You are on page 1of 19

This article was published in an Elsevier journal.

The attached copy is furnished to the author for non-commercial research and education use, including for instruction at the authors institution, sharing with colleagues and providing to institution administration. Other uses, including reproduction and distribution, or selling or licensing copies, or posting to personal, institutional or third party websites are prohibited. In most cases authors are permitted to post their version of the article (e.g. in Word or Tex form) to their personal website or institutional repository. Authors requiring further information regarding Elseviers archiving and manuscript policies are encouraged to visit: http://www.elsevier.com/copyright

Author's personal copy

Intern. J. of Research in Marketing 24 (2007) 224 241 www.elsevier.com/locate/ijresmar

Reaping relational rewards from corporate social responsibility: The role of competitive positioning
Shuili Du a,1 , C.B. Bhattacharya b,1 , Sankar Sen c,,1
c

School of Management, Simmons College, 300 The Fenway, Boston, MA 02115, United States School of Management, Boston University, 595 Commonwealth Avenue, Boston, MA 02215, United States Department of Marketing & International Business, Zicklin School of Business, Baruch College/CUNY Box B12-240, One Bernard Baruch Way New York, NY 10010, United States
b

Abstract This research examines the moderating influence of the extent to which a brand's social initiatives are integrated into its competitive positioning (i.e., a CSR positioning) on consumer reactions to CSR. We find that positive CSR beliefs held by consumers are associated not only with greater purchase likelihood but also with longer-term loyalty and advocacy behaviors. More importantly, we find that not all CSR initiatives are created equal: a brand that positions itself on CSR, integrating its CSR strategy with its core business strategy, is more likely than brands that merely engage in CSR to reap a range of CSR-specific benefits in the consumer domain. 2007 Elsevier B.V. All rights reserved.
Keywords: Corporate social responsibility; Competitive context; Product positioning; Loyalty; Advocacy

Corporate social responsibility (CSR) is at the forefront of corporate consciousness today. As the CSR debate shifts from whether to how, companies are embracing socially responsible ventures with unprecedented gusto (e.g., KPMG International Survey of Corporate Responsibility Reporting, 2005). Moral imperatives aside, this increasing commitment to CSR is spurred, at least in part, by the growing sense that consumers, a key stakeholder group, reward good corporate citizens through greater, more sustained patronage (Cone Corporate Citizenship Study, 2004; Lichtenstein, Drumwright, & Braig, 2004; Luo & Bhattacharya, 2006). Indeed, both marketplace polls (Cone Corporate Citizenship Study, 2004) and a growing body of primarily experimental studies (e.g., Brown & Dacin, 1997; Klein & Dawar, 2004; Sen & Bhattacharya, 2001; Sen, Bhattacharya, & Korschun, 2006) have demonstrated that, assuming all else is constant, consumers are more likely to purchase from companies that engage

Corresponding author. Tel.: +1 646 312 3302; fax: +1 646 312 3271. E-mail address: sankar_sen@baruch.cuny.edu (S. Sen). 1 All three authors contributed equally to this paper. 0167-8116/$ - see front matter 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.ijresmar.2007.01.001

in CSR actions, particularly in domains that consumers deem appropriate (e.g., a high degree of fit between the company and the CSR cause) and personally relevant. Importantly, however, little is constant in the real marketplace. For instance, consumers' CSR-related reactions to a company/brand are determined not only by its actions in this domain, but also by those of its stakeholder groups (e.g., activists, the media), which are typically beyond the company's control (Bhattacharya & Sen, 2003; Brown, Dacin, Pratt, & Whetten, 2006). Thus, there is a need for CSR research to move beyond the often rarefied, controlled empirical contexts to paint a more externally valid picture of the forces determining consumer reactions to CSR initiatives. One such force is the competitive context in which a company's CSR strategy is viewed by consumers. Given the widespread prevalence of CSR in today's marketplace, effective CSR strategies need to take into account the competitive context in which a particular set of CSR actions are likely to be enacted. In other words, much as the competitive context impacts the marketing mix, a company, in formulating its CSR strategy, needs to understand how consumers perceive and react to its CSR actions not in isolation but in the context of different CSR actions, if any, taken by its competitors (Bhattacharya & Sen, 2004).

Author's personal copy

S. Du et al. / Intern. J. of Research in Marketing 24 (2007) 224241

225

A key element of the competitive context is the relative positioning of each brand along the CSR dimension. Brands vary in the extent to which they rely on their CSR activities to position themselves, relative to their competitors, in the minds of consumers. More specifically, while many brands affiliate themselves with causes, some, such as Stonyfield Farm, Body Shop, and Ben and Jerry's, go beyond just engaging in CSR to position themselves wholly in terms of CSR, becoming known as the socially responsible brand (i.e., the CSR brand) in a category. For example, in the US supermarket category, the Whole Foods Market, positioned on CSR, espouses the core value of caring about our communities and our environment. Moreover, this value pervades virtually every aspect of its business, from organic and sustainable sourcing to environmentally-sensitive retailing, from devoting at least 5% of its annual profits to a variety of causes to encouraging community service among its employees on company time. Such a positioning has contributed to the company's extraordinary success, creating consumers who act as ambassadors for the company (Springer, 2006). The CSR brand in the footwear category, Timberland, has integrated a similar set of CSR principles into every aspect of its business (e.g., environmental consciousness, fair and humane labor practices). Since 2005, every Timberland product bears a nutritional label, informing consumers of its environmental and community impact. Such an integrated CSR approach has resulted in, not only the company's number six ranking on Business Ethics' 100 Best Corporate Citizens in 2006, but also the unwavering loyalty of its consumers (Gillentine, 2006). Notably, the need to understand the unexamined role of CSR positioning in consumer reactions to a company's CSR actions is underscored by the unique nature of such a positioning strategy compared to those along more conventional dimensions such as operational excellence, product innovativeness, or customer intimacy. This is because it is a company's actions in the CSR domain, as opposed to other, more product-related ones, that truly reveal its values (Turban & Greening, 1997), soul (Chappell, 1993), or character (Brown & Dacin, 1997), comprising the company's identity (Bhattacharya & Sen, 2003). Such a CSR-based identity is not only fundamental and relatively enduring, but also more distinctive than identities based on innovativeness and engineering expertise by virtue of its idiosyncratic bases (e.g., environmental stewardship, sponsorship of social causes, fair labor policies). This, in turn, renders such an identity not only more memorable but also more anthropomorphic, enabling consumers to identify with it more readily than with others based on more conventional positioning strategies. In other words, unlike other positioning strategies, CSR positioning humanizes a company or brand, encouraging consumers to not just like, respect or admire the company but actually identify with it (Bhattacharya & Sen, 2003). In turn, the benefits of such identification to the brand are strong, numerous and enduring (Bergami & Bagozzi, 2000; Lichtenstein et al., 2004; Sen & Bhattacharya, 2001). Specifically, these go beyond the transactional benefits to the company (i.e., sales) of any consistent, coherent positioning to the rarer, longer-term relational benefits such as loyalty and advocacy (e.g., positive

word-of-mouth, resilience to negative brand information), which prior research (Hoeffler & Keller, 2002; Klein & Dawar, 2004; Sen et al., 2006) suggests may be the primary payback of CSR. This research contributes to our extant understanding of the strategic benefits of CSR by providing insight into its transactional and relational outcomes in a real-world, competitive context. Specifically, we build on the field study conducted by Sen et al. (2006) to examine the role played by the competitive positioning of three major brands of yogurt in both the formation of consumers' CSR beliefs (i.e., beliefs that a company/brand2 is socially responsible) about these brands, as well as the extent to which these beliefs are linked to both brand choice (i.e., transactional outcomes) and the set of longer-term brand advocacy behaviors, such as positive word-of-mouth and resilience to negative brand information (i.e., relational outcomes). Our findings suggest that several advantages accrue to a brand that is positioned on CSR over those that, while engaging in CSR, are positioned on other traditional, product specific dimensions such as quality. Consumers are not only more aware of what the CSR brand is doing in terms of social initiatives but also make more favorable inferences about why the brand is doing so. This not only is reflected, as expected, in stronger beliefs that the brand is socially responsible but can also spill over, positively, to consumers' beliefs about the brand's performance on dimensions unrelated to CSR (e.g., beliefs about corporate ability; Brown & Dacin, 1997). In turn, consumers' CSR beliefs regarding the CSR brand are linked more strongly to coveted relational outcomes such as brand loyalty and advocacy. Interestingly, however, such competitive effects are not obtained for the basic transactional outcome (i.e., purchase). By demonstrating these differences, we attempt to establish the boundary conditions of extant CSR findings, obtained primarily from experimental investigations (e.g., Brown & Dacin, 1997; Klein & Dawar, 2004; Sen & Bhattacharya, 2001), in an externally valid setting. Next, we draw upon relevant literatures to propose a set of predictions regarding the determinants and outcomes of consumers' CSR beliefs in a competitive context. We then present a field study that tests our predictions. We conclude with a discussion of our findings. 1. Conceptual background Since Brown and Dacin's (1997) pioneering study of the corporate associations held by consumers, much research has attested to the pivotal role of consumers' beliefs about the extent to which a company/brand is socially responsible (i.e., CSR beliefs) in their reactions to CSR.3 Consumers' CSR beliefs
2 Since consumers' CSR beliefs can be at the company level, brand level or both, in this paper we use company and brand interchangeably to capture the range of companybrand relationships (i.e., from corporate brands to standalone brands/individual brands). 3 In this research, we refer to the brand in an integrative sense (Stern, 2006) i.e., consumers' mental associations that can take on both positive as well as negative meanings.

