THE EFFECTIVENESS OF CORPORATE SOCIAL RESPONSIBILITY (CSR) CAMPAIGNS ON CONSUMER RESPONSES TO BRANDS IN SOCIAL MEDIA: IMPRESSION MANAGEMENT PERSPECTIVES By Hyun Ju Jeong A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Media and Information Studies 2011 ABSTRACT THE EFFECTIVENESS OF CORPORATE SOCIAL RESPONSIBILITY (CSR) CAMPAIGNS ON CONSUMER RESPONSES TO BRANDS IN SOCIAL MEDIA: IMPRESSION MANAGEMENT PERSPECTIVES By Hyun Ju Jeong The purpose of the present study was to examine the effectiveness of two Corporate Social Responsibility (CSR) campaigns—the Cause-Related Marketing (CRM) campaign and the Corporate Sponsorship (CS) campaign—on the favorable response of consumers to the brands in a social media setting, such as Facebook. From the perspective of impression management theory, the current study suggested that consumers would be mobilized to present socially favorable images to their friends by participating in joining and inviting activities on a social media brand page featuring the CSR campaigns. Furthermore, the present study integrated the roles of friends and brand image into the impression management context in social media. A 3 (the CSR campaign condition: CRM campaign, CS campaign, or control) × 2 (the type of self-friend gender composition: same versus different) × 2 (the type of brand: symbolic versus practical) between-subjects experiment design was employed. The dependent variables were the consumer intentions to join the brand page and invite friends to the brand page in social media. The outcome expectancy of consumers to look favorable to their friends was explored as a mediator. To ensure the ecological validity of the experimental setting, all of the experiment procedures were done online. A total of 720 college students randomly selected at Michigan State University participated as the final sample. The results of the Analysis of Covariance (ANCOVA) indicated that CRM campaigns resulted in the more favorable response of consumers to the brand than CS campaigns, even though any kind of CSR campaign, either CRM or CS, led to a greater response as opposed to the absence of CSR campaigns. It was further found that even though consumers were less likely to join the practical brand pages than the symbolic brand pages in general, the effectiveness of the CSR campaign was more pronounced on the practical brand pages than on the symbolic brand pages. The mediation analyses further confirmed that the consumer expectancy to look favorable was a significant mediator working as a psychological mechanism underlying the relationship between the CSR campaign conditions and the consumer response to brands. The findings of the present study are consistent with the previous suggestion that the public display nature of social media encourages consumers to deliver favorable images of themselves to their friends. The findings of the current study further indicate that the impression management perspective could be one of the promising theoretical approaches to enhance our understanding of consumer participation in CSR campaigns in social media. Such findings are expected to initiate scholastic interest in the effectiveness of CSR campaigns on the favorable response of consumers to the brand in the social media setting. Given the fact that today‘s marketers are eager to extend their CSR campaigns into social media channels, the findings of the present study would provide practical answers for the marketers, particularly pertaining to how to use social media strategically to enhance consumer participation in CSR campaigns. Therefore, the current experimental study would shed light on how the effective provision of CSR campaigns using social media channels could be a potential marketing tool to increase the favorable responses of consumers to the brand, and further cultivate the active engagement of consumers to the brand. Copyright by HYUN JU JEONG 2011 ACKNOWLEDGMENTS Even though my name appears on the cover of this dissertation, I have been richly blessed with the precious efforts and support from advisors, friends, and family throughout the process of completing my dissertation. First of all, I would like to express my most sincere appreciation to my treasured advisors. Dr. Hye-Jin Paek has shown great leadership and expertise throughout my dissertation process. She gave me the freedom to explore on my own, and at the same time the guidance to recover when my steps faltered. She is also my life mentor who has always offered full support and encouragement for my doctoral career. My co-advisor, Dr. Mira Lee, has always been there to listen and give advice, even though she could not finish the race with me on my dissertation journey. She has served as my personal ―guiding light‖ and helped me find direction whenever I lost the way to move on. I would like to thank these two special women in my life who help inspire me simply by their presence. My appreciation extends to my committee members. Dr. Hairong Li has always provided me with creative and insightful comments at different stages of my work, which helped me carve out my ideas. Dr. Wei Peng helped me find solutions to problems with valuable comments from conceptual development to metrological issues. Dr. Elizabeth Quilliam shared her expertise with me by providing extensive feedback that eventually strengthened the theoretical and methodological argument of the dissertation. Furthermore, I have to give special mention to the support given by Dr. Thomas Hove. He has provided an immeasurable amount of guidance and encouragement for my doctoral studies and academic career. His unconditional help for my growth has been a true blessing. v I would like to thank other doctoral students in the Ad and PR program: Dr. Mi Kyoung Kim, Dr. Sung-mi Lee, Hyegye Lee, Hung Jung Oh, and Sookyoung Kim. The enthusiasm I shared with these girls has helped me overcome setbacks and stay focused on my doctoral study. I would like to thank my friend, Dr. Hosun Kang, as well for her long and sincere friendship. Furthermore, I know that my best friend, Hyunjung Lee, who left this earth for heaven last year with her husband, Dr. Chanho Lim, and two adorable children, Sunghoon Lim and Songhyun Lim, has been with me, watching me grow from a distance during both good and bad times. Most importantly, none of this would have been possible without the love and sacrifice of my family. My husband, Dr. Doo Young Kim, and my daughter, Irene Kim, to whom this dissertation is dedicated, have given me the opportunity to follow my dream and showed constant love and care all these years. I deeply appreciate their belief in me and express my heart-felt gratitude to them. I would like to thank my parents, Jaerin Jeong and Sunny Kim for always loving me and believing in me. They gave me a wonderful childhood and great start in life, which allowed me to dream. My parents-in-law, Hoiseok Kim and Chunyoung Kang, have given me support and enthusiasm throughout this endeavor. I am indebted to all those people who have made this completion of my dissertation possible, and because of them my graduate experience has been one that I will cherish forever. vi TABLE OF CONTENTS LIST OF TABLES …………………………………………………………………………. ix LIST OF FIGURES …………………...…………………………………………………… x CHAPTER I INTRODUCTION ……………………………………………..……………………...…… Corporate Social Responsibility Campaigns of Brands ………………………...….... Social Media as Emerging Media Channels for CSR Campaigns …………….…….. Objectives of the Study …………………………………………………………….... Design of the Study …………………………………………………………………... Expectations and Implications of the Study …………………………………………. Organizations of the Chapters …………………………..…………………………… 1 1 2 4 8 9 13 CHAPTER II BACKGROUND LITERATURE ………………………………………………………….. Corporate Social Responsibility (CSR) Campaigns …………………………………. Types of CSR Campaigns ……………………………………………………………. Cause-Related Marketing (CRM) ……………………………………………… Corporate Sponsorship (CS) …………………………………………………… Types of CSR Campaigns in Social Media ………………………………………..… Cause-Related Marketing (CRM) in Social Media ……………………………. Corporate Sponsorship (CS) in Social Media …………………………………. Effectiveness of CSR Campaigns on Consumer Responses to Brands …………….. Factors Influencing the Effectiveness of CSR Campaigns on Consumer Responses to Brands ………………………………………………………………………………... Types of Self-Friend Gender Composition and Types of Brands ……………… 14 14 16 16 17 19 21 22 24 27 31 CHAPTER III THEORETICAL FRAMEWORK ………………………………………………………..... Impression Management Theory ……………………………………………………... Impression Management in Social Media ……………………………………………. Cause Support for Impression Management …………………………………………. Direct versus Indirect Tactics for Impression Management …………………………. 34 34 35 37 39 CHAPTER IV HYPOTHESIS DEVELOPMENT ……………………………………………………...….. Main Effects of CSR Campaign Conditions …………………………………………. Mediating Role of Outcome Expectancy ……………………………………………. Moderating Roles of Types of Self-Friend Gender Composition ……………………. Moderating Roles of Types of Brands ……………………………………………….. 46 46 48 51 55 CHAPTER V vii METHODS ………………………………………………………………………………… Overview ……………………………………………………………………………... Stimuli Development ………………………………………………………………… Selection of Brands: Pre-Test ………………………………………………….. Selection of Charitable Cause: Pre-Test ……………………………….……… Manipulations of CSR Campaign Conditions: Main Study ……………….….. Manipulations of e-mail Invitations for Self-Friend Gender Composition: Main Study …………………………………………………………………………… Sampling and Final Sample ………………………………………………………….. Procedures ……………………………………………………………………………. Measures ……………………………………………………………………………... Analytic Strategy ………………………………………………………………...…… Assumption Checks ………………………………………………………….………. Awareness Checks …………….……………………………………………….…….. Manipulation Checks ………………………………………………………...…..…... Hypothesis and Research Question Testing ……………………………………..…… 63 63 63 64 68 69 71 72 76 78 81 82 85 86 88 CHAPTER VI RESULTS …………………………………………………………………………….….... 94 Assumption Checks ………………………………………………………………….. 95 Awareness Checks ………………………………………………………………..…. 95 Manipulation Checks ………………………………………………………………… 95 Hypothesis Testing …………………………………………………………………… 96 Main Effects and Interaction Effects ………………………………….……….. 96 Main Effects of CSR Campaign Conditions on Intention to Join and Intention to Invite ……………………………………………………………………….... 96 Main Effects of CSR Campaign Conditions on Outcome Expectancy ………... 97 Interaction Effects of CSR Campaign Conditions and Types of Self-Friend Gender Composition …………………………………………………………… 98 Interaction Effects of CSR Campaign Conditions and Types of Brands ….….. 100 Mediating Effects of Outcome Expectancy …………………………….……… 104 CHAPTER VII GENERAL DISCUSSION AND CONCLUSION ………………………………………… Study Overview ………………………………………………………………………. Summary of Findings and Conclusion ……………………………………………….. Limitations and Future Studies ………………………………………………………. Theoretical Implications ……………………………………………………………… Practical Implications ………………………………………………………………… 112 112 116 123 125 128 APPENDIX A ……………………………………………………………………………... APPENDIX B …………………………………………………………………………... REFERENCES…………………………………………………………………………... 132 133 156 viii LIST OF TABLES Table 1. Summary of Key Concepts and Definitions ………………………….….….... 44 Table 2. Descriptive Statistics of Images of and Attitude toward Bottled Water Brands in the Pre-Test …………………………………………………………………...……. 68 Table 3. Characteristics of Final Sample of the Main Study and MSU Undergraduate Population ………………………………………………...………………………...…... 75 Table 4. Final Sample Sizes in Each Experimental Condition ……………….….….…. 76 Table 5. Descriptive Statistics for Variables ……………………………………….….. 81 Table 6. Summary of Analytic Strategy ……………………………………………..… 91 Table 7. Means and Standard Errors for Main Effects of CSR Campaign Conditions on Dependent Variables ………………...…………………………………………….… 98 Table 8. Means and Standard Errors for Interaction Effects of CSR Campaign Conditions and Types of Self-Friend Gender Composition on Dependent Variables …………………………………………………………..……………...…..... 99 Table 9. Means and Standard Errors for Interaction Effects of CSR Campaign Conditions and Types of Brands on Dependent Variables …………………..….…..….. 101 Table 10. Summary of ANCOVAs for Hypothesis and Research Question Testing ……………………………………………………………………………….…. 103 Table 11. Summary of Hypothesis and Research Question Testing Results ……..…….. 117 Table 12. Manipulations for CSR Campaign Messages by Brand Types ………….…... 132 ix LIST OF FIGURES Figure 1. Summary of the Hypotheses …………………………………………...…… 60 Figure 2. Interaction Effects of CSR Campaign Conditions and Types of Brands on the Intention to Join ……………………………………………………………….…... 100 Figure 3. Mediating Role of Outcome Expectancy for the Relationship between CSR Campaign Conditions on the Intention to Join (H3a) ……………………………….… 106 Figure 4. Mediating Roles of Outcome Expectancy for the Relationship between CSR Campaign Conditions on the Intention to Invite (H3b) ………………….……...……. 110 Figure 5. The Content of E-mail Sent from Male Friend who Invites Fiji Water Facebook Page With CSR Campaigns ……………...………………………………… 135 Figure 6. The Manipulation of Fiji Water Facebook Page with CRM Campaigns …… 136 Figure 7. The Content of E-mail Sent from Female Friend who Invites Aquafina Facebook Page with CSR Campaigns ………………………………………………… 142 Figure 8. The Manipulation of Aquafina Facebook Page with CS Campaigns ………. 143 Figure 9. The Content of E-mail Sent from Male Friend who Invites Fiji Water Facebook Page without CSR Campaigns ……………………………………………... 149 Figure 10. The Manipulation of Fiji Water Facebook Page without CSR Campaigns... 150 x CHAPTER I INTRODUCTION Corporate Social Responsibility Campaigns Brands Over the years, marketers have attempted to find a way to increase the number of consumers who are willing to be ―connected‖ with brands (Berkowitz 2010; ExactTarget 2010). Noting the growth in consumer skepticism toward advertising messages (Ford, Smith and Swasy 1990; Obermiller and Spangenberg 2000) and consumer development of persuasion knowledge over time (Friestad and Wright 1994), it becomes important for today‘s marketers to develop marketing strategies stimulating consumers‘ favorable responses to brands and further active engagement in brand communication. As one of the key efforts, marketers have focused on partnering their brands with charitable causes, which is referred to as, in general terms, corporate social responsibility (hereinafter CSR) campaigns. Defined as ―a commitment of [a corporation] to improve community well-being through discretionary business practices and contributions of corporate resources‖ (Kotler and Lee 2004, p. 3), CSR campaigns, thus, combine different promotional mixes such as advertising and public relations under the umbrella concept of integrated marketing strategies for brands (Tangari, Folse, Burton, and Kees 2010). Accordingly, CSR campaigns have been specified as various forms of marketing and philanthropic activities, possibly ranging from cause-related marketing, cause sponsorship, socially responsible employment, and socially responsible manufacturing practices, to employers‘ community involvement (Bhattacharya and Sen 2004; Seitanidi and Ryan 2007). However, CSR campaigns are performed typically as the campaigns of brands or companies to support charitable causes (Drumwright and Murphy 2001; Nan and Keo 2007; Seitanidi and Ryan 2007). Thus, frequently 1 in advertising and marketing literature, CSR campaigns are referred to as ―various forms of company involvement with charitable causes and the nonprofits that represent them (Lichtenstein, Drumwright, and Braig 2004, p. 16).‖ Therefore, in the current study, the terms CSR campaigns and marketers‘ cause campaigns are used interchangeably. How consumers respond to brands has been a significant criterion to evaluate the effectiveness of CSR campaigns in the marketplace. Marketers‘ partnering with charitable causes (hereinafter causes) is considered as providing consumers with a good reason to be connected with the brands in a favorable light (Morrissey 2008). The favorable response of consumers toward brands that perform CSR campaigns has been substantially documented throughout advertising, public relations, and marketing literature (Cornwell and Coote 2005; Gwinner and Eaton 1999; Dean 2003/2004; Nan and Heo 2007; Pope and Voges 2000). However, one of the areas that has not been investigated actively in academia is the effectiveness of CSR campaigns on brands in social media. Even though social media are suggested to facilitate consumer connection with brands in voluntary and public ways in industry (ExactTarget 2010; Morrissey 2008; Slutsky 2011), and are further expected to become potential media channels to elicit consumer participation in CSR campaigns (Abrams 2011; Maul 2011), scholastic efforts and investigations of CSR campaigns in social media have not been rigorously made. Social Media as Emerging Media Channels for CSR Campaigns Given the fact that social media tend to include various online media platforms, from social networking sites (hereinafter SNSs) to microblogging sites (Mangold and Faulds 2009), the working definition of social media is ―a group of Internet-based applications that build on the ideological and technological foundations of Web 2.0, and that allow the creation and exchange 2 of User Generated Count (Kaplan and Haenlein 2010, p. 6).‖ In the present study, SNSs, particularly Facebook, are mainly discussed. It is because SNSs are the most popular social media platforms, defined as Web-based services that allow individuals to (1) construct a public profile within a bounded system, (2) articulate a list of other users with whom they share a connection, and (3) view and traverse their list of connections and those made by others within the system (Boyd and Ellison 2007). Moreover, Facebook is the top-ranked social media or SNS channel in terms of the number of users (Nielson 2010; Quantcast 2010). Regarding the potential of social media as new channels for CSR campaigns, recent industry articles (Abrams 2011; Maul 2011; Slutsky 2011) considered that social media facilitated consumer participation in CSR campaigns. Unlike traditional media channels (e.g., television, radio, magazines, newspapers), social media were expected to provide consumers an opportunity to respond to brand messages actively (Abrams 2011; ExactTarget 2010; Maul 2011; Williamson 2010). For example, consumers can join brand pages, comment on marketers‘ posts, and even spread brand messages to other people once they are interested in the brands or when the brands are recommended by their friends (Williamson 2010). Similarly, the provision of CSR campaigns on brand pages in social media can enhance the willingness of consumers to be connected with the brands as a way to participate in the CSR campaigns (Abrams 2011). By simply clicking the ―Like‖ button on Facebook, consumers can become fans (i.e., join the brand pages) of specific brands promoting any CSR campaigns (Morrissey 2009; Slutsky 2011). As a more important feature, social media were further expected to provide consumers an opportunity to respond to brand messages publicly. For instance, clicking the ―Like‖ button on Facebook is publicized to their friends through their own profile pages and their friends‘ News Feed pages (Slutsky 2011). As such, once consumers participate in CSR campaigns on the brand 3 pages, their participation is visible to their friends. Such features of social media can encourage consumers to respond to the brands featuring CSR campaigns in a favorable light. Consumers can deliver a favorable impression to other people, possibly as those who are connected with the brands supporting causes. Indeed, a recent survey reported that ―expressing themselves with personal contribution to and endorsement of brands to their Facebook friends‖ (39%) was ranked as one of the top two reasons why Facebook users actually clicked the ―Like‖ button to join the Facebook brand pages (ExactTarget 2010). Likewise, it seems important to note that social media become the public settings in which consumers can publicize their responses to the brands (e.g., become members), and it should be particularly true as the brands support causes. The public setting was highlighted in charity literature as a key component of charity campaigns in order to maximize individuals‘ participation in the campaigns. The literature suggested that individuals were more likely to increase their support for cause campaigns when they realized that their support was done in the presence of others, or when it was visible to others (Ariely, Bracha, and Meier 2009; Bereczkei, Birkas, and Kerekes 2010; White and Peloza 2009). In a similar vein, since consumer participation in a CSR campaign is automatically publicized to others in social media, consumers would expect their participation in the CSR campaign to be an opportunity to give off a favorable image to their friends. Thus, the consumers would respond to brands favorably by becoming members of the brands, when the brands support any cause on their social media pages. Objectives of the Study As mentioned, the current study noted the potential of CSR campaigns in social media as new marketing practice to increase consumers‘ favorable responses to the brands. It was suggested that CSR campaigns in social media would provide an opportunity for consumers to 4 publicize their support of the brands featuring the CSR campaigns, to other people. Thus, the present study focuses on how marketers would maximize consumers‘ favorable responses to the brands that feature CSR campaigns in social media. More specifically, the current study explores whether, which, and when the CSR campaigns in social media yield a favorable response to brands among consumers. Accordingly, the objectives of the current study are threefold: (1) understanding whether CSR campaigns lead to favorable consumer responses to brands in social media; (2) understanding which types of CSR campaigns yield the more favorable consumer responses to brands in social media; (3) understanding when CSR campaigns are effective for favorable consumer responses to brands in social media. Here, consumer responses to brands with CSR campaigns in the social media setting are specified using two aspects: the intention to join the brand pages, and the intention to invite their friends to the brand pages. It is because CSR campaigns in social media are primarily conducted to encourage consumers to be actively connected with brands (Abrams 2011), typically by joining the brand pages and inviting their friends to the brand pages. The increase in membership on their brand pages in social media is essential for marketers to elicit viral marketing and consumer engagement in the brands (ExactTarget 2010; Williamson 2009; Syncapse 2010), and further increase brand sales (Maul 2011; Syncapse 2010). Therefore, academic efforts should be made to systematically investigate how to increase the number of consumers who are willing to be favorably connected with the brands in social media. The current study suggests that the effective development of CSR campaigns in social media could be one of the answers regarding how to elicit consumer engagement in brands. First, the present study explores whether CSR campaigns would yield favorable consumer responses to brands in the social media setting. The current study carries out the 5 baseline research to examine if the presence of CSR campaigns on the brand pages in social media would generate more favorable responses to brands among consumers, as opposed to the absence of CSR campaigns. To enhance our understanding of the effectiveness of CSR campaigns on brands themselves in the social media context, it is vital to conduct such a baseline experimental study to compare the presence and absence of CSR campaigns in terms of consumer responses to the brands. Second, the present study attempts to investigate which kind of CSR campaigns would be more effective in yielding favorable consumer responses to brands in social media. Here, the present study focuses on two types of CSR campaigns most frequently found in the social media setting: cause-related marketing (hereinafter CRM) and cause sponsorship (hereinafter CS). The CRM campaign in social media typically promises that the brand will provide a certain monetary (e.g., $1) or product (e.g., six meals) donation to a cause each time a consumer adds his/her contribution to the CRM campaign by clicking the ―Like‖ button, which leads the consumer to join the Facebook brand page. The CS campaign in social media promotes the brand‘s monetary and product donations to a cause during a certain period of time, typically by announcing the brand as the official sponsor of the cause. Specifying the relative effectiveness of these two different types of CSR campaigns is necessary, in order to obtain further insight concerning which CSR campaign strategy would be more effective on the favorable response of consumers to the brands in the social media context. Third, to understand when the effectiveness of CSR campaigns would be more highlighted, the present study focuses on two situational factors by which the effects of CSR campaigns in social media on the consumer response to brands would be moderated: types of self-friend gender composition and types of brands. 6 The notion of self-friend gender composition is taken from impression management literature (Leary et al. 1994), refined in the current study as the gender types of the consumers and their friends who invite them to the brand pages, and further specified as two types (i.e., same and different) based on whether both the consumers and the friends have the same or different gender types. The self-friend gender composition in social media could be a considerable factor among marketers, given the fact that social media are beneficial to marketers because of the usage of friendship within social media to promote brands (Morrissey 2010). Dwelling on the previous suggestion that individuals are likely to manage their impression differently in terms of gender composition between self and partners (Leary et al. 1994), the current study examines whether the different types of self-friend gender composition (i.e., same and different gender composition between consumers and their friends in social media) would moderate the proposed effectiveness of CSR campaigns on the consumers‘ response to the brands. As another situational factor influencing the impact of CSR campaigns on consumer responses to the brands, the present study suggests two types of brands: brands with symbolic images (hereinafter symbolic brands) and brands with practical images (hereinafter practical brands). Consumer psychologists suggested that each brand had its own image (Dolich 1969), and such brand image was reflected in the image of consumers who purchased or possessed the brand (Belk 1988; Sirgy and Danes 1982). Symbolic brands are defined as brands that satisfy consumers‘ symbolic (or self-expressive) needs such as desires for products that fulfill internally generated needs for self-enhancement, role position, group membership, or ego-identification; whereas practical brands are defined as those that tap into consumers‘ functional (or utilitarian) needs that motivate the search for products that solve consumption-related problems (Bhat and 7 Reddy 1998). Accordingly, the present study suggests that such symbolic and practical brand images should also be utilized for the identification of those who become members on their brand pages in social media. Furthermore, the current study expects that consumers would decide their response to the brands featuring CSR campaigns differently, based on the image they would obtain and publicize by becoming members on the brand pages. In combination, understanding the moderating mechanisms of types of self-friend gender composition and types of brand in the consumer response to CSR campaigns may be particularly important in the social media context. As suggested in communication literature (Boyd and Ellison 2007; Krämer and Winter 2008; Lamp, Ellison and Steinfield 2007), social media users tend to control the images they present to other people by frequently articulating their personal information seen by others. Thus, when consumers decide to become members of brands featuring CSR campaigns in social media as an effort to present favorable images of themselves to their friends, they would consider whether they were invited to the brand pages by the same or different gender friends, or whether the brands featuring the CSR campaigns have symbolic or practical images. In light of all this, the present study integrates several emerging research domains related to CSR campaigns, self-friend gender composition, and brand images in the social media context. All of these discussions are conceptually based on impression management theory (Goffman 1959; Guadagno and Cialdini 2007; Leary and Kowalski 1990; Tedeschi 1981), which will be discussed mainly in the following chapter. Design of the Study A total of 720 college students, who were enrolled in various departments at Michigan State University (hereinafter MSU), were randomly recruited through the Office of the Registrar 8 and participated in this experimental study as the final sample. This sample seems to represent MSU undergraduates in general, in that the overall characteristics of the final sample are similar to those of the MSU undergraduate population. The present study conducted online experimental research, using a 3 (condition of CSR campaign: CRM, CS, or control) x 2 (type of self-friend gender composition: same versus different) x 2 (type of brand: symbolic versus practical) between-subjects design. A total of six brand pages, modifying existing brand pages on Facebook, were created corresponding to three CSR campaign conditions for each of the two brands. To manipulate the type of self-friend gender composition, the participant was asked to imagine getting an e-mail from his/her male or female friends inviting him/her to the brand page. Thus, after comparing the gender types of each participant and his/her friend, each participant was categorized as either the same (i.e., when each had the same gender type) or different (i.e., when each had different gender types) selffriend gender composition. In addition, the measures were adopted from previous studies and found to be reliable to indicate the index of each variable. Expectations and Implications of the Study First, the present study investigates whether the presence of each of the CSR campaigns (i.e., a CRM campaign or a CS campaign) will be effective in yielding favorable consumer response to the brands. Based on impression management theory (Goffman 1959; Guadagno and Cialdini, 2007; Leary and Kowalski 1990), it is expected that the presence of any kind of CSR campaign will be effective on the favorable response to the brand in social media, because participation in CSR campaigns on the brand pages will provide consumers an opportunity to be seen as socially favorable to their friends, possibly like those who would be involved in the cause support of brands. 9 The findings of the present study would broaden academic studies about social media into a new public setting for consumer participation in CSR campaigns, and eventually elicit various theoretical investigations on the strategic use of social media as new cause marketing campaign channels. The findings would also suggest that marketers could increase the number of consumers who join their brand pages in social media by developing a CRM campaign, or simply by featuring information about a CS campaign, on their brand pages in social media. Second, the present study investigates which types of CSR campaigns will result in more favorable consumer responses to the brands in social media. Based on the previous discussion on the superior effect of direct tactics (over indirect tactics) for impression management (Tal-Or 2010), the current study expects that the CRM campaign (versus the CS campaign) will lead to more favorable consumer responses to the brands in social media. As a possible reason, it is suggested that joining and inviting activities on the brand pages featuring the CRM campaign (versus the CS campaign) will be considered a more direct way to support causes, and consequently yield the greater outcome expectancy of consumers to look favorable to others. The mediation analysis will confirm that such outcome expectancy will be a mediating mechanism underlying the superior effects of CRM campaigns (over CS campaigns and the control condition) in terms of consumer responses to brands in social media. The findings of the present study would provide academic evidence that an opportunity to look favorable to their friends could motivate consumers to be connected with certain brands, when the brands feature CSR campaigns in social media, and further when the brands provide an opportunity for consumers to participate in CSR campaigns more directly. Likewise, by merging cause marketing literature and consumer motivation literature, this study attempts to add further theoretical discussions to the growing body of impression management studies, which have thus 10 far received little attention in the fields of advertising and marketing. Particularly, this new theoretical approach would be helpful in enhancing our general understanding of how young consumers would participate in CSR campaigns in the social media setting. Such findings are also beneficial for marketers. According to the findings of the present study, marketers need to place an emphasis on developing message strategies for CSR campaigns in social media, possibly by highlighting the opportunity for consumers to be seen as cause supporters, and increasing the visibility of consumers‘ membership with brands when the brands support causes. Third, as an attempt to specify when the CSR campaigns will be more or less effective in the social media setting, two situational factors are examined as the moderating variables influencing the main effects of CSR campaign conditions on consumer responses to the brands: types of self-friend gender compositions (i.e., same and different) and types of brands (i.e., symbolic and practical). More specifically, first, the present study expects that when the self-friend gender composition is different (versus same), consumers will be more motivated to present themselves in a favorable way to their friends; thus, the main effect of CSR campaign conditions on consumers‘ favorable responses to brands will be more pronounced. The theoretical discussion on self-friend gender composition is based on social psychology literature, suggesting that communication with different-gender partners is more likely to motivate people to manage their impression than that with same-gender partners (Glass, Gottman, and Shmurak 1976; Leary 1983; Leary et al. 1994; White and Dahl 2006). Second, the present study expects that the proposed effects of CSR campaigns will be different based on the types of brands (symbolic and practical). Because, to the author‘s best knowledge, no decisive evidence exists regarding whether and how brand images will moderate 11 consumer responses to brands providing CSR campaigns on social media brand pages, the present study does not predict any direction; rather, it asks a research question. That is, the relative impact of CSR campaigns on consumer joining and inviting activities on social media pages for either symbolic or practical brands can be explained with different theoretical rationales. First, when the brands have symbolic (versus practical) images, consumers‘ joining and inviting activities on the social media brand pages would be seen by consumers as an opportunity to present their social images (e.g., look like good citizens) more saliently, in conjunction with the image-salient nature of symbolic brands. Thus, consumers would expect greater social rewards for their joining and inviting activities and further respond to the CSR campaigns more favorably, on the symbolic (versus practical) brand pages. On the other hand, the opposing argument should also be considered. That is, even though consumers‘ joining and inviting activities would be greater in general, the effectiveness of CSR campaigns on the joining and inviting activities would be less for symbolic brands than for practical brands. This is because consumers would be less likely to consider whether the presence and type of CSR campaign would have the capacity to make them look favorable to their friends; thus, their joining and inviting activities would be less likely to enhance their social image. Based on such arguments, it seems reasonable to consider that when brands have symbolic (versus practical) images, consumers would have less intention to join the practical brand pages and invite their friends. Since there has been no literature documented for the moderating role of brand type on CSR campaign effectiveness in social media, such a research question will provide clear answers about whether and how brand image would influence consumer participation in CSR campaigns in the social media setting. 12 All of the findings from the current experimental study would extend theoretical explanations on how brand images could be integrated within the cause marketing context, particularly to encourage consumers to be involved in the brand page in social media. By exploring consumer motivation to look favorable in terms of cause support, self-friend gender composition, and brand images, the current study would broaden scholastic inquiry into the effectiveness of CSR campaigns on the brands in social media. Such findings would further provide practical implications for marketers. The findings would suggest that marketers‘ CSR campaign strategy should take into consideration how to utilize the gender composition between social media users and whether the brands would be positioned as symbolic or practical among consumers. Organization of the Chapters In this dissertation, Chapter II will briefly discuss background literature on the definitions, the scopes, the types, and the effectiveness of CSR campaigns, as well as CSR campaigns in social media. The study will then discuss impression management theory in Chapter III and provide further theoretical discussions for conceptualization and hypothesis development in Chapter IV. Next, this study will present the methodologies explaining the experimental design and procedures used for the study in Chapter V, which will be followed by the results of hypothesis testing in Chapter VI. Lastly, the current study will summarize the findings and provide the conclusion with further discussions on study limitations, suggestions for future studies, theoretical implications, and managerial implications in Chapter VII. 13 CHAPTER II BACKGROUND LITERATURE Corporate Social Responsibility (CSR) Campaigns CSR campaigns are defined as ―a commitment of [a corporation] to improve community well-being through discretionary business practices and contributions of corporate resources (Kotler and Lee, 2004, p. 3).