THE DISNEY THEMED BUNDLE

(1)

THE DISNEY THEMED BUNDLE

The influence of using a Disney themed bundle in nudging

children in soccer canteens towards healthier eating

by

ESTHER KOOISTRA

(2)

THE DISNEY THEMED BUNDLE

The influence of using a Disney themed bundle in nudging

children in soccer canteens towards healthier eating

by

ESTHER KOOISTRA

University of Groningen Faculty of Economics and Business

MSc Marketing Management Master Thesis

Completion date: 14-06-2016

Esther Kooistra Jan van Galenstraat 4a

9726 HM Groningen 06 - 46783974

esther--kooistra@hotmail.com Student number: S2380218

(3)

ABSTRACT

For this study, the researchers conducted, commissioned by the KNVB and conducted in collaboration with Disney, a field experiment in a soccer canteen in the Netherlands. The aim of the research was to study how the use of a Disney themed bundle, containing a healthier drink and snack, could nudge children in soccer canteens towards healthier eating. This is done through analyzing the influence of children’s attitudes towards the Disney themed bundle on the purchase intention for the bundle. Furthermore, the moderating effect of involvement in Disney is studied. The results of this research show that the attitudes of children towards the bundle and towards popcorn positively influence the purchase intention for the bundle. Moreover, even though we did not find proof for the moderating effect of involvement in Disney, we did find a significant positive and direct effect of involvement in Disney on the purchase intention for the bundle. Lastly, we found a significant decrease in the sales of crisps, and a significant increase in the sales of DubbelFrisss Free.

(4)

EXECUTIVE SUMMARY

Even though an increasing number of people is becoming more aware of the healthiness of the products they are consuming, obesity and unhealthy food consumption are still worsening worldwide health problems. Since obesity is especially a worsening problem among children, we will particularly focus on young children aged four to fourteen years. Additionally because most children make decisions unconsciously, which enables us to change children’s eating environment in order to nudge them towards healthier eating. The increase in obesity patients and unhealthy food choices among children might be explained by the fact that health is not only determined by individual choice. Other changes have to be taken into account when trying to improve human health, such as changing environments of places where drinks and food can be bought. After discussing multiple classes of environmental cues affecting food choices of children, such as the food environment, social marketing, menu strategies, social influence, and atmospherics, we decided to focus on the 'attractiveness’ aspect of the CAN approach. To increase the attractiveness of healthy foods, we studied the effect of using a Disney themed bundle in the promotion of healthier foods, including DubbelFrisss Free, JIMMY’s popcorn, and an incentive in the form of a key ring. The children were not aware of the fact that the snack and the drink in the bundle were healthier alternatives. The research was commissioned by the KNVB, and conducted in collaboration with Disney.

(5)

promotion of healthy foods influence the purchase intention for the Disney themed bundle among children in soccer canteens?” and “Does involvement in Disney moderate the effect of the attitude of children towards the Disney themed bundle in the promotion of healthy foods on the purchase intention for the Disney themed bundle foods among children in soccer canteens?” In order to answer these research questions, a field research was conducted in the soccer canteen of VV Heino. Soccer canteens have the capacity to promote healthy lifestyles through both physical activity and offering healthy food, and they present messages to children regarding acceptable food choices. During our experiment, we conducted surveys with 95 children aged four to fourteen years in the canteen of VV Heino, and we counted the sales numbers of the products sold in this canteen. By conducting surveys, we were able to discover background information, opinions and attitudes of our participants. Through comparing the sales numbers, we were able to study the change in actual purchases of healthy and unhealthy products in the soccer canteen.

(6)

PREFACE

This thesis is the final part of my Marketing Management master’s studies at the University of Groningen. I have always been interested in healthy food choices, particularly in the motivations of consumers whether to choose for unhealthy- or healthy products. Therefore, I was excited when I got the chance to conduct a research in this topic. Because of the practical process, conducting this research and writing this thesis was (most of the times) fun and interesting, and I have learned a lot in the past few months.

