Brand Evangelism

(1)

Rijksuniversiteit Groningen

Brand Evangelism

“ Developing a measurement scale for Brand Evangelism”

By: M.W. van der Haar July 2011

(2)

University of Groningen, Faculty of Economics and Business, July 2011 Pagina 2

Brand Evangelism

by

M.W. van der Haar

University of Groningen

Faculty of Economics and Business

Msc Business Administration, Marketing Management

July, 2011

Marten van der Haar 1st Supervisor: Dr. K.J. Alsem

Graaf Adolfstraat 61B 2nd Supervisor: Prof. Dr. J.C. Hoekstra 9717ED 61B

06-28880859

(3)

Preface

This Master Thesis is my final work in order to graduate from the Faculty of Economics and Business with the Master Marketing Management at the University of Groningen. After finishing my HBO study “Management, Economics and Law” I had high expectations for studying at the University. Eventually, these expectations came through and the academic research has totally changed my perspective on conducting research.

Although there were some struggles while working on this Thesis, I really liked writing it. Especially since I was able to choose my own subject and work this out. Ever since I have read Martin Lindstrom’s book “Brand Sense”, I have been interested in Brand Religion. With this thesis I hope to make a contribution to the academic research about Brand Evangelism and in the understanding of Brand Religion.

Since I was not able to write this thesis alone I would like to thank some persons for their contribution. First of all I would like to thank Dr. K.J. Alsem for his guidance while I was writing this thesis, without his help I would not have been able to achieve this result. Furthermore, I want express my gratitude towards Prof. Dr. J.C. Hoekstra for her critical an clear feedback which helped to improve the quality of this thesis. Next I would like to thank Theije for the discussions about our thesises during the important coffee brakes. Moreover, I would also like to thank Theije, Gert-Jan, Erik and Jos for reading through this thesis to check for any inconsistencies or spelling mistakes. In addition I would like to express my gratitude towards my parents for there financial and moral support. Finally, I want to thank my friends for their support and understanding while I was writing this thesis.

I hope this thesis will bring some contribution towards the academic understanding of Brand Evangelism and finally, I hope that someday I will be able to use the findings of this thesis in my career as “Online Marketing Consultant”.

(4)

Management Summary

In the last decennia Brand Evangelism has been increasing in popularity, especially within business literature. However, until recently, there has been little academic literature focusing on Brand Evangelism. Therefore this thesis developed a measurement scale in order to measure Brand Evangelism. Furthermore, this research wants to see how high Brand Evangelism scores on Word of Mouth (WOM) compared to other popular WOM spreaders. According to Godes et al. (2005), the most commonly known WOM spreaders are: “Opinion Leaders, Market Mavens and Innovators”. According to this, the following research question was stated:

Does Brand Evangelism have a higher positive relationship with WOM than Opinion Leadership, Market Mavenism and Innovatorship?

In order to see if there were any factors influencing the relationship between Brand Evangelism and WOM, this research included Consumers Uniqueness, Brand Attachment and Network Breadth into the conceptual model. However, extensive testing did not show any significant results.

In the literature review, WOM is described as an exchange of information about a brand throughout all the possible channels and observations. Furthermore, the antecedents of WOM are described as following. The Innovator talks more about the introduction of new products and new technologies (Feick & Price, 1987), whereas the Opinion Leader focuses on the more popular products. In addition, The Market Maven knows and talks more about the whole market or industry (Hoyer & MacInnis, 2007).

(5)

University of Groningen, Faculty of Economics and Business, July 2011 Pagina 5 In order to answer the main research question this thesis used a internet survey which was spread amongst students from the University of Groningen. According to academic literature (Belk & Tumbat, 2005), Apple possesses a large pool of Brand Evangelists. Therefore, respondents could only participate if they owned one or more Apple products.

A Cronbach's alpha and factor analysis showed that the constructs of Brand Evangelism significantly measure something different than the antecedents of WOM. Thereby, this research proves that Brand Evangelism is different from the Opinion Leadership, Market Mavenism and Innovatorship scales. Furthermore, regression analyses showed higher scores on WOM for Brand Evangelism than the antecedents of WOM. Moreover, an ANOVA analysis showed that Brand Evangelism scores significantly higher on WOM than the other antecedents of WOM. Therefore, it can be proven that Brand Evangelism scores significantly higher on WOM than the antecedents of WOM.

Additionally, Brand Evangelism also had a significantly high relationship with “years of using Apple products”, “hours of using Apple products and “the amount of Apple products used”. Therefore it can be concluded that Brand Evangelism is formed over the years and increases with the amount of times the consumer uses the products. Unfortunately, there were not a lot of respondents who scored high on the only the Brand Evangelism or one of the antecedents scales. Therefore, future research should investigate whether the results are also significant for a larger group of respondents.

(6)

Index

Preface ... 3

Management Summary ... 4

Index ... 6

1. Introduction ... 8

1.1 Brand Evangelism ... 8 1.2 Problem statement ... 10 1.3 Structure ... 11

2. Conceptual model ... 12

3. Literature review ... 13

3.1 Word of mouth ... 13 3.1.1 Antecedents of WOM ... 14 3.2 Brand Evangelism ... 17 3.2.1 Brand Religion ... 17 3.2.2 Brand Evangelism ... 17 3.2.3 Consumer Devotion ... 18

3.2.4 Behavioral acts of devotion ... 20

3.2.5 Hypothesis ... 22

3.3 Brand attachment. ... 23

3.3.1 What is Brand attachment? ... 23

3.3.2 Measuring Brand Attachment ... 25

3.3.3 Influence of Brand attachment ... 26

3.4 Consumer uniqueness ... 27

3.5 Network breadth... 28

4. Research Methods ... 29

4.1 Research Design ... 29

4.2 Sample and data collection ... 29

4.3 Operationalisation of constructs ... 30 4.3.1 WOM ... 30 4.3.2 Antecedents of WOM ... 31 4.3.3 Brand Attachment ... 31 4.3.4 Uniqueness ... 31 4.3.5 Network breadth ... 31

(7)

University of Groningen, Faculty of Economics and Business, July 2011 Pagina 7 4.4.1 Collecting ... 32 4.4.2 Rival Brand ... 32 4.4.3 Rituals ... 32 4.4.4 Fandom ... 32 4.4.5. Gift giving ... 33 4.4.6 Dependence ... 33 4.5 Control variables ... 33 4.6 Data analyses ... 34 4.7 Reliability analysis ... 35

5. Research Results ... 36

5.2 Descriptive statistics ... 37 5.3 Hypothesis ... 38 5.4.1 Brand Evangelism ... 38 5.4.2 Brand Attachment ... 46 5.4.3 Consumer Uniqueness ... 47 5.4.4 Network breadth ... 49 5.5 Other results ... 50 5.5.1 Gender ... 50

