The effect of product-cause fit on consumer attitudes toward
luxury brands’ CSR initiations
Student: Yanqian Deng Student number: 11583584
Submission date: 22th June 2018 - Final version
Track: MSc. in Business Administration - Marketing Track University of Amsterdam
Statement of originality
This document is written by Student Yanqian Deng who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
Abstract
No previous research has examined the impact of product-cause fit on consumers attitudes towards CSR initiatives of luxury brands. This paper aims at investigating to what degree the CSR effectiveness of a luxury brand would be influenced by a high versus low product-cause fit and whether the impact could be neutralized by the perceived altruistic attribution of the brand. The hypotheses are tested by a 2(Brand type: Luxury vs. Non-luxury)×3 (High fit vs. Low fit vs. No CSR) between-subjects factorial experiment.
According to the research findings, luxury brands achieve more favorable attitudes from consumers with the absence of CSR initiative while luxury brands achieve less favorable consumer attitudes in the same case. Moreover, the product-cause fit does play a key role in luxury brands’ CSR initiatives, for a high fit serves to improve the consumers’ attitudes towards non-luxury brands while undermines their evaluation of those luxury ones. Also, it has also been found that the perceived altruistic attribution of the band leaves no significant mediating effects on the reputation of luxury brands. Interestingly, an additional result shows statistically significant mediating effects of the perceived altruistic attribution on the evaluations to non-luxury brands. The results provide valuable insights helpful for both luxury and non-luxury brand marketers to design their CSR initiatives and formulate their marketing strategies.
Keywords: CSR, Luxury brands, Non-luxury brands, Luxury-CSR paradox, CRM, Product-cause
Acknowledgment
My postgraduate studies at the University of Amsterdam are coming to an end this season, but this is just a comma in my life, which marks the beginning of another journey. With the active support of my teachers and friends, the six-month experience of paper wiring has been a hard journey with invaluable gains, and there are a lot of people to whom I would like to express my gratitude.
Among them, my supervisor, Marlene Vock is the one I would like to thank most. Her rigorous style of study, impressive erudition, profound thoughts, and broad vision have lifted me to a higher level. This paper has largely been my own work with she offering appropriate intervention when necessary. As a non-native English speaker, I frequently encountered many grammar issues initially, and it is Dr. Marlene who has always been patient with a smile to help me to improve my writing. Besides, she gave me very detailed and valuable feedback promptly each time on receiving my draft, allowing me to proceed with my work on time. I might not be her best student, but she is my most respected teacher!
Also, I want to thank my family and friends for their consistent accompany and support. Inevitably there are downturns in the course of my writing. They are the ones who help me to bounce back from a low point and regain my confidence!
I am indebted to the generous help offered by all the schoolfellows, friends and relatives as well. They helped me fill out and circulate the questionnaire during the past months. This paper writing work would not be completed without their support!
Table of Contents
Statement of originality... I Abstract...II Acknowledgment...III List of Figures... V 1. Introduction... 1 2.Literature Review...62.1 Corporate Social Responsibility (CSR)...6
2.1.1 Defining CSR... 6
2.1.2Typologies...7
2.2 Luxury and CSR...8
2.2.1 Defining luxury...8
2.2.2 Luxury brands are engaging in CSR initiatives... 10
2.2.3 Luxury-CSR paradox... 12
2.3 Fit...15
2.3.1 Definitions and types of fit... 15
2.3.2 Product-cause fit... 17
2.3.3 Consumer responses to the perceived fit...18
2.4 Perceived altruistic attribution...19
2.4.1 Attribution theory...19
2.4.2 Perceived altruistic attribution and luxury brand...20
3. Methodology...22 3.1 Research design... 22 3.2 Pretest...24 3.2.1 Pretest 1...24 3.2.2 Results of Pretest 1... 24 3.2.3 Pretest 2... 27 3.2.4 Results of Pretest 2... 28 3.3Main Experiment... 29 3.3.1 Questionnaire design... 29 3.3.2 Data Collection...30
4. Analysis and Results...31
4.1Data Preparation...31 4.1.1 Data screening...31 4.1.2 Reliability... 31 4.1.3 Manipulation checks...32 4.2 HypothesesTesting... 34 4.2.1No CSR vs. CSR... 34
4.2.3 Moderated mediation...43
5. Additional Results...45
6. Discussion and Conclusion... 48
6.1 General Discussion...48
6.2 Implications... 52
6.2.1Theoretical implications... 52
6.2.2 Managerial implications... 54
6.3 Limitations and Further Research...54
Reference... 56
List of Appendix
Appendix A Treatments...62Appendix B Pretest 1 ...62
Appendix C Pretest 2 Item-Total statistics...64
Appendix D Condition texts...64
Appendix E Questionnaire... 68
Appendix F Independent Samples T-Test...71
Appendix G Two-way ANOVA H1 Output...72
Appendix H Two-way ANOVA out put H2...73
Appendix I Output of Mediation Model in luxury condition...73
Appendix J Output of Mediation model in non-luxury condition...75
List of Tables
Table 1: Experiment design-all treatments...22Table 2: Reliability Analysis in pretest 1: Cronbach’s Alpha...25
Table 3: Means Overview of all stimuli in pretest 1... 26
Table 4: Independent Samples Test in pretest 1: Luxury scale... 26
Table 5: Independent Samples Test in pretest 1: Fit scale...27
Table 7: Respondents in each condition in main experiment...30
Table 8: Cronbach’s Alpha value... 31
Table 9: Compared means and SDs in brand types (N = 326) and fit levels (N = 222)...33
Group Statistics... 33
Table 10: Means Overview... 34
Table 11: Univariate ANOVA output in H1... 35
Table 12: Descriptive statistics in Luxury/Non-luxury scenarios... 37
Table 13: Contrast test in Luxury condition...38
Table 14: Contrast Test in Non-luxury condition...38
Table 15 Tests of Between-Subjects Effects in relation to consumer attitude... 39
Table 16: Descriptive statistics in high/low fit scenarios...41
Table 17: Contrast test in Luxury brand condition...42
Table 18: Contrast Test in CSR condition...42
Table 19: Mediation model output... 44
Table 20: Total, direct and indirect effect in mediation model...45
Table 21: Mediation model output in non-luxury condition... 46
Table 22: Total, direct and indirect effect in non-luxury condition...47
Table 23: Hypotheses and the related results... 49
List of Figures
Figure 1: Conceptual Framework...21Figure 2: Relationship between brand-type and CSR-type relative to consumer attitude36 Figure 3: Relationship between brand-type and CSR-type relative to consumer attitude36 Figure 4: Relationship between brand-type and fit level relative to consumer attitude.40 Figure 5: Relationship between brand-type and fit level relative to consumer attitude.40 Figure 6: Statistical Mediation Model in luxury condition N =110...44
1.
Introduction
Defined as a commitment to benefit society in a long-term through making contributions by means of corporate resources (Mohr and Webb, 2005), corporate social responsibility (CSR) has been given priority in the global agenda of corporations around the world, including those managing luxury brands. In fact, a variety of luxury brands have already launched their CSR initiatives: i.e., to express sympathy and condolence to the victims of the mass shooting1at a high school in Florida, there has been a $500,000 donation from Gucci in
support of the gun control policy; A cooperation program has been established between Omega and Orbis——an international charity dedicated to restoring sight to the blind around the world; Cartier, Prada, LV partnered with UNICEF on improving availability of education services to African children (Fredette, 2018). CSR initiatives of Luxury brands have attracted considerable attention among the population nowadays.
