The influence of stakeholder culture on stakeholder activism : a quantitative research paper focusing on companies in the apparel industry and food industry

(1)

- Master Thesis -

Entrepreneurship and Management in the Creative Industries

The Influence of Stakeholder Culture on Stakeholder Activism

A Quantitative Research Paper focusing on Companies in the Apparel Industry and

Food Industry

Leeuwen van, Willemijn

Student number: 10885994 21st of August, 2015

(2)

Statement of Originality

This document is written by Student Willemijn van Leeuwen 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.

(3)

Abstract

Theory suggests that a stakeholder culture of companies is based on the extent to which companies emphasize the interest for stakeholders. However, the extent to which companies emphasize the interest for stakeholders differs between companies. As a result, stakeholders perceive that companies integrate stakeholders’ interests differently – while some companies fully integrate stakeholders’ interests in the companies strategy, other companies don’t integrate their interests. This might clarify why activists consider targeting particular companies over other companies.

Nonetheless, researchers haven’t considered whether the concept of stakeholder culture can actually reduce stakeholder activism. In an attempt to explore the determinants of stakeholder activism, I study how stakeholder culture can influence stakeholder activism in both the apparel industry and food industry. I do this by performing a database research. The results of this study, however, indicate that stakeholder culture doesn’t influence stakeholder activism. The reason for this outcome is that activists look at the CSR practices of companies when they consider targeting them and these practices don’t fully correspond to a stakeholder culture of companies. In overall, I aim to explain whether stakeholder culture can influence the determinants that activists consider as important when they decide to target

companies. Therefore, my study contributes to a broader understanding of the relationship between a companies’ stakeholder culture and stakeholder activism.

(4)

Index of Figures and Tables

FIGURE 1: STAKEHOLDER CULTURES 19

FIGURE 2: CONCEPTUAL MODEL 29

FIGURE 3: CORRELATION MATRIX 59

TABLE 1: OVERVIEW OF HYPOTHESES 30

TABLE 2: OVERVIEW OF VARIABLES 44

TABLE 3: DESCRIPTIVE SUMMARY FOR CONTROL AND CATEGORICAL VARIABLE 50 TABLE 4: DESCRIPTIVE SUMMARY FOR CONTINUOUS VARIABLES 51

TABLE 5: RESULTS OF HYPOTHESE 1 61

(7)

Introduction

Although scholars argue that companies engage in CSR nowadays, their motives behind CSR engagement differ. Some companies engage in CSR out of profit – or shareholder concerns, while other companies engage in CSR out of value – or stakeholder concerns (Jones, Felps & Bigley, 2007). It is intuitively plausible that CSR out of profit motives creates a backlash and can increase stakeholder activism (Jones, Felps & Bigely, 2007; Waldron, Navis & Fisher, 2013). So far though, it isn’t clear how the motives behind CSR engagement influence stakeholder activism. Therefore, my research question is: ‘how does a stakeholder culture influence

stakeholder activism?’

Smith (2003) argues that the motives for CSR engagement define to what extent companies actually respond to stakeholders’ interests. This, in turn, influences the behaviour of stakeholders (Morsing & Schultz, 2006; Peloza & Shang, 2010).

Namely, companies’ motives for CSR engagement do motivate stakeholders to either pressure companies or not (Fernandez-Feijoo, Romero & Ruiz, 2013). Therefore, companies wonder what particular motives for CSR engagement reduce stakeholder pressure: do they relate more or less to stakeholder concerns or to a company’s own self-interest? In this respect, scholars believe that companies that fully respond to stakeholders’ interests positively influence stakeholder behaviour and, as a result, reduce the amount of stakeholder pressure. On the other hand, companies that don’t meet stakeholders’ interests can face increasing stakeholder pressure (Peloza & Shang, 2010; Alniacik, Alniacik & Genc, 2011). As a result of increasing stakeholder pressure, stakeholder activism can occur (Rehbein, Waddock & Graves, 2004;

Rowley and Moldonveanu, 2003).

A comparison between Nike and LEGO group illustrates how different companies respond to the interests of their stakeholders and, furthermore, how this can trigger

(8)

hand, LEGO group fully responds to stakeholders’ interests. This positively influences stakeholder behaviour and, furthermore, limits negative stakeholder pressure. This may be one of the reasons why LEGO group is able to maintain its strong reputation.

Up till now the motives behind CSR engagement seem to influence stakeholder activism. However, scholars haven’t studied whether and how this can contribute to lower social pressure and, thus, reduce stakeholder activism. So far, studies only investigated these concepts separately or linked each of these concepts to other relevant business topics. For instance, Jones, Felps and Bigley (2007) use ethical theories to create corporate stakeholder cultures. Furthermore, Waldron, Navis and Fisher (2013) explain the differences in companies’ responses to stakeholder activism. Although these papers don’t search for a relationship between the motives for CSR engagement and stakeholder activism, it seems that these journals provide information to link these concepts. Therefore, I study this gap to determine how companies should consider the motives for CSR engagement that correspond to both company’s goals and stakeholders’ interests.

This research helps to understand how the motives for CSR engagement can either trigger or limit the amount of stakeholder activism. Therefore, I wonder whether activists target companies based on their stakeholder culture. In other words, I study whether a stakeholder culture is as important as other aspects that activists evaluate when they decide to target companies. If so, the consideration between different motives for CSR engagement can become the solution to reduce activism and,

furthermore, can make this concept a potential source of competitive advantage. This let companies be able to create superior value for their stakeholders and superior profit for themselves.

I choose a qualitative research method to examine if the motives for CSR engagement influence stakeholder activism. Therefore, I take a sample of companies from the apparel industry and food industry. Especially, these companies are known for their active engagement in CSR and also face relatively more activism compared to

(9)

obtain data of companies from these industries through the use of different databases that provide data for the variables I describe in this research.

This research consists of several sections. It starts with a section in which I make arguments based on existing literature about the concepts of the motives for CSR engagement, a stakeholder culture and stakeholder activism. Furthermore, I present several hypotheses that contribute to the research question. Based on these

hypotheses, I present a conceptual model in order to let people understand the

different relationships I test in this research. Thirdly, I describe how I collect data for each of the concepts in order to test the hypotheses. Therefore, I also present a table to define every variable. After this section, I analyse and interpret the obtained results by the use of SPSS. Then, a discussion will follow about the existing literature and the outcomes of the hypotheses. At last, I provide a conclusion in which I answer my research question.

(10)

Theoretical Framework

Introduction

In this section, the theoretical background of this research is given by first defining the concept of CSR. This discussion, then, turns into arguments why companies actually engage in CSR. Then, I explain how these motives can influence stakeholder behaviour and, as a result, can trigger stakeholder activism. Furthermore, other concepts, such as company size and R&D investments, are also introduced in this section to strengthen the arguments for a possible relationship between stakeholder culture and stakeholder activism.

