Sponsorship in the Netherlands
– The antecedents and consequences of
Sponsorship Perception –
Master Thesis, MScBA, specialization Marketing Research University of Groningen, Faculty of Marketing
July 1st, 2011
Malin Iren Haug Student number: 2049929
Supervisors:
Sponsorship in the Netherlands
!!"#$!%&'$($)$&'*!%&)!(+&*$,-$&($*!+."
Sponsorship Perception –
Groningen July 1st 2011 Master thesis
Master of Business Administration Marketing Research
Department of Marketing, Faculty of Economics and Business University of Groningen,
The Netherlands
MSc in Strategic Marketing Management Department of Marketing,
Norwegian School of Management, Norway
Malin Iren Haug Sagveien 1 1816 SKIPTVET Norway Telephone: +47 414 12 806 E-mail: maliniren@gmail.com Student number: 2049929 Supervision:
University of Groningen, Faculty of Economics and Business, Department of Marketing First supervisor: Prof. dr. P.C (Peter) Verhoef, Professor of Marketing
Preface
Dear reader,
This thesis marks the end of a Master of Science in Strategic Marketing Management at BI Norwegian School of Management, as well as a Master of Science in Business Administration: Marketing Research at The University of Groningen. These five years have been spent both at home and abroad and I have obtained a great amount of knowledge within business management and marketing research, which are reflected in this work about sponsorship perception.
I would like to thank everyone who has taken part in this study, in particular prof. dr. Peter C. Verhoef, my dedicated supervisor, for his effort, time and helpful contributions to this thesis, and prof. L. Voerman for comments on my work.
I would also like to thank family and friends for being supportive throughout the period of working with this thesis. In addition, I would like to give my thanks to Monique Bakker and Nastassja Tijssen for helping me translate and re-translate the survey from English to Dutch, as well as giving comments on my work. I am very grateful for their contributions.
Hope you enjoy the reading!
Groningen, The Netherlands, July 2011
Executive summary
In this thesis we have tested the antecedents and consequences of sponsorship perception, and their relation to loyalty intention. There is a considerable amount of research on sponsorship, but as far as we know there is no research that considers the antecedents level of sponsorship investment and number of sports, as well as the consequences of brand equity, corporate social responsibility, customer involvement and customer satisfaction. In addition to this, we also included a mediation function, that brand equity, CSR and customer satisfaction mediates the relationship between sponsorship perception and loyalty intention. As well as the mediator function, we analysed if customer involvement moderates the relationship between sponsorship perception and loyalty intention. The data consisted of 850 consumers from 9 different companies, and we applied several research methods on the data, such as t-test, correlation, multiple regression, as well as latent class analysis, to test the proposed hypotheses.
The results show that level of sponsorship investment is positively related to sponsorship perception, which implies that if a company has a high level of investment they will achieve higher sponsorship perception. The companies in our study were sponsoring one or several sports, and when testing this variable, the analysis did not find support for a positive relationship between number of sports and sponsorship perception.
We found corporate social responsibility (CSR) to be an important variable in our study, and to our knowledge, the mediation function of CSR is not found in earlier research. This variable mediates the relationship between sponsorship perception and loyalty intention, as well as being positively related to both variables. We recommend companies to integrate and increase their CSR strategy, given its influence on sponsorship perception and loyalty intention.
We applied a latent class analysis to distinguish segments based on consumers’ sponsorship perception. Corporate social responsibility and brand equity were found significant, and were included as covariates. We found four clear distinguished segments where the consumers differ in sponsorship perception.
Suggestions for future research were to expand this survey to more companies, and thereby further test the hypotheses. The overall findings contribute to a new aspect of sponsorship, which is not found in previous research. It will therefore be interesting to follow the research and practice on this area in the coming years.
Table of Content
1 INTRODUCTION ... 1" 1.2RESEARCH QUESTION... 2" 2 LITERATURE REVIEW ... 3"
2.1 Definition of Sponsorship Perception ... 4" 2.2 The History of Sponsorship... 4" 2.3Variables of interest ... 6" 2.4ANTECEDENTS OF SPONSORSHIP PERCEPTION... 6"
Level of sponsorship investment ... 6" Number of sports... 7" 2.5CONSEQUENCES OF SPONSORSHIP PERCEPTION... 8"
Brand Equity ... 9" Corporate Social Responsibility ... 10" Customer involvement... 11" Customer Satisfaction ... 12" 2.6SEGMENTATION... 13"
Segmentation based on Sponsorship Perception ... 13" 2.7CONCEPTUAL FRAMEWORK... 13" 3 COLLECTION OF THE DATA... 16"
3.1 Demographics ... 16"
4. RESEARCH DESIGN... 17" 4.1OPERATIONALIZATION OF VARIABLES... 17" Loyalty intention ... 17" Level of Sponsorship Investment... 17" Number of Sports ... 18" Sponsorship Perception ... 18" Brand Equity ... 18" Corporate Social Responsibility ... 19" Customer involvement... 19" Customer Satisfaction ... 19" 4.2SCALE... 20" 4.3RELIABILITY... 20" 4.3.1 Cronbach’s Alpha ... 21" 4.4METHOD... 21" Multiple regression ... 21" Latent Class analysis ... 22" 4.5MODEL VALIDATION... 23"
4.5.2 Heteroscedasticity... 23" 4.5.3 Normality ... 23"
5. ANALYSIS ... 24"
Introduction... 24" 5.1 Plan for analysis ... 24" 5.2RESULTS OF ANTECEDENTS OF SPONSORSHIP PERCEPTION... 25"
5.3EFFECT OF SPONSORSHIP PERCEPTION’S ON BRAND EQUITY,CORPORATE SOCIAL RESPONSIBILITY AND
CUSTOMER SATISFACTION... 27" 5.4EFFECTS OF SPONSORSHIP PERCEPTION ON LOYALTY INTENTION... 28" 5.5EFFECTS OF BRAND EQUITY AND CSR ON LOYALTY INTENTION... 28"
5.6THE MEDIATION EFFECT OF BRAND EQUITY,CSR AND CUSTOMER SATISFACTION... 29"
5.7ASSESSING THE JOINT IMPACT OF ALL CONSIDERED ANTECEDENTS OF LOYALTY INTENTION32"
5.8THE MODERATOR EFFECT OF CUSTOMER INVOLVEMENT... 33"
5.8.1 Analysis I... 34" 5.8.2 Analysis II ... 35" 5.9FINDING SEGMENTS BASED ON SPONSORSHIP PERCEPTION... 36" 5.9.1 Interpretation ... 37" 5.9.2 Cluster size... 37" 5.9.3 Significance of covariates ... 37" 5.9.4 Results ... 38" Segment 1 - The average segment... 39" Segment 2 - The sponsorship conscious segment... 39" Segment 3 - The sponsorship, brand and CSR conscious segment... 39" Segment 4 - The CSR and brand conscious segment ... 39" 5.10CONCLUSION... 40" 6. DISCUSSION... 41" 6.1RESULTS OF ANTECEDENTS OF SPONSORSHIP PERCEPTION... 41"
6.2EFFECT OF SPONSORSHIP PERCEPTION’S CONSEQUENCES ON BRAND EQUITY AND CORPORATE SOCIAL
RESPONSIBILITY... 42"
6.3EFFECTS OF SPONSORSHIP PERCEPTION ON LOYALTY INTENTION... 42" 6.4EFFECTS ON BRAND EQUITY AND CSR ON LOYALTY INTENTION... 43" 6.5THE MEDIATION EFFECT OF BRAND EQUITY,CSR AND CUSTOMER SATISFACTION... 43" 6.6THE MODERATOR EFFECT OF CUSTOMER INVOLVEMENT... 44"