However, luxury brands are traditionally regarded as incompatible with CSR, for they usually are associated with “elitism” and “self-enhancement”, whereas CSR put emphasis on “altruistic” spirit and thus requires self-transcendence (Kapferer, 2012). According to Schwartz (1992), what lies at the heart of the “self-enhancement” is one's self-interest, which conflicts with the notion of “self-transcendence” ——a term focuses on the benefits of the society as a whole. The two opposing values represent two opposite orientations within a corporation and thus make the consumers more skeptical about its motivation, eventually leading to negative brand evaluations (Torelli, Monga, and Kaikati, 2012). By this “Luxury-CSR paradox” theory, Wong and Dhanesh (2016) thus claimed that “Luxury-CSR initiatives of luxury brands might backfire when the consumers see no altruistic attributions shown by those big
1Mass shooting in Florida: A former student unleashed a hail of gunfire at Marjor Stoneman Douglas High School in
names.
Though the “Luxury-CSR paradox” has been widely recognized by researchers, some still believe in the essential role of CSR in the luxury market. From the consumers’ perspective, Khan and Dhar (2006) found that luxury shopping often provokes a sense of guilt and calls for a justification of the indulgence. Hagtvedt, Henrik and Patrick (2016) then asserted that the negative feeling could be mitigated by being aware of the brand’s CSR initiatives. Moreover, a conclusion has been drawn by Janssen, Vanhamme, Lindgreen, and Lefebvre (2013) that CSR practices prove to be quite suitable for industries like jewelry, as the value of conservation characterized by enduring products like jewelry fits quite well with the underlying value of CSR—— self-transcendence. In their article, Janssen et al., (2013) explained that it is the scarce and ephemeral nature of the luxury goods that make the consumers more apt to regard luxury brands as socially irresponsible, which means a higher level of social responsibility would be ascribed to enduring product brands by consumers because of a higher Luxury-CSR fit (Janssen et al., 2013).
The fit here refers to the degree of perceived compatibility between a company and a cause, and a high-fit means a strong connection between these two entities (Lafferty et al., 2004; 2007). To better investigate the impacts of fit, Bigné et al. (2012) differentiated fit into two dimensions: product-cause fit and image fit. Product-cause fit refers to the similar features shared between a cause and a product whereas image fit puts more emphasis on the general image of the two entities. For example, in TOMS shoes’ well-known plan “buy one give one”, the company promised a new pair of shoes for a child in need with each pair of shoes being sold (TOMS, 2018). It leads to higher perceived product-cause fit because the company’s voluntary donation meets a certain kind of social need in this way.
Most of the previous studies identified a critical role of product-cause fit as positively shaping the consumers’ evaluation of a non-luxury brand. (Menon and Kahn, 2003; Bhattacharya and Sen,2001; Enrique et al., 2012). More specifically, Cornwell et al. (2005) claimed that higher product-cause fit makes consumers more readily recall the CSR activities associated with the relevant brands, and thus contributes to establishing a positive image of the companies. In contrast with Bhattacharya and Sen (2004) who assumed a mediating effect of the fit, Enrique, Rafael, and Joaquín (2012) emphasized that the consumers’ attitudes tend to be more moderate when seeing a product-fit between the brand and CSR activities and found that the higher level of fit would enhance the positive impacts of CSR associations on the generation of a brand image among consumers. Moreover, Bloom et al. (2006) found that the effectiveness of the fit could be explained by how strong have the consumers perceived the company’s internal motivation in its CSR commitments. Unlike external motivation which refers to a desire of maximizing benefits (Amatulli, Caputo and Guido, 2011), internal motivation amounts to intrinsically altruistic concerns over the well-being of others (Ellen et al., 2006). What’s more, Drumwright (2006) conducted research investigating the manager's opinions on cause-related marketing and found that “when there was too close a relationship, managers pointed to cynical reactions from consumers about the company ’ s motive, who perceived it to be exploitative or not altruistic” (p. 71).
However, the effect of product-cause fit on a luxury industry remains unexamined. Luxuries distinguish itself from non-luxury goods by its very nature (e.g., stronger symbolism and experiential benefits), thus compared with non-luxury ones, it is reasonable to expect a different level of product-cause fit between them and their CSR activities (Vickers and Renand, 2003). As it is mentioned above, most of the CSR activities of luxury brands do not fit with its products very well, typical instances of these are the $500,000 donation from
Gucci to support gun control and Cartier, Prada, LV’s partnership with UNICEF on supporting children’s education in undeveloped countries. These cases show the relatively low level of product-cause fit due to the absence of an apparent association between the luxuries and the charities (Lafferty, 2007). Why do luxury brands engage in low fit CSR activities? Will the impact of product-cause fit run differently in luxury market compared with what it does on non-luxury one, where it has been found that higher product-cause fit would contribute to a more positive consumer attitude (Ellen et al., 2006; Becker-Olsen et al., 2006)? This paper would attempt to answer the above questions from a brand new perspective.
To sum up, this thesis aims to answer the following question: How does the high versus low product-cause fit influence consumer attitudes towards luxury brands’ CSR initiatives? Will this effect be mediated by the perceived altruistic attribution?
There are several sub-questions: 1. What is CSR？
2. What are a luxury brand and the Luxury-CSR paradox?
3. What is product-cause fit and why it could moderate the effectiveness of CSR actions? 4. To what degree does product-cause fit leave an impact on the consumers’ attitudes towards luxury brands’ CSR initiatives, compared with what it does in a non-luxury market?
5. Will the effect of product-cause fit on consumers’ evaluative attitudes towards a luxury brand’s CSR initiatives be mediated by the perceived altruistic attribution?
The research questions are highly relevant as little research has been done in the field of CSR of luxury brands. This paper would make theoretical contributions to this issue from several aspects. First, research on the CSR would be supplemented by an empirical study on a
previously unexamined industry—— the luxuries, as a response to the demand for a clearer clarification of Luxury- CSR relationship from Davies et al. (2012). Second, this study would serve as an extension of previous luxury-related literature by initiatively analyzing the effect of perceived product-cause fit. Previous research on luxuries focused more on the symbolic meaning of luxury brands (Kapferer et al., 2014; Wong and Dhanesh, 2017), more specific issues such as counterfeiting (Marticotte & Arcand, 2017) or brand positioning (Bartikowski & Cleveland, 2017) while no article takes the product-cause fit into consideration. This paper aims to specify how could the consumer attitudes towards the luxury brand be moderated by the product-cause fit between luxury brand and its CSR activities.
This study has practical significance as well, for many luxury companies are struggling to maintain the balance between CSR activities and its commercial business. One of the primary purposes of this study is to help luxury marketers find out an optimal strategy when taking part in CSR initiatives. To be specific, the following questions will be addressed with the help of the insights offered by this paper: which approach (high versus low product-cause fit) will allow luxury brands engage in CSR initiatives more effectively?
The following section is the literature review where related notions and theories will be presented in detail consecutively. A comprehensive theoretical framework will be proposed in this chapter with hypotheses being established in the end. The next part consists of the methodology and the results analysis. The final stage of this paper will begin with a discussion of the research findings and draws a brief conclusion with clarification of its theoretical and managerial implications. Limitations of this study will be listed at the end with suggestions for future improvements.
2.Literature Review
2.1 Corporate Social Responsibility (CSR)
2.1.1 Defining CSR
The definition of Corporate Social Responsibility (CSR) was firstly introduced as “a risk management tool for the company” by Schnepp and Bowen (1954) in his book "Social Responsibilities of the Businessman". With the constant development of technology and economy, CSR has acquired new meanings. For instance, as described by the European Commission (2001, p.6), CSR was “a notion whereby companies integrate social & environmental concerns or voluntary behaviors in their core business and their communication with stakeholders”. In 2005, Kotler and Lee referred to CSR as “a promise to improve societal welfare by arbitrary business actions and contributions of company resources” (p.32). Apart from that, CSR, as defined by Heal (2004), was also a voluntary action to cut down externalized costs or to relieve distributional conflicts. Banerjee (2008) even pointed out the following three essential characteristics of CSR activities. First of all, CSR is a commitment regulated via regulations and behaviors of companies; secondly, CSR behaviors transcend the mere requirement of laws; lastly, CSR activities are voluntary. To sum up, CSR is, as defined by Beltratti (2003), the behavior of enterprises in reducing any effect that is harmful to the society and maximizing social welfare in the process of abiding by laws and moral standards, treating employees fairly, protecting environment, and supporting charity.