5.5.2 Years using Apple products ... 51

5.5.3 Hours using Apple products per week ... 52

5.5.4 Amount of Apple products ... 52

5.6 Evaluation ... 53

5.6.1 Brand Evangelism ... 53

5.6.2 Brand Attachment ... 54

5.6.3 Consumer Uniqueness ... 54

5.6.4 Network Breadth... 55

6. Conclusion and Recommendations ... 56

6.1 Conclusions... 56

6.2 Limitations and further research ... 57

6.3 Managerial recommendations ... 58

References ... 59

Appendix I... 67

(8)

1. Introduction

During the past decennia the involvement of consumers with brands has increased rapidly. Throughout the information age it became clear that brands are becoming more than a utility. Although fan clubs were also present before PC’s entered our homes, the internet was able to join more fanatical/devoted consumers together in weblogs and internet communities. From hereon rallies were organized, joining hundreds of fanatics together to express their feelings about their brand. The Lego building contest and gatherings organized by Lego fanatics are some good examples. Since the last ten to twenty years, companies (like Lego) are starting to recognize the potential of these communities and are beginning to give them more attention (Collin and Murphy, 2009). Another good example is how Apple created a group of enthusiastic, paid software developers who were turned into Evangelists to spread the word and grow their developers' pool (Kawasaki, 2007). Although not intended, the word spread rapidly, whereby it was also turning customers into fanatics and Evangelists. Consumers who score high on Brand Evangelism are considered to be Brand Evangelists. These Evangelists are supposed to have a high positive relation with Word of Mouth (WOM), since Evangelists are motivated to tell everyone their story (Matzler, Pichler and Hemetsberger, 2007). This could be a powerful tool for companies to grow their brand. Therefore, in the last decennia several books and papers were written on how companies can turn these loyal costumers into a voluntary salesforce (Lindstrum, 2005; 2008; McConnell and Huba, 2007). However, until recently there has been little academic literature about Brand Evangelism (Collins and Murphy, 2009). This thesis will find out if it is possible to create a sound Brand Evangelism scale and if Brand Evangelism has a positive relationship with WOM.

1.1 Brand Evangelism

(9)

University of Groningen, Faculty of Economics and Business, July 2011 Pagina 9 has with its consumer. In addition, Lindstrom (2005) adds that when such a brand would experience any difficulties, consumers will rush to its rescue to save their brand. Although this sounds unrealistic, it is not very unlikely. When you look at brands like Harley Davidson and Lego, it is clear that the devoted consumers are just as much part of the company as the board of directors. When either of these companies are threatened to go bankrupt, this will result in a many unhappy consumers. More recently, Lindstrom takes his vision a step further in his next book, Buy-ology (2008). Where he measures, by means of neuroscientific tools, the reactions of consumers' brains towards multinational brands and religious symbols. Remarkably, the reactions of a devoted consumer as compared to a devoted Christian where considerably similar. Although these findings are remarkable, there has not been much literature about Brand Evangelism and how it can be measured. According to Collins and Murphy (2009), this lack of research suggests that academia neither supports nor disproves the business literature that is written about Brand Evangelism. For instance, there is no research supporting the proposition that Brand Evangelism has a positive relationship with WOM. In addition, literature exists which suggests that not all Brand Evangelists are intrinsically motivated to promote their brand (Macinnis, 2009). It could be that some consumers rather keep their favorite brand for themselves. Moreover, Brand Evangelism also supports rivalry between brands (Pimentel & Reynolds, 2004; Belk et al., 1989; Chung et al., 2008), whereby consumers openly speak out against consumers who prefer a different brand. This kind of rivalry is explicitly shown between the world’s biggest soft drinks suppliers: Coca-Cola and Pepsi.

(10)

University of Groningen, Faculty of Economics and Business, July 2011 Pagina 10 Opinion Leaders to promote their brand. This could be a good strategy for growing a brand or company. However, it is not clear how strong the relationship really is between Opinion Leaders and WOM. It could be that Brand Evangelists have a significantly higher positive relationship with WOM than Opinion Leaders. Therefore, this thesis will compare the scores of Brand Evangelism on WOM with Opinion Leadership, Market Mavenism and Innovatorship.

In addition, it is interesting to see if there are any moderating variables influencing these relationships. According to Godes et al. (2005) some of the primary moderating variables influencing WOM are consumer uniqueness and network breadth. In their research, they describe that consumer uniqueness is a way for consumers to express themselves. When consumers have a high level of uniqueness, they could be reluctant to engage in WOM, because they want to be one of the few who own that product. Furthermore, network breadth is seen as a primary moderating variable. According to Godes et al. (2005) consumers who exhibit a high network breadth are more likely to engage in WOM. Furthermore, Park et al. (2010) state that Brand Attachment has a high relationship with the promotion of a brand and therefore with WOM. They state that consumers with high brand attachment were more likely to promote their brand than consumers who were not attached to this brand. In this thesis, these three moderating variables will be tested to see if they have an influence on the relationship between Brand Evangelism and WOM. In order to measure these relationships, the following problem statement has been formulated.

1.2 Problem statement

This thesis is going to explore if there is a positive relationship between Brand Evangelism and

(11)

University of Groningen, Faculty of Economics and Business, July 2011 Pagina 11 constructs. Later on, these questions will form the Brand Evangelism construct which will be compared with WOM and more specifically, with the antecedents of WOM. This leads us to the following research question:

Does Brand Evangelism have a higher positive relationship with WOM than Opinion Leadership, Market Mavenism and Innovatorship?

1.3 Structure

(12)

2. Conceptual model

From here on the following conceptual model is determined, see fig. 1. In this model, the relationship between Brand Evangelism and Word Of Mouth (WOM) will be tested. This research hypothesizes that there will be a positive relationship (H1a). Furthermore, it will test if Brand evangelism is significantly different from the antecedents of WOM (Opinion Leadership, Market Mavenism and Innovatorship) (H1b). Thereafter, this research will test if respondents who score high on Brand Evangelism engages more in WOM than the antecedents of WOM (H1c). Finally, the moderating variables Brand attachment, Uniqueness and Network breadth will be tested to see if these influence the relationship of Brand Evangelism and WOM (H2;H3;H4).

After constructing the conceptual model, the next chapter will discuss the literature review where the hypotheses are derived from.

H3 H4 H2 H1b H1c The Antecedents of Word of mouth + - + Brand Evangelism Opinion Leadership Market Mavenism Innovatorship Brand Attachment Uniqueness Word of mouth Network Breadth + H1a

(13)

3. Literature review

Within this chapter the literature framework of this thesis will be described. First of all, the dependent variable WOM will be set apart, along with the antecedents of WOM. Secondly, the main construct Brand Evangelism is discussed thoroughly. Finally, the moderating variables Brand Attachment, Consumer Uniqueness and Network Breadth are described, along with their relationship with WOM.