6.7CONCLUSION... 44" 7. FURTHER RESEARCH AND LIMITATIONS ... 45" 8. REFERENCES ... 47"
Other ... 54"
9. APPENDIX... 55" APPENDIX 1THE COMPANIES USED IN THIS STUDY... 55" APPENDIX 1SAMPLE SIZE... 55" APPENDIX 2DEMOGRAPHICS OF SAMPLE... 56"
APPENDIX 3QUESTIONNAIRE... 57"
APPENDIX 4INTER-ITEM CORRELATION... 58"
APPENDIX 5CRONBACH’S ALPHA... 58"
APPENDIX 6HOTELLING’S T-SQUARED TEST... 58"
APPENDIX 7OLS ASSUMPTIONS – NORMALITY PLOT OF DEPENDENT VARIABLE... 59" APPENDIX 8MULTICOLLINEARITY... 59" APPENDIX 9HETEROSCEDASTICITY:RESIDUALS PLOT FOR OVERALL REGRESSION... 60" APPENDIX 10ANOVA ... 60"
APPENDIX 11RESULT T-TEST... 61"
APPENDIX 12RESULT MULTIPLE REGRESSION... 61"
APPENDIX 13:MODEL SELECTION... 61"
APPENDIX 14:CLUSTER SIZE 5-CLUSTER... 62"
1 Introduction
A sponsorship is a mutual relationship between a sponsor and the sponsored in which both parties gain advantages, often seen as increased brand equity for the sponsor and economic benefits for the sponsored. There are several reasons why a company chooses to sponsor an event. The main sponsor of the World Championship in skiing 2011, Statoil, states that the goal of the sponsorship is to contribute to internal pride in the company, as well as building a positive reputation. BMW, one of the international sponsors for the same event, claims that they want to use the event to extensively promote the driving system of BMW. While Intersport, also one of the international sponsors, states that they want to use the event to build a stronger brand image, as well as internal motivation in the company. As the companies above state, sponsorship is used mainly to build brand image externally and build motivation internally. A company invests a rather large amount of money in sponsorship, and it is therefore of interest for the company to fully explore the advantages of doing so.
Sponsorship is not only important for the companies but also for the event, team or person that is sponsored. As Aksel Lund Svindal, World Champion in alpine skiing, states: “I’m very proud, and grateful, of my cooperation with Longines (watches).” Performers being sponsored gain the advantage of being able to live by only performing their sports, and they build their personal brand value through being “the face” of the sponsor. One of the best-known examples of this are football players: David Beckham, for example, has a great brand value being “the face” of e.g. Gillette, Addidas and Pepsi.
Both the intensity of the sponsorship, i.e. is the company an active sponsor, and the strength of a company’s sponsorship will be measured, i.e. is the company a well-known sponsor.
1.2 Research question
The general research objective in this thesis is to understand how sponsorship affects customer performance. In this thesis we look at customer performance as increased brand loyalty. As mentioned in the previous chapter, a company want to gain knowledge about how their sponsorship investment relate to customer
performance. We will therefore explore the following research question:
1) How do the antecedents ‘level of sponsorship investment’ and ‘number of sports,’ and the consequences ‘brand equity’, ‘customer satisfaction’, ‘customer involvement’ and ‘corporate social responsibility’ affect the relationship between Sponsorship Perception and Loyalty Intention?
Another relationship of interest is the relationship between sponsorship and brand loyalty. How does a company’s sponsorship investment relate to brand loyalty? How do the customers respond to sponsorship, and does this effect brand loyalty? Based on these questions, we will consider through which mechanisms
sponsorship affects brand loyalty. Therefore, the following research questions will be considered:
2) How does sponsorship influence sponsorship perception?
3) What is the effect of sponsorship perception - brand loyalty and through which variables does sponsorship perception have an effect on brand loyalty?
We find this interesting research questions since this can increase knowledge about sponsorship, and thereby increase companies’ outcome of sponsorship investment. We believe that the outcome of our research can learn companies how consumers perceive companies’ sponsorship. This will help companies gain the most advantage from their investment, and it can also help companies determine the amount of money they should invest.
customer will most likely purchase again from the company, and thereby generate profit for the company. In addition to the dependent variable, we have taken several other variables into account: (1) level of sponsorship investment, which can influence sponsorship perception; (2) number of sports, where we distinguish between companies sponsoring one sport or several sports; (3) brand equity, where several research have found a relationship with sponsorship; (4) customer satisfaction, which has a known relationship with loyalty intention; (5) corporate social responsibility, which has increasing importance, but still has not been related to sponsorship, and last (6) customer involvement that has a known relationship with sponsorship perception and loyalty. In addition to this, to help companies in their segmentation, we will research if it is possible to define segments based on consumers’ sponsorship perception. With the inclusion of these variables we will get a more nuanced and clearer picture of sponsorship perception and thus help to bring in more precise, and useful knowledge.
2 Literature Review
A sponsorship agreement with an important event or performer is expensive for a company, and the company therefore expects positive outcomes of the investment. Even if the company is expecting a positive outcome, few companies know exactly what to expect of a sponsorship other than “enhanced image” (Sherry, 1998). Having knowledge of the consequences of sponsorship will help managers increase the outcome of their investments. “It is important for managers to be able to determine what impact a particular sponsorship will have on (…) brand preference, brand loyalty, and ultimately, profitability” (Aaker, 1991; Keller, 1998). This thesis will look into variables effecting sponsorship perception, and in that way hopefully increase the sponsor’s knowledge, and in doing so increase the outcome of a sponsorship agreement.
2.1 Definition of Sponsorship Perception
There are several definitions of sponsorship, of which nearly all include the variables image transfer, investment in objectives to enhance value, and access to the event (Gardner and Shuman, 1988; Cornwell et al., 2005; Ukman, 1995.) In this thesis, the definition of Cornwell et al. (2005) is adopted:
“Sponsor (i.e. a brand or firm) providing cash and/or other compensation in exchange for access to an object’s commercial potential (i.e. exposure and association with the cause, event, organization or individual related to a sport, cultural, and/or non-profit entity).”
In this thesis we look at sponsorship perception, where “perception” is defined as “a way of regarding, understanding or interpret something“ (Oxford Dictionary, 2008). In this thesis, we therefore define Sponsorship Perception as “the understanding and interpretation of a sponsorship”.