Although the definitions of CSR are broad, these definitions share a common ground, that is, CSR is about “demonstrating the integration of social and environmental concerns in business operations and in interactions with stakeholders” (Van Marrewijk and Werre, 2003,
p. 107). Consumers, as a major part of stakeholders, are regarded as a critical role in evaluating CSR initiatives of companies, according to Mohr, Webb and Harris (2001). From the perspective of consumers, this paper employs the definition of corporate social responsibility of Mohr et al. (2001), which referred CSR as a company project voluntarily committed to supporting one or more social programs and promoting social welfare for stakeholders (p.47).
2.1.2 Typologies
Several scholars including Bhattacharya and Sen (2004), Seitanidi and Ryan (2007) suggested that CSR covers various categories, each of which represents a separate voluntary action. According to Bhattacharya and Sen (2004), CSR activities involve six dimensions: community support (i.e., helping to solve health and education issues for economically vulnerable groups), diversity (i.e., supporting and appealing for gender equation), employee support (i.e., promoting the well-being of employees), environment (emphasizing the behaviors that are eco-friendly and environment protective), non-U.S. operations (i.e., ensuring fair labor practices) and product (i.e., focusing on product safety, research & development and innovation).
Alternatively, Seitanidi and Ryan (2007) pointed out four main categories of CSR activities: corporate philanthropy, socio-sponsorship, cause-related marketing (CRM), and long-term business & non-profit partnerships. Corporate philanthropy indicates cash or in-kind donations or support that is given as “gifts” rather than an exchange of interests. For example, Gucci donated $500,000 in the support of a gun control organization. Socio-sponsorship is symmetric that a company sponsors in money or in-kind for a project with the purpose of attaining the potential commercial interests (i.e., during the 2014 World Cup, Adidas sponsored the two final teams: Argentina and Germany, and obtained the rights to
design the tournament ball, which helped the company to expand its influence). Cause-related Marketing (CRM) refers to the practice of donating a fixed amount of sales revenue from a product or service to charities or other nonprofit organizations, establishing a win-win relationship between companies and nonprofit organizations. For instance, Cartier would donate to the UNICEF $100 for each purchase of a single bracelet with the cyan-blue cord, and $200 a double bracelet (Fredette, 2018). As for the long-term business & non-profit partnerships, due to the difficulty in solving massive social issues for just one or two organizations, it aims at building a long-term partnership such as knowledge sharing, technological support.( Seitanidi & Crane, 2008).
Varadarajan and Menon (1988) pointed out that Cause-related Marketing (CRM), as an increasingly prevailing strategy among corporate social responsibility (CSR) initiatives, provides consumers with the opportunity to engage in altruistic acts beneficial for a social cause. The fundamental role of the CRM is donating for a particular social cause; consumers’ way of participating the cause and organizations’ trading behaviors are closely linked, as Varadarajan and Menon (1988) identified, the most fundamental aspect is that donations are depending on specific commodity sales. In addition, CRM is an effective way for companies to win, in a short time, a positive social image in practicing social responsibilities (Dickinson and Barker, 2007).
2.2 Luxury and CSR
2.2.1 Defining luxury
Despite numerous attempts to define luxury (e.g., Yeoman and McMahon-Beattie, 2006; Heine and phan, 2011; ), little consensus has been reached about the exact meaning of the luxury since it is manifold and elusive (Kapferer and Michaut-Denizea, 2017). For instance,
Yeoman & McMahon-Beattie (2006) defined “luxuria”, the origin of luxury, as extravagance, sensuality, and pomp, while (Heine and Phan, 2011) identified luxury from the aspects of the extra-ordinariness and ancestral heritage, thus giving luxury a new meaning: “a desirable item more than necessary” (p.109). Scholars summarized that luxury has several core characteristics such as scarcity (Kapferer, 2015), superior quality, historical inheritance, high price (Heine and phan, 2011), as well as craftsmanship and prestige (Janssen et al., 2013). However, there still exists a major difficulty in forming a unified definition of luxury and luxury brands since consumers perceive that “what is luxury to one may just be ordinary to another” (Phau & Prendergast, 1999, p. 123).
Chandon, Laurent, and Valette-Florence (2016) pointed out from the perspective of consumers that the luxury sector is tied to consumers’ point of views, and they highlighted that “luxury brands weaved a dream attached to their name, value and created a symbolized world for consumers”(p. 301).
In existing literature, there is little information regarding to clear differentiation between luxury and non-luxury brand in spite of numerous attempts to define luxury. According to Kapferer (2012b), non-luxury brands can be identified by the exact opposite attributes associated with luxury brands. The reason why luxury differs from non-luxury and functional products is that it owns the extra value of art and beauty (Kapferer, 2009). Tynan et al. (2010) also concluded that goods or services of luxury brands are with top qualities and high prices, and are perceived to be scarce, prestigious and extravagant by consumers. According to Barnier (2012), there are seven common elements of luxury: superior quality, hedonism (contrary to utilization), high price, rarity, selective distribution and personalized products/services, prestige/privilege, and creativity. To differentiate luxury goods from non-luxury ones, these seven criteria are needed (Kapferer, 2012).
Whether a consumer perceives a brand as a luxury one depend partly on abstract concepts—or “unique, abstract meanings” (Park, Milberg, & Lawson, 1991, p.186)—that he or she associates with that specific brand. Additionally, Vanhamme et al. (2015) put forward the following features of luxury brand goods: “made of rare & precious materials”, “unique”, “luxurious”, “expensive”, and “of superior quality”. This article adopts the above indicators. According to Kapferer (2009), the exact opposite attributes that are associated with luxury brands can be applied to non-luxury brands. Therefore, non-luxury products score low in these indicators, and to be specific, they are “made of normal and ordinary materials”, “common”, “not luxurious”, “affordable”, and “of average quality”.
2.2.2 Luxury brands are engaging in CSR initiatives
Brain & Co (2018) argued that since 2009, the luxury industry has been enjoying a remarkable growth rate of over 10%. In 2014, this sector generated a revenue of around $250 billion, making it one of the most robust sectors. However, apart from its economic contributions, its performance in corporate social responsibility (CSR) has also begun to attract public attention. In 2014, BCG surveyed consumers’ perceptions of luxury products, the result of which demonstrates that 13% of the interviewees can associate CSR with luxury products, and this is a sharp contrast to the figure in 2013, which is 5% (Fredette, 2018). Hagtvedt & Patrick (2016) explained that consumers might feel guilty when purchasing luxury products and they regard social cause as a means to reduce that sense of guilt. In this sense, it seems that the question should no longer focus on whether luxury brands should engage in CSR actions but on how to invest in CSR (Janssen et al., 2013).
Previous studies focused on the critical role of CSR in the non-luxury market as a risk management tool (Bowen, 1950), by which reputational risks can be mitigated, and this is
particularly effective through Cause-related Marketing (Heal and Garret, 2004). Epstein (2012) also summarized that CSR could benefit a business in the following six aspects:1. Innovation (i.e., Unilever developed a hair conditioner with less water consumption); 2. Brand differentiation, which constitutes one of the primary reasons of companies embracing CSR (Hur, Kim and Woo, 2013); 3. Cost saving (emphasizing more about sustainability, i.e., after installing energy monitoring meters on Ga. Plants, General Mills saved $600,000); 4. Long-term thinking; 5. Improving consumer engagement (i.e., to arouse consumer resonance, Walmart ran an advertisement regarding to environment protection awareness); 6. Facilitating employee engagement (i.e., establishing sustainability action network to activate employees).
