Corporate Social Responsibility and Perceived
Motives
The Impact of Company-Cause Fit, Commitment, Source and
Preexisting Beliefs on Perceived Motives
Jelles Wester
10004531
Amsterdam, January 16
th2015
Qualification: MSc. Business Studies - Marketing Supervisor: prof. dr. J.H.J.P Tettero
Second supervisor: drs. ing. A.C.J. Meulemans Academic year: 2014-2015
1
Foreword
This thesis is written as part of my Msc Business Studies at the University of Amsterdam (UvA). I want to thank some people who helped me in the process of writing this thesis. First, I want to thank my thesis supervisor Prof. Dr. Tettero. I want to thank him for his supporting role and his feedback that helped me to improve the quality of the thesis. Secondly I want to thank the people who helped me in distributing the questionnaires of this thesis. Their support was essential in obtaining the data as quickly as possible. Of course I want to thank my respondents as well for taking the time to fill in the questionnaires. Furthermore I want to thank some friends and family for checking the entire thesis, especially Mr. Benda who checked whether the storyline of this thesis was logical. At last, I want to thank the rest of the people who helped me to think differently about certain aspects of the thesis and offered me a fresh perspective. Overall I think the support of all these people together improved the quality of the final version.
2 Contents Foreword ... 1 1. Abstract ... 3 2. Introduction ... 4 3. Literature review ... 7
3.1 The emergence of CSR and the definition ... 7
3.2 Why companies undertake CSR activities and the benefits ... 10
3.3 The effectiveness of CSR on performance outcomes ... 14
3.4 Perceived motives about CSR initiatives ... 16
3.5 Factors influencing perceived motives ... 17
3.5.1 Brand-cause fit ... 17
3.5.2 Level of commitment ... 18
3.5.3 Type of source ... 19
3.5.4 Preexisting beliefs ... 20
3.5.5 Conceptual framework ... 22
4. Data and Method ... 23
4.1 Sample ... 23
4.2 Research instrument(s) and procedures ... 24
4.3 Potential methodological limitations ... 28
5. Results ... 30 5.1 Descriptive statistics ... 30 5.2 Factor analysis ... 33 5.3 Factorial MANOVAs ... 37 5.4 Further analysis ... 39 6. Discussion ... 42 6.1 Limitations ... 42 6.2 Contributions ... 44
6.3 Practical and managerial implications ... 46
7. Conclusion ... 48
8. References ... 49
9. Appendices ... 56
9.1 Appendix A: The Questionnaire ... 56
9.2 Appendix B: Descriptive Statistics - Items perceived motives ... 67
3
1. Abstract
Corporate Social Responsibility (CSR) is increasingly important for companies’ reputations and financial performance. However, engaging in CSR does not necessarily lead to positive financial outcomes. Some variables seem to influence the effectiveness of CSR in gaining positive organizational outcomes. This thesis looks at four factors that could influence the perceived motives of consumers about why companies engage in CSR: the level of company-cause fit, the level of commitment to the company-cause, the type of source via which consumers hear about the CSR activity, and the preexisting beliefs of consumers about the company. The results confirm the existence of more sophisticated perceived motives, instead of just self-serving and other-self-serving motives. The more sophisticated motives include values-driven, stakeholder-driven, strategic, and egoistic motives. Furthermore, high fit and neutral source both had a positive impact on strategic motives. Preexisting beliefs seem to have a direct instead of a moderating impact on the perceived motives. This direct impact should be explored further in future research.
4
2. Introduction
Nowadays the importance of Corporate Social Responsibility (CSR) to business is growing. Although it is a centuries’ old thought that business should address society’s concerns, the attention to it from both business and the research community has increased over the last couple of decades (Carroll, 1999). Forbes magazine (2012) states that companies should consider engaging in CSR since it can bring many benefits such as innovation, costs savings, brand differentiation, long-term thinking, and customer and employee engagement. Currently, large companies are even being ranked on a CSR reputation rank (Reputation Institute, 2013). A CSR reputation study shows that 73% of the 55.000 consumers surveyed were willing to recommend companies perceived to be delivering on CSR. Contrary to that, just 17% are willing to recommend a company perceived as poorly delivering on its CSR programs (Reputation Institute, 2013).
In short CSR is, according to the Commission of the European Communities (2001), ‘a concept whereby companies integrate social and environmental concerns in their business
operations and in their interaction with their stakeholders on a voluntary basis’. Since
companies rely on resources from society and have an impact on society, they should address the concerns of society as well. Companies are important for our economic development and can have tremendous social power (Davis, 1967). With social power, come social responsibilities (Davis, 1967).
In the end, the most important objective for companies is to sustain. In order to sustain, companies need to make profits and therefore they need to sell their products. Reputation and image are becoming increasingly important (Kurucz et al., 2008). Overall there seems to be a positive relationship between corporate reputation and financial performance (Roberts & Dowling, 2002). Thus, companies that have a better reputation will
5 likely perform better financially, and ultimately have a better chance to sustain and be successful.
The CSR activities that a company is undertaking are becoming an important element of their reputation (Reputation Institute, 2013). The study by the Reputation Institute (2013), reports that people’s willingness to recommend, work for and invest in a company is driven 60% by their perceptions of the company. However, not every company succeeds in their CSR attempts (Du et al., 2010).
Szykman et al. (2004) state, that consumers can be highly skeptical about the CSR intentions of companies. They might think that companies engage in CSR just for profit motives and are not being sincerely concerned about the societal issues they address (Fein, 1996). Therefore it is interesting for companies to understand what triggers the suspicion and even more interesting what factors should companies take into account in designing effective CSR initiatives. Since there exist differences in findings about the impact of CSR on financial performance, Carroll and Shabana (2010) suggest that future research should take a look at some mediating variables and situational factors that could influence this relationship.
Ellen et al. (2006) looked at two potential elements of the CSR offer that might influence the perceived motives of consumers about why companies engage in CSR. These two elements include the fit between the company and the cause they support, and the level of commitment of the company in supporting this cause. Their results show that both can be factors impacting on the perceived motives (Ellen et al., 2006). Therefore it is interesting to look at potential other factors that could influence the consumers’ perceived motives. This thesis will address this need to look at other potential factors that influence the relationship between the CSR offer of a company and the perceived motives the consumers have about this offer. Thereby it will repeat the research of Ellen et al. (2006) to test the generalizability of their findings.
6 The first potential factor to add is suggested by Ellen et al. (2006) themselves. They state that their study did not take into account preexisting beliefs about the motives of specific industries to engage in CSR. They pose the question that if consumers have preexisting beliefs that firms intentionally engage in CSR to make up for their shortcomings, CSR might lose its viability as a marketing tool.
Another interesting factor can be derived out of the article of Yoon, Gürhan-Canli and Schwarz (2006) about companies with bad reputations engaging in CSR. They made a distinction in sources via which the consumers heard about the social efforts of the companies. The results indicated that a company source is less convincing than an independent source (Yoon et al, 2006). This thesis will explore this second factor as well and look at the effect on the perceived motives of consumers.
