Computers in Human Behavior 28 (2012) 10541062

Contents lists available at SciVerse ScienceDirect

Computers in Human Behavior

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

The role of self-construal in consumers electronic word of mouth (eWOM) in social networking sites: A social cognitive approach
Doohwang Lee a,, Hyuk Soo Kim b, Jung Kyu Kim c
a

Department of Journalism and Communication, Kyung Hee University, 1 Hoegi-dong, Dongdaemun-g, Seoul 130-701, Korea Department of Advertising and Public Relations, The University of Alabama, USA c Department of Advertising and Public Relations, Kyungsung University, Busan, Korea
b

a r t i c l e

i n f o

a b s t r a c t
The current study reconceptualized self-construal as a social cognitive indicator of self-observation that individuals employ for developing and maintaining social relationship with others. From the social cognitive perspective, this study investigated how consumers self-construal can affect consumers electronic word of mouth (eWOM) behavior through two cognitive factors (online community engagement self-efcacy and social outcome expectations) in the context of a social networking site. This study conducted an online experiment that directed 160 participants to visit a newly created online community. The results demonstrated that consumers relational view became salient when the consumers self-construal was primed to be interdependent rather than independent. Further, the results showed that such interdependent self-construal positively inuenced consumers eWOM behavioral intentions through their community engagement self-efcacy and their social outcome expectations. 2012 Elsevier Ltd. All rights reserved.

Article history: Available online 8 February 2012 Keywords: Online community Social networking sites Electronic word of mouth Social cognitive theory

1. Introduction The online brand community has been paid much attention by both marketing practitioners and scholars. With the recent popularity of social networking sites (e.g., Facebook, MySpace, Twitter, etc.), the online brand community has expanded its boundary to the virtually networked society. In such a social networking site, both marketers and consumers can easily create online communities and promote the consumerbrand relationships with current and potential community members (Hung & Li, 2007; Lee et al., 2011; Schau, Muniz, & Arnold, 2009; Leimeister, Sidiras, & Krcmar, 2006). For example, in Facebook, Chevrolet is promoted in online brand communities by marketers (http://www.facebook.com/#!/ chevrolet) and consumers (http://www.facebook.com/#!/pages/ Chevrolet/112714735407038). Consequently, the rapid emergence of the online brand community empowers marketers to encourage community members to voluntarily engage in various community behaviors such as membership intention, recommendation, active participation, and so forth (Algesheimer, Dholakia, & Herrmann, 2005; Casalo, Flavian, & Guinaliu, 2011; Pai & Tsai, 2011). However, not all online brand communities are always successful in motivating consumers to participate in the desirable community activities. Some consumers actively exchange opinions, ideas and information on the brand, product, or company in some online

Corresponding author. Tel.: +82 2 961 9383.

E-mail address: doohwang@gmail.com (D. Lee). 0747-5632/$ - see front matter 2012 Elsevier Ltd. All rights reserved. doi:10.1016/j.chb.2012.01.009

brand communities, but others may just want to observe ongoing discussion and come and go without any active participation, leading the communities to be ghost towns. Establishing active online brand communities should be challenging for both consumers and marketers. One particular approach to promoting the active membership participation is to encourage the community members to advertise the positive benets of the membership to potential members by spreading their words. This line of approach can be termed electronic word of month (eWOM), one of the most reliable sources of information for consumers (Bickart & Schindler, 2001; Brown, Broderick, & Lee, 2007; Hung & Li, 2007). For example, Hung and Li (2007) recognized the possibility that online communities can provide a dynamic, multimedia, and social interactive platform for eWOM. (Brown et al., 2007) also highlighted the importance of eWOM in the setting of online brand communities, suggesting that eWOM can be a major driving force to promote active community membership participation. Similarly, Bickart and Schindler (2001) demonstrated that consumers eWOM behaviors provided high credibility, empathy and relevance to other community members. Because successful online brand communities rely on consumer consumer relationship rather than consumerbrand relationship, online brand communities should promote consumers autonomous eWOM behaviors. Active eWOM participation can be evoked by how consumers see themselves in relation to other members in their online brand communities. This relational view can be termed interdependent self-construal (Cross, Bacon, & Morris, 2000). Previous studies

D. Lee et al. / Computers in Human Behavior 28 (2012) 10541062

1055

have examined the effect of the relational self-concept on various marketing outcomes such as consumerbrand relationships (Swami nathan, Page, & Grhan-Canli, 2007), consumes attitude toward the advertising (Han & Shavitt, 1994), consumers evaluation on brand (Escalas & Bettman, 2005) and purchase behavior (Zhang & Shrum, 2009). In the current context, consumers with interdependent self-construal are likely to engage in cooperative and supportive behavior (van Baaren, Horgan, Chartrand, & Dijkmans, 2004) such as eWOM. Thus, consumers with interdependent selfconstrual are more likely to help other consumers by passing along information on the internet. Consequently, this self-construal should be understood as the internal factor motivating consumers to engage in eWOM behavior. However, little empirical research has examined how consumers interdependent self-construals motivate their active eWOM engagement behavior particularly in a context of social networking sites. Based on Banduras (1991, 2001) social cognitive theory (SCT), this study attempts to explore whether and how interdependent self-construal elicit consumers motivations to engage in eWOM behavior. Specically, from the social cognitive perspective, this study expects that consumers eWOM behaviors for online brand communities will be a function of consumers cognitive judgment on their capabilities to engage in activities for online brand communities (self-efcacy in social cognitive terms) and their assessment of the social benets associated with their eWOM behaviors (social outcome expectations). Further, this study predicts that such community engagement self-efcacy will be directly instigated by consumers interdependent self-construal. 2. Literature review 2.1. eWOM behavior in online brand community According to Hennig-Thurau, Qwinner, Walsh, and Gremler (2004) electronic world of mouth (eWOM) is dened as any positive or negative statement made by potential, actual, or former customers about a product or company, which is made available to a multitude of people and institutions via the Internet (p. 39). The eWOM has long been an important topic among marketing researchers and practitioners (Balasubramanian & Mahajan, 2001; Bickart & Schindler, 2001; Hennig-Thurau et al., 2004). Studies have shown that eWOM is more inuential than traditional marketer-created information because eWOM is perceived to be convenient and have higher credibility (Bickart & Schindler, 2001). Several studies have suggested that consumers are likely to participate in eWOM communication because consumers have concerns for others consumption behavior in an altruistic manner (Dichter, 1966; Engel, Blackwell, & Miniard, 1993; Lee et al., 2011). For example, Dichter (1966) showed that consumers WOM behaviors satised their need to share something important to the other consumers. Also, Engel et al. (1993) reported that consumers passed along information to others because they had a genuine desire to help others make a better purchase decision and save others from negative experiences. Similarly, Lee, Kim, and Kim (2011) investigated consumers attribution style in online brand communities and found that consumers attribution to the communities intrinsic motivations of altruism increased their intention to engage in brand community behaviors including eWOM communication. Consumers perception on online social tie (Sun, Youn, Wu, & Kuntaraporn, 2006), social trust (Hau & Kim, 2011), and similarity with others (Chu & Kim, 2011) have been found to be the important antecedents affecting eWOM behaviors. For example, Sun et al. (2006) found that, online social tie with others was positively related to eWOM behavior because social ties played an important role in the ow of information from consumer to consumer. Previously, strong social ties between consumers were found to be more