(11)

CSR and Motives for CSR Engagement

Introduction

Research into corporate social responsibility (CSR) has been around for more than 70 years (Freeman & Hasnaoui, 2011). However, it is only very recent that CSR has transformed into a fundamental matter in the literature of international business management. In line with increasing business studies on CSR, public debates about CSR have intensified as well. As a result, companies have assigned more value to CSR in order to respond to public debates (Pinkston & Carroll, 1996). Even stronger, nowadays most companies engage in CSR to respond to public debates (Freeman & Hasnaoui, 2011). In response to the increase of companies that engage in CSR, scholars also attempt to understand what particular motives drive companies to engage in CSR.

Definition of CSR

In order to study how the motives to engage in CSR influence stakeholder activism, I first assume a definition of CSR that is given by Smith (2003): ‘CSR refers to the

obligations of the company to society or, more specifically, the company’s

stakeholders - those affected by corporate policies and practices’. Compared to other

scholars, Smith (2003) pays more attention to the behaviour of stakeholders; he even makes it part of the definition of CSR. Knowing this, companies that engage more in CSR aim to meet stakeholder obligations and, as a result, try to control the attitudes and behaviour of them (Smith, 2003). For this reason, improving stakeholder obligations is usually a reason for companies to engage in CSR. Yet, it is still not clear to what extent this can actually reduce stakeholder activism.

Motives for CSR Engagement

(12)

governmental relationship (Kang & Hustvedt 2014; Alniacik, Alniacik & Genc, 2011). Other-regarding motives, on the other hand, relate to concerns about

transparent communication to stakeholders (Fernandez-Feijoo, Romero & Ruiz, 2014; Kang & Hustvedt, 2014).

Companies communicate other-regarding motives because they believe it is necessary to meet stakeholder obligations (Morsing & Schultz, 2006; Alniacik, Alniacik & Genc, 2011). These companies feel ethically concerned for their stakeholders and, therefore, inform their stakeholders that they are socially responsible (Peloza & Shang, 2011). On the other hand, companies communicate self-regarding motives for CSR engagement if they believe it is necessary to improve their own economic performance (Jones, Felps & Bigley, 2007; Doh & Guay, 2006; Fernandez-Feijoo, Romero & Ruiz, 2014). This means that the actual purpose of companies - meeting stakeholder obligations versus improving the economic performance - let companies decide what motives for CSR engagement they communicate to their stakeholders. Therefore, companies consider different motives for CSR engagement (Jones, Felps & Bigley, 2007).

Smith (2003) considers the integration of stakeholders’ interests in the decision-making of companies as a precondition to obtain increasing CSR performance. Therefore, it is likely that communicating other-regarding motives for CSR engagement causes increasing CSR performance. This means that companies

communicate more other-regarding motives respond to stakeholders’ interests and, as a result, are likely to obtain increasing CSR performance compared to companies communicating self-regarding motives (Li, Zhao, Shi & Li, 2014; Freeman &

Hasnaoui, 2011). In this respect, companies that aim for increasing CSR performance should communicate other-regarding motives for CSR engagement, rather than communicating self-regarding motives.

However, some companies do only engage in CSR for self-regarding motives. Their behaviour is only motivated by their corporate self-interest (Friedman, 1962). Knowing this, some companies only respond to their environmental and social responsibilities in order to improve their economic position and, furthermore,

(13)

These economic benefits can, furthermore, result in an improved reputation or increased access to capital. Indeed, Kolk (2004) argues that aspects, such as an improved reputation, have been found important for companies to engage in CSR.

CSR performance can positively influence the reputation of companies (Kang & Hustvedt, 2014; Doh & Guay, 2006). In other words, increasing CSR performance can be the result of meeting stakeholders’ interests and, as a result, can strengthen the reputation of companies. Indeed, CSR engagement is often used as a standard to measure the reputation of companies (Ellen, Web & Mohr, 2006). However, companies do face the risk of reputational damage when they lack to meet the interests of stakeholders (Smith, 2003). For instance, companies can face negative stakeholder pressure if they don’t meet stakeholders’ interests (Smith, 2003). This example can negatively influence the economic position of companies (Baron & Diermeier, 2007). Knowing this, companies consider different motives for CSR engagement, ranging from self-regarding motives to other-regarding motives.

Investments for CSR Engagement

Companies consider self-regarding or other-regarding motives for CSR engagement and, as a result, consider different types of investment. Some companies focus more on short-term investments, while other companies focus more on R&D and long-term investments. Whether companies invest on the short-term or the long-term depends on the motives companies have for CSR engagement. Therefore, the motives for CSR engagement reflect the type of investments companies make (Surroca, Tribo & Waddock, 2006).

McWilliams, Siegel and Wright (cited in Arrigo, 2013) argue that decisions about CSR are considered as a form of strategic investment. A strategic investment drives

(14)

sustain their economic position define their shareholders as their obligation and, therefore, are likely to invest on the short-term (Waldron, Navis & Fisher, 2013).

Communication Approach for CSR Engagement

Since companies have different motives to engage in CSR, they also differ in their communication approach about their CSR efforts (Jones, Felps & Bigley, 2007). Indeed, scholars emphasize that companies differ in their communication about their CSR efforts to stakeholders (Morsing & Schultz, 2006). For this reason, companies develop a communication strategy that is in line with their motives for CSR

engagement (Frostenson, Helin & Sandström, 2011). In this respect, companies consider a communication strategy that is in line with either other-regarding motives or self-regarding motives for CSR engagement. At first sight, companies may fare better if they communicate other-regarding motives to stakeholders because these motives emphasize the involvement of stakeholders’ interests in their operation. It, thus, indicates that these companies aim to meet the interests of their stakeholders. However, shareholders care more about the economic performance of companies. Therefore, companies also need to take into account the interests of shareholders and, as a result, also communicate about their performance. For these reasons, companies consider different motives for CSR engagement (Jones, Felps & Bigley, 2007). Furthermore, these motives influence the behaviour of different stakeholders (Jones, Felps & Bigley, 2007, Waldron, Navis & Fisher, 2013). This influence is discussed in the next section about stakeholder behaviour.

(15)

Stakeholder Behaviour

Introduction

Companies communicate different motives to engage in CSR (Jones, Felps & Bigley, 2007). Companies can either communicate other-regarding motives or self-regarding motives for CSR engagement. Since these motives determine what communication approach companies use, it also influences the behaviour of stakeholders (Morsing & Schultz, 2006; Peloza & Shang, 2010). Thus, the motives for CSR engagement influence stakeholder behaviour (Jones, Felps & Bigley, 2007, Waldron, Navis & Fisher, 2013). In order to broaden our scope about this topic, I explain how the motives to engage in CSR influence the behaviour of stakeholders in this section.

Stakeholder Expectations

Again, stakeholder behaviour is influenced by the extent to which companies meet stakeholder obligations. Thus, positive stakeholder behaviour is the result of meeting stakeholder obligations, which in turn satisfies the expectations of stakeholders

(Roberts, 1992; Smith, 2003; Freeman & Hasnaoui, 2010). As a result, companies can maintain or even strengthen their reputation. On the other hand, companies that don’t meet stakeholders’ interests, but rather communicate self-regarding motives, face negative pressure. For instance, companies that only communicate motives for CSR engagement to improve their economic performance are at greater risk of being involved in a boycott by stakeholders. Thus, the motives that companies communicate can either satisfy or dissatisfy stakeholders and, therefore, influence the behaviour of stakeholders (Jones, Felps & Bigley, 2007; Peloza & Shang, 2011).

