Charity Highlight and Product Characteristics in

(1)

5/24/2012

Sjoukje Bakker

BE GOOD,

D

O GOOD

,

GET GOOD:

The Influence of Company/Cause Relationship,

Charity Highlight and Product Characteristics in

CSR on Willingness to Pay and Purchase

(2)

1

Master Thesis

By

Sjoukje Bakker

University of Groningen

Faculty of Economics and Business

MSc Business Administration – Marketing Management and Marketing Research

(3)

2

Abstract

As Corporate Social Responsibility (CSR) initiatives by companies are becoming more and more common, more research into specific variables that influence consumers’ behaviour regarding CSR is desirable. From other studies it is known that a CSR initiative by a company positively affects consumers’ purchase intentions and willingness to pay. This research aims to discover the influence of moderating variables on this relationship. A choice based conjoint questionnaire was set up to measure willingness to pay, followed by a list of product combinations measuring purchase intentions, by showing product descriptions which contain combinations of the moderators. The product stimulus used in this study is yoghurt. Demographics and level of altruism were used to describe classes.

From this study it is again confirmed that a CSR initiative of a company increases consumers’ purchase intentions and willingness to pay. Furthermore, the moderating role of length of relationship of the company and the good cause, either a social alliance (long term) or Cause Related Marketing (short term) was investigated. A company being in a social alliance elicited a higher willingness to pay for the product. The relationship with purchase intentions appears to be significant on a 90% confidence level, where a social alliance elicited higher purchase intentions. Secondly, the moderating role of advertisement highlights, either an individual being highlighted, or just a donation to a group was investigated. An individual highlight in an advertisement increases willingness to pay. The relationship with purchase intentions was significant on a 90% confidence level, with increased purchase intentions for an individual highlight in an advertisement. Finally, the moderating role of type of product characteristic, either hedonic or utilitarian was investigated. This moderating role of product characteristic did not appear to be significant. A latent class analysis has been performed, which generated 3 classes. The first class (41.8% of respondents) are the individual caring youngsters: young persons with an average age of 26 years, who prefer hedonic products. When it comes to CSR they prefer an individual highlight and a company being in a social alliance. The second class (8.3% of respondents) are the self-centred older working class: they have an average age of 37 years, are lower educated and prefer utilitarian products without any form of CSR. The third class (48.8%) are the functional big spenders. They are willing to pay the highest prices and prefer utilitarian products. When it comes to CSR they prefer an individual highlight and a company being in a social alliance.

(4)

3

Preface

While writing this preface, I am finally realizing that this thesis is the end of my student time. A great and inspiring time, where I eventually found out what I like most in business: Marketing, also the field of research of this thesis.

I would like to use this preface to thank several persons whom have made this thesis possible. Firstly I want to thank my supervisor dr. Marielle Non for all her help, precise feedback on my writing, her excellent and sharp statistical insights and also because it was really nice and fun working together with her. I would also like to thank dr. Jenny van Doorn, who gave feedback on the content of the literature part. Finally, I would like to thank my second supervisor drs. Niels Holtrop for his feedback on my thesis. Next to the help of the professionals, I would also like to thank my family, friends and boyfriend for their support and for filling out and forwarding my questionnaire.

Sjoukje Bakker

(5)

4

1. Introduction

1.1 Background

Companies that support and engage in green and social initiatives are becoming more and more common nowadays. More than ever, companies do not only focus on obtaining economic benefits, but they seek to pursue social and environmental goals as well (Chabowski, Mena and Gonzalez-Pardon, 2010). The concept of “triple-bottom line performance” (Elkington, 1997), where the success of a company is not only measured by its economic performance (profit), but also by its environmental (planet) and social performance (people), has gained interest from companies. Businesses can participate in green, environmental activities (planet), where the company engages in environmentally friendly activities, but they can also employ activities focusing on social welfare (people), or do both. This research paper aims at the social aspect of “doing good of companies”, focusing on Corporate Social Responsibility.

Corporate Social Responsibility has become practically common among companies, with 95% of the 250 largest companies in the world (G250 companies) that report corporate responsibility activities (KPMG International Survey of Corporate Social Responsibility Reporting, 2011). Moreover an estimated amount of $1.7 billion has been spent on good causes by companies in 2010 in the United States. This represents a growth of 6.7% as opposed to the spending in 2009 (IEG Sponsorship report, 2010).

These high expenses on CSR are partly done for social and ideological reasons of the company, supporting worthy causes, but also because companies expect a positive influence from it on their business results. In response to the on-going growth of Corporate Social Responsibility in business, CSR has become an important field of research in understanding the importance of CSR activities in the area of marketing and many other fields (Varadarajan and Menon, 1988).

(8)

7 In the field of marketing research, researchers have demonstrated that CSR positively influences many of these customer related outcomes. When assuming that all other factors stay constant, consumers show more positive attitudes towards companies that engage in CSR actions, and they are more likely to purchase from these companies (Sen and Bhattacharya, 2001; Barone, Miyazaki and Taylor, 2000). Consumers also show increased loyalty when a company engages in CSR (Du, Bhattacharya and Sen, 2007) and they are willing to pay premium prices (Creyer and Ross, 1996; De Pelsmacker, Driesen and Rayp, 2005). Moreover CSR initiatives have a positive influence on marketing campaign evaluations (Sen and Bhattacharya, 2001) and result in enhanced reputation of the company (Brown et al. 1997). Each of these marketing outcomes in turn supports financial performance (Peloza and Shang, 2010).

1.2 Goal and Relevance of Research

From previous studies it has become apparent that different forms of CSR initiatives by a company have positive effects on customer related outcomes, leading to improved performance of the company. It is important to know for managers who want to implement a CSR initiative by what factors this relationship is driven, in order to specifically design and implement the optimal CSR initiative. It is therefore important to know: a) what type of relationship between the company and the cause is best, for b) which forms of a CSR claim, and c) for what type of product characteristics have the optimal positive effect on consumers. Thus, what possible moderators influence the relationship between a CSR initiative and customer related outcomes.