inuential than weak ties in shaping eWOM communication process (Sun et al., 2006). Hau and Kim (2011) also reported that social trust toward others signicantly inuenced consumers intention to share information in online community. In a similar fashion, Chu and Kim (2011) also found that similarity with others inuenced ones willingness to provide their message to other consumers. These ndings suggest that consumers perception on the selfother relationship with other members in the virtual community should motivate their active participation in eWOM communication. 2.2. The role of self-construal in online brand community As such, how consumers perceive themselves in relation to other members can be an important individual factor predicting their eWOM behavior. Such an individual sense of self in relation to others can be termed self-construal (Cross et al., 2000; Lewis, Goto, & Kong, 2008; Markus & Kitayama, 1991). Two types of self-construal have been identied: the independent and the interdependent (Markus & Kitayama, 1991). Individuals with independent self-construal are likely to value their individual uniqueness and autonomy because they primarily view themselves as an independent individual entity, as distinct from the group. With this self-concept, individuals essentially perceive themselves separate from others and their independent self-construal becomes salient consequently, those with independent self-construal are likely to pursue their own goals of expressing individuality regardless of any social context (Tramow, Triandis, & Goto, 1991; Ybarra & Tramow, 1998). On the contrary, individuals with interdependent self-construal are likely to value connectedness and group harmony because they predominantly view themselves as part of a larger group. With this self-concept, individuals perceive themselves in connection to others, their interdependent self-construal becomes prominent. Consequently, those with interdependent self-construal tend to seek to achieve their goals of social cohesion in particular social contexts (Tramow et al., 1991; Ybarra & Tramow, 1998). As such, differences in self-construal have distinctive relevance to the self-related processes of how individuals think, feel, and behave regarding their social relationship to others. In particular, interdependent self-construal has been shown to positively inuence individuals cooperative and supportive behaviors (Holland, Roeder, van Baaren, Brandt, & Hannover, 2004; van Baaren et al., 2004). The individuals with an interdependent selfconstrual were also found to have a preference for relationship with others, both psychologically and emotionally (Aron, Aron, & Smollan, 1992; Gardner, Gabriel, & Hochschild, 2002), suggesting that interdependent self-construal are likely to inuence eWOM communication requiring reciprocity between community members. In this sense, it is likely that consumers with interdependent self construal would pass along information to others to help to make a better consumption decision. In fact, several studies have demonstrated that consumers self construal inuences social consequences in the marketing context (Escalas & Bettman, 2005; Han & Shavitt, 1994; Swaminathan et al., 2007; Zhang & Shrum, 2009). For example, Swaminathan et al. (2007) found that the effects of self-concept connection on consumerbrand relationship varied based on consumers selfconstrual. Han and Shavitt (1994) also showed that consumers with interdependent self-construal and independent self-construal responded differently to the advertisements with individualistic versus collectivistic appeals. In a similar fashion, Escalas and Bettman (2005) demonstrated that consumers with independent selfconstrual evaluated the brands more negatively when the brands images are not consistent with in-group. Zhang and Shrum (2009) also revealed that consumers with independent self-construal engaged in more impulsive consumption than ones with interdependent self-construal.

1056

D. Lee et al. / Computers in Human Behavior 28 (2012) 10541062

2.3. A social cognitive perspective on eWOM behavior in online brand community Banduras (1991, 2001) social cognitive theory (SCT) is particularly relevant to this study because it can explain human motivation for eWOM behavior as a function of the reciprocal interactions involving behavior, cognition, and social environmental inuences. In the present context of eWOM behavior, the potential three interdependent factors are consumers eWOM behavior (behavior), their assessment of the likely social consequences of the eWOM behavior and their perceived ability to engage in online brand community activities (cognition), and the technical features of the social networking site that may be able to promote consumers interdependent self-construal (social environmental inuences). In particular, SCT highlights the important functions of humans cognitive abilities. Here, one important cognitive factor is social outcome expectations, dened as ones judgment of the likely social consequences of eWOM behavior, which provide social incentives for enacting such behavior. For example, by recommending their online brand communities to others, consumers expect that they are likely to feel more afliated with their own communities and members. Such social outcome expectations may be shaped by personal experiences (enactive learning in social cognitive terms) associated eWOM behaviors, and by their observation of the eWOM behaviors enacted by others (vicarious learning), or by those persuasive promotion messages found in online brand communities. In fact, individuals tend to get involved in a variety of community behaviors because they expect meaningful social support from the community, In particular, previous studies have found that many of the traditional benets of social belonging, social support and social acceptance are associated with online communities (Cummings & Sproull, 2002; McKenna & Bargh, 1998) and identied social interactions as the primary motivation for consumers engagement with online brand community (Algesheimer et al., 2005; Bagozzi & Dholakia, 2006; Dholakia, Bagozzi, & Pearo, 2004). For example, Bagozzi and Dholakia (2006) argued that a brand community is a group of consumers with a well-developed social identity, whose members engage jointly in group actions to accomplish collective goals. Dholakia et al. (2004) also suggested that consumers identication with the online brand community is closely associated with the community consumers social need for group belonging especially as the consumer views himself or herself as a member of the group. Consequently, when consumers actively participate in online brand communities, consumers are likely to expect desirable social outcomes. From this perspective, social support from online brand communities should be considered as perceived social outcomes through which ones social needs are satised. In SCT, self-efcacy is another important cognitive factor motivating a course of human behavior. Dened as the belief in ones capabilities to organize and execute the courses of action required to produce given attainments (Bandura, 1997), p. 3, it reects the degree to which individuals are condent about executing a behavior with the skill they have rather than the measures of the skill that they possess. For example, when consumers are condent in engaging in certain behaviors for the online brand community, they are likely to be motivated to participate in eWOM behaviors and promote the community. In this sense, those who perceive themselves to be highly efcacious with reference to a particular eWOM communication will be more likely to invest sufcient levels of effort to achieve successful outcomes for online brand communities than those with low levels of self-efcacy. Self-efcacy has been empirically examined in relation to perceived social outcome expectations as well. Such relationship demonstrates the causal relationship among the cognitive constructs within SCT. That is, the stronger a persons self-efcacy judgment

