Decision Support Systems 55 (2013) 669–678

Contents lists available at SciVerse ScienceDirect

Decision Support Systems

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

Exploring the effect of e-WOM participation on e-Loyalty in e-commerce

Chul Woo Yoo a,⁎, G. Lawrence Sanders b, 1, Junghoon Moon c, 2
a
Department of management science and systems, School of management, University at Buffalo, 325 Jacobs Management Center Buffalo, NY 14260-4000, United States
b
Department of management science and systems, School of management, University at Buffalo, 325G Jacobs Management Center Buffalo, NY 14260-4000, United States
c
Department of Regional Information, Seoul National University, 200-8203B, Shinlim-Dong, Gwanak-Gu, Seoul, 151-742, Republic of Korea

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

Article history: The idea behind the concept of Electronic Word of Mouth (e-WOM) is very important to the visibility of
Received 11 June 2012 individuals and businesses seeking exposure on the Internet. e-WOM is defined as “any positive or negative
Received in revised form 12 November 2012 statement made by potential, actual, or former customers about a product or company, which is made avail-
Accepted 1 February 2013
able to a multitude of people and institutions via the Internet” [39]. The purpose of this study is to understand
Available online 8 February 2013
how electronic word of mouth influences the online shopping patterns of customers. This study draws on
Keywords:
motivation theory and identification theory to understand how e-WOM influences online shopping behavior.
Electronic Word of Mouth The results from the empirical examination are very encouraging. The proposed research model is supported
Online participation as well as the accompanying hypotheses. The findings illustrate that the intrinsic motives (e.g., desire to help
e-Loyalty other customers) are more important than the extrinsic ones (e.g., monetary rewards) as the antecedents of
Social identification e-WOM participation. It was also found that personal site identification has a greater impact on e-loyalty
e-Commerce when compared to social site identification.
Customer review © 2013 Elsevier B.V. All rights reserved.

1. Introduction far reaching as e-WOM because they are restricted by time, space
and relationships. This is in contrast to the e-WOM environment
Virtual community participation and User Generated Content (UGC) where there are very few restrictions in terms of time and place of
are the prevailing social activities in virtual spaces. Along with such interaction due to the fact that the Internet is always available. In es-
trends, word-of mouth (WOM) transactions have also increased dra- sence, range and speed of electronic word-of-mouth is wider and
matically. WOM information on the Internet (henceforth e-WOM) has faster than traditional word-of-mouth communication [17].
become an important information source for shopping. e-WOM is de- Most e-commerce sites attempt to encourage their customers to
fined as “any positive or negative statement made by potential, actual, produce more e-WOM because consumer-produced information pro-
or former customers about a product or company, which is made avail- vides potential customers with a sense of trust [35]. e-WOM is con-
able to a multitude of people and institutions via the Internet” [39]. sidered a viable marketing strategy to attract new customers and to
e-WOM is known to have a strong impact on purchase decisions and retain existing customers. Researchers have recognized the growing
is a key factor in the Internet shopping market [61]. importance of e-WOM in e-business. For example, there have been
Consumers now use blogs, search engines, internet communities, several studies on e-WOM participation involving the motives of cus-
social media, consumer review systems and online instant messenger tomers (e.g., [14,39]), the effects of negative e-WOM (e.g., [13]), and
services to gather and disseminate product information. These tools the relationship between revenue and e-WOM (e.g., [15,35]). Most
are the basis for e-WOM networks and are used for information of these studies deal with the role of e-WOM as an information source
searching, decision-making, and forming brand images [13,38]. The for customers, or focus on how to motivate customers to participate
emergence of these e-WOM tools has changed the way customers in- in the process. However, very little is known about the relationship
teract with other customers as well as how they engage with busi- between e-WOM participation and the participant's next purchase.
nesses. Today's online shopper compliments and complains in a The purpose of this study is to investigate the role of e-WOM partic-
more aggressive and bolder way in this brave new world of faceless ipation as it relates to loyalty to an online shopping site.
e-commerce [37]. Traditional word-of-mouth interactions are not as In order to investigate the influence of e-WOM participation on
e-loyalty, we developed an integrated framework that consists of the-
ories of motivation, participation and identification. Cermak et al. and
Wind and Rangaswamy found that customer participation in service
⁎ Corresponding author. Tel.: +1 716 418 5210; fax: +1 716 645 6117. specification or mass customization leads to an increase in customer
E-mail addresses: [email protected] (C.W. Yoo), [email protected]
(G.L. Sanders), [email protected] (J. Moon).
satisfaction, loyalty, and repurchase [12,66]. Creating e-WOM by
1
Tel.: +1 716 645 2373; fax: +1 716 645 6117. writing comments or posting a review is a type of extra-role behavior
2
Tel.: +82 2 880 4772; fax: +82 2 873 5080. carried out by the buyer. Our study uses motivation and identification

0167-9236/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.dss.2013.02.001
670 C.W. Yoo et al. / Decision Support Systems 55 (2013) 669–678