In sum, this thesis will repeat some elements of the research of Ellen et al. (2006) and will add two other elements – preexisting beliefs and different types of sources – that potentially can have an impact on the relationship of CSR on the perceived motives. The research question in this thesis is, whether and how do preexisting beliefs and different
sources impact on the relationship between CSR initiatives and perceived motives?
To give structure to this thesis, first, in the third paragraph existing literature will be reviewed. In this paragraph the factors to be investigated will be explained and hypotheses will be made. This results in a conceptual framework, which will be mapped to make it more clear. In the fourth paragraph the research methods will be explained. The fifth paragraph will handle the analysis of the data and will give the results. These results will be discussed in the sixth paragraph. The limitations of this research, the contributions, the practical implications and some avenues for further research will be given. And finally, the last paragraph will handle the conclusion.
7
3. Literature review
3.1 The emergence of CSR and the definition
In order to understand the factors that could influence how consumers perceive CSR initiatives, it is first important to have a clear understanding of CSR. This section will focus on the emergence and the meaning of CSR. Thereafter a definition will be presented that forms the foundation for this thesis.
In the literature the evolution of the CSR concept can be traced back to the beginning of the 1950s (Carroll, 1999). The body of literature on CSR has a varied history and is quite extensive (Carroll, 1999). Traces of evidence of business community’s concern for society could be found prior to the 1950s as well. However formal writings about the subject started at the beginning of the 1950s (Carroll, 1999).
During the 1950s until the 1970s, most research focused on defining CSR and outlining its antecedents (Carroll, 1999). Numerous definitions were proposed and a couple antecedents were identified. An influential one is proposed by Carroll (1979) and holds that the social responsibility of business encompasses the economic, legal, ethical and discretionary expectations that society has of organizations at a given point in time.
This definition explains that society expects several levels of responsibility from companies. At the very least a company should provide economic welfare to a community, which is the main reason of existence for businesses (Carroll, 1999). Thereby companies should obey the law, which is explained by the legal responsibility (Carroll, 1999). Governments could, for example, pose laws that protect communities, employees or the environment. The third level of responsibility is the ethical level, which represents the kinds of behaviors and ethical norms that society expects businesses to follow (Carroll, 1999). This ethical responsibility goes beyond just obeying the law and involves adjustment to social norms. The discretionary responsibility is the last level and represents voluntary roles that
8 business assumes, but for which society does not provide clear norms (Carroll, 1999). This could stem from the company’s or employees’ own concerns.
During the 1980s and 1990s other themes related to CSR started to emerge, such as stakeholder theory and Corporate Social Performance (Carroll, 1999). Moir (2001) focused in his article on what is meant by CSR. There is some disagreement among researchers what CSR actually means. Moir (2001) outlined three important theories related to CSR. He states that in order to understand the meaning of CSR it is necessary to have some understanding of stakeholder theory, social contracts theory and legitimacy theory (Moir, 2001). Stakeholder theory holds that the firm can be described as a series of connections of stakeholders that the firm’s management attempts to manage (Moir, 2001). It means that the company relies on several groups of stakeholders (Donaldson & Preston, 1995; Freeman, 1994). A stakeholder is any group or individual who can affect or is affected by the achievement of the organization’s objectives (Freeman, 1984, p. 46). Stakeholders can be divided into primary and secondary stakeholders (Moir, 2001). Primary stakeholders have a direct connection to the company and secondary stakeholders include individuals or groups that can be affected by the operations of the company but do not have a direct connection with the company (Moir, 2001). Primary stakeholders include for example shareholders and investors, employees, customers and suppliers (Clarkson, 1995). The bottom-line of stakeholder theory is that companies take into account various stakeholders because it is in their commercial interest. They have to do it in order to operate and survive (Moir, 2001).
The essence of social contracts theory in the context of CSR is that business acts in a responsible manner, because it is part of how society implicitly expects business to operate, not because of the commercial interests (Moir, 2001). Legitimacy theory is defined as ‘a generalized perception or assumption that the actions of an entity are desirable, proper, or appropriate within some socially constructed system of norms, values, beliefs and definitions
9 (Suchman, 1995, p. 574)’. In the context of CSR this clearly connects to social contracts theory in the sense that society might expect responsible behavior on the part of business (Moir, 2001). The overlapping notion between the three theories is that it is important to take into account the concerns of various stakeholders. This can be important for business and social interest (Moir, 2001).
Dahlsrud (2006) reviewed a total of 37 different definitions of CSR in order to give some clarity about the large amount of different existing definitions. He explains that many definitions of CSR are biased towards specific subjects related to CSR. If competing definitions have diverging biases, people will talk differently about CSR and this will thus prevent productive engagements (Dahlsrud, 2006, p. 1). Therefore it is important to understand how the various definitions differ. Dahlsrud (2006) reviewed the definitions and put forward five dimensions which overarch all the definitions. These dimensions include the stakeholder dimension, social dimension, economic dimension, voluntariness dimension and environmental dimension (Dahlsrud, 2006). Dahlsrud (2006) concludes that the challenge is to understand how CSR is socially constructed in a specific context and not so much how to define CSR in general.
The definition of CSR this thesis uses is the definition of the Commission of the European Communities (2001) which states that ‘CSR is a concept whereby companies
integrate social and environmental concerns in their business operations and in their interaction with their stakeholders on a voluntary basis’. This definition is chosen, because it
10
3.2 Why companies undertake CSR activities and the benefits
In order to understand the perceived motives of consumers about why companies engage in the social initiatives, it is important to understand what possibly could trigger these companies to engage in such initiatives. In other words, this section will focus on the perspective of the company and their motivation to undertake CSR.
As a starting point, Campbell’s (2007) institutional theory of CSR will be discussed. eight propositions that explain the conditions under which corporations are likely to conduct CSR activities. He argues that there are some mediators in the relationship between basic economic conditions and the way corporations behave (Campbell, 2007). These mediators include public and private regulation, the presence of nongovernmental and other independent organizations that monitor corporate behavior, institutionalized norms regarding appropriate corporate behavior, associative behavior among corporations themselves, and organized dialogues among corporations and their stakeholders (Campbell, 2007, p. 946).
Regarding the economic conditions, Campbell (2007) proposes that companies will be less likely to act socially responsible when their financial performance is weak and when the economic environment is unhealthy, simply because they do not have the necessary resources to invest in socially responsible activities (Waddock & Graves, 1997). Thereby he argues that companies with either too much or too little competition will display less socially responsible behavior (Campbell, 2007). In a situation when a firm is experiencing severe competition, the profit margins are likely to be low and therefore the firm is more likely to make cost savings (Campbell, 2007). Investments in CSR are then likely to be stopped. In the opposite situation when a firm is experiencing almost no competition, things like corporate reputation or customer loyalty are not likely to affect sales, profitability or survival (Campbell, 2007). Therefore investing in CSR activities would make no sense. It would only decrease the profit margins (Campbell, 2007).
11 Regarding the mediators in the relationship between the economic conditions and the corporation’s behavior, they all have a positive effect on the degree of socially responsible behavior. This means that when they are present and high enough, they will lead to more socially responsible behavior on the part of the corporations. For example, when there are public and private regulations regarding social behavior, it naturally will lead to increased socially responsible behavior (Campbell, 2007). The same principle applies to the presence of nongovernmental organizations that monitor behavior, institutionalized norms regarding behavior, associative behavior among corporations themselves, and organized dialogues among corporations and their stakeholders (Campbell, 2007, p. 946).