3.1 Word of mouth

Word of mouth (WOM) is widely known within the marketing and consumer behavior literature. According to several authors, WOM is the ultimate product success factor which helps sell your product without almost any cost (Day, 1971; Katz and Lazarsfeld, 1955; Feick and Price, 1987). De Matos and Rossi (2008), cite in their paper that WOM communications are defined as “informal communications directed at other consumers about the ownership, usage or characteristics of particular goods and services and/or their sellers” (Westbrook, 1987, p. 261). Another popular cited definition about WOM is given by Arndt (1967), where WOM is defined as “oral, person-to-person communication between a perceived non-commercial communicator and a receiver concerning a brand, a product or a service offered for sale”. However, Godes et al. (2005) argue that these traditional WOM definitions, similar to the ones provided by Katz and Lazarsfeld (1955) and Granovetter (1973) are too narrow to capture the whole phenomenon. Godes et al. (2005) states that the one-to-one and face-to-face exchanges of information is outdated with the use of modern technologies like e-mail, mobile communication and the internet. Furthermore, they argue that it also ignores consumers who observed other consumers and thereby imitate the observed persons. Finally Godes et al. (2005) argues that WOM can be positive and negative. Since there are numerous definitions about WOM, it is hard to give one sound definition. This thesis will use the definition of Godes et al. (2005) as the central definition for WOM. Furthermore, WOM will be measured using the measurement scale of Harrison-Walker (2001).

(14)

University of Groningen, Faculty of Economics and Business, July 2011 Pagina 14 of characteristics who are more likely to distinguish information, namely Opinion Leadership, Market Mavenism and Innovatorship (Feick and Price, 1987). These characterizations of consumers are also referred to as the antecedents of WOM (Godes et al., 2005). From here on, when this thesis addresses the antecedents of WOM, we refer to measurement scales of Opinion Leadership, Market Mavenism and Innovatorship. Furthermore, consumers who score high on these scale are considered to be Opinion Leaders, Market Mavens or Innovators. These antecedents of WOM are described in the next paragraph.

3.1.1 Antecedents of WOM

In everyday life, numerous consumers engage in WOM. One might argue that every consumer, during their life, has somehow participated in WOM about some kind of company. However, there are some consumers who significantly engage more in WOM then others. As described previously, marketing literature distinguishes three characteristics of consumers who are more likely to engage in WOM than normal consumers, namely: Opinion Leadership, Market Mavenism and Innovatorship (Feick and Price, 1987; Godes et al., 2005). These types of consumers are described in the following subsections.

Opinion Leadership

(15)

University of Groningen, Faculty of Economics and Business, July 2011 Pagina 15 “Opinion Leadership occurs when individuals try to influence the purchasing behavior of other consumers in specific product fields, whereas opinion seeking happens when individuals search out advice from others when making a purchase decision” (Flyn et al., 1996 p. 138). According to Hoyer and MacInnis, (2007) Opinion Leaders can score high on both factors, simply because they serve as an information broker wherein they ask as much as they spread information across the field. In this thesis, the focus will be on information providing, hence Opinion Leadership. To measure Opinion Leadership the scale from Flynn et al. (1996) will be used (this scale is also found in Hoyer and MacInnis, 2007). Finally, Godes and Mayzlin (2007) found that loyal Opinion Leaders create more WOM than other loyal customers. Therefore, this thesis proposes that Opinion Leaders with high Brand Attachment will score higher on the WOM scale than Opinion Leaders who score low on Brand Attachment.

Market Mavenism

(16)

University of Groningen, Faculty of Economics and Business, July 2011 Pagina 16 Innovatorship

The third popular spreaders of WOM are Innovators. Because Innovators are the first consumers who adopt certain products, they often tell others about their experiences with the product. This often results in product reviews on the Internet. The Innovators are part of the diffusion process proposed by Rogers (1962). He described that Innovators are the first consumers (2,5 percent of the market) who adopt a new product. They are characterized by their enthusiasm about technology and wanting to be the first to use a new high tech product, even though these could contain some bugs or inefficiencies (Hoyer and MacInnis, 2007). After the innovators comes the early adopters, these consumers mainly look at the abilities these new technologies can provide them, in order to make their lives easier. Although these consumers know that new products will become less expensive in the future, they refuse to wait before buying them (Hoyer & MacInnis, 2007). Both Innovators and early adopters provide WOM about new products. In the literature of WOM these two consumer groups are often combined, since they posess similar characteristics (Feick & Price, 1987). In this thesis, the concept of Innovators will refer to both Innovators and early adopters. According to Feick & Price (1987) Innovators talk about products because of its novelty, the desire to look like a pioneer, the experience and involvement that comes from using the product, and to confirm their assessment of the product. One important notion about Innovators is that, contrary to Opinion Leaders and Market Mavens, not all Innovators actively engage in WOM. Furthermore, early adopters are product specific, which means there is no general early adopter (Robertson, 1971). In this thesis, Innovators are suggested to engage the least in WOM compared to Opinion Leaders and Market Mavens. Additionally, it is likely that Innovators have a positive relationship with Opinion Leaders and Market Mavens. In order to measure Innovatorship the scale of Goldsmith and Hofacker (1991) will be used.

(17)

3.2 Brand Evangelism

This chapter describes the construct of Brand Evangelism. First the overall topic Brand Religion will be explained, then the definition of Brand Evangelism will be given, next a literature review of Consumer Devotion will follow and finally the behaviors of Brand Evangelism will be described.

3.2.1 Brand Religion

According to Martin Lindstrum (2005) Brand Religion is the ultimate goal brands should pursue. In his book, he describes that Religions are the ultimate brands, in a sense that consumers connect with religions on a different level. In his more recent book, Lindstrum (2008) tries to find a connection between religions and multinational brands. In what he calls the biggest neuromarketing study ever conducted, he explores the reasons of why consumers buy products. This project was called buy-ology. During this study he and two top researchers (Gemma Calvert and Richard Silberstein) found out that there are some brand symbols who provoke the same neurological response in people’s brains as would religious symbols, like a cross or a bible. In this book he states: “Bottom line, there was no discernible difference between the way the subjects’ brains reacted to powerful brands and the way they reacted to religious icons and figures” (Lindstrum, 2008 p. 124-125). This proves that brands can become a religion-like following. Furthermore, according to McConnell and Huba (2007), these Brand Evangelists can be converted into a voluntary sales force by promoting the brand without receiving any monetary value. These evangelic consumers essentially become missionaries for the brand and thereby spreading positive WOM about the brand. In this chapter the behaviors of these “religious followers” will be discussed which leads to a measurement scale to recognize Brand Evangelism.

3.2.2 Brand Evangelism

(18)

University of Groningen, Faculty of Economics and Business, July 2011 Pagina 18 (2007). In addition, they state that “consumers who evangelize are passionate about their brand and feel the need to share their emotions with others”. Whereas, WOM focuses on the positive or negative communications about the ownership, usage or characteristics of the Brand (Westbrook, 1987), Brand Evangelism engages more in the evangelistic part of convincing consumers to use the Brand. Furthermore, Matzler et al. 2007 states that the evangelistic part of WOM is underestimated and that consumers who score high on Brand Evangelism are far more likely to engage in evangelic forms of persuading people to use their Brand. Moreover, Matzler et al. (2007) found that there is a positive relation between extravert consumers and brand evangelism, making extravert consumers effective brand advocates, especially when they are passionate about the brand. These brand advocates are also know and referred to as Brand Evangelist. Therefore, consumers who score high on the Brand Evangelism scale are considered to be a Brand Evangelist (Collins and Murphy, 2009).