theory to investigate the effect of e-WOM participation on partici- pathway that would allow the customer to actually investigate the
pant's e-loyalty towards a web site. product being offered [22].
In the next section of this paper, we will review previous studies on Through the literature review, we found that most studies on
e-WOM; we will then present hypotheses and develop a structural e-WOM show how e-WOM influences future customers and how it
equation model based on the theoretical framework, which is followed benefits online shopping malls. However, few studies have attempted
by introducing the measurement instruments and the survey items. to reveal the relationship between a customer's e-WOM participation
Finally, we will present the empirical results and discuss them. and his/her next purchase, or illuminate the psychological aspect of
e-WOM participation and the process of enhancing e-loyalty. The ob-
2. Literature review jective of this study is to fill this gap. The next section delves into the
research model used in this study.
Compared to the offline environment, there exists a large amount of
e-WOM information about specialized activity and product and service 3. Theory building and hypotheses development
communities, along with consumer reviews [64]. Chatterjee constructed
an imaginary internet shopping mall, and using an imaginary consumer Based on the above discussion, the conceptual framework illus-
survey, investigated the effects of e-WOM [13]. The results show that a trated in Fig. 1 will be used to examine the importance of e-WOM
negative consumer review has a critical influence on the consumer's in generating awareness. This framework draws on motivation theo-
trust and purchase intention at the internet shopping mall. Moreover, ry, participation theory and identification theory. Motivational theory
in the case of bargain shopper who prefers to purchase at a lower cost, is the force behind the process of e-WOM participation. Secondly,
the negative word-of-mouth impact is even greater. It appears that on- participation explains how the customer's e-WOM posting behavior
line shoppers are much more sensitive to negative e-WOM and this is leads to positive and negative attitudes toward certain online shop-
particularly true of low-priced products. As this finding is more promi- ping malls. Finally, identification theory is employed as a mechanism
nent in the case of small online shopping malls, Chatterjee pointed out to explain the process of the e-WOM participants' increased loyalty to
that the management of negative e-WOM is crucial for those companies online shopping malls.
solely dependent on their cost advantage when they do not have brand
power. 3.1. Motivational theory
Hennig-Thurau and Walsh studied the customer motives for reading
online word of mouth and the resulting effects. Their study found that The first stage of our conceptual framework, as shown in Fig. 1, is
the most important motivations for reading reviews were: diminution the customer's motivation to participate in e-WOM. In this stage, we
of risk, lowering of search time, understanding the various methods discover significant motives that make the customer participate in
for how to consume a product, reduction or curtailment of purchase re- writing e-WOM. It is important to examine motive in this study, be-
gret, compensation from an imaginary community, searching for new cause it provides practical and theoretical implications to boost
products, and increasing social status [39]. Diminution of decision mak- e-WOM participation. Motivation has been studied in a wide range
ing time and better purchase decision making motivation were shown of research studies and is considered an important theoretical vari-
to be the most significant factors motivating the customer to read able [21]. Woolfolk defines motive as a psychological state that in-
e-WOM. In addition, Hennig-Thurau et al. studied the motivation to duces, directs and maintains human behavior [68]. Deci separates
post e-WOM on the internet [40]. These results can be summarized in motive into intrinsic and extrinsic motives [20]. On the one hand, in-
terms of consumers' desire for social interaction, economic incentives, trinsic motives trigger behavior by reason of interest, enjoyment, and
and increase in social status. satisfaction caused by the behavior itself. On the other hand, extrinsic
Gruen et al. viewed e-WOM as a kind of know-how exchange among motives are followed by reason of compensation obtained by result of
the customers. And that the development of the know-how or e-WOM external activity. These intrinsic and extrinsic motives simultaneously
was preceded by opportunity, motivation and ability [36]. Furthermore, influence the intention of the individual.
when e-WOM is well managed, it was shown that it has positive effects In the context of information systems, Davis et al. introduced mo-
on the customer evaluation of the company and on intentions to tivation theory in terms of technology acceptance [19]. Consistent
repurchase. with previous studies examining other behavior, their study found
Bickart and Shindler allowed customers to gather information on 5 that both intrinsic and extrinsic motives were key factors in behavior-
product categories for a period of 12 weeks [10]. The results show al intention, leading to actual acceptance of new technology. In their
that customers actually pay more attention to the information provided study, intrinsic motive refers to perceived enjoyment—the extent to
by other customers rather than those of the salesperson or marketers which using a computer is perceived to be enjoyable distinct from
because they have used the product. In short, the information they pro- any performance outcomes that might be obtained. Empirical support
vide is more appealing and meaningful to the potential customers than for enjoyment as a determinant of behavioral intention to use a spe-
that of the salesperson or marketers. Chevalier and Mayzlin also studied cific technology has been demonstrated for both computer games
the effects of e-WOM on actual sales by examining online book reviews [50] and computer technology in the workplace [19]. Extrinsic motive
posted on Amazon and Barnes and Noble [15]. The results illustrate that refers to the use of a specific technology in the workplace by linking it
more positive e-WOM is posted on online book stores than offline, with to more productively performing job-related activities. Individuals
Amazon having far more positive e-WOM in both quantity and quality are often rewarded for good and productive behavior with raises, bo-
than Barnes and Noble. Overall, this study demonstrated that high nuses, etc. Therefore, if a technology is perceived to be useful in facil-
levels of e-WOM increases book sales. itating the individual's productivity, s/he is likely to have extrinsic
Dellarocas stressed that game theory, which is based on evaluation motivation to use a given technology [19]. The consistent results of
and trust, may be the most useful method for explaining the importance Garrity et al. also showed that ‘task support satisfaction’ and ‘shop-
of e-WOM interactions. Pavlou and Dimoka analyzed text-oriented ping enjoyment,’ adopted as extrinsic and intrinsic motives of ‘intent
e-WOM [53]. Their study shows e-WOM in the form of text, as opposed to use an electronic commerce web site’ are significant factors [33].
to simple customer ratings, constructs more trust and creates price pre- In this study, we examined four motives (i.e., concern for other
miums on the products. One problem with e-Bay is that the interactions customers, self enhancement, social benefit and economic incentives)
between sellers and buyers are very constrained. This leads to trust that were posited to facilitate e-WOM participation behavior. Con-
problems that are caused by the lack of interaction between a salesper- cern for other customers refers to other-oriented emotions evoked
son and a customer in e-commercial transactions, as well a lack of a by and congruent with the perceived welfare of someone in need
 C.W. Yoo et al. / Decision Support Systems 55 (2013) 669–678 671

Fig. 1. Conceptual framework for enhancement of e-Loyalty.

[7]. Self enhancement works as a type of motivation that helps to make who has low levels of participation tend to attribute service failure to
people feel good about themselves and to improve self-esteem by par- the service provider, while the customer who has high levels of par-
ticipating in e-WOM [56]. Social benefit is a type of motivation wherein ticipation do not [30]. To summarize, these studies do not directly ex-
customers want to interact with other customers by posting e-WOM plain the causal relationship between participation and customer
[40]. Economic incentives refer to a desire to participate in making loyalty; however, they do indirectly support the flow of our conceptu-
e-WOM because of potential monetary rewards. al framework, which posits that e-WOM participation behavior en-
In this study, the four motives are categorized as being intrinsic or hances the participant's e-loyalty.
extrinsic motives. Concern for others, self enhancement, and social
benefit are categorized primarily as intrinsic motives. Whereas eco-
nomic incentives are categorized as being extrinsic. Intrinsic motives 3.3. Identification theory
lead to e-WOM participation because they are associated with posi-
tive sensations of enjoyment, friendship, and pleasantness and are Research on identification theory is a remarkably important bridge
beneficial to themselves and other customers. Extrinsic motives can between participation and e-loyalty and helps to highlight how psycho-
be facilitated by offering monetary benefits, such as gift certificates, logical variables explain the connection between participation and
extra air miles and other prizes. Intrinsic and extrinsic motives are as- e-loyalty. The phenomenon of identification is often studied in organi-
sumed to drive or generate e-WOM participation, leading to the fol- zation research. Organization identification means that employees
lowing two hypotheses: perceive a sense-of-connectedness to an organization and identify
themselves in terms of it [48]. Cable and Judge show that high suitability
H1a. Intrinsic motives will have a positive impact on e-WOM between person and organization reduces separation intention and in-
participation. creases the intention to recommend the organization [11]. These results
imply a relationship between e-WOM and identification.
H1b. Extrinsic motives will have a positive impact on e-WOM
Bhattacharya and Sen examined the idea that a customer who has
participation.
a deep involvement in a company or product becomes a supporter of
that product or company [9]. As a result, based on social identity the-
3.2. Participation ory and organization identification theory, a customer identifies him/
herself with a company or product in order to meet his/her self-
The second construct examined in this model relates to customer's definition, which is based on needs. Underwood et al. proposed a
participation behavior. In this study, participation is operationalized mechanism of combined emotions between brand and customer,
as the actual level of involvement and frequency in e-WOM writing which he called social identity [63]. Social identity theory is based
and reviews (i.e., how much time and how much effort was invested on the self-concept, which refers to the awareness of and feelings
in writing e-WOM text). The importance of consumer participation for individuals and consists of personal identity and social identity.
has been conceptually and theoretically substantiated in previous Those phenomena have been investigated in the context of brand
studies as there have been a number of studies on the relationship be- identification in the marketing area [25,27,49]. Researchers in market-
tween consumer participation and productivity and profitability [29]. ing have studied consumers' perceptions and they have found that the
Related studies have been conducted on the motives and effects of perception of a brand goes beyond the functional attributes and benefits
customer participation in service production and delivery [4,23], on and includes nonfunctional, symbolic qualities (often referred to as the
psychological aspects and the impact on service quality of customer “brand image”) [6]. The concept of a brand personality is often used as a
participation [18], on the impact of customer participation regarding component of the brand image [49]. Brand personality is usually de-
the level of satisfaction, on repurchase and preference [26], and final- fined as the portfolio of human characteristics that are associated with
ly, on the effect of customer participation on the service provider [42]. a brand [1]. A number of studies have shown that a consumer's attitude
We found that most studies on participation demonstrate a strong is influenced by matching the perceived product image with their own
connection to positive attitudes, satisfaction, and preferences for a self-concept [59]. Brand identification happens when a consumer sees
brand or for a company. his or her own self-image as overlapping with the brand's image [4].
There are conflicting opinions about how customer participation Recent research findings further support the assertion that consumers
influences service evaluation. One is that when customers are not construct their self-identity and present themselves to others through
satisfied, the unsatisfactory service result is usually attributed to the their brand choices based on the congruency between brand-user asso-
service provider. In other words, because a high degree of participa- ciations and self-image associations [27].
tion is costly to the customer, the customer expects high quality ser- Rio et al. divide brand identification into two categories: personal
vices that compensate for his or her participation costs. Therefore, and social brand identification. Personal brand identification begins
if the service is poor in terms of the participation cost, the service with a customer's affinity for a brand [55]. Social brand identification
failure causes the customer to place blame on the service provider is subsequently enhanced when a customer is a member of a commu-
rather than him/herself. nity [47]. In brief, when the brand image enriches, matches and pro-
It should be noted that customers having a high degree of partici- tects a customer's self-image, this leads to a better evaluation result
pation tend to attribute unsatisfactory service to him/herself rather for the brand. Consequently, a better evaluation of a product or ser-
than to the service provider [58]. It appears that customers who par- vice leads to increased brand awareness [41].
ticipate actively in e-WOM feel responsible for a poor product or ser- We used the work of Rio et al. and classified identification in terms
vice. Such an assertion indirectly supports the idea that the customer of personal identification and social identification [55]. Both personal
672 C.W. Yoo et al. / Decision Support Systems 55 (2013) 669–678