However, according to Bjordal-Aven & Landsvik (2010), CSR initiatives exert influence on consumers not only in the non-luxury market but also in the luxury market, and they affect consumers’ response to products, the image of the company, and the purchase intentions. Kapferer (2017) also points out the importance for luxury brands to engage in CSR initiatives as CSR can be a marketing tool for risk management. He insists that, for instance, even though luxury consumers attach little attention to CSR during their purchase, they may still have a negative response when they hear a dissonant message about luxury brands, such as using child labor or animal’s furs.
In fact, many luxury brands have already taken part in CSR campaigns so far. For example, Bulgari supported the campaign of “rewrite the future”, becoming the largest independent company that defends and advocates children’s right. The company also manufactured a unique ring for this campaign, 20% of the sales of which were donated for charitable purposes. Besides, Burberry’s corporate responsibility web page highlighted its effort in “inspiring sustainable action” and its supporting of “the next generation of creative thinkers” (Buberry, 2018). Luxury brands like Prada, Gucci, Cartier have also been dedicated
to CSR in various forms, such as establishing a partnership with UNICEF and building foundations for the welfare of the society.
However, it is also challenging for luxury bands to carry out CSR initiatives. CSR promotional communication dilemma, as introduced by Coombs and Holladay (2012), points out the fact that stakeholders like consumers are willing to see luxury brands making more achievements in CSR, but they would also worry that these companies have put too much effort or given too many resources for CSR, and thus their core business would be distracted. In some cases, CSR actions may exert a negative influence on luxury brands when consumers started to question the real motive behind CSR programs (Ellen et al., 2004). What is worse, according to Torelli(2012) and Kathuria(2013), the most challenging factor is the luxury-CSR paradox, the detail of which will be explained in the next section.
2.2.3 Luxury-CSR paradox
As it is shown in the last section, there remain tensions between luxury brands and CSR initiatives as a consequence of the Luxury-CSR paradox (Torelli, 2012). A Luxury-CSR paradox refers to a value conflict between the spirit of CSR and the orientations of luxury brands (Kathuria, 2013). A variety of values underlying the activities of luxury companies and CSR initiatives have been identified by previous studies, many of which could be found in Schwartz’s (1992) circular theory of human values.
With respect to the circular theory, Schwartz (1992) identified ten motivational brand values perceived by consumers and then further divided them into four categories: self-enhancement (emphasizing one's own interest), self-transcendence (caring more about others), conservation (focusing on the protection of current status) and openness (provoking innovative ideas and hedonism). He pointed out that conflicts would occur between these
value categories, such as self-enhancement against self-transcendence, and openness against conservation. On the basis of this theory, the relationship between the two competing value groups is termed by Torelli et al. (2012) as the Luxury-CSR paradox. They claimed that the values of self-enhancement and openness emphasized by luxury brands are incompatible with the values of self-transcendence and conservation lying at the heart of CSR activities.
Moreover, Janssen et al. (2013) classified luxury goods into enduring ones ( products like jewelry and diamond, which are not perishable) and ephemeral ones (luxuries with a limited life such as designer clothes) and argued that it is the brands of enduring luxuries that match quite well with CSR activities, contrary to the ephemeral ones. The authors explained that brands of ephemeral luxuries put emphasis on conventions, craftsmanship, long-time producing, and inheritance between generations, which are embodiments of the values of conservation, thus being inconsistent with CSR’s spirit of self-transcendence (Kapferer & Bastien, 2009).
According to Torelli et al. (2012), each brand has its own brand concept, which is usually a term with symbolic meanings and has its origin in Schwartz’s (1992) motivational values. The consumers’ behaviors could be unconsciously motivated and guided by the goals implicit in the brand concept (Park et al., 1991). For luxury brands, it is the desire for prestige or a sense of dominance over others that fuel the consumers’ purchase. While non-profit organizations are driven by the determination to improve the public well-being. The value conflict underlying the brand concept of luxury companies and CSR activities can also be unconsciously situated in the consumers’ minds (Torelli, 2012). When information about its CSR activities are available, the participants of the study conducted by Torelli et al. (2011) showed the more negative evaluation of a luxury brand emphasizing self-enhancement. The research finding shows that the consumer attitudes towards a luxury brand which labels itself
as the pursuer of perfection would not be improved by the company’s engagement in CSR activities.
On the other hand, the awareness of CSR activities shows no effect on the consumers’ evaluation of the brands associated with the spirit of openness or conservation (Janssen et al., 2013). As the idea of self-enhancement is commonly represented by luxury brands (Kathuria, 2013). The conflicts between the value orientations of CSR activities and luxury brands would become apparent in this research, which, however, is not a problem for non-luxury companies. Therefore, on the basis of the study conducted by Torelli et al. (2012), the following hypotheses are established:
H1a: Consumers respond to the luxury brand more favorably if there are no CSR initiatives than if there are CSR initiatives.
H1b: Consumers respond to the non-luxury brand less favorably if there are
no CSR initiatives than if there are CSR initiatives "
Though the “Luxury-CSR paradox” has been widely recognized by researchers, some still believe there is a way managing Luxury-CSR paradox in the luxury market. Based on Schwartz’s (1992) circular theory, values can be ascribed to different value categories, i.e., elitism and hedonism are representatives of “self-enhancement” values; equality and universalism belong to “self-transcendence” values; excess and emotions are categorized as “openness” values; moderation and rationality are classified as conservation values. The Luxury-CSR paradox thus can be further differentiated into four categories: elitism-equality paradox, hedonism-universalism paradox, excess-moderation paradox, and emotion-rationality paradox (Wong and Dhanesh, 2017).
Also, it has been found that the Luxury-CSR paradox could be treated with two discursive strategies, namely “harmonious coexistence” and “convergence”. The coexistence
strategy refers to convincing consumers that the core idea of luxuries——self-enhancement, is compatible with that of CSR—— altruistic, to mitigate the tension between hedonism and universalism. For instance, Tod’s——a luxury shoe brand, displays its altruistic concerns by showing its universalistic care for the employees’ well-being. On the contrary, to mitigate the emotion-rationality paradox requires us to adopt the convergence strategy, which is an approach of synthesis, aiming at dissolving differences and integrating opposition (Wong and Dhanesh, 2017). Cause-related marketing is one of the most effective instances of convergence strategy. For example, Omega — — a high-end watch manufacturer, has cooperated with Orbis — — an international charity, to help people with eye diseases and published the number of patients who have received medical treatments on their official website (Wong and Dhanesh, 2017). This article gives the current research a clue that Luxury-CSR paradox may be well managed by cause-related marketing (CRM). Therefore, the essential role of the fit between a social cause and luxury brand will be introduced in next section.
2.3 Fit
2.3.1 Definitions and types of fit
There are a few scholars arguing that stakeholders sometimes want CSR initiatives conducted by companies perfectly match their overall business strategies (Cone, 2007; Hamlin and Wilson, 2004). This “matching” or “similarity” is termed “CSR fit”, that is, the perceived consistency between a company’s brand and social actions (Du et al., 2010; Bhattacharya and Sen, 2004). In a broader view, the notion “fit” is also cited by authors in various terms, such as “link”, “relevancy”, and “congruence” (Fleck and Quester, 2007). Bigné et al. (2012) put forward that if two entities fit each other, their relationship shall be
“congruent”, “compatible”, “complementary”, “make sense” and “logical”. This paper adopts the definition of fit suggested by Bigné et al. (2012) and extends that, a high fit should score high in the above indicators.