Author's personal copy

226

S. Du et al. / Intern. J. of Research in Marketing 24 (2007) 224241

coexist with their CA (i.e., corporate ability) beliefs, or their beliefs regarding a firm's expertise in producing and delivering products/services. While consumers' transactional and relational responses will depend on both their CA and CSR beliefs, this paper focuses on the latter. Specifically, we try to understand not only the real-world determinants and consequences of consumers' CSR beliefs but also how differences in CSR positioning influence these relationships (summarized in Fig. 1). 1.1. Determinants of CSR beliefs At present, little is known about how consumers' CSR beliefs are formed. Because extant research has been concerned primarily with the consequences of CSR, it has either manipulated consumers' CSR beliefs in an experimental context (e.g., Brown & Dacin, 1997, study 1; Klein & Dawar, 2004; Sen & Bhattacharya, 2001) or assessed such beliefs without attention to its antecedents (e.g., Brown & Dacin, 1997, study 2; Lichtenstein et al., 2004, study 1). Based on prior research, we identify two interactive determinants of CSR beliefs: consumers' awareness of a brand's CSR actions (CSR awareness) and their attributions regarding the brand's motives for engaging in these CSR activities (CSR attributions). Though researchers have alluded to the importance of CSR awareness (Bhattacharya & Sen, 2004; McWilliams & Siegel, 2001) as a precursor of consumers' CSR beliefs, this link has only begun to be examined. Specifically, in a field experiment examining consumer reactions to a specific corporate philanthropic initiative, Sen et al. (2006) show that consumers' awareness of the philanthropic initiative is, not surprisingly, a precondition for their beliefs that the company is socially responsible. Of course, awareness, in and of itself, will not necessarily lead to favorable CSR beliefs. Favorability is likely to be contingent on attributions consumers make about the motives underlying a company's CSR actions, which prior research (Forehand & Grier, 2003; Menon & Kahn, 2003; Sen et al., 2006; Webb & Mohr, 1998) suggests consumers

frequently generate when exposed to CSR information. These ascribed motives are likely to be of two kinds: extrinsic and intrinsic (Batson, 1998). Extrinsic or self-interested motives have the ultimate goal of increasing the brand's own welfare (e.g., increase sales/profits or improve corporate image), whereas intrinsic or selfless motives have the ultimate goal of doing good and/or fulfilling one's obligations to society (e.g., benefit the community or cause that the CSR actions focus on). Notably, extrinsic and intrinsic motives are not viewed as two ends of a continuum; prior research (Ellen, Webb, & Mohr, 2006) indicates that a brand's CSR actions can be attributed to both intrinsic and extrinsic motives. What is the role of competitive positioning in determining consumers' CSR awareness, attributions and beliefs, as well as the relationships among these? Research on the effects of product positioning on brand equity and, more broadly, brand associations (Keller, 1993; Park, Jaworski, & MacInnis, 1986; Punj & Moon, 2002) suggests that when a brand is positioned as a CSR brand rather than a brand that just engages in CSR, it is likely to accrue several consumer-specific benefits. For instance, it is likely that consumers' CSR awareness levels will be higher for a CSR brand than its competitors because such a brand is likely to have more explicit and sustained communications regarding its CSR activities. Also, given that the brand has taken the more extreme, relatively uncommon and perhaps risky stance of positioning itself on CSR rather than just engaging in such activities, consumers are likely to make stronger intrinsic and weaker extrinsic attributions about its motives (Gilbert & Malone, 1995). More generally, there will be greater consistency for such a brand between consumers' prior expectations of it and its CSR actions, reinforcing over time the brand's positioning. These together are likely to lead to more favorable CSR beliefs for the CSR brand than for its competitors. Thus, H1. Consumers will have (a) higher levels of CSR awareness, (b) higher intrinsic attributions, (c) lower extrinsic attributions,

Fig. 1. Conceptual framework.

Author's personal copy

S. Du et al. / Intern. J. of Research in Marketing 24 (2007) 224241

227

and (d) more favorable CSR beliefs for a CSR brand than for its competitors. Sen et al. (2006) demonstrate that the intrinsic attributions consumers make about a company's CSR efforts are a positive moderator of the link between their CSR awareness and CSR beliefs. Thus, we expect consumers' awareness of a brand's CSR record to be associated with more positive CSR beliefs when they make intrinsic attributions regarding the underlying motives. On the other hand, research on suspicion (Fein, 1996) and, more specifically, the persuasion knowledge model (Campbell & Kirmani, 2000; Friestad & Wright, 1994) suggests that perceived extrinsic motives underlying a brand's CSR actions will diminish the favorability of its CSR beliefs. Thus, even though consumers often expect companies' CSR actions to be motivated at least in part by self-interest (Ellen et al., 2006), such attributions are likely to be associated with the belief that the firm is not truly socially responsible. In other words, extrinsic attributions are likely to be a negative moderator of the CSR awarenessCSR beliefs relationship. How might these moderating effects of attributions vary with the brands' competitive positioning? We expect consumers to be more sensitive to and rely more heavily on their causal attributions, both intrinsic and extrinsic, in ascertaining their beliefs about a CSR brand than that of its competitors. This is the case for three related reasons. First, because consumers are more likely to engage in causal attributions for the CSR brand due to its extreme and, at least in today's marketplace, atypical positioning, these salient causal cognitions are more likely to be used as inputs in consumers' subsequent judgments regarding such brands. As well, research suggests that consumers are likely to make more dispositional rather than situational attributions regarding the CSR brand than its competitors (Fiske & Taylor, 1991; Nesbitt & Ross, 1980) and therefore be more susceptible to a correspondence bias (see Gilbert & Malone, 1995 for a review), enhancing the relevance of these attributions to their CSR judgments regarding the brand. Finally, compared to its competitors, the CSR activities of a CSR brand are more likely, by virtue of their magnitude and close link to the brand, to be viewed by consumers as an inherent and important aspect of the brand rather than a tangential, irrelevant one. Thus, consumers are likely to process the CSR brand's CSR information, compared to that of its competitors, in a more systematic rather than a peripheral or heuristic manner (Chaiken & Trope, 1999), deliberating on not just the what but also the why. Hence, we propose that: H2a. The moderating role of intrinsic attributions in the CSR awarenessCSR beliefs relationship will be stronger for a CSR brand than for its competitors. H2b. The moderating role of extrinsic attributions in the CSR awarenessCSR beliefs relationship will be stronger for a CSR brand than for its competitors. 1.1.1. Influence on CA beliefs Finally, we also expect that consumers' brand associations for such CSR positioned brands will result in a stronger association between their CSR awareness for such brands and

their CA beliefs regarding these brands. Because such brands are likely to focus on CSR as an integral part of their core business objectives and competencies (i.e., their core business strategy), indeed their raison d'etre, over time consumers are likely to link, in their brand associations, the CSR actions of such brands to their ability to make good products and services. For instance, the Body Shop's use of natural ingredients and its environmentally friendly practices have become associated with perceptions of high quality products (Rummell, 1999). Relatedly, prior research (e.g., Sen & Bhattacharya, 2001) suggests that consumers are sensitive to the extent to which a company's CSR activities have direct implications for its ability to produce high quality products (i.e., its CA); some CSR initiatives are considered more CA-relevant than others. To the extent that the CSR brands' CSR activities are more integrated with their overall business strategy, consumers are more likely to observe a link between the CSR initiatives and the CA of such brands compared to their competitors. This is also consistent with Hoeffler and Keller's (2002) assertion that a considered engagement in CSR can enhance consumers' perceptions of brand credibility, of which CA-related expertise is a key component. H3. The relationships between consumers' CSR awareness and attributions and their CA beliefs will be stronger for a CSR brand than for its competitors. 1.2. Consequences of CSR beliefs Lichtenstein et al. (2004) document the transactional benefits of corporate philanthropy in a field study: consumers with more positive CSR beliefs about a grocery chain buy more from that chain. Moreover, they confirm the finding of prior research (e.g., Sen & Bhattacharya, 2001) that consumercompany identification (CC identification), which refers to consumers' psychological attachment to a company based on a substantial overlap between their perceptions of themselves and their perceptions of the company (Bergami & Bagozzi, 2000; Dutton, Dukerich, & Harquail, 1994), is a key internal consequence of such CSR beliefs. Interestingly, such identification does not appear to play a mediating role in consumers' transactional reactions to CSR (Lichtenstein et al., 2004, Table 1). However, this is not entirely surprising, given the growing sense (e.g., Bhattacharya & Sen, 2003) that such identification may be more likely to drive their longer-term, relational reactions to the company/brand rather than (just) product purchase. In particular, prior identification research (Bhattacharya & Sen, 2003; Maignan & Ferrell, 2004) suggests that CC identification causes people to become psychologically attached to and care about the company and its products, motivating them to commit to the achievement of its goals and expend more voluntary efforts on its behalf. Thus, CSR-based identification is likely to be associated with a range of relational behaviors that go beyond product purchase (Lichtenstein et al., 2004) to consumers' loyalty to the company's existing products (i.e., customer retention), their willingness to try its new products (i.e., cross-selling opportunities), favorable word-of-mouth, and

Author's personal copy

228 Table 1 Sample characteristics a Whole sample n = 11,000 Gender Male Female Age Less than 30 3140 4150 5160 Over 60 Education Some high school Graduate high school or G.E.D Graduated from college Graduate degree Household income Less than $25,000 $25,000$49,999 $50,000$74,999 $75,000 or more
a

S. Du et al. / Intern. J. of Research in Marketing 24 (2007) 224241

All respondents n = 3465 20.4% 78.5% 17.6% 29.5% 28.0% 19.5% 4.3% 1.4% 33.9% 41.6% 21.5% 7.8% 23.5% 24.3% 37.1%

Reduced set n = 1062 22.5% 77.1% 16.4% 33.7% 26.5% 18.8% 4.0% 1.2% 33.0% 42.9% 21.7% 7.4% 22.6% 25.5% 38.5%

behaviors for the brand (i.e., willing to try new products, engage in favorable word-of-mouth, and display greater resilience to negative information) when it is a CSR brand than when it is not. H5. The relationship between consumers' CSR beliefs and its consequences: (a) CC identification, (b) loyalty, and (c) advocacy behaviors, will be stronger for a CSR brand than for its competitors. Next, we describe the study that tests these hypotheses. 2. Method 2.1. Research design We conducted a web-based survey of consumers' reactions to three competitors in the yogurt category using Zoomerang (www.zoomerang.com), a leading software for web-based surveys. Zoomerang is particularly appropriate for a category-bound study such as ours because it allows sampling from its over two million member panel in a rapid, cost effective, and flexible manner using a wide range of selection criteria from demographics to category-specific purchase and consumption behaviors (Miller, 2006). At the same time, our theory-testing objective minimizes the general concern that web-based samples (i.e., Internet users) are not representative of the population as a whole (Hunt, 1990; Morgan & Hunt, 1994). We focused on the yogurt category for two main reasons, which emerged from the two focus groups we conducted with consumers in a large Northeast city in the United States as qualitative precursors to our survey. First, yogurt was a commonly purchased product in this population with several well-known brands in the market. As importantly, the three yogurt brands that we included in our survey, Dannon, Stonyfield Farm and Yoplait, all engage in CSR activities, but with requisite variance in the characteristics of their CSR strategies to allow tests of our predictions regarding the effects of the competitive context. Specifically, Dannon, a dominant player in this market, is positioned primarily on CA (Consumer Reports, 2002; Thompson, 2001). While the other two players, Yoplait and Stonyfield, are both more focused on CSR, it is only Stonyfield that is truly and clearly positioned as a CSR brand, as evidenced by the business press (e.g., GlorieuxBoutonnet, 2004; Smith, 2002), the web sites of the three brands, and, most importantly, the responses of our focus group participants. In other words, these three brands comprise a continuum in terms of CSR positioning, with a CA focused player (i.e., Dannon) at one end, and a CSR brand (i.e., Stonyfield) at the other. 2.1.1. CSR initiatives of the three brands The three focal brands engage in various CSR activities (Dannon Cares, 2005; Earth Actions, 2005; Yoplait is Committed to Fighting Breast Cancer, 2005). Specifically, Dannon has been active in fighting hunger in America, working with America's Second Harvest, the nation's largest hunger-relief organization, to end childhood hunger in the U.S.