identification and social identification with an online shopping site loyal [3]. Moon et al. also verify this assertion, where they found that
are investigated in this study. Personal site identification is the rela- blog use increases social identity with the blog and in turn strongly in-
tionship between an online shopping site and the customer, whereas fluences loyalty formation towards the blog service provider [51]. Based
social site identification occurs among customers through an e-WOM on these assertions and previous research findings, we suggest the fol-
system. Personal site identification refers to the extent to which a lowing hypotheses:
customer thinks the image of an online shopping site matches his/
her own image. Social identification refers to the identification that H3a. Personal site identification has a positive impact on e-loyalty.
a customer feels with respect to interactions, via the e-WOM system,
H3b. Social site identification has a positive impact on e-loyalty.
with other customers on the same online shopping site. Social identi-
fication also includes the degree to which the customer thinks these
relationships are valuable and how the relationships facilitate the 4. Research model
achievement of an identity in the online customer community.
When a customer engages in an e-WOM activity, this participation Based on the discussion of the six hypotheses, we also developed a
can be an opportunity for him or her to interact with an internet structural model that is presented in Fig. 2. This model is designed to ver-
shopping mall, and to understand the identity of the online social ify the relationships among motives, e-WOM participation, identifica-
community [8]. According to Bhattacharya, participation with a web tion and e-loyalty in an e-commerce environment. As discussed earlier,
site via e-WOM and other activities is the most effective method for intrinsic and extrinsic motives are set as preceding factors of e-WOM
promoting identification. Thus, we can infer that interactions with in- participation. Personal and social site identifications are set as preceding
ternet shopping malls, such as expressing satisfaction, giving advice factors of e-loyalty. In order to verify our research model, a research sur-
and getting feedback from a seller through an e-WOM system, can en- vey was conducted and a structural modeling analysis tool was used to
hance the participant's personal site identification. Meanwhile, Dick substantiate the research model.
et al. discuss extra-role behavior as having a positive impact on the
formation of identity in the group [24]. Feather and Rauter also sug-
gest that extra-role behavior enhances a sense-of-connectedness 5. Research method
and promotes social identity in the group [28]. Therefore, we can as-
sume that customers, who frequently participate in e-WOM, have a 5.1. Measurement development
higher sense-of-connectedness and membership than those who do
not. In essence, increased levels of customer participation are related The unit of analysis of this study is the individual user of e-WOM
to higher levels of recognition. After all, e-WOM participation may systems. Measurement items for the focal constructs were derived
enhance a customer's social site identification. Based on the literature from prior research and adjusted for our study. The survey was devel-
review, we propose these hypotheses: oped for the Korean online market space and careful attention to
measurement issues were taken in order to insure psychometric
H2a. e-WOM participation has a positive impact on personal site validity. After measurement item development, items were translated
identification. into a Korean version of the survey. Since there are language-oriented
differences such as nuance and interpretation, close attention was
H2b. e-WOM participation has a positive impact on social site
given to the translation. In order to check cleanness of items, reliabil-
identification.
ity and construct validity of the adjusted items, the pilot test for these
items (i.e., Korean version) was conducted with 26 graduate students
3.4. e-Loyalty in Korea.
As noted earlier, the intrinsic motive construct is a multidimensional
It is the ultimate goal of this paper to confirm the existence of a construct [32].We therefore set intrinsic motive as second order factor
causal relationship between e-WOM participation and e-loyalty. In consisting of the following first order factors: concern for other cus-
other words, we attempt to find out whether policies that online tomers, self enhancement, and social benefit. The relationship between
shopping malls use to stimulate and encourage customers to write the second order factor and first order factors are reflective because sep-
e-WOM are effective in increasing customer loyalty or e-loyalty. arate motives are manifestations of the construct, intrinsic motive [43].
e-Loyalty refers to customer loyalty in the internet market [54]. Garver and Mentzer pointed out that when researchers consider a hier-
Since internet shopping malls have different structural features, such archical structure for their measurement model, they need to use two
as no human element and material relating to non-existent facilities, approaches [34]. If the correlation coefficients between two constructs
the concept of customer loyalty in the internet market is also different. are higher than 0.70, the measurement model is encouraged to be spec-
Generally, customer loyalty represents the attitude and preference of a ified as a second order factor along with its first order factors, which
customer for a particular company and product. However, e-loyalty is are classified as two sub-constructs. However, the criterion 0.70 is not
conceptualized as a kind-of revisit attitude or revisit behavior in relation an absolute value, and thus, if theoretical basis strongly supports a
to a particular internet site [2]. Several recent studies have conceptual- sub-construct model, it too should be set as a second order factor rather
ized e-loyalty in a similar manner [2,60]. than a first order factor.
According to the extant literature [41,55], there is a strong relation- The items that were employed to measure personal site identifica-
ship between a customer's e-loyalty and identification with a brand. tion and social identification were adapted to our research setting
Several studies also explain the psychological aspects of this process [48,55]. Participation was adjusted to investigate the respondents' fre-
[9,57,69]. A customer's high site identification means that s/he under- quency and involvement in e-WOM writing. We then consulted the re-
stands the site well and has an affinity with it [9]. A brand image that search of Cermak et al. [12] and constructed tools to measure e-WOM
decidedly matches the customer's self-image makes the transaction sat- participation. Lastly, items from Anderson and Srinivasan [2] were se-
isfactory [57]. In addition, the customer tends to increase the amount of lected to measure e-loyalty. These constructs demonstrated substantial
purchasing activity when there are high levels of brand identification reliability and internal consistency. All details of the items are attached
[69]. This in turns leads to high levels of personal site identification, in Table 1. One question, “Do you agree with the opinion that the inter-
which in turn translates to stronger e-loyalty. net connection fee via mobile phone is likely to be raised in the future?”
Ashforth and Mael suggest that a community member with a strong is added as a marker variable to examine common method bias in the
sense of social identity among other members becomes devoted and questionnaire survey.
 C.W. Yoo et al. / Decision Support Systems 55 (2013) 669–678 673