Regarding the crucial role of CSR of a corporate, Yuan et al. (2011) classified seven dimensions of fit from “born CSR oriented” to “cooperating”. Born CSR oriented means CSR is put at the priority position of business from the beginning while cooperating means that companies conduct CSR activities through partnering with others. In line with Yuan et al. (2011), Aguinis and Glavas (2012b) divided fit in embedded fit and peripheral fit based on whether the CSR initiatives are embedded or peripheral. Moreover, Simmons and Becker-Olsen (2006) introduced “natural fit” and “created fit”. The former means that there is an evident link between a social cause and a company while the latter is created by the company to make linkage with a cause (i.e. use the same endorsers). Fit can also be distinguished into conceptual fit (based on brand positioning) and perceptual fit (based on appearance such as color, logo.). Additionally, Lafferty et al. (2007) asserted that the source of “fit” can be traced from business partners, i.e., brand-name fit.
According to Bigné et al. (2012), CSR fit originates from numerous aspects such as product associations (e.g., a clothing store donates its clothes to underdeveloped areas short of clothes); company image linkage (e.g., developing alliance with non-profit organizations having similar brand perception or with target groups sharing universal values). Based on different characteristics that a cause is linked to a brand, Berger et al. (2004) categorized fit into nine types, including mission fit, target market fit, product/cause fit, workforce fit, resource fit. For instance, mission fit refers to whether the company shares a common vision or mission with the cause, target market fit is met when two entities have the same target group, workforce fit is embodied in the volunteer support from a business while
product/cause fit refers to a compatible positioning between the products and a cause. Stepping further, Bigné et al. (2012) divided fit into two main categories: functional fit and image fit. The former is similar to product-cause fit which is determined by the compatibility of the product’s features and its use, as well as the cause. In contrast, image fit depends on how compatible the brand reception, perception, and the cause are. This thesis adopts this classification with an in-depth analysis of the critical role of product-cause fit acted in the luxury brands’ CSR initiatives.
2.3.2 Product-cause fit
In line with Bigné et al. (2012), Davis & Guzman (2017) defined product-cause fit as the perceived relatedness of a social cause’s functional needs with a specific product or service. A high product-cause fit indicates a high degree of similarity between the products and the cause (Lafferty, 2007). Tom's "one for one" plan can illustrate this point, in which the company matches every pair of shoes purchased with a new pair of shoes donated for a child in need (Toms, 2018). Another example is FedEx’s support for the distributing supplies in a disaster relief effort. In the above-mentioned cases, the product-cause fit is high since the brand’s products or services contributed to its related social needs. Additionally, Louis Vuitton linked traditional artwork with its products, setting the LV Foundation exhibiting the most valuable collections to art lovers. Defining its products as “heirlooms”, LV highlighted the strong convergence between its products and preservation. The product-cause fit in this case between Art Foundation and LV products is also high.
It appears more meaningful to investigate the effect of product-cause fit instead of image fit regarding luxury brands’ CSR initiatives for the following reasons. First, Luxury brands are commonly thought to be incompatible with the social cause because intrinsically, they are always considered as enhancement” and “excess” in conflict with the values
“self-transcendence” and “altruistic” of CSR. Therefore the image fit between luxury and CSR is low as the image fit refers to the compatibility of the overall image, value, perception between the brand and the cause (Bigné et al. 2012). Second, the product-cause fit however, can be both high and low in luxury-CSR initiations. For instance, unlike its collaboration with UNICEF, Prada's "going to fur-free" plan initiated in 2017 demonstrates a relatively high product-cause fit since consumers attach great attention to whether the clothes that they are to purchase are made by killing wild animals.
2.3.3 Consumer responses to the perceived fit
Fit is of paramount importance since it not only influence brands' images but also consumers’ attitude towards the brand (Menon & Kahn, 2003). As for non-luxury brands, consumers are prone to have a positive response to CSR initiatives when the product-cause fit is high, as numerous studies suggest (Becker-Olsen et al.,2006; Lee et al., 2013; Menon and Kahn, 2003; Sen and Bhattacharya,2001; Enrique et al., 2003). For instance, Tom's "buy one give one" strategy helps it obtain great praise from the public as the company directly donates its products. Furthermore, Becker-Olsen et al. (2006) also found that promotion of high-fit, initiatives improves consumers goodwill towards companies, while promotion of low-fit, profit-motivated initiatives has the opposite effect. Congruence theory also affirms that high fit would exert a positive influence on consumers’ attitudes towards a brand because they would recall the brand’s CSR initiation readily if it has a high fit (Hamlin and Wilson, 2004).
For luxury brands, however, no studies have addressed the effect of the perceived product-cause fit between them and CSR. As stated in the last section, image fit between luxury and CSR is low because of their conflict values. According to luxury-CSR paradox theory (Torelli, 2012; Kathuria, 2013), this low image fit hurts attributions and attitudes of a
luxury brand (Bigné et al., 2012). According to Forehand and Grier (2013), a brand with a low image fit might even arouse suspicion from consumers about whether the company is genuinely altruistic or not. Especially when consumers believe that the company carry out the luxury-CSR campaign with other intentions, the high product-cause fit might even render consumers’ attitudes to the luxury company lower (Yoon et al., 2006). Kekhili, Tagbata and Achabou (2013), through investigating the sustainability of the raw materials, found that luxury consumers dislike the recycled cotton in Herme (a luxury fashion brand) as recycled materials signify the loss of prestige, which is in sharp contrast with “scarcely” feature of luxury, even though the product-cause fit is rather high in this case.
Therefore, the author offers the following hypotheses:
H2: Different levels of product-cause fit interacting with distinct brand type (Luxury and Non-luxury) to influence the effectiveness of CSR initiations.
H2a: A high (low) product-cause fit leads to negative (positive) consumer attitudes towards the luxury brand.
H2b: A high (low) product-cause fit leads to positive (negative) consumer attitudes towards the non-luxury brand.
2.4 Perceived altruistic attribution
2.4.1 Attribution theory
Attribution theory explains that consumers may care less about what a firm is doing than about why it is doing it (Ellen et al., 2006). In other words, consumers try to figure out a company’s motives by interpreting its CSR behavior (Kelley and Michela, 1980). In line with the research of Ellen et al. (2006), many studies suggested that the perceived company motivation in CSR engagements helps explain the relationship between CSR fit and its effectiveness (Korschun, Bhattacharya and Swain, 2006). According to Korschun et al.
(2006), the evaluation of the motives of a brand’s CSR actions depends on the understanding of consumers. Ellen et al. (2006) found that motivations behind a company’s CSR initiative can be divided into two bipolar groups: Self-centered motives (including egoistic motive and strategic motive) and Other-centered motives (comprise stakeholder-driven and values-driven). Moreover, Consumers respond to the strategic motive and stakeholder-driven more favorably while responding to the egoistic and values-driven less favorably (Ellen et al., 2006). By conducting two studies, the authors found that CSR attributions are involved and that companies linked with more strategic & values-driven motives and few egoistic motives will raise more favorable evaluations from consumers (Ellen et al., 2006). Besides, those attributions (motives) mediate the relationship between fit and consumers’ attitude to CSR. In the study 2, the authors investigated to what extent different levels of fit with high versus low commitment would affect consumer responses and found that high fit exerts a win-win effect for both the brand and the cause as high fit leads to more strategic & values-driven motives while fewer egoistic motives (Ellen et al., 2006). However, the authors choose the Shell (a famous gas company) as an example, and this brand is not representative as it is remote from luxury sector.
