–39. pp. 1–11.
[54] J. Nahapiet, S. Ghoshal, Social capital, intellectual capital, and organizational [80] M.-Y. Wu, Hofstede’s cultural dimensions 30 years later: a study of Taiwan and
advantage, Academy of Management Review 23 (2), 1998, pp. 242–266. the United States, Intercultural Communication Studies 2006, pp. XV33–42.
[55] S.L. Pan, G. Pan, J.W. Chen, M.-H. Hsieh, The dynamics of implementing and [81] Y.C. Xu, C. Zhang, L. Xu, Measuring product susceptibility in online product review
managing modularity of organizational routines during capability development: social network, Decision Support Systems 2013, http://dx.doi.org/10.1016/
insights from a process model, IEEE Transactions of Engineering Management 54 j.dss.2013.01.009.
(4), 2007, pp. 800–813. [82] Zujewski, B., 2010. Defining Social Commerce First Post in a Social Commerce
[56] S.L. Pan, G. Pan, M.-H. Hsieh, A dual-level analysis of capability development Series. ATG. https://community.atg.com/community/business/blog/2010/02/11/
process: a case study of TT&T, Journal of the American Society for Information defining-social-commerce-first-post-in-a-social-commerce-series.
Science and Technology 57 (13), 2006, pp. 1814–1829.
[57] S.L. Pan, G.S.C. Pan, M. Newman, D. Flynn, Escalation and de-escalation of Celeste S-P Ng is currently an Assistant Professor of the
commitment to information technology projects: insights from an approach- Department of Information Management at Yuan Ze
avoidance punctuated equilibrium model, European Journal of Operational Re- University. Dr. Ng has published research work and
search 173 (3), 2006, pp. 1139–1160. articles in the Journal of Information Technology,
[58] R.A. Ping Jr., On assuring valid measures for theoretical models using survey data, Information & Management, Journal of Systems and
Journal of Business Research 57 (2), 2004, pp. 125–141. Software, Journal of Software Maintenance and Evolution:
[59] P.M. Podsakoff, S.B. MacKenzie, J.-Y. Lee, N.P. Podsakoff, Common method biases Research and Practice, Information Systems Frontier
in behavioral research: a critical review of the literature and recommended among others. Dr. Ng’s research interests include social
remedies, Journal of Applied Psychology 88 (5), 2003, pp. 879–903. media, social commerce, and Enterprise Resource
[60] I. Qureshi, Y. Fang, E. Ramsey, P. McCole, P. Ibbotson, D. Compeau, Understanding Planning (ERP) maintenance and upgrade issues.
online customer repurchasing intention and the mediating role of trust—an
empirical investigation in two developed countries, European Journal of Infor-
mation Systems 18 (3), 2009, pp. 205–222.