‖ Even though CSR campaigns are by no means new ideas among marketers (Bhattacharya and Sen 2004; Smith 2003), today‘s companies are more likely to use their skills and resources for the CSR campaigns than ever before (Bhattacharya and Sen 2004; Drumwright and Murphy 2001; KPMG, 1999, 2002, 2005; Nan and Heo 2007). According to a series of reports from KPMG (1999, 2003, 2005) — a U.S. audit, tax, and advisory service firm — the Fortune top 250 companies tended to engage in CSR campaigns with a continued growth over time, showing that 53% of these companies engaged in CSR campaigns in 2005, compared to 45% in 2002, and 35% in 1999. More interestingly, such growth in the top-ranked marketers‘ investment in CSR campaigns seems unlikely to be significantly impacted by a recent financial crisis in the U.S. Indeed, it was found that the Fortune top 500 companies considered the CSR campaigns as important commitments they should sustain and expand, even under the current economic conditions, as a long-term business strategy (Delevingne 2009). The special interest in CSR campaigns among marketers had begun with the marketers‘ awareness of the CSR campaigns as ―an ethical or ideological imperative (Bhattacharya and Sen 2004, p. 9).‖ Marketers had been involved in social well-being originally based on their recognition that ―in a free society any business operates only as long as societal members continue to grant it that right (Vardarajan and Menon 1988, p. 59).‖ However, such a philanthropic approach to CSR campaigns extended into a new phase, weighing the marketers‘ 14 commercial aspects of performing CSR campaigns (Bhattacharya and Sen 2004; Delevingne 2009; Vardarajan and Menon 1988). Marketers became increasingly aware of commercial benefits, such as enhancement of brand or company reputation (Gwinner and Eaton 1999; Dean 2003/2004; Nan and Heo 2007; Pope and Voges 2000), and increases in brand sales (Cornwell and Coote 2005; Pope and Voges 2000). As such, more recent types of CSR campaigns are placed somewhere on the middle ground between philanthropic service and marketing practice (Vardarajan and Menon 1988). Therefore, it should be concluded that ―companies today are increasingly aware of both the normative and business cases for engaging in CSR campaigns; not only is doing good, the right thing to do, but it also leads to doing better through its positive effects on key stakeholder groups (Bhattacharya and Sen 2004, p. 9).‖ Accordingly, CSR campaigns are practiced as different forms, in terms of the symmetrical level of the two philanthropic and marketing purposes. As CSR campaigns put more weight on philanthropic purposes than marketing purposes (i.e., the asymmetrical level between two purposes), marketers tend to be involved, for example, in socially responsible employment (e.g., job security), manufacturing practices (e.g., product safety), and purely altruistic giving to charitable causes (e.g., community support without any advertisement) (Polonsky and Speed 2001). These forms of CSR campaigns are committed in the domain of broader corporate social performance, not only philanthropic support for causes (Polonsky and Speed 2001). On the contrary, as CSR campaigns put equal focus on both philanthropic and marketing purposes (i.e., the symmetrical level between two purposes), marketers are likely to make partnerships, typically with charitable causes or charitable organizations (Bhattacharya and Sen 2004; Polonsky and Speed 2001; Seitanidi and Ryan 2007). That is, when marketers develop CSR campaigns, they attempt to use their associations with causes strategically to achieve their 15 marketing purposes, rather than simply offer altruistic help for the causes (Polonsky and Speed 2001). Indeed, marketers‘ CSR campaigns were found to be effective in enhancing consumer evaluation of brands (Cornwell and Coote, 2005; Gwinner and Eaton, 1999; Dean 2003/2004; Nan and Heo, 2007; Pope and Voges, 2000), even though they were also effective in making significant contributions to the causes (Creyer and Ross 1997). Of the various types of CSR campaigns, CRM campaigns and CS campaigns are most frequently practiced in the marketplace (Drumwright and Murphy 2001; Lichtenstein et al. 2004; Kotler and Lee 2004; Dean 2003/2004; Nan and Keo 2007; Polonsky and Speed 2001; Seitanidi and Ryan 2007). In this paper, these two types of CSR campaigns are mainly discussed. Types of CSR Campaigns Cause-Related Marketing (CRM) As the first type of CSR campaign to support causes, CRM should be noted. CRM is defined as ―the process of formulating and implementing marketing activities that are characterized by an offer from the firm to contribute a specific amount to a designated cause when customers engage in revenue-providing exchanges that satisfy organizational and individual objectives (Varadarajan and Menon 1988, p. 60).‖ The cause campaign of American Express in the early 1980s is a notable example from which the term CRM was coined (Kotler and Lee 2004). In 1983, American Express, a major credit card company, initiated a CRM campaign for the ―Statue of Liberty Restoration Project,‖ announcing that 1¢ for each use of the American Express credit card and $1 for each new credit card issued would be given to this cause. In four months, about $1.7 million was delivered to restore the statue through the CRM campaign which also resulted in a 25% increase in the use of the credit card during the fourth quarter of 1983 (Wall 1984). As another well-known example of CRM, Yoplait‘s ―Save Lids to 16 Save Lives‖ which was initiated in 1999, has donated 10¢ to the Susan G. Komen Breast Cancer Foundation, a leading charitable organization in the U.S., for each yogurt product‘s lid returned by consumers. This long-running CRM campaign has led to up to $2 million in contributions every year to the Susan G. Komen Breast Cancer Foundation (the Web site of Yoplait, n. d.), and the significant increase in the brand‘s sales (e.g., 14% increase in Yoplait sales, even in 2003) (Matthew 2004). The CRM campaign is one of the most prevalent types of CSR campaigns in the marketplace today (Nan and Heo 2007; Polonsky and Speed 2001). It might be attributable to the commercial benefits of CRM campaigns, which are embedded in the nature of CRM campaigns (Seitanidi and Ryan 2007). Over time, marketers have attempted in large part to enhance the effectiveness of CSR campaigns as marketing practices, by initiating and implementing CRM campaigns (Kotler and Lee 2004; Polonsky and Speed 2001; Varadarajan and Menon 1988). Corporate Sponsorship (CS) Another type of CSR campaign, CS, is also frequently performed by marketers. CS is defined as ―an investment in cash or in kind [of corporations] in activities in return for access to the exploitable commercial potential associated with those activities (Meenaghan 1991, p. 36).‖ As seen from this definition, marketers attempt to make associations with causes by sponsoring them, primarily in exchange for brand recognition. Even though the CS campaign supports various social and cultural issues or events (e.g., sports programs, movie productions), many of them are likely to represent the brands or the companies‘ support for the causes (Rifon, Choi, Trimble, and Li 2004). One notable example of a CS campaign is the ―Race for the Cure‖ campaign, which has recently been sponsored by several companies, such as Ford and Bank of America. These 17 marketers have sponsored the local and national walk or run races by providing financial contributions, product donations, or marketing services to the campaign. They attempted to raise public awareness to support breast cancer research by the Susan G. Komen Breast Cancer Foundation (the Web site of Susan G. Komen for the Cure 2011), at the same time hoping that their sponsorship could influence consumer recognition of the brands. Another example is the CS campaign for Special Olympics sponsored by a list of marketers, including Wal-Mart and AT&T. The marketers have provided various kinds of contributions for the cause (the Web site of Special Olympics 2011), and been given an opportunity to increase the visibility of their brands through multiple channels of the event (e.g., pamphlets, newsletters, Web sites) during the CS campaign (Seitanidi and Ryan 2007). The CS campaign was initiated by marketers through their altruistic donations to causes in the 1980s; however, it became more pronounced over time as its commercial benefit to marketers became more apparent (Seitanidi and Ryan 2007). The CS campaign is thus known as ―first and foremost a commercial activity (Polonsky and Speed 2001, p. 1361)‖ of marketers. It is also suggested that ―one of the focal characteristics of [corporate] sponsorship is… the compensation rewards that accomplished each sponsorship deal (Seitanidi and Ryan 2007, p. 250),‖ which results in measurable marketing performances (Meenaghan 1991; Seitanidi and Ryan 2007). As such, it is highly common for marketers to expect tangible (e.g., obtaining advertising space or time for their brands) or intangible (e.g., increasing the reputation of their brands) benefits as the marketing outcomes of their involvement in CS campaigns (Seitanidi and Ryan 2007). Therefore, similar to the CRM campaign, the CS campaign is based on the symmetrical relationship between these two entities of marketers and charitable cause organizations in nature, 18 which consequently results in benefits to both entities (Seitanidi and Ryan 2007). Meenaghan (1991) suggests that marketers are likely to use the CS campaign for both philanthropic and marketing activities, for example by providing community involvement; countering adverse publicity; building goodwill among staff, opinion formers, and decision makers; increasing company, product, and brand awareness; reinforcing or altering perceptions of products or brands; identifying products or brands with particular market segments; and indirectly increasing product sales. On the other hand, along with practicing CSR campaigns using traditional media channels (e.g., television, radio, magazines, newspapers), marketers are now involved in both CRM campaigns and CS campaigns using social media channels (Abrams 2011; Williamson 2010). Types of CSR Campaigns in Social Media As social media users have now reached a critical mass in the U.S. (Williamson 2010), marketers‘ efforts to utilize CSR campaigns strategically have recently been extended into the new media context, particularly social media. The definition of social media varies as a large number of social media channels have emerged in the marketplace. Thus, the recent attempt to define social media takes the position of the broad definition as, ―a group of Internet-based applications that build on the ideological and technological foundations of Web 2.0, and that allow the creation and exchange of User Generated Count (Kaplan and Haenlein 2010, p. 6).‖ Accordingly, social media encompass various online media platforms from social networking sites (hereinafter SNSs) (e.g., Facebook, MySpace), video-sharing sites (YouTube, Flickr), and collaborative Web sites (e.g., Wikipedia), to microblogging sites (e.g., Twitter) (Mangold and Faulds 2009). Out of such different social 19 media platforms, the present study is concerned with SNSs. As the most popular social media platforms, SNSs are defined as Web-based services that allow individuals to (1) construct a public profile within a bounded system, (2) articulate a list of other users with whom they share a connection, and (3) view and traverse their list of connections and those made by others within the system (Boyd and Ellison 2007). Facebook is mainly examined here, in that Facebook is the top-ranked social media or SNS channel in terms of the number of users (Nielson 2010; Quantcast 2010). For example, a recent report of Nielson (2010) announced Facebook as the top global social media channel. Similarly, Quantcast, a media measurement and Web analytics firm, rated Facebook (i.e., 141 million monthly users in 2010) as the most popular social media channel in the U.S., followed by YouTube (i.e., 136 million) and Twitter (i.e., 99.4 million) (Quantcast 2010). Indeed, by using social media channels, marketers now adopt various types of CSR campaigns in a creative way. For example, marketers encourage consumers to vote for the causes to which the marketers would donate (e.g., ―Chase Community Giving‖ which was featured on Facebook by J. P. Morgan Chase Bank), or ask consumers to recommend specific local causes the marketers could help (e.g., Pepsi‘s ―Refresh Project‖ in several social media channels). On the other hand, marketers promise consumers that they will make a product donation to a specific cause every time a consumer joins their brand page (e.g., Kraft Foods‘ cause campaign), or every time a consumer joins the charitable organization‘s social media page (e.g. ―Haiti Hope Project‖ on the Facebook page of Odwalla). Out of these various types, the present study focuses on two types of CSR campaigns, which are most frequently practiced in the social media setting: the CRM campaign and the CS campaign. 20 Cause-Related Marketing (CRM) in Social Media Since the CRM campaign in social media (e.g., Kraft Foods‘ CRM campaign, Target‘s CRM campaign) is performed somewhat differently than the traditional CRM campaign (e.g., Yoplait‘s CRM campaign), it is critical to define the CRM campaign in social media. Thus, the present study defines the CRM campaign in social media as the marketing campaign of firms using their social media brand pages, mainly designed to enhance consumer engagement in their brands in the social media setting, and typically performed by monetary or product donations from the firms to designated causes as an exchange for customers voluntarily becoming members of their social media brand pages. Unlike the traditional CRM campaign, the CRM campaign in social media does not promote consumers to purchase the brands (Kerwin 2010; Zmuda and York 2010). However, consumer participation in a CRM campaign in social media is still vital for the implementation and effectiveness of the CRM campaign. That is, rather than focusing on brand sales directly through the CRM campaign, the marketers attempt to connect their brands with a growing number of social media users, typically by encouraging consumers to join their social media brand pages (Kerwin 2010; Zmuda and York 2010). Marketers pay special attention to the effective provision of CRM campaigns in social media (Morrissey 2008, 2009). They expect that these CRM campaigns would elicit consumers‘ favorable responses to brands, particularly by increasing the number of consumers who join the brand pages (Kerwin 2010; Morrissey 2008, 2009). Indeed, a growing number of consumers participate in CRM campaigns in social media. For instance, Kraft Foods developed a CRM campaign in 2008, announcing to consumers on its Facebook page that it donated six meals to hungry families through Feeding America, one of the leading charitable organizations in the U.S., whenever a consumer joined Kraft Foods‘ Facebook page. During the span of this two- 21 week campaign, about 25,000 people joined the Kraft Foods‘ Facebook page, and accordingly, 150,000 meals were donated. Similarly, Target‘s 2010 Facebook CRM campaign promised donations to five selected causes on the basis of the votes received from those who joined Target‘s Facebook brand page during the campaign. When this CRM campaign had closed after two weeks, over 900,000 million people had joined Target‘s Facebook page to participate in voting for the selection of the causes to be supported. Cause Sponsorship (CS) in Social Media Since CS campaigns in social media have not been defined in literature yet, the current study defines the CS campaign in social media as the marketing campaign of firms using their social media brand pages, primarily designed to enhance consumer engagement in their brands in the social media setting, and mainly performed by monetary or product donations from the firms to designated causes as a way to promote their philanthropic commitment to the cause. Even though the performance of CS campaigns in social media does not require consumers to join their brand pages, the CS campaign in social media promotes corporate sponsorship activities to consumers directly through their own social media pages, which is expected to be effective in getting consumers to join their brand pages. Marketers often use social media channels as one of the tools to promote their support for a cause. With the growing number of social media users, marketers expect they would increase exposure of their CS campaigns to consumers by using the social media channels strategically. One example is Starbucks‘ CS campaign on Facebook. The Starbucks Facebook page indicated that the brand was an official sponsor of the charitable cause ―RED‖ to help Africa, eventually in the hopes of increasing the number of consumers who join their brand page on Facebook. 22 Likewise, joining the brand pages is not required, but it is promoted for the sake of the performance of the CS campaign in social media. In combination, it should be stressed that both CRM campaigns and CS campaigns are consistent in that they were initiated for both marketing and philanthropic purposes (Polonsky and Speed 2001; Seitanidi and Ryan 2007). These two CSR campaigns became important marketing tools, to elicit consumer responses to the brands in a favorable way (Dean 2003/2004; Nan and Heo 2007; Polonsky and Speed 2001; Seitanidi and Ryan 2007). The same discussion would apply to the CSR campaigns in the social media setting. In the CRM campaign, the contribution by marketers to causes is tied to the actions of the consumers (Polonsky and Speed 2001; Seitanidi and Ryan 2007), typically the joining activity of consumers on the brand pages. On the other hand, for the CS campaign in social media, even though the marketers‘ contributions to causes do not depend on the actions of consumers directly, consumers‘ joining activities on the brand pages are also promoted (Polonsky and Speed 2001; Seitadidi and Ryan 2007). Thus, consumer responses to brands (e.g., joining the brand page) should be a key criterion to evaluate whether the provisions of CRM campaigns and CS campaigns in the social media setting are successful or not. Encouraging consumers to join their brand pages in social media is important for marketers to promote their brands today more than ever before. A recent report (Syncapse 2010) suggested that the increase of membership in brand pages resulted in the increase of viral marketing, brand loyalty, and brand sale among consumers, particularly those who joined their brand page. It indicated that determining how to increase the number of consumers who joined their brand pages as fans of the brands became a part of the integrated marketing strategies among today‘s marketers (Williams 2011). As such, the effective development of CSR 23 campaigns in social media channels would be a potential marketing practice that would facilitate the long-lasting efforts of marketers to enhance consumer responses to their brands, and further encourage consumer engagement in their brands. Effectiveness of CSR Campaigns on Consumer Responses to Brands As noted briefly, the effectiveness of CRM campaigns and CS campaigns is mostly determined by whether or not consumers favorably respond to brands supporting the causes. In traditional media, the effectiveness of CSR campaigns on the brands is evaluated in terms of consumer purchase of the brands, whereas in the social media setting they are decided by the joining activities of consumers on the brand pages. A substantial number of empirical studies documented that CSR campaigns using traditional media were effective marketing practices in general, in increasing favorable responses of consumers toward brands (Chang 2008; Cone 2008; Cornwell, Bettina and Leonard 2005; Dean 2003/2004; Edelman 2009; Lafferty and Goldsmith 2005; Nan and Heo 2007; Strahilevitz and Myers 1998; Tangari et al. 2010). For instance, a list of national surveys provided an overall view of how U.S. consumers evaluated CSR campaigns: a 2008 survey by Cone, Inc., a Bostonbased strategic marketing communication firm reported that 85% of U.S. consumers indicated they would buy the brands supporting causes more than those without the cause support; 79% answered they liked to switch from one brand to another associated with the causes, if the price and quality were the same; furthermore, 38% indicated that they had purchased a product sponsoring the cause during the past 12 months (Cone 2008). A more recent survey (Edelman 2009) showed similar outcomes: 58% of U.S. consumers indicated that they would promote brands if there were causes behind the brands, and 64% expressed that they would recommend the brands that supported the causes. Furthermore, scholars suggested that consumers tended to 24 see the CRM campaign positively as a socially responsible activity of marketers (Ross, Paterson and Stutts 1992), and were likely to purchase the brands supporting causes (Smith and Alcorn 1991). However, not all of the previous studies reported the positive side of CRM campaigns. It was suggested that consumers could have unfavorable opinions of brands, for instance, when the brands donated a small amount of money to causes (Endacott 2004) or when an unclear advertising message describing the brands‘ donation was featured (Chang 2008). It should be noted that such criticism seems to be intensified when the CRM campaign is perceived among consumers as being a marketing tactic to increase brand sales rather than philanthropic help for causes. Indeed, recent studies have investigated the effectiveness of CRM campaigns on brands or marketers in terms of consumer attribution to the motives of brands or marketers to perform the CRM campaign (Dean 2003/2004; Webb and Mohr 1998). For example, Webb and Mohr (1998) suggested in their qualitative research that even though half of their respondents perceived the CRM campaign as being altruistic to support causes, the other half was mostly skeptical toward the marketers‘ motive behind the implementation of the CRM campaign. They further suggested that consumers with a high level of skepticism toward the marketers‘ selfinterest would be less likely to respond positively to the CRM campaign as opposed to those with a low level of skepticism. Dean (2003/2004) further added the new findings that such consumer attribution to the CRM campaign would be different by the brand reputation consumers possessed previously. The findings suggested that when the brand had a poor reputation already among consumers, the CRM campaign yielded a negative effect on the brand; however, when the brand had a good or average level of reputation for consumers, the CRM campaign had neither a negative nor a positive influence on consumers‘ attitude toward the brand. Thus, the study 25 indicated that the average brand would not risk a loss of public goodwill by implementing the CRM campaign (Dean 2003/2004). Similarly, to explore the effectiveness of CS campaigns on brands, empirical studies have juxtaposed both philanthropic and marketing purposes underlying the provision of CS campaigns in the marketplace. The studies investigated whether consumer responses to the brands and the marketers featuring a CS campaign would be different, depending on how consumers inferred the motive of marketers to practice the CS campaign (Rifon et al. 2004; Szykman, Bloom, and Blazing 2004). Rifon and her colleagues (2004) suggested two psychological motives of the marketers to perform the CS campaign (i.e., altruistic sponsor motive versus self-interest sponsor motive) to clarify the inconsistent responses of consumers to marketers (i.e., brands) sponsoring a health and disease-related charitable cause. They found that such different motives of marketers yielded different consumer evaluations according to the levels of congruence between the brand and the cause. They suggested that congruence of the brand and the cause was more likely to generate consumer attribution of altruistic sponsor motives, thereby enhancing marketer credibility and consumer attitude toward the marketer, than incongruence between the brand and the cause. As a similar approach, Szykman et al. (2004) proposed the notions of society-serving motive and self-serving motive, as main motives of sponsorship, and further explored how these motives were integrated with consumer perceptions of social messages featured on cause campaigns and attitudes toward sponsoring marketers. The study suggested that consumers were likely to perceive the socially desirable messages on the cause sponsorship (e.g., anti-drinking and driving message) more favorably, particularly as having society-serving motives of the sponsor, when this cause was sponsored by a non-profit organization (e.g., Mothers Against Drunk Driving [MADD]); on the other hand, consumers tended to perceive the same messages 26 less favorably, as being self-serving by the sponsor, when this cause was sponsored by a corporation (e.g., Budweiser). The perception of the socially desirable message was also found to affect the consumers’ attitude toward the sponsors themselves (i.e., NPO and corporate). In short, previous approaches suggest that there have been mixed responses among consumers toward brands, when the consumers are exposed to the brands‘ CSR campaigns. Such different consumer responses to the brands featuring CSR campaigns seem to be based on the dual practice of the CSR campaigns—a marketing activity versus altruistic help. Consequently, consumer responses to brands could be different, based on whether the CSR campaigns place more weight on a marketing purpose or a philanthropic purpose. Likewise, the consumer responses were the significant criteria to determine the effectiveness of CSR campaigns. On the other hand, after an investigation to understand whether the CSR campaigns were effective on favorable consumer responses to brands, another effort has been made in academia to identify the factors determining the effectiveness of CSR campaigns on the brands. Factors Influencing the Effectiveness of CSR Campaigns on Consumer Responses to Brands While earlier studies increased our understanding about the influence of CSR campaigns on consumer responses to the brands, later research has been devoted to predicting factors increasing and decreasing such effectiveness. Several factors were suggested, including the types of causes such as disaster causes versus ongoing causes (Cui et al. 2003), or local causes versus national causes (Ross et al. 1992); the importance of causes (Strahilevitz and Myers 1998); the types of products such as hedonic versus utilitarian (Chang 2008; Strahilevitz and Myers 1998); the congruence of brand-cause (Lafferty and Goldsmith 2005; Nan and Heo 2007; Rifon et al. 2004; Simmons and Becker- 27 Olsen 2006); and the types of donation messages framed (Chang 2008). For example, Cui and his colleagues (2003) suggested that cause types (e.g., disaster causes versus ongoing causes) could be a factor in determining the effectiveness of the CRM campaign, particularly among the young college-age generation. They indicated that the CRM campaign was more likely to be evaluated positively among this demographic group when it supported disaster causes rather than ongoing causes. On the other hand, Chang (2008) investigated multiple factors determining the effectiveness of a CRM campaign supporting World Vision, a global non-profit organization, such as product type, product price, donation framing, and donation magnitude. The effectiveness of CRM campaigns was assessed by the intention to participate in the CRM campaign. This consisted of the favorableness of the promoted product, the likelihood of purchasing the product, and the probability of recommending the product to others. The study found that the CRM campaign was more effective when the affective products (e.g., shampoo) (versus practical products; e.g., toilet paper) featured the CRM campaign, when the product price was low (versus high), when a higher percentage (e.g., 25% of a sale price) in donation magnitude (versus a lower percentage; e.g., 5% of a sale price) was promoted, or when the absolute dollar (versus the percentage of a sale price) was framed as the amount of donation to the cause. On the other hand, the effectiveness of the CS campaign, compared to that of the CRM campaign, on consumer responses to brands or marketers has been explored relatively less in cause marketing literature (Cornwell and Maignan 1998). One of the reasons is attributed to the fact that the effectiveness of CS campaigns on brands is relatively more difficult to measure in the marketplace than that of CRM campaigns (Harvey 2001). Nevertheless, as a potential marketing tool, the CS campaign was expected to enhance consumers‘ evaluation of the brands, 28 possibly by creating a link between the brand and the cause (Erodogan and Kitchen 1998). Particularly, as a factor influencing the effectiveness of CS campaigns, scholars suggested the pre-existing evaluation of the brands. For example, Lafferty and Goldsmith (2005) found that, even though the CS campaign was likely to yield negative attitudes toward the brand based on a low congruence between the brand and the cause, such outcomes would not occur when the brand had high familiarity and relatively positive attitudes among consumers; the effectiveness of brand-cause congruence on the brand attitudes was not moderated by the cause familiarity. In a similar vein, Dean (2003/2004) found that the company (i.e., the brand) with an average degree of reputation in the marketplace would benefit from the enhancement of their reputation through the development of CS campaigns in the marketplace. Thus, based on these findings, it seems reasonable to conclude that once the brands are perceived as having at least an average degree of reputation or image by consumers, marketers could effectively utilize CS campaigns to enhance consumer responses to these brands. Of these factors influencing the effectiveness of CSR campaigns, the most frequently explored notion throughout cause marketing literature is brand-cause congruence (Lafferty and Goldsmith 2005; Nan and Heo 2007; Rifon et al. 2004; Simmons and Becker-Olsen 2006). Even though there is still disagreement over its definition, brand-cause congruence is frequently defined as functional or image-based similarities between the brands (i.e., the marketers) and the causes (McDonald 1991; Murphy 1999). As such, brand-cause congruence occurs when the brands are used for the cause campaigns (i.e., functional similarities), and when the core images consumers have for the brands are consistent with those of the causes the brands support (i.e., image-based similarities) (McDonald 1991; Murphy 1999). The main promise of such brandcause congruence literature is that consumers are more likely to respond to the CSR campaigns 29 when they perceive congruence (e.g., a heart-related cause supported by an orange juice brand) between the brand sponsoring and the cause sponsored, than when they perceive incongruence (e.g., a traffic safety-related cause supported by an orange juice brand). This is further suggested as being a strong tendency, even when the brand-cause congruence is created by the advertising message that explains how the brand is congruent with the cause. For example, Alpo, which sponsored Special Olympics, made a connection with Special Olympics by stating in its advertisement that caring for pets increased the self-esteem of the mentally disabled (Simmons and Becker-Olsen 2006). Furthermore, Nan and Heo (2007) suggest that the brand-cause congruence is effective in increasing consumer responses to CRM, only for those high in brand consciousness, indicating that brand consciousness is another brand-related feature in the CRM context. These previous studies on brand-cause congruence provided significant implications for cause marketing literature, particularly pertaining to how the selection of brands could be a matter for the effective provision of CSR campaigns. Nevertheless, there has been no clear answer in academia regarding whether the images of brands could be a consideration in selecting the brands, which would modify the effectiveness of CSR campaigns in the social media setting. Moreover, consumer psychology literature suggested that the brand image could be one of the key factors to mobilize consumers‘ brand usage and consumption (Dolich 1969; Graeff 1996; Sirgy and Danes 1982), particularly in a public setting (Graeff 1996; Sirgy and Danes 1982). Further, as noted earlier, the present study emphasizes the previous findings, that consumers in social media are willing to be connected with brands and express such connections with these brands to other people, typically by joining the brand pages (ExactTarget 2010). The work of 30 Yoon and Kim (2008) also suggests that consumer consciousness of their public image was a significant factor predicting their responses to CRM campaigns. Taken together, the current study suggests that consumer participation in CSR campaigns in social media would be moderated by several situational factors, which may be related to consumers‘ motivation to manage their impression in the social media context. To obtain a more complete picture about when consumers are motivated to participate in CSR campaigns by joining the brand pages and inviting friends to the brand pages featuring CSR campaigns in social media, it seems vital to explore these situational factors more rigorously in academia. Here, two situational factors, based on the roles of friends and brand images in the social media setting, are under study: types of self-friend gender composition and types of brands. Types of Self-Friend Gender Composition and Types of Brands As the first potential factor, the current study considers the types of self-friend gender composition: whether consumers are invited to the brand pages by the same- or different-gender friends. The benefit of social media in promoting any persuasive messages, even for the brands, may reside in the usage of friendship (Morrissey 2010). Thus, the current study focuses on the role of friends, particularly the gender types of friends. It further suggests that whether consumers are invited to the brand pages by the same- or different-gender friends would be one of the key contextual factors potentially influencing the intention of consumers to participate in CSR campaigns by responding to the brands positively. Theoretical rationales for the moderating role of self-friend composition could be provided by social psychology, particularly impression management literature (Glass et al. 1976; Leary 1983; Leary et al. 1994; White and Dahl 2006). Social psychologists suggest that in our daily communication with the same- or different-gender target, we are more motivated to deliver 31 a favorable impression when we interact with the different-gender target, as opposed to when we interact with the same-gender target (Glass et al. 1976; Leary 1983; Leary et al. 1994; Rind and Benjamin 1994). Accordingly, the present study suggests that the proposed effectiveness of CSR campaigns on consumer responses to brands would be different whether the consumers are invited by different-gender friends or same-gender friends. Second, as another factor moderating the effectiveness of CSR campaigns, the present study proposes the types of brands: whether the brands featuring CSR campaigns are perceived as having symbolic or practical images among consumers. As stated briefly, consumer psychology literature categorized brands into two types based on their two distinctive images — symbolic and practical (Bhat and Reddy 1998; Park, Jaworski, and Maclnnis 1986). Further, consumers are likely to express themselves in terms of the images projected by the brands they use and purchase in the public setting (Graeff 1996; Sirgy and Danes 1982) and in the social media setting (ExactTarget 2010). Likewise, the present study investigates whether consumer participation in CSR campaigns in social media would differ based on the types of brands; namely, whether the brands are perceived as having symbolic or practical images. The investigation of two such contextual factors (i.e., the types of self-friend gender composition and the types of brands) may be more meaningful for the CSR campaigns that target college students. College students actually account for a considerable number of social media users: for example, 40% of Facebook users, 37% of YouTube users, and 38% of Twitter users were college students in 2010 (Quantcast 2010). Furthermore, as suggested previously, in social media such as Facebook and MySpace, college students are likely to manage the image they present to other people and control their profile content, which is set for others to view (Lampe et al. 2007; Krämer and Winter 2008). Thus, it is highly probable that the motivation of college 32 students to look favorable to their friends would increase their favorable responses to brands that support causes. Findings of the present study could therefore be beneficial for scholars and marketers to make more accurate predictions of when young consumers are likely to be motivated to participate in CSR campaigns and eventually connect with brands in the social media context. All such discussions will be based on impression management theory as a theoretical framework, and all of the hypotheses will be refined according to this theory in the next chapters. 