on community activities, the more likely he or she is to attempt to achieve the desired social outcomes, whereas the weaker a persons self-efcacy, the less likely he or she is to try to achieve the desired social outcomes. For example, when individuals feel selfefcacious about their abilities to promote their communities, they tend to expect meaningful social support from the communities. In online setting, several studies have also demonstrated that both self-efcacy and social outcome expectations have been positively associated with each other e.g., (LaRose & Eastin, 2004) as well as a variety of online behaviors e.g., (LaRose & Rifon, 2007; Lee, LaRose, & Rifon, 2008). For example, Eastin (2005) showed that Internet self-efcacy is positively related with positive social expectancy which, in turn, generated a positive relationship with social Internet use among teens. In the context of online communities, Hsu, Ju, Yen, and Chang (2007) demonstrated that social outcome expectancy has a positive effect on members behavior such as knowledge sharing in online communities. 2.4. A social cognitive perspective on the role of self-construal in eWOM behavior From the social cognitive perspective, self-construal might be construed as a social cognitive indicator of self-observation that individuals employ for developing and maintaining social relationship with others (Bandura, 1991). In particular, when the interdependent self is activated, such activation of interdependent self construal allows consumers to actively observe themselves in online brand communities (self-observation in social cognitive terms) and to judge their behaviors against personal and social norms in the social context (normative judgment). In the social cognitive process, consumers may become attentive to their self-reactive capabilities to engage in eWOM behaviors for online brand communities that are particularly relevant to social interactions among the community members (self-reactive inuence). Consequently, such selfreactive incentives motivate consumers with interdependent selves to participate in eWOM communication with other members in online brand community. This reasoning is consistent with Cross and Madsons (1997) claim that individuals with interdependent selfconstrual attempt to connect with others through social interactions such as helping others. Applying this social cognitive perspective to eWOM behavior in relation to online brand communities, we expect that consumers with interdependent self-construal are likely to have greater intrinsic interest and deep engrossment in online brand communities so that they can heighten and sustain their efforts to make the communities successful. In this process, such self-concept would motivate consumers to set and accomplish their goals by instigating their perceived self-efcacy to engage in behaviors for online brand communities. Consequently, we predict that consumers interdependent self-construal will increase their consumers community engagement self-efcacy, which in turn will affect their eWOM behaviors for the online brand community directly as well as indirectly through their social outcome expectations associated with the particular behavior for the online brand community. Based on the reasoning, a proposed model of the current study is presented in Fig. 1 and the following hypotheses are formally stated: Hypothesis 1. Interdependent self-construal priming will generate higher online brand community engagement self-efcacy than will independent self-construal priming.

Hypothesis 2. Consumers online brand community engagement self-efcacy will positively affect consumers eWOM intention.

D. Lee et al. / Computers in Human Behavior 28 (2012) 10541062

1057

Fig. 1. Hypothesized model of eWOM engagement behavior.

Hypothesis 3. Consumers online brand community engagement self-efcacy will positively affect their social outcome expectations associated with eWOM behavior. Hypothesis 4. Consumers social outcome expectations will positively affect consumers eWOM intention. 2.5. Community type as a potential moderator in the effects of selfconstrual Both marketers and consumers are empowered to create virtual brand communities and conveniently manage consumerbrand relationships on social networking sites. Marketer-created communities can be strategically used to promote particular products and services to current and potential members, and consumer-created communities can be used to lead members to voluntarily engage in various community behaviors such as membership intention, recommendation, active participation, and so forth (Hung & Li, 2007; Lee et al., 2011; Schau et al., 2009; Leimeister et al., 2006). As such, online brand communities on social networking Web sites serve as one of the primary brand promotion platforms for both marketers and consumers. From Banduras (1991, 2001) social cognitive perspective, the different types of online brand communities may be construed as an environmental factor that could differently affect the likelihood that consumers will engage in future eWOM behaviors as well as their expectations regarding consequences associated with particular behaviors. That is, the different environmental agent can instigate consumers to reect on how their future eWOM behaviors will bring about positive and negative outcomes in their everyday lives and to employ their self-reactive strategies to voluntarily participate in eWOM behaviors in the future. In the current context, it is expected that marketer- and consumer-created communities provide different levels of social incentives that promote consumerbrand relationships, in turn motivating different levels of eWOM behaviors. For example, when consumers browse consumer-created online brand communities, they are more likely to expect intrinsic motives of altruism than extrinsic motives of prot exploitation from the communities. Conversely, when consumers experience marketer-created online brand communities, their expectation regarding the particular communities should be the other way around. As a result, their expectations of positive and negative outcomes evoked by the different types of online brand communities will likely motivate them to engage in different levels of eWOM behaviors. As such, the community type may differentiate the social cognitive process in which eWOM behaviors can be guided by the environmental factor that motivates consumers to exercise