2.4.2 Perceived altruistic attribution and luxury brand
Many scholars suggest that consumers respond positively towards CSR actions when the perceived motivation is altruistic (Brown & Dacin, 1997). What ’ s more, the perceived altruistic motivation in CSR engagements helps explain the relationship between CSR fit and its effectiveness (Davis and Ahonkhai). However, most studies only investigated in non-luxury brands and whether the perceived altruistic attribution mediates the relationship between fit and consumer attitude towards luxury brands’ CSR initiatives is unknown.
According to Forehand and Grier (2013), a luxury brand might stimulate consumers’ suspicion of the company not being altruistic as the brand owns a low image fit with CSR initiatives. When consumers determined that the motivation of luxury-CSR campaign not be pure, a high product-cause fit might even lower consumers’ attitudes towards the luxury company (Yoon et al., 2006). Drumwright (1996), who also holds similar views with Forehand and Grier (2013), believes that a high fit may generate the perception that the company is taking advantage of the social cause. In this case, a low product-cause fit might be considered more selfless and altruistic. As the current thesis has assumed a low (high) product- cause fit might bring more positive (negative) consumer attitude towards a luxury brand in the last section. Ellen et al. (2006) also stressed that perceived altruistic attribution would lead to favorable evaluations of a brand. As a result, the following hypothesis is generated:
H3: Consumer perceived altruistic attribution mediates the relationship between CSR fit and Consumers’ attitude towards a luxury brand.
The conceptual framework is shown in Figure 1 below.
3. Methodology
3.1 Research design
The purpose of this research is to examine, when the product-cause fit between brand and CSR is high or low, how consumers respond to luxury brands’ CSR activities and whether or not the perceived altruistic attribution mediates the effectiveness of CSR. The following variables are needed to be examined in this experiment: the independent variable, that is, the brand type (luxury brand vs. non-luxury brand); the dependent variable: consumer attitude; the moderator: fit level (high fit, low fit); and the mediator: perceived altruistic attribution. For this study, it is suitable to conduct an experiment since it, by proving the occurrence of the dependent variables when some factors are manipulated, helps to explain the cause-and-effect (Freedman and David, 2009).
This survey will be conducted through a 2 (brand type: luxury vs. non-luxury) × 3 (CSR initiative: no CSR, low fit, high fit) factorial between-subjects experimental design. Apart from that, the no CSR group, as a control group, is free of manipulation. This group is entirely necessary as a comparison with the other two fit level groups. Table 1 displays all the conditions.
Table 1: Experiment design-all treatments
High fit Low fit No CSR
Luxury brand Luxury & High fit Luxury & Low fit Luxury & No CSR Non-luxury brand Non-luxury & High fit Non-luxury & Low fit Non-luxury & No CSR
Even randomization is quite essential since it eliminates the systematic variation between experimental conditions (Field, 2013) In this experiment, in order to remove all possible
effects that might interfere with the final results, all participants are assigned randomly to the above six conditions.
To avoid prior knowledge and bias, during the pretest, all participants were presented with carefully constructed written scenarios about a fictitious shoe brand and its CSR initiatives. The reasons of this study chooses a shoe brand are as follows. Firstly, everyone needs a pair of shoes and they all have experience in the consumption of shoes; secondly, respondents are likely to be familiar with both luxury and non-luxury brands; lastly, the shoe itself has no negative image, to be specific, consumers will not respond negatively to shoes (compare to junk food, for example). Therefore, selecting a hypothesized shoe brand makes in the experiment is more related to consumers’ life and thus more credible. As for CSR initiatives, the author chooses two real non-profit organizations for the cause-related marketing of the fictitious shoe brands. The chosen high-fit-NPO is “ It’s from the sole (IFTS)”, an organization committed to collecting, refurbishing, and offering tons of second-hand shoes and clothing to the homeless. Water.org is the selected as a low-fit NPO dedicated to bringing safe water and sanitation to the world, and to undeveloped areas in particular. The collaboration form is the same in both CSR scenarios, which is described as “‘Manzoni’ plans to design a special collection of shoes and will donate 10% of the sale of each pair to the NPO”.
In sum, the luxury scene and non-luxury scene were of great similar to each other in spite of different description (i.e., price, quality, and craftsmanship.) based on the definition of luxury (Vanhamme et al., 2015). The non-profit organizations in the experiment are the real but they are not well-known so as to reduce bias from the participants.
of pre-tests were conducted to select the optimal set of stimuli.
3.2 Pretest
3.2.1 Pretest 1
The pretest 1 is a within-subjects questionnaire that shows to every participant all treatments. Appendix A shows the stimuli that were presented to the participants in the pretest 1, and 28 respondents completed this survey through Qualtrics online. Moreover, the questionnaire has both English and Mandarin versions so that it could reach to more respondents. The questionnaire begins with a welcome part that introduces the purpose of the research. Luxury is measured by a 6-item 7-point Likert scale, which is introduced by Vanhamme et al. (2015) (e.g., “Manzoni shoes have superior quality/ are made of rare and precious materials/ are unique/ are expensive/ are luxurious/ are desirable”). These questions are repeated in both the luxury brand and the non-luxury brand scenarios. In the following part, there are five questions with five items to measure the product-cause fit level (Beckr-Olsen; Bigné et al.，2002). For instance, “I believe the combination of brand Manzoni and the charity described above is congruent/ complementary/ meaningful/ make sense/ logical.” These questions are also repeated for the two CSR initiatives, that is, high fit and low fit. All of these questions employ a 7-point Likert Scale ranging from strongly disagree (1) to strongly agree (7) to collect respondents’ opinions. See Appendix B for all these questions. At the end of the pretest 1, there are some demographic questions (e.g., gender, age, nationality).
3.2.2 Results of Pretest 1
3.2.2.1 Reliability
Reliability analyses were executed to reduce the data set. As stated previously, six items are employed to question the luxury trait for both brand types (luxury and non-luxury scenarios), followed by five questions that measure the CSR fit. Reliability analyses have
been done to explore the internal reliability between the items measuring the same variable. The results have been analyzed through SPSS (Statistical Package for Social Sciences), and the statistical outcomes are displayed in Appendix B.
Firstly, the Cronbach’s Alpha has been calculated for both luxury scale and fit scale, the results of which are displayed in table 2 below. The internal consistency is stronger when the value of Cronbach's alpha is higher. Previous studies have suggested that if Cronbach's alpha coefficient is more than 0.7, the unity among items can be determined as good. In this study, the Cronbach’s alpha surpasses 0.7 in both luxury scale and the fit scale, indicating that these items can be combined into one single construct.
Table 2: Reliability Analysis in pretest 1: Cronbach’s Alpha Reliability Statistics Cronbach's Alpha Cronbach's Alpha Based on Standardized Items N of Items Cronbach’s Alpha if Item Deleted Luxury scale 0.937 0.937 6 0.905* Fit level 0.882 0.879 5 0.882**
* Cronbach’s Alpha if item 6 deleted (“ Manzoni shoes are desirable”).
**Cronbach’s Alpha if item 3 deleted (“I believe that the combination of Manzoni and the charity described above is meaningless - meaningful”)
However, after participants reading the brand scenarios, if item 6 “Manzoni shoes are desirable” is deleted, the Cronbach’s alpha would be increased to 0.963. Nevertheless, Cronbach’s alpha would reduce if either of the remaining items was deleted. After participants reading the CSR text, if item 3 “……the cooperation of Manzoni and the charity described above is meaningless - meaningful” is deleted, the Cronbach’s alpha would rise to 0.905, the value of which is much larger than the original value of 0.882. As for the remaining items, Cronbach's Alpha would be diminished if either of the items is eliminated. It can be concluded that the item “Manzoni shoes are desirable” and the item “ … … the
cooperation of Manzoni and the charity described above is meaningless - meaningful” shall be removed in the final survey.
3.2.2.2 Means Overview
Table 3 provides an overview of the means of all stimuli.