33 CHAPTER III THEORETICAL FRAMEWORK Impression Management Theory Impression management is defined as ―the process by which individuals attempt to control the impressions others form of them‖ (Leary and Kowalski 1990, p. 34). Individuals‘ behaviors performed in a public setting seem to be constrained by their awareness that the behaviors may be visible to other people (Bushman 1993; Leary and Kowalski 1990), and by their concern that the impression others form of them may not be favorable (Schlenker and Leary 1982). Consistent with the majority of social psychology literature (Goffman 1959; Guadagno and Cialdini 2007; Leary and Kowalski 1990), the current study uses the terms impression management and self-presentation interchangeably. The general goal of impression management is to enhance individuals‘ images to other people in a social context by performing behaviors based on the belief in how others evaluate the self (Leary and Kowalski 1990). For a further scholastic effort in conceptualizing impression management, Leary and Kowalski (1990) suggested a model of impression management consisting of two components: impression motivation and impression construction. Impression motivation is defined as the degree to which the desire to control the image projected to others is activated. Impression construction is defined as the selection of what impression is appropriate to convey to others and how to convey that impression. According to this model, the present study proposes that in the social media setting, consumers are first motivated to monitor the favorable impression they will convey to their friends by joining the brand pages featuring CSR campaigns, and then decide their responses to the brands (e.g., joining the brand pages, inviting others to the brand pages) in social media. 34 However, individuals are motivated to present specific images (e.g., competent images, warm images) based on a given situation (Jones and Pittman 1982). That is, the impressions individuals like to present are frequently determined by the characteristics of the targets (e.g., boss, romantic partner) and the goal of impression management (e.g., career success, romantic relationship) (Jones and Pittman 1982). Nevertheless, it should also be noted that most people attempt to look desirable in any social context (Guadagno and Cialdini 2007; Leary and Kowalski 1990). Thus, impression management can occur together with social desirability (Crowne and Marlowe 1960; Guadagno and Cialdini 2007). As a close construct to impression management, social desirability is defined as an individual‘s tendency to describe themselves and behave in a manner in which they believe they will be viewed favorably for a situation (Crowne and Marlowe 1960). Given the fact that any kind of support for charitable causes tends to be considered a good citizenship behavior, the favorable responses of consumers to the brands that support causes (e.g., joining the brand page, inviting others to the brand pages) could also be socially desirable activities that would be seen favorably in general to others. Thus, the present study suggests that consumers expect to look favorable to other people by joining Facebook brand pages that support causes. Impression Management in Social Media Impression management has been examined in offline contexts such as career success (e.g., job interviews) (Judge and Bretz 1994; Kacmar, Delery, and Ferris 1990; Wayne and Ferris 1990), health-related behaviors (e.g., unnecessary dieting) (Leary, Tchividijian, and Kraxberger 1994), and prosocial behaviors (e.g. blood donation) (Ariely et al. 2009; Lacetera and Macis 2008; White and Peloza 2009). With the popularity of user-generated content within online media (e.g., blogs, social networking sites), scholars have started to examine impression 35 management in an online setting. Researchers have suggested that self-presentation in online media platforms is made to be more asynchronous (e.g., people can edit and update selfpresentational cues deliberately over time), malleable (e.g., they can simply manipulate such cues), and selective (e.g., they can select specific cues to enhance their impression) than is selfpresentation in an offline, face-to-face communication setting (Walther 1996). Due to these characteristics, online media platforms are expected to provide people with a greater opportunity for impression management (Walther 1996). Among the various types of online media channels, social media, particularly SNSs, have frequently been considered a media channel for impression management (Boyd and Ellison 2007; Lamp et al. 2007; Krämer and Winter 2008), probably due to the ―public display‖ nature of SNSs (Boyd and Ellison 2007). Boyd and Ellison (2007) noted the public display of connection within SNSs as a unique feature of SNSs, which enabled users to articulate and make visible their social networks. Indeed, impression management is suggested as a main motive for using SNSs (Boyd and Ellison 2007). SNS users attempt to control the impressions that others form of them by making selective personal information public and visible to others on their profile pages. For example, Lampe and his colleagues (2007) found that many Facebook users provided personal information on their profile pages such as ―Likes and Interests,‖ and the act of populating these profile pages was strongly related to their number of friends on Facebook. Lampe et al. (2007) suggested that people can ―articulate many preferences that shape the public persona they are trying to present to others‖ (p. 6) on the profile pages to signal their identity to other people. Similarly, Krämer and Winter (2008) reported that individuals‘ impression management behaviors were prevalent on SNSs. They found that individuals, especially those high in selfefficacy with regard to impression management, had a desire to create and modify their profile 36 pages (e.g., update the level of profile details, such as a profile picture and membership, on a regular basis) to deliver a positive image to other SNS users. Cause Support for Impression Management Consumers are likely to present their connections with brands to others in the social setting, particularly on Facebook (ExactTarget 2010). The present study expects that such likelihood to associate themselves with brands would be higher when the brands support causes in social media. Consumers would be given an opportunity to present favorable images to others, as those who are connected with brands supporting causes (e.g., by becoming fans of the brands). Accordingly, the study proposes that consumer intentions to join brand pages and invite friends to the brand pages would be intensified when the brands feature CSR campaigns on their brand pages. Indeed, giving rewards has been found to increase support for charities (Ariely et al. 2009; Lacetera and Macis 2008). Such positive effects of rewards on prosocial behaviors are more pronounced when the rewards are not monetary, but symbolic (Deci, Koestner, and Ryan 1999; Lacetera and Marcis 2008). Studies have shown that material rewards (e.g., money, food, cup) for the support of cause campaigns often undermined individuals‘ altruistic motivation and subsequently decreased their support (Benabou and Tirole 2006); whereas symbolic rewards (e.g., medal, verbal announcement, public attention) tended to increase the supporters‘ selfesteem and encouraged them to support the cause campaigns (Deci, Koestner, and Ryan 1999; Lacetera and Marcis 2008). Similar to symbolic rewards, an opportunity to publicize such a ―connection‖ with a brand can offer consumers the social incentive of ―looking favorable‖ to others, when the brand features CSR campaigns on their social media pages. 37 Consumers are likely to increase their support for cause or charity campaigns when they realize that their participation in the campaigns is visible to other people (Ariely et al. 2009; White and Peloza 2009) or covered by media (Lacetera and Macis 2008). For example, Lacetera and Macis (2008) found that when medals were given to blood donors publicly (e.g., pictures and names of donors covered in a local newspaper), people were more likely to participate in a blood donation campaign than when the medals were given to the donors privately (e.g., the donors picked up the medals at the charity office). White and Peloza (2009) also suggested that a helpothers appeal in charity advertisements (e.g., ―help those less fortunate‖) was more effective than a help-self appeal (e.g., ―build your resume by developing and practicing job skills‖) in increasing consumer intention to volunteer, when consumers were motivated to manage their impression by situational and psychological factors: when consumers were in a public condition (e.g., they were asked to discuss their volunteering intentions with other participants in the experiment setting) (Study1); when consumers were aware that the helping behaviors would be visible to others (e.g., they were aware that their responses were recorded with a video camera) (Study2); and when consumers were chronically high in public-self consciousness (Study3). In a similar vein, many cause campaigns attempt to maximize public displays of campaign participation, such as publicizing supporters‘ names, positions, and pictures in the media (Beckwith 2006). Although the effect of publicity on cause campaign participation has been extensively investigated in cause literature (e.g., Lacetera and Macis 2008; White and Peloza 2009), little is known about the role of publicity in motivating consumers to respond to brands featuring CSR campaigns in social media. However, marketers put special interest in social media as a new marketing channel facilitating the active engagement of consumers to the brands (Abrams 2011; 38 ExactTarget 2010; Maul 2011; Williamson 2010). Indeed, on Facebook, once consumers join the brand pages with CSR campaigns, that joining behavior is displayed automatically on their profile pages and their friends‘ News Feeds. Similarly, as consumers invite their friends to brand pages with CSR campaigns, these invitation activities are also visible to their friends who receive the invitation. These features available on social media may encourage consumers to boost the impression that their friends form of them in a favorable light, by making themselves members of the brands in the social media channels. Direct versus Indirect Tactics of Impression Management It is further suggested that people are likely to adopt direct or indirect tactics of impression management in a social setting (Cialdini 1989; Richardson and Cialdini 1981; Rosenfelf, Giacalone, and Riordan 1995). Direct tactics of impression management refer to a direct method by which information related to one‘s own traits, abilities, and accomplishments is purposely presented to others as a way to influence one‘s image on others (Richardson and Cialdini 1981). Jones and Pittman (1980) identified five direct tactics people frequently utilized to control their impression in a positive light: ingratiation, self-promotion, exemplification, intimidation, and supplication. The ingratiation tactic indicates boosting an individual‘s attractiveness to others who control desired rewards, possibly by showing an interest in his/her colleagues‘ personal lives to show others that s/he is friendly. The self-promotion tactic, defined as emphasizing a certain positive point of his/her personality or ability, is demonstrated, for instance, by making others aware of his/her accomplishments. The exemplification tactic indicates a class of strategic behaviors that elicit the impression of moral worthiness and integrity while creating a feeling of guilt among others. For instance, people may stay at work late to 39 exemplify their hard working quality to their colleagues. The intimation tactic refers to a way to arouse fear among other people to control them and exercise his/her powers, possibly such as letting others know s/he can make things difficult for them if they push him/her too far. Finally, the supplication tactic is defined as a class of strategic behaviors advertising a certain weak part of his/her personality and abilities in order to get sympathy from others. The supplication tactic is easily detected, for example, when people act like they know less than they do, so that others will help them out (Jones and Pittman 1980). Furthermore, each of these tactics is expected to be employed under different motivational mechanisms. For example, people are likely to use ingratiation (e.g., flattery) in order to have others like them, self-promotion (e.g., showing skills) to make others think they are competent, exemplification (e.g., arousing guilt) to let others think they are morally worthy, intimidation (e.g., arousing fear) to make others fear them, and supplication (e.g., showing weakness) to have others feel sorry for them (Jones and Pittman 1982). According to Johns (1990), ingratiation is probably the most frequently used direct tactic of impression management in a social setting. People tend to be perceived as likeable by using various ingratiation tactics, such as showing similar opinions as others, complimenting others‘ opinions, or showing fondness for others (Johns and Pittman 1982; Rosenfield, Giacalone, and Riordan 1995). The psychological mechanism for the effectiveness of ingratiation on making others like us is that we tend to like others who show favor to us or treat us in a favorable manner (Johns and Pittman 1982; Seiter 2007). Along with ingratiation, self-promotion was also noted in literature as another direct tactic frequently carried out, particularly in organizational settings (Judge and Bretz 1994; Kacmar, Delery, and Ferris 1990; Rosenfield, Giacalone, and Riordan 1995; Wayne and Ferris 1990). Judge and Bretz (1994) compared these two direct tactics (i.e., ingratiation and 40 self-promotion) among employees in a company, finding that the ingratiation tactic (e.g., immediately agreeing with supervisors‘ ideas) had positive influences on the extrinsic (e.g., an increase in salary) and intrinsic (e.g., overall job satisfaction) success of the employees who adopted this tactic, whereas the self-promotion tactic (e. g., making others aware of their accomplishments in their job) was not effective in achieving such career success. On the other hand, indirect tactics of impression management have not been actively explored in impression management literature (Guadago and Cialdini 2007; Tal-Or 2010). Cialdini and his colleagues (1976) were the first scholars who documented the usage of indirect tactics to control individuals‘ impressions as seen by others. They suggested that indirect tactics of impression management involve the presentation of information that is not directly related to their own traits, abilities, or accomplishments, but rather indirectly associated with successful or favorable others, groups, or things (e.g., celebrities, sports teams) — even in quite remote and tenuous ways—to others (Cialdini 1986; Cialdini et al. 1976). As a typical example of indirect tactics, Cialdini and his colleagues (Cialdini 1989; Cialdini et al. 1976) have introduced the notion of ―Basking in Reflected Glory,‖ indicating that people are likely to be associated with the glory of successful or favorable others, and then announce that association publicly, even though the glory has neither relevance to nor influence on them. Thus, indirect tactics are also called a created association (Guadagno and Cialdini 2007). Several studies have empirically demonstrated that young people use indirect tactics for their impression management in a social setting (Cialdini 1989; Cialdini et al. 1976; Guadago and Cialdini 2007; Tal-Or 2010). For example, Cialdini and his colleagues (1976) attempted to provide the utilization of indirect tactics among college students for the purpose of impression management. They found that the students at six different universities were more likely to wear 41 apparel that identified their university affiliation (e.g. jackets or t-shirts that displayed the university name, team nickname or mascot, or university insignia) on Monday morning when their university teams had won the football game on the preceding Saturday, than when their university teams had lost (Experiment 1); furthermore, the college students more frequently used the pronoun ―we‖ to describe the outcomes of their university teams (e.g., ―we won,‖ ―we got beat‖) after the game, when their university teams won than when theirs lost (Experiment 2). These findings indicate that people are likely to control their impression by publicizing their association with favorable others, even when there is only a slight link between themselves and the favorable others. Indeed, it was suggested that even such a weak association between self and the favorable others could be perceived as being similar to each other. For example, when an audience was informed that a current player came from the same town as a famous football player, the current player was perceived more favorably among the audience (Gilovich 1981). As such, in an attempt to deliver their impression favorably in a social setting, people are likely to use not only direct tactics by presenting the information about their own skills or favorable traits, but they also tend to utilize indirect tactics by associating themselves with those who are perceived to be skillful or possess favorable traits among the public. One key difference between the direct tactic and the indirect tactic of impression management is that the direct tactic focuses on presenting more specific, concrete, and particularistic information about the self, than the indirect tactic (Carter and Sanna 2008; Pratkanis 2007). For example, individuals in an organizational setting were likely to directly present specific and favorable descriptions of their own accomplishments in order to look competent (Judge and Bretz 1994). On the other hand, the indirect tactics emphasized the more global, abstract, and associated presentation of the self, than did the direct tactics (Carter and Sanna 2008; Pratkanis 2007). As seen in the findings of Cialdini 42 and his colleagues (1976), college students were likely to wear the university team clothing or use the pronoun ―we‖ to indirectly associate them with the university team that won the game. The present study applies the notions of direct and indirect tactics of impression management for the consumer responses to two different types of CSR campaigns (i.e., CRM and CS) in social media. This study argues that joining and inviting activities on the brand pages featuring CRM campaigns would be adopted as a direct tactic to present consumers‘ own contribution to the cause (e.g., adding a $1 donation). By participating in the CRM campaign, consumers would consider themselves as those who add their own help for the cause and further expect to be seen as cause supporters by their friends in social media. On the other hand, it is also suggested that the joining and inviting activities of consumers on the brand pages featuring CS campaigns would be utilized by consumers as an indirect tactic of impression management. Becoming members of brand pages featuring CS campaigns would be helpful for consumers to enhance their image, but such enhanced images are associated with the brands‘ favorable images, rather than their own image as cause supporters. As such, since CS campaigns tend to be made regardless of consumer responses to the brands (e.g., joining the brand pages), consumers who join the brand pages would be seen by their friends as someone who likes the brand‘s cause support, rather than someone who contributes to the brand‘s support. Further details will be provided in the following chapter, along with the relative effectiveness of CSR campaigns on the intention to join and the intention to invite. Table 1 shows the summary of key concepts and their working definitions in the present study. 43 Table 1 Summary of Key Concepts and Definitions Key Concept Corporate social responsibility (CSR) campaigns Definition A commitment of [a corporation] to improve community wellbeing through discretionary business practices and contributions of corporate resources. Cause-related marketing (CRM) The process of formulating and implementing marketing activities that are characterized by an offer from the firm to contribute a specific amount to a designated cause, when customers engage in revenue-providing exchanges that satisfy organizational and individual objectives. Cause-related marketing (CRM) in social media The marketing campaign of firms using their social media brand pages, which is mainly designed to enhance consumer engagement in their brands in the social media setting and typically performed by monetary or product donations from the firms to designated causes, as an exchange of customers‘ membership of their social media brand pages. Corporate sponsorship (CS) An investment in cash or in kind [of corporations] in activities in return for access to the exploitable commercial potential associated with those activities. Corporate sponsorship (CS) The marketing campaign of firms using their social media brand in social media pages which is mainly designed to enhance consumer engagement in their brands in the social media setting, and mainly performed by monetary or product donations from the firms to designated causes as a way to promote their philanthropic commitment in the cause. Social media A group of Internet-based applications that build on the ideological and technological foundations of Web 2.0, and that allow the creation and exchange of User Generated Content. Social networking sites (SNSs) Web-based services that allow individuals to (1) construct a public or semi-public profile within a bounded system, (2) articulate a list of other users with whom they share a connection, and (3) view and traverse their list of connections and those made by others within the system. Impression management The process by which individuals attempt to control the impressions others form of them. Impression motivation The degree to which the desire to control an individual‘s image projected to others is activated. 44 Table 1 (cont’d) Impression construction The selection of what impression is appropriate to convey to others and how to convey that impression. Indirect tactics of impression management The presentation of information that is not directly related to one‘s own traits, abilities, or accomplishments, but rather indirectly associated with successful or favorable others, groups, or things (e.g., celebrities, sports teams) — even in quite remote and tenuous ways — to others. Direct tactics of impression management The presentation of direct method by which information related to one‘s own traits, abilities, and accomplishments is purposely presented to others as a way to influence one‘s image on others. Outcome expectancy The perceived probability of successfully claiming the [self] image. Self-friend gender composition The gender composition (i.e., same and different) between the consumers and the friends who invite the consumers to the brand pages in social media. Symbolic brands The brands that satisfy consumers‘ symbolic (or self-expressive) needs such as desires for products that fulfill internally generated needs for self-enhancement, role position, group membership, or ego-identification. Practical brands The brands that tap into consumers‘ functional (or utilitarian) needs that motivate the search for products that solve consumption-related problems. 45 CHAPTER IV HYPOTHSIS DEVELOPMENT Main Effects of CSR Campaign Conditions As discussed earlier, consumer response to a brand is considered as the main criterion to determine the effectiveness of CSR campaigns in the marketplace. In the current study, the main effects of CSR campaigns were examined according to two aspects: (a) whether CSR campaigns (i.e., the presence versus the absence of CSR campaigns) would be effective and (b) which types of CSR campaigns (i.e., CRM campaigns versus CS campaigns) would be effective on the favorable responses of consumers to a brand. Again, the consumer response to a brand is specified in the current study as the consumer intention to join the brand page in social media and their intention to invite friends to the brand page in social media. First, this study proposes that consumers would be more likely to click the ―Like‖ button to join a brand page and invite their friends to a brand page that promotes either a CRM or a CS campaign, compared to a brand page that does not have such a campaign. Indeed, the ―Like‖ button was developed for Facebook users to share specific content they like with their friends (Facebook Like Button Developers 2010). For example, once consumers click the ―Like‖ button on a brand page containing a CRM or CS campaign, they end up joining the brand page as fans of the brand. Further, basic descriptions of what the consumers like (e.g., information about the brand‘s CSR campaigns) appear on their own profile pages and the friends‘ News Feed pages (Facebook Like Button Developers 2010). These features available in social media provide consumers with an opportunity to present favorable images to their friends as those who join brand pages that support causes. Furthermore, inviting others to the brand page could also be utilized by consumers to manage their impression in a favorable light in social media. 46 Consumers could be seen by their friends as those who encourage others to join the brand pages supporting the causes. Therefore, the provision of any kind of CSR campaign (e.g., a CRM or CS campaign) on the brand pages in social media is expected to lead consumers to respond to the brands more favorably, resulting in greater intentions to join the brand pages and invite friends to the brand pages among consumers, than the absence of CSR campaigns on the brand pages. As a further investigation, the current study focuses on the relative effectiveness of these two different types of CSR campaigns (i.e., a CRM campaign and a CS campaign) on the intentions to join the brand page and invite friends to the brand pages. As noted previously, the study argues that consumer responses to the brands (i.e., joining and inviting on the brand pages) featuring a CRM campaign can be adopted as a way to present themselves in a favorable way, as those who make a contribution to the brand in support of the causes, possibly by adding a $1 donation. Thus, in the present study, the CRM campaign is considered as providing consumers with an opportunity to manage their impression directly to others by showing their own contribution to the cause. On the other hand, consumer responses to the brand page featuring a CS campaign can also be considered as a way to manage their impression to others favorably. However, such a favorable image may be limited in that it is associated with the positive image of the brands that sponsor the causes, rather than being reflected directly from their own positive image as a cause contributor. Thus, the present study considers the CS campaign in social media as providing consumers with an opportunity to manage their impression indirectly, by associating themselves with the brands that support causes (e.g., their liking of the brand‘s CS campaign). Taking all of this into account, consumers may consider joining the brand page with the CRM campaign (versus the CS campaign) and inviting others to the brand page with the 47 CRM campaign (versus the CS campaign) as a more direct and effective way to deliver a favorable impression to their friends. Taken together, as illustrated in Figure 1, the current study posits that consumers would be more likely to join the brand page and invite their friends to the brand page, when the social media brand page features any kind of CSR campaign (versus when the brand page provides none of the CSR campaigns). It further posits that when the brand page provides the CRM campaign (versus the CS campaign), the intention to join and the intention to invite will be more intensified. Therefore, the study suggests the following hypothesis for the main effects of CSR campaign conditions on consumer responses to the brands in social media (H1a-b). H1a-b. Consumers who are invited to a brand page with the CRM campaign will generate (a) the highest intention to join the brand page and (b) the highest intention to invite others to the brand page, followed by those who are invited to the brand page with the CS campaign, and those who are invited to the brand page without either the CRM campaign or the CS campaign. Mediating Role of Outcome Expectancy For further investigation, the present study explores the consumer expectation that their joining and inviting activities on the brand pages would result in the projection of images they attempt to convey to their friends, and how this would mediate the proposed main effects of CSR campaign conditions on the intention to join the brand page and their intention to invite their friends to the brand page. That is, prior to deciding their intention to join and their intention to invite, consumers may consider whether these activities would help them deliver a positive image to their friends as someone who makes their own contribution to the cause through the CRM campaign, or someone who likes the brand with the CS campaign. As indicated earlier in Table 1, the notion of outcome expectancy is defined as ―the perceived probability of successfully claiming the [self] image (Greenberg, Pyszczynski, and 48 Stine 1985, p. 2)‖ in the impression management context. Originally, the notion of outcome expectancy was proposed from the concept of self-efficacy (Bandura 1977). As the belief that certain behaviors would or would not lead to given outcomes, the outcome expectancy was considered as one key component to motivate and guide the anticipated behaviors (Bandura 1977). The notion of outcome expectancy was further adopted into the expectancy-value theory (Fishbein 1967; Rotter 1954; Vroom 1964), suggesting that the expectation that certain behaviors can yield specific outcomes will increase the value of outcomes and the motivation to perform the behaviors. Furthermore, when it was later translated into impression management contexts, the outcome expectancy was suggested to influence the motivation and attempt of people to present a particular image in a given situation (Greenberg et al. 1985; Leary, Kowalski, and Campbell 1988; Schlenker 1980; Schlenker and Leary 1982). It is thus noted that the socially favorable impressions people want to project to others are determined by their perceived, anticipated, or imagined outcome expectancy (Schlenker 1980; Schlenker and Leary 1982). Several studies about individuals‘ online behaviors (e.g., dating, shopping) have also explored the outcome expectancy. Gibbs and his colleagues (2006) suggested that the outcome expectancy of success in online dating had an influence on online dating site users‘ impression management behaviors. Gibbs and his colleagues (2006) found that online dating site users who expected positive outcomes from online dating (e.g., long-term relationship involving anticipated face-to-face interactions) were more likely to engage in disclosing personal information in intentional, conscious, and positive manners than those who did not predict such positive outcomes. In a similar vein, LaRose and Eastin (2002) suggested that consumers were likely to anticipate their online shopping outcomes in the e-commerce setting (e.g., convenient with cost savings) prior to their online shopping. They found that such outcome expectancy was positively 49 related to consumers‘ online shopping activities such as the amount spent on online purchases and the number of products purchased online. The current study proposes that when brand pages in social media provide consumers with an opportunity to be involved in either a CRM campaign or a CS campaign, consumers may expect a positive social outcome from joining the brand page or inviting friends to the brand page, such as looking favorable among their Facebook friends. The extent to which they expect the outcome of their joining and inviting behaviors to be positive, may dictate their intention to join the brand page and their intention to invite friends to the brand page. Thus, an opportunity on the brand page to publicize both their support for the cause via the CRM campaign or their liking of the brand that provides the CS campaign, would increase consumer expectancy to be seen as favorable to their friends, which would in turn increase consumer intentions to join and invite on the brand page. Furthermore, regarding the relative effects of both CSR campaigns, the present study predicts that a direct contribution (e.g., adding $1) to the cause via participation in the CRM campaign would be more likely to be effective for consumers to expect that they can be seen favorably as cause supporters to their friends, as opposed to simply an expression of their liking of the brand that provides the CS campaign. Thus, the superior effect of CRM campaigns (over CS campaigns) was hypothesized for the outcome expectancy (H2); in turn, such expectancy to look favorable as the outcome of joining the brand pages and inviting their friends to the brand pages, was posited to influence the consumer intention to join the brand page and the consumer intention to invite their friends to the brand page. Accordingly, to investigate a mediating mechanism of outcome expectancy in the relationship between the CSR campaign conditions and the consumer response to the brand empirically, the present study proposes a mediation model, suggesting that the outcome expectancy will mediate the proposed main effects 50 of CSR campaign conditions on the intentions to join the brand page and invite their friends to the brand page in social media. H2. Consumers who are invited to a brand page with the CRM campaign will be most likely to expect that their joining or inviting on the brand page will make them look favorable to others, followed by those who are invited to the brand page with the CS campaign, and those who are invited to the brand page without either the CRM campaign or the CS campaign. H3a-b. Outcome expectancy to look favorable will mediate (a) the main effect of CSR campaign conditions on the intention to join the brand page (as seen in H1a), and (b) the main effect of CSR campaign conditions on the intention to invite friends to the brand page (as seen in H1b). Next, the present study attempts to identify when consumers are more or less motivated to look favorable, join the brand page that supports a cause, and invite their friends to the brand page supporting a cause. As the moderating variables influencing the proposed main effects of CSR campaign conditions on the dependent variables, two situational factors often encountered in the Facebook setting are suggested in the present study: (1) the types of self-friend gender composition; (2) the types of brands. Moderating Roles of Types of Self-Friend Gender Composition The present study proposes that the self-friend gender composition is one of the potential factors that might moderate the effectiveness of CSR campaigns on brands in social media. Gender composition is defined in a social context as the identification of the gender types (i.e., male or female) of two parties who communicate with each other, and specified as same- and different- gender composition in terms of the comparison of the gender types of two parties (Leary et al. 1994). Accordingly, the self-friend gender composition could be defined in the current study as the gender types of the consumers and their friends who invite them to brand 51 pages, and further specified as two types (i.e., same and different) based on whether the consumers and the friends have the same or different gender types. Self-friend gender composition between social media users could be a considerable factor among marketers, given the fact that social media are beneficial for marketers because of the usage of friendship within social media to promote brands (Morrissey 2010). Indeed, sending an invitation note to a specific brand page is prevalent and becomes a key benefit for marketers to use social media to advertise their CSR campaigns (Morrissey 2009). A large number of social media users are likely to invite their friends and be invited by their friends to brand pages. In fact, Facebook users shared more than 30 billion pieces of content, such as Web links, news stories, and posts, each month in 2010 (see Facebook Statistics (2011) for more). Facebook brand pages frequently encourage consumers to send their friends an invitation note via multiple links such as ―Share,‖ ―Suggest to Friends,‖ ―Send to Friends,‖ and ―Invite Friends.‖ The current study considers that invitations from friends may make consumers feel that their joining and inviting behaviors are monitored and visible by their friends in social media; and furthermore, when the different-gender (versus the same-gender) friend sends the invitation e-mail, such a feeling will be more (versus less) heightened. Thus, it should be reasonable to posit that consumers will be more likely to join the brand pages and invite friends to the brand pages with CSR campaigns, when they are dealing with different-gender friends (versus samegender friends). Some literature on impression management (Glass et al. 1976; Leary 1983; Leary et al. 1994; White and Dahl 2006) has consistently demonstrated that people are more motivated to manage their impression when they interact with different-gender partners, as opposed to samegender partners. Leary and his colleagues (1994) found that when college students had an 52 encounter with the different-gender target, they were likely to increase their concern for impression management by focusing on how the target would perceive them. The students were then motivated to be perceived as being favorable by the target, regardless of how familiar they were with the target. On the contrary, when they interacted with the same-gender target, such tendencies were detected only when they interacted with the same-gender target with whom they were unfamiliar. Similarly, Rind and Benjamin (1994) found that in the presence (versus the absence) of a female companion, male shoppers in a shopping mall food court were concerned about their public image as seen by their female companion. This produced more compliance to buy raffle tickets, after the male shoppers were informed that the seller needed the most tickets to win a cash prize which would be spent helping causes (Rind and Benjamin 1994). For possible reasoning of such effects of self-partner gender composition on impression management, scholars have explained that the encounter with the different-gender partner is expected to have greater social outcomes, in general, than that with the same-gender partner (Glass et al. 1976; Leary and Kowalski 1997; Leary et al. 1994). For example, people tend to expect that they could have their future romantic, sexual, or marital relationship from their interaction with the different-gender partner; however, such an expectation typically does not occur when they interact with the same-gender partner (Leary et al. 1994). Thus, more concern and self-regulatory attention is paid to convey the desirable impression when people interact with different-gender individuals. Particularly, for unmarried adults such as college students, such patterns seem to be intensified. Outcomes for the encounter with the different-gender individuals are considered by unmarried people as being so valuable, that the unmarried people are likely to focus on leaving favorable impressions on the different-gender individuals (Leary et al. 1994; Leary and Kowalski 1997). 