forethought, reect on the particular behavior, and apply self-reactive inuences (Bandura, 2001). Indeed, Sung, Kim, Kwon, and Moon (2010) classied online brand communities into consumer- and marketer-generated brand communities and found that consumers community participation intention was related to different motives depending on the community type. That is, consumers consumergenerated community participation intention was signicantly related to the motives of brand likability and interpersonal utility (e.g., social relationship), whereas their marketer-generated community participation intention was signicantly associated with the motives of brand liability, convenience seeking, and incentive seeking (e.g., coupon). Similarly, Lee et al. (2011) expected that different types of communities would generate different levels of motives and demonstrated that the consumer-created online brand community generated stronger intrinsic motives of altruism than the marketer-created online brand community, thus leading to consumers higher intention to engage in online brand community participation behaviors. These ndings suggest that the community type should be regarded as an important environmental variable that determines the levels of eWOM participation behaviors. It is important to note that in the current context, the community type may also moderate how consumers self-construal operates in their eWOM behaviors. That is, consumers with interdependent self-construal are more likely to be activated when the community is created by consumers than when it is created by marketers. As self-construal is heightened, such self-concept is more likely to instigate consumers to associate the community with intrinsic motives of altruism to help other existing or would-be members share their knowledge about the brand or product with each other. In the social cognitive process, consumers interdependent self-construal will increase the likelihood that they will engage in eWOM behaviors indirectly through their social outcome expectations associated with the particular behavior for the online brand community as well as their self-efcacy belief about community engagement. Consequently, consumers with interdependent self-construal are willing to perform eWOM behaviors to promote the community and the brand, especially when they are exposed to a consumer-created online brand community. Conversely, it is also expected that the effects of interdependent self-construal on eWOM behaviors may lessen when consumers are exposed to marketer-created online brand communities. That is, when the community is created by marketers, consumers interdependent self-construal is not likely to be activated as much. In this situation, even consumers with interdependent self-construal tend to associate the marketers effort to create the online brand community with the marketers extrinsic motives to increase sales or enhance the image of the corporation. In this process, consumers interdependent self-construal is not likely to increase their community engagement self-efcacy and social outcome expectations

1058

D. Lee et al. / Computers in Human Behavior 28 (2012) 10541062

associated with eWOM behaviors. As a result, consumers in marketer-created online brand communities are not likely to voluntarily perform eWOM behaviors that promote the community or brand even when their interdependent self-construal is activated. Based on this reasoning, the moderating role of the community type is hypothesized as follows. Hypothesis 5. The effect of interdependent self-construal on eWOM behavioral intention will be moderated by the community type.

Hypothesis 5a. The effect of interdependent self-construal will be more salient when the online brand community is created by consumers. Hypothesis 5b. The effect of interdependent self-construal will disappear when the online brand community is created by marketers.

3. Methods 3.1. Research design, participants and procedures The present study used a 2 2 between-subjects factorial design for an online experiment. The rst between-subject factor was two levels of self-construal (independent versus interdependent) and the second between-subject factor was two types of online brand communities (consumer versus marketer-created). As a result, four experimental conditions were created. For the online experiment, 160 college students were recruited at a large Southern university in the United States. All participants were directed to visit a newly created online brand community on Facebook and 40 participants were randomly assigned to each of the four stimulus Web sites. The stimulus Web site demonstrates how Facebook users can access, read, and postopinions about a ctitious digital camera brand and product, Carleton Digital Camera. Content for the site was drawn and adapted from existing Facebook digital camera online brand communities. Upon visiting the stimulus Web site, the participants were rst asked to read a short story for the self-construal priming in advance. After reading the story, they were asked to go to the next page and browse the stimulus online brand community. When they nished browsing the community, they were asked to ll out a follow-up online survey questionnaire. Most participants were Caucasian (88.8%) and female (61.2%). Subjects median age was 20 years. 3.2. Stimulus material 3.2.1. Self-construal priming A self-construal priming story was presented before the participants were asked to browse the stimulus Facebook online brand community. Adopted (Tramow et al., 1991), the manipulation asked all participants to read a story concerning a general named Sostoras who had to choose a warrior to send to the king. In the independent priming condition, the story ended when the general chose the warrior on the basis of his individual merits. In the interdependent priming condition, however, the story ended when the general chose the warrior on the basis of his familys well being. Specically, rst, the priming story had the same rst two paragraphs: Sostoras, a warrior in ancient Sumer, was largely responsible for the success of Sargon I in conquering all of Mesopotamia. As a result, he was rewarded with a small kingdom of his own to rule. About 10 years later, Sargon I was conscripting warriors for a new war. Sostoras was obligated to send a detachment of soldiers

to aid Sargon I. He had to decide who to put in command of the detachment. When the participants were assigned in the independent self-construal priming condition, they continued to read the following story: After thinking about it for a long time, Sostoras eventually decided on Tiglath who was a talented general. This appointment had several advantages. Sostoras was able to make an excellent general indebted to him. This would solidify Sostorass hold on his own dominion. In addition, the very fact of having a general such a Tiglath as his personal representative would greatly increase Sostorass prestige. Finally, sending his best general would be likely to make Sargon I grateful. Consequently, there was the possibility of getting rewarded by Sargon I. However, when the participants were in the interdependent self-construal priming condition, they continued to read the following story: After thinking about it for a long time, Sostoras eventually decided on Tiglath who was a member of his family. This appointment had several advantages. Sostoras was able to show his loyalty to his family. He was able to cement their loyalty to him. In addition, having Tiglath as the commander increased the power and prestige of the family. Finally, if Tiglath performed well, Sargon I would be indebted to the family. This manipulation has been shown to signicantly alter self-descriptions in the expected fashion in numerous prior studies (Gardner et al., 2002; Howard, Gardner, & Thompson, 2007; Tramow et al., 1991). 3.2.2. Online brand community type Two versions of the online brand community were created to reect two types of the Facebook online brand community (consumer- versus marketer-created online brand community). The type was manipulated by the creator and manager of the community through four prototypical pages of Facebook community Info, Wall, Photos, and Discussions. Specically, the info page of the marketer-created online brand community indicated that This community is administrated by Carleton Camera Inc. with descriptive statement, Carlton is the company that manufactures digital cameras. Carlton Company created this community for sharing ideas and stories of Carleton digital cameras with Facebook users. In the same fashion, the consumer-created online brand community indicated a ctitious name David Smith on whole creator information spaces. All features and contents were identical in the two versions of the online brand community, except for the manipulations of the community type. 3.3. Operational measures Consumers online brand community eWOM intention was considered the primary dependent variable. Consumers social outcome expectations and online community engagement self-efcacy were considered the two intervening variables that mediate the effects of self-construal on consumers eWOM intention. 3.3.1. Manipulation check for interdependent self-construal priming Four items were used to check the effectiveness of the manipulation, on the 7-point Likert-type scale with 1 representing strongly disagree and 7 representing strongly agree. Specically, after participants read the short story for the self-construal priming and browsed the online brand community, participants were asked to respond to the following statements: (1) my happiness depends on the happiness of those around me, (2) it is important for me to maintain harmony with my group, (3) i will sacrice my selfinterest for the benet of the group I am in, (4) i often have the feeling that my relationships with other are more important than my own accomplishments, and (5) it is important to me to respect decisions made by the group. This scale exhibited an acceptable degree of internal consistency (M = 4.83, SD = .99, Cronbachs alpha = .75).