To further clarify the term, some information about the measurement will follow: Sponsorship perception is measured based on how the consumers perceive the company as a sponsor. The questionnaire asks questions regarding consumers’ knowledge about companies’ sponsorship investment. The outcome of the questionnaire forms how the consumers perceive sponsorship, and thereby the term sponsorship perception.
2.2 The History of Sponsorship
The history of sponsorship goes back to the mid-1980s and early 1990s. The use of sponsorship as a marketing tool is quite new, and according to Meenagham (1994), an exponential growth has been seen. The exponential growth has lead to an increased and intensive academic interest, resulting in several empirical studies referred to throughout this thesis.
marketing.” The research of Cornwell et al. (2001), Javalgi et al. (1994), McDonald (1991), Quester (1997), Turco (1995) and Witcher et al. (1991) argue along similar lines. This has given an increase in the amount of money used for sponsorship, which increased (worldwide) from $2 billion in 1984 to $28 billion in 2004 (Carrillat et al, 2005; Koo et al, 2006). This amount increases each year; the annual worldwide expenditures for sport sponsorship were $30 billion in 2008 (source: The World Sponsorship Monitor 2008).
According to Meenagham (1994), companies invest in sponsorship to get access to an event, team or person in order for the company to be associated with the specific event, team or person. Gwinner (1997) argues that companies want an image transfer from the sponsor object to the company. A company can choose to sponsor different events, such as culture, art and sports. Even though companies can decide upon a wide range of events, “roughly two-thirds of the sponsorships in the United States are associated with athletic events” (sponsorship.com). Advertising and sponsorship both have the same objectives, namely increasing brand awareness and building brand equity, but these objectives are achieved in different ways. Erdogan and Kitchen (1998) distinguish advertising and sponsorship as follows:
“Advertising is more direct, explicit and can be more easily controlled, while sponsorship can overcome certain communication barriers and has practically unlimited target selection possibilities”
Sponsorship has more credibility than other marketing tools, according to Mizerski et al. (1979) and Grewald et al. (1994), which makes it an effective tool to use to increase brand equity and achieve loyal customers. Research of Crimmins and Horn (1996) found that a majority of NASCAR fans would “almost always” purchase the sponsor’s products compared to a competitor’s products.
Sponsorship is a new area and the research in this area therefore shows a lack of diversity. The majority of the research focuses on how sponsorship is related to brand image. We found, during the literature review, that research outside the relationship between sponsorship and brand equity is minor. We therefore believe that this study will contribute to clarify what influences consumers’ sponsorship perception.
2.3Variables of interest
A considerable amount of research has explained the relationship between sponsorship and purchase intention, and several empirical studies have discovered the relationship between the two variables to be significant. In this study, we will look into variables affecting sponsorship perception. The objective of this study is to determine the antecedents and consequences of sponsorship perception, and thereby providing companies with the possibility to achieve higher outcomes from their sponsorship. We consider level of sponsorship investment, and number of sports as antecedents of sponsorship perception. For consequences of sponsorship perception we focus on brand equity, customer satisfaction, customer involvement and corporate social responsibility. We will in the following sections look into the exciting literature related to our variables, as well as defining hypotheses.
2.4 Antecedents of Sponsorship Perception
This part of our conceptual framework will investigate the role of antecedents of sponsorship perception. There are different constructs that work as antecedents for sponsorship perception, and in this study we are considering level of sponsorship investment and number of sports. In the following part we will elaborate on these constructs, as well as integrate them into our model.
Level of sponsorship investment
help the company capture the attention of consumers, reinforce brand image and increase brand loyalty (Hughes and Shank, 2004). The amount of money the company provides to get access to the sponsored object differs depending on what channels (e.g. media, performers, communication channels) the sponsor get access to, in addition to the market value of the event or performer. To be one of the main sponsors for the World Championship in skiing 2011 a company invests approximately !3 million Euros.
In this study, the level of sponsorship investment is of interest. We will argue that by investing more in the sponsorship, thus increasing level of investment, it will increase brand equity and thereby loyalty intention. Quester and Thompson (2001) found level of investment to have a clear positive relationship with audience awareness of the sponsor, and improvements between pre- and post-event attitudinal measures. Grosh et al. (2004) found that by having more exposure of the brand during an event will increase brand awareness and brand recall, which again relates to brand equity. To increase exposure related to sponsorship, the level of investment is usually high. Based on practice, e.g. from the World Championship in skiing 2011, we assume that this relationship is true. Hensler et al. (2007) support this assumption by arguing that level of investment can substitute low fit and still obtain the same outcomes from the sponsorship.
We argue that level of investment will affect the impact of the variables in the model. We argue that higher investment leads to more exposure of the company’s name and thereby a possibility to increase brand equity, which again relates to loyalty intention. Based on the argumentation above, the following hypothesis is of interest:
H1: Level of Sponsorship investment is positively related to sponsorship
perception
Number of sports
unknown sports as horse ball, to better-known sports such as cricket and basketball. Football (soccer) is the world most popular sport, according to Palacios-Huerta (2004), with high salaries for players, huge sponsor budgets and an enormous amount of people playing the sport. The criteria used to decide what type of sport a company chooses to sponsor can be the sport’s popularity in society, coverage in media and number of performers at International level. Soccer is a sport with high popularity, given that the World Cup 2002 for instance, drew a cumulative audience of 42.5 billion people (Palacios-Huerta 2004), which implies that it is easy for the sport to draw sponsors. Other sports, e.g. skydiving and weight lifting, find it hard to acquire sponsors to cooperate with. Even so, some companies choose to sponsor less popular sports, such as Nissan (car company) that sponsor base-jumping in Norway.
Even though there are different criteria for choosing what kind of sport and number of sports the company would sponsor, we will argue that the desired outcome is the same, namely increased sponsorship perception. Some companies choose to sponsor a wide range of sport, while other companies choose to sponsor a more narrow range, and often only one sport. What we find of interest is if there is any difference in sponsorship outcome related to number of sports a company sponsors. Will a company that is sponsoring several sports have increased sponsorship perception? If the company is spreading their sponsorship investment, will this have a positive influence on sponsorship perception? We find these questions interesting, and we are therefore focusing on the following hypothesis:
H2: the Number of sports a company is sponsoring is positively related to
sponsorship perception.
2.5 Consequences of Sponsorship Perception
Brand Equity
The outcome of sponsorship can be numerous, but most research has looked at the effect sponsorship has on brand equity. Brand equity is defined in different ways of which Keller’s (1993) definition is the most cited, and also the one used in this thesis: brand equity is “the differential effect of brand knowledge on consumer response to the marketing of the brand.”