As with defining CSR, there are numerous theories about why companies undertake CSR activities. Garriga and Melé (2004) offered with their paper an overview of the most important theories. They categorized the most influential theories into instrumental, political, integrative and ethical theories. Instrumental theories argue that CSR should mainly be used as a strategic tool to achieve economic objectives and at the end wealth creation (Garriga & Melé, 2004, p. 53). This comes back to the view of Friedman (1970) that ‘the only responsibility of business towards society is that they maximize the profits that flow to their shareholders, within the boundaries of the legal framework’. CSR should be used as a strategic instrument to gain a competitive advantage (McWilliams & Siegel, 2001).
Another highly influential paper is that of Davis (1967) to which Garriga and Melé (2004) refer in their political theories section. Davis (1967) presented two well-known principles: the social power equation principle and the iron law of responsibility. The social power equation principle states that ‘social responsibilities of businessmen arise from the amount of social power that they have’ (Davis, 1967, p. 48). This could be translated into ‘the more power companies have, the more social responsibilities they will get’. The iron law of responsibility states that when companies do not use their social power in a way that society
12 considers responsible, it is likely that they will lose this power (Davis, 1960, p.63). This is because other groups eventually will interfere and take over some of this power and the associated responsibility (Garriga & Melé, 2004).
Carroll and Shabana (2010) took another perspective and wrote an article about the business case for CSR. They looked at the underlying arguments and rationales supporting the investment in socially responsible initiatives. They took a company perspective (Carroll & Shabana, 2010) instead of an economical process perspective as Campbell (2007) did. The four main business-case arguments they outlined were based on the work of Kurucz, Colbert and Wheeler (2008). The four main arguments include cost and risk reduction, gaining competitive advantage, developing reputation and legitimacy, and seeking win-win outcomes through synergistic value creation (Kurucz et al., 2008).
CSR initiatives could lead to cost and risk reductions in a couple of ways (Carroll & Shabana, 2010). First, the demands of stakeholders could pose a threat to the viability of the company. By undertaking CSR activities, companies could decrease this threat by responding to the stakeholder demands (Kurucz et al., 2008). A second way of cost and risk reduction may be achieved when the CSR activities address the concerns about the natural environment (Carroll & Shabana, 2010). Being environmentally proactive could result in cost and risk reduction (Berman et al., 1999). For example, when a company decides to use less packaging materials, this could save costs and address the demands of society. A third way of cost and risk reduction may be achieved when CSR activities are directed at managing community relations (Berman et al., 1999). Having the community on your side may decrease the amount of regulation to which a company must adhere (Carroll & Shabana, 2010).
Engaging in CSR could lead to competitive advantages as well (Kurucz et al., 2008; Smith, 2007). Where cost and risk reduction refers to a cost leadership strategy, gaining a competitive advantage refers to a differentiation strategy (Carroll & Shabana, 2010). Smith
13 (2003) argues that CSR activities may lead to competitive advantages. If the social responsibility strategy of a company is unique, it may help to differentiate that company from their competitors in the minds of the consumers (Smith, 2003). CSR activities can also strengthen the relationships between the company and their customers (Carroll & Shabana, 2010).
The third argument refers to the development of reputation and legitimacy (Kurucz et al., 2008). This argument implies that when a company undertakes CSR initiatives, it may enhance the company’s reputation and increase its legitimacy. The fourth argument refers to seeking win-win outcomes through synergistic value creation, what basically means that companies should turn social problems into economic opportunities and benefit from these (Drucker, 1984). The win-win perspective means that both stakeholders’ demands and companies’ objectives will be fulfilled (Carroll & Shabana, 2010; Kurucz et al., 2008).
In conclusion, Brønn and Vidaver-Cohen (2009) gave in their article an overview of the examined motives for companies to engage in CSR. They explained that there are two motives why companies undertake CSR activities: strategic and moral. The first is the strategic motive (Brønn & Vidaver-Cohen, 2009), which is subdivided into instrumental motives and institutional motives. Instrumental motives hold that companies belief that engaging in social initiatives can have a positive impact on profitability (Brønn & Vidaver-Cohen, 2009). This view connects clearly to the article of Kurucz and colleagues (2008) and focuses on the benefits for the company. The institutional motives suggest that companies engage in social initiatives primarily due to institutional pressures (Brønn & Vidaver-Cohen, 2009). This view clearly connects to the ‘Iron Law of Responsibility’ (Davis, 1960) and the view that business is pressured by several factors to behave socially responsible (Campbell, 2007). The second motive for companies to undertake CSR activities is the moral motive, of
14 which the main idea is that business has an ethical duty to ‘give back’ to society (Brønn & Vidaver-Cohen, 2009).
So, there are multiple theories about why companies undertake CSR activities. There are some pressures from society to take social responsibility (Campbell, 2007; Garriga & Melé, 2004; Davis, 1967). These societal pressures can come in the form of regulations, monitoring organizations, institutionalized norms, and consumers’ buying decisions (Campbell, 2007; Davis, 1960). On the other side CSR can give companies some great benefits, such as the reduction of costs and risks, a competitive advantage, enhanced reputation and legitimacy, and it can create a win-win situation for both the company and society (Carroll & Shabana, 2010). And next to this there is the ethical duty to ‘give back’, which Brønn and Vidaver-Cohen (2009) referred to.
3.3 The effectiveness of CSR on performance outcomes
It is clear that undertaking CSR activities has the potential to bring benefits to both companies and society (Carroll & Shabana, 2010). The question however remains whether and how effective CSR is (Carroll & Shabana, 2010). A large number of researchers investigated whether there is a relationship between CSR activities and Corporate Financial Performance (CFP), and their results differ quite substantially (Carroll & Shabana, 2010).
Some found the expected positive impact of CSR on financial performance (Orlitzky et al., 2003; Cochran & Wood, 1984), and some did not found any relationship at all (Aupperle et al., 1985; McWilliams & Siegel, 2000). The article of McGuire, Sundgren and Schneeweis (1988) found some results that indicate that the relationship might be the other way around. Their research shows a positive relationship between prior financial performance and social responsibility. This suggests that, at least in some situations, social responsible behavior is a result of financial performance instead of a cause (McGuire et al., 1988;
15 Stanwick & Stanwick, 1998). There is obviously some inconsistency in the findings regarding the effect of CSR on CFP (Carroll & Shabana, 2010).
Overall there seems to be a positive relationship between CSR activities and buying behavior (Sen & Bhattacharya, 2001). However, as Mohr, Webb and Harris (2001) pointed out, not all consumers respond in the same way to CSR activities. Their results show that there is a small but articulate group of consumers that are displaying socially responsible consumer behavior. These consumers are active and interested in buying products from companies that engage in CSR (Mohr et al., 2001). However, not all consumers respond in the same manner. Werder (2008) looked at the effect of CSR on behavioral intention as well. She was not able to find an effect of CSR initiatives on behavioral intentions.