D. Lee et al. / Computers in Human Behavior 28 (2012) 10541062

1059

3.3.2. Consumers eWOM intentions Three 7-point Likert-type items measured eWOM intention, with 1 representing very unlikely and 7 representing very likely. Specically, the items asked subjects to indicate how likely or unlikely they would be to participate in the following online brand community activities in future: (1) saying positive things about the online brand community to other people in future, (2) recommending this online brand community to anyone who seeks advice about the brand in future, (3) encouraging other people to join the online brand community in future. This scale exhibited a high degree of internal consistency (M = 4.49, SD = 1.32, Cronbachs alpha = .90). 3.3.3. Consumers social outcome expectations Four 7-point Likert-type items measured social outcome expectations, with 1 representing very unlikely and 7 representing very likely. Specically, the four items asked the participants to indicate how likely or unlikely they would be to (1) maintain a relationship I value in the community, (2) feel afliated with the community, (3) feel attached to the community, and (4) feel like I belong to the community. This scale showed a high degree of internal consistency (M = 4.21, SD = 1.32, Cronbachs alpha = .95). 3.3.4. Consumers online community engagement self-efcacy Five 7-point Likert-type items measured perceived efcacy that the online brand community will generate, with 1 representing strongly disagree and 7 representing strongly agree. Specically, the ve items asked the participants to indicate their perceived abilities to engage in online brand community activities, such as (1) posting new information about the online brand community, (2) supporting the community members opinions, (3) encouraging other people to join the online brand community, (4) recommending the online brand community to anyone who seeks advice about the brand, and (5) reaching the online brand communitys goals. This scale showed a high degree of internal consistency (M = 5.63, SD = .95, Cronbachs alpha = .91).

4.3. The direct and indirect effects of interdependent self-construal on consumers eWOM intentions Using AMOS 6, this study performed a structural equation modeling (SEM: Bollen, 1989) to further test the hypothesized direct and indirect relationships between all the variables simultaneously in the proposed model. Correlation matrix of all variables used to test the model is shown in Table 1. As hypothesized, consumers eWOM communication is positively related to interdependent self-construal, (r = .19) online brand community self-efcacy (r = .33), and social outcome expectations (r = .49). The highest correlation coefcient among these relationships was .49, suggesting that no serious multicollinearity problems were found in the variables. The t indices of the initial path analysis indicated that the path model was consistent with the data (AGFI = .947, NFI = .954, RMSEA = .067, v2 = 3.4 (df = 2), p = .18). As expected, the path model supported H1 (b = .22, p < .01), conrming the main effect of the interdependent self-construal on community engagement self-efcacy. In addition, a two-way analysis of variance (ANOVA) analysis also conrmed the signicant main effect of independent self-construal on consumers online brand community engagement self-efcacy in the hypothesized direction, with the interdependent self-construal priming (M = 5.85, SD = .84) engendering higher community engagement self-efcacy than the independent self-construal priming (M = 5.42, SD = 1.00), F = 8.27, p < 0.01, g2 = .05, thus providing support for H1. The path model also found that the self-efcacy affected consumers online brand community engagement intentions directly (b = .25, p < .01) as well as indirectly through social outcome expectations associated with eWOM behaviors. In the indirect process, self-efcacy inuenced social outcome expectations (b = .22, p < .01), which, in turn, also affected consumers eWOM intentions (b = .43, p < .001). As a result, the ndings of the path analysis supported H1, H2, H3, and H4. As recommended by Holbert and Stephenson (2003), the direct, indirect and total effects of the variables are examined as well. First, interdependent self-construal was found to have signicant effects on dependent variables either directly or indirectly. The variable had direct effects on community engagement self-efcacy (.22), but it had only indirect effects on social outcome expectations (.06) and eWOM behavioral intention (.07). Therefore, self-construal accounted for all total effects on self-efcacy (.22) directly and on social outcome expectations (.06) and eWOM behavioral intention (.07) indirectly. Self-efcacy also had substantial direct effects on social outcome expectations (.25) and eWOM intention (.22). It was also found that self-efcacy had signicant indirect effects on eWOM behavioral intention (.11), which accounted for about one third of its total effect on that variable (.33). As the SCT suggested, social outcome expectations was also found to be another important variable in this study as well. The variable accounted for the total effect on eWOM intention (.43) directly. Overall, 28% of the total variance of eWOM intention was jointly explained by all the vari-

4. Results 4.1. Manipulation check for interdependent self-construal priming A two-way analysis of variance (ANOVA) analysis revealed a statistically signicant effect for the manipulation, F (1, 158) = 4.73, p < .05, with the interdependent self-construal priming (M = 4.99, SD = .97) engendering higher interdependent self-construal perception than the independent self-construal priming (M = 4.66, SD = .98). The results indicate the success of the manipulation. 4.2. Effect of interdependent self-construal on eWOM intention A two-way analysis of variance (ANOVA) analysis found that interdependent self-construal and community type had a positive main effect on consumers eWOM intentions while controlling for the effect of community type. Specically, the results demonstrated that the interdependent self-construal priming (M = 4.73, SD = 1.25) generated higher eWOM intentions than the independent self-construal priming (M = 4.25, SD = 1.36), F = 5.68, p < 0.05, g2 = .04. In addition, the results showed that the consumer-created online brand community (M = 4.71, SD = 1.25) engendered higher eWOM intentions than the marketer-created online brand community (M = 4.26, SD = 1.36), F = 4.93, p < 0.05, g2 = .03. However, there was no strong statistical evidence that community type moderated the main effect of the interdependent self-construal (p > .1). Therefore, H5 was not supported.