Fig. 2. Research model.

5.2. Data collection internet shopping malls in Korea. Its online review system is used
well by customers. This site is considered to have an appropriate en-
In order to target online users, we employed a web-based survey. vironment for this study.
Online surveys have several advantages over traditional paper-based First we sorted customers who have at least one transaction re-
surveys: (1) the sample is not restricted to a single or local geograph- cord within a recent 6 months to avoid fake accounts or lost accounts.
ical location, (2) lower costs accrue, and (3) faster responses are like- Within the customer database, we randomly selected 300 customers
ly. We collected data in Korea since the e-commerce environment of as our informants, and invitation emails were sent to them. The ques-
Korea is very well developed, and customer's involvement in the In- tionnaire consisted of an instruction page that opened a separate Web
ternet was prevalent. For example, Korea has the largest percentage browser window containing the items to be assessed. At the last page,
of wireless broadband users in the world [52]. Korean e-commerce we placed an advertisement, which offered a prize to the respon-
environment is considered to represent general customer behavior dents, to try to improve the number of high-quality responses. The re-
online. A particular internet shopping mall (K company) was targeted spondents were asked to answer all the questions based on their
and requested to conduct our survey. K company is one of the popular experience, using a 1–7 Likert-type scale with an anchor of 1 for
“strongly disagree” to 7 for “strongly agree.” Answering all the ques-
tions was required to click the submit button.
A total of 257 responses were gathered. The response rate was
Table 1 85.6% (257/300). Seven of the 257 were printed out rather than
Measurement instruments. sent back electronically. Participants received about $5 worth of in-
Items Measures centives for completing the questionnaire. Fifty-nine percent of the
respondents were female, and the majority of respondents were peo-
Intrinsic motive (I write comments on customer review because…)
IM1 …I want to help others with my own positive experience. ple in their 30s and who had more-than-10-times experience with
IM2 …I want to give others the opportunity to buy the right product. e-commerce. During the analysis process all 257 responses were
IM3 …This way I can express my joy about a good buy. used because specific outlier and omitted responses were not found.
IM4 …I feel good when I can tell others about my buying successes.
IM5 …I believe a chat among like-minded people is a nice thing.
IM6 …It is fun to communicate this way with other people in the community. 6. Results

Extrinsic motive (I write comments on customer review because…) Assessment of the research model was conducted using PLS (Partial
EM1 …of incentives I receive (e.g., cyber money). Least Square). PLS is a structured equation modeling technique that can
EM2 …I receive a reward for my writing.
analyze structural equation models (SEMs) involving multiple-item
EM3 …I can get discount coupons on my next purchase on the site.
constructs, with direct and indirect paths. PLS works by extracting
e-WOM participation successive linear combinations of the predictors and is effective in
EP1 I frequently write a customer review. explaining both response and predictor variation [16]. PLS is a powerful
EP2 When I leaving a customer review, I elaborately write it.
tool for analyzing models because of the minimal demands on measure-
EP3 I spend much effort in posting review.
ment scales, sample size, and residual distributions. In addition, PLS
Personal site identification avoids two serious problems, inadmissible solutions and factor indeter-
PI1 This online shopping mall is associated with my life style well. minacy [31]. SEM approaches, such as LISREL and AMOS, are not able to
PI2 I have positive feeling at this online shopping mall. deal with non-normal distributions, and they can yield non-unique or
PI3 This online shopping mall suit with my self-image.
otherwise improper solutions in some cases [31]. PLS is not as suscepti-
Social site identification ble to these limitations [67]. The emphasis of PLS is on predicting the
SI1 I interact with other customers through review system. responses as well as in understanding the underlying relationship
SI2 I have positive feeling with other customers. between the variables [62].
SI3 I become important person to other customer through review system.
SI4 Because of product review I wrote, other customers respect me.
6.1. Measurement assessment
e-Loyalty
EL1 I seldom consider switching another to internet shopping mall. A PLS analysis involves two stages: (1) the assessment of the mea-
EL2 I believe that this is my favorite internet shopping mall. surement model, including the reliability and discriminant validity
EL3 When I need to purchase, this shopping mall is my best choice.
of the measures, and (2) the assessment of the structural model.
674 C.W. Yoo et al. / Decision Support Systems 55 (2013) 669–678

Table 2 Table 3
Internal consistency and convergent validity. Correlation of latent variables.