Carroll & Shabana (2010) state that overall, research seems to suggest a positive effect of CSR on performance. However, these findings are inconsistent. They argue that these inconsistencies may not merely be the result of methodological differences and interpretation biases (Carroll & Shabana, 2010). The differences in findings could be the result of the existence of some mediating variables and situational factors that influence the relationship between CSR activities and financial performance (Carroll & Shabana, 2010). They suggest that future research should explore these potential mediators and situational factors (Carroll & Shabana, 2010). The article of Du, Bhattacharya and Sen (2010) explains that undertaking CSR activities could lead to a number of positive company outcomes, such as enhanced corporate image, strengthened stakeholder-company relationships, and enhanced stakeholders’ advocacy behaviors. However, stakeholders’ low awareness of and unfavorable attributions towards the CSR activities can have a negative impact on the outcomes of the CSR activity (Du et al., 2010). The next section will look into detail at these attributions.
16
3.4 Perceived motives about CSR initiatives
The second paragraph of the literature review of this thesis discussed why companies engage in CSR activities and what benefits it can bring to them. The paragraph discussed the question why should we do it, from a company perspective. This section will take a consumer perspective. It will look at what motives consumers attribute to a company’s CSR initiatives. This anticipates on the need to explore potential mediators and situational factors in the relationship between CSR initiatives and organizational outcomes (Carroll & Shabana, 2010).
Attribution theory in the CSR literature focuses on consumers questioning ‘why’ companies engage in CSR (Gilbert & Malone, 1995). Instead of looking at the motives explained by the firms and managers themselves, it takes a consumer perspective and focuses on consumer attributions (Ellen et al., 2006). Early work in this research area looked at the distinction between altruistic and egoistic motives (Handelman & Arnold, 1999; Webb & Mohr, 1998). Altruistic motives are other-serving and are related to the moral motives, of which the main idea is that business has an ethical duty to ‘give back’ to society (Brønn & Vidaver-Cohen, 2009). Egoistic motives are self-serving and are focused on getting better out of the investments in CSR (Ellen et al., 2006). Drumwright (1996) explained that firms can have dual motives. They can obtain both economic and social objectives. He undertook a qualitative research and interviewed a large number of managers. The results indicated these managers believed that consumers are simplistic in their judgments about CSR initiatives and view them as serving either economic ends or reflecting sincere social concerns (Drumwright, 1996).
However, the results of Ellen et al. (2006) explain that consumers can have more complicated attributions about the motives of why companies undertake CSR initiatives. They showed in their exploratory research that self-serving motives can be divided into strategic and egoistic motives, whereas other-serving motives can be divided into values-driven and
17 stakeholder-driven motives (Ellen et al., 2006). They explain that just as Whetten and Mackey (2002) suggested, attributions related to strategic goals of getting and keeping customers, needed for their existence are widely accepted and therefore are perceived as positively. The negative counterpart of self-serving motives is reflected by the egoistic motives, such as taking advantage of a cause in order to make profits. Values-driven motivations reflect helping others for the sake of helping them, and is perceived as positive, whereas stakeholder-driven motivations reflect helping others because of public pressure and are perceived as insincere and negative (Ellen et al., 2006). The results of Ellen et al. (2006) showed that the attributions play a mediating role between CSR initiatives and organizational outcomes, such as purchase intent.
3.5 Factors influencing perceived motives
This section will discuss the four factors presented in the introduction that potentially can have an impact on the perceived motives discovered by Ellen et al. (2006). These factors include: brand-cause fit, level of commitment to the cause, type of source of CSR communication, preexisting beliefs about the company of interest.
3.5.1 Brand-cause fit
Ellen et al. (2006) looked at two elements of the CSR offer that could have an impact on the motives consumers attribute to companies that engage in these CSR initiatives: brand-cause fit and commitment. Brand-cause fit is researched quite extensively by numerous researchers (Bigné et al., 2012; Drumwright, 1996; Ellen et al., 2006; Fein, 1996; Hoeffler & Keller, 2002; Nan & Heo, 2007). A high brand-cause fit basically means that the company who supports a certain cause is logically related to that cause in the minds of consumers. Thus, in the minds of consumers it makes sense (Ellen et al., 2006). Fein (1996) suggests in his article that a high brand-cause fit compared to a low fit, will lead to less suspicion, because the firm
18 is not doing things far outside of its main business. According to the literature, it is recommended for firms to engage in CSR closely related to their main businesses (Cone et al., 2003; Varadarajan and Menon, 1988). Nan and Heo (2007) looked at the impact of brand-cause fit on consumer attitudes. Their results suggest that a brand-cause-related marketing message with a high brand-cause fit, leads to more favorable consumer attitudes towards the ad and the brand, compared to low brand-cause fit messages (Nan & Heo, 2007). The results of Ellen et al. (2006) showed that a high fit led to strategic and values-driven attributions, while a low fit led to egoistic attributions. They were unable to show any effect of brand-cause fit on stakeholder-driven attributions. This thesis will test their original hypothesis (including stakeholder-driven attributions) regarding brand-cause fit again.
Hypothesis 1: High (low) fit between a cause’s mission and a company’s core business will increase (decrease) (a) values-driven and (b) strategic attributions, while decreasing (increasing) (c) egoistic and (d) stakeholder-driven attributions (Ellen et al., 2006, p. 5).
3.5.2 Level of commitment
The second factor Ellen et al. (2006) investigated was the level of commitment. In their study they measured the level of commitment by the length of time a certain company engages in a certain cause. They made a distinction between companies engaging in a certain cause for a short period of time, namely one month, versus a long period of time, namely seven years (Ellen et al., 2006). Their results indicated that short commitment led to stakeholder-driven attributions, which had a negative impact on purchase intent (Ellen et al., 2006). The study was unable to show any results on the part of the long commitment. Ellen et al. (2006) state that this is possibly due to the fact that they only took into account the length of time of the commitment and no other factors of commitment. This thesis will elaborate more on the
19 commitment variable by taking into account all the criteria of commitment outlined by Dwyer (1987).
The three criteria of commitment include inputs, durability and consistency (Dwyer, 1987). The input criterion explains that the parties provide high levels of inputs to the association (Blau, 1964). Significant economic, communication, and/or emotional resources may be exchanged (Dwyer, 1987). In the case of CSR this could mean that a company invests a huge amount versus a very small of money in a social initiative. The durability criterion is the one Ellen et al. (2006) implemented in their research, and contains the length of time a certain commitment lasts. There should be some durability of the association over time (Dwyer, 1987). The third criterion explains that there is at least some level of consistency with which the inputs are made to the association (Dwyer, 1987), in such a way that the other party knows what to expect. In the case of CSR this could mean that a company invests on a regular basis in a certain cause. This thesis will take these factors into account and will test the same hypothesis tested in the research of Ellen et al. (2006).
Hypothesis 2: High (low) commitment to a cause will increase (decrease) (a) values-driven attributions, while decreasing (increasing) (b) strategic, (c) egoistic, and (d) stakeholder-driven attributions (Ellen et al., 2006, p. 5).
3.5.3 Type of source
The third factor to be tested in this thesis is the type of source from which the consumers hear about the CSR initiatives of the company. Yoon et al. (2006) investigated in their research the effects of CSR on companies with bad reputations. They looked at two factors that potentially could influence the positive effect of CSR these companies were aiming at.