53 For general discussions, impression management literature also indicates that characteristics of others are key elements that influence one‘s motivation for impression management (Schlenker and Leary 1982). Usually, ―people who are powerful, esteemed, attractive, expert, or high in status can mediate a variety of worthwhile outcomes; they can often mediate material gains or losses (e.g., promotions, raises, awards, dismissal, physical harm); their approval, respect, friendship, and assistance are highly valued; and their opinions are viewed as salient in confirming or disconfirming one‘s purported strengths or weaknesses (Schlenker and Leary 1982, p. 647).‖ Based on these previous suggestions, the present study proposes that when consumers are invited by the other-gender friend (versus the same-gender friend) to the Facebook brand page, they are more likely to increase their concern for their impression as reflected in the eyes of the friend, and they are more (versus less) motivated to present their socially favorable images to them. Therefore, the current study suggests that when invited by the different-gender friend (versus the same-gender friend), consumers would expect greater outcomes from their joining and inviting activities, and consequently generate greater intentions to join the brand page and invite their friends to the brand page, when the brand page features either a CRM campaign or a CS campaign. Furthermore, regarding the relative effects of two types of CSR campaigns, the CRM campaign would be expected to have greater outcomes from joining and inviting behaviors on the brand page, than the CS campaign. With regard to the CRM campaign, consumers are likely to be seen by the different-gender friend as those who make a direct contribution to the cause through their participation in the CRM campaign, whereas in the case of the CS campaign, consumers are merely seen as those who are associated with a brand that provides the CS 54 campaign. However, when invited by the same-gender friend, their status as those directly contributing to the causes or being indirectly associated with the brand‘s sponsorship would not be as significant among consumers. Therefore, the proposed main effects of CSR campaign conditions on the outcome expectancy, the intention to join, and the intention to invite will be more pronounced when consumers are invited to the brand page by the different-gender friend, as opposed to when they are invited by the same-gender friend (H4). Furthermore, to elaborate on the mediating role of outcome expectancy in the hypothesized interaction model, the present study posits the moderated mediation model, predicting that the outcome expectancies will mediate the proposed interaction effects of the CSR campaign conditions and the types of self-friend gender composition on the intention to join and the intention to invite (H5). Figure 1 demonstrates the visualized summary for these hypotheses. H4a-c. The proposed main effects of CSR campaign conditions on (a) the outcome expectancy to look favorable (as seen in H2), (b) the intention to join the brand page (as seen in H1a), and (c) the intention to invite their friends to the brand page (as seen in H1b), will be more pronounced when consumers are invited to the brand page by the different-gender friend (versus the same-gender friend). H5a-b. The outcome expectancy to look favorable will mediate the proposed interaction effect between the CSR campaign conditions and the types of self-friend gender composition on (a) the intention to join the brand page (as seen in H4b) and (b) the intention to invite their friends to the brand page (as seen in H4c). Moderating Roles of Types of Brands As the second situational factor moderating the proposed main effects of CSR campaign conditions on the outcome expectancy, the intention to join, and the intention to invite, the present study proposes the types of brands. Consistent with consumer psychologists (Bhat and Reddy 1998; Park et al. 1986), the present study suggests that brands can be categorized into two types based on their two distinctive images: symbolic or practical. 55 Symbolic brands tend to satisfy consumers‘ symbolic (or self-expressive) needs such as ―desires for products that fulfill internally generated needs for self-enhancement, role position, group membership, or ego-identification (Park et al. 1986, p. 136).‖ Thus, symbolic brands are mostly consumed for the enhancement of self-image and social image, and their practical usage is only incidental (Bhat and Reddy 1998). On the contrary, practical brands are likely to tap into consumers‘ functional (or utilitarian) needs that ―motivate the search for products that solve consumption-related problems (Park et al. 1986, p. 136).‖ Thus, practical brands are primarily consumed on the basis of immediate and practical needs (Bhat and Reddy 1998). It should also be stated that in many cases, brands within the same product category (e.g., watch) are distinctively divided into either symbolic (e.g., Rolex) or practical (e.g., Timex) brands (Bhat and Reddy 1998; Park et al. 1986). Once the image of the brand is projected in the marketplace, it tends to last over the brand‘s life for the sake of consistency (Park et al. 1986). It should be noted that both symbolic (e.g., Gucci, Chanel, Chantecaille) and functional (e.g., Colgate, Tide, Kraft) brands in various product categories (e.g., fashion, cosmetics, banking, retail stores, drinks, toothpaste, food, detergents) are widely involved in CSR campaigns on Facebook. To the author‘s best knowledge, however, the role of brand image has rarely been examined in advertising or marketing literature, particularly as it pertains to the effectiveness of CSR campaigns on the brands within social media. The role of brand image, however, could be critical in understanding consumer willingness to be connected with the brands in the social media setting, because consumers can publicize such connections with the brands to other people in social media (ExactTarget 2010; Slutsky 2011). Symbolic brands may be more effective than practical brands, in general, at increasing the number of fans for their brands in social media. Even though each brand possesses its own 56 overall abstract image (Dolich 1969), the symbolic brands are more effective in expressing the image of those who purchase or use them, than the practical brands (Bhat and Reddy 1998; Park et al. 1986). Actually, the symbolic brands are divided into two distinctive sub-dimensions, namely prestige and personality expression (Bhat and Reddy 1998). Here, the current study suggests that joining and inviting others to the symbolic brand page in social media may be different from actually owning or consuming the symbolic brand in terms of the prestige image of the brand, but similar in terms of the self-expression image of the brand. Thus, the selfexpressive nature of the symbolic brand may be salient both for consumers who purchase or use the symbolic brand, and for consumers who join the symbolic brand page on Facebook. Thus, the present study explores whether becoming a fan of the symbolic brand page (for self-expression) would be more effective in presenting self-impression to friends in social media, as opposed to becoming a fan of the practical brand page. It should follow the same reasoning that ownership and consumption of symbolic brands are more effective in expressing the impression of owners and consumers in a public setting, than in a private setting (Dolich 1969; Graeff 1996). Therefore, it may be true that the symbolic brands would have a greater number of consumers who become fans of the brands in social media than the practical brands. Accordingly, it seems reasonable for the present study to predict that when consumers are invited to a CSR campaign on a social media page for a symbolic (versus practical) brand, they would expect more pronounced effects of their joining and inviting activities on their impression management, based on the benefit of symbolic brands, such as giving off the image of those who become members of the brands saliently. Consequently, such positive expectations would result in the favorable responses of consumers to join the symbolic brand page and invite friends to the symbolic brand page. That is, when the brands are perceived as symbolic (versus practical), 57 affect transfer may be processed. Affect transfer is the process wherein consumers‘ preexisting affective feelings or images associated with one object are transferred to a closely related object, toward which consumers may not possess previous affective feelings or images (Shimp 1981). Nan and Heo (2007) introduced the notion of affect transfer for the effectiveness of CRM campaigns, suggesting that consumers were likely to have a favorable attitude toward brands sponsoring the cause for NPOs because consumers‘ overall favorable attitude toward NPOs was transferred to the sponsored brands. Similarly, the current study posits that when brands are perceived as symbolic (versus practical), consumers will be more likely to see that the symbolic image of the brands could be transferred to the image of those who become members or invite friends on the symbolic brand pages. Thus, consumers will be more motivated to manage their impression by joining the brand page and inviting their friends to the brand page, when they are invited to the symbolic (versus practical) brand page in social media. On the other hand, the present study also takes into account the counterargument that the effects of CSR campaigns on consumer responses to brands will be greater when the brands have practical images (versus symbolic images). That is, even though symbolic brands are expected to elicit greater intentions of consumers to join their pages and invite others to the pages in social media, joining symbolic brand pages and inviting friends to the symbolic brand pages themselves may be considered by consumers to be a satisfactory condition to make their social image salient to their Facebook friends. Thus, when the symbolic brands feature either CRM campaigns or CS campaigns on their social media Web pages, consumers may not expect their joining and inviting behaviors to be significantly effective in enhancing their images in the eyes of others. This might be because supporting the CRM campaign or liking the brand‘s CS campaign on the symbolic brand page would not provide consumers with added value in terms of their impression 58 management. Likewise, this study explores the ideas that when the brand has symbolic images, consumers‘ outcome expectancy to be seen as favorable in the eyes of their friends, consumers‘ intentions to join, and consumers‘ intentions to invite would not be significantly influenced by the presence and the types of CSR campaigns. Thus, the proposed main effects of CSR campaign conditions on the dependent variables would not be shown significantly on the symbolic brand page. On the other hand, a counterargument would be applied for the situation when the brand pages are for practical brands. The self-expressive image from the practical brand itself may not be saliently considered among consumers. In turn, consumers‘ favorable responses to brands, such as joining and inviting for brands that support causes in social media, may be considered as offering consumers an added benefit to control their social image. Thus, on the practical brand pages, the decisions of consumers regarding their outcome expectancies to look favorable, their intentions to join, and their intentions to invite would be primarily impacted by whether the brand pages present any CSR campaigns, and further, what kind of CSR campaigns are available on the brand pages. As such, the proposed main effects of CSR campaign conditions on the dependent variables will be strengthened on the practical brand pages. Taken together, the present study considers these two arguments reasonable to provide the theoretical accounts for the moderating role of brand types (symbolic and practical) on consumer responses to brands featuring CSR campaigns in social media. Therefore, the present study posits research questions, rather than hypotheses, to ask whether the previously proposed main effects of CSR campaigns on the outcome expectancy — and subsequently the intention to join, and the intention to invite — will be different whether the brands are practical or symbolic (RQ1). Furthermore, the moderated mediation model was asked to explore the mediating 59 mechanism of outcome expectancies in the proposed interaction effects of the CSR campaign conditions, and the types of brands, on the intention to join and the intention to invite (RQ2). RQ1a-c. Will the types of brands moderate the proposed main effects of CSR campaign conditions on (a) the outcome expectancy to look favorable (as seen in H2), (b) the intention to join the brand page (as seen in H1a), and (c) the intention to invite their friends to the brand page (as seen in H1b); and if so, which type of brands, symbolic versus practical, will have the moderation effect? RQ2a-b. Will the outcome expectancy to look favorable mediate the proposed interaction effect between the CSR campaign conditions and the types of brands on (a) the intention to join the brand page (as seen in RQ1b) and (b) the intention to invite their friends to the brand page (as seen in RQ1c) and if so, which type of brands, symbolic versus practical, will have the moderation effect? Figure 1 illustrates the summary of the proposed hypotheses and research questions. Figure 1 Summary of the Hypotheses and Research Questions H1. The main effects of CSR campaign conditions on (a) the intention to join and (b) the intention to invite. (a) Intention to Join (b) Intention to Invite CSR Campaign Conditions H2. The main effects of CSR campaign conditions on the outcome expectancy. Outcome Expectancy CSR Campaign Conditions 60 Figure 1 (cont’d) H3. The mediating role of outcome expectancy in the main effect of CSR campaign conditions on (a) the intention to join and (b) the intention to invite [mediation model]. Outcome Expectancy (a) Intention to Join (b) Intention to Invite CSR Campaign Conditions H4. The moderating roles of types of self-friend gender composition in the main effects of the CSR campaign conditions on (a) the outcome expectancy, (b) the intention to join, and (c) the intention to invite. Types of Self-Friend Gender Composition (Same [-] and Different [+]) (a) Outcome Expectancy (b) Intention to Join (c) Intention to Invite CSR Campaign Conditions H5. The mediating role of outcome expectancy in the interaction effects of CSR campaign conditions and types of self-friend gender composition on (a) the intention to join and (b) the intention to invite [moderated mediation model]. Types of Self-Friend Gender Composition (Same [-] and Different [+]) Outcome Expectancy CSR Campaign Conditions (a) Intention to Join (b) Intention to Invite 61 Figure 1 (cont’d) RQ1. The moderating roles of types of brands in the main effects of CSR campaign conditions on (a) the outcome expectancy, (b) the intention to join, and (c) the intention to invite. Types of Brands (Symbolic and Practical) (a) Outcome Expectancy (b) Intention to Join (b) Intention to Invite CSR Campaign Conditions RQ2. The mediating role of outcome expectancy in the interaction effects of CSR campaign conditions and types of brands on (a) the intention to join and (b) the intention to invite [moderated mediation model]. Types of Brands (Symbolic and Practical) Outcome Expectancy (a) Intention to Join (b) Intention to Invite CSR Campaign Conditions 62 CHAPTER V METHODS Overview To test the previously discussed hypotheses, a 3 (CSR campaign conditions: CRM, CS, or control) x 2 (types of self-friend gender composition: same versus different) x 2 (types of brands: symbolic versus practical) between-subjects experimental design was applied. The present study has one independent variable (i.e., CSR campaign condition), two moderating variables (i.e., type of self-friend gender composition, type of brand), two dependent variables (i.e., intentions to join brand pages, intentions to invite friends to brand pages), and one mediating variable (i.e., outcome expectancy). All of the independent and moderating variables were manipulated. For the sample of the current experimental study, the college students were recruited and then participated by clicking a hyperlinked URL on the research Web site. By contacting the Office of the Registrar at MSU, a total of 5,000 undergraduate students enrolled in various programs at the university were randomly selected and asked to take part in the study. A total of 875 students (17.5% response rate) randomly participated in one of 12 conditions. Finally, 720 students (82.29% out of 875) remained as the final sample after removing those who did not correctly answer at least one of the screening questions to check whether the participants recognized the messages for CRS campaign conditions and the gender types of friends who invited to the brand pages. (See Analytic Strategies for the processes and wording of screening questions.) Stimuli Development 63 Two stimuli were created for the main study: the brand pages in social media and the invitation e-mails sent from friends. First, to manipulate the brand pages in social media, three main components were considered: two types of brands (i.e., symbolic and practical) were selected; a cause supported by the selected brands was chosen; and the manipulations of three conditions of CSR campaigns (i.e., two CSR messages for a CRM campaign and a CS campaign and one neutral message for the control condition) were developed for each of the selected brands. Second, the invitation e-mails were created as another stimulus to manipulate the same and different self-friend gender compositions. Thus, the participants were asked to imagine receiving an e-mail invitation to either one of the six Facebook pages sent from male friends, or one of the six Facebook pages sent from female friends. The gender types of friends were compared with those of the participants, and each of the participants was categorized in either the same or different self-friend gender composition group. In short, the stimuli of the main study consist of the following four components: (1) the selection of brand; (2) the selection of cause; (3) the manipulation of CSR campaign conditions; (4) the manipulation of e-mail invitations for self-friend gender composition. The first two components (i.e., brand and cause) were selected through the pre-test, and the last two (i.e., CSR campaigns conditions and e-mail invitation) were manipulated for the main study. Selection of Brands: Pre-Test Prior to creating the brand pages in social media, two brands and one charitable cause were selected. A pre-test was carried out to select two brands with symbolic and practical images, respectively, under the same product category, and find a charitable cause congruent with these selected brands. A total of 33 college students from MSU were recruited to participate in the pre-test for extra course credit. 64 To identify a product category at first, the participants were asked to indicate their perceived familiarity and perceived relevance with four products (i.e., bottled water, retail stores, soft drinks, sporting goods). These four products were adopted from the existing studies as those familiar and relevant to college students (Anderson Analytics 2006, 2010; Paek, Choi and Nelson 2010; Tobin 2010). The selection of products relevant and familiar to study participants could be important in increasing the accurate responses of the participants to the stimuli. One single item was provided to measure perceived familiarity with each product on a seven-point scale, with endpoints of ―not familiar (1)‖ and ―familiar (7).‖ Similarly, another single item was asked to measure perceived relevance to each product with 1 being ―not relevant‖ and 7 being ―relevant.‖ These single items were adopted from Nan and Heo (2007). The results showed that, out of the four products, bottled water had the highest mean in terms of both perceived familiarity (M = 6.30) and perceived relevance (M = 5.61). Furthermore, the participants were asked to indicate how they perceived the brand images for each product. Under each product, the most popular four to five brands were selected based on recent industry reports (Interbrand 2010; Vitrue 2010): five brands for bottled water (i.e., Aquafina, Evian, Dasani, Ice Mountain, Fiji Water); four brands for retail stores (i.e., Kmart, Meijer, Walmart, Target); four brands for soft drinks (i.e., Dr Pepper, Mountain Dew, Monster Energy Drink, Red Bull); and five brands for sporting goods (i.e., Adidas, Fila, New Balance, Nike, Puma). The popularity of the brands was considered here to avoid any chance that the participants would not recognize the brand image because of unfamiliarity with the brands. To measure the symbolic image of the brands, two items, adopted from Bhat and Reddy (2008) were asked for each brand: ―People drink X brand as a way of expressing their personality‖; and ―Drinking X brand says something about the kind of person you are.‖ To measure the practical 65 image of the brand, two other items (Bhat and Reddy 2008) were provided: ―X brand is for people who are down-to-earth‖; and ―Drinking X brand is practical.‖ Each item was measured on a seven-point Likert scale, anchored from 1 being ―strongly disagree‖ and 7 being ―strongly agree.‖ The results of a paired-samples t-test demonstrated that only pairs of bottled water brands showed significant differences in their symbolic and practical images, whereas such findings were not detected for other product categories (e.g., retail stores, soft drinks, sporting goods). Thus, the present study confirmed bottled water brands as the target brands used for the main test. Furthermore, each brand of bottled water was compared for its symbolic and practical images, using a paired-samples t-test. The findings showed that Fiji Water was more likely to have symbolic images (M = 4.99) than practical images (M = 3.46) [t (32) = 3.22, p < .005]; Aquafina was less likely to have symbolic images (M = 3.03) than practical images (M = 4.09) [t (32) = -3.37, p < .005]; Dasani was less likely to have symbolic images (M = 3.33) than practical images (M = 4.48) [t (31) = -3.66, p < .005]; Ice Mountain was less likely to have symbolic images (M = 3.03) than practical images (M = 4.36) [t (32) = -4.25, p < .001]; there was no significant difference in symbolic (M = 4.07) and practical images (M = 3.42) for Evian [t (32) = -1.54, n.s.]. Since four of these brands (i.e., Fiji Water, Aquafina, Dasani, Ice Mountain) were found to have either symbolic or practical images, further analyses were carried out to compare pairs of these brands in terms of their images, using paired-samples t-tests. The findings demonstrated that Fiji Water (M = 4.99) was perceived as being more symbolic than Aquafina (M = 3.03) [t (32) = -5.78, p < .001], Ice Mountain (M = 3.03) [t (32) = -5.16, p < .001], and Dasani (M = 3.33) [t (32) = -4.46, p < .001]; Fiji Water (M = 3.46) was perceived as being less practical than Aquafina (M = 4.09) [t (32) = -2.59, p < .05], Ice Mountain (M = 4.36) [t (32) = - 66 3.07, p < .005], and Dasani (M = 4.48) [t (32) = -2.18, p < .05]. Therefore, Fiji Water seemed to be appropriate for a brand with a symbolic image, whereas Aquafina, Ice Mountain, or Dasani were reasonable choices for brands with practical images. Given the additional finding that Ice Mountain is a regional brand that is mainly distributed in the Midwest (e.g., IL, MI, IN) (see the official Web page for Ice Mountain for more), Ice Mountain was removed from the list. Furthermore, in order to rule out any possible impact of brand attitudes on the dependent variables (Sirgy et al. 1997), it is important to select two brands toward which consumers have a similar degree of attitude. In this pre-test, attitudes toward each brand were also asked with three items on a seven-point semantic differential scale (Nan and Heo 2008) anchored by ―Unfavorable (1) / Unfavorable (7),‖ ―Bad (1) / Good (7),‖ and ―Dislike (1) / Like (7).‖ The findings of a paired-samples t-test showed no significant mean difference in attitude toward brands between Fiji Water (M = 5.83) and Aquafina (M = 5.39) [t (31) = -1.75, n.s.], but a significant mean difference between Fiji Water and Dasani (M = 4.97) [t (32) = -3.10, p < .005]. Thus, Fiji Water was selected as a brand with symbolic images (hereinafter symbolic brand), while Aquafina was selected as a brand with practical images (hereinafter practical brand) for the main study. In addition, in this pre-test, two forms of questionnaires were provided, listing the brand names and the product names in a different order. This was designed to control any possible order effect on participants‘ perception of products, brands, and attitude toward brands. However, none of the order effects was significantly detected. Table 2 shows the descriptive statistics for all bottled water brands examined in the pretest, in terms of brand images and attitude. 67 Table 2 Descriptive Statistics of Images of and Attitude toward Bottled Water Brands in the Pre-Test Aquafina Brand Image Symbolic Practical Brand Attitude Dasani Evian Fiji Water Ice Mountain 3.03 (1.61) 4.09 (1.22) 5.39 (1.54) 3.33 (1.44) 4.48 (1.24) 4.97 (1.70) 4.07 (1.94) 3.42 (1.46) 5.20 (1.68) 4.99 (1.87) 3.46 (1.55) 5.83 (1.73) 3.03 (1.44) 4.36 (1.54) 4.76 (1.79) Note: Numeric values are mean scores which were measured on a seven-point scale; higher scores represent greater perceived symbolic and practical images of the brands, and more favorable attitude toward the brands; numeric values in parentheses are standard deviations; Ice Mountain was removed from the final list owing to its limited supply to specific regions; n = 33. Selection of Charitable Cause: Pre-Test Along with the selection of symbolic and practical brands, the selection of a charitable cause was made. To find a cause that would be featured on each of the selected brand pages in social media, the same pre-test was adopted. The participants in the same pre-test (n= 33) were asked to list the names or the descriptions of causes that they perceived as being congruent with each product shown on the pre-test. A list of causes suggested as being congruent with the brand/product in previous cause marketing literature (Gupta and Pirsch 2006; Zdravkovic 2010) was given to the participants. As a result, a charitable cause for recycling was most frequently mentioned as a cause type that was perceived to be congruent with bottled water brands: recycling accounted for 27% of all answers. Therefore, recycling was selected as the cause that would be featured on Fiji Water and Aquafina social media brand pages. A fictitious cause campaign, ―Go Recycling,‖ was used in the main study. As suggested by Nan and Heo (2007), 68 the hypothetical nature of the cause campaign could reduce variations in participants‘ previous experiences with the cause campaign. Manipulations of CSR Campaign Conditions: Main Study After selecting brands and a cause, the manipulations of CSR campaign conditions were followed. For the manipulations of CSR campaign conditions, three different types of written messages were created for the CRM campaign condition, the CS campaign condition, and the control condition. All of these messages were featured as the main content of the brand pages. Thus, reflecting the mixed combination of three CSR campaign conditions (i.e., CRM, CS, and control) and two types of brands (i.e., symbolic and practical), a total of six official brand pages on Facebook were created for use in the main study: (1) the symbolic brand page with the CRM campaign, (2) the symbolic brand page with the CS campaign, (3) the symbolic brand page in the control condition (i.e., no CSR campaigns), (4) the practical brand page with the CRM message, (5) the practical brand page with the CS message, and (6) the practical brand page in the control condition. These six official brand pages were made by modifying existing bottled water brand pages on Facebook. These six Facebook brand pages were created identically for all content (e.g., information about the brand, information about the cause, and messages soliciting consumer responses), except for the manipulations of CSR campaign conditions. For example, on all of the three Facebook pages for either Fiji Water or Aquafina, the Facebook logo, the brand logo, and the brand catchphrase were all placed at the top of the page. Next, the brief statement showing the reason why the brand supported the cause was presented on the following line across all pages: ―Fiji Water (or Aquafina) supports ‗Go Recycling,‘ a national charitable cause. We take responsibility for the bottle packaging that we place in the market. ‗Go Recycling‘ is committed to our recycling efforts to prevent global climate changes and save our 69 national resources.‖ Furthermore, a soliciting message on the bottom of these three Facebook pages was consistent for each brand: ―Click the ‗Like‘ button to join our Facebook page.‖ Finally, the picture of the brand was placed on the left across all six pages. The manipulations of two types of CSR campaigns (i.e., CRM and CS) were made by modifying the CRM campaign and the CS campaign on the existing Facebook brand pages. For example, for the CRM campaign, the following written statements were presented on the center of the band page: ―As every person clicks the ‗Like‘ button to join our Facebook page, we will donate $1 to ‗Go Recycling‘ up to $100,000 during the next two weeks. You Join  We donate $1. Once you click the ‗Like‘ button to join our Facebook page, the following message will be seen on your profile page: John Smith (i.e., your name) and Fiji Water (or Aquafina) donate $1 to ‗Go Recycling.‘‖ For the CS campaign, the following messages were shown on the center of the brand page: ―We will donate $100,000 during the next two weeks. We proudly sponsor ‗Go Recycling.‘ Once you click the ‗Like‘ button to join our Facebook page, the following message will be seen on your profile page: John Smith (i.e., your name) likes Fiji Water‘s (or Aquafina‘s) donation to ‗Go Recycling.‘‖ For both the CRM campaign and the CS campaign, the amount of money donated (i.e., $100,000) and the campaign period (i.e., two weeks) were manipulated based on the recent trends of CSR campaigns in social media to secure ecological validity (Brewer 2000) of the manipulation of the CSR campaigns under study. As another attempt to possess ecological validity of the manipulation of the control condition, the written statements for health information on drinking water were suggested as neutral messages (not related to brands and causes) on the brand page: ―From head to toe, our bodies rely on water for essential functions. Sip smarter by learning more about water. Water is our bodies‘ principal component, making up 60% of our body weight, 85% of our brain, and 70 10% of our teeth. Water flushes toxins out of vital organ, carries nutrients to our cells, and provides a moist environment for our cell tissues. Every day we lose water through our breath, perspiration, urine, and bowel movements. For our body to function, we must replenish its water supply. See our Facebook page to learn more about water‖ (see Appendix A and Appendix B). To ensure whether the message readability held constant across three messages for CSR campaign conditions, the readability check was carried out, as a similar approach of Kim and Paek (2009). The Flesch-Kincaid Grade Level Scores for both the CRM campaign and CS campaign were measured as 10, indicating that these two CSR campaign messages were understandable at a 10th-grade level. The score for the control group message was 7. Additionally, the Flesch Reading Ease Index Scores were 54% for the CRM campaign, 53% for the CS campaign, and 64% for the message in the control group. These findings indicated that all three messages for the CSR campaign conditions should be considered as having the standard reading level. Therefore, the findings concluded that the messages of the three CSR campaign conditions were readable for college students (i.e., the sample of the current study); and in particular, the equal levels of readability between CRM campaigns and CS campaigns were considered. Additionally, the number of words was also calculated to ensure that the message length was consistent across the three CSR campaign conditions: 128 words for the CRM message, 112 words for the CS message, and 117 words for the control condition. Thus, the message length seems to be consistent overall across the three conditions of CSR campaigns. Manipulations of e-Mail Invitations for Self-Friend Gender Composition To create same and different self-friend gender compositions, 12 invitation e-mails were manipulated, corresponding to the six brand pages (i.e., those with three conditions of CSR campaigns for each of the two brands) sent from male friends, and the other six sent from female 71 friends. These invitation e-mails to the brand pages featuring either CRM campaigns or CS campaigns specified the gender types of friends (i.e., male or female), the brand names (e.g., Fiji Water or Aquafina), brief information about the cause, and a note inviting participants to the Facebook brand pages (i.e., ―Your male (or female) casual friend sent an e-mail invitation to you! Hi, X (Fiji Water or Aquafina), a bottled water brand, supports „Go Recycling,‟ a national charitable cause to save our natural resources. Please join the official Facebook page of X to support the cause campaign of X. Your male (or female) casual friend. Next, click ‗Next Page‘ to visit the official Facebook page of X.‖). Furthermore, these invitation e-mails to the brand pages for the control condition included the gender types of friends (i.e., male or female), the brand name (e.g., Fiji Water or Aquafina), and a note inviting participants to Facebook brand pages; however, no information on cause was provided (i.e., ―Your male (or female) friend sent an email invitation to you! Hi, I like X (Fiji Water or Aquafina), a bottled water brand. Please join the official Facebook page of X. Your male (or female) friend. Next, click ‗Next Page‘ to visit the official Facebook page of X.‖). Later, to create the two types (i.e., same and different) of self-friend gender composition, the gender types of the e-mail senders (i.e., friend) were compared with those of the participants of the main study (i.e., self): if the participants had the same gender type as the e-mail senders, they were categorized into the group of same self-friend gender composition, whereas if the participants had a different gender type than the e-mail senders, they belonged to the group of different self-friend gender composition. The Appendices present the study stimuli used for the main study. Sampling and Final Sample 72 The current study recruited college students currently enrolled in various departments at MSU. College students seem to be appropriate for the sample of this study. It has been found that about 40% of Facebook users are college students (Quantcast 2010 ) and college students rank Facebook as the Web site they use most frequently (Tech Crunchies 2008). Moreover, a recent study (Bhagat, Loeb, and Rovner 2010) reported that Generation Y, born between 1981 and 1991, tended to actively use the Internet for cause-supporting behaviors such as forwarding messages about cause campaigns to friends or joining cause Web pages. Even though the current study recruited college students for the sample, it used a random sampling method as an attempt to represent the whole undergraduate population at MSU. More specifically, upon approval of the Institutional Review Boards (IRBs) within the university, a total of 5,000 undergraduate students were randomly selected by the Office of the Registrar at the university and received two times of recruitment e-mails, including a URL for the research Web site. Their e-mail addresses were purchased by the Office of the Registrar. In the recruitment e-mails, the students were asked to visit the research Web site by clicking the hyperlinked URL within two weeks upon receipt of the first e-mail. Finally, a total of 875 students participated in the study during a two-week period, indicating a 17.5% response rate. To increase student participation, lottery prizes were announced as a participation incentive in the recruitment e-mail. Out of those who completed the questionnaire, two students were randomly selected as winners of the Grand Prizes (i.e., $99.99 Amazon gift card for each), and 10 students as winners of the Second Prizes (i.e., $20.00 Amazon gift card for each). The prizes were given to all winners via e-mail one week after closing the research Web site. After removing 155 participants who answered incorrectly in the 73 screening questions to check participant awareness of manipulated contents, a total of 720 students remained as the final sample of the study. As seen in Table 3, the average age of participants in the final sample was 21 years old (SD = 2.58). Women (59.9%) outnumbered men. A majority of the students were Americans (95.4%). They varied in terms of studying areas and class standings. About half of them (46.74%) had used Facebook for about three to four years, and the other half (42.4%) had used it for more than five years. These characteristics of the final sample seemed to be consistent in general with those of the undergraduate population at MSU. Table 3 illustrates the demographic characteristics of both the final sample under study and the MSU undergraduate population. Furthermore, as noted previously, the present study employed using a 3 (CSR campaign conditions: CRM, CS, or control) x 2 (types of self-friend gender composition: same versus different) x 2 (types of brands: symbolic versus practical) between-subjects experimental design. Thus, each participant was randomly exposed to one of these 12 conditions using JavaScript. The following Table 4 shows the number of the final sample in each of the 12 experimental conditions. Even though the original sample sizes across the conditions were found to be almost equal, the final sample sizes — obtained after removing 155 participants who incorrectly answered at least one of the screening questions to check manipulated contents — did not seem to be equal in general across the conditions. Nevertheless, the ratio of the largest to the smallest group variance was less than 4:1 (i.e., 3.27:1 for the intention to join, and 2.69:1 for the intention to invite) across the 12 conditions, which indicated that this data did not seriously violate the homogeneity of variance (Moore 1995). 