The recognition of the importance of brand equity started in the 1950s (e.g. Gardner and Levy, 1955), followed by the research of Farquhar (1989), stating that “brand equity is the added value endowed by the brand to the product”, and has developed further through several empirical studies (e.g. Meenaghan, 1996). We will argue, based on Vogel et al. (2008), that brand equity is important, since it will lead to positive attitudes such as satisfaction and loyalty. Holehonnur et al. (2009) found support for the hypothesis that the higher the brand awareness, the higher the consumer’s brand equity, and Vogel et al. (2008), as referred to in Prins and Verhoef (2007), argues: “if customers judge a particular brand as strong, unique, and desirable, they experience high brand equity.”
H3a Sponsorship perception is positively related to brand equity
H3b Brand equity is positively related to loyalty intention
H3c Sponsorship perception is positively related to loyalty intention
H3d Brand equity mediates the relationship between sponsorship perception and
loyalty intention
Corporate Social Responsibility
The focus of corporate social responsibility (CSR) has increased during the last few years, both in research and in media, and in the government and for the consumer, and thereby increasing the necessity for companies to focus on CSR. The effect of CSR has been discussed in research (e.g. Kanji and Chopra, 2010) and the number of companies that have a CSR strategy is increasing. The focus on green business, sustainability of the environment and working conditions has given rise to empirical research in this area. The research has found a relationship between CSR and brand attitude (e.g., Klein and Dawar, 2004), as well as purchase behaviour (e.g., Webb and Mohr, 1998.) The definition of CSR used in this study is from Pride and Ferrell (2006) who define it as:
“Corporate social responsibility as a company’s obligation to exert a positive impact and minimize its negative impact on society.”
To further describe CSR, we also apply the definition of Carroll (1991; 1999): “CSR is the economic, legal, ethical and philanthropic responsibilities of companies.”
We will argue that sponsorship can be a tool to build a positive CSR image and increase the social responsibility of the company. Wagner, Lutz and Weitz (2009) propose that the way a company is communicating to consumers affect the perception of the company. According to Davis and Tsiantas (2008), sponsors of the Olympics believed that their support in the Olympic Games had “resulted in their being perceived as good corporate citizens demonstrating social responsibility.“ We will draw this further, and argue that since sponsorship is a way of communicating with customers, sponsorship can be used as a way to affect customers’ perceptions and thereby behaviour. When consumers are exposed to the sponsorship, their image of the company/brand will be altered. This impression will be processed into the overall image of the company, allowing the company to use sponsorship as a tool to build good corporate responsibility. In addition to this, we would argue that CSR could mediate the relationship between sponsorship perception and loyalty intention. We will therefore test the following hypotheses in our study:
H4a Sponsorship perception is positively related to Corporate Social
Responsibility.
H4b Corporate Social Responsibility is positively related to Loyalty Intention.
H4c Corporate Social Responsibility mediates the relationship between
Sponsorship Perception and Loyalty Intention.
Customer involvement
Another variable of interest is customer involvement. In this study we adopt the definition of involvement from Mitchel (1980) who defines involvement as:
“an internal state variable, at the individual level that indicates the amount of arousal, interest or drive evoked by a particular stimulus or situation.“ Involvement can be related to involvement in the sponsored event, or involvement in the brand that is sponsoring the event. In this thesis, involvement in the brand is of interest. Thus, the involvement with the brand is measured and analyses (i.e. how interested the consumer is in the particular brand and brand category).
influence the ability to accurately recognize the sponsor” and thereby have an effect on brand equity. Research has also found that high degree of involvement in the brand make consumers more likely to have positive behaviour toward the brand. Other outcomes of customer involvement are increased loyalty intention, positive Word-of-Mouth and increased interest in the brand. As shown, several empirical studies have examined customer involvement, and we will contribute to earlier research by studying customer involvement in a sponsorship context.
An interesting question relates to the moderator effect of customer involvement, thus we argue that how involved the customers are in a brand could affect the direction and strength of the relationship between sponsorship perception and loyalty intention. According to Baron and Kenny (1986), a “moderator variable specifies when certain effects will hold”, and we would like to research if customer involvement moderates the relationship between sponsorship perception and loyalty intention. Based on this argumentation, we will test the following hypothesis:
H5: Customer Involvement moderates the relationship between sponsorship
perception and loyalty intention.
Customer Satisfaction
Customer satisfaction is a well-known and established concept in several sciences, ranging from marketing (Fornell and Werneldt, 1987; 1988), to economic psychology (Johnson and Fornell, 1991) and economics (Van Raaij, 1981; Wärneryd, 1988). We adopt the definition of Andreassen and Lindstad (1996) and Oliver (1980), and define customer satisfaction:
“as the customer is evaluating the service performance, and the result is compared to expectations prior to purchase or consumption, where any discrepancy leads to disconfirmation; i.e. positive disconfirmation increases or maintains satisfaction and negative disconfirmation creates dissatisfaction.”
satisfaction on loyalty intention is well known. Even though there is much research in this area, we would like to explore this area further. We find it interesting to study if customer satisfaction mediates the relationship between sponsorship perception and loyalty intention. We will therefore test the following hypothesis:
H6 Customer Satisfaction mediates the relationship between Sponsorship
Perception and Loyalty Intention.
2.6 Segmentation
Segmentation based on Sponsorship Perception
To further research consumers’ sponsorship perception, and increase the knowledge about sponsorship perception, we will look into if it is possible to define segments based on the consumers’ sponsorship perception. The analysis will indicate whether it is possible to divide customers into segments based on their sponsorship perception. We would argue that sponsorship perception, being the main variable of interest, should be used to define segments in the data set. The main question of interest is if there are segments that can be defined, and differ, based on consumers’ sponsorship perception. The following hypothesis is therefore of interest:
H7: Segments can be found based on sponsorship perception.
2.7 Conceptual Framework
i. Antecedents consist of three constructs: level of sponsorship investment, number of sports, and sponsorship perception.
ii. Consequences consist of five constructs: corporate social responsibility, brand equity, customer satisfaction, category involvement and loyalty intention.
The conceptual idea is that the antecedents will influence the consequences, and finally loyalty intention. (1) Sponsorship perception is the main independent variable and (2) antecedents of sponsorship and (3) consequences of sponsorship mediates the relationship between the main independent variable, sponsorship perception, and the dependent variable, loyalty intention.
! "#$%&%'%#$(! )*#(%+,%#&%(! -.*#(*/(01.!1#2%($3%#$! "#" " 4,35%/!*6!(.*/$(! #" " 7/8#'!9+,1$:! " #" )*/.*/8$%!-*&18;!<%(.*#(151;1$:! " #" ),($*3%/!-8$1(68&$1*#! " #" ),($*3%/!=#2*;2%3%#$! " #"
3 Collection of the data
Data for this research was collected through phone interviews, where a research company, MetrixLab, contacted consumers by phone. The respondents where asked to answer a questionnaire consisting of 160 questions about different service companies in a broad range of sectors. The majority of the questions consisted of statements about companies ranging from perception about social responsibility to customer service, and the consumers were asked to rate these statements. In addition to this, questions regarding consumers’ usage of services were asked to determine loyalty and purchase intention.