(16)

and, as a result, reflects how and whether companies respond to their stakeholders (Jones, Felps & Bigley, 2007; Smith, 2003). Indeed, companies that fully involve stakeholders in their decision-making act different towards stakeholders than companies that don’t involve stakeholders in their decision-making (Jones, Felps & Bigley, 2007). Therefore, companies deal differently with their responsibilities towards stakeholders and, thus, maintain a different stakeholder culture (Waldron, Navis & Fisher, 2013).

Jones, Felps and Bigley (2007) study the aspects of different stakeholder cultures. As a result of their study, these authors present five stakeholder cultures that range from an altruist stakeholder culture to an agency stakeholder culture (Figure 1). The authors argue that a stakeholder culture is determined by the extent that companies care for their stakeholders (Jones, Felps & Bigley, 2007). In other words, stakeholder culture depends on the extent of stakeholder involvement in companies’ decision-making (Jones, Felps & Bigley, 2007; Waldron, Navis & Fisher, 2013). Hence, these stakeholder cultures seem in line with the motives to engage in CSR. Namely, companies that maintain an altruist culture fully are likely to involve stakeholders in their decision-making. As a result, these companies respond to their stakeholders’ interests and, thus, engage in CSR for other-regarding motives (Jones, Felps & Bigley, 2007). On the other hand, companies that maintain an agency culture aren’t likely to involve stakeholders in their decision-making, but rather act in their self-interest. As a result, these companies don’t respond to their stakeholders’ interests and, thus, engage in CSR for self-regarding motives (Jones, Felps & Bigley, 2007). Therefore, I argue that the motives to engage in CSR correspond to a stakeholder culture of companies.

(17)

Figure 1 - Stakeholder Cultures along one Continuum

Stakeholders can observe to what extent companies respond to their interests and, therefore, they can perceive what stakeholder culture companies have. Thus, companies that don’t meet stakeholder obligations, but rather communicate self-regarding motives towards stakeholders, are likely to maintain an agency culture. As stakeholders perceive self-regarding motives to engage in CSR, stakeholders can develop a negative perception of companies. As a result, stakeholders can consider targeting these companies in order to call attention for their interests. Therefore, it is likely that companies acting in their own self-interest are likely to be targeted by stakeholders. On the other hand, companies that fully respond to stakeholders’

interests don’t face stakeholder pressure because they aim to meet their obligations. In summary, these arguments lead to the following hypothesis:

Hypothesis 1: Companies with a self-regarding stakeholder culture are more likely to be targeted by activists than companies with an other-regarding stakeholder culture.

(18)

better meet stakeholder obligations can obtain increasing CSR performance. On the other hand, companies acting in their own self-interest lack to respond to stakeholder obligations and, therefore, are likely to obtain decreasing CSR performance.

Therefore, a stakeholder culture can reflect what CSR performance companies aim for. In summary, these arguments lead to the following hypothesis:

Hypothesis 2: Companies with a self-regarding stakeholder culture have lower CSR performance than companies with an other-regarding stakeholder culture.

Furthermore, companies with decreasing CSR performance are more likely to face stakeholder activism and, as a result face the risk of loosing their competitive advantage and their reputation (Peloza & Shang, 2010; Arrigo, 2013). Meaning that decreasing CSR performance usually leads to negative perceptions among

stakeholders, which can be at the expense of the economic performance and reputation of companies (Kang & Hustvedt, 2014). Therefore, decreasing CSR performance let stakeholder consider targeting companies (Kang & Hustvedt, 2014). Therefore, it is likely that decreasing CSR performance triggers stakeholder activism. In summary, these arguments lead to the following hypothesis:

Hypothesis 3: Companies with lower CSR performance are more likely to be targeted by activists than companies with a higher CSR performance.

The concept of CSR has become a matter of pubic debate (Baron & Diermeier, 2007). Since stakeholder activists are responsible for public debates with regard to CSR matters, their role is more important nowadays (Rehbein, Waddock & Graves, 2004; Waldron, Navis & Fisher, 2013). As a result, activists can run campaigns towards companies and raise awareness to stakeholders’ interests. Indeed, the aim of

stakeholder activists is to influence the motives of companies for CSR engagement by running campaigns (den Hond & de Bakker, 2007; Burchell & Cook, 2013).

Therefore, it is likely that the motives of companies for CSR engagement can trigger stakeholder activists to run campaigns. Therefore, I explain how the motives for CSR engagement influence stakeholder activism in the next section.

(19)

Stakeholder Activism

Introduction

Stakeholder behaviour is influenced by the motives that companies communicate to engage in CSR. These motives can encourage stakeholder activism. Yet, it isn’t clear how the motives for CSR engagement can reduce stakeholder activism because the impact of these motives on stakeholder activism isn’t studied yet (den Hond & de Bakker, 2007). Therefore, I explain in this section how the motives to engage in CSR influence the actions undertaken by activists.

Activist Considerations: the Perception of Stakeholders about Companies

The motives that companies communicate shape the perception of society about companies and, as a result, generate discussions in society (Kang & Hustvedt, 2014). Therefore, the motives for CSR engagement cause discussions in society (Baron & Diermeier, 2007). These discussions trigger either positive or negative word-of-mouth and, thus, influence the behaviour of stakeholders in a positive or a negative way (Fernandez-Feijoo, Romero & Ruiz, 2014; Kang and Hustvedt, 2014).

Negative stakeholder behaviour increases stakeholder pressure and, as a result, triggers activists to come into action (Baron & Diermeier, 2007; den Hond & de Bakker, 2007). Therefore, companies usually try to reduce negative stakeholder behaviour in the first place (Alniacik, Alniacik & Genc, 2011; Jones, Felps & Bigley, 2007). In turn, activists are motived by social or ethical concerns of stakeholders in order to raise awareness for stakeholder obligations (Baron & Diermeier, 2007; Smith, 2003). In this respect, activists aim to respond to negative stakeholder pressure by calling attention for stakeholder obligations. Furthermore, their aim is to put

pressure on the operation of companies to force companies to change their motives for CSR engagement. Thus, the motives for CSR engagement cause discussions, which

(20)

engagement (den Hond & de Bakker, 2007). Namely, companies communicating self-regarding motives do usually not respond to stakeholder obligations, and, therefore, can cause negative stakeholder behaviour. As a result, activists consider coming into action. Therefore, they call attention for the motives that these companies

communicate in order to change their motives for CSR engagement.