Item Weight Loading Mean SD Alpha CR AVE IM EM EP PI SI EL

Intrinsic motive IM1 0.209 0.854 3.86 1.40 0.920 0.937 0.714 Intrinsic motive (0.845)
IM2 0.208 0.858 3.86 1.38 Extrinsic motive 0.672 (0.914)
IM3 0.154 0.766 4.99 1.17 e-WOM participation 0.699 0.618 (0.872)
IM4 0.153 0.768 4.99 1.16 Personal identification 0.393 0.265 0.384 (0.859)
IM5 0.226 0.911 4.72 1.32 Social identification 0.615 0.461 0.604 0.511 (0.853)
IM6 0.224 0.901 4.72 1.29 CMB marker 0.252 0.076 0.135 0.161 0.167 0.075
Extrinsic motive EM1 0.416 0.916 4.56 1.53 0.903 0.939 0.836 e-Loyalty 0.392 0.251 0.314 0.682 0.487 (0.944)
EM2 0.359 0.934 3.94 1.58
Note: The number in parenthesis is the square root of AVE.
EM3 0.319 0.893 3.95 1.53
e-WOM EP1 0.423 0.907 3.99 1.60 0.841 0.905 0.762
participation EP2 0.387 0.875 4.60 1.53
6.2. Structural model assessment
EP3 0.333 0.833 4.37 1.74
Personal PI1 0.373 0.860 4.73 1.03 0.809 0.894 0.738
identification PI2 0.422 0.908 5.00 1.07 The structural equation model proposed for this study is assessed, as
PI3 0.367 0.806 4.94 1.32 shown in Fig. 3. With an adequate measurement model, the hypotheses
Social SI1 0.300 0.861 4.51 1.30 0.875 0.914 0.728
were tested by examining the structural model. The R 2 value was used
identification SI2 0.297 0.844 4.71 1.21
SI3 0.300 0.842 4.73 1.34 to assess the proportion of variance in the endogenous constructs that
SI4 0.276 0.865 4.29 1.26 could be explained by the antecedent constructs. Approximately 53%
e-Loyalty EL1 0.361 0.945 5.41 1.02 0.939 0.961 0.892 of the variance in e-WOM participation was explained by the two moti-
EL2 0.359 0.954 5.24 1.11 vational factors, and 49% of the variance in e-loyalty was explained by
EL3 0.338 0.934 5.34 1.09
the two identification factors, making the interpretation of the path co-
CMB marker CMB na na 4.25 1.23 na na na
efficients meaningful. The results also confirmed that intrinsic motive is
Note: CMB marker is an item to check the common method bias in the study.
a more influential motivational factor affecting e-WOM participation
than extrinsic motive (b= 0.517, p b 0.01), e-WOM participation has
more influence on social site identification than personal identification
Individual item loadings and internal consistency were examined as a (b= 0.604, p b 0.01), and personal site identification has a stronger im-
test of reliability. Individual item loadings that are greater than 0.7 pact on e-loyalty than social site identification does in the e-commerce
are considered to be adequate. As shown in Table 2, loadings for all environment (b= 0.586, p b 0.01). The path coefficients from the PLS
measurement items are above 0.7, indicating that there is sound in- analysis are shown in Fig. 3. The coefficients in the model represent
ternal reliability. In addition, all the weights are statistically significant standardized regression coefficients.
at p b 0.01. The almost uniformly distributed weights show each item After computing path estimates in the structural model, the PLS
contributes to each construct equivalently. In addition, we also investi- software was used to perform a bootstrap to obtain the correspond-
gated Cronbach's alpha for internal consistency. Table 2 shows that ing t-values. Support for each hypothesis can be determined by exam-
Cronbach's alpha for all constructs was greater than 0.7. The Average ining the sign (positive or negative) and the statistical significance for
Variance Extracted (AVE) was also calculated. This shows the variance the t-value for its corresponding path. All hypotheses were accepted
that a construct captures from its indicators, relative to the variance with a significance level of 0.01, as represented in Fig. 3.
contained in measurement error. This statistic is generally interpreted
as a measure of reliability for the construct and as a means of evaluating 6.3. Testing mediating effect
discriminant validity [5]. All AVEs for the constructs in our study were
greater than 0.7. This indicated that 70% of the variance of the indicators After the structural model assessment, we investigated the mediating
could be accounted for by the latent variables. Also, if all composite re- effect of ‘Site Identification’ at the relationship between ‘e-WOM Partici-
liability values are higher than 0.7, it can be concluded that that mea- pation’ and ‘e-Loyalty.’ First of all we tested Model A of Fig. 4, a direct re-
surement has both internal consistency and convergent validity [65]. lationship of ‘e-WOM Participation’ and ‘e-Loyalty.’ As a result, the
According to the results shown in Table 2, all result values in this relationship is significant but the R2 is relatively low (0.101). In Model
study are higher than 0.8, which means that the measurement model B, and C of Fig. 4, we tested the mediating roles of ‘Personal Site Identifi-
of this study has suitable composite reliability. cation’ and ‘Social Site Identification.’ Fig. 4 shows that ‘Site Identifica-
The AVE is also used to assess discriminant validity. The square root tion’ fully mediates the relationship between e-WOM participation and
of AVE should be greater than the correlations among the constructs; e-loyalty because significances of direct relationship disappear in both
that is, the amount of variance shared between a latent variable and cases. Furthermore, each R2 of the dependent variable is increased
its block of indicators should be greater than the shared variance be- more than two and four times respectively, compared to the direct rela-
tween the latent variables. Table 3 shows the inter-correlations of the tionship. Consequently, these results indicate the importance of the role
constructs and variance shared between the latent variables and their of site identification in converting the consumer's behavior into loyalty
indicators. The diagonal elements in Table 3 are the square root of the to the site.
AVE. This showed that the square roots of each AVE value were greater
than the off-diagonal elements. The measurement model, thus, had a 6.4. Common method bias
reasonable degree of discriminant validity among all of the constructs.
The results of the measurement analysis also indicated that all the We investigated the presence of common method bias because de-
constructs and measures have acceptable discriminant validity. pendent and independent variables come from the same source. The
In order to verify the discriminant validity in more detail, a cross- partial correlation technique was adopted by using a marker variable
loading analysis was also conducted. As can be seen in the cross- to control the common method bias [45,46]. Lindell and Whitney [46]
loading table (Table 4), all the loadings of the constructs of the latent showed that if a variable can be identified theoretically, it should not
variables were over 0.7 and all the factor loadings were significant at a be related to other variables included in the study. Thus, it can be
confidence level of 0.01. Furthermore, for other constructs the loadings used as a marker variable. Therefore, any associations between marker
(cross-loadings) were below 0.7. This shows that the measurement and any of the other variables can be assumed to be due to the common
model of this study has strong discriminant validity. method bias.
 C.W. Yoo et al. / Decision Support Systems 55 (2013) 669–678 675