In our study we used a cross sectional design, which involves multiple subjects measured at one point in time. In total 100 companies in 7 different sectors where involved in the study, and 7000 consumers participated. The sample used in our study involves 9 of those companies, and our total sample size is 849 (see appendix 1).
3.1 Demographics
4. Research Design
This chapter will be structured as follows: First, the operationalization of the variables will be discussed and defined. Next the scale used in this study will be elaborated on. Thereafter reliability of the concepts will be discussed. There are several methods used in this thesis, and this will be discussed after that. Lastly, the validation of the data will be examined.
4.1 Operationalization of variables
In this chapter we will elaborate on how the variables in our study are measured, as well as the scale used. The specific questions used can be seen in appendix 3.
Loyalty intention
Loyalty intention is the dependent variable in our research. Loyalty can be measured in multiple ways, but most common is measuring the likelihood that a customer will recommend the service/product to a friend, colleague or family. Homburg et al. (2009) measure customer loyalty based on Homburg and Giering (2001), by asking customers about the likelihood of using the company in the future and recommending the company to others. Battacharya and Sen (2003) use questions regarding purchase behaviour to establish a loyalty dimension. The Norwegian Customer Satisfaction Barometer measures loyalty by the following variables: likelihood of retention, likelihood of speaking favourably about the company to others, and likelihood of recommending the company to others (Johnson et al, 2001). Our research is measuring loyalty intention by the same variables as the previous mentioned research, thus we would argue that it is an appropriate way to measure the dependent variable.
Level of Sponsorship Investment
previous year, and we assume that the companies provided us with the correct value of their investment.
Number of Sports
The second hypothesis in our thesis argues that number of sports is positively related to sponsorship perception. This variable was measured by studying what kind of sport, and how many sports, the companies are sponsoring. The companies are divided into two groups: (1) companies sponsoring one sport, and (2) companies sponsoring several sports/culture The two groups are thereafter used in further analysis throughout this thesis. The groups are divided based on information from the companies regarding what kind and amount of sport or culture they are sponsoring, and we make the assumption that the information provided is correct and reliable.
Sponsorship Perception
To measure the variable sponsorship perception we included two questions in the questionnaire: (1) is the company a known sponsor of sports, and (2) is the company an active sponsor in sport. These two questions are used to determine the sponsorship perception of consumers and are consistent with metrics used in previous research, such as the research from Grosh et al. (2004) that measure sponsorship perception in relation to the World Championship in skiing.
Brand Equity
There are several studies in which authors look into brand equity (e.g. King and Grace, 2008; Park and Srinivasan, 1994). King and Grace (2008) argue that there are two ways to measure brand equity, namely non-financial and financial. Financial measures relate to the future earnings or the market share. Non-financial measures, on the other hand, relate to name awareness, brand loyalty, and brand associations, and this is how we applied the measurement of brand equity in our study. Another way to measure brand equity is conjoint analysis (Keller, 1993; Green and Srinivasan, 1990). We are not using this method in this study, given that the data was collected based on other objectives.
react to and perceive the brand, thus a favourable response to brand attributes will give greater brand equity. Brand equity in our study is based on the above research, and we therefore asked the respondents to rate statements related to their perception of the brand in question. The specific questions used can be found in appendix 3.
Corporate Social Responsibility
Even though the literature on how to measure CSR on a company level is evolving rapidly (Clarkson, 1995), there is still no generally established method (Gjølberg 2009). CSR is a relative new term and the focus on metrics has only been brought to attention in the past few years. In practice, the most common metrics to measure a company’s use of CSR is the Global Reporting Initiative, in addition to three other standards: Social Accountability International’s SA8000, International Organizations of Standardization’s ISO 14000, and the Institute of Social and Ethical Accountability’s AA1000 (Lecture note Gordon, 2009). These measures are measuring companies’ contribution to CSR in an objective way. In our study we are focusing on consumer’s perception, and these metrics are therefore not applicable. In this thesis CSR is measured through statements where the respondents rate how they perceive the company’s contribution to the society, as well as taking public and social responsibilities.
Customer involvement
It is common to measure involvement with a brand by asking the customers how interested they are in the brand. Our research uses the same approach to measure customer involvement. In the questionnaire two questions are used to measure this variable: (1) customer’s interest in services from the company, and (2) the importance of this kind of service for the customer. As these questions are similar to previous research, we argue that the questions are appropriate to measure customer involvement.
Customer Satisfaction
most research measures overall satisfaction by asking the question about how satisfied the customer is with the experience with the company. The Norwegian customer satisfaction barometer uses the metric for overall satisfaction: performance vs. the customer’s ideal service provider in the category. Tse and Wilton (1988) measure customer satisfaction by asking the respondents about product attitude, purchase intentions and perceived performance. Homburg et al. (2009) measure customer satisfaction, based on Bettencourt (1997) and Bitner and Hubbert (1994), by asking the customers how satisfied they are overall with the company. In the questionnaire used on this study, we used the same approach, and asked the respondents the same question, namely how satisfied they are with the company.
4.2 Scale
In the questionnaire, respondents were asked to rate statements. The questionnaire use a 7-point Likert scale, with the endpoints of “Strongly Agree” and “Strongly Disagree” for each of the independent variables corporate social responsibility, customer satisfaction, and brand equity. The independent variables sponsorship perception and customer involvement were measured on a scale from 1 (not known) to 10 (known). The respondents specified their level of agreement to different statements proposed.
4.3 Reliability
4.3.1 Cronbach’s Alpha
We performed a Cronbach’s Alpha test to test for internal consistency reliability, as suggested by Hair et al. (2006). Cronbach’s Alpha is a diagnostic measure to assess the consistency of the entire scale, thus the Cronbach’s Alpha should be higher than 0.70 to ensure that there is consistency throughout the entire scale (George and Mallery, 2003). We performed a Cronbach’s Alpha for the independent variables in our study, and the results are displayed in appendix 5. The table shows that the independent variables are above 0.70 and therefore meeting the criteria. We therefore conclude that the internal consistency in our measures is acceptable.
Another statistical test that was conducted was the Hotelling’s T square test, which should be significant. This test controls for multiplicity by comparing, and applying a vector function, to the two questions measuring an independent variable. The Hotelling’s T square test results in significant values for all of our variables (p<0.05) (see appendix 6).
4.4 Method
Multiple regression
When performing an OLS regression, there are several assumptions that need to be met. To diagnose the validation of the assumptions we made a residuals plot to show the relationship between the residuals and explanatory variables, as according to Leeflang et al. (2000). The residuals plot shows that the fitted relationship is not linear and the variance is not constant across the explanatory variables. It therefore meets the assumptions of OLS regression (see appendix 7). OLS regression has several additional assumptions, which we will elaborate on in the following chapter.