Activist Considerations: Company Size

Activists usually target prominent companies for several reasons (den Hond & de Bakker, 2007). They are highly visible in society, their operations raise concerns in society and, furthermore, these companies have a large impact on society (Rehbein, Waddock & Graves, 2004; den Hond & de Bakker, 2007). In addition, activists make use of the vulnerability of the reputation of prominent companies (van Cranenburgh, Liket & Roome, 2013). Indeed, scholars argue that prominent companies are highly dependent on their reputation in order to make higher profits (den Hond & de Bakker, 2007; Cranenburgh, Liket & Roome, 2013). For these reasons, activists target

prominent companies to achieve their goal, such as affecting the reputation or financial performance of prominent companies.

Since scholars argue that activists target prominent companies, I assume that prominent companies communicate their motives for CSR engagement with more care compared to smaller companies. Furthermore, I assume that prominent companies care more for positive CSR performance in order to reduce activism. Therefore, I assume that company size mediates the relationship between the motives for CSR engagement and stakeholder activism. In addition, I assume that company size influences the relation between CSR performance and stakeholder activism. These arguments lead to the following hypotheses:

Hypothesis 4: Companies with a self-regarding stakeholder culture are more likely to be targeted by activists than companies with an other-regarding stakeholder culture. This relationship will be stronger for larger companies than for smaller companies.

(21)

Hypothesis 5: Companies with lower CSR performance are more likely to be targeted by activists than companies with a higher CSR performance. This relationship will be stronger for larger companies than for smaller companies.

Activist Consideration: Investments for CSR Engagement

The type of investment can indicate what motives companies have to engage in CSR. Therefore activists look at the type of investments companies make when they decide to target companies (Barnett & Salomon, 2010). Knowing this, decisions about CSR relate to R&D and long-term investments, while decisions with regard to improving the economic position mainly relate to short-term investments (Arrigo, 2013; Waldron, Navis & Fisher, 2013; Surroca, Tribo & Waddock, 2006). As a result, companies that focus on short-term wealth maximization are more likely to be targeted by activists. These particular companies only view shareholders as their primary obligation and, thus, engage in CSR for their own self-interest. For this reason, activists aim to target these companies in order to destroy their financial performance (Waldron, Navis & Fisher, 2013). On the other hand, companies that invest in R&D are less likely to be targeted by activists. These companies invest in R&D because they aim to respond to stakeholders’ interests and, thus, engage in CSR for other-regarding motives. Thus, the amount of R&D investments can indicate to what extent companies aim to respond to stakeholders’ interests. In summary, these arguments lead to the following hypothesis:

Hypothesis 6: Companies with a self-regarding stakeholder culture are more likely to be targeted by activists than companies with an other-regarding stakeholder culture. This relationship will be stronger as R&D investments increases.

(22)

Concluding Remarks

A stakeholder culture depends on the beliefs, values and practices companies have to solve problems or manage stakeholder relationships (Waldron, Navis & Fisher, 2013). Therefore, companies can have a different stakeholder culture and respond differently to stakeholders’ interests (Jones, Felps & Bigley, 2007). Knowing this, some

companies prioritize to stakeholders’ interests, while others don’t take stakeholders’ interests into consideration (Jones, Felps & Bigley, 2007). Previous studies suggest that companies respond to stakeholders’ interests by communicating particular motives for CSR engagement (Jones, Felps & Bigley, 2007; Waldron, Navis & Fisher, 2013). Companies can face negative pressure from various stakeholders, as their motives don’t fully correspond to the interests of their stakeholders. As a result, stakeholder activism can occur (Smith, 2003; den Hond & de Bakker, 2007).

However, it isn’t clear whether responding to stakeholders’ interests actually reduce the extent of stakeholder activism. Therefore, this research tests whether a stakeholder culture can influence stakeholder activism.

(23)

Conceptual Model

In figure 2, I present the purposed hypotheses in a conceptual model. Furthermore, I also present the hypotheses in the table 1. These five hypotheses are analyzed and interpreted in order to answer whether a stakeholder culture can influence stakeholder activism.

(24)

Hypothesis 1: Companies with a self-regarding stakeholder culture are more likely to be targeted by

activists than companies with an other-regarding stakeholder culture.

Hypothesis 2: Companies with a self-regarding stakeholder culture have a lower CSR performance

than companies with an other-regarding stakeholder culture.

Hypothesis 3: Companies with a lower CSR performance are more likely to be targeted by activists

than companies with a higher CSR performance.

activists than companies with an other-regarding stakeholder culture. This relationship will be stronger for larger companies than for smaller companies.

Hypothesis 5: Companies with a lower CSR performance are more likely to be targeted by activists

than companies with a higher CSR performance. This relationship will be stronger for larger companies than for smaller companies.

activists than companies with an other-regarding stakeholder culture. This relationship will be weaker for companies that make R&D and long-term investments.

(25)

Methodology

Introduction

In this thesis I study how a stakeholder culture influences stakeholder activism. Therefore, I describe in this section how I collect data in order to answer this question and, furthermore, how I am going to test the purposed hypotheses. In order to do so, I divide this section into several sub-sections. I describe in a sub-section the research design; I also describe what sample I study and I also describe the data collection procedures of every relevant variable. As a result, the reader recognizes the industries I analyse in this research. Furthermore, the reader understands what databases I use to obtain data of the concepts I study and, in addition, the reader can distinguish the independent variables from the dependent, moderating and control variables that are used in this research.

(26)

Research Design

The design of this research is based on a database research. This quantitative research method enables me to do a statistical analysis and test the hypotheses. Furthermore, it enables me to look at the relationship between particular variables, such as the

relationship between a stakeholder culture and stakeholder activism. Furthermore, this type of analysis allows me to involve a relatively large sample. This enhances the generalisation of the results in my study (Saunders, Lewis & Thornhill, 2009). Since quantitative research is usually done in a short time period, I can spend more time on analysing and interpreting the obtained results with regard to the hypothesis.

However, this type of research has also a few limitations. At first, a database is sometimes hard to access (Saunders, Lewis, & Thornhill, 1997). Furthermore, the initial outcomes of qualitative research can be difficult to read and understand; it can be difficult for an average reader to distinguish statistical aspects (Saunders, Lewis, & Thornhill, 1997). This makes the process of describing the findings of this study time-consuming. Yet, a database research suits this research paper very well. Namely, there are databases that provide data to measure the given variables, such as CSR

performance, stakeholder activism, company size and R&D investments. Furthermore, a stakeholder culture is determined by coding the motives that companies communicate to engage in CSR. Then, the data is collected in order to perform several regression analyses with regard to the hypotheses.

(27)

The Sample Description

A control variable - In order to test the research topic through database research, I

select two widely discussed industries in society. These industries are the apparel industry and the food industry. These industries operate in advertising-intensive consumer markets and are, therefore, well known by consumers (van Cranenburgh, Liket & Roome, 2013; Fernandez-Feijoo, 2013). Furthermore, these industries can exert a direct impact on consumer behaviour (Peloza & Shang, 2011). Meaning that companies from these industries offer products directly to consumers and, therefore, influence the behaviour of consumers. Thus, one small mistake of these companies does also influence the behaviour of consumers (den Hond & de Bakker, 2007). As a result, discussions in society occur and, furthermore, activists can decide to target such companies (Smith, 2003; van Cranenburgh, Liket & Roome, 2013). For this reason, companies from these industries consider carefully how they translate stakeholders’ interests in their practices and decision-making with regard to CSR engagement (Jones, Felps & Bigley, 2007). As a result, they aim to avoid the chance of being targeted by activists and, furthermore, strive for maintaining their reputation in society.