24.8% 75.2% 14.1% 27.7% 28.6% 22.7% 6.9% 1.0% 38.3% 39.0% 21.7% 7.6% 21.3% 21.7% 36.9%

The percentages may not add up to 100% exactly because of missing values.

resilience in the face of negative information about the company, such as in a product-harm crisis (e.g., Klein & Dawar, 2004). Given that favorable CSR beliefs are likely to be a key driver of CC identification (Bhattacharya & Sen, 2003; Lichtenstein et al., 2004), such beliefs can be expected to strengthen the consumerbrand relationship, leading to a range of advocacy behaviors. As in the case of the determinants of CSR beliefs, we propose that the consequences of such beliefs will also vary with a brand's competitive positioning. In particular, the CC identification literature suggests two basic differences. First, we expect consumers of a CSR brand to display higher levels of identification and, consequently, loyalty and advocacy behaviors. Second, we expect the relationship between CSR beliefs and its relational consequences to be stronger for a CSR brand than for its competitors. This is because a brand that integrates its CSR activities into its core business strategy and makes CSR the essence or soul of the brand will present a clearer, more sincere CSR identity to the socially conscious consumer. Given that a CSR-based brand identity, in contrast to conventional CAbased positioning platforms such as quality or price, is likely to be more effective at engendering identification (Bhattacharya & Sen, 2003), the CSR beliefs of such a brand will be more strongly associated with CC identification and, consequently, loyalty and advocacy behaviors among its target consumers. On the other hand, brands that lack a systematic, integrated approach toward CSR, though attractive to consumers, are less likely to be candidates for identification and, therefore, less likely to engender voluntary advocacy behaviors. Thus: H4. Consumers of a brand are more likely to (a) identify with the brand, (b) be loyal to the brand, and (c) engage in advocacy

Author's personal copy

S. Du et al. / Intern. J. of Research in Marketing 24 (2007) 224241

229

Yoplait has been committed to battling breast cancer. In 1998, it launched Save Lids to Save Lives, its signature pink-lid promotion, and has been the National Series Presenting Sponsor of the Susan G. Komen Breast Cancer Foundation's Race for the Cure since 2001. Stonyfield Farm, positioned as the brand For a Healthy Planet, is focused primarily on environmental responsibility. The brand's numerous environmental initiatives include: (1) organic farming: over 80% of Stonyfield Farm sales are from certified organic products which are produced using earth-friendly practices, (2) environmental packaging: reducing excessive packaging by switching to lighter weight plastic, and (3) Profits for the Planet (PFP) program: under this program, the company gives 10% of its profit to efforts that help protect and restore the earth through non-profit organizations such as The Sustainability Institute, Children's Health Environmental Coalition, and The Rainforest Foundation. Notably, the spirit of CSR pervades every aspect of Stonyfield's business: the corporate mission, production and packaging, human resources management and importantly, marketing, including its communication strategy. Not surprisingly, it has won several CSR awards including the Green Cross Millennium Award for Corporate Environmental Leadership from Global Green USA/Green Cross International and the National Award for Sustainability in the Category of Atmosphere and Climate from the President's Council on Sustainable Development and Renew America (Earth Actions, 2005). 2.2. Data collection A web-link to the questionnaire was emailed to 11,000 Zoomerang panelists that satisfied our selection criteria, i.e., panelists who resided in the North Eastern region of the United States, were employed, and purchased yogurt. We restricted ourselves to the North East as all three brands enjoy full-scale distribution in this region and we used the employment screen to get a reasonable proportion of consumers who purchase branded yogurt. As a token of appreciation and an incentive to participate in our survey, respondents who answered every question in the survey were entered in a drawing for one of two $100 gift certificates for Amazon.com. Overall, we received 3465 completed questionnaires representing a 31.5% response rate. 2.2.1. Sample characteristics As shown in Table 1, our sampling method succeeded in obtaining respondents who varied sufficiently on demographic (i.e., gender, age) and socioeconomic (i.e., education, income) status. A majority of our respondents were female, which is not surprising given that women are not only heavier consumers of yogurt but also more likely to do the grocery shopping. To ascertain non-response bias, we compared our survey respondents to the larger sample they belong to (i.e., the 11,000 Zoomerang panelists). Characteristics of the survey respondents and the whole sample are comparable (Table 1); the differences in gender, age, education and income between the respondents

and the whole sample, if any, are small suggesting that nonresponse bias is not a significant problem.4 2.3. Measures Given our predictions, the key variables in our study were: CSR awareness, CSR attributions, CA and CSR beliefs, identification, brand choice, loyalty and advocacy (measure details in Appendix A). We also measured consumers' support for the issues defining the CSR of these brands (CSR support). Brand choice was measured in terms of the brand last purchased. The remaining variables, except CSR support, were measured using five-point Likert scales with verbal anchors (i.e., strongly disagree, disagree, neither agree nor disagree, agree, and strongly agree). CSR support items were anchored by 1=do not support at all, and 5=strongly support. In keeping with our study objectives, measures of CSR awareness and CSR beliefs also had an NA (i.e., Not Aware) option to allow consumers to express their inability to respond to these items due to their lack of familiarity with a particular brand's CSR status. All scale items were developed based on the prior literature as well as the responses of our focus group participants. The measures for CA and CSR beliefs were informed primarily by our focus group discussions. Focus group respondents characterized CA beliefs in terms of the quality, taste, nutrition and value perceptions of the different brands whereas CSR beliefs were expressed in terms of the extent to which a brand is socially responsible and has made a difference to the social cause it supports. In line with prior attribution research (Ellen et al., 2006; Forehand & Grier, 2003) and our focus group discussions, we measured intrinsic CSR motives as the brand's genuine concern in being socially responsible and extrinsic CSR motives in terms of the competitive pressures the brand faces to engage in CSR activities. The identification measure was taken from Bhattacharya and Sen (2003) and the advocacy behaviors were captured by three items: intention to try new products of the brand (Bhattacharya & Sen, 2003), favorable word-of-mouth (Aaker, 1996; Zeithaml, Berry, & Parasuraman, 1996), and resilience to negative information (Klein & Dawar, 2004; Peloza, 2006). Respondents first indicated their brand choice behavior in this category by selecting the brand they purchased most recently. Next, respondents also indicated the brand they purchased most frequently. Those purchasing brands other than the three focal competitors selected the option Other. Respondents then provided loyalty, advocacy and identification ratings for all three brands. Next, respondents provided their CA and CSR beliefs regarding these three brands. Respondents then indicated their support for the issues underlying the CSR of each brand and their awareness of each brand's CSR initiatives. Following this, respondents were informed of the CSR initiatives of each brand and asked to provide their attributions regarding the motives underlying each brand's CSR activities.
4 Again, while our respondents or our chosen sample may not be representative of the entire population, this is not a major concern given our interest in theory testing rather than in the generalizability of our findings to the entire population.

Author's personal copy

230

S. Du et al. / Intern. J. of Research in Marketing 24 (2007) 224241

Respondents' age, gender, education and income levels were collected at the end of the survey. 2.4. Preliminary checks and controls Not surprisingly, and in line with prior research (Sen et al., 2006), we found that respondents' accurate awareness of a particular brand's CSR initiative(s) was generally low: about 20% of respondents checked either 4 [agree] or 5 [strongly agree] on the relevant CSR awareness items. At the same time, there was greater support for the CSR issues of breast cancer (Mean = 3.44; 51% of respondents checked either 4 [support] or 5 [strongly support]) and the environment (Mean = 3.34; 49% of respondents checked either 4 [support] or 5 [strongly support]) than for the eradication of hunger (Mean = 3.01; 37% of respondents chose either 4 [support] or 5 [strongly support]). We speculate that the majority of our respondents being female may account, at least partially, for the relative popularity of the fight against breast cancer cause. Finally, the correlation between intrinsic and extrinsic attributions was low (r = .05), supporting the notion that these are two distinct dimensions of consumers' CSR attributions rather than two ends of a one-dimensional construct. In line with our research objectives to test our predictions regarding the moderating effect of competitive positioning on the determinants and consequences of CSR beliefs we excluded from our analysis respondents who chose NA (i.e., Not Aware) on measures of CSR awareness and CSR beliefs (n = 2163), and who had missing values on other key variables (n = 240) reducing the total number of respondents to 1062. The demographic information of these respondents (Reduced Set in Table 1) indicates that they are comparable to both the original respondent set (i.e., n = 3465) and the overall sample (i.e., n = 11,000). As a preliminary check of our basic premise, we also contrasted the responses of those who indicated awareness of the brands' CSR (i.e., CSR Aware group) to those who were not (i.e., CSR Unaware group). As expected, identification, loyalty and advocacy levels were greater among the CSR Aware group than the CSR Unaware one. Not surprisingly, while there is no difference in the purchase likelihood of Dannon or Yoplait between the CSR Aware group and CSR Unaware group, the CSR Aware group is more likely to purchase Stonyfield than the CSR Unaware group. These differences are consistent with the fact that awareness of a company's CSR initiatives is at least a partial prerequisite for identification and the consequent relational and transactional behaviors. More generally, these differences are probably also a reflection of the varying levels of category involvement across the two groups. 2.5. Measure validation Using the reduced dataset (n = 1062), we validated our multiitem measures by assessing the item-to-total correlations for the items in each scale. As shown in Appendix A, the measures are internally reliable. Before testing our hypotheses, we also checked for common-method variance, which can be a problem in a study such as ours where data on both antecedents and consequences are collected from the same respondent using

similar types of response scales (i.e., Likert scales). We checked for this potential problem using Harman's single-factor test (Podsakoff, MacKenzie, Lee, & Podsakoff, 2003; Podsakoff & Organ, 1986), which suggests that a substantial amount of common-method variance is present if (a) a single unrotated factor solution emerges from an exploratory factor analysis or (b) one general factor accounts for the majority of the covariance among the measures. With our data, the unrotated factor solution revealed five factors with eigen values greater than one, accounting for 64.3% of the total variance (the first factor accounts for 31.6% of the total variance). In other words, there was no general factor in the unrotated structure, which suggests that common-method variance does not pose a significant problem. 2.6. Hypothesis tests Our use of single item measures for many of our constructs to minimize respondent fatigue precluded hypothesis tests based on structural equations models. Consequently, we combined differences of means tests (for H1 and H4) with regression analyses (for H2, H3, and H5) to test our predictions regarding brand-specific differences in the determinants and consequences of CSR beliefs. Brand-specific differences in the regression
Table 2 Summary of hypothesis test results Hypothesis Key findings H1 Consumers have higher levels of CSR awareness, stronger intrinsic attributions, weaker extrinsic attributions, and more favorable CSR beliefs for a CSR brand than for its competitors. Intrinsic attributions are a positive moderator of the CSR awarenessCSR beliefs relationship, and this moderating role is stronger for a CSR brand than for its competitors. Extrinsic attributions are a negative moderator of the CSR awarenessCSR beliefs relationship, and this moderating role is stronger for a CSR brand than for its competitors. The relationships between consumers' CSR awareness and intrinsic attributions and their CA beliefs are stronger for a CSR brand than for its competitors. Consumers of a brand identify with the brand to a greater extent, are more loyal, and engage in more advocacy behaviors when it is a CSR brand than when it is not. The relationships between consumers' CSR beliefs and their CC identification, loyalty, and advocacy behaviors are stronger for a CSR brand than for its competitors. Hypothesis supported a

H2

Hypothesis supported

Hypothesis supported

H3

Hypothesis partially supported b Hypothesis supported

H4

H5

Hypothesis supported

The hypothesized brand-specific differences regarding CSR awareness and intrinsic attributions are supported in the Stonyfield vs. Dannon/Yoplait (pooled) and Stonyfield vs. Dannon comparisons but not in Stonyfield vs. Yoplait comparison. b There is stronger link between intrinsic attributions and CA beliefs for Stonyfield than for Dannon or Yoplait; stronger link between CSR awareness and CA beliefs is only found in Stonyfield vs. Dannon comparison, not in Stonyfield vs. Yoplait comparison.