Latent Class analysis
In addition to performing multiple regression, a latent class analysis will be performed. Latent class analysis is a classification technique, which assigns a case (consumer) to the class to which the consumers probably belong by looking at the probability consumers have to belong to a certain segment (Mooijaart and Van der Heijden, 1992). This study aim, thorough a Latent Class Analysis (LCA), to detect relevant subgroups of people, based on their sponsorship perception.
Latent Gold is the most popular statistic method for performing latent class analysis. A preliminary Principal Component Analysis (PCA) can be used before performing a LCA in order to reduce data. This would be especially convenient when there are many measurement items and there is a (strong) reason to believe that there is an underlying structure. For this study, no preliminary PCA was conducted. We do not have (too) many variables and therefore we use our original set of variables for the Latent Class Analysis, which is common practice.
would argue that this disadvantage is no concern for our data, since our data has a considerable sample size.
4.5 Model validation
The overall goal of the analysis “consists essentially of trying to explain variation in the criterion variable by fluctuations in the predictor variables” (Leeflang et al., 2000). To ensure that this goal is met, there are several criteria that a model should meet (Ibid). A wide range of statistical tests is performed to test the validation criteria, which will be elaborated on here.
4.5.1 Multicollinearity
Multicollinearity can make parameter estimates unreliable (Leeflang et al., 2000), thus we conducted an analysis to be sure that multicollinearity is not present. By performing a regression analysis we found VIF values ranging from 1.0 to 1.4 for the independent variables (see appendix 8). Leeflang et al. (2000) argue that VIF values under 10 do not indicate multicollinearity, and we therefore conclude that this is not present in our data.
4.5.2 Heteroscedasticity
Heteroscedasticity is another problem that can give unequal variance of the disturbance term, and result in wrong estimates of variance of effects (Hair et al., 2006). To detect the problem, we searched for homogeneity of variance by plotting the residuals of the independent variables against all levels of the dependent variable. By doing this, we found that heteroscedasticity is not present in the data (see appendix 9).
4.5.3 Normality
supports that normality of the error terms in the dataset is present. Lastly, we conducted an ANOVA analysis to be sure that the dependent variable is normally distributed (Hair et al., 2006). The ANOVA analysis showed significant results, thus we conclude that the data meet the validation assumptions for a regression analysis (see appendix 10).
5. Analysis
Introduction
In this part of the thesis we will do several in-depth analyses to be able to confirm the hypotheses proposed in previous chapters. First, we will analyze the antecedents of sponsorship, respectively hypothesis H1 and H2. Second, we will analyze the consequences of sponsorship, which are hypotheses H3a, and H4a. We will thereafter analyse the relationship between sponsorship perception and loyalty intention (hypothesis H3c), before looking at the mediators of sponsorship perception, respectively hypothesis H3d, H4c and H6. Thereafter, we will analyze the moderator effect of customer involvement on sponsorship perception (hypothesis H5). Lastly, we will analyse if there are segments in the data based on sponsorship perception (hypothesis H7).
5.1 Plan for analysis
The list under indicates the order of which the analysis will be performed. In addition, it gives an overview of which analysis will be conducted, and which equations that will be tested.
1. Correlation: Sponsorship investment – Sponsorship perception 2. T-test: Number of sports - Sponsorship perception
3. Multiple regression Antecedents of Sponsorship Perceptions:
Sponsorship Perceptionj = !0 +!1 Level of Investmentj +!2 Number of
sportsj + "j
4. Multiple regression: Sponsorship perception – Brand Equity 5. Multiple regression: Sponsorship perception – Corporate Social
Responsibility
6. $%&'()&*"+*,+*--(./"011*2'-".1"3)./-.+-4()")*+2*)'(./"./"&.56&'5" (/'*/'(./7
7. Multiple regression: Brand Equity – Loyalty intention
8. Multiple regression: Corporate Social Responsibility – Loyalty Intention 9. Multiple regression: The mediation effect of Brand Equity, CSR and
Customer Satisfaction:
I: Loyalty Intentionj = !0 +!1 CSRj + !4 Sponsorship Perceptionj+ "j
II: Loyalty Intentionj = !0 +!1 Brand Equityj + !2 Sponsorship
Perceptionj+"j
II: Loyalty Intentionj = !0 +!1 Customer Satisfactionj + !4 Sponsorship
Perceptionj+ "j
10. Multiple regression: The joint impact of all considered antecedents of Loyalty Intention:
Loyalty Intentionj = !0 +!1 Brand Equityj +!2 CSRj + !3 Customer
Satisfactionj + !4 Sponsorship Perceptionj+ "j
11. Multiple regression: The Moderator effect of Customer Involvement: Loyalty Intentionj = !0 +!1 Brand Equityj +!2 CSRj + !3 Customer
Satisfactionj +!4 Customer Involvementj * Sponsorship perceptionj + !5
Sponsorship Perceptionj+ "j
12. Latent class analysis
5.2 Results of antecedents of sponsorship perception
H1: Level of Sponsorship investment is positively related to sponsorship
perception
groups of which the correlation coefficient in this analysis belongs to a medium level. Our correlation coefficient indicates a medium strong, positive relationship between the two variables, thus level of investment is positively related to sponsorship perception. We therefore conclude that hypothesis H1 is supported.
H2: the Number of sports a company is sponsoring is positively related to
sponsorship perception
Hypothesis H2 relates to the number of sports a company is sponsoring. In this thesis the variable “number of sports” is divided into two categories: (1) one sport, and (2) several sports. We therefore applied a t-test to test this hypothesis. A t-test is usually performed to “assess the statistical significance of the difference between two sample means for a single dependent variable” (Hair et al., 2006). There are several t-tests that can be performed, but an independent samples t-test is appropriate for our data, given that the data has one categorical independent variable (number of sports) and one continuous dependent variable (sponsorship perception). The independent samples t-test will tell whether there is a statistically significant difference between sponsoring one sport or several sports in terms of increased sponsorship perception.
The t-test results in a non-significant value of 0.074 (p>0.05), which indicates that there is no significant difference between the two groups, thus there is no difference in sponsorship perception related to how many sports a company is sponsoring. We also calculated the effect size as proposed by Cohen (1988). The calculation shows a small effect of 0.003. This indicates that number of sports a company is sponsoring explains 3 % of the variation in sponsorship perception, which we would argue is quite small. In conclusion, there was no significant difference in sponsorship perception when sponsoring one sport (M=4.98, SD= 1.96), or several sports (M=5.23, SD= 217.) The magnitude of the differences in the means (mean difference -0.25) was very small (eta squared= 0.003) (see appendix 11), thus we conclude that hypothesis H2 is not supported.
Sponsorship Perceptionj = !0 +!1 Level of Investmentj +!2 Number of sportsj + "j
The multiple regression analysis shows an adjusted R2 of 17.3 %. Level of investment was significant (0.000), while number of sports were not significant (0.373). The regression shows a F-value of 89.36, and a beta value of respectively 0.910 (level of investment), and 0.116 (number of sports) as seen in the Table 2.
Beta value Sign
Level of investment 0.910 0.000
Nr. of sports 0.116 0.373
Table 2: Parameter estimates for regression model explaining sponsorship perception.