Table 4 First of all, we found that both intrinsic and extrinsic motives have a
Cross-loading matrix. significant impact on e-WOM participation. This result is consistent with
IM EM EP PI SI EL the results of Davis et al. [19] and Garrity et al. [33]. Our results com-
bined with their results suggests that motivation theory can explain
IM1 0.854 0.665 0.621 0.307 0.520 0.297
IM2 0.858 0.657 0.619 0.310 0.522 0.306 e-WOM participation. We compared path coefficients and found that
IM3 0.766 0.384 0.458 0.418 0.521 0.384 intrinsic motives have a stronger influence on e-WOM participation
IM4 0.768 0.393 0.453 0.416 0.529 0.385 than extrinsic motives. This difference implies that customers are
IM5 0.911 0.619 0.672 0.306 0.536 0.333
more motivated by considering e-WOM participation as a ‘kind’ and
IM6 0.901 0.612 0.664 0.303 0.515 0.328
EM1 0.678 0.916 0.638 0.285 0.469 0.310 ‘good’ behavior to themselves and to other customers; furthermore,
EM2 0.591 0.934 0.551 0.257 0.450 0.246 they expect that other customers’ e-WOM participation will help them
EM3 0.560 0.892 0.489 0.170 0.328 0.107 later. The previous literature also discusses the fact that intrinsic motives
EP1 0.693 0.650 0.907 0.331 0.563 0.276 play a more important role than extrinsic motives in education and
EP2 0.609 0.508 0.875 0.334 0.557 0.253
learning [21]. In addition, it has been found by other researchers that
EP3 0.510 0.441 0.833 0.346 0.452 0.298
PI1 0.375 0.285 0.319 0.860 0.445 0.561 strong external compensation weakened intrinsic motives [44]. Hence
PI2 0.408 0.278 0.380 0.908 0.515 0.625 we can motivate customers to participate in e-WOM by stimulating
PI3 0.222 0.113 0.286 0.806 0.348 0.569 their intrinsic motives through, for example, advertising e-WOM partic-
SI1 0.578 0.398 0.560 0.453 0.861 0.382
ipation as ‘kind’ and ‘benevolent’ behavior, as well as promoting extrinsic
SI2 0.510 0.375 0.459 0.507 0.844 0.498
SI3 0.499 0.350 0.530 0.395 0.842 0.419 motives through mileage, bonus points, or other compensation. In other
SI4 0.509 0.453 0.511 0.385 0.865 0.357 words, to make the customer aware of the ‘goodness’ and ‘kindness’
EL1 0.374 0.260 0.295 0.660 0.464 0.944 of e-WOM is more effective than to incite them by simple external
EL2 0.379 0.232 0.305 0.656 0.464 0.954 compensation.
EL3 0.358 0.218 0.289 0.614 0.449 0.934
We also found that e-WOM participation behavior enhances social
identity among customers. The findings of this study also suggest that
social identity plays a role in using the e-WOM system. It is necessary
for practitioners to maintain an e-WOM system from the perspective
First, we computed the correlation coefficient between the marker of communication media so that they can develop strong social iden-
and other variables (see Table 3). And we tested the significance tity on the site through enhanced interaction with other customers.
of the relationships among our core variables after partialing out As discussed earlier, attempts such as increasing economic incentives,
the impact of marker variable. Next, we checked every relationship enhancing interaction with a manager's active feedback, and promo-
among constructs, which remained significant. As a result, we con- tion of campaigns inspiring altruism in terms of e-WOM helps to im-
firm that the correlation coefficient between the marker and other prove the interactions between customers so that they can develop a
variables is not significant and the relationships among our core vari- strong social identity. Also, installation of an e-WOM system through
ables, after partialing out the impact of the marker, still remain which customers can post e-WOM using richer media, such as images
significant. and videos, is also helpful in enhancing the quality of the interaction
between customers. The net result is that this will lead customers to
build e-loyalty towards the online site.
7. Discussion and implications Finally, both personal and social site identifications turned out to
have a significant influence on customer e-loyalty. However, personal
This study attempted to examine the impact of e-WOM participation identification has a stronger path coefficient and higher mean value
on e-loyalty from the perspective of the participant, not the reader. We than social site identification. Personal site identification was measured
introduced motivation theory to explain e-WOM participation, and by the degree of fitness between the image achieved by products and
employed identification theory to analyze the psychological aspect of the service that the site provides, and the image of the customer. The
the e-WOM participant. To achieve the goal of this study, we conducted take-away is that when customers identify themselves by continuing
a survey and proposed a model including 6 accompanying hypotheses. to participate in electronic-word-of-mouth, then e-loyalty emerges.
Implications from the results for academia and practitioners are dis- This result is consistent with previous studies (e.g., [57,69]). Moreover,
cussed as follows. this study expanded identification theory from the offline brand to the

Fig. 3. Structural model results.

676 C.W. Yoo et al. / Decision Support Systems 55 (2013) 669–678

Fig. 4. Mediation test results.

online shopping site. Practically speaking, this result stresses the im- when they purchase heterogeneous goods (e.g., fresh food, clothes)
portance of the customization of internet services and site interactivity. rather than homogeneous goods (e.g., CDs, DVDs, books) from the inter-
To promote personal site identification, customized content service and net; that is, customers who purchase heterogeneous goods have a
proper feedback replies from shopping sites are required. Customized greater tendency to rely more on e-WOM. Eventually, the role of
content service and high interactivity of site enhance customers' satis- e-WOM for heterogeneous goods must be more critical than that of
faction and give them opportunities to understand the shopping site. the others. It would be worthwhile to study the effect of the interactiv-
In order to corroborate these aspects, machine learning technology ity of an online shopping site on different product types. The results of
that customizes sites and extends communication channels, can be such a study would have profound practical implications for online
employed. shopping mall management.
This study demonstrates that these strategies (e.g., customization
of internet services and site interactivity enhancement) increase
personal site identification, and ultimately make customers loyal. Fur- 8. Conclusion
thermore, the interaction among customers promotes the formation
of social site identification. Loyal customers are recognized as those We believe that e-WOM has become an important part of the online
who believe themselves to be important to other customers on the shopping experience. Understanding the phenomena is essential to
site. Thus, as discussed earlier, the installation of an e-WOM system electronic commerce systems and user behavior. This study has helped
that allows customers to easily interact with others must be one of to bring additional insight to the relationship between customer partic-
the key success factors of the loyalty program of an e-commerce ipation in e-WOM and consumer purchases. The empirical findings sug-
site. Furthermore, the development of a communication channel gest that the level of motivation encourages customers to participate in
connecting customer to customer is a requirement, as well as a chan- review systems, which is in line with prior e-WOM research. Specifically,
nel between the shopping site and the customers. For example, it is it was found that internal motivation is more powerful in promoting
helpful for enhancing social site identification to encourage cus- e-WOM systems. At the same time, the results of this study also reveal
tomers to reply to other customers' e-WOM or to offer them an exog- that customer's review participation has a significant impact on site
enous reward for doing so. identification building and should not be ignored by online retailers.
This study provides several opportunities for future research. First, These insights not only help researchers better understand how incen-
the surveys were mainly conducted in a single online shopping mall. tives for review posting and customer's review posting behavior work
Thus, in order to further generalize our results, we need to collect data in electronic markets, but also provide guidelines for online retailers to
from a variety of online shopping malls. Secondly, we investigated better adjust their IT strategies to improve customer retention.
the respondents' satisfaction with a shopping site. There were a few It is hoped that future research will be conducted on user review
respondents that had bad feelings about them. Customers who have posting behavior and site identification. IS theory could be extended
negative feelings may write negative e-WOM comments because they from the typical context of using the system to purchase goods to situa-
do not like the online shopping experience. This behavior can be differ- tions in which users make voluntary contributions to systems. Given that
ent from that of the respondents who are satisfied with the online shop- electronic markets have become an irrevocable part of modern society,
ping experience. Thirdly, the research model and analysis methods of further research on identification and user participation is warranted.
this study were designed based on a cross-sectional data set. This has
the apparent limitation of not allowing for the consideration of the
lifecycle of social identity. Thus, time-series analysis methods will be References
necessary to identify dynamic relationships between participation [1] J.L. Aaker, Dimensions of brand personality, Journal of Marketing Research 34 (3)
behavior and online social identity development. (1997) 347–356.
Finally, according to product type, the customers' site identification [2] R.E. Anderson, S.S. Srinivasan, E-satisfaction and e-loyalty: a contingency frame-
work, Psychology and Marketing 20 (2) (2003) 123–138.
might be different depending on product types (e.g., CDs, DVDs, books [3] B.E. Ashforth, F. Mael, Social identity theory and the organization, Academy of
vs. fresh food, clothes) because they feel much more uncertainty Management Review 14 (1) (1989) 20–39.
 C.W. Yoo et al. / Decision Support Systems 55 (2013) 669–678 677