The first factor was the level of benefit salience. They argued that, for example, if a tobacco company would support the National Cancer Association, the consumers are
20 presumably aware that the company tries to improve its negative image and they are going to question the sincerity (Yoon et al., 2006). The second factor they looked at was whether the source was a company source or a neutral (independent) source, via which consumers observe the social initiatives of companies with bad reputations. The results indicated that high benefit salience in combination with a company source led to a backfire effect of CSR (Yoon et al., 2006). On the other hand, low benefit salience in combination with a neutral source led to more positive evaluations (Yoon et al., 2006). Their results confirmed that consumers infer less sincere motives when they learn about CSR activities from a company source rather than a neutral source.
In this thesis, it will be investigated whether the type of source leads to certain attributions. The expectations are similar to the expectations regarding brand-cause fit. This leads to the following hypothesis.
Hypothesis 3: Learning about a company’s social intentions via an independent neutral (company) source will increase (decrease) (a) values-driven and (b) strategic attributions, while decreasing (increasing) (c) egoistic and (d) stakeholder-driven attributions.
3.5.4 Preexisting beliefs
The fourth factor to be tested in this thesis is whether preexisting beliefs of consumers about companies engaging in CSR have an impact on their perceived motives. Ellen et al. (2006) suggest that if a consumer has a preexisting belief that a firm intentionally uses CSR to make up for a shortcoming, it might lose its effectiveness. Prior research shows that there seem to be CSR strategies that can improve companies’ bad reputations (Yoon et al, 2006). Strahilevitz (2003) examined in his research the effect of the initial perception of the ethical nature of a firm. He compared ethical, unethical and ethically neutral firms engaging in CSR and their effect on the degree to which consumers expect ulterior motives for them to engage
21 in CSR. The results showed that consumers are less skeptical when an ethical firm engages in CSR compared to an unethical one (Strahilevitz, 2003). Strahilevitz (2003) looked in his research at the effects on altruistic versus ulterior motives, but did not take into account the fact that self- and other serving motives can be subdivided into strategic, egoistic, values-driven and stakeholder-values-driven attributions. Therefore, this thesis tries to relate these factors to the preexisting motives. It will be expected that preexisting beliefs will have a moderating role in the relationship between the three element of the CSR offer described above (brand-cause fit, commitment, and type of source) and the attributions. This leads to the following hypotheses.
Hypothesis 4: Positive (negative) preexisting beliefs about a company’s social intentions will have a moderating role in the relationship between brand-cause fit and the attributions, and will increase (decrease) (a) values-driven and (b) strategic attributions, while decreasing (increasing) (c) egoistic and (d) stakeholder-driven attributions.
Hypothesis 5: Positive (negative) preexisting beliefs about a company’s social intentions will have a moderating role in the relationship between the level of commitment and the attributions, and will increase (decrease) (a) values-driven and (b) strategic attributions, while decreasing (increasing) (c) egoistic and (d) stakeholder-driven attributions.
Hypothesis 6: Positive (negative) preexisting beliefs about a company’s social intentions will have a moderating role in the relationship between the type of source and the attributions, and will increase (decrease) (a) values-driven and (b) strategic attributions, while decreasing (increasing) (c) egoistic and (d) stakeholder-driven attributions.
22
3.5.5 Conceptual framework
Figure 1 shows the conceptual model which will be tested in this thesis. The model that this thesis adds to the literature is partly built on prior research, especially on the work of Ellen et al. (2006). The model tries to extend the generalizability of some of their results. Thereby it will investigate more in-depth the role of commitment by taking into account all three criteria of commitment outlined by Dwyer (1987), instead of just the length of time of the commitment. Furthermore it will extend the work of Ellen et al. (2006) with the variable of different types of sources. It will add a potential moderator, preexisting beliefs, on the relationship between CSR and consumer attributions as well. Prior research on this factor explains partially the effect of this moderator, but it is not yet tested in combination with the attributions discovered by Ellen et al. (2006).
23
4. Data and Method
4.1 Sample
The population to be researched in this thesis is quite broad. Basically it includes all types of consumers, who are exposed to news and communications about companies engaging in CSR. In that sense, there is no obvious sampling frame, such as when your population is a certain university, in which your sampling frame would be a list of all students of that university. A certain list for all consumers does not exist. Since the population is so large, it is necessary to attract as much respondents as possible. To make generalizability a little more feasible, this study will focus on the population of Dutch consumers instead of all consumers worldwide. We assume that consumers to the age of sixteen are not subject to CSR communications of companies and basically everyone above that age does. An estimation of the population for this thesis includes the total population in the Netherlands above an age of sixteen. According to the CBS bureau (CBS, 2014), the total population of the Netherlands above the age of sixteen (17 years and older) is rounded off to 13.6 million people.
This thesis aims at getting a sample size of 300 respondents. However, getting more respondents would be better to improve the generalizability of the findings. This number is based on the fact that for each of the tested ‘main effect’ variables (fit, commitment, and type of source) there will be around 100 respondents. The ‘interaction effect’ variable (preexisting beliefs) will be tested for all 300 respondents. The generalizability to the population could improve if a more specific group was chosen, for example the citizens of a certain city or the employees of a certain company. However, due to time and budget constraints it is more convenient to take the entire Dutch population as the group of focus. If the population would be narrowed down, it would be less feasible to attract the targeted sample size.
According to Saunders et al. (2011, p. 234) in deciding on the appropriate sampling technique a number of questions should be answered. The first question is whether a suitable
24 sampling frame is available. This is not the case, so a non-probability sampling technique will be used. The second question than is whether the sample must be representative (Saunders et al., 2011). This is the case, which leads to the question whether there are relevant quota variables available. It could be argued that the sample should represent the Dutch population on some demographics, for example age and level of education. By doing this it could have increased potential to generalize the findings to the population (Saunders et al., 2011, p. 235). However, due to time and budget constraints it will be difficult to get enough respondents that represent the Dutch population on several demographics. Therefore a convenience sampling technique will be used (Saunders et al., 2011).
The thesis will use a survey containing multiple questionnaires in order to test the three different ‘main effect’ variables. Actually, the respondents will be divided over a couple of separate sections of the questionnaire. The next subparagraph will explain in more detail the structure of the questionnaire. The questionnaires will be administered via www.qualtrics.com to the selected respondents. When this method will not yield sufficient respondents, the questionnaires will be administered in shopping areas and perhaps some companies interested in this topic can be contacted to help spreading the questionnaires.
4.2 Research instrument(s) and procedures
This thesis will use a questionnaire in the form of an experiment. It basically means that the respondents will be divided over a couple of scenarios. In the analysis these groups will be compared to see whether they react differently on these scenarios. Ellen et al. (2006) had a response rate of 44.7%. Their questionnaire was mailed to a random sample of staff employees working at a university. This thesis will expect a certain response rate as well.
The outline of the questionnaire is quite complicated. Figure 2 shows how the questionnaire will be structured. First, all respondents will be thanked for taking the time to
25 fill in the questionnaire. They will be ascertained that their responses will be used for this thesis only and that their responses will be used anonymously and with confidentiality. They also will be asked to take enough time to fill in the questions as accurately as possible. After the introduction, the respondents will read a short text about the content and outline of the questionnaire.