We therefore conclude that level of sponsorship investment is positively related to sponsorship perception, while there is no support for the hypothesis that number of sports is positively related to sponsorship perception.
5.3 Effect of Sponsorship perception’s on Brand Equity, Corporate Social Responsibility and Customer Satisfaction
H3a: Sponsorship perception is positively related to brand equity.
Hypothesis H3a argues that sponsorship perception is positively related to brand equity. We conduct a regression analysis with brand equity as the independent variable and sponsorship perception as the dependent variable. The result shows an adjusted R2 of 6.8%, F-value 61.2, significance 0.000, and beta value 0.499 (see appendix 12). We therefore conclude that hypothesis H3a is supported.
H4a: Sponsorship perception is positively related to Corporate Social
Responsibility.
variables (0.000), an adjusted R2 of 9.0% and a beta value of 0.490 (see appendix 12). We therefore conclude that hypothesis H4a is supported.
The two analyses above indicate that sponsorship perception is positively related to brand equity and corporate social responsibility, and we will elaborate on the impact of this in section 6.
5.4 Effects of Sponsorship perception on loyalty intention
H3c: Sponsorship perception is positively related to loyalty intention.
To test the effect of sponsorship perception on loyalty intention, we performed a multiple regression analysis with the two variables, as shown in the equation below:
Loyalty Intentionj = !0 +!1 Sponsorship Perceptionj+ "j
The regression analysis is performed to see if the independent variable, sponsorship perception, is positively related to the dependent variable, loyalty intention. The multiple regression analysis shows a significant relationship (0.000) between sponsorship perception and loyalty intention. The explanatory power, adjusted R2, is 5.0%. This is a low R2 square, but according to Leeflang et al. (2000) “low R2 values should not be interpreted as indication of unacceptable or useless results”. The beta value of sponsorship perception is 3.44, and F-value 45.6 (see appendix 12). Based on the analysis we conclude that hypothesis H3c is supported, thus sponsorship perception effects loyalty intention, and there is a positive relationship between the two variables. Note, that so far the analysis did not account for the effects of brand equity, corporate social responsibility and satisfaction on loyalty intention. This will be discussed in section 5.7.
5.5 Effects of Brand Equity and CSR on loyalty intention
H3b: Brand equity is positively related to loyalty intention.
For hypothesis H3b we conducted a regression analysis to see if the independent variable, brand equity, is positively related to the dependent variable, loyalty intention. The analysis shows that there is a significant relationship between the two variables (sign 0.000). Brand equity explains 8.0% of the variation in loyalty intention (see appendix 12), and given the results of the analysis, we conclude that hypothesis H3b is supported.
H4b: Corporate Social Responsibility is positively related to Loyalty Intention.
To test this hypothesis we perform a regression between the independent variable, corporate social responsibility, and the dependent variable loyalty intention. The hypothesis showed a significant relationship (0.000), an adjusted R2 of 10,1%, and a beta value of 7.85 (see appendix 12). This indicates that CSR has a high influence on loyalty intention, and hypothesis H4b is therefore supported.
Based on the analysis, we conclude that both brand equity and corporate social responsibility are positively related to loyalty intention. We will elaborate on these findings in section 6.
5.6 The mediation effect of Brand Equity, CSR and Customer Satisfaction
In this part we will test the mediation effect of the independent variables brand equity, CSR and customer satisfaction on the relationship between sponsorship perception and loyalty intention. First, we will test the mediator function of brand equity, second the mediator function of customer satisfaction, and thereafter the mediator function of Corporate Social Responsibility. Lastly, we will test the mediator function in a multiple regression where brand equity, CSR, customer satisfaction and sponsorship perception are included.
H3d: Brand equity mediates the relationship between sponsorship perception and
loyalty intention.
Loyalty Intentionj = !0 +!1 Brand Equityj + !2 Sponsorship Perceptionj+ "j
To be able to confirm a mediator function, all of the independent variables need to be significant (Baron and Kenny, 1986). In addition to this, the beta value for the relationship between sponsorship perception and loyalty intention should decrease. Short and Bolger (2002) state that there should be a decomposition of the total effect of the relationship between sponsorship perception and loyalty intention into the indirect effect of the independent variables and the direct effect between sponsorship perception and loyalty intention (see Table 3).
! 7%$8!28;,%!
3)./-.+-4()"!"&.56&'5"" 89::" 3)./-.+-4()"!";+6/<"*=%('5"!"&.56&'5" >9:?"
>%&/%8(%!! ?@AB!
Table 3: Decrease in beta value Brand Equity
The analysis shows that the requirements from Baron and Kenny (1986) are met. The independent variables are significant (see Table 4), and there is a decrease in the beta value (see Table 3). Brand equity mediates the relationship between sponsorship perception and loyalty intention, and we therefore conclude that Hypothesis H3d is supported.
" @*'6"A6&%*" 3(,/"
3)./-.+-4()")*+2*)'(./" >9:?" B9BB" @+6/<"0=%('5" C9BD" B9BB"
Table 4: Parameter estimates for regression model explaining sponsorship perception.
H4c: Corporate Social Responsibility mediates the relationship between
Sponsorship Perception and Loyalty Intention.
To test hypothesis H4c, the independent variables CSR and sponsorship perception, and the dependent variable loyalty intention, are included in a regression analysis (see equation under).
The variables are significant, and according to Baron and Kenny (1986), a mediator function is present if all of the independent variables are significant (see Table 5).
" @*'6"A6&%*" 3(,/"
3)./-.+-4()")*+2*)'(./" >9EC" B9BB"
F3G" H9CD" B9BB"
Table 5: Parameter estimates for regression model explaining loyalty intention.
In addition to this, there is a decrease in the beta value, which also implies a mediator function (see Table 6). The results show that CSR mediates the relationship between the independent variable sponsorship perception, and the dependent variable loyalty intention. Hypothesis H4c is therefore supported.
" 7%$8!28;,%!
3)./-.+-4()"!"I.56&'5"" 89::" 3)./-.+-4()"!"F3G""!"I.56&'5" >9EC"
>%&/%8(%!! C@DE!
Table 6: Decrease in beta value Corporate Social Responsibility
H6: Customer Satisfaction mediates the relationship between Sponsorship
Perception and Loyalty Intention.
Hypothesis H6 argues that sponsorship perception is mediated through customer satisfaction, and a multiple regression is performed to test the hypothesis. The independent variables customer satisfaction and sponsorship perception are included, as well as the dependent variable loyalty intention.
Loyalty Intentionj = !0 +!1 Customer Satisfactionj + !4 Sponsorship Perceptionj+"j
The results show an R2 of 12.3%, F-value of 60,5, and the independent variables are significant (see Table 7).
" 7%$8!28;,%! -1F#!