Although Brand Evangelism is increasing in popularity, there is still a lack of academic literature about the subject. According to Collins and Murphy (2009), this could suggest that academia neither support nor disprove the business literature that is written about the subject. However, there is a lot of literature about Consumer Devotion. Because Brand Evangelism and Consumer Devotion are quite similar, the next paragraph will discuss how Consumer Devotion evolved and why it can be important for business.

3.2.3 Consumer Devotion

(19)

University of Groningen, Faculty of Economics and Business, July 2011 Pagina 19 According to Pimentel and Reynolds (2004), Consumer Devotion can be seen as a possible extension to brand equity. While brand loyalty is an important asset contributing to a brand’s equity, Aaker (1991) describes five levels of brand loyalty, wherein the highest level is the “committed buyer”. Pimental and Reynolds, (2004) propose, that the “ultimate committed consumer” is a “devoted consumer”. In their paper, they state that as the secularization of religion and the sacralization of the secular takes place, consumers' brands may take on aspects of “the sacred” for their devoted followers (Belk et al., 1989). An example of this phenomenon comes forth when the logo of a brand becomes a symbol, whereas the consumer will display the logo on items from product categories different from that of the branded product, as for example an automobile logo on a T-shirt (Pimental & Reynolds, 2004). Furthermore, they argue that with “Consumer Devotion” the commitment of the consumer towards the brand is essentially permanent.

According to Pimentel and Reynolds (2004), Consumer Devotion can be compared to a cult following (Carr, 1996). They state that Harley Davidson is often mentioned as a brand that has gained intense consumer loyalty. Furthermore, some television shows like The X-Files and Star Trek have gained similar loyalty form their fans. Kozinets (2001) discovered that several parallels between religion and the sacred exist within the Star Trek fan clubs. These Star Trek clubs profess a goal of helping to bring about the idealized future portrayed in the television series and movies, wherein fans cite the show as a source of moral and spiritual guidance (Kozinets, 2001). In their study, Pimentel and Reynolds (2004) discovered that the same attitudes, affects and behaviors were displayed by fans of sports teams, particularly in college football. They developed a model of Consumer Devotion based on these kind of fans. Herein, they propose that Consumer Devotion occurs when there is internalized affective commitment accompanied by proactively sustaining rituals. This means that consumers have an affective commitment with the brand and are proactively engaging in rituals in order to sustain this relationship. Therefore, consumer devotion can be defined as: an affective commitment accompanied by proactive sustaining rituals (Pimentel and Reynolds, 2004).

(20)

University of Groningen, Faculty of Economics and Business, July 2011 Pagina 20 Depending on the intensity of the components, different facets and the intensities of devotion that might exist (Pichler & Hemetsberger, 2007). To conclude, Pichler and Hemetsberger, (2007) state that “acts of devotion perpetuate and refuel the loving feeling for the loved, sacred object”.

3.2.4 Behavioral acts of devotion

In the following, several acts of Consumer Devotion, described in the literature above, will be discussed. These behaviors and characteristics are used to distinguish Brand evangelist from normal consumers. Consumers who score high on these dimensions of Brand Evangelism may be considered to be Brand Evangelists.

Collecting

One of the most reoccurring behaviors in Consumer Devotion is the tendency to collect products from the brand to which the consumer is devoted (Belk et al., 1989; Pimentel & Reynolds, 2004; Chung, Beverland, Farrelly & Euster, 2008). Although not all of the collected items are regarded as sacred, the collection as a whole is considered very special and unique (Belk et al., 1989). Furthermore, Chung et al. (2008) found that collecting certain products from a certain brand can be seen as an end in itself, rather than a means to an end (Lehmann, 1987). Meaning, consumers sometimes are collecting for the sake of collecting.

Rival brand

Devotion towards a brand often means active opposition towards a rival brand (Pimentel & Reynolds, 2004; Belk et al., 1989; Lindstrum, 2005). These rivalries reflect the property of sacredness termed opposition to the profane, which means that the sacred brand is seen as better then the opposition profane rival brand (Pimentel & Reynolds, 2004). Furthermore, the presence of a rival brand unites the devoted consumers (Lindstrum, 2005).

Rituals

(21)

University of Groningen, Faculty of Economics and Business, July 2011 Pagina 21 that some consumers used their product with an extreme level of care. This could be seen as a ritual for using the product.

Fandom

Furthermore, several articles have shown that devoted consumers are often depicted as fans (Pimentel & Reynolds, 2004; Kozinets, 2001; Chung et al., 2008). Although some devoted consumers would not depict themselves as a fan, they often show signs of being one (Chung et al., 2008). Among these fans, there is a distinction between devoted fans and fanatical fans. While devoted fans have high commitment and attachment towards a team (or brand), fanatical fans are somewhat more extreme in their attitudes and behaviors (Pimentel & Reynolds, 2004). According to Chung et al. (2008), fanatics surround themselves with products or reminders of their interest. In addition, fans also show obsessive and addictive-like behavior (Redden and Steiner, 2000). In some literature, fanatical fans are seen as dysfunctional (Hunt, Bristol & Bashaw, 1999) and should be avoided. However, fanatical fans could still be very valuable for a brand since they are highly committed (Chung et al., 2008).

Brand Love

Consumers can also develop loving feelings towards brands or products (Fournier, 1998; Ahuvia, 2005; Pichler & Hemetsberger, 2007). According to Ahuvia (2005), love objects receive much attention and dedication from consumers. Because individuals put a lot of energy in these objects, it makes them more meaningful and helps integrate them into their sense of identity (Ahuvia, 2005). In addition, Carrol and Ahuvia (2006) found that there is a significant positive relationship between Brand Love and positive WOM. Furthermore, their research also finds that consumers with greater brand love are more brand loyal (Carrol & Ahuvia, 2006). Finally, devoted consumers are considered to exhibit more loving feelings towards brands than normal consumers (Pichler & Hemetsberger, 2007).

Commitment

(22)

University of Groningen, Faculty of Economics and Business, July 2011 Pagina 22 (Pimental and Reynolds, 2004). MaccInnis (2009) also states that Brand Evangelism strongly relates to commitment. The five item scale of Park et al. (2009a) will be used to measure commitment. These items are based on the measurements used by Beatty, Homer and Kahle (1988).

Gift giving

According to MacConel and Hubba (2007), Brand Evangelists like the product so much that they will also buy it as a present for their friends or family. Furthermore, Belk et al. (1989) describes that sacralization of products can take place trough gift giving. Hence, gifts can become sacred products for consumers. Therefore, when a Brand Evangelist offers their favorite product as a gift to his friends or family, they intend to share their love for the product with them.