Table 3: Means Overview of all stimuli in pretest 1
Stimuli N Mean Std.Deviation
luxury 28 5.75 0.82
Non-luxury 28 2.81 0.99
Low fit 28 5.73 0.78
High fit 28 4.35 1.38
3.2.2.3 Independent-Samples T Test
An independent-samples t-test indicates that there is an extreme significance between luxury stimuli and non-luxury ones, t(45) = 10.52, p < 0.01, d = 0.17. However, it has been found to be statistically non-significant in the fit scale, t(33) = 4.08, p = 0.08, d = 0.6. As a result, except the item “desirability”, the remaining five items that measure luxury can be employed in the final survey. In addition, the difference is not obvious in fit stimuli since, in this within-subject experimental design, participants may be under the influence of the brand information given earlier, that is, prior knowledge or bias to luxury or non-luxury brands’ CSR initiatives. Therefore, a second pretest solely for the fit stimuli is necessary to be carried out. Table 4 and Table 5 display statistical output of two scales in this T-test.
Table 4: Independent Samples Test in pretest 1: Luxury scale Independent Samples Test
Levene's Test for Equality
F Sig. t df
Sig.
(2-tailed) DifferenceMean Std. Error Differen ce 95% Confidence Interval of the Difference Lower Upper
luxury Equal variances
assumed .199 .658 10.519 45 .000 2.784 .265 2.251 3.317
Equal variances
not assumed 10.391 40.911 .000 2.784 .268 2.243 3.325
Table 5: Independent Samples Test in pretest 1: Fit scale Independent Samples Test
Levene's Test for Equality of
Variances t-test for Equality of Means
F Sig. t df Sig. (2-tailed)
Mean Differen ce Std. Error Differen ce 95% Confidence Interval of the Difference Lower Upper
fit Equal variances
assumed 14.582 .020 4.173 44 .000 1.366 .327 .706 2.025
Equal variances
not assumed 4.078 32.524 .080 1.366 .335 .684 2.047
3.2.3 Pretest 2
The second pre-test aims to figure out whether the fit stimuli are optimal for the final experiment and how respondents perceive the level of fit between the CSR initiatives and the hypothesized brand “Manzoni”. To avoid the prior bias towards the brand, the luxury texts are not presented in this pretest. Results from pretest 1 indicate that item 3 (…the cooperation of Manzoni and the charity described above is meaningless - meaningful ) should be deleted as it reduces the Cronbach’s Alpha value of the luxury scale. In this sense, only the fit texts are shown to the participants, followed by a 4item (incongruentcongruent, incompatible -compatible, not complementary - complementary, illogical - logical) and 7-Likert scale used by Bigné et al. (2002). The pretest 2, like pretest 1, includes the same introduction part as
well as demographic question and is also conducted through Qualtrics online. This pretest is completed by 26 respondents.
3.2.4 Results of Pretest 2 3.2.4.1 Reliability
Four items, as stated before, are questioning the fit scale for two fit scenarios (high product-cause fit and low product-cause fit). Reliability analyses were conducted to ensure the internal reliability between fit items. This pretest also adopts SPSS (Statistical Package for Social Sciences) as the analysis tool.
In this pretest, the Cronbach’s Alpha was 0.905 (> 0.7), indicating that those items could be combined into one single construct. Then, when several items are removed, the Cronbach’s Alpha has been checked as to whether it would be increased to over 0.905. The results demonstrate that the Cronbach’s Alpha reduces no matter which item has been eliminated. It can be summarized that there is a high internal congruity between objects.
3.2.4.2 Independent-Samples T Test
An independent-samples t-test shows that scores were significantly higher for high fit (M = 5.84, SD = 0.68) than for low fit (M = 3.01, SD = 1.01), t(26) = 8.07, p < .001, d = 0.33. Table 6 below displays the analysis output. Therefore, proved by pretest 2, this fit item can be coded into a variable, that is, fit type.
Table 6: Independent Samples Test in pretest 2: Fit scale Independent Samples Test
Levene's Test for Equality of
Variances t-test for Equality of Means
F Sig. t df Sig. (2-tailed)
Mean Differen ce ErrorStd. Differen 95% Confidence Interval of the Difference
ce Lower Upper fit Equal variances
assumed 8.62 .005 8.344 44 .000 2.812 .337 2.132 3.491
Equal variances
not assumed 8.067 26.241 .000 2.812 .349 2.095 3.528
3.3Main Experiment
3.3.1 Questionnaire design
The questionnaire contains an introduction, a brand-related situation, and questions measuring the dependent variable, the mediator, the independent variable, the moderator, the control variables and the demographic statistics. Six scenarios have been refined or modified in the main experiment based on the results of the previous two pre-tests. Apart from the manipulation, the scenarios(i.e., brand name, CSR campaigns, layout, etc.)are identical. The treatments are displayed in Appendix D.
The measuring items for the variables have been drawn from earlier studies, and they have been adjusted slightly to better serve for this research. The dependent variable, that is, consumer attitude towards the brand is measured by five items (appealing, pleasant, favorable, good image, desirable) referring to the previous research of Spears and Singh (2004). As for the independent variable: brand type (luxury vs. non-luxury), as suggested by Vanhamme et al. (2015) has been measured by five items like “Manzoni shoes are of superior quality/ made of rare and precious materials/ unique/ expensive/ luxurious”. All of these questions adopt a 7-point Likert Scale, in which 1 represent strongly disagree and 7 strongly agree.
In view of the manipulation check of pretest 2, the main experiment adjusts the fit stimuli with a four-item scale (i.e., “I believe the combination of brand Manzoni and the charity described above is Incongruent- Congruent/ Incompatible- Compatible/ Not
complementary-Complementary/ Illogical- Logical”) drawn from Bigné et al. (2002). Another critical variable is the mediator, that is, perceived altruistic attribution that is measured by a three-item scale including, for instance, “I think Manzoni is motivated by self- interest or social-interest/ profit-motivated or social-motivated/ altruistic or egoistic” (Becker-Olsen et al., 2006). In the above questions, bipolar responses that are measured on seven-point scales have been applied.
After manipulation check measures, the control questions adapted from Dekhili et al. (2013) have been presented to explore consumer opinions and their awareness of CSR’s significance. Four demographic questions concerning the respondents’ gender, age, education, and nationality have been displayed at the end of the questionnaire. See the final questionnaire in Appendix E.
3.3.2 Data Collection
Most of the questionnaires were randomly distributed through the researchers’ social networks, and people could participate by clicking the link or scanning the QR code from the posts on various social media platforms. Since all respondents were asked to forward the post, a snowball sampling method might take effect. Other participants were randomly allocated via student. uva email. This survey finally collected 366 respondents, and 326 of them filled out the entire questionnaire. Among them 59.1% are female ( M = 1.47, SD = 0.656), the average age of the participants is 31 (M = 1.48, SD = 0.66), and 256 of them have a Bachelor’s degree (M = 3, SD = 0.71). Table 7 below demonstrates the number of responses under all conditions.
Table 7: Respondents in each condition in main experiment
Conditions Luxury-High fit Luxury-Low fit Luxury-No CSR Non-luxury &
N 55 55 51 57 55 53
4. Analysis and Results
4.1Data Preparation
4.1.1 Data screening
It is vital to screen and record the data before analyzing the reliability and validity of the items. Missing values and errors in data entry can be easily spotted and excluded through running frequency distribution in the SPSS. After conducting frequency distribution, no missing values were shown in the output.
Another essential process is to recode counter-indicative items. In this experiment, since none of the questions used negatively-keyed items2, there is no need recoding
counter-indicative items. In this sense, the primary data are prepared for further analysis. 4.1.2 Reliability
According to Saunders, Lewis, and Thornhill (2009), reliability examines whether or not a variable is capable of being interpreted consistently across various conditions. Table 8 presents the Cronbach’s Alpha value.