74 Table 3 Characteristics of Final Sample of the Main Study and MSU Undergraduate Population Percent or Mean MSU Undergraduate Population (n=33,767) Percent 59.9% 40.1% 21 (SD = 2.58) 52.8% 47.2% N/A 32.2% 31.1% 19.5% 17.2% 32.9% 25.5% 21.3% 20.4% 27.5% 23.7% 17.2% 16.7% 26.2% 21.6% 20.3% 17.3% 14.4% 14.6% 80.4% 9.4% 6.7% 3.5% 71.6% 5.7 % 7.2 % 2.8% Final Sample of the Main Study (n=720) Characteristics Gender Female Male Age Class Standing Senior Junior Sophomore Freshman Studying Area (Major) Natural Science (e.g., Biology, Physics) Business (e.g. Accounting, Management) Social Science (e.g., Communication, Education) Engineering (e.g., Civil Engineering, Chemical Engineering) Arts and Letters (e.g., Literature, Music) Race White/Caucasian Asian/Pacific Islander African American Hispanic Facebook Usage About 1–2 years About 3–4 years About 5–6 years About 7–8 years 10.8% 46.7% 38.7% 3.7% Note: Each percentage for the MSU undergraduate population is based on the enrollment data in Spring 2011, which is the same academic semester at which the present study recruited the sample; the studying area may not completely match the official college names at MSU. For example, ―Education‖ is a college unit at MSU but in this study it was categorized into Social Science in order to indicate Studying Areas parsimoniously; in the MSU undergraduate population, even though there was an additional 13.24% of the students categorized as ―others‖ (e.g., mixed race), it was not found in the final samples possibly because the present study asked the participants to select the type of race most representing themselves. 75 Table 4 Final Sample Sizes in Each Experimental Condition Same Self-Friend Gender Composition CSR Campaign Conditions CRM CS Control Symbolic Brand 61 47 74 Practical Brand 76 57 49 Different Self-Friend Gender Composition Symbolic Brand 53 45 65 Practical Brand 76 46 70 Note: n = 720; these 720 students in the final sample remained after removing 155 participants, out of a total of 875, who answered incorrectly on the screening questions regarding the manipulations. Procedure All procedures were done online, in particular by sending an e-mail that contained the hyperlinked URL for the research Web site. Upon clicking the URL, participants were asked to read the consent form on the first page. Participants were then asked to indicate their gender type on the next page. Considering an event in which female participants would outnumber male participants, the stratified sampling method by gender types was applied in order to obtain the consistent ratio of self-friend gender composition (i.e., same or different) between men and women. Once the participants clicked their gender type, the JavaScript of the Web site randomly assigned each participant, under each stratified gender group, to one of the 12 manipulated conditions: (1) a symbolic brand page with a CRM campaign, invited by a different-gender friend, (2) a symbolic brand page with a CRM campaign, invited by a same-gender friend, (3) a symbolic brand page with a CS campaign, invited by a different-gender friend, (4) a symbolic brand page with a CS campaign, invited by a same-gender friend, (5) a symbolic brand page for the control condition, invited by a different-gender friend, (6) a symbolic brand page for the 76 control condition, invited by a same-gender friend, (7) a practical brand page with a CRM campaign, invited by a different-gender friend, (8) a practical brand page with a CRM campaign, invited by a same-gender friend, (9) a practical brand page with a CS campaign, invited by a different-gender friend, (10) a practical brand page with a CS campaign, invited by a samegender friend, (11) a practical brand page for the control condition, invited by a different-gender friend, (12) a practical brand page for the control condition, invited by a same-gender friend. Then, participants were asked to answer the pre-questionnaire items including three covariates (i.e., personal involvement in issues, brand preference, and frequency of using the ―Like‖ button on Facebook). Next, the participants were asked to imagine a hypothetical situation, in which they received an e-mail from one of their male or female friends inviting them to either the Fiji Water or Aquafina brand page on Facebook. The participants were also randomly assigned to one of the six invitation e-mails from their male casual friends or one of another six invitation e-mails from their female casual friends, using the JavaScript. Additionally, to control for any influence of self-friend relationship on the dependent variables, each participant was randomly asked to imagine the e-mail sender as his/her casual friend (i.e., male casual friend or female casual friend). To make the casual friend figures salient, each participant was also asked to write down the name of his/her male or female casual friend, prior to getting the invitation e-mail. The description of casual friends, adopted from Holiday and Kerns (1999), was provided to the participants to increase their understanding of a casual friendship: ―Male (or female) casual friends are male (or female) people with whom you enjoy spending time but do not know as well as and so not feel as close to as your ‗close personal friends.‘ You occasionally communicate with the male (or female) casual friends on Facebook and via e-mail. Your friendship seems to be based upon the circumstances in which you find 77 yourself (e. g., you work together, are in a class together). You plan on remaining friends as long as circumstances permit. Your friendship would dissolve if your circumstances were to change.‖ After being asked to read the assigned invitation e-mail carefully, the participants were directed to the corresponding Facebook brand page and asked to review the page thoroughly. Next, the participants were asked to complete the questionnaire, consisting of a list of questions about their outcome expectancy to look favorable, their intention to join the brand page, and their intention to invite their friends to the brand page. Appendix B presents the study stimuli and the questionnaires used for the main study. Measures As mentioned earlier, the present study has one independent variable (i.e., CSR campaign condition), two moderating variables (i.e., type of self-friend gender composition, and type of brand), one mediating variable (i.e., outcome expectancy), and two dependent variables (i.e., the intention to join brand pages, and the intention to invite friends to brand pages). The independent and moderating variables were manipulated, whereas the mediating variables and dependent variables were measured. Three covariates (i.e., personal involvement in issues, brand preference, and frequency of using the ―Like‖ button on Facebook) were also measured. Intention to Join To measure the intention to join a Facebook brand page, participants were asked to indicate the likelihood that they would click the ―Like‖ button to join the brand page, using four items adopted by Bearden, Lichtenstein, and Teel (1984) on a seven-point semantic differential scale: ―How likely are you to click the ‗Like‘ button to join the Fiji Water (or Aquafina) Facebook page?‖: ―Unlikely (1) / Likely (7),‖ ―Improbable (1) / Probable (7),‖ ―Impossible (1) / Possible (7),‖ and ―Uncertain (1) /Certain (7).‖ Cronbach‘s alpha reliability indicated reasonable 78 internal consistency of these items (α = .90). All four items were thus averaged to construct the variable of the intention to join. Intention to Invite The current study measured the intention to invite friends to a Facebook brand page, by asking participants to indicate the likelihood that they would invite people to the brand page. Four items, adopted by Bearden, Lichtenstein, and Teel (1984) were asked on a seven-point semantic differential scale: ―How likely are you to invite other people to the Fiji Water (or Aquafina) Facebook page?‖: ―Unlikely (1) / Likely (7),‖ ―Improbable (1) / Probable (7),‖ ―Impossible (1) / Possible (7),‖ and ―Uncertain (1) /Certain (7).‖ Cronbach‘s alpha coefficient showed this measure was reliable (α = .81). These items were averaged for the index of the intention to invite. Outcome Expectancy The outcome expectancy to look favorable to their friends by joining a Facebook brand page or inviting their friends to a Facebook brand page was assessed on a seven-point Likert scale, anchoring from 1 being ―Strongly disagree‖ to 7 being ―Strongly agree‖ with four items adopted from Arbuthnot (1977) and Sweeney and Soutar (2001): ―What do you think about your joining the Fiji Water (or Aquafina) Facebook page or inviting your friends to this page? ‗Joining the brand page or inviting your friends to the brand page will help me make … to my Facebook friends‖‘—―a good impression‖; ―an altruistic impression‖; ―a socially responsible impression‖; ―an environmentally knowledgeable impression.‖ The measure was acceptable (α = .81). The items were aggregated and averaged to create the index of outcome expectancy. Personal Involvement 79 Given the previous findings that individuals‘ involvement in social issues is likely to influence their decision making in cause marketing contexts (Robin, Reidenbach, and Forrest 1996), personal involvement in cause-related issues was tested as a covariate in the model. Thus, participants were asked to indicate their personal involvement in the issue of recycling, with five items modified from past studies (Mittal 1995; Zaichkowsky 1994) on a seven-point semantic differential scale: ―To me, the issue of recycling is…‖—―Unimportant (1) / Important (7),‖ ―Of no concern (1) / Of concern to me (7),‖ ―Means nothing to me (1) / Means a lot to me (7),‖ ―Does not matter to me (1) / Does matter to me (7),‖ and ―Insignificant (1) / Significant (7).‖ The Cronbach‘s alpha coefficient was acceptable (α = .95). Thus, the five items were averaged to construct personal involvement in issues. Brand Preference Oftentimes, brand-related features are factors predicting consumer responses to causerelated marketing (Cornwell and Coote 2005; Lafferty and Goldsmith 2005). Thus, the current study added brand preference as a possible covariate, which may influence consumer intentions to join and invite on the brand pages in social media. Brand preference was suggested as an umbrella concept influencing consumers‘ major responses to the brands, such as consumer attitude, perception, and behavior intention regarding brands (Sirgy et al. 1997). In the current study, brand preference was measured using the four-item, seven-point Likert scale (1 = ―Strongly disagree‖ and 7 = ―Strongly agree‖) (Sirgy et al 1997): ―I like Fiji Water (or Aquafina) better than other bottled water brands,‖ ―I would use Fiji Water (or Aquafina) more than I would use other bottled water brands,‖ ―Fiji Water (or Aquafina) is my preferred brand over the other bottled brands,‖ and ―I would be inclined to put Fiji Water (or Aquafina) over other bottled 80 water brands. The Cronbach‘s alpha coefficient was acceptable (α = .96). Four items were thus averaged for subsequent analyses. Frequency of Using the “Like” Button The frequency of using the ―Like‖ button on Facebook to join the brand page was added in this study as another covariate. One single item on a seven-point scale was adopted: ―How often do you click the ―Like‖ button to join a Facebook page in general?‖ with 1 being ―Never‖ and 7 being ―Quite Often.‖ Descriptive statistics for the variables and scales under study are shown in Table 5 (see Appendix B for the complete questionnaires). Table 5 Descriptive Statistics for Variables Variable (The Number of Items) Intention to Join (4) Intention to Invite (4) Outcome Expectancy (4) Personal Involvement (5) Brand Preference (4) Frequency of Using the ―Like‖ Button (1) Cronbach Alpha Coefficient .90 .81 .81 .95 .96 Not Available Mean SD n 3.84 2.47 3.44 5.44 2.63 3.02 720 720 720 720 720 720 1.75 1.36 1.61 1.22 1.65 2.05 Note: All variables are measured on a seven-point scale; higher mean scores represent greater intention and greater perceptions. Analytic Strategy The main study was carried out based on different analyses for each of the different tests: the assumption checks, the awareness checks, the manipulation checks, and the hypotheses testing. 81 Assumption Checks Even though impression management literature conceptually suggests that direct tactics of impression management have more specific, concrete, and particularistic representations than indirect tactics (Carter and Sanna 2008; Pratkanis 2007), there have been no empirical findings to investigate such differences between direct and indirect tactics. Thus, the present study checked the assumption that joining and inviting activities on the brand pages featuring a CRM campaign (versus a CS campaign) would be seen by others in a social media setting, as making more (versus less) direct contributions to the cause. First Assumption Check. Two questions were asked on the questionnaire of the main study. The first question was to examine whether the participants considered their joining and inviting activities on these brand pages as a way to manage their impression in social media: ―Do you think your joining and inviting on the Fiji Water (or Aquafina) Facebook page would help you make a good impression on your Facebook friends? (select one) – Yes; No.‖ Each answer was coded with 1 = ―Yes,‖ and 2 = ―No‖ selections. The second question was provided only for those who answered that they considered their joining and inviting activities as a way of impression management (i.e., those who selected ―Yes‖). It was asked whether they believed that they would be seen as direct or indict contributors to the cause by their friends in social media: ―If you answered ‗Yes‘ in the previous question, please indicate why you think so by clicking the number that represents your opinion most — ‗Because my ___ support for the brand cause campaigns is indirect (1) / direct (7), passive (1) / active (7), secondhand (1) / firsthand (7).‖ Three items, which were asked on a seven-point semantic differential scale, were found to be one single item with a total variance of 76.1% explained, and the Cronbach‘s alpha coefficient was 82 reliable (α = .84). Thus, these three items seemed to be appropriate to represent the index of direct-indirect contribution to the cause. The independent samples t-test was conducted to compare the mean difference in participants‘ belief about direct-indirect contributions to the causes through their participation in CSR campaigns (i.e., the second question). The Q-Q plot supported the assumption of normality. The Levine‘s Test for Equality of Variance also showed that two types of CSR campaigns had equal variances on participants‘ belief about direct-indirect contributions to the causes (p = .84, n.s.), which led to the conclusion that the assumption of homogeneity of variance was also supported. Thus, the independent samples t-test seemed to be appropriate to test the second question. However, regarding the first question, the present study raised an issue pertaining to the wording of the question. The findings of the first question showed that 69.4% of the participants (i.e., 53.0% in the CRM campaign condition, 66.8% in the CS campaign condition, and 83.3% in the control condition) selected ―No,‖ which might result in inaccurate outcomes owing to the considerable loss of data. More importantly, it was noted that when they were asked to describe their own reason for why they selected ―No,‖ the participants expressed negative feedback about their impression management on the brand pages, particularly indicating that they were not interested in impression management or they did not manage their impression to their friends. Such unexpected results from the first question on the assumption check test, led to further speculation about any possibility that the participants might be biased to answer in a socially desirable way. Literature suggested that researchers could frequently face difficulty in measuring accurate opinions from survey respondents, when the participants were biased by their perception of what was a socially desirable answer and thus when they tended to indicate their 83 answers based on such a perception (Schoderbek and Deshpande 1996). Particularly, when the participants were asked to indicate their ethical behaviors, such a tendency to express their opinion in a socially desirable way could be related to the impression management tendency (i.e., purposeful self-presentation) (Randall and Fernandes 1991). Moreover, such patterns were strengthened when anonymity was not completely provided for the participants in the study setting (Randall and Fernandes 1991). In the present study, the participants were asked to indicate their own impression management in a cause-supportive context (which could be a somewhat ethical issue), and were further asked to provide personal information such as full name and e-mail address, to enter themselves into the lottery pool. Accordingly, the second assumption check was strongly recommended as a follow-up study, which consisted of a newly developed questionnaire. Second Assumption Check. The second assumption check was carried out after the main study, by recruiting 20 undergraduate students conveniently in the main library at MSU. Two forms of questionnaires were developed, including the screenshots of both the CSR campaign stimuli for each brand (i.e., Fiji Water‘s Facebook page with both the CRM campaign and the CS campaign, and Aquafina‘s Facebook page with both the CRM campaign and the CS campaign). Each participant was randomly exposed to either of the two types of CSR campaigns by Fiji Water or those by Aquafina. On the first page they were asked to examine the stimuli carefully; on the next page, they were asked to indicate their opinion about how people in general would evaluate the effectiveness of the two types of CSR campaigns on the joining and inviting activities on the brand pages in Facebook: ―Out of these two examples, which seems to be more effective in encouraging Facebook users in general to join this brand page on Facebook and invite their 84 friends to this brand page on Facebook? (select one): Example A (CRM campaign), Example B (CS campaign).‖ The answer was coded with 1 = ―CRM campaign‖ and 2 = ―CS campaign.‖ Furthermore, they were asked to write down their reasons for why either type of campaign would be more effective than the other among people, on open-ended questions. Then, each answer was coded by a researcher. Here, to rule out any potential that participants would answer in a socially desirable manner, participants were asked to indicate their belief of general people‘s joining and inviting behaviors for the purpose of impression management, rather than directly indicate their own joining and inviting activities. Additionally, no personal information was asked, in order to increase the anonymity of the participants. Awareness Checks The awareness checks were conducted to determine whether or not the participants recognized the main contents of two manipulations used for the current experimental study: three messages of CSR campaign conditions (i.e., CRM, CS, or control), and two gender types of friends who invited the participants to brand pages (i.e., male or female). Messages of CSR Campaign Conditions. One screening question was given at the end of the questionnaire, asking each participant to identify the message s/he reviewed on the brand page, out of three sentences adopted from the messages of CSR campaign conditions: ―In the previous section, you were invited to Fiji Water‘s (or Aquafina‘s) Facebook page, featuring one of the following messages. Please select the one that best describes the message (select one)‖: ―(1) ‗As every person clicks the ―Like‖ button to join our Facebook page, we will donate $1 to ―Go Recycling‖ up to $100,000 during the next two weeks,‘ (2) ‗We will donate $100,000 during the 85 next two weeks,‘ (3) ‗Water is our bodies‘ principal chemical component, making up 60% of our body weight, 86% of our brain, and 10% of our teeth,‘ (4) ‗Don‘t know/don‘t remember.‘‖ Gender Types of Friends. The participants were given another screening question, asking them to select the gender types of the friends who invited them to the brand pages: ―At the beginning stage of this study, you were asked to imagine getting an e-mail from one of the following persons inviting you to the Facebook page for a bottled water brand (select one): (1) Male casual friend, (2) Female casual friend, (3) Don‘t know/don‘t remember.‖ In combination, those who did not provide the correct answers across two screening questions were removed from the final sample, because they should be considered as those who did not understand the key content of the manipulations. Table 6 demonstrates the summary of awareness checks. Manipulation Checks The present study checked two manipulations used in the experiment: brand types and cause type. Brand Types. As discussed earlier, this study used two types of bottled water brands: Fiji Water for a symbolic brand, and Aquafina for a practical brand. To ensure the efficacy of the symbolic and practical images of these two selected brands, two items for symbolic brand image and one item for practical brand image (Bhat and Reddy 2008) were measured: ―People drink Fiji Water (or Aquafina) as a way to express their personality,‖ and ―Drinking Fiji Water (or Aquafina) says something about the kind of person you are‖ [for symbolic images of brands]; ―Drinking Fiji Water (or Aquafina) is practical‖ [for practical images of brands]. The first two items were found to be one single construct as the index of symbolic brand images (r = .72, p < .001 and 86% of a total variance explained). 86 As indicated in Table 6, independent samples t-tests, using the SPSS 19 program, were adopted to compare the mean scores of perceived symbolic and practical images for two brands. The independent samples t-test seemed appropriate to test the mean difference between two types of brands in terms of perceived symbolic and practical images. The assumption of normality was confirmed by the Q-Q plot. Further, the assumption of homogeneity of variance was supported by the results of Levine‘s Test for Equality of Variance, showing that two types of brands had equal variances on perceived symbolic image (p = .14, n.s.) and perceived practical image (p = .20, n.s.). Cause Type. The present study expected that the fictitious charitable cause campaign ―Go Recycling‖ was congruent with the selected two brands, Fiji Water and Aquafina. To check additionally whether this cause campaign was appropriately selected for the brands under study, perceived congruence between Fiji Water (or Aquafina) and the cause campaign ―Go Recycling‖ was measured at the end of the questionnaire. Four items adopted from Kamins and Gupta (1994) and Till and Busler (2000) were asked of participants in the CRM campaign condition and the CS campaign condition, on a seven-point semantic differential scale: ―Not compatible (1) / Compatible (7),‖ ―Bad fit (1) / Good fit (7),‖ ―Irrelevant (1) / Relevant (7),‖ and ―Bad match (1) / Good match (7).‖ The EFA result indicated these four items as one factor with 86% of a total variance explained, and the Cronbach‘s alpha score for these items was highly acceptable (α = .94). These items were thus an index of perceived brand-cause congruence. As shown in Table 6, an independent samples t-test, using the SPSS 19 program, was conducted to compare the mean scores of perceived brand-cause congruence. The use of independent samples t-test was confirmed by supporting the assumption of normality (e.g., the Q-Q plot) and the assumption of homogeneity of variance (e.g., the Levine‘s Test for Equality of 87 Variance showed that two types of brands had equal variances on perceived brand-cause congruence (p = .84, n.s.)). Hypotheses Testing As indicated in Table 6, to test the hypotheses, the current study adopted three kinds of analyses: the Analysis of Covariance (ANCOVA) to test the proposed main effect of CSR campaign conditions, the moderating roles of types of self-gender composition, and the moderating roles of types of brands; the unconditional indirect effect (Preacher and Hayes 2008) for the mediation analysis; and the conditional indirect effect (Preacher, Rucker and Hayes 2007) for the moderated mediation analysis. First, to test the hypotheses for the main effect of CSR campaign conditions (H1a-b and H2) and the interaction effect of CSR campaign conditions, and either the type of self-friend gender composition (H4a-c) or the type of brands (RQ1a-c) on the dependent variables, the ANCOVAs were conducted using the SPSS 19 program. The multivariate analysis of covariance (MANCOVA) seemed to be the most appropriate statistical method to test these hypotheses, dealing with the situation where there were multiple dependent variables (i.e., intentions to join, intentions to invite, outcome expectancy) and three factors (i.e., CSR campaign conditions, types of self-friend gender composition, types of brands), as well as supporting use of continuous control variables as covariates (i.e., personal involvement, brand preference, frequency of using the ―Like‖ button) (Nelder and Wedderburn 1972; Gill 2001). However, it was found in the present study that the covariates were significantly different among the dependent variables (Box M = 101.215, p < .001), which resulted in a failure to meet assumptions necessary for the performance of a MANCOVA (Hair et al. 1998). 88 Thus, two sets of ANCOVAs for each of the dependent variables seemed to be appropriate as the alternative analyses to test these hypotheses. Regarding the ANCOVA assumption that each group of the factors has equal variance, the present study found that this assumption was violated [F (11,707) = 6.29, p < .001 in Levene‘s Test for the intention to join; F (11,707) = 5.50, p < .001 in Levene‘s Test for the intention to invite]. However, literature suggests that ANCOVA is robust for small and even moderate departures from homogeneity of variance (Box 1954). Moore (1995) further suggested that if the ratio of largest to smallest group variances was 4:1 or less, the failure to meet this assumption was not fatal to ANCOVA. In the present study, the ratio of largest to smallest group variances was less than 3.27:1 (i.e., 1.48 ~ 3.98 for the range of variance) for the intention to join, and 2.69:1 (i.e., .85 ~ 2.78 for the range of variance) for the intention to invite. Therefore, it was found that the violation of this assumption was not critical in applying ANCOVA for the hypotheses for the proposed main and interactions effects. Additionally, three covariates were found to be significant for the dependent variables on the ANCOVA models. More specifically, three different results of ANCOVAs showed that personal involvement in issues had no significant influences on the outcome expectancy [F (1, 704) = .03, n.s.], whereas it had significantly positive influences on the intention to join [F (1, 704) = 9.82, p < .01] and the intention to invite [F (1, 714) = 4.20, p < .05]; brand preference had significantly positive influences on the outcome expectancy [F (1, 704) = 78.29, p <. 001], the intention to join [F (1, 704) = 42.49, p < .001], and the intention to invite [F (1, 714) = 36.08, p < .001]; the frequency of clicking the ―Like‖ button had significantly positive influences on the outcome expectancy [F (1, 704) = 27. 60, p <. 001], the intention to join [F (1, 704) = 78.48, p < .001], and the intention to invite [F (1, 714) = 15.20, p < .001]. Taken together, ANCOVA, 89 adding these three covariates into the model as continuous control variables, was confirmed to be appropriate to test the proposed hypotheses. Second, to examine the mediating role of outcome expectancy in the relationship between the CSR campaign conditions on the dependent variables (i.e., mediation models) (H3a-b), the present study adopted the mediation analysis, using the SPSS Macro (i.e., the unconditional indirect effect), which was recommended by Preacher and Hayes (2008). Recent literature (McKinnon, Krull, and Lockwood 2000; Preacher and Hayes 2008; Zhao, Lynch, and Chen 2010) emphasized the superiority of the mediation analysis by Preacher and Hayes (2008), in that it allowed one, or more than one, mediator and adjusted all paths for the potential influence of covariates, even those not proposed in the model (Preacher and Hayes 2008; Zhao et al., 2010). Thus, as suggested by Preacher and Hayes (2008), the mediating role of outcome expectancy was examined after controlling the covariates. Third, to examine the moderated mediating role of outcome expectancy on the interaction effects of CSR campaign conditions and either types of self-friend gender composition or types of brands on the dependent variables (i.e., moderated mediation models) (H5a-b and RQ2a-b), the moderated mediation analysis, using the SPSS Macro (i.e., the conditional indirect effect) which was suggested by Preacher et al. (2007), was applied. This relatively new analysis was superior in estimating indirect effects including the Sobel test and percentile-based, biasedcorrected, and accelerated bootstrap confidence intervals for the moderated mediation effect (Preacher et al. 2007). In addition, for both the mediation analysis and the moderated mediation analysis, two of three manipulated conditions of CSR campaigns were coded to examine the different effect between two CSR campaign conditions. That is, the CRM campaign was coded as ―1‖ and the 90 control condition was coded as ―0‖ for the first comparison of two CSR campaign conditions (hereinafter CRM over Control); the CS campaign was coded as ―1‖ and the control condition was coded as ―0‖ for the second comparison (hereinafter CS over Control); finally, the CRM campaign was coded as ―1‖ and the CS campaign was coded as ―0‖ for the third comparison (hereinafter CRM over CS). Accordingly, three sets of mediation analyses and another three sets of moderated mediation analyses could be conducted, as suggested by Preacher and Hayes (2008) and Preacher et al. (2007) respectively. Table 6 summarizes the analytic strategies adopted for different tests in the main study. Table 6 Summary of Analytic Strategy Testing Purpose Analyses Assumption Checks To check the assumption whether people could be seen as either direct or indirect contributors to the cause through their participation in CSR campaigns in social media. Awareness Checks To check whether participants recognize correctly the types of CSR campaign messages 91  An open-ended question - “Why do you think one of the brand pages seems to be more effective in encouraging Facebook users in general to join this brand page on Facebook and invite their friends to this brand page on Facebook?”  Each answer was coded.  A screening question - “Please select the one that best describes the message [you reviewed] (select one)”: “(1) „As every person clicks the “Like” button to join our Facebook page, we will donate $1 to “Go Recycling” up to $100,000 during the next two weeks,‟ (2) „We will donate $100,000 during the next two weeks,‟ (3) „Water is our bodies‟ principal chemical component, making up 60% of our body weight, 86% of our brand, and 10% of our teeth,‟ Table 6 (cont’d)  To check whether participants correctly recognized the gender types of friends who sent them invitation e-mails   Manipulation Checks Hypotheses Testing To ensure whether the manipulation of brand types (i.e., symbolic and practical) was successful. To ensure whether the selected cause was congruent with the selected brands. To test the main effects of CSR campaign conditions (i.e., CRM, CS, control) on the intention to join (H1a), the intention to invite (H1b), and the outcome expectancy (H2). To test the interaction effects of CSR campaign conditions (i.e., CRM, CS, control) and the types of self-friend gender composition (i.e., same and different) on the outcome expectancy (H4a), the intention to join (H4b), and the intention to invite (H4c). 92  (4) „Don‟t know/don‟t remember.‟ Those who answered incorrectly were removed from the final sample. A screening question - At the beginning stage of this study, you were asked to imagine getting an e-mail from one of the following persons inviting you to the Facebook page for a bottled water brand (select one): (1) Male casual friend, (2) Female casual friend, (3) Don‟t know/don‟t remember.” Those who answered incorrectly were removed from the final sample. Independent samples t-tests.  Independent samples t-test.  ANCOVA  Covariates: personal involvement in the cause, brand preference, frequency of using the ―Like‖ button.  ANCOVA  Covariates: personal involvement in the cause, brand preference, frequency of using the ―Like‖ button. Table 6 (cont’d) To test the interaction effects of CSR campaign conditions (i.e., CRM, CS, control) and the types of brands (i.e., symbolic and practical) on the outcome expectancy (RQ1a), the intention to join (RQ1b), and the intention to invite (RQ1c). To test the mediating roles of outcome expectancy in the relationship between the CSR campaign conditions (i.e., CRM, CS, control) for either the intention to join (H3a) or the intention to invite (H3b). To test the mediating role of outcome expectancy in the interaction effects of CSR campaign condition (i.e., CRM, CS, control) and the types of self-friend gender composition (i.e., same and different) on the intention to join (H5a) and the intention to invite (H5b); the interaction effects of CSR campaign conditions (i.e., CRM, CS, control) and the types of brands (i.e., symbolic and practical) on the intention to join (RQ2a) and the intention to invite (RQ2b). 93  ANCOVA  Covariates: personal involvement in the cause, brand preference, frequency of using the ―Like‖ button.  The mediation analysis, by Preacher and Hayes (2008) (i.e., unconditional indirect effect)  Covariates: personal involvement in the cause, brand preference, frequency of using the ―Like‖ button.  The moderated mediation analysis by Preacher et al. (2007) (i.e., conditional indirect effect)  Covariates: personal involvement in the cause, brand preference, frequency of using the ―Like‖ button. CHAPTER VI RESULTS Assumption Checks As mentioned in analysis strategies, the independent samples t-test was applied for the second question in the first assumption check, to compare the mean difference in direct-indirect contributions to the cause through joining and inviting activities in the CRM campaign condition and in the CS campaign condition. The findings showed no statistically significant difference in consumer responses on these brand pages [t (188) = .35, n.s.], even though joining and inviting activities in the CRM campaign (M = 4.36, SD = 1.30) were perceived as making a more direct contribution to the cause than those in the CS campaign (M = 4.29, SD = 1.25). As explained previously, several issues were raised regarding the null effects at the first assumption check; therefore, the second assumption check was followed. The findings showed that out of these 20 students who participated in the second assumption check, 18 participants (90%) answered that the brand page featuring the CRM campaign would be more effective than the page with the CS campaign, with regard to joining and inviting activities on the brand pages among general people. Further, after coding the reasons why they considered the brand page featuring the CRM campaign (versus the one featuring the CS campaign) as being more effective in enhancing Facebook users‘ joining and inviting behaviors, the present study found two major reasons with the highest frequencies: ―people can donate in their own names‖ (72%) and ―people can make their help visible‖ (11%). On the other hand, the two students who selected the brand page featuring the CS campaign as being more effective in enhancing people‘s joining and inviting activities, explained their reasons: ―people can get a more altruistic feeling,‖ and ―a fixed amount of money may be donated,‖ respectively. 94 All of these findings, particularly the results of the second assumption check, seem likely to support the basic assumption of the present study that consumers would be seen as those who contribute to the cause more (versus less) directly, when they join the brand pages and invite their friends to the brand pages that feature CRM campaigns (versus CS campaigns). Awareness Checks Messages of CSR Campaign Conditions and Gender Types of Friends The results of two screening questions showed that out of a total of 875 participants, 155 (17.71%) answered incorrectly or selected ―don‘t know/do not remember,‖ in at least one of these two manipulations for CSR campaign conditions and gender types of friends who sent the invitation e-mails. These 155 participants were excluded from the final sample, to secure accurate results from hypothesis testing. Thus, a total of 720 participants remained in the final sample (see Sampling and Table 4 for the final sample characteristics). Manipulation Checks Brand Types The findings of independent samples t-tests showed a significant mean difference between Fiji Water and Aquafina in terms of their symbolic (MFiji Water = 3.50, MAquafina = 2.61) [t (642) = 7.71, p < .001] and practical images (MFiji Water = 2.71, MAquafina = 3.36) [t (641) = - 4.91, p < .001]. Thus, the manipulation of brand types was successful. Cause Type The findings of independent samples t-tests showed no significant mean difference between Fiji Water and Aquafina in terms of perceived brand-cause congruence (MFiji Water = 4.76, MAquafina = 4.84) [t (460) = -.59, n.s.]. Thus, these two brands were considered equally 95 congruent in general with the charitable cause of ―Go Recycling.