Dependence

Finally, devoted consumers are also known for their enduring involvement with brands (Pimentel & Reynolds, 2004). This means that these consumers are not able to live without their brand and make sacrifices to obtain the brand. Moreover, devoted consumers are also known for their heavy usage and purchase patterns.

3.2.5 Hypothesis

After defining the measurement scale for Brand Evangelism, the construct will be compared with WOM to see if there is a positive relationship. According to several authors Brand Evangelists should actively participate in positive WOM (Pimental and Reynolds, 2004; Lindstrum, 2005; Matzler et al., 2007; Pichler & Hemetsberger, 2007; MacConel and Hubba, 2007; Chung et al., 2008). However, MaccInnis (2009) states that not all Brand Evangelists have intrinsic interest in brand advocacy (WOM). Furthermore, she also states that Brand evangelism is different from Opinion Leadership and Market Mavenism because consumers who score high on Brand Evangelism are attached to the brand. This thesis is going to research whether consumers who score high on Brand Evangelism significantly engage more in WOM than consumers who score high on Opinion Leadership, Market Mavenism and Innovatorship. From hereon, the following hypotheses are derived.

Hypothesis 1a

(23)

University of Groningen, Faculty of Economics and Business, July 2011 Pagina 23 Hypothesis 1b

Brand Evangelism is significantly different from Opinion Leadership, Market Mavenism and Innovatorship.

Hypothesis 1c

Consumers who score high on Brand Evangelism engage more in Word of mouth than consumers who score high on Opinion Leadership, Market Mavenism and Innovatorship.

3.3 Brand attachment.

The first moderator that will be discussed on the engagement in WOM is Brand Attachment. First, the construct will be explained, then the measurement will be discussed, and finally the influence of the construct will be portrayed.

3.3.1 What is Brand attachment?

According to Park, Macinnis & Prieter (2006) the definition of Brand Attachment is “the strength of the cognitive and emotional bond connecting the brand with the self”. In their research, attachment examines the extent to which an individual’s relationship to an attachment object is described as strong or weak, instead of focusing on the primary relationship experiences that create secure, ambivalent or avoidant attachment styles (Park et al., 2006). According to the definition of Park et al. (2006), Brand Attachment consists of two parts. Firstly, it describes the connectedness between the brand and the self. Secondly, it states that there is a cognitive and emotional bond, which measures the strength of the readiness to allocate one’s processing resources towards the brand. Both parts are described below.

(24)

University of Groningen, Faculty of Economics and Business, July 2011 Pagina 24 from real or imagined personal experiences that create autobiographical memories, personalized meanings and trust. The more the brand is able to create these connections, the more it is regarded as an extension of the self (Belk, 1988; Klein and Bakker, 2004). Park et al. (2006) states that the more the attachment is viewed as an extension of the self, the more distress and sadness is experienced from the prospect of losing the brand. Wherein separation, distress, and depression are strong concomitants of humans’ attachment to people. According to Park, Macinnis & Prieter (2008) a brand becomes connected to the self when it is included in the self memory. This takes place while during the formation of a relationship with the brand the other person makes his or her resources readily available to the self. This in turn leads to cognitive organization that makes that the other’s resources seem as if they were one’s own, and thus coming to take on the other’s perspectives and identities as one’s own (Aron, Mashek, Mclaughlin-Volpe, Wright, Lewandowski & Aron (2005).

The bonds that connect the brand to the self are both cognitive and emotional. The cognitive bond is derived from the personalized experiences and autobiographical memories of the brand, which evoke rich cognitive schemata (Berman and Sperling, 1994) with links connecting the brand with personalized elements of the self. These links are relevant and have strong self-implications, which imply that the connections are emotional (Mikulincer and Shaver, 2005). Because of this connection, the emotional property implicates a `hot affect’ (Ball and Tasaki, 1992; Thomson et al., 2005). According to Park et al. (2006) this hot effect induces desire for the brand, satisfaction with its acquisition, frustration at its lack of availability, fear over its potential loss, sadness over its actual loss, and hope for its future acquisition.

(25)

University of Groningen, Faculty of Economics and Business, July 2011 Pagina 25 3.3.2 Measuring Brand Attachment

In their research, Park, Priester, MacInnis and Wan, (2009) (see also, Park, MacInnis, Priester, Eisingerich and Iacobucci, 2010), develop a parsimonious, valid and reliable measurement scale of brand attachment. In order to measure brand attachment they give the following definition: “Brand attachment is the strength of the bond connecting the brand with the self”. The strength of the bond is measured by two critical indicators: (1) Brand-self connectedness and (2) Brand Prominence (the prominence of brand relevant thoughts and feelings).

Brand self-connection

As described above, brand self-connection occurs when consumers feel strongly attached to a brand and feel as if the brand is a part of them. Park et al. (2009) view brand self-connection as the strength of connectedness a consumer has to a given brand. The more the brand has been incorporated into one’s sense of self and the greater the personal connection the consumer feels between the self and the brand, the stronger is their brand attachment. Furthermore, the idea that attachment involves a bond (with the brand included as part of the self) suggests that a critical aspect of attachment involves the cognitive and emotional connection between the individual and the self ( Park et al., 2010). Though cognitive in representation, this brand linkage is inherently emotional (Mikulincer and Shaver, 2007; Thomson et al., 2005). To conclude, Park et al. (2010) states that, “consumers can be connected to a brand because it represents who one is (e.g. an identity basis) or because it is meaningful in light of goals, personal concerns, or life projects (an instrumentality basis, Mittal (2006))”.

Brand prominence

(26)

University of Groningen, Faculty of Economics and Business, July 2011 Pagina 26 connects the brand to the self. This salience is reflected by the perceived (1) ease and (2) frequency with which brand related thoughts and feelings are brought to mind. Therefore, consumers’ attachment in relation to two brands with the same degree of brand-self connection is greater for the brand that is perceived to be more prominent (Park et al., 2010).

According to Park et al. (2010), brand-self connection is a core component of attachment, since it centrally reflects the definition of attachment as the bond connecting the individual with the brand. However, they suggest that the inclusion of brand prominence adds precision in measuring the strength of the bond connecting the brand with the individual. In their research, they give two reasons for this. Firstly, when thoughts and feelings about the brand are highly accessible, prominence may exert a disproportionately strong influence on decision making (Alba and Marmorstein 1987) and ultimately consumer purchasing behavior (Akcura, gonul and petrova, 2004). Secondly, Park et al. (2010) argue that prominence may serve as an important indicator of attachment when consumers are connected to a brand given its instrumental value (i.e., one’s iMac is important to fulfilling entertainment- and work-related goals). This means, that when a brand has instrumental value, attachment should be stronger when brand-related thoughts and feelings are more prominent than when they are less prominent. To conclude, Park et al. (2010) states that, as prominence increases, brand-related thoughts and feelings are part of everyday life tasks, making brand attachment stronger.