(9)

for-8 profit company and a good cause but is a more long-term partnership, more similar to a marketing based strategic alliance, than to the relatively short-term relationship of Cause Related Marketing (Berger, Cunningham and Drumwright, 2006). Less is known about the effects of a company being in a social alliance on consumer related outcomes. Is there a difference in customer related outcomes between a more short term relationship, and a long term relationship? Webb and Mohr (1998) suggest that companies that show a long-term commitment towards a cause might be able to overcome consumers’ scepticism. It is therefore proposed that a long term commitment in the form of a social alliance will be more effective than a more short term commitment like cause related marketing. When deciding to engage in CSR, it is also important for managers to know what type of charity to support and how it should be communicated. Type of charity has been extensively examined in literature. Research has shown that a good fit between the company and the charity improves consumer attitudes towards the company and the company-cause alliance (Arora and Henderson, 2007; Barone et al. 2007; Ellen et al, 2000; Nan and Heo, 2007). Also supported charities that are more personally relevant for consumers, are rewarded by more positive attitudes: donations to local charities improve consumer attitudes and purchase decisions. Moreover, consumers who are higher involved with a charity show more positive attitudes towards the company (Grau and Folse, 2007). Also CSR initiatives that focus on a current disaster (versus an ongoing cause) are more preferred by consumers (Ellen et al, 2000). These studies could be expanded by finding out whether it matters, when communicating the support of the good cause, if one individual directly benefits from the supported good cause, as is recently done at Pampers, where one child in Africa gets a tetanus vaccination after each purchase? Or is just donating money to a good cause, without an individual directly benefitting from it, more effective? The moderating role of the type of charity, or cause that is supported has been examined by various researchers.

(10)

9 hedonic product in combination with donation to a charity instead of a cash refund, than a utilitarian product in combination with a donation.

1.3 Problem Definition and Research questions

This research aims to discover which type of company-cause relationship, for which forms of a CSR claim and for what type of products have the optimal positive effect on consumers. The effects in this research will be measured by two dependent variables. The willingness to pay and the purchase intentions of a consumer - which are in fact frequently used dependent variables in previous studies in the area of CSR- will be used as dependent variables.

Three moderators will be added to explain the relationship. Firstly whether the type of relationship between the company and the cause, either short-term or long-term, influences WTP and purchase intentions. The second moderator is whether a CSR claim highlights an individual, or a non-profit organization (collective). In an individually highlighted CSR claim, an individual will be supported. For instance, per sold product one child will get a tetanus vaccination. The collective highlighted CSR claim will only state that a donation to a charity will be made. Thus, without a direct individual benefitting destination for the money that is donated to the charity. Finally, the type of product aspect, either hedonic or utilitarian that is highlighted is one of the moderating variables

Hence, the main question of this study is:

“What is the effect of a CSR claim on consumers’ willingness to pay and purchase intentions?”

Including the following research questions:

 What is the influence of the type of relationship, either short term, or more long term oriented, between the company and the cause on consumers’ willingness to pay and purchase intentions?  What is the influence of the type of charity of the CSR initiative and whether it is individually or

collectively highlighted on consumers’ willingness to pay and purchase intentions?

(11)

10

1.4 Structure of thesis

(12)

11

2. Literature review

2.1. Literature on Corporate Social Responsibility

This section gives an overview of the literature on Corporate Social Responsibility, in order to build a theoretical framework for the hypotheses formed later on in this chapter.

2.1.1 Defining Corporate Social Responsibility

A definition of Corporate Social Responsibility needs to be made, in order to set boundaries for this research. This research focuses on CSR in marketing and will therefore narrow its definition into CSR in marketing. Moreover, in this research the focus will be on supporting a good cause in the area of a social initiative, rather than a green, environmental initiative. The definition will thus be made for CSR in marketing with a social focus.

Vaaland et al (2008) define CSR as: “Corporate social responsibility is the management of stakeholder

concern for responsible and irresponsible acts related to environmental, ethical and social phenomena in a way that creates corporate benefit” (Vaaland et al, 2008). They found that CSR is especially relevant

because of the focus it has on business benefits in combination with stakeholder groups, like customers. As can be seen from the definition, the emphasis of CSR mainly lies on corporate benefit, stakeholders, both responsible and irresponsible acts and the environmental and social phenomena. From CSR, companies basically pursue two main goals: improving business performance and support good causes (Varadarajan and Menon, 1988). The definition of Vaaland et al (2008) will be adopted in this paper, where for this research the focus is on the social phenomena used in this definition

(13)

12

2.1.2 Outcomes of CSR in literature

There is an extensive literature on the relationship between CSR initiatives of companies and customer- and business related outcomes. The most relevant findings will be shortly described in this sub section. The first benefit of companies engaging in CSR activities is that they are rewarded by consumers with more positive attitudes and evaluations towards the company (Nan et al. 2007; Brown and Dacin, 1997; Sen and Bhattacharya, 2001; Berger et al. 2006). Positive CSR beliefs of consumers result in enhanced product evaluations, whereas negative CSR beliefs can have a harmful effect on overall product evaluations (Brown et al, 1997). Moreover, CSR initiatives lead to more positive campaign evaluations and favourable consumer attitudes towards the non-profit cause (Sen et al, 2001; Berger et al. 2006). Also, Du et al (2007) find that when consumers form positive beliefs of a company, believing that a company is social responsible, the likelihood of purchasing a product, or service from that company will become greater. They found that consumers reward social responsible companies with longer-term loyalty and advocacy behaviour, such as positive word-of-mouth. Consumers show more resistance towards negative brand information. In line with this, companies that support social causes, for what consumers think to be genuine social motivations of the company, will be gratified with an increase in the choice of a company’s brand (Barone et al. 2000). Also Brown et al, (1997) find an increased purchase likelihood for companies exploiting a social initiative. There are studies suggesting that consumers are willing to pay premium prices for more ethical and socially responsible products (De Pelsmacker, 2005; Brown et al. 1997; Arora et al. 2007). Arora et al (2007) found that when a claim of a social cause is added to a product or service, the willingness to pay premium prices increases together with the quality perceptions of the product or service. Finally, Luo and Bhattacharya (2006) found that CSR initiatives positively influence customer satisfaction, which in turn positively mediates the market value of a company. Thus, CSR helps to create satisfied customers and this in turn helps to increase the market value of the company. An example is given by Luo et al (2006), that “given a company with an average market value of $ 48 billion, one unit increase of CSR ratings would result in approximately $ 17 million more profits on average in subsequent years, a substantial increase of financial returns” (Luo et al. 2006).