Author's personal copy

S. Du et al. / Intern. J. of Research in Marketing 24 (2007) 224241 Table 3a Comparison of CSR awareness, attributions and CSR beliefs: means and standard deviations Means (S.D.) Stonyfield All respondents, n = 1062 CSR awareness 3.34 (.71) Intrinsic attributions 3.56 (.77) Extrinsic attributions 3.29 (.76) CSR beliefs 3.31 (.59) Frequent buyers of a brand n = 130 CSR awareness 3.89 (.93) Intrinsic attributions 4.03 (.77) Extrinsic attributions 3.12 (.90) CSR beliefs 3.92 (.73) p b .01. Dannon 3.02 (.45) 3.32 (.66) 3.40 (.72) 3.11 (.47) Yoplait 3.43 (.80) 3.51 (.76) 3.38 (.77) 3.13 (.51) T statistics Stonyfield vs. Dannon & Yoplait (pooled) 4.25 5.08 3.66 9.33 Stonyfield vs. Dannon 12.09 7.58 3.53 8.46

231

Stonyfield vs. Yoplait 2.76 1.37 2.81 7.23

n = 410 3.07 (.49) 3.37 (.66) 3.40 (.74) 3.17 (.49)

n = 262 3.79 (.93) 3.64 (.77) 3.36 (.79) 3.31 (.54)

7.03 8.03 3.49 13.00

13.15 9.57 3.56 13.30

.99 4.77 2.65 8.28

analyses were modeled using the standard approach of indicator (i.e., dummy) variables. Specifically, we conducted two levels of analyses, each with Stonyfield as the reference brand. First, we looked at the effects for Stonyfield (the CSR brand) vs. that for Yoplait and Dannon, pooled (i.e., the non-CSR brands). The second level of analyses entailed two indicator variables to look at the effects, separately, for Stonyfield vs. Dannon and Stonyfield vs. Yoplait. While mean-centering the continuous predictor variables in moderated regressions does not, contrary to common belief, reduce potential multi-collinearity (Gatignon & Vosgerau, 2005), it does make it somewhat easier to interpret the regression coefficients for first order terms (Aiken & West, 1991): the coefficient of a first order predictor variable represents its effect on the dependent variable when the moderator variable is at its mean value (i.e., is zero because of the mean-centering). Therefore, we mean centered the continuous variables wherever appropriate. Finally, we included CSR Support as a covariate in all analyses. However, since this variable was not significant in any of the analyses, we report findings from analyses that did not include CSR support as a covariate. Table 2 provides a summary of findings based on the hypothesis tests. Consistent with our expectations, our findings indicate that the CSR brand does reap more relationship benefits from its social initiatives than do its competitors. More specifically, consumers tend to be more aware of and make more favorable attributions regarding the CSR initiatives of Stonyfield, considering it to be more socially responsible than its competitors, Yoplait and Dannon. Furthermore, these CSR beliefs held by consumers are associated with greater identification, loyalty and advocacy behaviors for Stonyfield than for the two brands not positioned on CSR. Interestingly, we also find a stronger link between consumers' CSR-related perceptions (i.e., CSR awareness, attributions) and their CA-related beliefs (e.g., product quality, taste, value) for Stonyfield, the CSR brand, than for the non-CSR brands. We discuss these results in detail next. 2.6.1. Determinants of CSR beliefs We tested H1 using difference of means tests. Specifically, we expect consumers' awareness of and attributions regarding

Stonyfield's (i.e., the CSR brand) CSR activities to be greater and more favorable than that of its competitors. These, in turn, are expected to yield more favorable CSR beliefs for Stonyfield than for its competitors. Comparisons of respondents' CSR awareness, intrinsic and extrinsic attributions, and CSR beliefs for Stonyfield with those for Dannon and Yoplait, pooled (i.e., Dannon/Yoplait), revealed significant differences in the predicted direction (Table 3a). Specifically, respondents were more aware of Stonyfield's CSR activities than that of Dannon/Yoplait, made stronger intrinsic attributions and weaker extrinsic attributions in the case of Stonyfield than in the case of its competitors, and have more favorable CSR beliefs for Stonyfield than for its competitors. Notably, comparisons of these variables between Stonyfield and each of its competitors (Table 3a), separately, reveals that consumers make less extrinsic attributions and have more positive CSR beliefs about Stonyfield compared to both Dannon and Yoplait. In the case of CSR awareness and intrinsic attributions, however, such brand-specific differences were only obtained between Stonyfield and Dannon. This is due possibly to the disproportionate number of women in our sample, who are more involved with and informed about Yoplait's CSR issue (i.e., fighting breast cancer). To further understand these brand-specific differences, we compared the CSR awareness, attributions and belief levels of each brand among only its frequent buyers (i.e., those who reported buying the brand most frequently). The results are provided in Table 3a. Of the 1062 respondents, 410 bought Dannon, 130 bought Stonyfield, and 262 bought Yoplait most frequently. The remaining respondents either reported buying another brand or buying more than one brand most frequently.5 Compared to Dannon's frequent buyers, those of Stonyfield displayed higher awareness levels, stronger intrinsic attributions, weaker extrinsic attributions and more favorable CSR beliefs for their brand. Similarly, compared to the Yoplait's frequent buyers, Stonyfield's frequent buyers displayed stronger intrinsic attributions, weaker extrinsic attributions, and more
5 Analyses with and without the respondents who frequently purchase more than one brand yielded similar results. Thus, we report results from data that includes respondents who buy multiple brands most frequently.

Author's personal copy

232

S. Du et al. / Intern. J. of Research in Marketing 24 (2007) 224241

Table 3b Comparison of CSR awareness, attributions and CSR beliefs among each brand's frequent buyers: means and standard deviations Means (S.D.) Stonyfield Stonyfield's frequent buyers, n = 130 CSR awareness 3.89 (.93) Intrinsic attributions 4.04 (.77) Extrinsic attributions 3.12 (.90) CSR beliefs 3.92 (.73) Dannon's frequent buyers, n = 410 CSR awareness 3.27 (.61) Intrinsic attributions 3.50 (.74) Extrinsic attributions 3.35 (.77) CSR beliefs 3.20 (.51) Yoplait's frequent buyers, n = 262 CSR awareness 3.32 (.63) Intrinsic attributions 3.48 (.72) Extrinsic attributions 3.27 (.67) CSR beliefs 3.31 (.55) p b .01. p b .05. + p b .10. Dannon 2.88 (.59) 3.23 (.79) 3.47 (.78) 2.92 (.56) Yoplait 3.42 (.80) 3.62 (.85) 3.44 (.79) 3.00 (.61) T statistics Stonyfield vs. Dannon & Yoplait (pooled) 8.43 6.99 3.74 14.07 6.47 2.68 .77 .12 11.02 4.46 .53 1.86+ Stonyfield vs. Dannon 10.39 8.32 3.32 12.42 5.25 2.48 1.02 .87 6.89 2.58 1.37 .16 Stonyfield vs. Yoplait 4.36 4.21 3.01 11.07 6.55 2.26 .32 .66 10.89 5.43 .47 3.19

3.07 (.49) 3.37 (.66) 3.40 (.74) 3.17 (.49)

3.36 (.78) 3.49 (.76) 3.39 (.78) 3.15 (.51)

3.10 (.45) 3.31 (.64) 3.39 (.68) 3.18(.44)

3.79 (.93) 3.65 (.77) 3.36 (.79) 3.31 (.54)

favorable CSR beliefs for their brand. No differences were obtained, however, for CSR awareness. We also find some notable asymmetries in frequent buyers' perceptions of the brands they do not buy (Table 3b). For example, as expected, Stonyfield's frequent buyers are more aware of Stonyfield's CSR initiatives, and make more intrinsic attributions and less extrinsic attributions for Stonyfield than is the corresponding case for Dannon or Yoplait. However, Dannon's frequent buyers are less aware of Dannon's CSR initiatives than Stonyfield's or Yoplait's CSR initiatives. They also make less intrinsic attributions for Dannon than for Stonyfield or for Yoplait. Similarly, while Dannon's and Yoplait's frequent buyers consider Stonyfield to be as socially responsible as their own brand, Stonyfield's frequent buyers believe Stonyfield is more socially responsible than both Dannon and Yoplait. In sum, these analyses suggest that compared to the consumers, particularly the frequent users, of a CSR brand's competitors, those who consume the CSR brand not only respond more positively to its CSR actions, as reflected in their awareness, attributions and CSR beliefs, but also less positively to those of the competition. To test H2, we ran the regression models shown below. The overall model (1) is primarily a replication of prior research, whereas model (2) looks at the effect for Stonyfield vs. Dannon/ Yoplait and (3) models the effects for the three brands individually: CSR Beliefs Intercept b1 CSR Awareness b2 Intrinsic Attributions b3 Extrinsic Attributions b4 CSR Awareness Intrinsic Attributions b5 CSR Awareness Extrinsic Attributions e 1

CSR Beliefs Intercept b1 D b2 CSR Awareness b3 Intrinsic Attributions b4 Extrinsic Attributions b5 CSR Awareness D b6 Intrinsic Attributions D b7 Extrinsic Attributions D b8 CSR Awareness Intrinsic Attributions b9 CSR Awareness Extrinsic Attributions b10 CSR Awareness Intrinsic Attributions D b11 CSR Awareness Extrinsic Attributions D e 2 CSR Beliefs Intercept b1 D1 b2 D2 b3 CSR Awareness b4 Intrinsic Attributions b5 Extrinsic Attributions b6 CSR Awareness D1 b7 CSR Awareness D2 b8 Intrinsic Attributions D1 b9 Intrinsic Attributions D2 b10 Extrinsic Attributions D1 b11 Extrinsic Attributions D2 b12 CSR Awareness Intrinsic Attributions b13 CSR Awareness Extrinsic Attributions b14 CSR Awareness Intrinsic Attributions D1 b15 CSR Awareness Intrinsic Attributions D2 b16 CSR Awareness Extrinsic Attributions D1 b CSR Awareness Extrinsic Attributions D2 e
17