Figure 2: Structure of the questionnaire
Then all respondents will be introduced to a certain company. This company will be a well-known existing company, so that it is likely that the respondents already have perceptions about that company. Three types of companies will be introduced. On the one hand a company of which it is likely that the respondents have relatively positive preexisting beliefs and on the other hand a company of which it is likely that the respondents have relatively negative ones. The third company that will be used is a company of which the consumer perceptions most likely will fluctuate between positive and negative ones. The
26 companies will be presented by name, logo, and a short one-sentence explanation about their business.
In order to test the variable preexisting beliefs, it is important that most of the respondents are familiar with the companies presented in the questionnaire. The companies that will be used for this thesis include Albert Heijn, the NS, and Apple. Albert Heijn is the largest supermarket company in the Netherlands. They are well-known, present throughout the entire country, and quite popular. Overall perceptions of Albert Heijn, therefore, will most likely be positive. The NS is an acronym for ‘Nederlandse Spoorwegen’, which means Dutch railways. The NS is responsible for the timetables and schedules, and physical transport of almost all trains in the Netherlands. The NS is criticized very often and they have a relatively bad reputation in the Netherlands. Overall perceptions of the NS, therefore, will most likely be negative. The third company is Apple, which is a well-known international electronics company. They are famous for their products, such as the Ipod, Ipad, Iphone and Mac computers. Apple has both true loyal fans, and determined opponents. It will be expected, therefore, that the preexisting beliefs for Apple will fluctuate between negative and positive.
After the division of the respondents over the three different companies, they will answer two questions about whether they know the company and how well they know it. Then, the respondents are asked about their perceptions of this company on a 5-point Likert-type scale ranging from extremely negative to extremely positive.
Subsequently the respondents will be further divided over 7 different scenarios: high fit, low fit, high commitment, low commitment, neutral source, company source, and a control group. In order to be able to test the hypotheses, different scenarios have to be developed. In the analysis the different groups will be compared to test whether the scenario they read, influenced their perceived motives. For example, the respondents in the low fit group will be compared to the respondents in the high fit group on their perceived motives.
27 Thus 21 different scenarios have to be developed: seven different scenarios multiplied by three different companies.
In the high fit scenario, it is important that there is an obvious link between the company and the cause they support (Bigné et al., 2012; Drumwright, 1996; Ellen et al., 2006; Fein, 1996; Hoeffler & Keller, 2002; Nan & Heo, 2007). Albert Heijn is coupled to the Voedselbank, which is a nonprofit organization that distributes free food packages to the poorest people of the Netherlands. The scenario explains that Albert Heijn donates its leftover products to the Voedselbank, in order to give these to the poor people. The NS is coupled to the Voedselbank as well. At first sight, the link between these two organizations might be unclear. However, the scenario explains that the NS will help in distributing the food packages over the entire country. This explanation makes the link between the NS and the Voedselbank strong and clear. Apple is coupled to a fictitious company called Start Education Now (SEN), which is committed to help provide education to children. The link is explained in more detail in order to make it clear and strong in the minds of the respondents.
In the low fit scenario, all companies are coupled to the Aids Fonds, which is a nonprofit organization that focuses on the fight against the aids disease. The links with the Aids Fonds and a supermarket company, railway company, and electronics company are unclear and therefore these links are chosen to represent the low fit conditions.
As mentioned in paragraph 3.5.2 Ellen et al. (2006) only took into account the length of time as part of the commitment variable. This thesis will use all the three criteria of commitment outlined by Dwyer (1987) in order to test whether this improves the relationship between commitment and perceived motives. The three criteria of commitment include inputs, durability and consistency (Dwyer, 1987). In the high commitment scenarios, it is explained that the companies will commit to the causes for a period of seven years, and the helping processes will get daily attention and have a high priority. In the low commitment
28 scenarios, it is explained that the companies will commit to the cause for just one month and make a small one-time donation of 25.000 euro/dollar.
In order to make a difference between the neutral and company source scenarios, different layouts were used. The text of the neutral source scenario is presented in the layout of NRC Handelsblad, which is a highly valued newspaper in the Netherlands. The text of the company source scenario is presented in the layout of the different websites of the companies. Appendix A shows the entire questionnaire, including the different scenarios.
The people in the control group will only read that the direction of the company decided to engage in CSR and that they want to donate money to a cause. It is not mentioned which cause, how committed they are, and the layout is neither company nor neutral. So the control scenarios are not influenced by any of the variables to be tested.
After reading the different scenarios of the different companies, all respondents will respond to four 7-point Likert-type scale items regarding purchase intentions. Subsequently, they will respond to sixteen 7-point Likert-type scale items regarding the perceived motives. These items are the same items that were used by Ellen et al. (2006). These sixteen items formed the strategic, egoistic, values-driven, and stakeholder-driven factors in the results of Ellen et al. (2006). This thesis will test whether the sixteen items will lead to the same factors.
At last, all respondents will fill in some demographic questions. They will report their gender, age, level of income, level of education, country of origin, and ethnicity.
4.3 Potential methodological limitations
One limitation can be the relatively small sample size. Since the population is extremely large and several variables are to be tested, a relatively small number of respondents per variable will be the result. This makes it difficult to generalize the findings.
29 A second limitation in the process of distributing the respondents over the different variables is that it is likely that the sample is less representative of the population in terms of demographics. It could be the respondents of one variable are significantly older than the respondents of the other variable.
For the preexisting beliefs variable, it could also be a limitation when the respondents are not familiar with the company presented in the questionnaire. Therefore the questions asking whether and how well the respondent knows the company were included. Also well-known companies were used to decrease the likelihood of this potential limitation.
Finally, another possible limitation could be that the direction of this questionnaire is too obvious and that the respondents will give desirable answers. They can guess what the questionnaire wants to test and therefore give answers in that direction.
30
5. Results
The first subsection will look at some descriptive statistics and the proportions of several groups within the total sample. The second subsection will focus on the factor analysis that will be used to test whether the same four perceived motives result from this group of respondents. The third subsection will discuss the factorial MANOVAs that test the impacts of fit, commitment, source, and preexisting beliefs on the perceived motives. The last section will focus on some further analysis based on the results of the factor analysis and factorial MANOVAs.
5.1 Descriptive statistics
First, the data was scanned for missing values or incorrect data. Some of the respondents were deleted, because they filled in some incorrect responses. For example, one respondent filled in that he was born in Alaska and that his ethnicity was Japanese. Thereby he claimed that he was zero years old, which is clearly incorrect. These kinds of respondents were deleted from the data file and there remained data of 302 respondents, which could be used for further analysis. The data was checked for counter-indicative items as well, but there were none of them present in the data file.