(14)

13 results in an impediment of benefitting from the positive business returns gained from CSR (Du et al. 2010). To overcome this, companies can more actively communicate their CSR activities, by targeting on consumers with specific interests in good causes. Targeting on these interests gives the highest chance in increasing consumers’ awareness of a company’s CSR activities (Pomering and Dolnicar, 2009). And although there is a positive relationship between evaluation of the CSR initiative and the purchase intentions (Ciu, Trent, Sullivan and Matiru, 2003), consumers need to actually believe in the company’s good intentions for engaging in CSR (Sen et al. 2006), that is, consumers need to believe that the company is engaging in CSR because it is genuinely concerned about the good cause. Du et al. (2007) find that a company that positions itself as a CSR embraced company, with CSR as its core business strategy (intrinsic motivated), is more likely to profit from positive customer related outcomes, than a company that just exploits a CSR initiative because it has to, due to financial or competitive pressure (extrinsic motivated). This is because consumers tend to favour CSR initiatives more of companies that, in consumers’ mind are genuinely concerned in being socially responsible, than CSR initiatives that are exploited by more egoistic motives (Du et al. 2007). Negative attributions formed by consumers about a CSR initiative, lead to less consumer trust and support (Vlachos et al. 2009). In line with this, Ellen et al (2006) find that consumers that believe in the company’s good intentions, by viewing a company’s motives to exploit a CSR initiative as values driven, strategically oriented and other-centered, have an increased purchase likelihood. Research has also shown that the reputation of a company in terms of how socially responsible it is influences how consumers judge the CSR initiative and the company (Dean, 2004). According to Dean (2004), firms with a good reputation gain little advantage from using a CSR claim with respect to consumers’ attitudes. Companies with an irresponsible reputation gain slightly in favour from consumers once they initiated a CSR claim. However they attributed negatively stated motives about, what they thought, were the company’s reasons to exploit a CSR initiative. Whereas Yoon et al (2006) found that consumers form more positive attitudes towards a CSR initiative from a company that they believe has a good reputation when it comes to social responsibility.

(15)

14 When it comes to communicating a CSR initiative, the fit between the company and the social cause is an important factor. Consumers often expect companies to only sponsor those social issues that have a good fit, or a logical association with the core company activities (Du et al. 2010). Some researchers find no relationship between the fit of the company-cause and customer related outcomes such as positive attitudes and enhanced purchase intentions (Lafferty, 2007), or a partial relationship (Nan et al, 2007) where brand/cause fit has only effect for consumers with a high brand consciousness. However, most researchers found a positive relationship. Results of Barone et al. (2007), Basil et al (2006), Ellen et al. (2000) suggest that a good fit between the company and the cause, generally improves consumer attitudes, towards the company-cause alliance, the CSR campaign, and towards the company itself. Moreover purchase intentions are improved by a good fit.

Next to a good fit, consumers need to find the charity personally relevant. Consumers’ choice for a product is influenced by CSR initiative of companies that are involved with their favourable charities (Arora et al, 2007). When communicating a CSR initiative, also the source of communicating the CSR initiative is of importance. Yoon, Gurhan-Canli and Schwarz (2006) found that consumers’ reactions were more positive when they became aware of a company’s CSR initiative from a neutral and independent source. Companies also need to take into account that employees are an important source for communicating the company’s principles; employees have shown to be active in advocacy behaviour like word of mouth (Dawkins, 2004). In line with this, Drumwright (1996) found that campaigns with a social aspect are also effective in motivating employees and that they in turn will communicate the social mission of a company.

Moreover, the choice of the supported cause moderates the effects that CSR produces. As already mentioned, there needs to be a good fit between the supported cause and the company. But also the type of relation between the company and the cause matters. A long-term commitment of a company to a charity is seen as a sincere concern of the company for the cause, which shows more favourable consumer attitudes, whereas short-term commitment is seen as a way of exploiting CSR for reasons of making profit. (Webb and Mohr, 1998; Ellen et al, 2006).

(16)

15 different segments of consumers varying in their reactions towards CSR initiatives. Firstly, there are the “Sceptics”, who are driven by scepticism and are afraid of being taken advantage of by the company and its CSR initiative. The second segment includes the “Balancers” who are fairly positive towards the CSR initiative, but not highly involved in CSR activities. The third segment consists of “Attribution Oriented” consumers, who are most concerned about the motives of a company’s CSR initiative, and they are willing to believe that there are at least some good motives behind the company’s engagement in CSR. Finally, there are the “Socially Concerned” consumers, who are positive about CSR initiatives and highly involved. Especially the first group of consumers is an issue to work on for companies. Suspicious, or sceptic consumers form mainly negative attributions about the CSR initiative, which in turn lead to less internal consumer responses, like trust and support in the company as well as less behavioural consumer responses, like recommendation intentions (Vlachos et al. 2009). Moreover, the positive business effects of a CSR initiative can become negative when consumers do not believe that the company engages in CSR because of genuine concern of the cause, due to suspiciousness of consumers (Du et al, 2010).

From the above results, it can be concluded that is important to design the right CSR initiative, taking into account many factors that influence consumers’ attitudes, in order to evoke positive attitudes from consumers. Marketing communications tools can play a major role in effectively communicating a company’s CSR activities (Jahdi and Acikdilli, 2009).

2.2 Dependent Variables

(17)

16

2.2.1 Willingness to pay

One of the dependent variables to measure the successfulness of a CSR initiative is the willingness to pay (WTP) for a product. “The WTP is the maximum amount of money a consumer is willing to spend on a product or service” (Cameron and James 1987). Thus, “WTP is a measure of the value that a person assigns to a consumption or usage experience in monetary units” (Homburg et al, 2005). WTP has been examined extensively in marketing literature, among others in the areas of advertising (effectiveness) and consumer dealing patterns (Homburg et al, 2005). Creyer and Ross (1997) found that the ethicality of a firm’s behaviour is an important consideration when it comes to consumers’ purchase decisions. They found that consumers are willing to pay higher prices for products when the company behaves ethical. Also, consumers punish unethical firms. Consumers are still willing to buy from an unethical firm, but they punish unethical companies with a lower willingness to pay. Also Arora et al (2007) found that when a social cause is added to a product or a service, the willingness to pay premium prices increases. Moreover, research of De Pelsmacker (2005) and Brown et al. (1997) suggests a higher willingness to pay premium prices for ethical and socially responsible products. From this, it is expected that there is a positive relationship between an expression of a company engaging in CSR, and the willingness to pay for a product of the company.

2.2.2 Purchase intentions

(18)

17 Based on the findings that show a positive relationship between a CSR initiative and willingness to pay and purchases intentions, the first hypothesis is stated as:

H1 Including a CSR claim in an advertisement positively affects consumers’: a) Purchase intentions; b)

Willingness to pay.