Table 1 Zero-order correlation, means, and standard deviation (N = 160). Variable 1. 2. 3. 4. eWOM behavioral intention Social outcome expectations Self-efcacy Self-construal 1 1 .49** .33** .19* 2 1 .25** .04 3 4 M 5.63 4.21 4.49 .5 SD .95 1.32 1.32 .5

1 .23**

Note: Self-construal is coded into independent (coded 0) and interdependent selfconstrual (coded 1). * p < .05. ** p < .01.

1060

D. Lee et al. / Computers in Human Behavior 28 (2012) 10541062

Fig. 2. Final path model of eWOM engagement behavior.

ables in the proposed model. Fig. 2 shows the nal model of the present study.

5. Discussion The results of the present study empirically demonstrated how consumers self-construal instigated their motivations to participate in eWOM behavior through social cognitive variables of selfefcacy and social outcome expectations. Specically, consumers community engagement self-efcacy was found to have a signicant inuence on their eWOM behavior intentions directly and indirectly through their cognitive assessment of the potential social outcomes associated with the behaviors for their online brand communities. Further, the results also demonstrated that the social cognitive process of eWOM behavior was strongly instigated when consumers self-construal was primed to be interdependent rather than independent. Consequently, the results conrmed the causal linkage between interdependent self-construal, community engagement self-efcacy, social outcome expectations, and eWOM intentions from the social cognitive perspective. As such, this study offers a new theoretical perspective on consumers eWOM communication because it highlights self-construal as one important construct in evoking the self-regulatory mechanisms of eWOM behavior in social networking sites. That is, based on Banduras SCT (Bandura, 1991, 1997), self-construal was reconceptualized as a social cognitive indicator of self-observation that may differently instigate the self-regulatory behavior associated with eWOM depending on how consumers view themselves in relation to other community members. In particular, when consumers interdependent self-construal is activated, consumers are motivated to actively observe themselves with other members eWOM behaviors in the online community (self-observation), to judge their own eWOM behaviors against personal and social norms in the context of social networking sites (normative judgment), and to employ their self-reactive strategies to voluntarily participate in eWOM behaviors (self-reactive inuences). Consequently, interdependent self-construal can play a pivotal role in initiating consumers to connect with other members by promoting personal, social, and self-reactive incentives. The results of the current study also have practical implications for marketers, who should recognize the importance of consumers eWOM behavior. However, it may be also true that marketers may not realize how they can promote community members eWOM communication. For example, consumers may quickly lose interest in passing along information to others unless their interdependent self-construal becomes salient. If that is the case, consumers may

be less likely to perceive social benets and lose eWOM intention. From a practical point of view, online brand communities should be able to increase consumers eWOM engagement behaviors by encouraging their members interdependent self-construal. In order for this strategy to be successful, online brand communities should be perceived as interdependent rather than independent communities. Identifying and developing the features and services that promote consumers interdependent self-construal would be valuable for marketing practitioners as they create online brand communities. One possible approach that online brand communities can employ is to develop and utilize interactive message cues that can easily prime members to interdependent self-construal, such as by using the pronouns we, our, ourselves, and ours in the community platform could easily and effectively prime such interdependent self-construal. Indeed, several studies suggested that interdependent self-construal can be primed by using those pronouns (Stapel & Koomen, 2001). Another possibility is to use visual images to stimulate group members interdependent self-construal. For example, when Americans are exposed to their national ag, their interdependent selfconstrual is likely to be activated (Briley & Wyer, 2002). In this sense, the visual images representing particular groups in the online community can be used to instigate the group members interdependent self-construal. As such, community group members identity can be strategically employed to induce their interdependent self-construal because it is an integral part of their self-construal (Druckman, 1994). Calling attention to this group identity can generate ones desire and behavior to promote group interests for individuals of either individualistic or collectivist cultures (Davis, 1999). These simple strategies can successfully promote active eWOM behaviors among the community members in the online brand communities. Given that eWOM behavior is regarded as the most reliable and credible communication strategy for consumers, such message strategy will positively inuence consumers product choice (Lewis et al., 2008) and post-purchase product perceptions (Bone, 1995). Notably, this study also demonstrated that the community type created for the manipulation signicantly inuenced consumers eWOM behavioral intentions regardless of the two types of self-construal (interdependent versus independent). Specically, consumers eWOM intention was higher in the consumer-created online brand community than in the marketer-created online brand community, and the effect size of interdependent self-construal on eWOM behavioral intention did not disappear even when the brand community was created by marketers. The results suggest that whether consumers are primed to be either interdependent or inde-