Where D = 1 if the brand is Dannon or Yoplait, D = 0 if the brand is Stonyfield. D1 = 1 if the brand is Dannon, else D1 = 0. D2 = 1 if the brand is Yoplait, else D2 = 0. And is the error term. Note that since all the variables in the above regressions are measured at the level of individual consumer

Author's personal copy

S. Du et al. / Intern. J. of Research in Marketing 24 (2007) 224241

233

for each of the three brands, we omit the consumer and brand subscripts for expositional ease. The results of these regressions are presented in Table 4. Regression model (1) tests the overall moderating role of CSR attributions in the CSR awarenessCSR beliefs relationship, across the three brands. Replicating prior findings, both types of attributions moderate the CSR awarenessCSR beliefs relationship. Specifically, the significant, positive interaction between CSR awareness and intrinsic attributions (4 = .03, p b .05) indicates that the link between CSR awareness and CSR beliefs is stronger for those respondents making more intrinsic attributions regarding a brand's CSR activities. Similarly, the significant, negative interaction between CSR awareness and extrinsic attributions (5 = .04, p b .05) indicates that the link between CSR awareness and CSR beliefs is weaker for those respondents making more extrinsic attributions regarding a brand's CSR activities. H2a and H2b propose that these moderating roles of both intrinsic and extrinsic attributions will be stronger for a CSR brand relative to its competitors. The results of regression models (2) and (3) (Table 4) provide support for these hypotheses. Comparing Stonyfield to Dannon/Yoplait (i.e., model 2), we find that the coefficient of CSR Awareness Intrinsic Attributions is .12 ( p b .01) and that of CSR Awareness Intrinsic Attributions D is .15 ( p b .01). The significant three way interaction indicates that there is a significant difference in the strength of CSR Awareness Intrinsic Attributions between
Table 4 Determinants of CSR beliefs: unstandardized regression coefficients (T statistics) a, b Independent variables Model 1 Overall CSR awareness Intrinsic attributions Extrinsic attributions CSR awareness Intrinsic Attributions CSR awareness Extrinsic Attributions CSR awareness D Intrinsic Attributions D Extrinsic Attributions D CSR awareness Intrinsic Attributions D CSR awareness Extrinsic Attributions D CSR awareness D1 CSR awareness D2 Intrinsic Attributions D1 Intrinsic Attributions D2 Extrinsic Attributions D1 Extrinsic Attributions D2 CSR awareness Intrinsic Attributions D1 CSR awareness Intrinsic Attributions D2 CSR awareness Extrinsic Attributions D1 CSR awareness Extrinsic Attributions D2 Adjusted R2 d.f. Model F .33 (26.57) .21 (18.90) .02 ( 2.31) .03 (2.94) .04 ( 3.42) Model 2

Stonyfield and Dannon/Yoplait. Specifically, since D = 1 if the brand is Dannon or Yoplait and D = 0 if the brand is Stonyfield, we can infer that the coefficient of CSR Awareness Intrinsic Attributions is .12 for Stonyfield and .12 + ( .15)= .03 for Dannon/Yoplait. Consistent with H2a, this suggests that the moderating role of intrinsic attributions is more positive for Stonyfield than for its competitors. Similarly, we find that in model 2, the coefficient of CSR Awareness Extrinsic Attributions is .07 ( p b .01) and that of CSR Awareness Extrinsic Attributions D is .07 ( p b .01). Again, this indicates that there is a significant difference in the strength of two way interaction CSR Awareness Extrinsic Attributions between Stonyfield (i.e., D = 0) and Dannon/ Yoplait (i.e., D = 1). Specifically, we infer that the coefficient of CSR Awareness Extrinsic Attributions is .07 for Stonyfield and zero (i.e., .07 + .07) for Dannon/Yoplait, suggesting, in line with H2b, that the moderating role of extrinsic attributions is more negative for Stonyfield than for its competitors. Notably, separate comparisons of Stonyfield vs. Dannon and Stonyfield vs. Yoplait yield similar results (see model 3 in Table 4). In sum, our regression results provide unequivocal support for H2a and H2b. 2.6.1.1. Influence on CA beliefs. H3 suggests that consumers' CSR awareness and attributions are more likely to spill over to their CA beliefs for Stonyfield than for its competitors. To test this, we ran the following overall regression model (4)

Model 3 Stonyfield vs. Dannon and vs. Yoplait .31 (14.14) .26 (14.41) .03 ( 1.48) .12 ( 6.15) .07 ( 3.96)

Stonyfield vs. Dannon & Yoplait (pooled) .31 (13.98) .26 (14.25) .03 ( 1.47) .12 ( 6.08) .07 ( 3.92) .00 (.07) .10 ( 4.43) .02 ( 1.07) .15 ( 6.21) .07 (3.18)

.39 (5, 3180) 400.72

.41 (11, 3174) 204.06

.18 (4.88) .07 ( 2.35) .11 ( 3.99) .09 ( 3.37) .05 (1.94) .02 (.64) .11 ( 3.47) .14 ( 5.42) .12 (3.84) .04 (1.72)+ .43 (16, 3169) 148.36

p b .01. p b .05. + p b .10. a D = 1 if brand is Dannon or Yoplait, D = 0 if brand is Stonyfield. b D1 = 1 if brand is Dannon, else D1 = 0; D2 = 1 if brand is Yoplait, else D2 = 0.

Author's personal copy

234

S. Du et al. / Intern. J. of Research in Marketing 24 (2007) 224241

Table 5 Determinants of CA beliefs: unstandardized regression coefficients (T statistics) a, b Independent variables Model 4 Overall CSR awareness Intrinsic attributions Extrinsic attributions CSR awareness Intrinsic Attributions CSR awareness Extrinsic Attributions CSR awareness D Intrinsic Attributions D Extrinsic Attributions D CSR awareness Intrinsic Attributions D CSR awareness Extrinsic Attributions D CSR awareness D1 CSR awareness D2 Intrinsic Attributions D1 Intrinsic Attributions D2 Extrinsic Attributions D1 Extrinsic Attributions D2 CSR awareness Intrinsic Attributions D1 CSR awareness Intrinsic Attributions D2 CSR awareness Extrinsic Attributions D1 CSR awareness Extrinsic Attributions D2 Adjusted R2 d.f. Model F .16 (8.60) .25 (14.86) .01 (.66) .04 (2.63) .00 ( .21) Model 5 Stonyfield vs. Dannon & Yoplait (pooled) .20 (5.99) .31 (11.15) .01 (.34) .07 (2.20) .02 ( .60) .06 ( 1.45) .09 ( 2.50) .01 ( .16) .05 ( 1.34) .03 (1.01) Model 6 Stonyfield vs. Dannon and vs. Yoplait .20 (6.00) .31 (11.15) .01 (.34) .07 (2.20) .02 ( .60)

.15 (5, 3180) 109.18

.17 (11, 3174) 60.81

.10 ( 1.82)+ .04 ( .85) .08 ( 1.87)+ .08 ( 1.98) .02 (.53) .01 ( .28) .02 ( .46) .08 ( 1.91)+ .10 (2.06) .02 (.45) .17 (16, 3169) 42.30

p b .01. p b .05. + p b .10. a D = 1 if brand is Dannon or Yoplait, D = 0 if brand is Stonyfield. b D1 = 1 if brand is Dannon, else D1 = 0; D2 = 1 if brand is Yoplait, else D2 = 0.

and its brand-specific extensions (5 and 6 in Appendix B), which are analogous to models (2) and (3):
CA Beliefs Intercept b1 CSR Awareness b2 Intrinsic Attributions b3 Extrinsic Attributions b4 CSR Awareness Intrinsic Attributions b CSR Awareness Extrinsic Attributions
5

e:

4 The results of regression models (4)(6) are reported in Table 5. As mentioned earlier, note that the coefficients of the main effects in these models represent the effects of these variables only when the moderator variables are their mean value. Results of the overall model (4) suggest that there is a positive link between CSR awareness and CA beliefs (1 = .16, p b .01) and this positive relationship is stronger when consumers make greater intrinsic attributions (4 = .04, p b .01). Additionally, there is a positive relationship between intrinsic attributions and CA beliefs (2 = .25, p b .01).6 Comparing
6 We conducted a simple slope analysis (Aiken & West, 1991) to examine the relationship of CSR awareness and intrinsic attributions to CA beliefs. Our analysis shows that there is a positive relationship between CSR awareness and CA beliefs across the range of intrinsic attributions, and a positive relationship between intrinsic attributions and CA beliefs across the range of CSR awareness.

Stonyfield to Dannon/Yoplait (model 5), we find a more positive relationship between intrinsic attributions and CA beliefs for Stonyfield than for its competitors (3 = .31, p b .01; 6 = .09, p b .01). No brand-specific difference in the CSR awarenessCA beliefs link was found. Model 6 allows us to compare Stonyfield to Dannon and to Yoplait separately. We find that there is a more positive link between CSR awareness and CA beliefs for Stonyfield than for Dannon (6 = .10, p b .10), and a more positive link between intrinsic attributions and CA beliefs for Stonyfield than for Dannon (8 = .08, p b .10) or Yoplait (9 = .08, p b .05). In sum, we find supportive evidence that there are brand-specific spillover effects of intrinsic attributions and CSR awareness on CA beliefs. 2.6.2. Consequences of CSR beliefs H4 proposes that consumers of a CSR brand (i.e., Stonyfield) are more likely than consumers of non-CSR brands (i.e., Dannon and Yoplait) to identify with, be loyal to and be advocates for their respective brands. To test this hypothesis, we use a series of difference of means tests. A key issue pertains to how we operationally define the consumers of a brand. Specifically, consumers of a brand could be operationalized as (i) those who reported buying the particular brand most frequently (i.e., frequent buyers), (ii) those who bought the brand most recently (i.e., recent buyers), and/or (iii) those who think of the brand when purchasing yogurt (i.e., brand salience). As the results