2.3 Moderating Variables

The relationship between a CSR claim and the dependent variables will be moderated by three variables. Firstly whether the type of relationship between the company and the cause is long-term, or short-term. Second, whether the charity used in the CSR claim is individually or collectively highlighted. And finally whether the product offered in the advertisement is highlighted with a hedonic, or a utilitarian characteristic. The three moderating variables will each be explained, followed by the hypotheses.

2.3.1 Relationship between company and cause

The type of relationship between the company and the cause possibly moderates the effect of a CSR claim on consumers’ willingness to pay and purchase intentions. Do consumers favour a long-term relationship more than a short-term relationship? A distinction will be made between more short-term oriented relationships as cause related marketing, and more long-term oriented relationships like social alliances. This will be explained further in this section.

(19)

18 Similar to marketing alliances, social alliances are close, mutually beneficial partnerships, where both parties share their knowledge, resources and capabilities (Berger et al, 2006). In a social alliance, the company benefits mainly from improved social image and publicity, whereas the charity receives financial contribution, managerial knowledge, or volunteers from the company in the alliance(Liu and Ko, 2011).

Thus, a social alliance is a more long-term oriented relationship between the company and the cause, whereas cause related marketing is a more temporary, short-term oriented relationship. Research has shown that a long-term commitment of a company to a good cause is seen by consumers as a genuine concern of the company for the cause, which in turn leads to more favourable consumer attitudes. A short-term commitment is seen a way of exploiting CSR for reasons of making profit (Webb and Mohr, 1998; Ellen et al, 2006). Moreover, companies that position themselves as a CSR embraced company, with CSR as a core business strategy, are more likely to be rewarded by consumers, by several consumer related outcomes (Du et al., 2007). Du et al distinguish two different motives, consumers think a company has for exploiting a CSR initiative. Intrinsic motives are interpreted by consumers as the company’s genuine concern in being socially responsible. Extrinsic motives are interpreted by consumers as motives of the company engaging in CSR for competitive pressure and financial benefits. These are more egoistic motives aimed at increasing the company’s own welfare (Du et al. 2007). Consumers thinking that a company has egoistic motives, can raise their suspicion and lower the trust in the company, resulting in less recommendation intentions (Vlachos et al. 2009).

A long term relationship might contribute to what consumers think are genuine motives for a company exploiting a CSR initiative. In line with this, research has shown that longer term relationships can lead to positive advocacy behaviour by employees, which can in turn lead to more positive brand associations. Research has shown that the long-term social alliances make employees identify themselves more closely with their organization, which in turn enhances organizational identification (Berger et al. 2006). It is therefore expected that the more long-term oriented social alliance will have a more positive influence on consumers’ WTP and purchases intentions.

(20)

19

H2 A long-term relationship between the company and the cause in the form of a social alliance, will

positively moderate the relationship between the CSR initiative and a) Purchase intentions and b) Willingness to pay.

2.3.2 Individual charity versus Collective Charity

The way in which the philanthropy activity in the CSR claim is highlighted possibly moderates the relationship between a CSR claim and WTP and purchase intentions. It could make a difference whether the CSR initiative claimed by a company highlights an individual that will be directly helped, or that a donation to an organization (collective) is made.

In an individually highlighted CSR claim, an individual will be supported. For instance per sold product one child will get a tetanus vaccination (as is recently done at Pampers). A collective highlighted CSR claim will state that a donation to a charity will be made (for instance to Save the Children). Thus, without a direct individual benefitting destination for the money that is donated to the charity. The main difference between these two claims is that the individual claim assigns a more direct purpose to the charity in the CSR claim. Namely an individual that directly benefits from the money donated. The collective claim does not have a visible direct purpose, since the only thing known is that the donation goes to an organization. And it is therefore not visible that one individual benefits directly from the donation made to the charity.

(21)

20 In line with this, Simon, Stürmer and Steffens (2000) did research into the self-categorization theory. This theory is mainly based on the fact that people are more willing to help other people, when they see the other people as members of their own group (in-group), than other people from outside the group (out-group). This effect is stronger for people with a high collective identification, meaning that people strongly identify themselves with groups as opposed to people with an individual identification, who do not have this group feeling and see themselves as unique individuals.

Small and Loewenstein (2003), studied the effect that victims in a CSR initiative have on persons. They researched whether persons care more about helping a victim that they can identify (by making it personal), than victims they only get some statistical and descriptive information about. They manipulated this by exposing a charity that benefits a family that had been selected from a list and was thus identifiable. As opposed to being exposed to a charity helping a family that would later be selected from a list. It appeared that people are more willing to contribute to a charity that benefits a family that is identifiable. Elaborating on this, Kogut and Ritov (2005), found that specific information about one victim is perceived as more alive and emotionally distressing, than only information given about a group of unidentified victims. Specifically they found that one identified victim created significantly more willingness to contribute to a good cause, than a victim that was not identified. A group of victims was found not to have any effect on the willingness to contribute to a good cause. They concluded that the emotional reaction respondents have to the victim exposed, has an effect. It could therefore be useful to know whether an individual, identified victim elicits higher willingness to pay and purchase intentions than an unidentified group.

(22)

21 It is thus expected that an individual highlighted charity will positively moderate the relationship between the CSR initiative and Willingness to Pay and Purchase intentions. From this, the following hypothesis is formulated:

H3 An individually highlighted claim will positively moderate the relationship between the CSR initiative

and a) Purchase intentions and b) Willingness to pay.

2.3.3 Hedonic and utilitarian product characteristics

A difference in type of product could influence consumers’ willingness to pay and purchase intentions. A distinction in product type can be made by categorizing products into either hedonic consumed products, or utilitarian consumed products. Hedonic products anticipate on more experiential consumption, responding to motives of fun, pleasure and excitement. “It relates to the multisensory, fantasy and emotional aspects of the product usage experience” (Hirschman and Holbrook, 1982). Examples of these products are designer clothes, luxurious jewellery, or sports cars. Utilitarian products on the other hand, are mainly instrumental, functional and goal oriented, like personal computers or kitchen supplies (Hirschman and Holbrook, 1982).

It is known that consumers are more likely to respond positively to a single opportunity to consume a hedonic good than to a single opportunity to consume a utilitarian good. But when a purchase opportunity arises with both utilitarian and hedonic products, for which rational thinking and justifying a choice is needed, consumers tend to choose the utilitarian product. This because when fun and pleasure related products are consumed, this raises issues of guilt and need for justification in consumers mind. Therefore, consumers tend to have a higher need for justifying a hedonic choice, than a utilitarian choice (Okada, 2005).