Table 8: Cronbach’s Alpha value Reliability Statistics
Variable Cronbach’s Alpha N of items
Consumer Attitude 0.936 5
Luxury Scale 0.942 5
Image Fit 0.928 4
Product Fit 0.912 4
Perceived Altruistic 0.920 3
All scales are highly reliable with Cronbach’s Alpha greater than 0.9, which indicates that those items in each scale are capable of being combined into one single construct. In addition, the corrected item-total correlations in each scale all surpass 0.3, and this suggests a good correlation. Also, none of the items would substantially affect the reliability even though they are deleted. In this experiment, product fit and image fit share the same questions based on earlier studies introduced by Bigné et al. (2002).
4.1.3 Manipulation checks
After analyzing and examining the reliability of all instruments, the next step is to compute the scale means. In this study, items measuring the luxury scale were recoded into a new variable: IVluxury TOT; items measuring consumer attitude were recoded into a new variable: DVTOT; items measuring the perceived altruistic attribution were recoded into a new variable: MediatorTOT; items measuring the CSR fit were recoded into new variables: ImagefitTOT and ProductfitTOT. All the new variables recorded the means of the original items. What’s more, dummy variable “ condition ” was established to divided data into different 6 conditions (luxury & high fit/ luxury & low fit/ luxury & no CSR/ non-luxury & high fit/ non-luxury & low fit/ nonluxury & no CSR). Dummy variables “Brand_type” and “Fit_level” were also established for further analysis. The Table 9 below shows the means overview of the above variables. According to the rule of thumb, 0 is within the bounds of the 95% CI if the data is normally distributed. As is shown in this figure, none of the variables has a significant skewness or a high kurtosis. Therefore, all the scales were normally distributed and can be used in further analysis.
4.1.3.1 Independent-Samples T- Test
Fitness level and brand type in this experiment have been manipulated in six treatments. As image fit and product-cause fit are not measured in control groups (no CSR text), it should be taken into consideration both high fit level and low fit one. An independent- Samples T-Test has been executed for two levels of CSR fit and brand type (luxury vs. Non-luxury) according to the means overview (in Table 9) of the variables mentioned above.
Table 9: Compared means and SDs in brand types (N = 326) and fit levels (N = 222) Group Statistics
Fit level N Mean Std. Deviation Std. Error Mean
Imagefit high fit 112 6.03 1.03 .097
low fit 110 5.08 1.03 .099
Productfit high fit 112 6.34 .75 .071
low fit 110 2.57 .78 .075
Brand_type luxury 161 6.38 .35 .028
Non-luxury 165 2.05 .34 .026
Note: The Std. Deviation is significant at the 0.05 level
There is a extreme significance between luxury stimuli (M = 6.38, SD = 0.35) and non-luxury stimuli (M = 2.05, SD = 0.34), t(323) = 114.18, p < 0.001. Meanwhile, when measuring the product fit, this test has been found to be statistically significant between hifh fit level (M = 6.34, SD = 0.75) and low fit level (M = 2.57, SD = 0.78), t(220) = 36.62, P < 0.001. With regard to image fit, there is also an extreme significance between high fit level (M = 6.03, SD = 1.03) and low fit level (M = 5.08, SD = 1.03), t (220) = 6.85, P < 0.001 though the mean difference was only 1.
Therefore, after Examination of the overall mean scores of the brand type and Fit level showed that manipulation texts were perceived as expected in brand type and fit level. What’s
more, it is quite impressive that in low fit level, respondents see product fit much lower than image fit (M= 2.58, SD=0.78 versus M = 5.08, SD = 1.03). As this paper investigates the effect of product-cause fit, therefore, the results are as intended because the difference is significant between product-cause fit levels (p < 0.05). See the Independent-Samples T Test output in Appendix F. Table 10 shows the means overview and skewness & Kurtosis of all dependent variable and independent variables.
Table 10: Means Overview Descriptive Statistics
N Minimum Maximum Mean SD Skewness Kurtosis
Statistic Statistic Statistic Statistic Statistic Statistic SD Statistic SD
IVLuxuryTOT 326 1.00 7.00 4.19 2.20 .025 .135 -1.219 .269 DVTOT 326 1.00 7.00 5.16 1.36 -0.436 .135 -.610 .269 MediatorTOT 326 1.00 7.00 4.6503 1.77497 -.383 .135 -.728 .269 ImagefitTOT 222 2.50 7.00 5.5586 1.13433 -.334 .163 -.819 .325 ProductfitTOT 222 1.00 7.00 4.4718 2.03520 -.050 .163 -1.150 .325 Valid N (listwise) 222
4.2 Hypotheses Testing
4.2.1No CSR vs. CSRThe first hypothesis (H1a, H1b) assumes that when there is no CSR activity, consumers would respond positively (negatively) to the luxury (non-luxury) brand, but they would act adversely when a luxury (non-luxury) brand conducts CSR initiatives. To test this hypothesis, a two-way ANOVA was adopted for the effect of independent variables, that is, brand type (luxury and non-luxury) and CSR (with CSR and without CSR) on the dependent variable, that is, consumer attitude. Control questions have been taken into account as well. The results are displayed in the following Table 11.
Table 11: Univariate ANOVA output in H1 Dependent Variable: Consumer Attitued
Source Sum of Squares df Mean Square F Sig. Partial Eta Squared
Corrected Model 247.22a 8 30.90 27.90 .000 .413 Intercept 56.34 1 56.34 50.34 .000 .138 control_1 .11 1 .11 .10 .750 .000 control_2 .05 1 .05 .05 .830 .000 control_3 12.67 1 12.67 11.44 .001 .035 Gender .05 1 .05 .05 .826 .000 Age 2.95 1 2.95 2.66 .104 .008 CSR type 16.05 1 16.05 14.49 .000 .044 Brand type 98.06 1 98.06 88.55 .000 .218 CSR * Brand type 118.09 1 118.09 106.64 .000 .252 Error 351.06 317 1.11 Total 9267.29 326 Corrected Total 598.28 325
a. R Squared = .413 (Adjusted R Squared = .398) b. Computed using alpha = .05
Levene’s test of equality of error variances shows a significance of .000, which means that the assumption is not homology.There is a significant main effect of CSR type on consumer attitude, F (1, 326) = 14.49, p < 0.001, ƞ2 = 0.044. Also, a extremely significant
main effect of brand type on consumer attitude appears in the results, F (1, 326) = 88.55, p < 0.001, ƞ2= 0.218. With regard to the interaction effect between CSR and brand type, there is
a statistically significance in between, F ( 1, 326) = 106.64, p < 0.001, ƞ2= 0.252. Besides, A
statistically significant difference between control question 3 (regarding to the importance of CSR) and consumer attitude is shown in above Table 12, F (1,326) = 11.44, p = 0.001, ƞ2 =
0.035. Figure 2 and Figure 3 below show the interaction relationship between CSR and No-CSR initiatives in different brand type. The estimated marginal means of the dependent
variable in the varying conditions are shown.
Figure 2: Relationship between brand-type and CSR/No-CSR relative to consumer attitude
Figure 3: Relationship between brand-type and CSR/No-CSR relative to consumer attitude
The Two-way ANOVA analysis also found differences between the means and SDs in different conditions. Regarding to a luxury brand, there is a difference of consumer attitude between CSR engagements (M = 5.35, SD = 1.21) and No-CSR engagements (M = 6.13, SD = .68). For a non-luxury brand, this difference is also significant in that CSR engagements score much higher (M = 5.36, SD = 1.16) than No-CSR engagements (M = 3.38, SD = 0.91).The means and standard deviations of different conditions are shown in Table 12 below.