I take a sample from these two industries to test the hypotheses. The number of the sample is equal to the number of companies that is mentioned in databases that I use. In other words, I select companies that are mentioned in all the databases. This means that I select companies that are mentioned in a database with data about CSR

performance (KLD database), a database with data related to stakeholder activism

(BHRRC database) and a database with data about financial and statistical

information (Compustat). Knowing this, the total sample of this research consists of

134 companies. Subsequently, I create a variable, known as industry, to distinguish apparel companies from food companies. In this respect, companies from the food industry get a 0, while companies from the apparel industry get a 1.

(28)

Data Collection Procedure

CSR and motives for CSR Engagement

An independent variable and a dependent variable - In order to collect data for CSR

performance (CSP), I use a database that provides an overview of companies´ corporate social performances. The KLD database provides such data. As a result of using KLD, I can argue whether companies have either high or low CSR

performance. Since the KLD database provides most of the data for the year 2012, it is tested how CSR performance in 2012 is influenced by a stakeholder culture maintained earlier in time.

The KLD database presents 159 different CSR factors that measures CSR

performance of companies. These CSR factors refer to either strengths or weaknesses towards stakeholders and company performance. In this respect, a positive strength indicates how well companies perform in meeting the expectation of one or more stakeholders. On the other hand, a positive weakness indicates how poorly companies perform in meeting the expectations of one or more stakeholders. As a result, KLD coded a positive strength as 1, while it also codes a positive weakness as a 1. Thus, both a positive strength and a positive weakness are coded as 1, while both a negative strength and a negative weakness are coded as 0. By obtaining scores that relate to positive strengths (1) and negative weaknesses (0), companies can get higher CSR performance.

I select all the CSR factors of the KLD database to determine an average of CSR performance. In order to calculate an average of CSR performance, I first transform the scores of weaknesses in such a way that they can be interpreted as scores of strengths. Namely, the initial outcomes of strengths and weaknesses should now be interpreted the other way around. Therefore, I reverse a weakness of 0 into 1 and I reverse a weakness of 1 into 0. As a result, negative weakness is now coded as a 1, while positive weakness is now coded as a 0. Subsequently, I take all the positive strengths (1) and transformed weaknesses (1) to calculate an average of CSR performance. Since KLD rates the CSR factors as 0 or 1, the average of CSR

(29)

performance close to 1 indicates positive CSR performance. On the other hand, an average of CSR performance close to 0 indicates negative CSR performance.

Stakeholder Culture

An independent variable - In order to determine a stakeholder culture of companies, I

use a coding scheme to identify a stakeholder culture of companies. Four bachelor students in Business Administration at the University of Amsterdam created this coding scheme. Therefore, they identified keywords that relate to profit-driven motives, value-driven motives and stakeholder-driven motives for CSR engagement (Appendix 1-3). In this respect, they identified keywords that several scholars

consider as a profit-driven motive, value-driven motive or stakeholder-driven motive for CSR engagement (Surroca, Tribo & Waddock, 2006; Jones, Felps & Bigley, 2007; Smith, 2003; Barnett & Salomon, 2010; Alniacik, Alniacik & Genc, 2011; Waldron, Navis & Fisher, 2013; Burchell & Cook, 2013; Kang & Hustvedt, 2014). For

instance, keywords that they consider as profit-driven motives for CSR engagement can be ‘shareholder’, ‘cost savings’ or ‘reputation’. They relate ‘green is good’, ‘save our earth’, ‘volunteerism’ to value-driven motives for CSR engagement. Keywords that they consider as stakeholder-driven motives for CSR engagement are ‘employee engagement’, ‘stakeholder value’ and ‘better respond to’.

Then, the students and I counted the number of each motive (Appendix 1 – 3) that companies mention in their annual reports 2011. In order to do so, the Bachelor students and I divided the number of companies so that everyone coded an equal number of annual reports. Since the sample consists of 134 companies (66 food; 68 textile), each of us coded either 26 or 27 companies. In this respect, I coded 27 annual reports.As a results, we can explain what particular motives each company

(30)

Subsequently, I only consider the number of profit-driven motives and stakeholder-driven motives to determine a stakeholder culture of companies (Appendix 1 and 3). In this respect, I define profit-driven motives as self-regarding motives for CSR engagement, while I define stakeholder-driven motives as other-regarding motives for CSR engagement. Based on the number of profit-driven motives and stakeholder-driven motives that companies communicate in their annual report, I determine whether companies respond to stakeholders’ interests or not. Thus, companies respond to stakeholders’ interests as they mention many stakeholder-driven motives in their annual report. On the other hand, companies act in their self-interest as they mention many profit-driven motives in their annual report. For instance, company X communicates 15 profit-driven motives and 5 stakeholder-driven motives in its annual report. As a result of this outcome, I conclude that company X has a self-regarding stakeholder culture.

Based on the analysis of annual reports, I do create another variable. This variable is called stakeholder culture. In this respect, companies using profit-driven motives in their annual report receive a 0, while companies using stakeholder-driven motives in their annual report receive a 1. This means that companies with a self-regarding stakeholder culture get a 0, while companies with an other-regarding stakeholder culture get a 1. Companies that use an equal number of motives receive a 0.5. Namely, these companies can’t be allocated to either one of these stakeholder cultures, but rather score in between these two.

Stakeholder Activism

A dependent variable - In order to determine stakeholder activism, I use a database

that provides data about companies that are targeted by activists. The BHRRC (Business and Human Rights Resource Centre) database provides such data. This database communicates information of companies that are targeted by activists for at least one time. By the use of this database, I can distinguish companies that are targeted by activists from companies that have never been targeted by activists. Since CSR performances are collected for the year 2012, stakeholder activism is determined from January 2012 till June 2015 in order to test a cause-effect relationship between

(31)

The BHRRC database mentions companies from different industries that are targeted by activists for at least one time. By the use of this database, I can count the number of activist campaigns towards companies. Subsequently, I can determine whether these campaigns aim to strengthen the reputation of these companies or aim to damage the reputation of these companies. Namely, the BHRRC database indicates the purpose of each activist campaign. Therefore, I classify every activist campaign as either a negative campaigns or a positive campaign. In this research, however, I test whether a stakeholder culture and CSR performance influence the number of negative campaigns. Yet, I also mention the number of positive campaigns and, as a result, the total number of campaigns. This total number of activist campaigns indicates the overall level of stakeholder activism towards companies. Furthermore, these variables may contribute to the explanation of the outcomes of several hypotheses in this research.

In order to determine stakeholder activism, I assume activist groups that the BHRRC database mentions. These activist groups include both national activist groups as (international) NGOs. These groups are also presented in appendix 4. As a result, I can now test how and whether stakeholder culture and CSR performance can trigger and, thus, influence these activist groups.