3.3.3 Influence of Brand attachment

(27)

University of Groningen, Faculty of Economics and Business, July 2011 Pagina 27 Furthermore, Godes and Mayzlin (2007) state that loyal customers have a higher persuasiveness when it comes to WOM. Hence, respondents who score high on brand attachment will also score high on WOM praise. However, it is unclear if there will be a significant difference between the WOM activity and the WOM praise from highly attached consumers. According to these findings the following hypothesis is stated.

Hypothesis 2

Brand Attachment strengthens the positive relationship between Brand Evangelism and Word of mouth.

3.4 Consumer uniqueness

Another moderator that affects the engagement in WOM is the need for Uniqueness. Below, this construct is discussed and the relationship between this construct and WOM will be given.

(28)

University of Groningen, Faculty of Economics and Business, July 2011 Pagina 28 Hypothesis 3

Consumer uniqueness decreases the positive relationship between Brand Evangelism and Word of mouth.

3.5 Network breadth

Finally, network breadth is also a moderator within this thesis that effects the engagement in WOM. This paragraph addresses the construct “Network Breadth” and its relationship with WOM.

Network breadth refers to the amount of people a consumer has in its network. When the network breadth is big, consumers are able to talk with more people about the product, hence create more WOM (Godes et al., 2005; Godes and Mayzlin, 2007). Recently, some popular literature has focused on this concept, wherein Rosen (2000) distinguishes “expert hubs” and “network hubs” and Gladwell (2000) talks about “connectors” and “mavens”. These authors make a distinction between Opinion Leaders, Market Mavens and consumers with a large network (network hubs and connectors). Godes and Mayzlin (2007) propose that brands should also focus on consumers with a large network to enhance the diffusion process of their product (Rogers, 1962). This theory is in line with Granovetter’s (1973) theory on “strong ties” and “weak” ties. Granovetter, (1973) states that strong ties mostly belong to the same network while weak ties form the bridge between these networks. According to his theory, the distribution of information goes faster through weak ties, since people with strong ties tend to be in the same social network and thereby already possess knowledge about similar information. This research proposes that network breadth will have a positive influence on the amount of WOM. This means that Brand Evangelists who have a large network breadth will also engage more in WOM. In order to measure the degree of network breadth, the scale of Godes and Mayzlin (2007) is used. The following hypothesis is proposed.

Hypothesis 4

(29)

4. Research Methods

After defining several hypotheses in the literature review, the next step is to empirically test these hypotheses and draw conclusions. In order to test the hypotheses, empirical data is required. This chapter will set apart how this data will be collected. First, the research design will be addressed, after which the collection of samples will be described. Then the constructs are operationalised, followed by a discussion of the measurement scales and control variables. Finally, the data analyses will be set apart.

4.1 Research Design

According to Malhotra (2007), marketing research can be divided into several designs, for instance an exploratory research Designs and a Conclusive research designs. This thesis conducts a conclusive research design on Brand Evangelism. Furthermore, it will gather data from a given sample (see the following section) only once; therefore this is a `single cross-sectional research design' (Malhotra, 2007).

4.2 Sample and data collection

(30)

University of Groningen, Faculty of Economics and Business, July 2011 Pagina 30 Facebook fan pages to express their (loving) feelings for the brand to other users. There are numerous pages on Facebook where users can subscribe by `liking' the page and keep informed about new features and updates. Currently, the official page of the iPhone (Apple’s top selling product) has 2,4 million subscribers, and is still growing by the day. Furthermore, there are dozens of private fan pages of Apple with thousands of subscribers. Unfortunately, it proved difficult to find respondents through Facebook fan pages. Therefore, the questionnaire was eventually spread amongst students of the University of Groningen.

According to Malhotra (2007), the advantages of an internet-based questionnaire, from the perspective of respondents, is that they are able to choose when they want to participate, and these questionnaires are perceived as accessible and easy to use. Moreover, the data is collected relatively fast and is easy to process. In the next paragraph, the operationalisation of the constructs will be discussed.

4.3 Operationalisation of constructs

In order to measure the construct used in the conceptual model, these constructs have to be operationalised. After the operationalistation, questions are formed to measure the construct. As described in the literature review, most of the constructs already possess well-defined questions based on academic literature. These questions are pretested and academically sound, therefore the internal validity of these questions will be high (Malhotra, 2007). However, currently there is no known measurement scale for Brand Evangelism. Therefore the questions to measure this construct will be developed within this thesis. Next, the questions that will be used per construct will be given, after which the scale development of Brand Evangelism will be described.

4.3.1 WOM

(31)

University of Groningen, Faculty of Economics and Business, July 2011 Pagina 31 4.3.2 Antecedents of WOM

The antecedents of word of mouth are measured by scales developed by the following authors: Opinion Leaders by Flynn et al. (1996), Market Mavens by Feick and Price, (1987) and Innovatorsby Goldsmith and Hofacker, (1991). To fit the scales to this topic some of the questions from the Opinion Leader- and Innovator-scale are adjusted for electronic products. All of the questions are measured on a 7-point Likert-scale ranging from (1) totally disagree to (7) totally agree. The questions are found in the appendix I.

4.3.3 Brand Attachment

In order to measure Brand Attachment, the measurement scale of Park et al. (2009; 2010) will be used. All the items are measured on an 11-point scale anchored by 0 (= “not at all’) and 10 (= “completely”) (see appendix I).

4.3.4 Uniqueness

Uniqueness will be measured with the scale of Tian et al. (2001). In their research, they divide uniqueness in three components, namely: creative choice, unpopular choice, and avoidance of similarity. Unfortunately, this scale is very wide (31 items), therefore this thesis will use the top-three items per component. All the questions are measured on a 7-point Likert scale ranging form (1) totally disagree to (7) totally agree, which are found the appendix I.

4.3.5 Network breadth

Finally, to measure network breadth, the measurement scale from Godes and Mayzlin (2007) is used. For this scale two items are open questions and two items are measured on a 7-point Likert scale (see appendix I).

4.4 Measurement scale of Brand Evangelism

(32)

University of Groningen, Faculty of Economics and Business, July 2011 Pagina 32 disagree” to “totally agree”. In these questions respondents have to answer towards which degree they agree or disagree with the statements. Next, the constructs are operationalised and the questions that measure the constructs are given. Finally, the constructs of Brand Love and Commitment are measured by the measurement scales of Ahuvia (2005) and Park et al. (2009; 2010) (see Appendix 1).

4.4.1 Collecting

To measure the construct of Collecting, the following operational definition is defined: “The desire to own and cherish products from Apple”. Consequently, the following questions are defined.

1. I collect merchandise of Apple.

2. When Apple brings out a new product, I have to have it. 3. I cherish the products of Apple.

4.4.2 Rival Brand

For the construct Rival brand the next operational definition is defined: “The Brand that consumers try to avoid and which is perceived as inferior to Apple”. The next questions measure the construct Rival Brand.

1. I try not to use products of Microsoft. 2. I do not like Microsoft.

3. I believe products of Microsoft are inferior to products of Apple. 4.4.3 Rituals

The construct Rituals is operationalised as follows: “All the rituals that are used for the products of Apple”. This construct is measured by the following questions.