Author's personal copy

S. Du et al. / Intern. J. of Research in Marketing 24 (2007) 224241

235

from these alternative operationalizations are similarly supportive of the hypothesis (see Table 6), we detail below only the results with respect to the frequent buyers of each brand. Notably, we distinguish frequent buyers from loyal consumers of a brand. Whereas frequent purchase of a brand can be a result of behavioral inertia or promotional activity, loyalty, at least as conceptualized in this paper, reflects consumers' psychological affiliation with, or commitment to a brand (Jacoby & Chestnut, 1978; Oliver, 1999). As stated before, of the 1062 respondents, 410 buy Dannon most frequently, 130 buy Stonyfield most frequently, and 262 buy Yoplait most frequently. Because the tests for identification, loyalty, and advocacy are unaffected by the inclusion of those respondents who frequently purchase more than one brand, we discuss the results from data that includes them. Consistent with H4 (Table 6), frequent buyers of Stonyfield identify more with Stonyfield (M = 3.66) than frequent buyers of Dannon/Yoplait, do with their respective brands (M = 3.13, p b .01). Frequent buyers of Stonyfield are also more loyal to Stonyfield (M = 4.32) than frequent buyers of Dannon/Yoplait to their respective brands ( M = 3.87, p b .01). Similarly, frequent buyers of Stonyfield engage in more advocacy behaviors toward Stonyfield ( M = 4.20) than do frequent buyers of Dannon/Yoplait toward their respective brands (M = 3.94, p b .01). Separate comparisons of Stonyfield vs. Dannon and Stonyfield vs. Yoplait yielded comparable results (Table 6). H5 suggests that the positive relationships between consumers' CSR beliefs and (a) CC identification, (b) loyalty,

and (c) advocacy will be stronger for Stonyfield than for Dannon or Yoplait. To test these predictions, we ran multiple regressions for identification, loyalty and advocacy behaviors with CSR beliefs as the independent variable. Because these behaviors are surely based on CA beliefs as well, this variable was also included as a predictor in all analyses. Specifically, we ran the following regression models: Identification Intercept b1 CA Beliefs b2 CSR Beliefs e Intercept b1 CA Beliefs b2 CSR Beliefs e Intercept b1 CA Beliefs b2 CSR Beliefs e:

Loyalty

Advocacy

The brand-specific extensions of the above models (i.e., models 1015) are described in Appendix B. As expected, there exist positive relationships, overall, between CSR beliefs and CC identification (2 = .44, p b .05), loyalty (2 = .14, p b .05) and advocacy (2 = .08, p b .05). H5 also captures the advantage predicted to accrue to a CSR brand over its competitors in terms of consumers' relational behaviors. The results of regression (10)(15) (Table 7) provide support for this hypothesis. As before, given that

Table 6 Comparison of identification, loyalty, and advocacy: means and standard deviations Means (S.D.) Stonyfield Dannon
a, b

T statistics Yoplait n = 262 3.11 (.90) 3.97 (.90) 4.01 (.62) Stonyfield vs. Dannon & Yoplait (pooled) Stonyfield vs. Dannon Stonyfield vs. Yoplait

Consumers of a brand = frequent buyers n = 130 n = 410 Identification 3.66 ( 1.00) 3.14 (.81) Loyalty 4.32 (.92) 3.81 (.95) Advocacy 4.20 (.69) 3.90 (.65)

6.40 5.35 4.18

6.07 5.35 4.60

5.48 3.67 2.73

Consumers of a brand = those who bought the brand most recently a, b n = 148 n = 439 n = 278 Identification 3.80 (1.06) 3.13 (.79) 3.08 (.92) 4.60 Loyalty 4.14 (.96) 3.76 (.98) 3.86 (.97) 3.95 Advocacy 4.12 (.71) 3.89 (.64) 3.95 (.67) 3.50 Consumers of a brand = those who think of the brand when purchasing yogurt a n = 264 n = 779 n = 603 Identification 3.50 (1.01) 3.03 (.79) 3.06 (.84) 7.98 Loyalty 3.83 (1.00) 3.40 (1.01) 3.45 (1.03) 5.98 Advocacy 4.04 (.65) 3.72 (.66) 3.79 (.67) 6.33

4.17 4.16 3.68

4.10 2.92 2.50

.67 5.95 6.69

6.67 5.09 4.92

p b .01. p b .05. a Some respondents purchased more than one brand most frequently and/or most recently. Similarly, some respondents indicated that multiple brands came to mind when purchasing yogurt (i.e., chose 4 [agree] or 5 [agree strongly]). Since the results are similar with or without such respondents, the results reported here are based on data that include these respondents. b The cell sizes do not add up to 1062 because some respondents chose Other on items Which brand did you purchase most frequently? and Which brand did you purchase most recently?

Author's personal copy

236

S. Du et al. / Intern. J. of Research in Marketing 24 (2007) 224241

Table 7 Brand-specific effects in consequences of CSR beliefs: unstandardized regression coefficients (T statistics) a, b Identification Independent variables Model 10 Model 13 Loyalty Model 11 Model 14 Advocacy Model 12 Model 15

Stonyfield vs. Dannon & Stonyfield vs. Dannon Stonyfield vs. Dannon & Stonyfield vs. Dannon & Stonyfield vs. Dannon & Stonyfield vs. Yoplait (pooled) & vs. Yoplait Yoplait (pooled) vs. Yoplait Yoplait (pooled) Dannon & vs. Yoplait .30 (8.01) .52 (11.85) .05 (1.13) .13 ( 2.31) .08 (1.48) .03 (.55) .14 ( 2.00) .12 ( 1.93)+ .23 (5,3180) 190.18 .23 (8,3177) 118.94 .39 (5,3180) 410.73 .30 (8.01) .52 (11.85) .85 (19.36) .26 (5.15) .09 (1.75)+ .15 ( 2.37) .10 (1.64)+ .08 (1.33) .17 ( 2.19) .13 ( 1.80)+ .39 (8,3177) 256.69 .55 (5,3180) 787.64 .85 (19.36) .26 (5.15) .78 (28.28) .15 (4.62) .06 (1.73)+ .09 ( 2.09) .04 (1.07) .07 (1.89)+ .10 ( 1.92)+ .08 ( 1.64)+ .55 (8,3177) 492.12 .78 (28.27) .15 (4.62)

CA beliefs CSR beliefs CA beliefs D CSR beliefs D CA beliefs D1 CA beliefs D2 CSR beliefs D1 CSR beliefs D2 Adjusted R2 d.f. Model F

p b .01; p b .05; +p b .10. a D = 1 if brand is Dannon or Yoplait, D = 0 if brand is Stonyfield. b D1 = 1 if brand is Dannon, else D1 = 0; D2 = 1 if brand is Yoplait, else D2 = 0.

D = 0 for Stonyfield and D = 1 for Dannon/Yoplait, the negative coefficients for the interaction term CSR beliefs D suggest that the relationships between CSR beliefs and identification (5 = .13, p b .05), loyalty (5 = .15, p b .05), and advocacy (5 = .09, p b .05) are stronger for Stonyfield than for Dannon/ Yoplait. Comparisons of Stonyfield to Dannon and Yoplait separately (i.e., models 1315) support H5 as well. Again, given that Stonyfield is the base brand, the overall positive coefficients for CSR beliefs and the negative coefficients for the brand-specific interaction terms suggest that the relationship between CSR beliefs and identification is stronger for Stonyfield than for Dannon (7 = .14, p b .05) or Yoplait ( 8 = .12, p b .10). Similarly, the relationship between CSR beliefs and consumer loyalty is stronger for Stonyfield than for Dannon (7 = .17, p b .05) or Yoplait ( 8 = .13, p b .10). Finally, the relationship between CSR beliefs and advocacy is stronger for Stonyfield than for its individual competitors, Dannon (7 = .10, p b .10) and Yoplait ( 8 = .08, p b .10). In sum, these individual comparisons reveal differences in the predicted direction, particularly for Stonyfield vs. Dannon. Finally, we tested the effects of CSR on brand choice by estimating a multinomial logit model (Guadagni & Little, 1983) with CSR beliefs, CA beliefs and the brand-specific constants as explanatory variables. We define the probability of individual i choosing brand k as: Pik expUik = X
k

P where Uik j bjk xjik , and xjik = observed value of attribute j of brand k for customer i, and bjk = utility weight of attribute j of brand k. In line with prior research, CSR beliefs are positively related to likelihood of brand choice ( = .62, p b .05). Not surprisingly, CA beliefs are a significant predictor of brand choice as well ( = 4.80, p b .05). To investigate brand-specific differences in the CSR beliefsbrand choice relationship, we ran a multinomial logit model with D (as defined previously), CSR beliefs, CA beliefs, as well as the interaction of these beliefs measures with D as predictors. Notably, there were no brand-specific effects of CSR beliefs on choice. Individual comparisons of Stonyfield vs. Dannon and Stonyfield vs. Yoplait (i.e., multinomial logit model with D1, D2, CSR beliefs, CA beliefs, as well as the interaction of these beliefs measures with D1, D2 as predictors) yield similar results. In sum, compared to its competitors, the CSR brand, Stonyfield, seems to be reaping stronger relational benefits (i.e., identification, loyalty, and advocacy) but not transactional benefits (i.e., brand choice) from its CSR actions. 3. Discussion This research taps into the CSR-related thoughts, feelings and behaviors of consumers in the real marketplace to implicate a brand's competitive positioning on CSR as a key determinant of consumers' reactions to its CSR initiatives. Overall, the findings (Table 2) suggest that the market-based benefits of brandspecific CSR initiatives vary depending on the extent to which such initiatives are an integral part of the brand's positioning

expUik

16

Author's personal copy

S. Du et al. / Intern. J. of Research in Marketing 24 (2007) 224241

237

relative to its competitors. Several implications follow from our study findings. 3.1. The role of CSR positioning Virtually all prior CSR research in the consumer domain has focused on a single brand or company. By documenting brandspecific differences in consumers' reactions to CSR, this paper identifies an important moderator of such reactions in the real marketplace, where more and more companies are engaging in some form of CSR but with varying degrees of integration with their core competencies. Our findings suggest that the same CSR actions may elicit different reactions from consumers depending on the extent to which these are viewed as isolated endeavors vs. the very basis of the brand. In particular, we demonstrate three ways in which the strategic rewards of CSR for a brand that successfully positions itself along this dimension are greater than for those that merely engage in CSR, even when such engagement is known and liked (e.g., Yoplait's fight against breast cancer). First, consumers tend to have more favorable beliefs, make more charitable attributions and, consequently, reward CSR brands to a greater extent in terms of relational behaviors such as loyalty and advocacy. Second and perhaps more interesting, we observe that consumers are more responsive or sensitive to their CSR beliefs about a CSR brand in that an increase in consumers' favorable CSR beliefs is associated with greater incremental dividends in terms of loyalty and advocacy behaviors for such a brand than for its competitors. Finally, CSR brands benefit from a stronger spillover from their CSR actions to CA beliefs suggesting that perceptions of mainstream attributes such as quality and value can be favorably influenced to a greater extent when social responsibility is an integral aspect of what the brand represents to consumers. Notably, our findings also provide some support for the uniqueness of CSR as a positioning strategy in garnering these customer-driven benefits. Compared to the CSR brand in our study, Stonyfield, the CA-positioned brand, Dannon, not only reaps fewer rewards from its CSR activities but also does not appear to reap additional CA-based advantages. Specifically, our brand-specific analysis of the relationship of consumers' CA beliefs to key relational outcomes such as identification and advocacy (Table 7, models 13, 14, and 15) suggests that compared to CSR positioning (i.e., Stonyfield), CA positioning (i.e., Dannon) does not provide significant incremental returns to identification, advocacy or loyalty. In other words, consistent with our expectation, because a CSR positioning truly reveals the fundamental values, or character of a brand, it is more unequivocally associated with the coveted deeper, more long-term relationships with consumers than other CAbased positioning strategies such as product leadership. While a more thorough contrast of specific positioning strategies awaits further investigation, our findings complement research on customer-company identification (Bhattacharya & Sen, 2003; Lichtenstein et al., 2004) that highlights the potential relational benefits of an identity-revealing CSR focus.