Company Size

A moderating variable - In order to determine the size of companies, I use a database

that provides data about the size of companies for the year 2011. Compustat provides such data. Namely, this database provides data about the number of employee, which in turn explains the size of companies. Knowing this, companies with a large number of employees are defined as prominent companies (Rehbein, Waddock & Graves, 2004; den Hond & de Bakker, 2007). As a result of using Compustat and collecting

(32)

result, I can test whether the motives to engage in CSR are influenced by company size and, as a result, also influence stakeholder activism. In addition, I can test whether company size moderates the relationship between CSR performance and stakeholder activism. As a result, I can draw conclusions with regard to the moderator effect of company size.

R&D Investments

A moderating variable - In order to determine whether companies focus on short-term

investments or long-term investments, I also select Compustat to collect data from. Compustat provides data with regard to the amount of R&D investments companies make in 2011. As a result of using Compustat, I can determine whether R&D

investments moderates the relationship between a stakeholder culture and stakeholder activism.

I interpret R&D investments as a continuous variable in order to determine its moderating effect. Therefore, I take the real values of R&D investments that Compustat gives. As a result, I can test whether the motives to engage in CSR are influenced by R&D investments and, as a result, influence stakeholder activism. As a result, I can explain whether R&D investments moderate the purposed cause-effect relationship.

(33)

Overview of Variables

As I describe the methodology of this study in this section, I also present an overview in which I define every variable that is mentioned in the hypotheses. Therefore, I define each variable as either a control, dependent, independent or moderator variable. This overview is shown in figure 2.

Table 2 - Overview of Variables

activists than companies with an other-regarding stakeholder culture.

Independent Variable Industries Control Variable

Stakeholder Culture Stakeholder Activism Industries

Hypothesis 2: Companies with a self-regarding stakeholder culture have lower CSR performance

than companies with an other-regarding stakeholder culture.

Stakeholder Culture CSR Performance Industries

Hypothesis 3: Companies with lower CSR performance are more likely to be targeted by activists

than companies with a higher CSR performance.

(34)

Independent Variable Industries Moderator Control Variable

Stakeholder Culture Stakeholder Activism Company Size Industries

Hypothesis 5: Companies with lower CSR performance are more likely to be targeted by activists

than companies with higher CSR performance. This relationship will be stronger for larger companies than for smaller companies.

CSR Performance Stakeholder Activism Company Size Industries

activists than companies with an other-regarding stakeholder culture. This relationship will be stronger as R&D investments increase.

(35)

Analyses and Results

Introduction

I report the findings of the statistical analysis in this section. These findings are the result of the hypotheses tests and are performed by the use of the statistical program SPSS. Furthermore, several preliminary tests are done to identify sources of error or abnormalities.

The first chapter explains how the missing values are handled in this study.

Subsequently, I create a new variable based on the number of other-regarding motives and self-regarding motives. This variable is called stakeholder culture and allows me to perform hypotheses with regard to a stakeholder culture of companies. Then, a description of the categorical and continuous variables is presented by the use of descriptive statics. Furthermore, the most significant correlations between the variables are reported. Therefore, I take the Pearson correlation coefficients into account. Then, I test the violations of normality and homoscedasticity. After performing these tests, I can perform hierarchical multiple regression analyses that test the hypotheses. This enables me to interpret the hypotheses.

(36)

Missing Values

The presence of missing data in a dataset is a very common problem when performing quantitative research. There are only a few articles that explain how researchers handle missing data in a dataset (Myers, 2011).

In this respect, my dataset also presents missing values. This is due to some databases that I say. Some of these databases don’t give information for the companies that I measure. In order to handle the missing data in my dataset, SPSS offers tools that handle these values in a consistent way. Out of these methods, I select a tool that replaces the missing values with the mean for the entire series. In other words, the missing values are replaced by a series mean of the variable. For instance, all the missing values of self-regarding motives are replaced by the series mean of this variable.

In this research, there are quite a lot of missing values. There are missing 148 values (Appendix 5). These missing values occur along the different variables. Based on the frequency command in SPSS, I can check the number of missing values of each variable that is mentioned in my dataset (Appendix 5). As a result, appendix 5 shows that some variables don’t have values for many companies, while other variables present values for almost every company. After replacing these missing values, I can use all the companies of my sample to test the hypotheses. As a result, the

(37)

Descriptive Summary

In this study, I use only one control variable. This variable is industry. The variables of other-regarding motives and self-regarding motives are recoded into one

dichotomous variable that is called stakeholder culture. As I explained in the previous section, companies communicating most other-regarding motives get a 1, while companies communicating most self-regarding motives get a 0. Companies communicating an equal number of other-regarding motives and self-regarding motives get a 0.5. Then, the other 9 variables - other-regarding motives,

self-regarding motives, total motives, CSP, the number of negative campaigns, the number positive campaigns, the total number of campaign, the number of employees and R&D investments - are classified as continuous variables.

134 companies are tested in this study. Therefore, there is first collected data for each company. Subsequently, the complete dataset is checked for irregularities. Thus, all 10 continuous variables are checked for irregularities. Then, the missing values of each variable are replaced by a series mean of the particular variable. As a result of replacing the missing values by a series mean, no company is removed from the dataset and can, therefore, be used to test the hypotheses.

Table 3 presents a descriptive summary of the control variable and categorical variable. Out of 134 companies, 66 companies belong to the food industry (49.3%), whereas 68 companies belong to the apparel industry (50.7%). Furthermore, all companies have a particular stakeholder culture. 59 companies have an other-regarding stakeholder culture (44%), while 32 companies maintain a self-other-regarding stakeholder culture (23.9%). Out of these two stakeholder cultures, most of the

companies communicate an other-regarding stakeholder culture. Yet, there are a lot of companies that use a similar amount of other-regarding motives and self-regarding motives for CSR engagement. Namely, 43 companies have a stuck-in-the-middle

(38)

Table 3 - Descriptive Statistics for the Categorical Variable and Control Variable

Variable Levels N %

Industry 0 = Food Industry 1 = Apparel Industry

66 68

49.3 50.7 Stakeholder Culture 0 = Self-Other Regarding Stakeholder Culture

0.5 = Stuck-in-the-middle Stakeholder Culture 1 = Other-regarding stakeholder Culture

32 43 59 23.9 32.1 44

Apart from this descriptive summary, I use 9 continuous variables in this study. These are the dependent variables - the number of negative campaigns, the number of positive campaigns, the number of total campaigns, the independent variables – other-regarding motives, self-other-regarding motives and total motives and the moderating variables – the number of employees and R&D investments. The variable - CSP is both used as a dependent variable and an independent variable. A descriptive summary of these variables is presented in table 4.