I would like to express my gratitude to my supervisor Prof. Dr. Ir. Koert van Ittersum, Professor of Marketing and Consumer Well-Being, who provided me with useful guidance throughout the course of this research, and who gave my fellow researcher and me the opportunity to conduct a research commissioned by the KNVB. I would also like to thank the KNVB for giving us this great opportunity, and Disney, for their collaboration in this experiment. Lastly, I want to thank my thesis group, family, and friends who supported me during the process of writing this thesis.

(7)

1. INTRODUCTION

Nowadays, trends concerning the healthiness of people and their food consumption can be viewed from two sides. On the one hand, an increasing number of people is becoming more health conscious and more aware of the healthiness of the products they are consuming (CBL, n.d.; Nielsen, 2015). This health-consciousness can be noticed in the awareness of people of whether they are overweight: 49% of respondents worldwide believe that they are overweight, and a similar percentage (50%) is actively trying to lose weight (Nielsen, 2015). On the other hand, obesity and unhealthy food consumption are worsening worldwide public health problems (McFerran et al., 2010; Nourissat, Ciais, & Coudane, 2015; Quick et al., 2015).

1.1 Children

(9)

awareness concerning this issue. The main factors related to childhood obesity are poor eating behaviors (increase in energy intake) and lack of physical activity (Deckelbaum & Williams, 2001; Ebbeling, Pawlak, & Ludwig, 2002; Hall, Chai, & Albrecht, 2016; Kelly et al., 2008). In this research, we focus on improving eating behavior among children. Children generally make decisions unconsciously and are unaware of why they choose the foods they choose (Wansink, 2013). Therefore, we are able to change children’s eating environments so their quick, instinctive actions are biased towards healthier products. In previous literature, many influences on the healthiness of consumers´ choices in different contexts have been studied. These will be discussed more extensively in the literature review of this thesis.

1.2 Soccer Canteens

(10)

healthy lifestyles are promoted through sports, unhealthy food choices are being provided at sports canteens. Also, even though soccer canteens seem like a suitable opportunity to combine sports and healthy eating, it is for many of these canteens not obvious to offer healthy food. However, it is crucial that the seriousness of childhood obesity and unhealthy consumption is also recognized in local contexts, such as sports clubs, since starting small may encourage greater contexts, and activity there is mainly done on voluntary basis (Skille, 2008). Because of that, it is critical to make volunteers of soccer clubs aware of the fact that selling more healthy foods is socially responsible and that it does not have to be less profitable. However, although many children participate in sports, surprisingly little can be found in existing literature about food environments in sports settings. Therefore, we will focus on children in this context.

1.3 KNVB

In collaboration with the “Koninklijke Nederlandse Voetbalbond” (KNVB), food choices of children in a soccer canteen are studied. The KNVB is the biggest sports federation in the Netherlands with over 1.2 million members and more than 500 employees (KNVB, n.d.). Because of this large membership, the KNVB feels partly responsible for a sporty and healthy society and wants to use the power of soccer to help resolve social problems such as obesity (KNVB, n.d.). Their corporate social responsibility ambition is supported by two pillars: “KNVB Samen” and “KNVB Fit”. We focus on the last pillar since it is related to a healthy way of life through helping people to acquire healthier nutrition and exercise patterns.

1.4 Disney

(11)

Living Brochure, n.d.). Disney discontinued the use of its name and licensed characters on packages for products that do not meet their criteria for healthier food (Sharma, Teret, & Brownell, 2010). The Disney themed bundle is included since solely promoting healthy foods may not be effective as consumers have the impression that healthy foods will taste worse compared to relatively unhealthy foods (Berning, Chouinard, & McCluskey, 2010; Chandon, & Wansink, 2012; Horgen & Brownell, 2002; Raghunathan, Walker Naylor, & Hoyer, 2006). The use of a Disney themed bundle might make the healthy alternatives more appealing to children, without them knowing it is healthier. Little is known about the influence of Disney characters on promotional material of certain healthy foods. Therefore, our results can be used by sports clubs to increase the number of healthy foods sold in their canteens.