3.2. Relational vs. transactional benefits of CSR Our findings point to the primacy of company-specific longterm relational behaviors, such as brand loyalty and advocacy, over the transactional behavior of product purchase as the basic strategic dividend of CSR. Specifically, while we find that CSR beliefs are indeed positively associated with higher purchase levels, the advantages accruing to the CSR brand, Stonyfield, are in the relational domain and not the transactional one (i.e., purchase). This suggests that CSR positioning is less a shortterm sales generating mechanism as it is one that deepens customer relationships over time, creating brand advocates or champions. In other words, companies interested in using CSR to increase sales may need to be less concerned with their CSRrelated positioning as might companies that are interested in broader, deeper and longer-term consumer responses. In that sense, this research suggests that the real value of a CSR positioning may stem from its contribution to a company/ brand's long-term reputational capital. For such brands, CSR may be more like a long-term insurance policy that is there when you need it (Klein & Dawar, 2004, page 215) rather than one that just contributes to profitability. More generally, the finding that CSR initiatives, particularly when part of a brand's positioning, have the potential to transform consumers into long-term advocates of the brand has important implications for relationship marketing. Despite heavy investments in customer relationship management, the relationships between companies and consumers continue to be troubled (e.g., Dowling, 2002). For instance, Fournier, Dobscha, and Mick (1998) observed that relationship marketing, as it is currently practiced, has, instead of bringing customers closer to companies, driven them farther apart. They thus exhort companies to build relationships that are not empty, meaningless, or stressful. Our findings suggest that CSR might be one such way to build meaningful, long-term relationships: it satisfies consumers' essential self-definitional and selfenhancement needs, turning them from mere buyers of a brand into its champions. Relatedly, recent research suggests that the maximization of customer equity (Rust, Lemon, & Zeithaml, 2004), defined as the sum of lifetime values of all of a firm's customers, across all the firm's brands, may be a more meaningful strategic objective for a company than that of brand equity (i.e., the sum of customers' assessments of a brand's offerings above and beyond its objectively perceived value). To the extent that CSR promotes the relational outcomes examined in this paper, often at the company rather than the brand level, it is likely to be a key contributor to customer equity. Such a relationship is, in fact, suggested by Rust et al. (2004) in their conceptualization of the three determinants of customer equity (page 118). Specifically, they suggest that CSR contributes to brand equity, which along with value equity (i.e., objectively considered quality, price and convenience of the offering) and relationship equity (i.e., customers' switching costs from learning curves, user-community benefits and other considerations) determines customer equity. Our findings suggest that when a brand is positioned on CSR, its efforts in this domain may contribute to relationship

Author's personal copy

238

S. Du et al. / Intern. J. of Research in Marketing 24 (2007) 224241

equity as well, making it more difficult for customers to switch to competitors they do not feel as committed to. 3.3. Managerial implications 3.3.1. Positioning on CSR From a practitioner standpoint, CSR positioning may not be realistic in today's market, particularly for brands that target a broad segment of the population (e.g., the dominant brands in a category). However, it is interesting to note that brands that do so may also benefit in the CA domain, as is evidenced in the stronger spillover of CSR-related attributions and knowledge into CA for Stonyfield relative to its competitors. In this sense, even though CSR may position a brand more emotionally than cognitively (Mahajan & Wind, 2002) going straight to the heart rather than providing a product solution or benefit that none of its competitors provide such positioning may grant the brand a cognitive edge as well, enhancing consumers' beliefs regarding its ability to deliver key functional benefits. Thus, if and when CSR becomes a more prominent element of consumers' consciousness of the marketplace, marketers may do well to consider a CSR positioning if it makes sense based on factors such as product/market definition, competitive structure, product life cycle and consumer knowledge levels (Punj & Moon, 2002). 3.3.2. Communicating CSR For firms that do want to derive the most benefits from their CSR investments, our findings confirm the growing sense that while awareness of a brand's CSR initiatives is a key prerequisite for CSR beliefs, consumers' awareness of such initiatives outside of experimental contexts is rather low. In other words, CSR awareness, or the lack thereof, is a key stumbling block for companies looking to reap strategic benefits from their CSR initiatives (Bhattacharya & Sen, 2004). This points to a key deficiency in most CSR strategies (i.e., the lack of effective communication); since minimal benefits are likely to accrue to companies if their target market is unaware of their CSR initiatives, they need to work harder to increase CSR awareness. Furthermore, our findings regarding the roles of intrinsic and extrinsic attributions in shaping CSR beliefs suggest that companies need to also work smarter in communicating their CSR initiatives, ensuring that consumers attribute such efforts to intrinsic (e.g., genuine concern) rather than extrinsic (e.g., profit motive) motivations. Interestingly, our findings suggest that when a brand is positioned on CSR, consumers are not only likely to be more aware of its activities, but also make more charitable attributions regarding the company's motivations. However, there is a potential downside as well: consumers' reactions to a CSR brand are also more sensitive to their attributions regarding its CSR actions. Thus, a CSR company has more to lose than its competitors if consumers attribute its involvement to extrinsic rather than intrinsic motivations. In sum, companies looking to be more strategic in their CSR decisions need to ask several key questions, such as: What do we want to achieve through our CSR efforts (i.e., sales vs. long-term reputation)? How do we differentiate our CSR initiatives from

those of our competitors? How can we better integrate our CSR efforts with our core competencies? While most companies' CSR efforts are diffuse and unfocused (Porter & Kramer, 2002), our research suggests that from a consumer response standpoint, the best CSR strategies are those that are well integrated into a company's core business activities. As Smith (2003) said, developing the right CSR strategy requires an understanding of what differentiates an organizationits mission, values, and core business activities from its competitors. The presence of strongly held, favorably evaluated CSR beliefs that are unique to the brand is critical to the success of a CSR strategy. 3.4. Limitations and future research directions Our study has several limitations. First, it does not allow us to establish causal links between the focal constructs. While some of the theoretical links such as the one between CSR beliefs and identification have been supported in prior experimental research, other links, like the relationship between CSR beliefs and consumer advocacy, need corroboration in future experimental studies that control for confounds and alternative explanations. Second, although our common-method variance test indicates that common-method bias is not a problem in our study, future research should obtain data on consumers' actual behaviors (i.e., purchase, actual word-of-mouth, actual supportive behaviors during periods of negative media reports, etc.) rather than self-reported behaviors or behavioral intentions, allowing us to examine the effects of CSR on actual consumer advocacy behaviors. More generally, our study was conducted in a category where most brands engage in some form of CSR. The generalizability of our findings is contingent on validation of our findings in product categories with greater CSR diversity. This study also gives rise to a number of future research opportunities. First, a central theme from this study is the potential of CSR initiatives to build strong consumer relationships and generate relational consumer behaviors. Given the growing prominence of CSR on the corporate agenda and the emphasis on marketing productivity, a notable stream of additional research would be to examine the returns on CSR investment. One fruitful avenue for doing so might lie in using Rust et al.'s (2004) customer equity framework. More specifically, our study established a moderating role of competitive positioning, but did not delve into its underlying mechanisms. Further research on when and why CSR brands are likely to reap disproportionate rewards from their CSR activities is extremely germane. Future research needs to also focus on how our findings play out with brands (Camel, Wal-Mart) or categories (e.g., gasoline, cigarettes, fast food) that are increasingly viewed as bad citizens. Acknowledgements The authors thank the Hitachi Foundation for its generous support of this research and all the managers and researchers who have provided helpful comments at various conferences worldwide.

Author's personal copy

S. Du et al. / Intern. J. of Research in Marketing 24 (2007) 224241

239

Appendix A
Measures and descriptive statistics Brand Choice (1=Dannon, 2=Stonyfield Farm, 3=Yoplait, 4=Other) Which of the following yogurt brands did you purchase most recently? Brand Most Frequently Bought (1=Dannon, 2=Stonyfield Farm, 3=Yoplait, 4=Other) Which of the following yogurt brands did you purchase most frequently? Loyalty (1=strongly disagree, 5=strongly agree) M = 3.00 S.D. = 1.08 I am loyal to this brand. Advocacy (1=strongly disagree, 5=strongly agree) M = 3.42 S.D. = .79 = .77 I would like to try new products introduced under this brand name. I talk favorably about this brand to friends and family. If the maker of this brand did something I didn't like, I would be willing to give it another chance. Identification (1=strongly disagree, M = 2.93 S.D. = .82 5=strongly agree) My sense of this brand matches my sense of who I am. CSR beliefs (1=strongly disagree, M = 3.18 S.D. = .53 r = .55 5=strongly agree, 6=N/A) This brand is a socially responsible brand. This brand has made a real difference through its socially responsible actions. CA Beliefs (1=strongly disagree, 5=strongly agree) M = 3.60 S.D. = .68 = .83 Overall, this brand is of high quality. This brand tastes good. This brand is very nutritious. This brand provides good value for money. CSR Awareness (1=strongly disagree, M = 3.26 S.D. = .69 5=strongly agree, 6=N/A) This brand works for XX (the brand's CSR initiative) CSR Attributions (1=strongly disagree, 5=strongly agree) Intrinsic Attributions M = 3.46 S.D. = .74 This brand works for XX (the brand's CSR initiative) because it is genuinely concerned about being socially responsible. Extrinsic Attributions M = 3.36 S.D. = .75 This brand works for XX (the brand's CSR initiative) because it feels competitive pressures to engage in such activities. CSR Support (1=do not support at all, M = 3.25 S.D. = 1.24 5=strongly support) To what extent do you support XX (the CSR issues that these three brands work for)

CA Beliefs Intercept b1 D1 b2 D2 b3 CSR Awareness b4 Intrinsic Attributions b5 Extrinsic Attributions b6 CSR Awareness D1 b7 CSR AwarenessD2 b8 Intrinsic AttributionsD1 b9 Intrinsic AttributionsD2 b10 Extrinsic Attributions D1 b11 Extrinsic Attributions D2 b12 CSR AwarenessIntrinsic Attributions b13 CSR AwarenessExtrinsic Attributions b14 CSR AwarenessIntrinsic AttributionsD1 b15 CSR AwarenessIntrinsic AttributionsD2 b16 CSR AwarenessExtrinsic AttributionsD1 b17 CSR AwarenessExtrinsic AttributionsD2 e 6 Consequences of CSR beliefs
a,b

Identification Intercept b1 D b2 CA Beliefs b3 CSR Beliefs b4 CA BeliefsD b5 CSR BeliefsD e Loyalty Intercept b1 D b2 CA Beliefs b3 CSR Beliefs b4 CA BeliefsD b5 CSR BeliefsD e Advocacy Intercept b1 D b2 CA Beliefs b3 CSR Beliefs b4 CA BeliefsD b5 CSR BeliefsD e Identification Intercept b1 D1 b2 D2 b3 CA Beliefs b4 CSR Beliefs b5 CA Beliefs D1 b6 CA BeliefsD2 b7 CSR Beliefs D1 b8 CSR Beliefs D2 e Loyalty Intercept b1 D1 b2 D2 b3 CA Beliefs b4 CSR Beliefs b5 CA Beliefs D1 b6 CA BeliefsD2 b7 CSR Beliefs D1 b8 CSR Beliefs D2 e Advocacy Intercept b1 D1 b2 D2 b3 CA Beliefs b4 CSR Beliefs b5 CA Beliefs D1 b6 CA BeliefsD2 b7 CSR Beliefs D1 b8 CSR Beliefs D2 e
a

10

11

12

Appendix B Brand-specific models Determinants of CA beliefsa,b CA Beliefs Intercept b1 D b2 CSR Awareness b3 Intrinsic Attributions b4 Extrinsic Attributions b5 CSR Awareness D b6 Intrinsic AttributionsD b7 Extrinsic AttributionsD b8 CSRAwarenessIntrinsic Attributions b9 CSR AwarenessExtrinsic Attributions b10 CSR AwarenessIntrinsic AttributionsD b11 CSR AwarenessExtrinsic AttributionsD e 5

13

14

15

D = 1 if brand is Dannon or Yoplait, D = 0 if brand is Stonyfield. b D1 = 1 if brand is Dannon, else D1 = 0; D2 = 1 if brand is Yoplait, else D2 = 0.