3)./-.+-4()")*+2*)'(./" >9D>" B9BB" F%-'.J*+"36'(-162'(./" K98>" B9BB"
As can be seen from Table 8, there is a decrease in the beta value. Based on Hair et al. (2006) we therefore conclude that customer satisfaction mediates the relationship between sponsorship perception and loyalty intention. Based on this, hypothesis H6 is supported.
" @*'6"A6&%*"
3)./-.+-4()"!"&.56&'5"" 89::" 3)./-.+-4()"!"-6'(-162'(./"#"&.56&'5" >9D>"
>%&/%8(%!! ?@GD!
Table 8: Decrease in beta value customer satisfaction
5.7 Assessing the joint impact of all considered antecedents of Loyalty Intention
To study the joint impact of all the independent variables on the dependent variable loyalty intention, the following equation is tested in a multiple regression:
Loyalty Intentionj = !0 +!1 Brand Equityj +!2 CSRj + !3 Customer Satisfactionj + !4 Sponsorship Perceptionj+ "j
The beta values indicate that CSR (4.40) and customer satisfaction (3.55) are the two variables with the highest influence on loyalty intention. The independent variables brand equity (2.3) and sponsorship perception (1.96) are inferior (see Table 9). The independent variables are significant, and the R2 is 15.9%.
Variable Beta value Sign.
CSR 4.409 0.000
Brand Equity 2.311 0.043
Customer Satisfaction 3.550 0.000
Sponsorship Perception 1.960 0.000
Table 9: Parameter estimates for regression model explaining loyalty intention.
between sponsorship perception and loyalty intention should not be significant if sponsorship perception was fully mediated (Hair et al., 2006). We therefore conclude that brand equity, CSR and customer satisfaction partly mediate the relationship between sponsorship perception and loyalty intention.
Conclusion:
So far in the analysis, we have found several relationships to be significant (see Table 10). As can be seen, the analysis performed indicates positive relationships between the tested variables.
Sponsorship Perception Loyalty Intention
Brand Equity + +
CSR + +
Customer Satisfaction * +
Sponsorship Perception * +
* not tested
Table 10: Parameter estimates for regression joint impact of all considered antecedents of Loyalty Intention
5.8 The Moderator effect of Customer Involvement
H5: Customer Involvement moderates the relationship between sponsorship
perception and loyalty intention.
According to Baron and Kenny (1986), a moderator function is present when the interaction between X*Z is significant. In this analysis, a moderator function is present when the interaction between customer involvement and sponsorship perception is significant. To test if customer involvement has a moderator function in our model we will, in the following paragraph, test each of the four relationships (Analysis I):
C@ 79H!-.I!)=I!-JK)=! D@ -"LH!-.I!)=I!-JK)=! M@ )-<H!-.I!)=I!-JK)=! N@ O*:8;$:H!-.I!)=I!-JK)=!
5.8.1 Analysis I
To test the moderator function we perform a multiple regression analysis for each of the equations above.
Equation 1. First, we performed a multiple regression including brand equity as the dependent variable, while sponsorship perception, customer involvement and the moderator variable are independent variables. The analysis resulted in non-significant values for all independent variables (see Table 11).
Equation 2. Second, we performed the same analysis with the same independent variables, but this time with customer satisfaction as the dependent variable. The analysis, in this case as well, resulted in non-significant values (see Table 11). Equation 3. The third analysis uses the same independent variables, but this time CSR is the dependent variable. The results are similar to the two analysis performed above: non-significant independent variables (see Table 11).
Equation 4. The last analysis we performed to test for a moderator effect of customer involvement is using the same independent variables as above, but here loyalty is the dependent variable. The results showed that sponsorship perception is significant (0.021), while the other independent variables have non-significant values (see Table 11).
! 7/8#'!
9+,1$:!
),($*3%/! -8$1(68&$1*#!
)-<! O*:8;$:!
-.*#(*/(01.!J%/&%.$1*#!P-JQ! B9?>C" B9CE8" B9:>E" B9B>E"
),($*3%/!=#2*;2%3%#$!P)=Q! B9KB8" B9B?C" B9KK8" B9EED"
)=!K!-J! B9BHH" B9HKB" B9BDH" B9K?K"
<D! D9:L" 89HL" EE9>L" K9:L"
*sign. values displayed
Table 11: Parameter estimates for regression model explaining the moderator function
involvement does not moderate the relationship between sponsorship perception and loyalty intention. In the following section, we will test the moderator function further.
5.8.2 Analysis II
To further test the moderator function we perform a multiple regression where all the independent variables in this study are included. The following equation is tested:
Loyalty Intentionj = !0 +!1 Brand Equityj +!2 CSRj + !3 Customer Satisfactionj +!4 Customer Involvementj * Sponsorship perceptionj + !5 Sponsorship Perceptionj+ "j
The analysis gives an adjusted R2 of 15.9% and F-value of 32.9. The independent variables were significant, with the exception of the moderator (sign 0.821) (see Table 12). " -1F#!P.R?@?GQ! 7%$8S28;,%! @+6/<"0=%('5" B9B:>" >988:" F3G" B9BBB" :9:>H" 3)./-.+-4()"M*+2*)'(./" B9B>C" >9E:H" F%-'.J*+"36'(-162'(./" B9BBB" 89KKD" T*'%/8$*/! ?@UDC! S?@?ME!
Table 12: Parameter estimates for regression of overall equation testing for moderator effect
The adjusted R2 in this regression is equal to the adjusted R2 in the mediator regression performed earlier (see Table 13). The equation including the mediator does not include customer involvement, and when including customer involvement there is no change in the R2. This implies that the variable customer involvement does not explain any variation in the dependent variable loyalty intention.
! <D!
0=%6'(./"J*<(6'.+"*11*2'" EK9DL" 0=%6'(./"J.<*+6'.+"*11*2'" EK9DL"
The regression results in a non-significant value, and a negative beta value for the moderator variable (see Table 12). For an interaction effect to be present, the variables need to be significant and have a positive beta value (Leeflang et al., 2000). The result indicates that there is no interaction effect present. We therefore conclude that customer involvement does not have an interaction effect.
Customer involvement has a negative beta value, as well as being non-significant. This indicates that customer involvement has no main effect on the relationship between sponsorship perception and loyalty intention. Based on the analysis performed we conclude that hypothesis H5 is not supported, thus customer involvement does not moderate the relationship between sponsorship perception and loyalty intention.
5.9 Finding segments based on Sponsorship Perception
H7: Segments can be found based on sponsorship perception
We conduct a latent class analysis by using the software LatentGold, to test the hypothesis. We run a cluster analysis with 1-10 clusters solution to be able to find the best cluster solution. To determine the best cluster solution we will use the criteria proposed by Leelang et al. (2000), respectively Bayesian IC (BIC), Akaike’s IC (AIC) and Average Weight of Evidence (AWE). The AWE is important because this criterion combines likelihood and classification and is controlling for the degree of classification. The reported values indicate how well separated the classes are, namely the degree of class overlap. The optimal solution has the lowest value of all criteria, and when in doubt about which solution to use, cluster size and AWE will be used as the deciding variables.