Table 4 presents information for the dependent, independent and moderating variables. The values of other-regarding motives range between 0 and 24. This variable has a mean of 3.74 and a standard deviation of 3.13. The values of self-regarding motives range between 0 and 22. This variable has a mean of 3.43 and a standard deviation of 4.21. The values of the total number of motives range between 0 and 33. This variable has a mean of 5.88 and a standard deviation of 6.66. The values for CSP range between 0.24 and 0.43. This variable has a mean of 0.30 and a standard deviation of 0.03. Then, the values of the number of negative campaigns vary

between 0 and 117. This variable has a mean of 4.75 and a standard deviation 14.96. For the number of positive campaigns, the values vary between 0 and 70. This variable has a mean of 2.83 and a standard deviation of 8.77. The values of the total number of campaigns range between 0 and 187. This variable has a mean of 7.57 and a standard deviation of 23.33. For the total number of employees the values vary between 153 and 420000. It has a mean of 28080 and a standard deviation of 50087. At last, the values of R&D investments range between 0 and 1570000000. It has a a

(39)

Table 4 - Descriptive Statistics for Continuous Variables

Variable N Min. Max. Mean Std. Dev.

Other-Regarding Motives 134 0 22 3.43 4.21

Self-Regarding Motives 134 0 24 3.74 3.13

Total Number of Motives 134 0 33 5.88 6.66

CSP 134 0.24 0.43 0.30 0.03 Number of Positive Campaigns 134 0 70 2.83 8.77 Number of Negative Campaigns 134 0 117 4.75 14.96

Total Number of Campaigns 134 0 187 7.57 23.33

Number of Employees 134 153 420000 28080 50087

R&D Investments 134 0 1570000000 47239777 169353124

Based on the descriptive summary in table 4, there are quite some interesting findings that I notice. The first interesting finding is that companies communicate more other-regarding motives (3.74) than self-other-regarding motives (3.43). Indeed, the descriptive summary in table 3 shows that most companies have an other-regarding stakeholder culture (44%). Furthermore, the minimal score of CSP (0.24) and the maximal score of CSP (0.43) aren’t very widespread. Since a score close to 0 relates to lower CSP, companies from the sample score relative low on CSP. Subsequently, the average of the number of negative campaigns (4.75) and the average of the number of positive campaigns (2.83) are relatively close to each. These scores are low compared to their maximum obtained score. Namely, the maximal score of the number of negative campaigns is 117, whereas the maximal score of the number of positive campaigns is

(40)

the sample. Thus, companies from the sample differ in the amount they invest in R&D.

(41)

Normality and homoscedasticity

I use a normality test to measure sampling adequacy among the continuous variables other-regarding motives, self-regarding motives, the total number of motives, CSP, the number of negative campaigns, the number of positive campaigns, the total number of campaigns, number of employees and R&D investments. A descriptive summary of the normality analysis can be found in appendix 6. This summary presents how the original data is distributed.

The independent variable other-regarding motives has a mean value of 3.74 and a 5% trimmed value of 3.44. This indicates that the more extreme scores didn’t have a significant impact on the mean. The skewness value for other-regarding motives is positive (2.786) and the kurtosis value is also positive (13.573). This shows that the scores of this variable are clustered at the left side of the graph. Furthermore, it shows that the data of this variable isn’t normally distributed (Appendix 6 & Appendix 7). Then, the independent variable self-regarding motives has a mean of 3.43 and a 5% trimmed value of 2.90. This indicates that the more extreme scores have a slight significant impact on the mean. The skewness value of self-regarding motives is positive (1.994) and the kurtosis value is also positive (4.422). This shows that the scores of this variable are clustered at the left side of the graph. Furthermore, it shows that the data isn’t normally distributed (Appendix 6 & Appendix 8). Subsequently, the independent variable total motives has a mean of 5.88 and a 5% trimmed value of 5.09. It shows that the more extreme scores have a significant impact on the mean. The skewness value of total motives is positive (1.774) and the kurtosis value is also positive (3.298). This explains that the scores of this variable are clustered at the left side of the graph. In addition, the distribution of the data of this variable isn’t normal (Appendix 6 & Appendix 9).

(42)

A descriptive summary of the dependent variables – the number of negative

campaigns, the number of positive campaigns and the total number of campaigns – also show some interesting findings. At first, the negative number of campaigns has a mean value of 4.75 a 5% trimmed mean of 1.91. This explains that extreme values have a significant influence on the mean. Given the positive skewness value (5.293) and the positive kurtosis value (31.973), the scores of this variable are clustered at the left side of the graph. Furthermore, the distribution of the data of this variable isn’t normally distributed (Appendix 6 & Appendix 11). Secondly, the number of positive campaigns has a mean value of 2.83 and a 5% trimmed mean of 1.20. This indicates that extreme values have a significant influence on the mean. Given the positive skewness value (5.101) and the positive kurtosis value (30.709), the scores of this variable are clustered at the left side of the graph. The distribution of the data of this variable isn’t also normal (Appendix 6 & Appendix 12). Thirdly, the total number of campaigns has a mean of 7.75 a 5% trimmed value of 3.18. This indicates that

extreme values have a significant influence on the mean. Given the positive skewness value (5.299) and the positive kurtosis value (32.805), the scores of this variable are clustered at the left side of the graph. At last, the distribution of the data of this variable isn’t normal (Appendix 6 & Appendix 13).

The first moderating variable – the number of employees – has a mean value of 28080 and a 5% trimmed mean value of 19931. This shows that the extreme values have a significant impact on the mean. Bases on the positive skewness value (5.241) and the positive kurtosis value (34.025), the scores of number of employees are clustered to the left side of the graph. The distribution of the data of this variable isn’t normal (Appendix 6 & Appendix 14).

The second moderating variable – R&D investments – has a mean value of 47239777 an a 5% trimmed mean value of 17796435. This indicates that the more extreme scores have a large significant impact on the mean. Based on the positive skewness value (6.713) and the positive kurtosis value (52.868), the scores of R&D investments are clustered to the left side of the graph. At last, the data of this variable isn’t

(43)

Subsequently, I use the Kolmogorov-Smirnov statistics test to assess the normality of the distribution of the obtained scores. The results of this test are given in appendix 16. The values of the independent variables - other-regarding motives and self-regarding motives - are significant. This shows that the data of these variables isn’t normally distributed. Furthermore, the histograms of these variables also indicate that the data isn’t normally distributed (Appendix 7 & Appendix 8). Subsequently, the data of the other variables is also not-normally distributed. This means that the

skewness values are extremely high and the histograms also indicate that the data isn’t normally distributed (Appendix 6 & Appendix 9 - 16).

Due to the relatively high skewness value, I conclude that all the variables aren’t normally distributed (Appendix 6 - 16). In order to test the hypotheses, I need to transform this data so that the values of the variables become normally distributed. There are various transformation methods that I can use to correct for skewing. Out of these methods, the LOG10 (1+variable) seems the most suitable transformation (Appendix 17). Namely, this transformation method presents the most representative outcomes of the variables. As a result of performing this transformation, the skewness values of the variables decrease and, furthermore, the histograms look more

representative (Appendix 17 - 26). In addition, I used the Kolmogorov-Smirnov test again to assess the normality of the distribution of the scores. These scores are presented in appendix 27. The obtained scores in the Kolmogorov-Smirnov test still indicate that the data isn’t normally distributed. However, this transformation method is most representative compared to the other transformation methods. Therefore, I apply this transformation method in order to construct a correlation matrix and perform several regression analyses.