(23)

22 This suggests that “the affect derived from the consumption of a hedonic product, may be complementary to the utility derived from contributing to a good cause” (Strahilevitz et al. 1998). The pleasure and guilt that is felt by consumers when consuming a hedonic product, contributes to the positive effect that is the result of a good feeling towards giving money to a charity. When contributing to a good cause, the consumption of hedonic products may be justified (Strahilivetz et al. 1998; Van Doorn et al. 2011). Van Doorn et al. (2011) did a study in the organic food industry, and found a difference in the positive effect of an organic claim in virtue products and vice products. Vice products are products with a pleasure experience, similar to hedonic products. Virtue products are products less negative on long term consequences, similar to utilitarian products. They found that the positive effect of an organic claim on perceived pro-social product benefits is stronger for vice food products, than for virtue food products. This is partially motivated by the fact that an organic claim works as a guilt reducing complement, and thus will reduce the guilt that is evoked by the consumption of a vice product.

Thus, it is known that hedonic products require more need for justification, and it appears from previous research that a donation to a charity or an organic claim can work as a justifier for consumers. In order to make this research more compact, it is decided to only use one product in the questionnaire that has both hedonic and utilitarian ways to consume. In the questionnaire either a claim of the product highlighting the hedonic characteristic of the product, or a utilitarian characteristic will be highlighted. Based on the results of research on hedonic and utilitarian products in CSR, it is expected that the positive relationship between a CSR claim and the willingness to pay and purchase intentions is stronger for products that are highlighted with a hedonic-, than a utilitarian characteristic. This is due to the fact that consumers seek for guilt reducing attributes when consuming a hedonic product. A CSR claim can bring this justifying, guilt reducing claim to respondents when looking at an advertisement with a product with hedonic characteristics highlighted.

Therefore the following hypothesis is formulated:

H4 Highlighting a hedonic characteristic of a product positively moderates the relationship between a

CSR initiative and a) Purchase intentions and b) Willingness to pay.

2.4 Control Variables

(24)

23 segments. These variables and previous studies explaining the relationship of these variables with CSR will be explained in this chapter.

2.4.1 Demographics

A hypothetical moderating variable of the relationship CSR and the dependent variables is the lifestyle of consumers. The lifestyles are represented by consumers’ activities, interests, opinions and demographics (Plumer, 1974). Since the activities, interests and opinions are mainly qualitative data and will thus not fit in a model, only demographics will be used to explain consumer characteristics. Plumer (1974) distinguishes 9 characteristics of demographics: age, gender, education, income, occupation, family size, dwelling, geography, city size and life cycle stage. From previous studies it appears that females form more positive attitudes towards CSR initiatives than males (Ciu et al, 2003; Winterich, Mittal and Ross, 2009). There are also indications that income is a positive moderator (Mc Williams and Siegel, 2001). Moreover, in a research among students, a higher income of the parents positively influenced attitudes towards the CSR initiative (Ciu et al, 2003). Furthermore the type of education seems to have an influence: social science majors show more favourable attitudes towards CSR initiatives than other majors (Ciu et al, 2003). Also age positively moderates consumers’ attitudes, older people tend to support CSR more than younger people (Basil and Weber, 2006). From this, it is expected that demographics can predict class membership of CSR segments.

2.4.2 Social values of consumers

Another variable moderating the relationship between CSR and the dependent variables are the social values of consumers, specifically the level of altruistic personality. How concerned are respondents for the welfare of others? Are respondents with more altruistic personalities more likely to favour a company’s CSR initiative than respondents with a more selfish personality?

(25)

24 others and from these individuals it is known that they are more likely to help others (Van Lange et al, 1997)”. It is therefore expected that consumers that are more other-centered and social will respond more positively towards a CSR claim of a company, than more self-centered consumers.

Several studies developed scales to measure the level of social values of individuals, measuring the level of altruism. This means: how concerned of others’ welfare respondents are. Carlo et al. (1991) found contexts in which an altruistic personality might be most apparent. They use a combination of scales to measure altruism, which can be used for this research. Osterhus (1997) also developed a scale to measure the social norm of consumers toward the environment. Although the research of Osterhus (1997) aims at environmental initiatives rather than social initiatives, the scales of measuring social norms and responsibility towards the environment can easily be transformed into a scale measuring the same for social initiatives. Furthermore, Berkowitz and Lutterman (1968) developed a scale to measure how responsible and concerned about the welfare of others, consumers are.

2.5 Conceptual model

Below, the conceptual model of this research is displayed.

(26)

25

3. Research Design

This research design is the blueprint for this research. The information that is needed for this research is specified, as well as the design of the questionnaire with measurement- and scaling procedures. Furthermore it contains a plan of analysis and the results of a pre-test of the questionnaire.

3.1 Goal

In order to answer the research question and to test whether the hypotheses are supported, an experiment in the form of an online choice questionnaire made by the program Sawtooth SSI Web will be performed. The questionnaire aims to measure the following outcomes:

- The willingness to pay for each combination of the company/cause alliance, type of charity highlight of the advertisement and type of product characteristic.

- The purchase intentions for the company/cause alliance, type of charity highlight of the advertisement and type of product characteristic.

- To what extent the type of company alliance, type of charity highlight and type of product characteristic moderate the purchase intentions and willingness to pay.

- A segmentation of consumers with regard to their responsiveness to CSR initiatives, described by demographics and social values of consumers.

3.2 Design of the questionnaire

This research will be performed by means of an online choice questionnaire. The questionnaire consists of three parts. Firstly a choice questionnaire in the form of displayed advertisements texts of a product, from which consumers have to choose the ones they prefer. Secondly a section measuring purchase intentions, by means of product descriptions which contain combinations of the moderators. The final section contains questions measuring the altruism of consumers and demographic questions. The questionnaire can be found in appendix 1.

3.2.1 Choice based conjoint

(27)

26 information across respondents, it is possible to include interactions” (Sawtooth manual). Respondents in this research are asked to evaluate the stimuli; the advertisements that consist of combinations of the different moderators (attribute levels). Next to the proposed moderators, an additional attribute will be added; the price of the product offered in the advertisement. This will be later on used to measure the willingness to pay. Appendix 2 shows an overview of the scales used for the questionnaire.

The type of product that will be used in the questionnaire is yoghurt. As yoghurt can be both hedonic in relating to response of pleasure, fun and other emotional aspects, and utilitarian in relating to functional and instrumental needs (Hirschman and Holbrook, 1982). Yoghurt can be useful in that it is healthy (utilitarian), but also fun and delighting (hedonic). Moreover, for respondents it is easy to determine the price of yoghurt, they have a clear view of the price they want to pay for yoghurt.