D. Lee et al. / Computers in Human Behavior 28 (2012) 10541062

1061

pendent, the consumer-created community tend to instigate consumers to perceive the community as having intrinsic motives of altruism whereas the marketer-created community likely to trigger consumers to believe that the community has the extrinsic motive of prot exploitation. Consequently, consumers different attribution style to the community type may have generated the different level of eWOM behavior regardless of the types of self-construal. This explanation is consistent with the argument of Lee et al. (2011) that consumer-created online communities increase community engagement behaviors through consumers attribution of altruism. Several limitations from the current research warrant further investigation. First, this study employed college students as participants, which might limit the generalizability of the results to other user populations of social networking Web sites. Although college students are primary users of social networking Web sites, college students do not represent all consumers who participate in online brand communities of the sites. Another limitation of the study may be the product category of the stimulus online brand community (digital camera), which can be classied as the hedonic product, or something that is pleasant, fun, or appeals to consumers senses (e.g., perfume) (Dhar & Wertenbroch, 2000). Because a hedonic product is more likely to serve a value-expressive, socially visible function, the related online brand community might be more appealing to those who are interdependent rather than independent. Online brand communities related to utilitarian products, dened as something that is practical, functional, or helps consumers achieve a goal (e.g., pain killer), might be more persuasive to those who are independent rather than interdependent. Future researchers should investigate the possible different social cognitive process of eWOM behavior involving different types of products in the online brand community setting. It is also important to note that the concept of self-construal is closely related with different cultures (Singelis & Sharkey, 1995). Indeed, individual goals tend to be emphasized over group goals in individualistic cultures, whereas group goals are likely to be salient over individuals goals in collectivistic cultures. For example, consumers in several Asian countries e.g., (Han & Shavitt, 1994; Markus & Kitayama, 1991) and the Mediterranean region e.g., (Oyserman, Coon, & Kemmelmeier, 2002) as well as Hispanics e.g., (Oyserman et al., 2002) are likely to show interdependent self-construal. Given that the current study was conducted in the United States (an individualistic culture), future studies researchers should explore the possible different processes of eWOM behavior in relation to consumers self-construal in other countries representing collectivistic culture. Finally, this study employed eWOM intention instead of actual eWOM behavior as an outcome variable because it is assumed that a behavior is largely determined by the intention to perform the behavior (Ajzen & Fishbein, 1977). However, it should be noted that such relationship is often limited, especially in the setting of computer-mediated communication where consumers eWOM behaviors may lead to negative consequences of privacy violations, such as spam attacks, identify theft, or phishing. Indeed, the potential risk of privacy violations has been found to deter consumers willingness to engage in online community behaviors (Krasnova, Spiekermann, Koroleva, & Hildebrand, 2010; Posey, Benjamin, Roberts, & Ellis, 2010). Since the current study did not directly measure the participants eWOM behaviors, the possible discrepancy between eWOM intentions and actual eWOM behaviors may have limited the validity of the ndings. Therefore, it is recommended that future research should investigate the possible effect of consumers self-construal on eWOM behaviors in a more realistic setting.

References
Ajzen, I., & Fishbein, M. (1977). Attitudebehavior relations: A theoretical analysis and review of empirical research. Psychological Bulletin, 84, 888918. Algesheimer, R., Dholakia, U. M., & Herrmann, A. (2005). The social inuence of brand community: Evidence from European car clubs. Journal of Marketing, 69, 1934. Aron, A., Aron, E. N., & Smollan, D. (1992). Inclusion of other in the self scale and the structure of interpersonal closeness. Journal of Personality and Social Psychology, 63, 596612. Bagozzi, R. P., & Dholakia, U. M. (2006). Antecedants and purchase consequences of customer participation in small group brand communities. International Journal of Research in Marketing, 23, 4561. Balasubramanian, S., & Mahajan, V. (2001). The economic leverage of the virtual community. International Journal of Electronic Commerce, 5, 103138. Bandura, A. (1991). Social cognitive theory of self-regulation. Organizational Behavior and Human Decision Processes, 50, 248287. Bandura, A. (1997). Self-efcacy: The exercise of control. New York: Freeman. Bandura, A. (2001). Social cognitive theory: An agentive perspective. Annual Review of Psychology, 52, 126. Bickart, B., & Schindler, R. M. (2001). Internet forums as inuential sources of consumer information. Journal of Interactive Marketing, 15, 3140. Bollen, K. (1989). Structural equations with latent variables. New York: John Wiley & Sons. Bone, P. F. (1995). Word-of-mouth effects on short-term and long-term product judgments. Journal of Business Research, 32, 213223. Briley, D. A., & Wyer, R. S. (2002). The effect of group membership salience on the avoidance of negative outcomes: Implications for social and consumer decisions. Journal of Consumer Research, 29, 400415. Brown, J., Broderick, A., & Lee, N. (2007). Word of mouth communication within online communities: Conceptualizing the online social network. Journal of Interactive Marketing, 21, 220. Casalo, L. V., Flavian, C., & Guinaliu, M. (2011). Understanding the intention to follow the advice obtained in an online travel community. Computers in Human Behavior, 27, 622633. Chu, S. C., & Kim, Y. (2011). Determinants of consumer engagement in electronic word-of-mouth (eWOM) in social networking sites. International Journal of Advertising, 30, 4775. Cross, S. E., Bacon, P. L., & Morris, M. L. (2000). The relational interdependent selfconstrual and relationships. Journal of Personality and Social Psychology, 78, 791808. Cross, S. E., & Madson, L. (1997). Models of the self: Self-construals and gender. Psychological Bulletin, 122, 537. Cummings, J. N., Sproull, L., & Kiesler, S. (2002). Beyond hearing: Where the realworld and online support meet. Group Dynamics, 6, 7888. Davis, T. C. (1999). Revisiting group attachment: Ethnic and national identity. Political Psychology, 20, 25. Dhar, R., & Wertenbroch, K. (2000). Consumer choice between hedonic and utilitarian goods. Journal of Marketing Research, 37, 6071. Dholakia, U. M., Bagozzi, R. P., & Pearo, L. (2004). A social inuence model of consumer participation in network and small-group-based virtual communities. International Journal of Research in Marketing, 21, 241263. Dichter, E. (1966). How word-of-mouth advertising works. Harvard Business Review, 44, 147157. Druckman, D. (1994). Nationalism, patriotism, and group loyalty: A social psychological perspective. International Studies Quarterly, 38, 43. Eastin, M. S. (2005). Teen internet use: Relating social perceptions and cognitive models to behavior. Cyberpsychology and Behavior, 8, 6275. Engel, J. F., Blackwell, R. D., & Miniard, P. W. (1993). Consumer behavior (8th ed.). Fort Worth: Dryden Press. Escalas, J. E., & Bettman, J. R. (2005). Self-construal, reference groups, and brand meaning. Journal of Consumer Research, 32, 378389. Gardner, W. L., Gabriel, S., & Hochschild, L. (2002). When you and I are we, you are not threatening: The role of self-expansion in social comparison. Journal of Personality and Social Psychology, 82, 239251. Han, S., & Shavitt, S. (1994). Persuasion and culture: Advertising appeals in individualistic and collectivistic societies. Journal of Experimental Social Psychology, 30, 326350. Hau, Y. S., & Kim, Y. G. (2011). Why would online garners share their innovationconducive knowledge in the online game user community? Integrating individual motivations and social capital perspectives. Computers in Human Behavior, 27, 956970. Hennig-Thurau, T., Qwinner, P. K., Walsh, G., & Gremler, D. D. (2004). Electronic word-of-mouth via consumer-opinion platforms: What motivates consumers to articulate themselves on the Internet? Journal of Interactive Marketing, 18, 3852. Holbert, R. L., & Stephenson, M. T. (2003). The importance of indirect effects in media effects research: Testing for mediation in structural equation modeling. Journal of Broadcasting & Electronic Media, 47, 556572. Holland, R. W., Roeder, R. W., van Baaren, R. B., Brandt, A., & Hannover, B. (2004). Dont stand so close to me: Self-construal and interpersonal closeness. Psychological Science, 15, 237242. Howard, E. S., Gardner, W. L., & Thompson, L. L. (2007). The role of the self-concept and the social context in determining the behavior of power holders:

1062

D. Lee et al. / Computers in Human Behavior 28 (2012) 10541062 Pai, P. Y., & Tsai, H. T. (2011). How virtual community participation inuences consumer loyalty intentions in online shopping contexts: An investigation of mediating factors. Behavior and Information Technology, 30, 603615. Posey, C., Benjamin, L. P., Roberts, T. L., & Ellis, T. S. (2010). Proposing the online community self-disclosure model: The case of working professionals in France and the UK who use online communities. European Journal of Information Systems, 19, 181195. Schau, H. J., Muniz, A., & Arnold, E. J. (2009). How brand community practices create value. Journal of Marketing, 73, 3051. Singelis, T. M., & Sharkey, W. F. (1995). Culture, self-construal, and embarrassability. Journal of Cross-Cultural Psychology, 26, 622644. Stapel, D. A., & Koomen, W. (2001). I, we, and the effects of others on me: How selfconstrual level moderates social comparison effects. Journal of Personality and Social Psychology, 80, 766781. Sun, T., Youn, S., Wu, G., & Kuntaraporn, M. (2006). Online word-of-mouth (or mouse): An exploration of its antecedents and consequences. Journal of Computer-Mediated Communication, 11, 11041127. Sung, Y., Kim, Y., Kwon, O., & Moon, J. (2010). An explorative study of Korean consumer participation in virtual brand communities in social network sites. Journal of Global Marketing, 23, 430445. Swaminathan, V., Page, K. L., & Grhan-Canli, Z. (2007). My brand or our brand: The effects of brand relationship dimensions and self-construal on brand evaluations. Journal of Consumer Research, 34, 248259. Tramow, D., Triandis, H. C., & Goto, S. G. (1991). Some tests of the distinction between the private and the collective self. Journal of Personality and Social Psychology, 60, 649655. van Baaren, E. R., Horgan, T. G., Chartrand, T. L., & Dijkmans, M. (2004). The forest, the trees, and the chameleon: Context dependence and mimicry. Journal of Personality and Social Psychology, 86, 453459. Ybarra, O., & Tramow, D. (1998). How priming the private self or collective self affects the relative weights of attitudes and subjective norms. Personality and Social Psychology Bulletin, 24, 362370. Zhang, Y., & Shrum, L. J. (2009). The inuence of self-construal on impulsive consumption. Journal of Consumer Research, 35, 838850.

Self-construal in intergroup versus dyadic dispute resolution negotiations. Journal of Personality and Social Psychology, 93, 614631. Hsu, M.-H., Ju, T. L., Yen, C.-H., & Chang, C.-M. (2007). Knowledge sharing behavior in virtual communities: The relationship between trust, self-efcacy, and outcome expectations. International Journal of Human-Computer Studies, 65, 153169. Hung, K. H., & Li, S. Y. (2007). The inuence of eWOM on virtual consumer communities: Social capital, consumer learning, and behavioral outcomes. Journal of Advertising Research, 47, 485495. Krasnova, H., Spiekermann, S., Koroleva, K., & Hildebrand, T. (2010). Online social networks: Why we disclose. Journal of Information Technology, 25, 109125. LaRose, R., & Eastin, M. S. (2004). A social cognitive theory of internet uses and gratications: Toward a new model of media attendance. Journal of Broadcasting and Electronic Media, 48, 358377. LaRose, R., & Rifon, N. J. (2007). Promoting i-safety: Effects of privacy warnings and privacy seals on risk assessment and online privacy behavior. Journal of Consumer Affairs, 41, 127149. Lee, D., Kim, H., & Kim, J. K. (2011). The impact of online brand community type on consumers community engagement behaviors: Consumer-created vs. marketer-created online brand community in online social networking web sites. Cyberpsychology, Behavior, and Social Networking, 14, 5963. Lee, D., LaRose, R., & Rifon, N. J. (2008). Keeping our network safe: A model of online protection behavior. Behavior and Information Technology, 27, 445454. Leimeister, J. M., Sidiras, P., & Krcmar, H. (2006). Exploring success factors of virtual communities: The strategic implications of virtual communities of consumption. Journal of Organizational Computing and Electronic Commerce, 16, 277298. Lewis, R. S., Goto, S. G., & Kong, L. (2008). Culture and context: East Asian American and European American differences in P3 event-related potentials. Personality and Social Psychology Bulletin, 34, 623634. Markus, H. R., & Kitayama, S. (1991). Culture and the self: Implications for cognition, emotion, and motivation. Psychological Review, 98, 224253. McKenna, K., & Bargh, J. A. (1998). Coming out in the age of the Internet: Identity demarginalization through virtual group participation. Journal of Personality and Social Psychology, 75, 681694. Oyserman, D., Coon, H. M., & Kemmelmeier, M. (2002). Rethinking individualism and collectivism: Evaluation of theoretical assumptions and meta-analyses. Psychological Bulletin, 128, 372.