The second step to undertake was to have a look at the demographic variables. The demographic variables that were included in the questionnaire were the gender, age, level of education, level of income, country of origin, and ethnicity of the respondents. The ratio male/female was 40 percent male to 60 percent female, just like in the research of Ellen et al. (2006). The ages of the respondents ranged from 16 to 70 years old, with a mean age of 29.26 years. Subsequently, the respondents were categorized into different age groups. Table 1 and figure 3 give an overview of the number of respondents in the different age groups. Most respondents are in the age group of 20-24. This is most likely the result of the fact that the
31 student conducting this research falls into this age group as well. Due to the fact that a convenience sampling technique has been used, most respondents come from his direct network. Therefore most respondents are about the same age as he is. The mean age in this data file does not correspond with the mean age in the data file of Ellen et al. (2006). The mean age of their respondents was 38.7 years old.
Age groups Frequency Percentage (%) Cumulative Percentage (%) 15-19 28 9.3 9.3 20-24 184 60.9 70.2 25-29 39 12.9 83.1 30-34 12 4.0 87.1 35-39 4 1.3 88.4 40-44 4 1.3 89.7 45-49 3 1.0 90.7 50-54 11 3.6 94.4 55-59 12 4.0 98.3 60-64 2 0.7 99.0 65< 3 1.0 100.0 Total 302 100.0 100.0
Table 1: Age groups
Figure 3: Age groups
Probably due to the same reason, most of the respondents earned between €0 and €999 per month. However none of the questions asked it directly, probably most of the respondents are students and do not have a fulltime job. Therefore most respondents (192 out of 302) fall into the first category of income. Table 2 gives an overview of the distribution of the income
32 levels. Most respondents have a HBO or WO level of education (162 out of 302). Table 3 will give an overview of the distribution of the level of education. At last, most respondents claimed to be born in the Netherlands and most of them filled in that their ethnicity is Dutch.
Income (€) Frequency Percentage (%) Cumulative
Percentage (%) 0 - 999 192 63.6 63.6 1000 - 1499 40 13.2 76.8 1500 - 1999 30 9.9 86.8 2000 - 2499 26 8.6 95.4 2500 - 2999 3 1.0 96.4 3000 - 3499 4 1.3 97.7 3500 - 3999 3 1.0 98.7 4000 < 4 1.3 100.0 Total 302 100.0 100.0
Table 2: Level of Income
Level of Education Frequency Percentage (%) Cumulative
Percentage (%)
Basisschool 1 0.3 0.3
Lbo, mavo, vmbo, mbo-1, eerste 3 jaar van havo of vwo
18 6.0 6.3
Havo, vwo, mbo- 2,3,4 119 39.4 45.7
Hbo, Wo 162 53.6 99.3
Anders 2 0.7 100.0
Total 302 100.0 100.0
Table 3: Level of Education
Furthermore, from a total of 302 respondents, 106 respondents (35.1%) filled in a questionnaire with as focal company the NS, 104 respondents (34.4%) filled in one of Albert Heijn, and 92 respondents (30.5%) of Apple. The differences in percentages are a result of the fact that the software on www.qualtrics.com, equally divides the respondents over the three companies. However, the software does not take into account the fact that some of the respondents do not completely finish the questionnaire. Clearly this was the case with a number of respondents that filled in a questionnaire about Apple.
Most of the respondents knew the companies. Only 4 out of 302 respondents declared that they did not know the company of interest. Thereby most of the respondents stated that they knew the companies ‘good’. Two respondents indicated that they know the focal company ‘really bad’, seven know it just ‘bad’, 60 ‘average’, 140 ‘good’, and 93 ‘really
33 good’. The preexisting beliefs were tested by asking how the respondents felt about the focal company on a scale of one (extremely negative) to five (extremely positive). Overall 5 respondents felt extremely negative, 45 negative, 75 neutral, 143 positive, and 34 extremely positive. As expected most of the negative and extremely negative reactions came from NS (27 respondents) followed by Apple (15 respondents). As expected, most respondents were positive or extremely positive about Albert Heijn (76 out of 104). Table 4 gives an overview of the distribution of preexisting beliefs about the different companies.
Company Extremely Negative
Negative Neutral Positive Extremely
Positive
NS 4 23 39 39 1
Albert Heijn 0 8 20 62 14
Apple 1 14 16 42 19
Total 5 45 75 143 34
Table 4: Distribution of the Preexisting Beliefs
The distribution of respondents over the different scenarios is relatively equal. The least respondents read a low commitment scenario, whereas most of the respondents read a low fit or control group scenario. High fit accounted for 42 respondents, low fit for 48, high commitment for 44, low commitment for 39, neutral source for 41, company source for 40, and the control groups for 48.
5.2 Factor analysis
In order to analyze the conceptual model and test the according hypotheses, first the perceived motives variables had to be created (values-driven, stakeholder-driven, strategic, and egoistic). This section will focus on the factor analysis needed to create these variables. The resulting factors will be used as variables in the next section of the analysis, in which the impact of fit, commitment, source and preexisting beliefs will be examined.
Ellen et al. (2006) reduced sixteen items to four factors. These items were asked to the respondents of the questionnaire for this thesis as well. Table 5 gives an overview of these
34 items, including the means and standard deviations. These means and standard deviations are the ones of the total responses, including all the scenarios. For example, both high and low fit are integrated in these numbers. Overall, the means are relatively around the middle score of four (on a 7-point scale). ‘To get publicity’ had the highest mean score of 5.98, which means that many respondents thought that was certainly one of the reasons of why companies undertake CSR initiatives. Appendix B gives an overview of the descriptive statistics of all of the items of the perceived motives, including the means, standard deviations, variances, skewness’ and kurtosis’.
The company is undertaking this social initiative, because: Mean Standard Deviation
They feel morally obligated to help. 4.21 1.67 They have a long-term interest in the community. 4.84 1.53 Their owners or employees believe in this cause. 4.01 1.47 They want to make it easier for consumers who care about the cause to support it. 3.72 1.53 They are trying to give something back to the community. 3.68 1.61 They feel their customers expect it. 4.58 1.68 They feel society in general (i.e. consumers) expects it. 4.69 1.67 They feel their stockholders expect it. 3.96 1.69 They feel their employees expect it. 3.38 1.56 They are taking advantage of the nonprofit organization to help their own business. 5.06 1.57 They are taking advantage of the cause to help their own business. 5.25 1.46 They want it as a tax write-off. 4.07 1.65 They want to get publicity. 5.98 1.19 They will get more customers by making this offer. 5.63 1.42 They will keep more of their customers by making this offer. 5.56 1.38 They hope to increase profits by making this offer. 5.36 1.63
Table 5: Items perceived motives/attributions
Taking a look at the correlation matrix of the sixteen items shows that there are many correlations among the items. Some items correlate higher than other items, but overall it gives the impression that there are some underlying factors. The highest correlation is between item six and seven, and has a significant correlation of 0,741 (p < 0.01). Appendix C gives an overview of the entire correlation matrix.
Principal axis factoring with an oblimin rotation was used to examine the structure of the 16 attribution items. The reason why this type of factor analysis was chosen is because Ellen et al. (2006) used the same method. In order to be consistent with their research, the
35 same path had to be followed. The screeplot suggests that either two or four factors should be extracted. The eigenvalues greater than one suggest that four factors should be extracted. Since Ellen et al. (2006) came up with four factors as well, and the both the screeplot and eigenvalues greater than one suggest an extraction of four factors, this amount of factors was extracted. The results of the factor analysis were almost completely the same as the results of Ellen et al. (2006). Only one item switched: ‘they want to get publicity’. In the results of Ellen et al. (2006), this item loaded on the factor of egoistic attributions. In this thesis it switched to strategic attributions. Except for this item, all the other items loaded on the same factors. The results of Ellen et al. (2006) displayed a total percentage of variance extracted of 62.8. The total variance extracted of the factor analysis of this thesis was 53.4, thus slightly lower.