The following sections explain per variable how they are manipulated in the advertisements displayed in the choice questionnaire.

CSRClaim

The CSR initiative and the claim that are used for the product in the advertisement should have a good fit between the product and the cause (Basil et al. 2006). As explained earlier in this section, the product that will be chosen for this research is yoghurt. Thus, a prerequisite for the research is that the CSR initiative claimed in the advertisement should have a fit with the product yoghurt. It is manipulated in the advertisement texts by employing CSR projects that focus on dairy products (milk, yoghurt) for children in Africa.

(28)

27

Company/cause relationship and Charity highlight

The two CSR related moderators company/cause relationship and charity highlight will be combined into one attribute, in order to take into account an advertisement without CSR as well. Using these two moderators, which have two levels added by an option of an advertisement without CSR leads to an attribute of 5 levels:

1) An advertisement of a company without a CSR initiative;

2) An advertisement of a company with a CSR initiative that is in a social alliance, with an individually highlighted charity claim;

3) An advertisement of a company with a CSR initiative that uses cause related marketing, with an individually highlighted charity claim;

4) An advertisement of a company with a CSR initiative that is in a social alliance, with a collectively highlighted charity claim;

5) An advertisement of a company with a CSR initiative that uses cause related marketing, with a collectively highlighted charity claim.

The next two sections explain per moderator separately how they are manipulated.

Company/Cause relationship

(29)

28

Table 3.2.2.1 Stimuli companies

Company Stimulus in questionnaire

Company X Company X is a big, internationally operating dairy company that produces and sells among others yoghurt.

Company Y Company Y is a big, internationally operating dairy company that produces and sells among others yoghurt. The company has a long-term cooperation with Unicef. They jointly founded a charity: “Dairy for Africa”. The goal of this charity is to give children in Africa dairy products at school.

Company Z Company Y is a big, internationally operating dairy company that produces and sells among others yoghurt. The company financially supports Unicef this year, by donating a part of their profit to Unicef. With this money, Unicef starts projects giving children in Africa dairy products.

Type of Charity highlight

The two companies in the advertisement that do have CSR initiatives (company Y and Z) can have a charity claim with either an individual, or a collective highlight. Table 3.1.2.2 shows how the highlights are manipulated.

Table 3.2.2.2 Stimuli charity highlight

Charity highlight Claim

Individual When you buy one package of yoghurt, one child in Africa will receive free school milk during one day.

Collectively The company donates 10% of their profit to a project of Unicef focusing on dairy in Africa.

Type of product characteristic

Consumers will see an advertisement with a product that is highlighted with either a utilitarian

(30)

29 of yoghurt is manipulated by phrasing a claim focused on how nice and creamy it is, including pieces of chocolate. In relating to functional and instrumental needs, the utilitarian characteristic of yoghurt is manipulated by phrasing a claim related to the good bacteria it contains, which improves digestion. Results of the pretest showed that the differences at first were too small; respondents did not see a clear difference. Therefore, two brand names were created namely Noma yoghurt (hedonic) and Actifit (Utilitarian), in order for respondents to see quickly, clear differences between the products.

Willingness to pay

The willingness to pay, the maximum amount a consumer is willing to pay for a product, or service, will be measured by the attribute price in the choice set. Measuring willingness to pay by using a conjoint analysis is the preferred way (Green et al., 1978). The choice set contains three levels, thus using three prices for yoghurt, namely €1.29,- , €1.49,- and €1.69,-. These prices are based on the range of prices of similar types of yoghurt in Dutch supermarkets, ranging from relatively cheap, to relatively expensive. Measuring willingness to pay in a way that only one of three price options can be chosen, has limitations, because respondents are not obliged to really buy the product. However, for this research it is sufficient, because the researcher is more interested in the relative differences between willingness to pay, than the actual prices consumers are willing to pay for a product.

3.2.2 Purchase intentions

The purchase intentions of consumers are measured in a separate part of the questionnaire. All the different types of company/cause relationships, the charity highlights and the type of product characteristics are shown again. The attribute price is left out in this section, because the number of questions (30 questions) asking purchase intentions for each set of attributes would have been too large. Leaving the attribute price out, means that the respondent has to evaluate another 10 (5 CSR levels * 2 product characteristic levels) product advertisements related to their purchase intentions. Purchase intentions are measured on a three point Likert scale, as is done in the research of Sen et al. (2001). They are asked how likely it is they will buy the product as is advertised, giving them three options: 1) Not likely at all, 2) Maybe, or 3) Very likely.

3.2.3 Social values

(31)

30

3.2.4 Demographics

Based on Plumer (1974), a set of demographic questions will be asked, in order to describe segments later on. By means of multiple choice questions gender, education, income, source of income, family size, geography, city size and life cycle stage will be measured. The demographic question age will be measured by means of an open question, for which afterwards categories will be made. The other demographic questions will be measured by means of multiple choice questions.

3.3 Data collection and sampling

This research focuses on respondents in the Netherlands. It aims to find as many respondents with different demographics to obtain a representative sample. Everybody is allowed to participate in the questionnaire, but respondents do have to be 18 years, or older. The technique that will be used to find a sufficient amount of respondents is Snowball Sampling. This method aims to obtain a research sample through “referrals made among people who share, or know others who possess some characteristics that are of research interest” (Biernacki and Waldorf, 1981). An initial group of respondents is selected to participate in the research, afterwards these respondents are asked to find other respondents who could also participate in the research. The main advantage of this type of sampling is that it “increases the likelihood of locating the desired characteristic in the population, and that results in relatively low sampling variance and costs” (Malhotra, 2007). In this research the initial group of respondents that are selected are friends, family and acquaintances of the researcher. They are approached by email and the social media channels Facebook and Linkedin and are asked to send the questionnaire to their acquaintances. The minimum amount of respondents needed for this master thesis research is 200 respondents, as this number has been found to provide an acceptable margin of error (Hair et al., 2010).

3.4 Plan of analysis

This section explains the tests that will be performed in order to measure the outcomes of this research. The tests used for this research are firstly, the conjoint analysis by Sawtooth, which will measure the willingness to pay and the interaction effects. Sawtooth will also create segments, which will be described by Multinomial Logit, as Sawtooth does not have the option do describe segments. Secondly, the purchase intentions will be measured by means of Ordered Multinomial Logit.