‖ The selection of cause type was therefore successful. Hypotheses Testing Main Effects and Interaction Effects The present study posited the main effect of CSR campaign conditions (H1a-b, H2) and the interaction effect between the CSR campaign conditions and types of self-friend gender composition (H4a-c); and the interaction effect between the CSR campaign conditions and types of brands (RQ1a-c) on the intention to join, the intention to invite, and the outcome expectancy. As noted in Analytic Strategy, ANCOVAs were conducted to test these main and interaction effects. Main Effects of CSR Campaign Conditions on Intention to Join and Intention to Invite. H1 predicted significant main effects of CSR campaign conditions on consumer responses to the brand pages. Specifically, H1 posited that consumers in the CRM campaign condition would have the greatest intention to (a) join the brand page and (b) invite their friends to the brand page, followed by those in the CS campaign condition, and those in the control condition (i.e., pages without either a CRM campaign or a CS campaign). As seen in Table 7, the findings of the first ANCOVA model showed that the main effect of the CSR campaign conditions on the intention to join the brand page was significant [F (2, 704) = 73.75, p < .001]. The Bonferroni post-hoc tests revealed a significant mean difference between the CRM campaign condition (M = 4.78) and either the control condition (M = 2.97) (p < .001) or the CS campaign condition (M =3.63) (p < .001), and between the CS campaign condition (M =3.63) and the control group (M =2.97) (p < .001), in terms of the intention to join the brand page. Such findings indicated that the brand page with the CRM campaign resulted in 96 the greatest intention of consumers to join the brand page, followed by the brand page with the CS campaign, and the brand page for the control condition. Thus, H1a was supported. Regarding H1b, another ANCOVA model showed that the main effect of CSR campaign conditions on the intention to invite friends to the brand page was also significant [F (2, 704) = 21.99, p < .001]. Post-hoc analyses using Bonferroni adjustments revealed a significant mean difference in the intention to invite between the CRM campaign condition (M = 2.94), and either the control condition (M = 2.09) (p < .01), or the CS campaign condition (M = 2.33) (p < .001); however, there was no significant mean difference in the intention to invite between the CS campaign condition (M = 2.33), and the control condition (M = 2.09) (n.s). Such findings indicated that the brand page with the CRM campaign generated a significantly greater intention to invite, than the brand page with the CS campaign and the brand page for the control condition, whereas there was no statistically significant difference in the intention to invite between the CS campaign and the control condition. Therefore, H1b was partially supported. Main Effects of CSR Campaign Conditions on Outcome Expectancy. H2 predicted a significant main effect of CSR campaign conditions on the outcome expectancy to look favorable to friends by engaging in joining and inviting activities on the brand page. Particularly, it was hypothesized that consumers in the CRM campaign condition would have the greatest outcome expectancy, followed by those in the CS campaign condition, and those in the control condition. The findings of another ANCOVA test confirmed the significant main effect of CSR campaign conditions on the outcome expectancy [F (2, 704) = 61.19, p < .001]. Post-hoc analyses using Bonferroni adjustments showed a significant mean difference in terms of participants‘ outcome expectancy between the CRM campaign (M = 4.12) and either the control 97 condition (M = 2.59) (p < .001) or the CS campaign (M = 3.64) (p < .001), and between the CS campaign (M = 3.64) and the control condition (M = 2.59) (p < .001). Thus, as expected in H2, the brand page with the CRM campaign had the greatest outcome expectancy, followed by the brand page with the CS campaign, and the brand page for the control condition. Thus, H2 was supported. Table 7 shows a summary of means and standard errors for the main effects of CSR campaign conditions on the dependent variables. Table 7 Means and Standard Errors for Main Effects of CSR Campaign Conditions on Dependent Variables CSR Campaign Conditions CRM CS Control Mean (SE) Mean (SE) Hypothesis Dependent Variable Mean (SE) H1a (Supported) Intention to Join 4.78 (.10) H1b (Partially supported) Intention to Invite 2.94 (.09) H2 (Supported) Outcome Expectancy 4.12 (.10) a a a b 3.63 (.11) b 2.33 (.10) b 3.64 (.11) c 2.97 (.10) b 2.09 (.09) c 2.59 (.10) Note: Since covariates are included in the model, adjusted marginal means and standard errors are presented; means with different superscripts are significantly different from each other at the p < .05 level; each rating scale of the dependent variables is scored from 1 to 7; higher mean scores represent greater intentions and greater outcome expectancy. Interaction Effects of CSR Campaign Conditions and Types of Self-Friend Gender Composition H4 predicted that the types of self-friend gender composition would moderate the proposed main effects of CSR campaign conditions on the dependent variables. More 98 specifically, it was hypothesized that the effect of CSR campaign conditions on (a) the outcome expectancy, (b) the intention to join, and (c) the intention to invite, would be more pronounced, when the self-friend composition was different (versus same). According to results of ANCOVAs, the study did not find any significant interaction effect between the CSR campaign conditions and the types of self-friend gender composition on the outcome expectancy [F (2, 704) = .01, n.s.] (H4a), the intention to join [F (2, 704) = 1.25, n.s.] (H4b), and the intention to invite [F (2, 704) = .03, n.s.] (H4c). Thus, H4a, H4b, and H4c were not supported. Table 8 provides a summary of means and standard errors for the interaction effects of the CSR campaign conditions and the types of self-friend gender composition on the dependent variables. Table 8 Means and Standard Errors for Interaction Effects of CSR Campaign Conditions and Types of Self-Friend Gender Composition on Dependent Variables Same Gender Composition Different Gender Composition CRM CS Control CRM CS Control Hypothesis Dependent Variable Mean (SE) Mean (SE) Mean (SE) Mean (SE) Mean (SE) Mean (SE) H4a (Not supported) Outcome Expectancy 4.09 (.15) 3.64 (.16) 2.63 (.13) 4.15 (.12) 3.64 (.14) 2.52 (.14) H4b (Not supported) Intention to Join 4.56 (.16) 3.42 (.17) 2.95 (.14) 5.02 (.13) 3.85 (.15) 2.96 (.15) H4c (Not supported) Intention to Invite 2.87 (.13) 2.27 (.14) 2.04 (.11) 2.99 (.11) 2.39 (.14) 2.15 (.13) Note: Since covariates are included in the model, adjusted marginal means and standard errors are presented; each rating scale of the dependent variables is scored from 1 to 7; higher mean scores represent greater intentions and greater outcome expectancy. 99 Interaction Effects of CSR Campaign Conditions and Types of Brands RQ1 asked whether and which types of brands would moderate the main effects of CSR campaign conditions on the dependent variables. In particular, it was asked whether the proposed main effects of CSR campaign conditions on (a) the outcome expectancy, (b) the intention to join, and (c) the intention to invite, would be different when the CSR campaigns promoted practical brands (versus symbolic brands). As indicated in Table 9, the results demonstrated a significant interaction effect of the CSR campaign conditions and the types of brands on the intention to join [F (2, 704) = 8.50, p < .001] (RQ1b); however, such significant interaction effects were not found with regard to the intention to invite [F (2, 704) = .45, n.s.] (RQ1c) and the outcome expectancy [F (2, 704) = .07, n.s] (RQ1a). For further analyses of RQ1b, post-hoc analyses using Bonferroni adjustments were conducted for each type of brand. As illustrated in Figure 2, the findings showed that for practical brands, the CRM campaign condition (M = 4.49) led to greater intentions to join the brand page than either the CS campaign (M = 3.96) (p < .05) or the control condition (M = 2.81) (p < .001); further, the CS campaign condition (M = 3.96) had greater intentions to join than the control condition (M = 2.81) (p < .001). On the other hand, for symbolic brands, the CRM campaign condition (M = 5.21) yielded greater intentions to join than either the CS campaign (M = 3.24) (p < .001) or the control condition (M = 3.05) (p < .001); however, there was no significant mean difference between the CS campaign (M = 3.24) and the control condition (M = 3.05) (n.s.). Table 9 provides a summary of means and standard errors for the interaction effects of the CSR campaign conditions and the types of brands on the dependent variables. Figure 2 100 further visualizes how the proposed main effect of CSR campaign conditions on the intention to join was different for the two types of brands. Table 10 provides a summary of the findings of ANCOVAs that were tested for the proposed main and interaction effects. Table 9 Means and Standard Errors for Interaction Effects of CSR Campaign Conditions and Types of Brands on Dependent Variables Hypothesis RQ1a (Insignificant Interaction Effect) RQ1b (Significant Interaction Effect) RQ1c (Insignificant Interaction Effect) Dependent Variable Symbolic Brand CRM CS Control Mean Mean Mean (SE) (SE) (SE) Practical Brand CRM CS Control Mean Mean Mean (SE) (SE) (SE) Outcome Expectancy 4.28 (.16) 4.03 (.13) Intention to Join 5.21 (.16) Intention to Invite 2.85 (.15) 3.54 (.16) ab 3.24 (.15) 2.25 (.14) 2.59 (.15) c 3.05 (.14) 2.16 (.14) c 4.49 (.14) 2.96 (.11) 3.70 (.15) a 3.96 (.16) 2.46 (.13) 2.54 (.13) b c 2.81 (.13) 2.05 (.11) Note: Since covariates are included in the model, adjusted marginal means and standard errors are presented; means with different superscripts are significantly different from each other at the p < .05 level; each rating scale of the dependent variables is scored from 1 to 7; higher mean scores represent greater intentions and greater outcome expectancy. 101 Figure 2 Interaction Effects of CSR Campaign Conditions and Types of Brands on the Intention to Join Intention to join 5.21 5 4 4.49 Symbolic Brands 3.96 Practical Brands 3.05 3.24 3 2.81 CRM CS Control Note: Since covariates are included in the model, adjusted marginal means are presented; the rating scale of the intention to join is scored from 1 to 7; higher mean scores represent greater intentions to join. 102 Table 10 Summary of ANCOVAs for Hypothesis and Research Question Testing Dependent Variables Outcome Expectancy Intention to Join Main Effects CSR Campaign Conditions [H1, H2] Intention to Invite F (2, 704) = 61.19*** [H2] F (2, 704) = 73.75*** [H1a] F (2, 704) = 21.99*** [H1b] Types of Self-Friend Gender Composition [Not Hypothesized] F (1, 704) = .05 F (1, 704) = 5.78* F (1, 704) = .97 Types of Brands [Not Hypothesized] F (1, 704) = 8.62** F (1, 704) = 9.24** F (1, 704) = 1.21 F (1, 704) = .03 F (1, 704) = 9.82** F (1, 704) = 4.20* F (1, 704) = 78.29*** F (1, 704) = 27.60*** F (1, 704) = 42.49*** F (1, 704) = 78.48*** F (1, 704) = 36.08*** F (1, 704) = 15.20*** F (2, 704) = .01 [H4a] F (2, 704) = 1.25 [H4b] F (2, 704) = .03 [H4c] F (2, 704) = .07 [RQ1a] F (2, 704) = 8.50*** [RQ1b] F (2, 704) = .45 [RQ1c] F (1, 704) = .25 F (1, 704) = .14 F (2, 704) = .60 F (2, 704) = 1.32 Personal Involvement [Covariate] Brand Preferences [Covariate] Frequency of Using the ―Like‖ Button [Covariate] 2-Way Interaction Effects CSR Campaign Conditions x Types of Self-Friend Gender Composition [H4] CSR Campaign Conditions x Types of Brands [RQ1] Types of Self-Friend F (1, 704) = .27 Gender Composition x Types of Brands [Not Hypothesized] 3-Way Interaction Effects CSR Campaign F (2, 704) = .75 Conditions x Types of Self-Friend Gender Composition x Types of Brands [Not Hypothesized] Note: * p < .05, ** p < .01, *** p < .001; the covariates are included in the F-Statistics; each ANCOVA was conducted for each dependent variable. 103 Mediating Effects of Outcome Expectancy In the present study, H3a-b, H5a-b, and RQ2a-b predicted the mediating effect of outcome expectancy regarding the proposed main and interaction effects. Figures 3 and 4 illustrate the summary of the findings of the mediation models posted in H3a and H3b respectively. First of all, H3a posited the mediating role of the outcome expectancy on the main effect of presence and type of CSR campaigns on the intention to join the brand page. As illustrated in Figure 3, the first mediation analyses showed that the outcome expectancy significantly mediated the main effect of CRM over Control on the intention to join the brand page [F (5, 518) = 75.21, p < .001]. Specifically, the CRM campaign (over the control condition) had a significant influence on the intention to join the brand page, the dependent variable (coefficient = 1.74, p < .001); the CRM campaign (over the control condition) had a significant influence on the outcome expectancy, the proposed mediator (coefficient = 1.53, p < .001); the outcome expectancy, controlling for the CRM campaign (over the control condition), showed a significant influence on the intention to join the brand page (coefficient = .44, p < .001); finally, the effect of the CRM campaign (over the control condition) — controlling for the outcome expectancy — on the intention to join remained significant but at a weakened level (coefficient = 1.07, p < .001). Furthermore, the indirect effect of the CRM campaign (over the control condition) on the intention to join through the outcome expectancy was significant (coefficient = .68, p < .001 with a 95% confidence interval (CI) [.51, .87]). All of these findings indicate that the effect of the CRM campaign (over the control condition) on the intention to join the brand page was mediated by the outcome expectancy. As illustrated in Figure 3, the second mediation analyses showed that the outcome expectancy significantly mediated the main effect of CS over Control on the intention to join the 104 brand page [F (5, 448) = 70.16, p < .001]. That is, the CS campaign (over the control condition) had a significant influence on the intention to join the brand page, the dependent variable (coefficient = .72, p < .001); the CS campaign (over the control condition) had a significant influence on the outcome expectancy, the proposed mediator (coefficient = 1.07, p < .001); the outcome expectancy, controlling for the CS campaign (over the control condition), showed a significant influence on the intention to join the brand page (coefficient = .40, p < .001); finally, the effect of the CS campaign (over the control condition) — controlling for the outcome expectancy — on the intention to join remained significant but at a weakened level (coefficient = .29, p < .05). Furthermore, the indirect effect of the CS campaign (over the control condition) on the intention to join through the outcome expectancy was significant (coefficient = .43, p < .001, 95% CI [.29, .60]). Thus, it was concluded that the effect of the CS campaign (over the control condition) on the intention to join the brand page was mediated by the outcome expectancy. As also illustrated in Figure 3, the third mediation analyses showed that the outcome expectancy significantly mediated the main effect of CRM over CS on the intention to join the brand page [F (5, 456) = 46.39, p < .001]. That is, the CRM campaign (over the CS campaign) had a significant influence on the intention to join the brand page, the dependent variable (coefficient = .93, p < .001); the CRM campaign (over the CS campaign) had a significant influence on the outcome expectancy, the proposed mediator (coefficient = .40, p < .001); the outcome expectancy, controlling for the CRM campaign (over the CS campaign), showed a significant influence on the intention to join the brand page (coefficient = .60, p < .001); finally, the effect of the CRM campaign (over the CS campaign) — controlling for the outcome expectancy — on the intention to join remained significant but at a weakened level (coefficient = .68, p < .001). Furthermore, the indirect effect of the CRM campaign (over the CS campaign) on 105 the intention to join through the outcome expectancy was significant (coefficient = .24, p < .001, 95% CI [.08, .43]). All of these findings indicate that the effect of the CRM campaign (over the CS campaign) on the intention to join the brand page was mediated by the outcome expectancy. Taken together, the outcome expectancy was a significant mediator for the superior effects of the CRM campaign (over either the CS campaign or the control condition) on the intention to join the brand page, and the superior effects of the CS campaign (over the control condition) on the intention to join. Thus, H3a was supported. Figure 3 Mediating Role of Outcome Expectancy on the Relationship between the CSR Campaign Conditions on the Intention to Join (H3a) (i) Mediating Role of Outcome Expectancy on the Superior Effect of the CRM Campaign (over the Control Condition) on the Intention to Join 1.53 *** Outcome Expectancy CSR Campaign Conditions (CRM over Control) .44 *** Intention to Join 1.07 *** (1.74 ***) Note: The model was significant, F (5, 518) = 75.21, p < .001, controlling for three covariates; ―CRM over Control‖ indicates the first comparison of two CSR campaign conditions to show the different effect between the CRM campaign condition (coded as 1) and the control condition (coded as 0); the coefficient for the total effect is shown in parentheses; * p < .05, ** p < .01, *** p < .001. 106 Figure 3 (cont’d) (ii) Mediating Role of Outcome Expectancy on the Superior Effect of the CS Campaign (over the Control Condition) on the Intention to Join 1.07 *** Outcome Expectancy CSR Campaign Conditions (CS over Control) .40 *** Intention to Join .29 * (.72 ***) Note: The model was significant, F (5, 448) = 70.16, p < .001, controlling for three covariates; ―CS over Control‖ indicates the second comparison of two CSR campaign conditions to show the different effect between the CS campaign condition (coded as 1) and the control condition (coded as 0); the coefficient for the total effect is shown in parentheses; * p < .05, ** p < .01, *** p < .001. (iii) Mediating Role of Outcome Expectancy on the Superior Effect of the CRM Campaign (over the CS Campaign) on the Intention to Join .40 *** Outcome Expectancy CSR Campaign Conditions (CRM over CS) .60 *** Intention to Join .68 *** (.93 ***) Note: The model was significant, F (5, 456) = 46.39, p < .001, controlling for three covariates; ―CRM over CS‖ indicates the third comparison of two CSR campaign conditions to show the different effect between the CRM campaign condition (coded as 1) and the CS campaign (coded as 0); the coefficient for the total effect is shown in parentheses; * p < .05, ** p < .01, *** p < .001. Second of all, H3b posited the mediating role of the outcome expectancy on the main effect of CSR campaign conditions on the intention to invite friends to the brand page. As illustrated in Figure 4, the first mediation analyses showed that the outcome expectancy significantly mediated the main effect of CRM over Control on the intention to invite friends to 107 the brand page [F (5, 518) = 28.68, p < .001]. That is, the CRM campaign (over the control condition) had a significant influence on the intention to invite friends to the brand page, the dependent variable (coefficient = .88, p < .001); the CRM campaign (over the control condition) had a significant influence on the outcome expectancy, the proposed mediator (coefficient = 1.53, p < .001); the outcome expectancy, controlling for the CRM campaign (over the condition campaign), showed a significant influence on the intention to invite friends to the brand page (coefficient = .24, p < .001); finally, the effect of the CRM campaign (over the control condition)—controlling for the outcome expectancy—on the intention to invite remained significant but at a weakened level (coefficient = .51, p < .01). Furthermore, the indirect effect of the CRM campaign (over the control condition) on the intention to join through the outcome expectancy was significant (coefficient = .37, p <.001, 95% CI [.24 ~ .52]). All of these findings indicate that the effect of the CRM campaign (over the control condition) on the intention to invite friends to the brand page was mediated by the outcome expectancy. As illustrated in Figure 4, at the second mediation analyses showed that the outcome expectancy significantly mediated the main effect of CRM over CS on the intention to invite friends to the brand page [F (5, 456) = 23.75, p < .001]. That is, the CRM campaign (over the CS campaign) had a significant influence on the intention to invite friends to the brand page, the dependent variable (coefficient = .56, p < .001); , the CRM campaign (over the CS campaign) had a significant influence on the outcome expectancy, the proposed mediator (coefficient = .40, p < .01); the outcome expectancy, controlling for the CRM campaign (over the CS campaign), showed a significant influence on the intention to invite friends to the brand page (coefficient = .35, p < .001); finally, the effect of the CRM campaign (over the CS campaign) — controlling for the outcome expectancy — was not significant on the intention to invite (coefficient = .43, p 108 < .01.). Furthermore, the indirect effect of the CRM campaign (over the CS campaign) on the intention to join through the outcome expectancy was significant (coefficient = .14, p < .001, 95% [.04, .25]). All of these findings indicate that the effect of the CRM campaign (over the CS campaign) on the intention to join the brand page was mediated by the outcome expectancy. The third mediation analyses for the comparison of CS over Control was not conducted, owing to the insignificant finding from the earlier test for H1b: Post-hoc analysis for H1b showed there was no significant mean difference between the CS campaign and the control condition, in terms of the intention to invite friends to the brand page. Such an insignificant total effect of the independent variable on the dependent variable was disqualified for the mediation analysis (Preacher and Hayes 2008). Thus, the mediating role of outcome expectancy was not tested for the superior effect of the CS campaign (over the control condition) on the intention to invite. Taken together, the outcome expectancy was a significant mediator for the superior effects of either the CRM (over the control condition) or the CRM campaign (over the CS campaign) on the intention to invite friends to the brand page, whereas the mediating role of outcome expectancy in the relationship between the superior effect of CS (over the control condition) and the intention to invite was not tested. Thus, H3b was partially supported. 109 Figure 4 Mediating Role of Outcome Expectancy on the Relationship between the CSR Campaign Conditions on the Intention to Invite (H3b) (i) Mediating Role of Outcome Expectancy on the Superior Effect of CRM Campaign (over the Control Condition) on the Intention to Invite 1.53 *** Outcome Expectancy CSR Campaign Conditions (CRM over Control) .24 *** Intention to Invite .51 ** (.88 ***) Note: The model was significant, F (5, 518) = 28.68, p < .001, controlling for three covariates; ―CRM over Control‖ indicates the first comparison of two CSR campaign conditions to show the different effect between the CRM campaign condition (coded as 1) and the control condition (coded as 0); the coefficient for the total effect is shown in parentheses; * p < .05, ** p < .01, *** p < .001. (ii) Mediating Role of Outcome Expectancy on the Superior Effect of CRM Campaign (over the CS campaign) on the Intention to Invite .40 ** Outcome Expectancy CSR Campaign Conditions (CRM over CS) .35 *** Intention to Invite .43** (.56***) Note: The model was significant, F (5, 456) = 23.75, p < .001, controlling for three covariates; ―CRM over CS‖ indicates the third comparison of two CSR campaign conditions to show the different effects between the CRM campaign condition (coded as 1) and the CS campaign (coded as 0); the coefficient for the total effect is shown in parentheses; *p < .05, ** p < .01, *** p < .001. Furthermore, the present study predicted the moderated mediating models in H5a-b and RQ2a-b. Specifically, H5 proposed a mediating role of outcome expectancy in the interaction 110 effects of CSR campaign conditions and types of self-friend gender composition on (a) the intention to join and (b) the intention to invite; RQ2 posited a mediating role of outcome expectancy in the interaction effects of CSR campaign conditions and types of brand on (a) the intention to join and (b) the intention to invite. As noted in Analytic Strategies, moderated mediation analyses (i.e., conditional indirect effects) recommended by Preacher et al. (2007) seem to be appropriate to test H5a-b and RQ2ab. However, as seen in the earlier findings for H4a, there were no significant interaction effects of the CSR campaign conditions and the types of self-friend gender composition on the outcome expectancy. Likewise, since there was no significant interaction effect in the earlier test, further tests for moderated mediation analyses were disqualified (Preacher et al. 2007). Therefore, it would be reasonable to conclude that H5a and H5b were not qualified for the moderated mediation analyses. Thus, H5a and H5b were not tested. In a similar vein, as seen in RQ1a, there was no significant interaction between the CSR campaign conditions and the types of brands on the outcome expectancy, and consequently the moderating mediation analyses were not qualified to test RQ2a and RQ2b (Preacher et al. 2007). Thus, the statistic testing for RQ2a and RQ2b was not carried out. 111 CHAPTER VII GENERAL DISCUSSION AND CONCLUSION Study Overview A growing number of marketers are now involved in CSR campaigns: 58% and 78% of brands engaged in cause marketing in 2009 and 2010, respectively (PR Week and Barkely 2010). CSR campaigns are performed in the marketplace mostly by supporting charitable causes (Drumwright and Murphy 2001; Nan and Keo 2007; Seitanidi and Ryan 2007). This is because supporting charitable causes can provide an opportunity for marketers to obtain favorable responses from consumers for their brands (ExactTarget 2010; Morrissey 2008). Indeed, consumers show favorable opinions of marketers‘ cause support in general (Cone 2008; Edelman 2009). Such positive consumer responses are remarkable, considering the fact that today‘s consumers have developed knowledge about marketers‘ persuasive tactics over time (Friestad and Wright 1994) and tend to be skeptical toward advertising messages (Ford et al. 1990; Obermiller and Spangenberg 2000). It is also remarkable that 71% of consumers (which was 10% higher than the previous year) thought that marketers spent too much money on advertisements and should put more into supporting causes (Edelman 2009). Therefore, the effective provision of CSR campaigns could be a powerful marketing strategy to elicit consumer responses to brands in a favorable way, and further encourage consumers to be connected with brands. The current study stresses that such favorable consumer responses to brands supporting causes would be facilitated in the social media setting, because of the public display nature of social media (i.e., individuals‘ information and activities are seen to others in social media) (Boyd and Ellison 2007). For instance, on Facebook, consumers‘ joining and inviting activities 112 on the brand pages are publicized to their friends through their profile pages and their friends‘ News Feed pages (Slutsky 2011). As such, the public display nature of social media would encourage consumers to join brand pages and further invite others to these brand pages, once the brand pages feature CSR campaigns. The present study, thus, considers social media as the public setting in which consumers can express their connection with brands (e.g., become fans) to others, and it should be particularly true as the brands support charitable causes. Indeed, in communication research, social media users were found to be involved in impression management behaviors by controlling personal information seen by their friends in a social media setting in a favorable manner (Lampe et al. 2007; Krämer and Winter 2008). Moreover in charity literature, it was also suggested that consumers were more likely to increase their support for cause campaigns when they realized that their support was done in the presence of others or it was visible to others (Lacetera and Macis 2008; White and Peloza 2009). Taken together, it seems to be reasonable to suggest that CSR campaigns on brand pages in social media would elicit consumer motivations to look favorable to their friends and further publicize membership of the brand pages. Despite the growing potential of CSR campaigns in social media as a new cause marketing practice, little research has been carried out by advertising and marketing scholars. However, it should be essential to initiate a scholarly inquiry into the role of CSR campaigns in the social media setting in mobilizing consumers to respond to brands favorably. As an attempt to respond to such a call, the present study aims to explore whether, which, and when CSR campaigns in social media would be effective for the intention of consumers to be connected with brands through becoming fans. As the theoretical framework underlying all of these questions under study, the present study suggests the impression management theory (Goffman 113 1959; Leary and Kowalski 1990; Schneider 1981) — ―the process by which individuals attempt to control the impressions others form of them (Leary and Kowalski, 1990, p. 34).‖ More specifically, the first objective was to examine whether the presence of CSR campaigns on the brand pages in social media would be more effective in having consumers respond to the brands in a favorable light, than the absence of CSR campaigns. The current study expected that by favorably responding to the brands featuring any CSR campaigns, consumers would be given an opportunity to look favorable to their friends, possibly as those who supported the causes or liked the brands‘ cause campaigns. The second objective of the present study was to clarify which types of CSR campaigns would be more effective for positive consumer responses to brands in social media. Here, two types of CSR campaigns were compared in the social media setting: CRM and CS. In a typical situation, the CRM campaign in social media is performed when brands promise donations to specific charitable cause each time a consumer joins their brand pages. On the contrary, the CS campaign in social media is usually implemented when brands announce their official sponsorship of certain causes on their brand pages in social media, in the hope of encouraging consumers to join their brand pages. The study thus suggested that joining and inviting activities on the brand pages featuring the CRM campaign would help consumers manage their impression favorably by showing their own direct contribution to the causes (e.g., showing my $1 being added to the brand‘s donation to a cause) to their friends; on the contrary, it was suggested that the joining and inviting activities on the brand pages featuring the CS campaign would also be effective in presenting their favorable image to others, by expressing their indirect association with the brand‘s CS campaign (e.g., showing my liking of the brand‘s sponsorship for the cause). 114 As a theoretical account for the relative effectiveness of these two types of CSR campaigns, the present study suggested two tactics of impression management: direct and indirect tactics (Cialdini et al. 1976; Cialdini and Richardson 1980; Richardson and Cialdini 1981). For the direct tactic (e.g., ingratiation, self-promotion), for example, people attempted to present aspects of their own abilities or traits (e.g., likeability, competency) to potentially look favorable and competent to others (John and Pittman 1982). (See John and Pittman (1982) for more categories of direct tactics; e.g., ―I have a Ph. D. in Physics.‖) For the indirect tactic, people tried to publicize the mere connection with others to manage their impression (Cialdini 1989; Cialdini and Richardson 1980; Richardson and Cialdini 1981), typically by associating themselves with those favorable or successful and presenting such associations to others (Cialdini et al., 1976) (e.g., ―I grew up in the same town as Albert Einstein‖). Thus, the study examined which type of CSR campaign—both of which would be adopted as direct and indirect tactics of impression management among consumers—was more effective in motivating consumers to present their favorable images to their friends by publicizing their favorable responses to the brands in social media. As the third objective, this study was to understand when consumers were more or less motivated to look favorable in the eyes of their friends, and thus respond favorably to the brands featuring CSR campaigns on their social media pages. The study particularly focused on two situational factors, which are often encountered in the social media setting but neglected in cause marketing literature: (a) whether consumers were invited to the brand pages by same-gender friends or different-gender friends (i.e., types of self-friend gender composition); (b) whether the brands featuring the CSR campaigns were perceived as being symbolic or practical (i.e., types of brands). To support such arguments theoretically, further discussions were provided regarding 115 how impression management was motivated and constructed differently depending on the gender types of the target, and how the images of brands were reflected in the consumers who joined the brand pages in social media. To achieve these objectives, the present study employed a 3 (CSR campaign condition: CRM, CS, and control) x 2 (type of brands: symbolic and practical) x 2 (type of self-gender composition: same and different) between-subjects experimental design. All procedures were done online to increase the ecological validity of the experimental study (Brewer 2000). A total of 720 college students currently enrolled in a variety of departments at MSU participated in the study. Summary of Findings and Conclusion Table 11 summarizes the hypothesis testing and research question results. First, the present study explored the baseline research pertaining to whether the presence of each CSR campaign (i.e., CRM or CS) affects the favorable responses of consumers to the brands. The experiment found that, as expected, the brand pages under the presence of CRM campaigns yielded a greater intention to join the brand page (H1a) and a greater intention to invite their friends to the brand page (H1b), compared to those without a CSR campaigns (i.e., the control condition). Furthermore, the presence of CS campaigns on the brand pages showed mixed outcomes in terms of the pattern of consumer response to brands. That is, the brand pages with CS campaigns led to a greater intention to join, than those without CSR campaigns (H1a); however, the presence of CS campaigns was not effective on the intention to invite (H1b). Such findings indicate that the presence of any kind of CSR campaign (i.e., either the CRM campaign or the CS campaign) resulted in the greater intention of consumers to become fans of the brands in social media, as opposed to the absence of CSR campaigns. 116 Table 11 Summary of Hypothesis Testing and Research Question Results Hypotheses Testing Results Main Effects of CSR Campaign Conditions H1a CSR Campaign Conditions  Intention to Join Supported H1b CSR Campaign Conditions  Intention to Invite Partially Supported H2 CSR Campaign Conditions  Outcome Expectancy Supported Mediating Role of Outcome Expectancy for Main Effects H3a CSR Campaign Conditions  Outcome Expectancy Supported  Intentions to Join H3b CSR Campaign Conditions  Outcome Expectancy Partially Supported  Intentions to Invite Interaction Effects of CSR Campaign Conditions x Types of Self-Friend Gender Composition H4a CSR Campaign Conditions x Types of Self-Friend Gender Not Supported Composition  Outcome Expectancy H4b CSR Campaign Conditions x Types of Self-Friend Gender Not Supported Composition  Intentions to Join H4c CSR Campaign Conditions x Types of Self-Friend Gender Not Supported Composition  Intentions to Invite Mediating Role of Outcome Expectancy for Interaction Effects of CSR Campaign Conditions x Types of Self-Friend Gender Composition a H5a CSR Campaign Conditions x Types of Self-Friend Gender Not Qualified Composition  Outcome Expectancy  Intentions to Join a H5b CSR Campaign Conditions x Types of Self-Friend Gender Not Qualified Composition  Outcome Expectancy  Intentions to Invite Interaction Effects of CSR Campaign Conditions x Types of Brands RQ1a CSR Campaign Conditions x Types of Brands  Outcome Not Significant Expectancy RQ1b CSR Campaign Conditions x Types of Brands  Intentions Significant to Join RQ1c CSR Campaign Conditions x Types of Brands  Intentions Not Significant to Invite Mediating Role of Outcome Expectancy for Interaction Effects of CSR Campaign Conditions x Types of Brands a RQ2a CSR Campaign Conditions x Types of Brands  Outcome Not Qualified Expectancy  Intentions to Join a RQ2b CSR Campaign Conditions x Types of Brands  Outcome Not Qualified Expectancy  Intentions to Invite a Note: The moderated mediation analyses were not carried out owing to the insignificant interaction effects, as seen in H4a or RQ1a, according to the suggestion of Preacher et al. (2007). 117 However, only the presence of the CRM campaigns, not the presence of the CS campaigns, was persuasive in encouraging consumers to send invitation e-mails to their friends, in the social media context. Such different patterns of consumer responses to brands in terms of the types of CSR campaigns may be explained by the levels of consumer engagement in brands within social media. Despite the difficulties in defining consumer engagement in social media, scholars have frequently specified the levels of engagement — from participating and sharing, to networking online — by the degrees of consumer attention and interaction within social media (Ghuneim 2008). Accordingly, joining the brand pages may be understood as a relatively lower level of consumer engagement, possibly representing ―participating,‖ than inviting others to the brand pages, which may be represented by ―sharing‖ or ―networking.