1. I am very careful with the products that I just bought from Apple.

2. When I have bought a product from Apple it goes through some rituals before I start using it.

3. I have rituals for using products of Apple. 4.4.4 Fandom

Fandom can be operationalised by the next sentence: “The intensity of the fandom towards Apple”. The following questions measure the construct fandom.

(33)

University of Groningen, Faculty of Economics and Business, July 2011 Pagina 33 2. I always carry a product of Apple with me, wherever I go.

3. I am addicted to products of Apple. 4. Apple is like an obsession to me. 4.4.5. Gift giving

Gift giving can be operationalised by the following sentence: “The extent to which the products of Apple is perceived as a good gift”. In addition, gift giving is measured by the following questions.

1. Products from Apple are excellent gifts.

2. I would like to receive a product from Apple as a gift. 3. I gave several products of Apple to people close to me. 4.4.6 Dependence

Finally, the construct dependence is operationalised as following: “The extent to which the consumer is dependent on the products of Apple”. The following questions measure the construct dependence.

1. I cannot live without Apple. 2. I use products of Apple every day.

3. I will probably use products of Apple for the rest of my life.

4.5 Control variables

(34)

4.6 Data analyses

First of all, some univariate tests will be performed to describe the population and the main results on the constructs. In order to measure the proposed hypothesis the amount of questions per construct should be reduced. This will be done by measuring the Cronbach’s alpha of the questions that measure a certain construct. When the results of the Cronbach’s alpha are 0.6 or higher, the questions will be combined. Furthermore, to determine if the questions measure the right construct a factor analysis will be performed. Next, a cluster analysis will be executed to see whether different groups can be formed depending on the independent variables and are therefore, forming separate groups who either score high on Brand Evangelism, Opinion Leadership, Market Mavenism or innovativeness. Although the latter three constructs are empirically distinct from each other (Feick and Price, 1987), there could be some overlap. When it is possible to form distinct clusters between the dependent variables, this could prove that BE is a new antecedent of WOM.

After colliding the questions that measure the constructs, statistical test will be used to prove the hypothesis provided earlier. With the exception of the main questions about the demographics of the respondents, most of the questions are measured on a 7 or 10-point Likert-scale. According to Malthotra (2007) this data can be considered as semi-interval. Because the hypotheses all have a positive or negative direction between two variables, this thesis will make use of a bivariate asymmetric test. The following tests will be used to prove the hypothesis and test control variables: t-test, regression analyses, correlation analysis and ANOVA.

In order to measure the moderating variables Brand Attachment (BA), uniqueness (U) and Network Breadth (NB) in relation to Brand Evangelism (BE) with WOM, a multiple regression analysis (Malhotra, 2007) is used. According to Hayes and Matthes (2009), quantitative variables are measured in a linear regression model by multiplying the dependent variable with the moderating variable. Therefore, in addition to the normal variables the model also contains the multiplication of BE with the moderating variables. Hence, the following models will be used to test the moderating variables:

(35)

4.7 Reliability analysis

Before the research results will be explained, first it is wise to reduce the amount of questions by forming constructs according to the conceptual model. For this, Chronbach's Alpha analyses are performed to determine if the questions can be added up in order to form the new construct. According to Malhotra (2007) the questions can be combined when the Chronbach's Alpha score is 0,6 or higher. In table 1 the Chronbach's Alpha scores of the constructs are presented.

Construct Amount of questions Chronbach's Alpha score

Brand Evangelist 28 ,953 Opinion Leader 6 ,825 Market Maven 6 ,901 Innovator 6 ,763 Word of Mouth 6 ,862 Brand Attachment 4 ,920 Uniqueness 9 ,828 Network Breadth 2 ,666

Table 1, Chronbach's Alpha scores.

(36)

5. Research Results

In this chapter the sample results are shown and the hypotheses are accepted or rejected. First, the sample representativeness and demographics are given. Then, the reliability analyses will be described. After this some descriptive statistics will be shown. Finally, the hypotheses will be answered.

5.1 Sample Descriptives

This section will describe the demographics of the sample. Table 2 shows that the average student is 22 years old, with a standard deviation of 2,18. Furthermore, the minimum age is 18 years and the maximum age is 34. Table 3 shows that 64,9% of the respondents are male. Moreover, table 3 also states that the majority of the respondents are Dutch 80,9%. Finally, 96,9% of the respondents are students.

Mean Mode Std.

Deviation

Minimum Maximum N Missing

What is your Age? 22,43 22,00 2,18 18,00 34,00 223 2

Table 2, Age statistics.

Frequency Percentage

What is your Gender? Male 146 64.9%

Female 78 34.7%

What is your Nationality? Bulgarian 4 1,8%

Chinese 12 5,3%

Dutch 182 80,9%

German 6 2,7%

Mexican 3 1,3%

Other 18 8%

What is your Occupation? Student 218 96,9%

Employed 3 1,3%

Unemployed 3 1,3%

Other 1 0,4%

(37)

5.2 Descriptive statistics

Before conducting the analyses that can answer the hypothesis, some descriptive analyses will be discussed. Firstly, the means and distribution of all the constructs and questions will be given. Then a number demographics will be compared with the constructs from the conceptual model.

Construct N Missing Mean Mode Std. Deviation Minimum Maximum

Brand Evangelist 205 20 3,60 4,00 1,25 1,07 6,54

Opinion Leader 224 1 4,40 4,50 1,06 1,00 6,83

Market Maven 219 6 4,42 5,00 1,24 1,00 7,00

Innovator 224 1 4,44 4,50 1,04 1,33 7,00

Word of Mouth 224 1 3,21 3,83a 1,29 1,00 6,50

Brand Attachment 219 6 4,44 1,00 2,11 1,00 8,50

Uniqueness 221 4 3,50 3,89 1,01 1,00 6,56

Network Breadth 225 0 4,43 5,00 1,16 1,00 7,00

Table 4, Constructs mean scores.

In table 4, the means of all the constructs are shown. Herein it shows that most of the averages lie between three and four. Four is the center value of the Likert scale, except for the brand attachment construct, where five is the center value. In general, the scores are all quite close to the center value. Table 5 shows the rest of the questions from the questionnaire.

Question N Missing Mean Mode Std. Deviation Minimum Maximum

Amount of Apple products owned

225 0 1,938 1,00 2,110 ,00 27,00

Number of years using Apple products

221 4 3,889 3,00 2,187 ,00 15,00

Hours spent using Apple products

222 3 17,763 10,00 20,368 ,00 150,00

Amount of friends 224 1 8,321 5,00 5,482 ,00 40,00

Amount of acquaintances 220 5 47,305 20,00 72,989 ,00 600,00

(38)

5.3 Hypothesis

Within this paragraph, the hypotheses will be tested. To do this, several statistical tests will be used to see if the hypotheses are accepted or rejected. This thesis will use regression analyses, ANOVA, cluster analyses, factor analyses and multi-linear regression analyses in order to prove the hypotheses that are stated in the conceptual model above. Before a statistical test can be proven, a confidence level is needed. The confidence level states that a certain test is reliable and will have similar results after using a different sample. According to Malhotra (2007) a significance level of 95% is appropriate, therefore this thesis will use this level to conclude if the test are significant.