3.4.1 Conjoint Analysis

(32)

31 used to create segments of consumers based on CSR responsiveness. The program Sawtooth Latent Class will be used to perform Conjoint Analysis on both aggregate, and segment level.

“Conjoint analysis attempts to determine the relative importance consumers attach to salient attributes and the utilities they attach to the levels of attributes “(Malhotra, 2007). Conjoint analysis models consumer preferences using multi-attribute alternatives, in order to understand the development of consumers’ preferences for a product, based on the utility of the chosen attributes for the product (Green, 1978). In this research, conjoint analysis assesses the relative impact of each highlight of the charity, the form of relationship and the type of product characteristic in order to determine a consumers’ utility by evaluating the profiles. As described earlier, the method choice based conjoint analysis is chosen, because this method gives a greater robustness of the results, it shows more than one option (including a none option), which is more realistic. It gives a better, more realistic representation of the behavior of consumers in a store. It is more realistic because having consumers choose one option out of more given options is more realistic. Moreover, it has the ability to estimate interaction terms, in order to test for significance of the moderators. The choice based conjoint analysis is performed in Sawtooth, where the importances of the attributes are estimated, as well as the willingness to pay and whether or not the suggested moderators moderate the dependent variables. The stimuli will be constructed by the use of the program Sawtooth. This program ensures an efficient design of the conjoint analysis.

3.4.1.1 Conjoint Analysis design

The number of attributes in the conjoint analysis is three. Green and Srinivasan (1978) recommend a maximum number of 6 attributes in order to be clear enough for respondents, thus three falls within the range of the allowed number of attributes. Moreover, when a design contains 6 or less attributes, a full profile method showing all attributes per choice set is recommended, hence this will be used for this research.

(33)

32 chosen too often and to find out whether respondents find a choice set of 3 options and a none option comfortable to choose from.

In total, 10 choice sets are presented in the questionnaire. Although research has shown that up to 20 choice sets can be reliably answered (Hair et al, 2010), this research will only have 10 choice sets. This because a lot of information per choice set is displayed, as advertisement slogans and product descriptions are used. Moreover, after the choice sets, the respondents have to analyze another 10 choice sets on purchase intentions, which then in total adds up to 20 questions. All combinations of attribute levels may be shown together in the questionnaire, no prohibitions are used. It is chosen to use a randomized design for 9 out of 10 choice sets. Randomized designs are designs where every respondent will see a unique series of choice sets, which is set to 200 unique designs. One choice set is fixed, meaning that the design of this choice set will be the same in each questionnaire. It is used as a hold out choice task, which in turn will be used to calculate the hit rate. It is recommended to have a hold out set for calculating the hit rate, in order to assess model reliability (Hair et al, 2010). The hold out set in this research is designed manually, without the design option of Sawtooth. The hold out choice set is designed in a way that the clearest differences between attributes are shown. Also, when segmentation is done, the segments that are expected to be created are reflected in one of the three choice options in the hold out set.

The design of the attributes of the conjoint analysis is tested in Sawtooth SSI web for efficiency. The table with results of the test can be found in appendix 3. The report indicates that 200 versions of the CBC questionnaire were created. Each version of the questionnaire had 9 random choice tasks, so a total of 200 x 9 = 1800 choice sets were generated. The relative efficiency of the design is measured by a number, where zero indicates the worst efficiency and one the best. The efficiency numbers for the conjoint design of this research are all one, or very close to one, meaning that this choice questionnaire has an efficient design

(34)

33

Table 3.4.1.1 Attribute levels

Attribute Levels Preference function Company/Cause relationship + CSR highlight 1) No CSR 2) CSR, individually highlighted, social alliance 3) CSR, collectively highlighted, social alliance 4) CSR, individually highlighted, cause related marketing 5) CSR, collectively highlighted, cause related marketing Part worth Type of product characteristic 1) Hedonic 2) Utilitarian Part worth Price 1) € 1.29,- 2) € 1.49,- 3) € 1.69,-

Part worth or linear vector

3.4.1.2 Conjoint Analysis latent class

After Conjoint Analysis on aggregate level, Latent Class Analysis will be performed and in this research will be used to predict class membership based on CSR responsiveness in order to make consumer segments. Latent class analysis performs a probabilistic clustering approach, meaning that although each respondent is assumed to belong to one class, it does take into account that it is not completely certain that the respondent does belong to that class. This way of creating segments is more realistic. Also, latent class analysis instead of cluster analysis is chosen to make consumer segments, because unlike cluster analysis, it does contain measures of model fit.

3.4.2 Multinomial Logit

(35)

34 predicts the probability that a respondent is member to a class. A separate analysis, by means of Multinomial Logit in SPSS, will be used to link class membership of respondents to the demographic variables and the social value variable.

3.4.3 Ordered multinomial logit

In order to measure the purchase intentions of consumers for a given product, the method ordered multinomial logit will be used. The moderators company/cause relationship, charity highlight and product characteristic will be measured, the price is left out. The purchase intentions of consumers are measured on a three point Likert scale, from “not likely to buy” till “likely to buy”. As these options are discrete, instead of continuous variables, a logit model should be used. In the questionnaire, respondents face more than two options to choose from, so in order to find out the probability of the purchase intention, a multinomial logit model should be used (Franses and Paap, 2001). However, the 3 point Likert scale used in this questionnaire follows a discrete ordered pattern. This means that respondents face a ranked variable, for which an ordered multinomial logit model is suitable (Franses and Paap, 2001).

3.5 Pre-test

Before the actual research, a pre-test has been conducted to test the design of the questionnaire. A sample of 10 respondents has been used to find out whether the questionnaire was well-designed, understandable and whether the designed scales are appropriate.

Feedback from respondents showed that the differences between the levels of the attributes were too small and that the text in the choice options was too large. Brand names for the product have therefore been added to the questionnaire, to make it easier for respondents to quickly recognize the type of product. Moreover, the sentences describing the companies have been made shorter.

The number of choice options per choice set has been found to be the right number according to the respondents of the pre-test, since three options were easy, but still realistic to choose from. The pre-test has also been conducted to find out whether the none option had not been chosen too often. The none option in the choice sets has been perceived as pleasant and realistic, however the none option was not chosen at all by the respondents.

(36)

35 willingness to help others. The scale with 7 items firstly generated a Cronbach’s Alpha of 0.557, which indicates a questionable reliability of the scale. An additional Cronbach’s alpha table showed that when two items were deleted, a Cronbach’s alpha of 0.767 was generated, which is acceptable. These two items are:

6. “I do not think it is necessary to do something to help future generations”. 7. “I think that I should do as much as needed to help less fortunate people”.