[4] R.P. Bagozzi, U.M. Dholakia, Antecedents and purchase consequences of customer [36] T.W. Gruen, T. Osmonbekov, A.J. Czaplewski, Ewom: the impact of customer-to-
participation in small group brand communities, International Journal of Research customer online know-how exchange on customer value and loyalty, Journal of
in Marketing 23 (1) (2006) 45–61. Business Research 59 (4) (2006) 449–456.
[5] Y. Bakos, The emerging role of electronic marketplaces on the internet, Commu- [37] L.J. Harrison-Walker, E-complaining: a content analysis of an internet complaint
nications of the ACM 4 (8) (1998) 35–42. forum, Journal of Consumer Marketing 15 (5) (2001) 397–412.
[6] R. Batra, P. Lenk, M. Wedel, Brand extension strategy planning: empirical estima- [38] S. Helm, Viral marketing — establishing customer relationships by word of
tion of brand–category personality fit and atypicality, Journal of Marketing mouth, Electronic Markets 10 (3) (2000) 158–174.
Research 47 (2) (2010) 335–347. [39] T. Hennig-Thurau, G. Walsh, Electronic word-of-mouth: motives for and conse-
[7] C.D. Batson, The Altruism Question: Toward a Social–Psychological Answer, Lawrence quences of reading customer articulations on the internet, International Journal
Erlbaum Associates, Hillsdale, NJ, 1991. of Electronic Commerce 8 (2) (2003) 51–74.
[8] C.B. Bhattacharya, When customers are member: customer retention in paid [40] T. Hennig-Thurau, K.P. Gwinner, G. Walsh, D.D. Gremler, Electronic word-
membership contexts, Journal of the Academy of Marketing Science 26 (1) of-mouth via consumer-opinion platforms: what motivates consumers to artic-
(1998) 31–44. ulate themselves on internet? Journal of Interactive Marketing 18 (1) (2004)
[9] C.B. Bhattacharya, S. Sen, Consumer company identification: a framework for un- 38–52.
derstanding consumers' relationships with companies, Journal of Marketing 67 [41] M.K. Hogg, A.J. Cox, K. Keeling, The impact of self-monitoring on image congru-
(2) (2003) 76–88. ence and product/brand evaluation, European Journal of Marketing 34 (5/6)
[10] B. Bickart, R.M. Schindler, Internet forums as influential sources of consumer in- (2000) 641–666.
formation, Journal of Interactive Marketing 15 (3) (2001) 31–40. [42] A. Hsieh, A.C. Yen, The effect of customer participation on service providers' job
[11] D.M. Cable, T.A. Judge, Person-organization fit, job choice decisions and organiza- stress, Service Industries Journal 25 (7) (2005) 891–905.
tional entry, Organizational Behavior and Human Decision Processes 67 (3) (1996) [43] C.B. Jarvis, S.B. MacKenzie, P.M. Podsakoff, A critical review of construct indicators
294–311. and measurements model misspecification in marketing and consumer research,
[12] D.S.P. Cermak, K.M.K. File, P.A. Prince, Customer participation in service specifica- Journal of Consumer Research 30 (2) (2003) 199–218.
tion and delivery, Journal of Applied Business Research 10 (2) (1994) 90–98. [44] A. Kohn, Why incentive plans cannot work, Harvard Business Review 71 (5)
[13] P. Chatterjee, Online reviews: do consumers use them? Advances in Consumer (1993) 54–60.
Research 28 (1) (2001) 129–133. [45] M.K. Lindell, C.J. Brandt, Climate quality and climate consensus as mediators of
[14] C.M.K. Cheung, M.K.O. Lee, What drives consumers to spread electronic word of the relationship between organizational antecedents and outcomes, Journal of
mouth in online consumer-opinion platforms, Decision Support Systems 53 (1) Applied Psychology 85 (3) (2000) 331–348.
(2012) 218–225. [46] M.K. Lindell, D.J. Whitney, Accounting for common method variance in cross-
[15] J. Chevalier, D. Mayzlin, The effect of word of mouth on sales: online book sectional designs, Journal of Applied Psychology 86 (1) (2001) 114–121.
reviews, Journal of Marketing Research 43 (3) (2005) 345–354. [47] M.M. Long, L.G. Schiffman, Consumption values and relationships: segmenting
[16] W.W. Chin, Issue and opinion on structural equation modeling, MIS Quarterly 22 the market for frequency programs, Journal of Consumer Marketing 17 (3)
(1) (1998) 7–16. (2000) 214–232.
[17] E. Churchill, A. Girgensohn, L. Nelson, A. Lee, Blending digital and physical spaces [48] F. Mael, B.E. Ashforth, Alumni and their alma mater: a partial test of the
for ubiquitous community participation, Communications of the ACM 47 (2) reformulated model of organizational identification, Journal of Organizational
(2004) 38–44. Behavior 13 (2) (1992) 103–123.
[18] C. Claycomb, C.A. Lengnick-Hall, L.W. Inks, The customer as a productive resource: a [49] S. Magin, R. Algesheimer, F. Huber, A. Herrmann, The impact of brand personality
pilot study and strategic implication, Journal of Business Strategies 18 (1) (2001) and customer satisfaction on customer's loyalty: theoretical approach and findings
47–68. of a causal analytical study in the sector of internet service providers, Electronic
[19] F.D. Davis, R.P. Bagozzi, P.R. Warshaw, Extrinsic and intrinsic motivation to use com- Markets 13 (4) (2003) 294–308.
puters in the workplace, Journal of Applied Social Psychology 22 (14) (1992) [50] T.W. Malone, Toward a theory of intrinsically motivating instruction, Cognitive
1111–1132. Sciences 5 (4) (1981) 333–369.
[20] E.L. Deci, Intrinsic Motivation, Plenum, New York, NY, 1975. [51] J.H. Moon, J.P. Li, S. Jo, G.L. Sanders, Improving quality of life via blog and develop-
[21] E.L. Deci, R.M. Ryan, Intrinsic Motivation and Self-Determination in Human ment of a virtual social identity, Journal of Information Technology Management
Behavior, Plenum, New York, NY, 1985. 13 (3) (2006) 26–37.
[22] C. Dellarocas, The digitization of word of mouth: promise and challenges of online [52] C. Osborne, South korea hits 100% mark in wireless broadband, in: (C.NET, 2012).
feedback mechanisms, Management Science 49 (10) (2003) 1407–1424. [53] P.A. Pavlou, A. Dimoka, The nature and role of feedback text comments in online
[23] U.M. Dholakia, R.P. Bagozzi, L.K. Pearo, A social influence model of consumer par- marketplaces: implications for trust building, price premiums, and seller differen-
ticipation in network- and small-group-based virtual communities, International tiation, Information Systems Research 17 (4) (2006) 392–414.
Journal of Research in Marketing 21 (3) (2004) 241–263. [54] F. Reichheld, P. Schefter, E-loyalty: your secret weapon on the web, Harvard Busi-
[24] R.v. Dick, M.W. Grojean, O. Christ, J. Wieseke, Identity and the extra mile: rela- ness Review 78 (4) (2000) 105–113.
tionships between organizational identification and organizational citizenship [55] A.B.d. Rio, R. Vazquez, V. Iglesias, The effects of brand associations on consumer
behavior, British Journal of Management 17 (4) (2006) 283–301. response, Journal of Consumer Marketing 18 (1) (2001) 410–425.
[25] S. Dimitriadis, E. Papista, Linking consumer-brand identification to relationship [56] C. Sedikides, M.J. Strube, The multiply motivated self, Personality and Social Psy-
quality: an integrated framework, Journal of Customer Behavior 10 (3) (2011) chology Bulletin 21 (12) (1995) 1330–1335.
271–289. [57] B. Shiv, J. Huber, The impact of anticipating satisfaction on consumer choice, Jour-
[26] A.B. Eisingerich, S.J. Bell, Relationship marketing in the financial service indus- nal of Consumer Research 27 (2) (2000) 202–217.
try: the importance of customer education, participation and problem manage- [58] P. Silpakit, R.P. Fisk, Participating in the Service Encounter: A Theoretical Frame-
ment for customer loyalty, Journal of Financial Services Marketing 10 (4) (2006) work, American Marketing Association, Chicago, IL, 1985.
121–132. [59] M.J. Sirgy, D. Grewal, T.F. Mangleburg, Retail environment, self-congruity, and retail
[27] J.E. Escalas, J.R. Bettman, You are what they eat: the influence of reference groups patronage: an integrative model and research agenda, Journal of Business Research
on consumers' connections to brands, Journal of Consumer Psychology 13 (3) 49 (2) (2000) 127–138.
(2003) 339–348. [60] S. Srinivasan, R. Anderson, K. Ponnavolu, Customer loyalty in e-commerce: an ex-
[28] N.T. Feather, K.A. Rauter, Organizational citizenship behaviours in relation to job ploration of its antecedents and consequences, Journal of Retailing 8 (4) (2002)
status, job insecurity, organizational commitment and identification, job satisfac- 283–299.
tion and work values, Journal of Occupational and Organizational Psychology 77 [61] B. Stauss, Global word of -mouth: service bashing on the internet is a thorny
(1) (2004) 81–94. issue, Marketing Management 6 (3) (1997) 28–30.
[29] J.A. Fitzsimmons, Consumer participation and productivity in service operation, [62] R.D. Tobias, An Introduction to Partial Least Squares Regression, SAS Institute,
Interfaces 15 (3) (1985) 60–67. Cary, NC, 1999.
[30] V.S. Folkes, Customer reactions to product failure: an attributional approach, Journal [63] R. Underwood, E. Bond, R. Baer, Building service brands via social identity: lessons
of Consumer Research 10 (4) (1984) 398–409. from the sports marketplace, Journal of Marketing Theory and Practice 9 (1)
[31] C. Fornell, F.L. Bookstein, Two structural equation models: Lisrel and pls applied (2001) 1–13.
to consumer exit-voice theory, Journal of Marketing Research 19 (4) (1982) [64] L.C. Ward, A.L. Ostrom, The internet as information minefield: an analysis of the
440–452. source and content of brand information yielded by net searches, Journal of Busi-
[32] M. Gagne, C.B. Senecal, R. Koestner, Proximal job characteristics, feelings of ness Research 56 (11) (2003) 907–914.
empowerment, and intrinsic motivation: a multidimensional model, Journal of [65] C.E. Werts, R.L. Lin, K.G. Jöreskog, Intraclass reliability estimates: testing structur-
Applied Social Psychology 27 (14) (1997) 1222–1240. al assumptions, Educational and Psychological Measurement 34 (1) (1974)
[33] E.J. Garrity, J. O'Donnell, Y.J. Kim, G.L. Sanders, An extrinsic and intrinsic 25–33.
motivation-based model for measuring consumer shopping oriented web [66] J. Wind, A. Rangaswamy, Customerization: the next revolution in mass customi-
site success, Journal of Electronic Commerce in Organizations 5 (4) (2007) zation, Journal of Interactive Marketing 15 (1) (2001) 13–33.
18–38. [67] H. Wold, Causal flows with latent variables, European Economic Review 5 (1)
[34] M.S. Garver, J.T. Mentzer, Logistics research methods: employing structural equa- (1974) 67–86.
tion modeling to test for construct validity, Journal of Business Logistics 20 (1) [68] A.E. Woolfolk, Educational Psychology, 6th edition Allyn and Bacon, Boston, MA,
(1999) 33–57. 1995.
[35] D.K. Gauri, A. Bhatnagar, R. Rao, Role of word of mouth in online store loyalty: [69] Y. Yi, S. La, Brand personality-brand identification-brand equity model: an explor-
comparing online store ratings with other e-store loyalty factors, Communica- atory study on difference between users vs nonusers, Journal of Korean Market-
tions of the ACM 51 (3) (2008) 89–91. ing Association 17 (3) (2002) 1–33.
678 C.W. Yoo et al. / Decision Support Systems 55 (2013) 669–678