5.4.1 Brand Evangelism

Hypothesis 1a states that BE has a positive influence on WOM. Both constructs are semi- interval, since they are measured on a 7-point Likert scale. To test this hypothesis, a bivariate regression analysis is used to test the linear relationship. Before the results are given, Hypothesis 1a is formulated below.

Hypothesis 1a

Brand Evangelism has a positive influence on Word of Mouth.

Table 6 shows that a significantly positive relationship exists between Brand Evangelism and WOM, because B= 0,875 and P=0,000/2 < α 0,05. Furthermore, the test shows a high correlation exists between the variables (R=0,842). Finally, BE explains over two thirds of the variation in WOM.

R R Square B (constant) B Sig.

Brand Evangelism 0,842 0,709 0.082 0,875 0,000

Opinion Leadership 0,317 0,100 1,518 0,386 0,000

Market Mavenism 0,338 0,114 1,677 0,347 0,000

Innovatorship 0,250 0,062 1,852 0,308 0,000

(39)

University of Groningen, Faculty of Economics and Business, July 2011 Pagina 39 The regression analysis shows that a strong positive relationship exists between BE and WOM. In general, this model states that when Brand Evangelism increases, the amount of WOM also increases by 0,875, which is remarkable. Therefore, from these results it can be concluded that BE has a positive influence on WOM.

Table 6 also shows the results from the regression analyses between the other antecedents. From this it becomes clear that all variables have a significantly positive relationship with WOM. However, it is clear that the correlation of BE with WOM is higher than the other variables (0,842 vs 0,317, 0,338, 0,250). Furthermore, the explanation of the variation is also a lot higher for BE. In addition, BE also has a higher positive relationship (B=0,875 vs 0,386; 0,347; 0,308). Hence, it can be concluded that Brand Evangelism has a higher positive relationship with WOM than the Opinion Leader, Market Maven and Innovator.

Hypothesis 1b

Brand Evangelism is significantly different from Opinion Leadership, Market Mavenism and Innovatorship.

To find out if the Brand Evangelism construct is significantly different from the Antecedents of WOM a cluster analyses was performed. However, no significant groups were found after performing a cluster analysis on the questions that form the constructs of Brand Evangelism and the antecedents of WOM. After performing the cluster analysis, the difference between the first two groups was too large. Therefore, a factor analysis was performed to test the differences between the constructs.

(40)

University of Groningen, Faculty of Economics and Business, July 2011 Pagina 40 using a principal components analysis and a Varimax rotation method. Malhotra (2007) states that when a factor analysis is performed, first it is important to see if the analysis is appropriate. According to Malhotra (2007), the test is appropriate when the KMO measure of sampling adequacy is higher than 0,5 and when the Bartlett’s test of Sphericity is lower than α = 0,05. In table 7, it is shown that this is the case: KMO=0,875 and BTS=0,000 <0,05. Therefore, a factor analysis is appropriate.

KMO Measure of Sampling Adequacy 0,875

Bartlett’s Test of Sphericity App. Chi-Square Df Sig. 1169,305 55 ,000

Table 7, KMO and Bartlett's test of sphericity.

After establishing the appropriateness of the factor analysis, the amount of factors has to be established. There are three ways to determine the amount of factors (Malhotra, 2007). Firstly, the amount of factors is determined by an Eigen value of one. Secondly, the factor variance should be higher than five. Finally, the cumulative percentage of the variables should explain at least 60% of the total. Table 8 shows the results of the factor analysis. The table shows that the Eigen value is higher than one when using two factors. At five factors the percentage of the variance is higher than five. Thirdly, the cumulative percentage is higher than 60 when using two factors. Since the variance must be at least higher than five, two factors are chosen for the factor analysis, table 9 shows the rotated results.

(41)

University of Groningen, Faculty of Economics and Business, July 2011 Pagina 41 Hypothesis 1c

Consumers who score high on Brand Evangelism engage more in Word of mouth than consumers who score high on Opinion Leadership, Market Mavenism and Innovatorship.

After comparing the results of the regression analyses (table 6) this thesis is going to find out if these constructs score significantly different on WOM. In order to compare the different scores of WOM between the constructs of Brand Evangelism, Opinion leadership, Market Mavenism and Innovatorship these constructs have to be converted into one nominal variable. Therefore, the constructs have to be split in in high, low scores to see which responded belongs to which construct. Meaning that respondents who for example score high on the Brand evangelism scale are considered to belong to this construct. In order to test this two points are used. Firstly, a cutting point of four will be used since this is the middle point of a 7-point Likert scale. Then, a cutting point of five will be taken to ensure results that are more conservative. For example, all respondents who score higher than five on the Brand Evangelism scale are considered to be Brand Evangelists. In table 10 the results for the splitting point of four are shown. Here we see that there are 86 Brand Evangelists, 162 Opinion Leaders, 158 Market Mavens and 158 Innovators.

Components

Initial Eigen Values

Total Variance % Cumulative %

1 2 3 4 5 6 7 8 9 10 11 5,235 1,708 ,809 ,687 ,625 ,538 ,430 ,298 ,268 ,241 ,162 47,588 15,527 7,356 6,248 5,685 4,891 3,905 2,709 2,435 2,188 1,468 47,588 63,115 70,471 76,719 82,404 87,295 91,200 93,909 96,343 98,532 100,000 Factors 1 2 Collect Rival brand Rituals Fandom Brand Love Gift giving Dependence Commitment Opinion Leadership Market Mavenism Innovatorship ,840 ,620 ,662 ,884 ,800 ,553 ,795 ,823 ,090 ,197 ,128 ,172 ,022 ,056 ,196 ,116 ,341 ,172 ,249 ,879 ,849 ,726

(42)

University of Groningen, Faculty of Economics and Business, July 2011 Pagina 42 Furthermore, figure 2 shows that the Brand Evangelists score significantly higher on WOM than the other antecedents of WOM (4,31, P=0,00 < α=0,05). However, there is a big overlap between these groups of respondents. This means that there are a lot of respondents who score higher than four on several scales, or sometimes on even all the scales. Therefore a cutting point of five will result in less respondents per category and a more reliable result. Table 11 shows the amount of people who score higher than five on each of the constructs. Here we see that there are 33 respondents who score high on Brand Evangelist, 78 on Opinion Leader, 79 on Market Maven, and 67 respondents score high on the Innovator scale.

(43)

University of Groningen, Faculty of Economics and Business, July 2011 Pagina 43 Frequency Percentage Brand Evangelism >4 <4 86 169 38.2% 61,2% Opinion Leadership >4 <4 162 93 72% 28% Market Mavenism >4 <4 158 97 70,2% 29,8% Innovator-ship >4 <4 158 97 70,2% 29,8%

Table 8, Antecedents of WOM > 4 frequency table.