(37)

36

4. Analyses and results

This chapter describes the results of the analyses that will be performed in order to test the hypotheses. The chapter starts with descriptive statistics of the sample, and then begins with an aggregate conjoint analysis, followed by a latent class analysis testing the relationship of the moderators on willingness to pay. The classes are then described with use of multinomial logit. The chapter finalizes with an ordered multinomial logit testing the role of the moderators on the purchase intentions.

4.1 Descriptive statistics

This section describes the descriptive statistics of the sample and will test scale validity of the scales used.

4.1.1 Descriptive statistics Sample

After having the questionnaire online for 1.5 weeks, in total 407 respondents participated in the questionnaire. From this, 166 respondents did start the questionnaire, but did not complete it; therefore these respondents were left out the sample. As quite a large amount of respondents quit the questionnaire early, some research has been done into when they quit. Table 4.1.1.1 shows where in the process of the questionnaire the respondents quit. Apparently 88% of the respondents quit very early in the questionnaire, before finishing the conjoint part. All the 166 respondents should indeed be left out, since the questionnaire should be fully completed, in order to be useful for this research.

Table 4.1.1.1 Incomplete questionnaires

Part until respondent quit Number of respondents Percentage

Nothing 95 57%

Halfway or completed Conjoint 51 31%

Halfway or completed Purchase Intentions

8 5%

Halfway or completed Social Values

4 2%

Halfway Demographics 8 5%

(38)

37 has been used to find participants for the questionnaire. The initial group of respondents was already young, and these young respondents forwarded the questionnaire to their friends who are also young. The same yields for level of education, 86.3% has a HBO or University education level. Also a quite high percentage (61.5%) lives in the northern provinces of the Netherlands: Friesland, Groningen and Drenthe. This is due to the fact that the researcher lives in the northern part of the Netherlands.

Table 4.1.1.2 Descriptive statistics sample

Demographic Category Percentage

Gender Male Female 38.6% 61.4% Age 18-25 26-30 31-40 41-65 > 65 48,9% 25,3% 9.6% 15.8% 0.4%

Gros annual income <€20.000

€20.000-€40.000 €40.000-€60.000 > €60.000 53.1% 32.5% 10% 5.4%

Education Elementary school VMBO/MAVO MBO HAVO/VWO HBO/University 0.4% 1.7% 7.1% 4.6% 86.3%

Household structure Single

Living together without children Living together with children One parent family

58.1% 27.4% 13.7% 0.8%

Living rural or urban Urban Rural

80.1% 19.9%

Life stage Highschool student Student

Working/looking for job Retired

1.2% 44% 53.5% 1.2%

Source of income Mainly salary Mainly profit Mainly pension

Mainly Entitlement/ student loan

(39)

38 Noord-Holland Zuid- Holland Zeeland Noord-Brabant Limburg 7.5% 7.1% 0.4% 0.8% 0%

4.1.2 Descriptive statistics Conjoint Analysis

Table 4.1.2 shows per choice option relatively how many times it was chosen among all respondents. Choice option 1, 2 and 3 are about equally chosen, which makes sense, as the content of these options changes per choice set. Choice option 4 is the none option, and is chosen 10.93% by respondents. The none option has been chosen more than five times in the questionnaire by 8.8% of the respondents. 68.9% of the respondents never chose the none option in the choice part of the questionnaire.

Table 4.1.2 Descriptive statistics

Choice option Number chosen Percent chosen

1 602 27.75%

2 669 30.84%

3 661 30.47%

4 (none option) 237 10.93%

4.2 Scale validity

The 6 items measuring the level of altruism, that is how concerned of others’ welfare respondents are, are tested on internal consistency. The 6 items are recoded into the same scales, as question 2 and question 6 are reversed. A reliability analysis has been performed, giving a Cronbach’s alpha. The Cronbach’s alpha for the 6 items is 0.696, which is an acceptable level of consistency. The Cronbach’s alpha output “When item would be deleted” shows whether the Cronbach’s alpha would become higher when an item would be deleted (appendix 5). This table did not show a notably higher Cronbach’s alpha then the alpha for 6 items. Therefore the measurement of altruism will keep 6 items to measure one value. The mean of the 6 items has been used to measure how social people are. The mean of socialness on a scale of 1 to 7 is 3.93, with a standard deviation of 0.53.

4.3 Aggregate solution CBC latent class method

(40)

39

4.3.1 Method

Sawtooth Latent Class has been used to estimate a 1 class model. All attributes including the none option and prices are added. Firstly all attributes including price are assumed to be part worth. Effects coding has been used to estimate the parameters for the attributes, as this is the way Sawtooth Latent Class set the coding settings. Also interactions have been added to the model, to test the moderating effect of the moderators on the willingness to pay of respondents. This means that between the attributes product type and CSR activity interaction terms have been created, which results in 10 interaction terms.

Linear coding of the attribute price is preferred above part worth coding, since a linear model has only one derivative, which makes it less complicated to determine the price of the product for every utility level (Hair et al, 2010). It is therefore checked whether the part worth utilities of price show a linear pattern. A model has been created and the part worth utilities for price are displayed in figure 4.3.1.1. As can be seen, the attribute price shows a linear pattern, which gives already an indication of linearity of the price.

Figure 4.3.1.1 Price part worth

Also the model fit criteria have been compared for both a model with price part worth coded and a model with price linear coded. As can be seen in table 4.3.1.2 the model does not improve a lot when part worth coding is used. The percent certainty of the part worth price model is only a bit higher, as well as the CAIC and the chi square. The relative Chi square, which takes the number of parameters into account, shows a higher value for the linear price model, than for the part worth model. It is chosen to set the preference function for price to linear. This since the values of percent certainty, CAIC and chi square are only slightly lower for the linear price model and since the relative chi square for the linear price model is higher. Moreover, linear coding is preferred above part worth coding.

Table 4.3.1.2 Comparison partworth and linear

Charity Highlight and Product Characteristics in

5/24/2012

Sjoukje Bakker

BE GOOD,

D

O GOOD

,

GET GOOD:

The Influence of Company/Cause Relationship,

Charity Highlight and Product Characteristics in

CSR on Willingness to Pay and Purchase

Master Thesis

By

Sjoukje Bakker

University of Groningen

Faculty of Economics and Business

MSc Business Administration – Marketing Management and Marketing Research

Abstract

Preface

Contents

1. Introduction

2. Literature review

3. Research Design

4. Analyses and results