Thereafter the internal consistencies of the four factors were assessed by using Cronbach’s alpha. All Cronbach’s alphas were above .70 and therefore reliable. Table 7 displays the results of the factor analysis, including the eigenvalues, the percentage of variance extracted, the Cronbach’s alphas, and the KMO and Bartlett’s test. Table 6 displays the correlation matrix of the four factors. Especially the strategic and egoistic factors, and the values-driven and stakeholder-driven factors are correlated. This is probably due to the fact that Ellen et al. (2006) derived these factors from the original self-serving and other-serving factors outlined by Brønn and Vidaver-Cohen (2009). At last, the factors were computed into variables which could be used for further analyses.
Factor Strategic Values-driven
Stakeholder-driven Egoistic Strategic --- -.037 -.105 .537 Values-driven --- --- -.403 -.016 Stakeholder-driven --- --- --- -.198 Egoistic --- --- --- ---
36
The company is undertaking this social initiative, because: Strategic Values Driven Stakeholder Driven Egoistic
They will get more customers by making this offer. .98 They will keep more of their customers by making this offer. .68
They want to get publicity. .62
They hope to increase profits by making this offer. .61
Their owners or employees believe in this cause. .72 They feel morally obligated to help. .62 They are trying to give something back to the community. .60 They have a long-term interest in the community. .56 They want to make it easier for consumers who care about the cause to support it. .56
They feel their customers expect it. -.90
They feel society in general (i.e. consumers) expects it. -.82
They feel their stockholders expect it. -.52
They feel their employees expect it. -.39
They are taking advantage of the cause to help their own business. .88 They are taking advantage of the nonprofit organization to help their own business. .79
They want it as a tax write-off. .38
Eigenvalue 3.77 3.03 1.00 .75
Percentage of variance extracted 23.51 18.93 6.25 4.69
Cronbach’s alpha .86 .76 .80 .74
NOTE: Kaiser-Meyer-Olkin measure = .79; Bartlett’s test of sphericity = 2033.37; p < .01. Loadings <.35 suppressed.
37
5.3 Factorial MANOVAs
This section will use the four variables of the perceived motives created in the factor analysis to test the hypotheses. In order to test the separate impacts of fit (high versus low), commitment (high versus low), and source (neutral versus company) on the four main perceived motives with preexisting beliefs (positive versus negative) as a potential moderator, three separate factorial MANOVAs had to be performed. Performing one factorial MANOVA including all the independent variables at once was not possible, since all the different conditioned groups were independent of each other. For example, there were no respondents that both read a scenario regarding low fit and low commitment, or high fit and neutral source. The conditioned groups are all independent and therefore three separate factorial MANOVAs had to be performed.
First, the four independent variables were created. For example, regarding the variable of fit, the respondents that read a scenario of high fit, were assigned a one. The ones who read a low fit scenario were assigned a zero. In this manner all independent variables were created.
The first factorial MANOVA tested the impact of fit on the attributions, with preexisting beliefs as a potential moderator. A full factorial MANOVA analysis was performed, including the main effects of fit and preexisting beliefs, and the interaction effect between both. The results showed that there were no significant interaction effects between fit and preexisting beliefs, rejecting hypothesis 4. The influence of fit on attributions was significant, Wilks’s λ = 0.88, F(4, 83) = 2.847, p < .05, ɳ² = .12, as well as the influence of preexisting beliefs on attributions, Wilks’s λ = 0.83, F(4, 83) = 4.251, p < .05, ɳ² = .17. Respondents that read the high-fit scenario were more likely than those who read the low-fit scenario to attribute firm participation in the CSR offer to strategic and egoistic motives, strategic F(1, 86) = 4.644, p < 0.05, ɳ² = 0.052; egoistic F(1, 86) = 4.673, p < 0.05 ., ɳ² =
38 0.051, offering support for hypothesis 1b and the opposite effect for hypothesis 1c. Hypotheses 1a and 1d were not supported.
The second factorial MANOVA tested the impact of commitment on attributions, with preexisting beliefs as a moderator. The results of the second MANOVA showed that were no significant interaction effects between commitment and preexisting beliefs, rejecting hypothesis 5. The main effect of commitment on the attributions is not significant, just as the main effect of preexisting beliefs. Therefore, the second MANOVA offers no support for the second hypothesis.
The third factorial MANOVA tested the impact of source on attributions, with preexisting beliefs as a moderator. The results of the third factorial MANOVA showed that were no significant interaction effects between source and preexisting beliefs as well, rejecting hypothesis 6. The influence of source on attributions was significant, Wilks’s λ = 0.88, F(4, 83) = 2.847, p < .05, ɳ² = .12, as well as the influence of preexisting beliefs on attributions, Wilks’s λ = 0.83, F(4, 74) = 2.605, p < .05, ɳ² = .12. Respondents that read the neutral source scenario were more likely than those that read the company source scenario to attribute firm participation in the CSR offer to strategic motives, strategic F(1, 77) = 6.612, p < 0.05, ɳ² = 0.079, offering support for hypothesis 3b. Hypothesis 3a, 3c, and 3d were not supported.
39
5.4 Further analysis
The results show that only hypotheses 1b and 3b were supported, whereas most of the other hypotheses were rejected. Only the reversed version of hypothesis 1c was supported as well. Furthermore, the results suggest that there is a direct significant effect between preexisting beliefs and the perceived motives. In two of the three factorial MANOVAs, preexisting beliefs had a significant impact.
The results of the first factorial MANOVA state that the respondents that had positive preexisting beliefs were more likely than those who had negative preexisting beliefs to attribute firm participation in the CSR offer to values-driven and stakeholder-driven motives, values-driven F(1, 86) = 15.713, p < 0.05, ɳ² = 0.154; stakeholder-driven F(1, 86) = 5.319, p < 0.05 ., ɳ² = 0.058. The results of the third factorial MANOVA state that the respondents that had positive preexisting beliefs were more likely than those who had negative preexisting beliefs to attribute firm participation in the CSR offer to values-driven motives, values-driven F(1, 77) = 5.678, p < 0.05, ɳ² = 0.069.
As mentioned in paragraph 5.2, the factor analysis also could have extracted two factors. In order to see whether the factorial MANOVAs will yield more results, a factorial analysis was performed again to create two factors instead of four. Table 8 shows the results of this factor analysis. The two factors that came out of the analysis were the sums of the egoistic and strategic factors, and the values-driven and stakeholder-driven factors. These new two factors correspond to the factors of self-serving and other-serving outlined in the literature (Brønn & Vidaver-Cohen, 2009).
The two factors (self-serving and other-serving) were computed into new variables and the three factorial MANOVAs were performed in the same way as in paragraph 5.3, except that the dependent variables were replaced for the self-serving and other-serving variables.