Chul Woo Yoo is currently a PhD student in the Department of Management Science Junghoon Moon is an assistant professor of the Department of regional information at
and Systems in the School of Management at the State University of New York at the Seoul National University, Korea. He received his Ph.D. degree from the State Uni-
Buffalo. He holds a Master's degree in Economics from Seoul National University, Ko- versity of New York at Buffalo in 2006. He worked for several years as a system analyst
rea. His research interests include human factor in e-business, information privacy, pi- and consultant. In addition, he is an associate director of Auto-ID labs sponsored by
racy behavior, and agricultural information system. EPCglobal. Dr. Moon's research interests include Human factor in MIS/E-business,
Technology Management, E-government, Information Policy for Food Industry, and
G. Lawrence Sanders, PhD, is a professor in the Department of Management Science and Business Application using an Ubiquitous Sensor Network.
Systems in the School of Management at the State University of New York at Buffalo. He He has presented his studies at KMIS, IRMA, AMICS and HICSS. At AMICS 2006 one of
has also served as a department chair and the chair of the PhD program in the School of his studies was awarded as the best paper of the year. At HICSS 2007 one of his studies
Management. He has taught MBA courses in the People's Republic of China, India and was nominated as the best paper of the year. He has published articles in many
Singapore. His research interests are in the ethics and economics of digital piracy, systems journals, including Online Information Review, Asia Pacific Journal of Information
success measurement, cross-cultural implementation research, systems development and Systems, e-Business Studies, Journal of Information Technology Management, Information
decision processes. Systems Frontiers, and Technological Forecasting & Social Change.
He has over 50 conference proceedings and presentations and has also published over 50
referred research papers in outlets such as The Journal of Business, MIS Quarterly, Informa-
tion Systems Research, the Journal of Management Information Systems, the Journal of Strate-
gic Information Systems, the Journal of Management Systems, Decision Support Systems, and
Decision Sciences. He has also published a book on database design and co-edited two
other books. He is currently working on a book on technology development.