‖ The findings of the current study indicated that the CRM campaign in social media would be effective in encouraging consumers to invite other people to the brand page, whereas the CS campaign in social media would not reach such levels of consumer engagement in the brands. Second, the present study investigated which types of CSR campaigns would lead to the more favorable consumer responses to the brand in social media. The findings showed that the brand page with the CRM campaign yielded the greater intentions to join (H1a) and invite their friends (H1b), than the brand page with the CS campaign. Thus, the CRM campaign (versus the CS campaign) seems to be more effective in eliciting favorable responses from consumers toward the brand in social media. In order to understand the psychological mechanisms underlying the effects of CSR campaigns on the favorable consumer responses to the brands in social media, the current study examined a mediating role of outcome expectancy to look favorable to friends. The findings of 118 preliminary analyses demonstrated that, as predicted in H2, the brand page with the CRM campaign had the greatest outcome expectancy of consumers to look favorable, followed by that with the CS campaign, then that in the control condition. The mediation analysis further confirmed that such outcome expectancy was a significant mediator, working as a mechanism underlying the greatest effects of the CRM campaign followed by the CS campaign and then the control condition on the intention to join (H3a). As such, when consumers encountered either CRM campaigns or CS campaigns (versus none of the CSR campaigns) on the brand page in social media, they were more likely to expect that their joining activities would affect their favorable images to their friends. Consequently, such a higher expectation resulted in greater intentions to join the brand pages with any CSR campaign (versus none of the CSR campaigns). Another mediation analysis demonstrated that the outcome expectancy was a significant mediator underlying the superior effects of the CRM campaign (over the CS campaign) on the intention to invite (H3b). Such findings confirmed that featuring the CRM campaign, rather than the CS campaign, resulted in a greater expectation of consumers to look favorable to friends, which in turn led to a greater intention to invite the friends to the brand page. Taken together, the mediation analyses indicated that the impression management of consumers in social media could be one of the possible theoretical approaches to predict the more favorable responses of consumers to the brands, when the CRM campaign (versus the CS campaign and/or the control condition) was present on brand pages in social media. Third, in order to clarify when the CSR campaigns would be more or less effective in the social media setting, two situational factors were examined as the moderators influencing the main effects of CSR campaigns on the responses to the brands: type of self-gender composition (i.e., same versus different) and type of brand (i.e., symbolic versus practical). 119 Regarding the interaction effect between the CSR campaign conditions and the types of self-friend gender composition (i.e., same versus different) on the outcome expectancy to look favorable (H4a), the intention to join the brand pages (H4b), and the intention to invite their friends to the brand pages (H4c), the experiment demonstrated no significant findings. Several issues might be attributed to the insignificant findings: (1) the use of the scenario method in manipulating the types of self-friend gender composition; (2) the value of interaction with the different gender partner as a potential confounding variable. As the first possible reason, it should be noted that the current study employed a hypothetical situation to manipulate the email invitation from male or female friends. That is, the participants were instructed to imagine getting an e-mail invitation from one of their male or female friends at the beginning stage of the experiment. The scenario method seemed to have some limitations in manipulating the social context to examine the hypotheses (Ashworth, Darke, and Schaller 2005). Particularly, the present study considers that the participants‘ responses to these imagined situations (e.g., imagine getting e-mail invitations from their male or female friends) may not coincide with their responses to a real situation, such as getting actual e-mail invitations from their friends. As the second attribution to the insignificant outcomes, the value of interactions with different-gender friends should be noted with regard to the motive for impression management. Previous studies suggested that people tend to put more value on the interaction with the different-gender target than the interaction with members of their own sex (Leary et al. 1994; Wheeler and Nezlek 1997), because the friendship with the different-gender target might be more scarce (Leary et al. 1994). It was further suggested that as the valued resources were scarce, people were likely to engage in tactics of impression management (Pandey and Rastogi 1979; Pandey 1986). For example, Pandey and Rastogi (1979) found that impression management tactics (e.g., 120 ingratiation) increased in an organizational setting, as the chance to gain career success became more scarce in a competitive job situation. These previous findings led to the attribution that the self-friend gender composition might not provide sufficient contexts of impression management, possibly for those who were in a romantic relationship or married, because the encounter with different gender friends might not be considered as being valuable among them. Regarding the interaction effect between the CSR campaign conditions and the types of brands (i.e., symbolic versus practical), the present study found that such interaction effects did not emerge for the outcome expectancy to look favorable (RQ1a) and the intention to invite their friends to the brand pages (RQ1c), but did for the intention to join the brand pages (RQ1b). Furthermore, the effects of CSR campaigns on the intention to join the brand pages showed mixed findings, depending on the types of brands. That is, when the brands had practical images (e.g., Aquafina), as expected, the CRM campaigns generated the greatest intention to join the brand page, followed by the CS campaigns and the control condition. However, when the brands had symbolic images (e.g., Fiji Water), CRM campaigns yielded a significantly greater intention to join the brand pages than either the CS campaigns or the control condition; and furthermore, there was no significant difference between the CS campaigns and the control condition. These findings demonstrate that when the brands are practical, the presence of either a CRM campaign or a CS campaign (versus the absence of any CSR campaigns) yielded the higher intention to join the brand pages; in particular, the CRM campaign (versus the CS campaign) had the greater intention to join the brand pages. On the other hand, when the brands are symbolic, only the presence of a CRM campaign (versus the absence of any CSR initiative) resulted in the higher intention to join than the absence of CSR campaigns, whereas the presence of a CS campaign did not. 121 As the possible rationales for such findings, the present study suggests that when the brands are perceived as being practical among consumers, the role of CSR campaigns becomes more vital in encouraging consumers to join the brand pages and invite others to the brand pages. It‘s because, on the practical brand pages, consumers may not expect that becoming members of the practical brand pages themselves would provide them with added benefits for presenting their social image to friends saliently. However, when the practical brands feature any type of CSR campaign, particularly a CRM campaign, consumers may perceive that their connection with the brands could be utilized as a way to present their favorable images saliently. Thus, when the brands are practical, consumers are more likely to process external cues, such as the provision of CSR campaigns, than the brand images themselves, while deciding their membership of the brand pages in social media. In short, when the brands are practical, consumers may consider whether the presence and type of CSR campaigns have the capacity to make them look favorable to their friends. Such considerations would determine their intentions to join the social media pages for practical brands. On the other hand, when the brands are perceived as being symbolic among consumers, consumers seem to process two types of CRM campaigns in different ways, while deciding their joining intentions on the symbolic brand pages in social media. Thus, the null effects of CS campaigns on the intention to join would be attributed to the notion that the favorable images consumers could give off by joining the symbolic brand pages would be indirectly associated with the symbolic brands‘ cause support, rather than directly driven from their own support. Thus, joining the symbolic brand pages featuring CS campaigns may not be perceived among consumers as providing sufficient benefits for the enhancement of their social images. However, when the symbolic brands promote a CRM campaign, joining the brand pages seems likely to be 122 considered as a way for consumers to manage their impressions favorably and directly. Thus, consumers may consider that the positive images of cause supporters via their participation in the CRM campaign could be intensified because, as seen earlier in the affect transfer discussion, the symbolic features of brands would be transferred to the impression of those who join the brand pages. Therefore, when the brands are symbolic, consumers may focus on which types of CSR campaigns are featured on the brand pages in social media, rather than on whether the CSR campaigns are present. This may be the case because only the CRM campaigns would have the capacity to make them look favorable to their friends more saliently. Limitations and Future Studies The following limitations should be noted prior to providing the implications: (1) using the college student sample; (2) testing a single product category; (3) measuring behavioral intentions rather than actual behaviors; and (4) providing a hypothetical, rather than real, situation for the manipulation of the e-mail invitations from friends. First, one major limitation of this study is the use of college students as a sample. It is true that targeting college students as those who join the brand pages in social media, seems to be a vital strategy for marketers. Indeed, college students are active participants in various prosocial activities. Their volunteer rate increased by 20% (compared to the 9% increase of the all-adult volunteer rate) from 2001 to 2005 (Dote et al. 2006). Furthermore, 75% of young Americans who were born between 1982 and 1992 showed favorable opinions on cause marketing, indicating that marketers had the material resources to help causes; 60% thought that marketers had the knowledge to support causes (Esquivel, McMahon and Raymond 2010). However, it is also true that about 47% of Facebook users are non-college students, whereas only 40% of them are college students (Quantcast 2010). Therefore, to increase the potential generalizability of our 123 findings, future studies should be replicated among the non-college-student population with various age groups‘ use of one single product—bottled water. Although the choice of bottled water was considered appropriate, it is not representative of all products featuring CSR campaigns in social media. Several studies (Chang 2008; Strahilevitz and Myers 1998) suggested that in the CRM campaign context, hedonic products (e.g., ice cream) were more effective in increasing consumers‘ favorable responses to the CRM campaigns than utilitarian products (i.e., ink printer). Therefore, the consideration of symbolic and practical brand types under multiple product categories would increase confidence in the generalizability of the current study‘s findings. Third, this experimental study asked the participants to indicate their intentions to join the brand pages and invite their friends to the brand pages, rather than have them show their actual joining and inviting behaviors on the brand pages. The present study chose to measure behavioral intention, owing to the difficulties in measuring the actual behaviors of consumers directly in the experimental setting with non-working Facebook brand pages. Studies have found that the behavioral intentions made a significant impact on the actual behaviors in the experimental setting: Webb and Sheeran (2006), who conducted a meta-analysis of 47 experimental studies to examine how behavioral intention compared to actual behaviors, found that a medium-large change in intention (d = .66) led to a small-to-medium change in behavior (d = .36). Nevertheless, there is still a concern that the findings of the present study pertaining to the consumer responses to the brands would be different when they were asked to demonstrate their actual behaviors. To reproduce the findings of the current study, future studies should be followed by measuring the actual joining and inviting behaviors of consumers on the brand pages featuring CSR campaigns in social media. 124 Finally, the present study used a hypothetical situation to develop the gender composition stimuli. At the beginning of the experiment, all participants were asked to imagine getting an email from their male or female friends inviting them to the brand pages. As stated previously, the current study attributes null findings from the moderating role of types of self-friend gender composition, to such methodological issues in the experiment design. By employing the e-mail invitation from real friends for the study design, future studies may secure more accurate findings pertaining to the role of friendship in CSR campaign effectiveness in social media. In addition, the question of how consumer motivation to be connected with brands is open to future research. The present study controlled individual consumers‘ brand preferences as a covariate in order to explore the roles of CSR campaign presence and type purely in consumer responses to brands promoting CSR campaigns. However, consumers can express themselves in social media through connections with either causes or brands, or both. Future research should explore whether consumers join the brand pages featuring CSR campaigns in the social media setting, to be seen as supporters for a specific cause or endorsers for a specific brand. Such research will extend our findings about the interaction effect of CSR campaigns and brand types. It could further broaden our understanding about when and how consumers are mobilized to positively respond to brands featuring CSR campaigns in social media. Future studies could explore the extent to which public settings online and offline would influence consumer participation in CSR campaigns in a similar or dissimilar way. Such studies will enhance our understanding about the role of social media in consumers‘ impression management through consumer participation in CSR campaigns. Theoretical Implications 125 Despite the limitations mentioned above, the findings of this study provide theoretical implications for advertising and marketing scholars in several ways: (1) by exploring the role of social media as new media channels to facilitate consumer participation in CSR campaigns; (2) by applying impression management motivation to cause marketing literature; and (3) by providing a new theoretical account for any cause-supportive behaviors, particularly among today‘s young consumers. First, the present study extends the academic interest in social media as a new public setting for consumer participation in CSR campaigns. Even though literature (Ariely et al. 2009; Bereczkei et al. 2010; White and Peloza 2009) has stressed the public setting as the key contextual factor in determining individual participation in charity campaigns, little scholastic efforts have been made, to date, to examine the role of social media as a context in which consumers would be motivated to respond to CSR campaigns in a favorable way. Based on these calls for the studies, the present study attempts to initiate theoretical investigations on the strategic use of social media as new cause marketing campaign channels. Second, the present study elaborates on theoretical discussions of how the consumer motivation approach is integrated with the cause marketing context. The majority of previous approaches to consumer responses to CSR campaigns have merely explored the relation between brands or marketers and causes. Some of them have been marked by a scholastic interest in the brand-cause fit as a way to increase consumer responses to the brands (Cornwell and Coote 2005; Gupta and Pirsch 2006; Nan and Heo 2007; Simmons and Becker-Olsen 2006; Tangari et al. 2010; Zdravkovic 2010). Others have focused on exploring how consumers perceived the motives of marketers to perform CSR campaigns in the marketplace (Dean 2003/2004; Rifon et al. 2004; Szykman et al. 2004; Webb and Mohr 1998). The current study advances these 126 previous perspectives. By adopting impression management as a theoretical framework, the present study indicated that consumer motivation to look favorable in social media could be combined with their participation in CSR campaigns in social media, thereby showing positive responses to the brands that support the causes. Such new theoretical approaches will provide a more complete picture to understand consumer participation in CSR campaigns in the social media context. Furthermore, by merging cause marketing literature and impression management in social media, this study also provides more theoretical elaborations to the growing body of impression management studies in social media, which have thus far received little attention in the fields of advertising and marketing. Particularly, by adopting the notions of direct and indirect tactics, which have not been actively explored in academic research (Tal-Or 2010), the present study attempts to extend theoretical discussion about the relative effects of CSR campaign types among consumers. Third, the current study extended theoretical discussion about young consumers‘ impression management in social media into the cause-supportive context. Identifying the factors that influence individuals‘ responses to cause campaigns has indeed captured considerable interest in literature (Cornwell and Coote, 2005; Dean 2003/2004; Gwinner and Eaton, 1999; Nan and Heo, 2007; Pope and Voges 2000). Empathy-based altruism has been suggested as one strong internal factor that motivates people to engage in cause supportive behaviors (Batson et al. 1981; Eisenberg and Miller 1987). However, as indicated by recent research (O‘Brien, Hsing, and Konrath 2010), college students today are not as empathetic as those of the 1980s and the 1990s; rather, today‘s college students are assumed to be more self-centered, narcissistic, and competitive. Noting these traits, this study relies on the impression management theory to predict how the young generation would be motivated to participate in prosocial activities. The findings 127 of the current study, therefore, broaden theoretical explanation regarding how to enhance young people‘s involvement in the cause campaigns and further their engagement in brands in the marketplace. Practical Implications Along with the theoretical implications, the managerial implications should be stated as: (1) providing the evidence about whether or not CSR campaigns would be effective on the positive response of consumers to brands; (2) recommending to marketers which types of CSR campaigns would be more or less effective; and (3) identifying when the CSR campaigns would be effective, in social media. First, pertaining to whether the presence of CSR campaigns would result in more favorable responses of consumers in social media, the findings of the present study suggest that any kind of CSR campaign would be effective on the positive responses of consumers in social media. According to the findings of the current study, marketers can secure consumers as fans of their brands, either by developing a CRM campaign or simply by featuring information about the CS campaign on their brand pages in social media, even though the provision of the CRM campaign can be more beneficial to the marketers, than the CS campaign. Second, regarding which types of CSR campaigns (i.e., the CRM campaign or the CS campaign) would be more effective, the findings suggest that the development of CRM campaigns (versus CS campaigns) on the brand pages in social media would result in the more favorable responses of consumers to the brands. As indicated from the mediation analysis, marketers need to understand that consumers are more likely to respond to the brands with the CRM campaign (versus those with the CS campaign) because it could be perceived by consumers as a more effective way to show their own support for the causes. Furthermore, the 128 provision of a CRM campaign, more so than that of CS campaign, would be further beneficial for marketers, particularly to facilitate consumer engagement in the brand pages in a more active way, such as inviting their friends to the brand pages. Finally, regarding when CSR campaigns would be more or less effective, the present study suggests that the proposed effects of CSR campaigns on the intention to join would be more pronounced for practical brands, rather than symbolic brands. When brands are perceived as having practical images, marketers need to note that the provision of any kind of CSR campaign could become an opportunity for them to promote their brands among social media users, even though the CRM campaign can have greater effects in general than the CS campaign. That is, when encouraging consumers to become fans of the practical brands, the practical brands would benefit most from the development of a CRM campaign on their social media page, followed by the provision of a CS campaign, or the absence of any cause campaigns on their social page. On the other hand, when their brands are perceived as having symbolic images among consumers, marketers need to focus on the CRM campaign only in a social media setting, because, according to our finding, the provision of the CS campaign did not elicit the willingness of consumers to become members on their social media brand pages. As such, when marketers attempt to develop any CSR campaign, they need to take into consideration whether the brands are positioned as symbolic or practical among consumers. All of these results highlight that when marketers attempt to initiate CSR campaigns within the social media context, they need to understand consumer motivation to deliver their image in a favorable light to their friends in a social media channel. Noting the growing number of marketers now involved in CSR campaigns, it is clear that CSR campaigns in social media have emerged as a potential marketing practice to make the connection between the brands and a 129 growing number of social media users. Therefore, the current study helps to initiate scholastic and industry interest in the strategic use of social media as a way to develop consumers‘ favorable responses to the brands, and cultivate consumers‘ active engagement in new forms of brand communication such as CSR campaigns. 130 APPENDIX 131 APPENDIX A Table 12 Manipulations for CSR Campaign Messages by Brand Types a Cause-Related Marketing (CRM) Corporate Sponsorship (CS) Control a Symbolic Brand As every person clicks the ‗Like‘ button to join our Facebook page, we will donate $1 to ‗Go Recycling‘ up to $100,000 during the next two weeks. You Join  We donate $1. Practical Brand As every person clicks the ‗Like‘ button to join our Facebook page, we will donate $1 to ‗Go Recycling‘ up to $100,000 during the next two weeks. You Join  We donate $1. Once you click the ‗Like‘ button to join our Facebook page, the following message will be seen on your profile page, ―John Smith (i.e., your name) and Fiji Water donate $1 to ‗Go Recycling.‘‖ We will donate $100,000 during the next two weeks. We proudly sponsor ‗Go Recycling.‘ Once you click the ‗Like‘ button to join our Facebook page, the following message will be seen on your profile page, ―John Smith (i.e., your name) and Aquafina donate $1 to ‗Go Recycling.‘‖ We will donate $100,000 during the next two weeks. We proudly sponsor ‗Go Recycling.‘ Once you click the ‗Like‘ button to join our Once you click the ‗Like‘ button to join our Facebook page, the following message will be Facebook page, the following message will be seen on your profile page, ―John Smith (i.e., seen on your profile page, ―John Smith (i.e., your name) likes Fiji Water‘s donation to ‗Go your name) likes Aquafina‘s donation to ‗Go Recycling.‘‖ Recycling.‘‖ From head to toe, our bodies rely on water for essential functions. Sip smarter by learning more about water. Water is our bodies‘ principal component, making up 60% of our body weight, 85% of our brain, and 10% of our teeth. Water flushes toxins out of vital organ, carries nutrients to our cells, and provides a moist environment for our cell tissues. Every day we lose water through our breath, perspiration, urine, and bowel movements. For our body to function, we must replenish its water supply. See our Facebook page to learn more about water.‖ a Note: The bottled water picture for each brand is also shown on the left position on the brand Web page. 132 APPENDIX B EXAMPLES OF QUESTIONNAIRE Dear Students, You are randomly selected to participate in a research study about college students‘ general opinion of brand pages on Facebook. The current research study is to understand how MSU students feel and use a brand page on Facebook. All of your responses are confidential and will be used for educational purposes only. In return for your participation, we will offer the prizes a total of twelve students, two winners of Grand Prize (i.e., $ 99.99 Amazon gift card for each) and ten winners of the Second Prizes (i.e., $20.00 Amazon gift card for each). Out of all participants, twelve students will be randomly selected as the winners for the prizes and will be notified by email within 48 hours of being selected at the early week of May. The opportunity to participate in this study will end on the next Saturday, April 30, at 5:00 PM. If you have specific questions about the survey or have any technical difficulties, please contact me: Hyun Ju Jeong (jeonghyu@msu.edu). We greatly appreciate your participation! If you agree to participate, please click on the link below. http://www.msu.edu/~jeonghyu/dis/randomlink.html a Hyun Ju Jeong A doctoral candidate in Media & Information Studies Department of Advertising, Public Relations, and Retailing College of Communication Arts and Sciences Michigan State University Email: jeonghyu@msu.edu a Note: This link contains JavaScript by which each participant is randomly exposed to one of the twelve experimental conditions. Thus, one URL is developed. [Go to the next page] 133 [CRM Campaign Condition for the Symbolic Brand, Invited by a Male Friend] Please click the number that best represents your opinion on each statement. The following statements ask your opinion of a specific brand of bottled water – Fiji Water. Strongly Disagree (1) I like Fiji Water better than other bottled water brands. (2) I would use Fiji Water more than I would use other bottled water brands. (3)Fiji Water is my preferred brand over other bottled water brands. (4)I would be inclined to buy Fiji Water over other bottled water brands. Strongly Agree                             The following statements ask your opinion of the issue of recycling in general. Recycling is defined as a process of converting used materials into new products to prevent the waste of potentially useful resources. ―To me, the issue of recycling is…‖ Unimportant Of no concern Means nothing to me Does not matter to me Insignificant                                    Important Of concern to me Means a lot to me Does matter to me Significant Today, Facebook users tend to click the “Like” button on Facebook to join anything they like (e.g., Facebook pages for brands or celebrities). How often do you click the ―Like‖ button to join a Facebook page in general? Never    [Go to the next page] 134    Quite Often  Please follow each step of the directions carefully. 1. First, read the following definition of male casual friends. MALE CASUAL friends are ―male people with whom you enjoy spending time but do not know as well as and so not feel as close to as your ‗close personal friends.‘ You occasionally communicate with the male casual friends on Facebook and via email. Your friendship seems to be based upon the circumstances in which you find yourself (e.g., you work together, are in a class together, etc.). You plan on remaining friends as long as circumstances permit. Your friendship would dissolve if your circumstances were to change." 2. Now, think specifically one of your male casual friends and provide the first name of one of your male casual friends here. [Go to the next page] Now, image that you get an email from your male casual friend whom you just specified in the previous page. The following is the content of the email you get from your male friend. Please read the email content carefully and thoroughly. Figure 5 The Content of E-mail Sent from Male Friend who Invites Fiji Water Facebook Page With CSR Campaigns From mycasualfriend@msu.edu To me@msu.edu Title Invitation Your male casual friend sent an e-mail invitation to you! Hi, Fiji Water, a bottled water brand, supports „Go Recycling,‟ a national charitable cause to save our natural resources. Please join Fiji Water‟s official Facebook page to support the cause campaign of “Go Recycling.” Your male casual friend Next, click ‗Next Page‘ to visit the Fiji Water‘s official Facebook page [Go to the next page] 135 The following webpage is the official Facebook page of Fiji Water. Please examine every element in the following webpage carefully and thoroughly. Figure 6 The Manipulation of Fiji Water Facebook Page with CRM Campaigns (For Interpretation of the References to Color in This and All Other Figures, the Reader is Referred to the Electronic Version of This Dissertation) [Go to the next page] 136 Based on your review of the Fiji Water Facebook page in the previous section, please indicate your behavioral intention regarding each statement. Feel free to go back to previous page by clicking “Previous Page” to review the content of Fiji Water’s Facebook page again. 1. How likely are you to click the “Like” button to join the Fiji Water Facebook page? Unlikely Improbable Impossible Uncertain                             Likely Probable Possible Certain 2. How likely are you to invite other people to the Fiji Water Facebook page? Unlikely Improbable Impossible Uncertain                             Likely Probable Possible Certain 3. What do you think about your clicking of the “Like” button on the Fiji Water Facebook page? ―Clicking the ―Like‖ button to join the Fiji Water Facebook page will help me make ……… to my Facebook friends‖ (1) (2) (3) (4) Strongly Disagree         A good impression An altruistic impression A socially responsible impression An environmentally knowledgeable impression [Go to the next page] 137             Strongly Agree         You’re almost at the end. Please answer the following questions about the e-mail invitation and the Facebook page for a bottled water you just reviewed. 1. At the beginning stage of this study, you were asked to imagine getting an email from one of the following persons inviting you to the Facebook page for a bottled water brand (Select one).  Male casual friend  Female casual friend  Don‘t know/don‘t remember 2. At the beginning stage of this study, you were asked to visit the Fiji Water Facebook page. Please click the number that best describe your opinion of this brand. (1) People drink Fiji Water as a way of expressing their personality (2) Drinking Fiji Water says something about the kind of person you are. (3) Fiji Water is for people who are down-toearth. (4) Drinking Fiji Water is practical. 3. Strongly Disagree      Strongly Agree                        In the previous section, you were invited to the Fiji Water’s Facebook page, featuring one of the following messages. Please select one that best describes the message (Select one).  "As every person clicks the ―Like‖ button to join our Facebook page, we will donate $1 to ―Go Recycling‖ up to $100,000 during the next two weeks.‖  ―We will donate $100,000 to ―Go Recycling‖ during the next two weeks.‖  "Water is our bodies‘ principal chemical component, making up 60% of our body weight, 85% of our brain, and 10% of our teach."  Don‘t know/don‘t remember 4-1. Do you think your clicking of the “Like” button to join the Fiji Water Facebook page you just reviewed would help you make a good impression on your Facebook friends? (Select one).  Yes  No [Go to the next page] 138 4-2. If you answered as “Yes” in the previous question (4-1), please indicate why you think so by clicking the number that represents your opinion most. ―Because of my ____ support for the brand‘s cause campaign‖ Indirect Passive Secondhand                      Direct Active Firsthand [Go to the next page] 5. How do you perceive the relationship between the “Go Recycling” cause and the “Fiji Water” brand? Not compatible Bad fit Irrelevant Bad match                 [Go to the next page] 139             Compatible Good fit Relevant Good match Finally, please answer the following questions about yourself. 1. What is your age? years old 2. What is your gender?  Female  Male 3. What is your nationality?  American  Other (please specify) 4. How long have you used Facebook? (If you are not a Facbook user, please enter 0). year(s) month(s) 5. Please type the name of your major at MSU (e.g., chemistry, communication). 6. What is your class standing? (Select one)     Freshman Sophomore Junior Senior 7. Please type your name (for incentive). Last name First name 8. Please let us know your valid email address by which I can reach you (for incentive). E-Mail address ---- Thanks for your participation ---140 [CRM Campaign Condition for the Practical Brand, Invited by a Female Friend] Please click the number that best represents your opinion on each statement. The following statements ask your opinion of a specific brand of bottled water – Aquafina. Strongly Disagree (1) I like Aquafina better than other bottled water brands. (2) I would use Aquafina more than I would use other bottled water brands. (3) Aquafina is my preferred brand over other bottled water brands. (4) I would be inclined to buy Aquafina over other bottled water brands. Strongly Agree                             The following statements ask your opinion of the issue of recycling in general. Recycling is defined as a process of converting used materials into new products to prevent the waste of potentially useful resources. ―To me, the issue of recycling is…‖ Unimportant Of no concern Means nothing to me Does not matter to me Insignificant                                    Important Of concern to me Means a lot to me Does matter to me Significant Today, Facebook users tend to click the “Like” button on Facebook to join anything they like (e.g., Facebook pages for brands or celebrities). How often do you click the ―Like‖ button to join a Facebook page in general? Never    [Go to the next page] 141    Quite Often  Please follow each step of the directions carefully. 1. First, read the following definition of female casual friends. FEMALE CASUAL friends are ―female people with whom you enjoy spending time but do not know as well as and so not feel as close to as your ‗close personal friends.‘ You occasionally communicate with the female casual friends on Facebook and via email. Your friendship seems to be based upon the circumstances in which you find yourself (e.g., you work together, are in a class together, etc.). You plan on remaining friends as long as circumstances permit. Your friendship would dissolve if your circumstances were to change." 2. Now, think specifically one of your female casual friends and provide the first name of one of your female casual friends here. [Go to the next page] Now, image that you get an email from your female casual friend whom you just specified in the previous page. The following is the content of the email you get from your female friend. Please read the email content carefully and thoroughly. Figure 7 The Content of E-mail Sent from Female Friend who Invites Aquafina Facebook Page with CSR Campaigns From mycasualfriend@msu.edu To me@msu.edu Title Invitation Your female casual friend sent an e-mail invitation to you! Hi, Aquafina, a bottled water brand, supports „Go Recycling,‟ a national charitable cause to save our natural resources. Please join Aquafina‟s official Facebook page to support the cause campaign of “Go Recycling.” Your female casual friend Next, click ‗Next Page‘ to visit the Aquafina‘s official Facebook page [Go to the next page] 142 The following webpage is the official Facebook page of Aquafina. Please examine every element in the following webpage carefully and thoroughly. Figure 8 The Manipulation of Aquafina Facebook Page with CS Campaigns [Go to the next page] 143 Based on your review of the Aquafina Facebook page in the previous section, please indicate your behavioral intention regarding each statement. Feel free to go back to previous page by clicking “Previous Page” to review the content of Aquafina’s Facebook page again. 1. How likely are you to click the “Like” button to join the Aquafina Facebook page? Unlikely Improbable Impossible Uncertain                             Likely Probable Possible Certain 2. How likely are you to invite other people to the Aquafina Facebook page? Unlikely Improbable Impossible Uncertain                             Likely Probable Possible Certain 3. What do you think about your clicking of the “Like” button on the Aquafina Facebook page? ―Clicking the ―Like‖ button to join the Aquafina Facebook page will help me make ……… to my Facebook friends‖ Strongly Disagree         A good impression An altruistic impression A socially responsible impression An environmentally knowledgeable impression [Go to the next page] 144             Strongly Agree         You’re almost at the end. Please answer the following questions about the e-mail invitation and the Facebook page for a bottled water you just reviewed. 1. At the beginning stage of this study, you were asked to imagine getting an email from one of the following persons inviting you to the Facebook page for a bottled water brand (Select one).  Male casual friend  Female casual friend  Don‘t know/don‘t remember 2. At the beginning stage of this study, you were asked to visit the Aquafina Facebook page. Please click the number that best describe your opinion of this brand. (1) People drink Aquafina as a way of expressing their personality (2) Drinking Aquafina says something about the kind of person you are. (3) Aquafina is for people who are down-toearth. (4) Drinking Aquafina is practical. Strongly Disagree      Strongly Agree                        3. In the previous section, you were invited to the Aquafina’s Facebook page, featuring one of the following messages. Please select one that best describes the message (Select one).  "As every person clicks the ―Like‖ button to join our Facebook page, we will donate $1 to ―Go Recycling‖ up to $100,000 during the next two weeks.‖  ―We will donate $100,000 to ―Go Recycling‖ during the next two weeks.‖  "Water is our bodies‘ principal chemical component, making up 60% of our body weight, 85% of our brain, and 10% of our teach."  Don‘t know/don‘t remember 4-1. Do you think your clicking of the “Like” button to join the Aquafina Facebook page you just reviewed would help you make a good impression on your Facebook friends? (Select one).  Yes  No [Go to the next page] 145 4-2. If you answered as “Yes” in the previous question (4-1), please indicate why you think so by clicking the number that represents your opinion most. ―Because of my ____ support for the brand‘s cause campaign‖ Indirect Passive Secondhand                      Direct Active Firsthand [Go to the next page] 5. How do you perceive the relationship between the “Go Recycling” cause and the “Aquafina” brand? Not compatible Bad fit Irrelevant Bad match                 [Go to the next page] 146             Compatible Good fit Relevant Good match Finally, please answer the following questions about yourself. 1. What is your age? years old 2. What is your gender?  Female  Male 3. What is your nationality?  American  Other (please specify) 4. How long have you used Facebook? (If you are not a Facbook user, please enter 0). year(s) month(s) 5. Please type the name of your major at MSU (e.g., chemistry, communication). 6. What is your class standing? (Select one)     7. Freshman Sophomore Junior Senior Please type your name (for incentive). Last name First name 8. Please let us know your valid email address by which I can reach you (for incentive). E-Mail address ---- Thanks for your participation ---147 [Control Condition for the Symbolic Brand, Invited by a Male Friend] Please click the number that best represents your opinion on each statement. The following statements ask your opinion of a specific brand of bottled water – Fiji Water. Strongly Disagree (1) I like Fiji Water better than other bottled water brands. (2) I would use Fiji Water more than I would use other bottled water brands. (3) Fiji Water is my preferred brand over other bottled water brands. (4) I would be inclined to buy Fiji Water over other bottled water brands. Strongly Agree                             The following statements ask your opinion of the issue of recycling in general. Recycling is defined as a process of converting used materials into new products to prevent the waste of potentially useful resources. ―To me, the issue of recycling is…‖ Unimportant Of no concern Means nothing to me Does not matter to me Insignificant                                    Important Of concern to me Means a lot to me Does matter to me Significant Today, Facebook users tend to click the “Like” button on Facebook to join anything they like (e.g., Facebook pages for brands or celebrities). How often do you click the ―Like‖ button to join a Facebook page in general? Never    [Go to the next page] 148    Quite Often  Please follow each step of the directions carefully. 1. First, read the following definition of male casual friends. MALE CASUAL friends are ―male people with whom you enjoy spending time but do not know as well as and so not feel as close to as your ‗close personal friends.‘ You occasionally communicate with the male casual friends on Facebook and via email. Your friendship seems to be based upon the circumstances in which you find yourself (e.g., you work together, are in a class together, etc.). You plan on remaining friends as long as circumstances permit. Your friendship would dissolve if your circumstances were to change." 2. Now, think specifically one of your male casual friends and provide the first name of one of your male casual friends here. [Go to the next page] Now, image that you get an email from your male casual friend whom you just specified in the previous page. The following is the content of the email you get from your male friend. Please read the email content carefully and thoroughly. Figure 9 The Content of E-mail Sent from Male Friend who Invites Fiji Water Facebook Page without CSR Campaigns From mycasualfriend@msu.edu To me@msu.edu Title Invitation Your male casual friend sent an e-mail invitation to you! Hi, I like Fiji Water,a bottled water brand. Please join Fiji Water‟s official Facebook page. Your male casual friend Next, click ‗Next Page‘ to visit the Fiji Water‘s official Facebook page [Go to the next page] 149 The following webpage is the official Facebook page of Fiji Water. Please examine every element in the following webpage carefully and thoroughly. Figure 10 The Manipulation of Fiji Water Facebook Page without CSR Campaigns [Go to the next page] 150 Based on your review of the Fiji Water Facebook page in the previous section, please indicate your behavioral intention regarding each statement. Feel free to go back to previous page by clicking “Previous Page” to review the content of Fiji Water’s Facebook page again. 1. How likely are you to click the “Like” button to join the Fiji Water Facebook page? Unlikely Improbable Impossible Uncertain                             Likely Probable Possible Certain 2. How likely are you to invite other people to the Fiji Water Facebook page? Unlikely Improbable Impossible Uncertain                             Likely Probable Possible Certain 3. What do you think about your clicking of the “Like” button on the Fiji Water Facebook page? ―Clicking the ―Like‖ button to join the Fiji Water Facebook page will help me make ……… to my Facebook friends‖ Strongly Disagree         A good impression An altruistic impression A socially responsible impression An environmentally knowledgeable impression [Go to the next page] 151             Strongly Agree         You’re almost at the end. Please answer the following questions about the e-mail invitation and the Facebook page for a bottled water you just reviewed. 1. At the beginning stage of this study, you were asked to imagine getting an email from one of the following persons inviting you to the Facebook page for a bottled water brand (Select one).  Male casual friend  Female casual friend  Don‘t know/don‘t remember 2. At the beginning stage of this study, you were asked to visit the Fiji Water Facebook page. Please click the number that best describe your opinion of this brand. (1) People drink Fiji Water as a way of expressing their personality (2) Drinking Fiji Water says something about the kind of person you are. (3) Fiji Water is for people who are down-toearth. (4) Drinking Fiji Water is practical. Strongly Disagree      Strongly Agree                        3. In the previous section, you were invited to the Fiji Water’s Facebook page, featuring one of the following messages. Please select one that best describes the message (Select one).  "As every person clicks the ―Like‖ button to join our Facebook page, we will donate $1 to ―Go Recycling‖ up to $100,000 during the next two weeks.‖  ―We will donate $100,000 to ―Go Recycling‖ during the next two weeks.‖  "Water is our bodies‘ principal chemical component, making up 60% of our body weight, 85% of our brain, and 10% of our teach."  Don‘t know/don‘t remember 4-1. Do you think your clicking of the “Like” button to join the Fiji Water Facebook page you just reviewed would help you make a good impression on your Facebook friends? (Select one).  Yes  No [Go to the next page] 152 4-2. If you answered as “Yes” in the previous question (4-1), please indicate why you think so by clicking the number that represents your opinion most. ―Because of my ____ support for the brand‘s cause campaign‖ Indirect Passive Secondhand                      Direct Active Firsthand [Go to the next page] 5. How do you perceive the relationship between the “Go Recycling” cause and the “Fiji Water” brand? Not compatible Bad fit Irrelevant Bad match                 [Go to the next page] 153             Compatible Good fit Relevant Good match Finally, please answer the following questions about yourself. 1. What is your age? years old 2. What is your gender?  Female  Male 3. What is your nationality?  American  Other (please specify) : 4. How long have you used Facebook? (If you are not a Facbook user, please enter 0). year(s) month(s) 5. Please type the name of your major at MSU (e.g., chemistry, communication). 6. What is your class standing? (Select one)     Freshman Sophomore Junior Senior 7. Please type your name (for incentive). Last name First name 8. Please let us know your valid email address by which I can reach you (for incentive). E-Mail address ---- Thanks for your participation ---154 REFERENCES 155 REFERENCES Abrams, Karin von (2011), ―UK Multichannel Marketing New Paths to Consumer Engagement,‖ eMarketer Report. American College Health Association, ―American College Health Association: National College Health Assessment,‖ The Fall 2010 Reference Group Data Report. Anderson Analytics (2006, 2010), ―GenX Annual College Brand Survey,‖ An Anderson Analytic White Paper. Andrews, Jennifer, and Anita Bhappu (2010), "A Propositional Research Framework for the Conceptual and Technological Adoption of Digital Coupons in the US," Proceedings of the 43rd Hawaii International Conference on System Sciences Koloa, Kauai, Hawaii. Arbuthnot, Jack (1977), "The Roles of Attitudinal and Personality Variables in the Prediction of Behavior and Knowledge," Environmental and Behavior, 9(2), 217-232. Argo, Jennifer J., and Kelley J. Main (2008), "Stigma by Association in Coupon Redemption: Looking Cheap because of Others," Journal of Consumer Research, 35(4), 559-572. Ariely, Dan, Anat Bracha, and Stephan Meier (2009), "Doing Good or Doing Well? Image Motivation and Monetary Incentives in Behaving Prosocially," American Economic Review, 99(1), 544-555. Ashworth, Laurence, Peter R. Darke, and Mark Schaller (2005), "No One Wants to Look Cheap: Trade-Offs Between Social Disincentives and the Economic and Psychological Incentives to Redeem Coupons," Journal of Consumer Psychology, 15(4), 295-306. Bandura, Albert (1977), "Self-Efficacy: Toward a Unifying Theory of Behavioral Change," Psychological Review, 84(2), 191-215. _____________(1987), "Self-Regulation of Motivation and Action through Goal Systems," in Cognitive Perspectives on Emotion and Motivation. V. Hamilton, G. H. Bower, and N. H. Frijda, eds., Kluwer Academic Publishers, The Netherlands. Baron, Reuben M., and David A. Kenny (1986), "The Moderator–Mediator Variable Distinction in Social Psychological Research: Conceptual, Strategic, and Statistical Considerations," Journal of Personality and Social Psychology, 51(6), 1173–1182. Batson, C. Daniel, Bruce D. Duncan, Paula Ackerman, Terese Buckley, and Kimberly Birch (1981), "Is Empathic Emotion a Source of Altruistic Motivation?" Journal of Personality and Social Psychology, 40(2), 290-302. 156 Bearden, William O., Donald R. Lichtenstein, and Jesse Teel (1984), "Comparison Price, Coupon, and Brand Effects on Consumer Responses to Retail Newspaper Advertisements," Journal of Retailing, 60(2), 11-34. Beckwith, Sandra L. (2006), Publicity for Nonprofits: Generating Media Exposure that Leads to Awareness, Growth, and Contributions, Chicago, IL: Kaplan Publishing. Belk, Russell W. (1988), "Possessions and the Extended Self," Journal of Consumer Research, 15(2), 139-168. Bereczkei, Tamas, Bela Birkas, and Zsusanna Kerekes (2010), "The Presence of Others, Prosocial Triats, Machiavellianism," Social Psychology, 41(4), 238-245. Berkowitz, David (2010, March 30), "Facebook Gives ‗Like‘ More Love, ‗Fans‘ the Boot: Social Website‘s Shift in Language, User Actions Could Work to Marketers‘," Advertising Age. Bhagat, Vinay, Pam Loeb, and Mark Rovner (2010), "The Next Generation of American Giving: A Study on the Contrasting Charitable Habits of Generation Y, Generation X, Baby Boomers and Matures," White Paper by Convio, Edge Research, and Sea Change Strategies. Bhat, Subodh and Srinivas K. Reddy (1998), "Symbolic and Functional Positioning of Brands," Journal of Consumer Marketing, 15(1), 32-43. Bhattacharya, C. B., and Sanker Sen (2004), "Doing Better at Doing Good: When, Why, and How Consumers Respond to Corporate Social Campaigns," California Management Review, 47(1), 9-24. Box, G. E. P. (1954), "Some Theorems on Quadratic Forms Applied in the Study of Analysis of Variance Problems," The Annals of Mathematical Statistics, 25(2), 290-302. Bushman, Brad J. (1993), "What's in a Name? The Moderating Role of Public SelfConsciousness on the Relation between Brand Label and Brand Preference," Journal of Applied Psychology, 78(5), 857-857. Boyd, Danah M., and Nicole B. Ellison (2007), "Social Network Sites: Definition, History, and Scholarship," Journal of Computer Mediated Communication, 13(1), available at http://jcmc.indiana.edu/vol13/issue1/boyd.ellison.html (accessed September 15, 2009). Brewer, Marilynn B. (2000), "Research Design and Issues of Validity," In Reis, H. and Judd, C. eds.. Handbook of Research Methods in Social and Personality Psychology. Cambridge: Campridge University Press. Carter, Seth. E., and Lawrence J. Sanna (2008), "It's Not Just What You Say But When You Say It: Self-Presentation and Temporal Construal," Journal of Experimental Social Psychology. 44(5), 1139-1145. 157 Chang, Chun-Tuan (2008), "To Donate or Not to Donate? Product Characteristicis and Framing Effects of Cause-Related Marketing on Consumer Purchase Behavior," Psychology and Marketing, 25(12), 1089-1110. Cialdini, Robert R. (1989), "Indirect Tactics of Image Management: Beyond Basking," in Impression Management in the Organization, R.A. Giacalone and P. Rosenfield, eds., NJ, Hillsdale: Erlbaum. _______________, and Ricahrd J. Borden (1980), "Two Indirect Tactics of Image Management: Basking and Blasing," Journal of Personality and Social Psychology, 39(3), 406-415. _______________., ______________, Avril Thorne, Marcus Randall Walker, Stephen Freeman, and Lloyd Reynolds Sloan (1976), "Basking in Reflected Glory: Three (Football) Field Studies," Journal of Personality and Social Psychology, 34(3), 463-476. Cone, Inc. (2008), Cone Cause Evaluation Survey, available at http://www.coneinc.com/content1188 (accessed January, 30, 2010). Cornwell, T. Bettina and Isabelle Maignan (1998), "An International Review of Sponsorship Research," Journal of Advertising, 27(1), 1-21. ________________, and Leonard V. Coote (2005), "Corporate Sponsorship of a Cause: The Role of Identification in Purchase Intent," Journal of Business Research, 58(3), 268-276. Creyer, Elizabeth, and William Ross (1977), "The Impact of Corporate behavior on Perceived product Value," Marketing Letters, 7(2), 173-185. Crowne, Douglas P., and David Marlowe (1960), "A New Scale of Social Desirability Independent of Psychopathology," Journal of Consulting Psychology, 24(4), 349-354. Cui, Yanli, Elizabeth S. Trent, Pauline M. Sullivan, and Grace N. Matiru (2003), "Cause-related marketing: how generation Y responds," International Journal of Retail & Distribution Management, 31(6), 310-320 Dean, Dwane Hal (2003/2004), "Consumer Perception of Corporate Donations: Effects of Compnay Reputation for Social Responsibility and Type of Donation," Journal of Advertising, 32(4), 91-102. Deci, Edward L., Richard Koestner, and Richard M. Ryan (1999), "A Meta-Analytic Review of Experiments Examining the Effects of Extrinsic Rewards on Intrinsic Motivation," Psychological Bulletin, 125(6), 627-668. Delevingne, Lawrence (January 20, 2009), Surprising Survivors: Corporate Do-Gooders. CNN. Dolich, Ira J. (1969), "Congruence Relationships between Self Images and Product Brands," Journal of Marketing Research, 6(1), 80-84. 158 Dote, Lillian, Kevin Cramer, Nathan Dietz, and Robert Grimm (2006), "College Students Helping America," A Report of Corporation for National and Community Service. Drumwright, Minette, and Patrick E. Murphy (2011), "Corporate Social Markeing," in Handbooks of Marketing and Society, Paul N. Bloom and Gregory T. Gundlach, eds. Thousand Oaks, CA: Sage Publications. Edelman (2009), "Consumer New Media Study," Cone's Corporate Resposonsibility Report. Eisenberg, Nancy, and Paul A. Miller (1987), "The Relation of Empathy to Prosocial and Related Behaviors," Psychological Bulletin, 101(1), 91-119. Endacott, Roy William John (2004), "Consumers and CRM: A National and Global Perspective," Journal of Consumer Marketing, 21(3), 183-189. Erdogan, B. Zafer, and Philip J. Kitchen (1998). Managerial Mindsets and the Symbiotic Relationship between Sponsorship and Advertising. Marketing Intelligence & Planning, 16(6), 369 - 374. Esquivel, Eliza, Brian McMahon, and Susan Raymond (2010), "Social Activism 2.0: Young Adults & The Revival of The "Social" in Social Responsibility," White Paper of Advertising Age . ExactTarget (2010), "Facebook X-Factors," Subscribers, Fans, & Followers Research Series #5 of Exact Target Report. Facebook Like Button Developers (2010), availbale at http://developers.facebook.com/docs/reference/plugins/like (accessed January 3, 2010). Ferris, Gerald R., and Timothy A. Judge (1991), "Personnel/Human Resources Management: A Political Influence Perspective," Journal of Management, 17(2), 447-488. Fishbein, Martin (1967), Readings in Attitude Theory and Measurement. New York: Wiley. Ford, Gary T., Darlene B. Smith and John L. Swasy (1990), "Consumer Skepticism of Advertising Claims: Testing Hypotheses from Economics of Information," Journal of Consumer Research, 16(March), 433-441. Fredricksen, Clark (2011), "Traditional Media: Dollars and Attention Shift to Digital," eMarketer Report. Friestad, Marian, and Peter Wright (1994), "The Persuasion Knowledge Model: How People Cope with Persuasion Attempt," Journal of Consumer Research, 21(June), 1-31. _____________, and __________ (1995), " Lay People's and Researchers' Belief about the Psychology of Advertising," Journal of Consumer Research, 22(June), 62-74. 159 Ghuneim, Mark (2008), "Terms and Engagement: Measuring the Active Consumer," Wiredset Report. Gibbs, Jennifer L., Nicole B. Ellison, and Rebecca D. Heino (2006), "Self-Presentation in Online Personals : The Role of Anticipated Future Interaction, Self-Disclosure, and Perceived Success in Internet Dating," Communnication Research, 33(2), 152-177. Gill, Jeff (2001). Generalized Linear Models: A Unified Approach, Quantitative Applications in the Social Sciences: A Mathematical Overview of GLM., Thousand Oaks, CA: Sage Publications. Gilovich, Thomas (1981), "Seeing the Past in the Present," Journal of Personality and Social Psychology, 40(5), 797-808. Glass, Carol R., John M. Gottman, and Steven H. Shmurak (1976), "Response-Acquisition and Cognitive Self-Statement Modification Approaches to Dating-Skills Training," Journal of Counseling Psychology, 23(6), 520-526. Goffman, Erving (1959), The Presentation of Self in Everyday Life. New York: Doubleday. Graeff, Timothy R. (1996), "Image Congruence Effects on Product Evaluations: The Role of Self-Monitoring and Public/Private Consumption," Psychology and Marketing, 13(5), 481499. Greenberg, Jeff, Tom Pyszczynski, and Penny Stine (1985), "Social Anxiety and Anticipation of Future Interaction as Determinants of the Favorability of Self-Presentation," Journal of Research in Personality, 19(1), 1-11. Guadagno, Rosanna E., and Cialdini, Robert B. (2007), "Gender Differences in Impression Management in Organizations: A Qualitative Review," Sex Roles, 56(7-8), 483-494. Gupta, Shruti and Julie Pirsch (2006), "The Company-Cause-Customer Fit Decision in CauseRelated Marketing," Journal of Consumer Marketing, 23(6), 314-326. Gwinner, Kevin P., and John Eaton (1999), "Building Brand Image through Event Sponsorship: The Role of Image Transfer. Journal of Advertising, 28(4), 47-57. Hamlin, Robert P., and Toni Wilson (2004), ―The Impact of Cause Branding on Consumer Reactions to Products: Does Product/Cause ‗Fit‘ Really Matter?‖ Journal of Marketing Management, 20(7/8), 663–681. Harvey, Bill (2001), "Measuring the Effects of Sponsorship," Journal of Advertising Research, 41(January/February), 59-65. 160 Holladay, Sherry J., and Kathleen S. Kerns (1999), "Do Age Difference Matter in Close and Casual Friendship? A Comparison of Age Discrepant and Age Peer Friendship," Communication Report, 12(2), 101-114. Interbrand (2010). Best Global Brands: 2010 Ranking, available at http://www.interbrand.com/en/best-global-brands/best-global-brands-2008/best-globalbrands-2010.aspx (accessed at March 2011). Jones, Edward E. (1990). Interpersonal Perception. New York: W. H. Freeman and Company. _____________, and Thane S. Pittman (1982), "Toward a General Theory of Strategic SelfPresentation," in Psychological Perspectives on the Self, J. Suls, ed., Hillsdale, NJ: Erlbaum. Judge, Timothy A., and Robert D. Bretz (1994), "Political Influence Behavior and Career Success," Journal of Management, 20(1), 43-65. Hopper, Joseph. R. (1991), "Rcycling as Altrustic Behavior: Normative and Behavioral Strategies to Expand Participation in a Community Recycling Program," Environment and Behavior, 23(2), 195-220. Kacmar, K. Michael, John E. Delery, and Gerald R. Ferris (1990), "The Effectiveness of the Use of Impression Management Tactics by Applicants on Employment Interview Outcomes," Paper presented at the Annual Meeting of the Southern Management Association, Orlando. Kamins, Michael A., and Kamal Gupta (1994), "Congruece between Spokesperson and Product Type: A Matchup Hypothesis Perspective," Psychology and Marketing, 11(6), 569-589. Kaplan, Andreas M., and Michael Haenlein (2010), "Users of the World, Unite! The Challenges and Opportunities of Social Media," Business Horizons, 53(1), 59–68. Kerwin, Ann Marie (2010, June 28), "How to Get the Social-Media Generation Behind Your Cause: For Young Adults, Activism Can Be Hitting 'Like' on Facebook, but Brands Can Use This to Their Advantage," Advertising Age. Kim, Jooyoung, and Hye-Jin Paek (2009), "Information Processing of Genetically Modified Food Messages Under Different Motives: An Adoption of the Multi-Motive HeurtisicSystematic Model," Risk Analysis, 29(12), 1793-1806. Klein, H. M, & Willerman, L. (1979), "Psychological Masculinity and Femininity and Typical and Maximal Dominance Expression in Women," Journal of Personality and Social Psychology, 37, 2059-2070. Kotler, Philip and Nancy Lee (2004), Corporate Social Responsibility: Doing the Most Good for Your Company and Your Cause. Hoboken, NJ: Wiley & Sons, Inc. 161 KPMG (1999), "Interntaional Annual Review 1999," KPMG Annual Review Report. ______(2002), "Interntaional Annual Review 2002," KPMG Annual Review Report. ______(2003), "Interntaional Annual Review 2003," KPMG Annual Review Report. ______(2005), "Interntaional Annual Review 2005," KPMG Annual Review Report. Krämer, Nicole, and Stephan Winter (2008), "Impression Management 2.0: The Relationship of Self-Esteem, Extraversion, Self-Efficacy, and Self-Presentation within Social Networking Sites," Journal of Media Psychology: Theories, Methods, and Applications, 20(3), 106-116. Lacetera, Nicole, and Mario Macis (2008), Social Image Concerns and Pro-Social Behavior. IZA Discussion Paper No. 48109, Institute for the Study of Labor (IZA). Lafferty, Barbara A., and Ronald E. Goldsmith (2005), "Cause-Brand Alliances: Does the Cause Help the Brand or Does the Brand Help the Cause?" Journal of Business Research, 58(4), 423-429. Lampe, Cliff, Nicole Ellison, and Charles Steinfield (2007), "A Familiar Face (book): Profile Elements as Signals in an Online Social Network," Paper presented at conference on Human Factors in Computing System, San Jose, CA. LaRose, Robert, and Matthew S. Eastin (2002), "Is On-Line Buying out of Control? Electronic Commerce and Consumer Self- Regulation," Journal of Broadcasting and Electronic Media, 45(4), 549-564. Leary, Mark R.(1983), Understanding Social Anxiety: Social, Personality, and Clinical Perspectives. Berverly Hills, CA: Sage. ____________, John B. Nezlek, Deborah Downs, Julie Radford-Davenport, Jeffrey Martin, and Anne McMullen (1994), "Self-Presentation in Everyday Interactions: Effects of Target Familiarity and Gender Composition," Journal of Personality and Social Psychology, 67(4), 664-673. ____________., Lydia R. Tchividjian, and Brook Kraxberger (1994), "Self-Presentation Can Be Harzardous to Your Health: Impression Management and Health Risk," Health Psychology, 13(6), 461-470. ____________ and Robin M. Kowalski (1990), "Impression Management: A Literature Review and Two-Component Model," Psychological Bulletin, 107(1), 34-47. ______________., ______________, and Cynthia D. Campbell (1988), "Self-Presentational Concerns and Social Anxiety: The Role of Generalized Impression Expectancies," Journal of Research in Personality, 22(3), 308-321. 162 Linchtenstein, Donald R., Minnett Drumwright, and Bridgette M. Braig (2004), "The Effect of Corporate Social Responsibility on Consumer Donation to Corporate-Supportied Nonprofits," Journal of Marketing, 68(October), 16-32. Mangold, W.Glynn, and David J. Faulds (2009), "Social Media: The New Hybrid Element of the Promotion Mix," Business Horizons, 52(4), 357–365. Matthew, Markus (2004), "Yoplait's Got it Right," g-Think report. Maul, Kimberly (2011), "Facebook Marketing Strategies for Turning "Likes" into Loyalty," eMarketer Report. McDonald, Colin (1991), "Sponsorship and Image of the Sponsoz," European Journal of Marketing, 25(11), 31-28. McKinnon, David P., Jennifer L. Krull, and Chondra M. Lockwood (2000), "Equivalence of the Mediation, Confounding, and Suppression Effect," Prevention Science, 1(4), 173-181. Meenaghan, Tony (1991), "The Role of Sponsorship in the Marketing Communications Mix," International Journal of Advertising 10 (1), 35–47. Mittal, Banwari (1995), "A Comparative Analysis of Four Scales of Consumer Involvement," Psychology and Marketing, 12 (7), 663-682. Moore, D. S. (1995). The Basic Practice of Statistics. NY: Freeman and Co. Morrissey, Brian (December 3, 2008), "Kraft Gives Facebook Users Reason to Share: Kraft Donates Six Meals to Hungry Families for Each Friend a User Convinces to Add the Application," Adweek. ______________ (May 18, 2009), "Cause Marketing Meets Social Media: Brands Tap DoGooder Impulse to Encourage Consumers to Pass Along Marketing Messages," Adweek. Muller, Dominique, Charles M. Judd, and Vincent Y. Yzerbyt (2005), "When Moderation is Mediated and Mediation is Moderated," Journal of Personality and Social Psychology, 89(6), 852-863. Murphy, Claire (March 25, 1999), "Brand Values Can Build on Charity Ties," Marketing, 41-42. Nan, Xan, and Kwangjun Heo (2007), "Consumer Responses to Corporate Social Responsibility (CSR) Campaigns: Examining the Role of Brand-Cause Fit in Cause-Related Marketing," Journal of Advertising, 36(2), 63-74. Nelder, John, and Robert Wedderburn (1972), "Generalized Liner Models," Journal of the Royal Statistical Society, 135 (3), 370-384. 163 Nielson (June 15, 2010), "Social Networks/Blogs Now Account for One in Every Four and a Half Minutes Online," Nielson Report. Obermiller, Carl, and Eric R. Spangenberg (2000), "On the Origin and Distinctiness of Skepticism and Advertising," Marketing Letters, 11(4), 311-322. O'Brien, Ed, Courtney Hsing, and Sara H. Konrath (2010), "Changes in Dispositional Empathy in American College Students over Time: A Meta-Analysis," Paper presented in the annual convention of Association for Psychological Science, Boston, MA. Olson, Chester L. (1976), "On Choosing a Test Statistic in Multivariate Analysis of Variance," Psychological Bulletin, 83(4), 579-586. Paek, Hye-Jin, Choi Hojoon, Michelle R. Nelson (2010), "Product, Personality or Prose? Testing Functional Matching Effects in Advertising Persuasion," Journal of Current Issues & Research in Advertising, 32(2), 11-26. Pandey, Janak (1986), "Sociocultural Perspectives on Ingratiation," in Progress in Experimental Personality, Research, in B. Maher ed., 205-229. San Diego, CA: Academic Press Pandey, Janak, and Renu Rastogi (1979), "Machiavellianism and Ingratiation," Journal of Social Psychology, 108(2), 221-225. Park, C. Whan, Bernard Jaworski, and Deborah J. Maclnnis (1986), "Strategic Brand ConceptImage Management," Journal of Marketing, 50(4), 135-145. Polonsky, Michael Jay and Richard Speed (2001), "Linking Sponsoring and Cause Related Marketing: Complentarities and Conflict," European Journal of Marketing, 35(11-12), 1361-1385. Pope, Nigel Li, and Keven E. Voges (2000), "The Impact of Sport Sponsorship Activities, Corporate Image, and Prior Use on Consumer Purchase Intention," Sport Marketing Quarterly, 9(2), 96-102. Pratkanis, Anthony R. (2007), "Social Influence Analysis: An Index of Tactics," In Anthony R. Pratkanis (Ed.), The Science of Social Influence: Advances and Future Progress. Philadelphia: Psychology Press. Preacher, Kristiopher J., and Andrew F. Hayes (2008), "Asymptopic and Resampling Strategies for Assessing and Comparing Indirect Effects in Multiple Mediator Models," Behavior Research Methods, 40(3), 879-891. ________________, Derek D. Rucker, and ______________ (2007), "Assessing Moderated Mediation Hypotheses: Theory, Methods, and Prescriptions," Multivariate Behavioral Researcg, 42(1), 185-227. 164 Pracejus, John W. and Duglas Olsen, and Norman R. Brown (2003-2004), "On the Prevelence and Impact of Vague Quantifiers in the Advertising of Cause-Related Marketing (CRM)," Journal of Advertising, 32(4), 19-28. PR Wee and Barkley (2010), "Cause Survey 2010: The Male Perspective," PRWeek/Barkley PR Cause Survey. Richardson, K. D., and Robert B. Cialdini (1981), "Basking and Blasting: Tactics of Indirect Self-Presentation," in Impression Management Theory and Social Psychology Research, J. T. Tedeschi, ed., N.Y: Academic Press Inc. Rifon, Nora, Sejung M. Choi, Carrie S. Trimble, and Hairong Li (2004), "Congruence Effects in Sponsorship: The Mediating Roles of Sponsor Credibility and Consumer Attributions of Sponsor Motive," Journal of Advertising, 33(1), 30-42. Rind, Bruce, and Daniel Benjamin (1994), "Effects of Public Image Concerns and Self-Image on Compliance," Journal of Social Psychology, 134(1), 19-19. Robin, Donald P., R Eric Reidenbach, and P. J. Forrest (1996), "The Perceived Importance of an Ethical Issue as an Infleunce on the Ethical Decision-Making of Ad Managers," Journal of Business Research, 35(1), 17-28. Rosenfield, Paul., Rober A., Giacalone and Catherine A. Riordan (1995). Impression Management in Organizations. London: Routledge. Ross, John K., Larry T. Patterson, and Mary Ann Stutts (1992), "Consumer Perceptions of Organizations that Use Cause-Related Marketing," Journal of the Academy of Marketing Science, 20(1), 93-97. Rotter, Julian B. (1954). Social Learning and Clinical Psychology. Englewood Cliffs, NJ: Prentice-Hall. Schlenker, Barry R. (1980). Impression Management: The Self-Concept, Social Identity, and Interpersonal Relations. Monterey, CA: Brooks/Cole. ________________, Donelson R. Forsyth, Mark R. Leary,and Rowland D. Miller (1980), "A Self-Presentational Analysis of the Effects of Incentives on Attitude Change Following Counterattitudinal Behavior," Journal of Personality and Social Psychology, 39(4), 553577. Schlenker, Barry R., and Mark R. Leary (1982),"Social Anxiety and Self-Presentation: A Conceptualization Model. Psychological Bulletin," 92(3), 641-669. Schneider, David J. (1981), "Tactical Self-Presentations: Toward a Broader Conception," in Impression Management Theory and Social Psychology Research, J. T. Tedeschi, ed., N.Y.: Academic Press Inc. 165 Seitanidi, Maria May, and Annmarie Ryan (2007), "A Critical Review of Forms of Corporate Community Involvement: From Philanthropy to Partnerships," International Journal of Nonprofit and Voluntary Sector Marketing, 12(3), 247-266. Seiter, John S. (2007), "Ingratiation and Gratuity: The Effect of Complimenting Customers on Tipping Behavior in Restaurent," Journal of Applied Social Psychology, 37(3), 478-485. Shimp, Terry A. (1981), "Attitude Toward the Ad as a Mediator of Consumer Brand Choice," Journal of Advertising, 10(2), 9-15. Simmons, Carolyn J. and Karen L. Becker-Olsen (2006), "Achieving Marketing Objectives Through Social Responship," Journal of Marketing, 70(October), 154-169. Sirgy, M. Joseph, Dhruv Grewal, Tamara F. Mangleburg, Jae-Ok Park, Kye-Sung Chon, C.B. Claiborne, C. B., J. S. Johar, and Harold Berkman (1997), "Assessing the Predictive Validity of Two Methods of Measuring Self-Image Congruence," Journal of the Academy of Marketing Science, 25(3), 229-241. ______________ and Jeffrey E. Danes (1982), "Self-Image/Product-Image Congruence Models: Testing Selected Models," in Advances in Consumer Research , 9, Andrew Mitchell, ed., 556-561, Association of Consumer Research. Ann Abor: MI. Slutsky, Irina (January 25, 2011), "Facebook Turns the 'Like' Into Its Newest Ad: Say Something About Brands? It Could End Up as an Ad in Your Friends' Feeds," Advertising Age. Smith, Craig, N. (2003), "Corporate Social Responsibility: Whether or How?" California Management Review, 45(4), 52-76. Smith, Scott M., and Davis S Alcorn (1991), "Cause Marketing: A New Direction in the Marketing of Social Responsibility," Journal of Consumer Marketing, 8(3), 19-34. Snell, William E., Jr. (1989), "Willingness to Self-Disclose to Female and Male Friends as a Function of Social Anxiety and Gender," Personality and Social Psychology Bulletin, 15(1), 113-125. Sweeney, Jullian C., and Geoffrey N. Soutar (2001), "Consumer Perceived Value: The Development of a Multiple Item Scale," Journal of Retailing, 77(2), 203-220. Syncapse (June, 2010), "A Value of a Facebook Fan: An Empirical Review," A Report of Syncapse. Szykman, Lisa, R., Paul N. Bloom, and Jennifer Blazing (2004), "Does Corporate Sponsorship of a Socially-Oriented Message Make a Difference? An Investigation of the Effects of Sponsorship Identity on Responses to an Anti-Drinking and Driving Message," Journal of Consumer Psychology, 14(1/2), 13-20. 166 Strahilevitz, Michal, and John G. Myers (1998), ―Donations to Charity as Purchase Incentives: How Well They Work May Depend on What You Are Trying to Sell,‖ Journal of Consumer Research, 24(March), 434–446 Tal-Or, Nuri (2010), "Direct and Indirect Self-Promotion in the Eyes of the Perceivers," Social Influence, 5(2), 87-100. Tangari, Andrea Heintz, Judith Anne Garretson Folse, Scott Burton, and Jeremy Kees (2010), " The Moderating Influence of Consumers' Temporal Orientation on the Framing of Societal Needs and Corporate Responses in Cause-Related Marketing Campaigns," Journal of Advertising, 39(2), 35-50. Tedeschi, James T. (ed.) (1981), Impression Management Theory and Social Psychological Research. New York: Academic Press. Facebook (2011), Facebook Statistics, available http://www.facebook.com/press/info.php?factsheet#!/press/info.php?statistics (accessed January 03, 2011). Quantcast (2010), Facebook.com: Demographics, available at http://www.quantcast.com/facebook.com#demographics (accessed October 15, 2010). Tech Crunchies (2008), Top 10 Websites for US College Students, available at http://techcrunchies.com/top-10-websites-for-us-college-students (accessed September 15, 2009). Till, Brian D., and Michael Busler (2000), "The Match-Up Hypothesis: Physical Attraction, Expertise, and the Role of Fit on Brand Attitude, Purchase Intent, and Brand Beliefs," Journal of Advertising, 29 (3), 1-13. Tobin, Jim (June, 2010), Ignite Social Media Blog: Top 50 Branded Facebook Pages, available at http://www.ignitesocialmedia.com/facebook-marketing/top-50-branded-facebook-pages/ (accessed November 01, 2010). Varadarajan, P. Rajan, and Anil Menon (1988), "Cause-Related Marketing: A Coalignment of Marketing Strategy and Corporate Philanthropy," Journal of Marketing, 52(3), 58–74. Vitrue (2010). The Virture 100: Top Social Brands of 2010, available at http://vitrue.com/blog/2011/01/07/the-vitrue-100-consumer-electronics-reigned-supremein-2010/ (accessed at March 10, 2011). Vroom, Victor Harold (1964). Work and Motivation, New York: Wiley Wall, Wendy L. (June 21, 1984), "Companies Change the Ways They Make Charitable Donations," Wall Street Journal. 167 Walther, Joseph B. (1996), "Computer-Mediated Communication: Impersonal, Interpersonal, and Hyperpersonal Interaction," Communication Research, 23(1), 3-43. Wayne, Sandy J., and Gerald Ferris (1990), "Influence Tactics, Affect, and Exchange Quality in Supervisor-Subordinate Interactions: A Laboratory Experiment and Field Study," Journal of Applied Psychology, 75(5), 487-499. Webb, Deborah, and lois Mohr (1998), "A Typology of Consumer Responses to Cause-Related Marketing: From Skeptics to Socially Concerned," Journal of Public Policy and Marketing, 17(2), 226-238. Webb, Thomas L., and Paschal Sheeran (2006), "Does Changing Behavioral Intentions Engender Behavior Change? A Meta-Analysis of the Experimental Evidence," Psychological Bulletin, 132(2), 249-268. Wheeler, Ladd, and John Nezlek (1977), "Sex Differences in Social Participation," Journal of Personality and Social Psychology, 35(10), 742-754. White, Katherine, and Darren W. Dahl (2006), "To Be or Not Be? The Influence of Dissociative Reference Groups on Consumer Preferences," Journal of Consumer Psychology, 16(4), 404-141. White, Katherine, and John Peloza (2009), "Self-Benefit Versus Other-Benefit Marketing Appeals: Their Effectiveness in Generating Charitable Support," Journal of Marketing, 73(4), 109-124. Williamson, Debra Aho (2009), "Marketing on Social Networks: Branding, Buying and Beyond," eMarketer Report. ___________________ (2010), "Brand Interactons on Social Networks," eMarketer Report. ___________________ (2011), "eMarketer Webinar: Social Media Outlook for 2011," eMarketer Report. Yoon, Seounmi and Hyuksoo Kim (2008), "Antecedents of Consumer Attitudes toward CauseRelated Marketing," Journal of Advertising Research, 48 (March), 123-137. Zadrakovic, Srdan, Peter Magnusson and Sarah M. Stanley (2010), "Dimensions of Fit between a Brand and a Social Cause and Their Infleunce on Attitudes," International Journal of Research in Marketing, 27(2), 151-160. Zaichkowsky, Judith L. (1994), "The Personal Involvement Inventory: Reduction, Revision, and Application to Advertising," Journal of Advertising, 23(4), 59-70. 168 Zhao, Xinshu, John G. Lynch Jr, and Qimei Chen (2010), "Reconsideting Baron and Kenny: Myths and Trusts about Mediation Analysis," Journal of Consumer Research, 37(2), 197206. Zmuda, Natalie, and Emily Bryson York (March 10, 2010), "Cause Effect: Brand Rush to Save World One Deed at a Time: From Bisquick to Quacker, Pepsi to Prilosec, Marketers Turn to Microsponsorships," Advertising Age. 169