By performing these different tests, I can now construct a correlation matrix and test the hypotheses. Again, I use the transformed data to compute a Pearson correlation

(44)

Correlation Matrix

As I explained what descriptive statistics I use (Appendix 17 - 26), I can now prepare a correlation matrix. In order to prepare this matrix, I select the transformed values of all the continuous variables. These variables are transformed in the previous section. Furthermore, I select the values of the control variable and the categorical variable. This means that I select all the variables that I introduce in the hypothesis. I also select the variables that are the result of the initial data collection. For instance, I count the total number of campaigns by the use of the BHRRC database and then classify these campaigns as either a negative campaign or a positive campaign. Although I only use the number of negative campaigns to test some hypotheses, a correlation between a particular variable and the number of positive campaigns can contribute to the explanation of the outcome of several hypotheses. For this reason, I also explain other findings outside the findings that can be related to one of the hypotheses. Furthermore, I also mention the variables other-regarding motives and self-regarding motives separately in the correlation matrix. Namely, the size of these variables can influence the outcomes of stakeholder culture and, therefore, can also contribute to the explanation of the outcome of several hypotheses. Thus, I present the most important correlations with regard to my research topic and the given hypotheses.

The correlations between the research variables are investigated by analyzing the Pearson correlation coefficients. These correlations are shown in figure 3 at the end of this section. The significant correlations at a 5% level are labeled with one asterisk (*), while the significant correlations at a 1% level are labeled with two asterisks (**). A low significant level indicates a consistent relationship between variables. For instance, there is a consistent relationship between other-regarding motives and the total number of motives (P-value < 0.05). This correlation explains a significant amount of variance in the variable the total number of motives (P-value < 0.05). This outcome is somewhat expected. Namely, other-regarding motives can determine the total number of motives. Since the correlation matrix only indicates a correlation between variables, it is still not clear if other-regarding motives cause more variance in the total number of motives or if other-regarding motives causes less variance in the total number of motives.

(45)

Independent Variables and Control Variable

The first notable finding in figure 3 explains that the independent variable stakeholder culture correlates to the control variable industry, while the other independent

variable CSP doesn’t correlate to industry. Thus, stakeholder culture is negatively correlated to industry (Pearson correlation coefficient = -0.237**; N = 134; Sig. = 0.006), while CSP isn’t significant correlated with industry (Pearson correlation coefficient = -0.144; N = 134; Sig. = 0.096).

Independent Variables and Dependent Variables

The second notable finding in figure 3 is that stakeholder culture isn’t correlation with the number of negative campaigns (Pearson correlation coefficient = 0.039; N= 134; Sig. = 0.652). Although, there is no correlation between these two variables, the number of other-regarding motives positively correlates with the number of negative campaigns (Pearson correlation coefficient = 0.229**; N = 134; Sig. = 0.008), Then, the total number of motives is also positively correlated with the number of negative campaigns (Pearson correlation coefficient = 0.194*; N = 134; Sig. 0.025). However, the number of self-regarding motives isn’t significantly correlated with the number of negative campaigns (Pearson correlation coefficient = 0.130; N = 134; Sig. = 0.135).

The correlations in figure 3 also shows that stakeholder culture isn’t correlated with CSP (Pearson correlation coefficient = 0.020; N = 134; Sig. = 0.819). However, other-regarding motives is positively correlated with CSP (Pearson correlation coefficient = 0.209*; N = 134; Sig. = 0.00). Furthermore, CSP is also positively correlated with the number of negative campaigns (Pearson correlation coefficient = 0.527**; N = 134; Sig. = 0.00).

Independent Variables and Moderating Variables

(46)

Additional Findings

I only use the variable number of negative in my hypotheses. This means that I test the relationship between some independent variables and the dependent variable number of negative campaigns. However, the correlation matrix shows some interesting correlations between independent variables and the dependent variable number of positive campaigns. At first, other-regarding motives is positively correlated to the number of positive campaigns (Pearson correlation coefficient = 0.211*; N = 134; Sig. = 0.014). Then, CSP is also positively correlated to the number of positive campaigns (Pearson correlation coefficient = 0.594**; N = 134; Sig. 0.00). Thus, the other-regarding motives and CSP are positively correlated to the number of positive campaigns. These independent variables are positively correlated to the number of positive campaigns and the number of negative campaigns. Although, I don’t test any relationship between these independent variables and the number of positive campaigns, these findings reveal interesting information about a possible relationship between stakeholder culture, CSP and stakeholder activism.

Another interesting finding in the correlation matrix relates to CSP and the amount of R&D investments. Although, I don’t test the moderating effect of R&D investments on the relationship between CSP and the number of negative campaigns, the

correlation matrix presents a positive correlation between CSP and the amount for R&D investments (Pearson correlation coefficient = 0.264**; N = 134; Sig. 0.02). This result may indicate a possible relationship between CSP and the number of negative campaigns.

(47)

1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11 1 M ot iv es 0. 01 5 1 ot iv es 0. 24 5* * 0. 37 2* * 1 f M ot iv es 0. 04 2 0. 47 8* * 0. 44 2* * 1 tu re -0. 23 7* * 0. 15 1 -0. 74 2* * -0. 18 6* 1 -0. 14 4 0. 20 9* 0. 14 4 0. 15 8 0. 02 0 1 iti ve C am pa ig ns -0. 19 1* 0. 21 1* 0. 13 0 0. 29 3* * 0. 00 8 0. 59 4* * 1 at iv e C am pa ig ns -0. 22 2* * 0. 22 9* * 0. 09 4 0. 19 4* 0. 03 9 0. 52 7* * 0. 88 4* * 1 f C am pa ig ns -0. 21 8* 0. 22 6* * 0. 09 7 0. 22 8* * 0. 04 3 0. 56 4* * 0. 93 9* * 0. 98 1* * 1 pl oy ee s -0. 12 8 0. 18 7* 0. 04 4 0. 04 5 0. 09 3 0. 32 4* * 0. 30 9* * 0. 34 1* * 0. 36 5* * 1 en ts -0. 41 0* * 0. 05 6 -0. 04 1 -0. 20 8* 0. 08 1 0. 26 4* * 0. 10 7 0. 14 4 0. 13 9 0. 26 7* * 1 T ran sf or m ed D at a if ic an t a t a 0 .0 5 le v el ( 2-tai le d) . * * C or re lat ion s ign if ic an t a t a 0 .0 1 le v el ( 2-tai le d)

The influence of stakeholder culture on stakeholder activism : a quantitative research paper focusing on companies in the apparel industry and food industry

- Master Thesis -

Statement of Originality

Abstract

Table of Contents

Index of Figures and Tables

Introduction

Theoretical Framework

Conceptual Model

Methodology

Analyses and Results