Author's personal copy

240

S. Du et al. / Intern. J. of Research in Marketing 24 (2007) 224241 Glorieux-Boutonnet, A. (2004). Can the environment help boost your marketing? Australian Journal of Dairy Technology, 59(2), 8589. Guadagni, P. M., & Little, J. D. C. (1983). A logit model of brand choice calibrated on scanner data. Marketing Science, 2(3), 203238. Hoeffler, S., & Keller, K. L. (2002). Building brand equity through corporate societal marketing. Journal of Public Policy and Marketing, 21(1), 7889. Hunt, S. D. (1990). A commentary on an empirical investigation of a general theory of marketing ethics. Journal of the Academy of Marketing Science, 18(2), 173177. Jacoby, J., & Chestnut, R. W. (1978). Brand loyalty: Measurement and management. New York: Wiley. Keller, K. L. (1993). Conceptualizing, measuring, and managing customerbased brand equity. Journal of Marketing, 57, 122. Klein, J., & Dawar, N. (2004). Corporate social responsibility and consumers' attributions and brand evaluations in a product-harm crisis. International Journal of Research in Marketing, 21, 203217. KPMG International Survey of Corporate Responsible Reporting. (2005). Accessed August 03 2005, available at http://www.kpmg.com/Rut2000_ prod/Documents/9/Survey2005.pdf Lichtenstein, D. R., Drumwright, M. E., & Braig, B. M. (2004). The effects of corporate social responsibility on customer donations to corporate-supported nonprofits. Journal of Marketing, 68(4), 1632. Luo, X., & Bhattacharya, C. B. (2006). Corporate social responsibility, customer satisfaction, and market value. Journal of Marketing, 70(4), 118. Mahajan, V., & Wind, Y. (2002). Got emotional product positioning? Marketing Management, 11(3), 3641. Maignan, I., & Ferrell, O. C. (2004). Corporate social responsibility and marketing: An integrative framework. Journal of the Academy of Marketing Science, 32(1), 319. McWilliams, A., & Siegel, D. (2001). Corporate social responsibility: A theory of the firm perspective. Academy of Management Review, 26(1), 117127. Menon, S., & Kahn, B. E. (2003). Corporate sponsorships of philanthropic activities: When do they impact perception of sponsor brand? Journal of Consumer Psychology, 13(3), 316327. Miller, J. (2006). Online marketing research. In R. GRover & M. Vriens (Eds.), The handbook of marketing research (pp. 110131). Thousand Oaks, CA: Sage Publications. Morgan, R. M., & Hunt, S. D. (1994). The commitmenttrust theory of relationship marketing. Journal of Marketing, 58(3), 2038. Nesbitt, R. E., & Ross, L. (1980). Human inference: Strategies and shortcomings of social judgments. Englewood Cliffs, NJ: Prentice Hall. Oliver, R. L. (1999). Whence consumer loyalty? Journal of Marketing, 63(4), 3344. Park, C. W., Jaworski, B. J., & MacInnis, D. J. (1986). Strategic brand concept Image management. Journal of Marketing, 50(4), 135145. Peloza, J. (2006). Using corporate social responsibility as insurance for financial performance. California Management Review, 48(2), 5272. Podsakoff, P. M., MacKenzie, S. B., Lee, J., & Podsakoff, N. P. (2003). Common method biases in behavioral research: A critical review of the literature and recommended remedies. Journal of Applied Psychology, 88(5), 879903. Podsakoff, P. M., & Organ, D. (1986). Self-reports in organizational research: Problems and prospects. Journal of Management, 12, 531543. Porter, M. E., & Kramer, M. R. (2002). The competitive advantage of corporate philanthropy. Harvard Business Review, 80(12), 5768. Punj, G., & Moon, J. (2002). Positioning options for achieving brand association: A psychological categorization framework. Journal of Business Research, 55(4), 275283. Rummell, T. (1999). What's new at the body shop? Global Cosmetic Industry, 165(5), 1618. Rust, R. T., Lemon, K. N., & Zeithaml, V. A. (2004). Return on marketing: Using customer equity to focus marketing strategy. Journal of Marketing, 68(1), 109127. Sen, S., & Bhattacharya, C. B. (2001). Does doing good always lead to doing better? Consumer reactions to corporate social responsibility. Journal of Marketing Research, 38(2), 225243. Sen, S., Bhattacharya, C. B., & Korschun, D. (2006). The role of corporate social responsibility in strengthening multiple stakeholder relationships. Journal of the Academy of Marketing Science, 34(2), 158166.

References
Aaker, D. A. (1996). Measuring brand equity across products and markets. California Management Review, 38(3), 102120. Aiken, L. S., & West, S. G. (1991). Multiple regression: Testing and interpreting interactions. Thousand Oaks, CA: Sage Publications. Batson, D. C. (1998). Altruism and prosocial behavior. In D. T. Gilbert, S. T. Fiske, & G. Lindzey (Eds.), The handbook of social psychology (pp. 282316). New York, NY: McGraw-Hill. Bergami, M., & Bagozzi, R. P. (2000). Self-categorization, affective commitment and group self-esteem as distinct aspects of social identity in the organization. British Journal of Social Psychology, 39(4), 555577. Bhattacharya, C. B., & Sen, S. (2003). Consumercompany identification: A framework for understanding consumers' relationship with companies. Journal of Marketing, 67(2), 7688. Bhattacharya, C. B., & Sen, S. (2004). Doing better at doing good: When, why and how consumers respond to corporate social initiatives. California Management Review, 47(1), 924. Brown, T. J., & Dacin, P. (1997). The company and the product: Corporate beliefs and consumer product responses. Journal of Marketing, 61(1), 6884. Brown, T. J., Dacin, P., Pratt, M. G., & Whetten, D. A. (2006). Identity, intended image, construed image, and reputation: An interdisciplinary framework and suggested terminology. Journal of the Academy of Marketing Science, 34(2), 99106. Campbell, M. D., & Kirmani, A. (2000). Consumers' use of persuasion knowledge: The effects of accessibility and cognitive capacity on perceptions of an influence agent. Journal of Consumer Research, 20, 561574. Chaiken, S., & Trope, Y. (1999). Dual-process theories in social psychology. New York: Guilford Press. Chappell, T. (1993). The soul of a business: Managing for profit and the common good. Des Plaines, IL: Bantam. Cone Corporate Citizenship Study (2004). Accessed July 24, 2005, available at http://www.coneinc.com/Pages/pr_30.html Consumer Reports (2002). Cultural Affairs, 67(3), 3235. Dannon Cares (2005). Accessed on October 3, 2005, available at http://www.dannon. com/dn/dnstore/cgi-bin/ProdSubEV_Cat_240865_SubCat_241488_NavRoot_200. htm Dowling, G. W. (2002). Customer relationship management: In B2C markets, often less is more. California Management Review, 44(3), 87104. Dutton, J. M., Dukerich, J. M., & Harquail, C. V. (1994). Organizational images and member identification. Administrative Science Quarterly, 39(2), 239263. Earth Actions. (2005). Accessed on October 3, 2005, available at http://www. stonyfield.com/EarthActions Ellen, P. S., Webb, D. J., & Mohr, L. A. (2006). Building corporate associations: Consumer attributions for corporate socially responsible programs. Journal of the Academy of Marketing Science, 34(2), 147158. Fein, S. (1996). Effects of suspicion on attributional thinking and the correspondent bias. Journal of Personality and Social Psychology, 70(6), 11641184. Fiske, S. T., & Taylor, S. E. (1991). Social cognition. Boston, MA: McGraw Hill. Friestad, M., & Wright, P. (1994). The persuasion knowledge model: How people cope with persuasion attempts. Journal of Consumer Research, 21, 131. Forehand, M. R., & Grier, S. (2003). When is honesty the best policy? The effects of stated company intent on consumer skepticism. Journal of Consumer Psychology, 13(3), 349356. Fournier, S., Dobscha, S., & Mick, D. G. (1998). Preventing the premature death of relationship marketing. Harvard Business Review, 76, 4251. Gatignon, H., & Vosgerau, J. (2005). Moderating effects: The myth of mean centering. INSEAD working paper series. Gillentine, A. (2006, February 17). Timberland to provide nutritional label for all shoes. The Colorado Springs Business Journal, 1. Gilbert, D. T., & Malone, P. S. (1995). The correspondence bias. Psychological Bulletin, 17(1), 2138.

Author's personal copy

S. Du et al. / Intern. J. of Research in Marketing 24 (2007) 224241 Smith, P. A. (2002). A natural reputation. Dairy Field, 185(6), 1. Smith, N. C. (2003). Corporate social responsibility: Whether or how? California Management Review, 45(4), 125. Springer, J. (2006, February 13). Whole foods posts 22% sales growth. Supermarket News, 54(7), 8. Stern, B. B. (2006). What does brand mean? Historical analysis method and construct definition. Journal of the Academy of Marketing Science, 34(2), 216223. Thompson, S. (2001, January 29). Dannon resets lines. Advertising Age, 72(5), 18. Turban, D. B., & Greening, D. W. (1997). Corporate social performance and organizational attractiveness to prospective employees. Academy of Management Journal, 40(3), 658672.

241

Webb, D., & Mohr, L. A. (1998). A typology of consumer response to causerelated marketing: From skeptics to socially concerned. Journal of Public Policy and Marketing, 17(2), 226238. Yoplait is Committed to Fighting Breast Cancer. (2005). Accessed on October 3, 2005, available at http://www.yoplait.com/breastcancer_lids.aspx Zeithaml, V. A., Berry, L. L., & Parasuraman, A. (1996). The behavioral consequences of service quality. Journal of Marketing, 60(2), 3146.