1.5 Summary

The problem focused on in this research is that children in soccer canteens eat too unhealthy, and the overall main problem is the increase in childhood obesity. The main contribution is to increase the healthiness of the choices children make in soccer canteens. We expect that if we can demonstrate that the use of a Disney themed bundle can nudge children in soccer canteens towards healthier eating, this may lead to an incentive for initiatives in canteens of other sports clubs, schools etcetera. In this way, this research can eventually help to increase the overall level of consumer well-being and to decrease the number of childhood obesity patients.

1.6 Research Questions

(12)

the attitudes of the children who did buy the bundle, and the children who did not buy the bundle. Lastly, related to the collaboration with Disney, we are interested in the moderating effect of involvement in Disney. We formulated the following research questions:

- “How does the attitude of children towards the Disney themed bundle in the promotion of healthy foods influence the purchase intention for the Disney themed bundle among children in soccer canteens?”

- “Does involvement in Disney moderate the effect of the attitude of children towards the Disney themed bundle in the promotion of healthy foods on the purchase intention for the Disney themed bundle foods among children in soccer canteens?”

1.7 Thesis Outline

The remainder of this thesis is structured as follows: The next section reviews relevant literature concerning childhood obesity, healthy eating behavior, individual factors, and environmental cues. It also includes the hypotheses and the conceptual model. Section three is the methodology; in this section the data collection method and study design, the participants, and the data analysis method will be described. Section four shows the results of this research, followed by the discussion in the fifth section, and limitations and recommendations for future research in the sixth section. Finally, section seven consists of the conclusion.

2. LITERATURE REVIEW

(13)

articles which will be discussed in this section can be found in the table in ‘Appendix 1’. The table presents the context, environmental cue(s), and study design of each relevant article.

2.1 Healthy Eating Behavior

Concerning healthy foods, the reduction of saturated fat and energy intake, and the promotion of fruit, vegetable and fiber intake is important (Brug et al., 2008). It is known that fruit and vegetables are essential components of healthy diets (Hoffman et al., 2010; Tapper, Horne, & Fergus, 2003). They contain high levels of vitamins, minerals, fiber and antioxidants (Tapper, Horne, & Fergus, 2003), which are important components of healthy foods. However, according to Striegel-Moore et al. (2006), over 95% of girls failed to meet the Healthy People 2010 recommendations for daily fruit and vegetable intake. Therefore, it is important to be able to change the environment in such a way that the consumption of fruit, vegetables, and other relatively healthy foods among children increases. The traditional approach to increase the intake of healthy foods by children is through health education campaigns in which children and parents are informed of what they should and shouldn’t be eating (Tapper, Horne, & Fergus, 2003). However, this approach had limited success in the past since ‘knowing’ unfortunately often does not mean ‘doing’.

(14)

of choices consumers make, there are also multiple individual and environmental factors which can be influential in smaller contexts. These will be discussed in the following sections.

2.2 Individual Factors

Ways to prevent obesity and weight gain already have received much attention in previous studies (Sobal & Wansink, 2007). It has been proposed that the health behavior of individuals, and thus children, is determined by three categories: Motivations, abilities and opportunities (Brug et al., 2008). The first two categories, motivations and abilities, are individual factors. An individual is most likely to engage in health behavior when they are motivated to make healthy choices, and have the ability to engage in healthy behavior (Brug et al., 2008).

Previous literature discusses multiple individual factors which influence eating behavior among children. Individual factors which have a particularly negative influence on childhood obesity are low socioeconomic status (SES; a measure of income, education and employment) (Hall, Chai, & Albert, 2016), consumption of sugar-sweetened, carbonated drinks (James et al., 2004; Ludwig, Peterson, & Gortmaker, 2001), unfortunate generic-, and perinatal factors (Ebbeling, Pawlak, & Ludwig, 2002), lack of physical activity and excessive inactivity (particularly television viewing) (Ebbeling, Pawlak, & Ludwig, 2002; Warburton, Nicol, & Bredin, 2006), emotional eating (Karlsson et al., 2000), body dissatisfaction (Stice, 2002), stress, depression and boredom (Ganley, 1989), and family factors (paternal obesity), demographics, lifestyle in early childhood, and maternal smoking during pregnancy (Ebbeling, Pawlak, & Ludwig, 2002; Reilly et al., 2005), and lack of sleep (Quick et al., 2015; Reilly et al., 2005). Lack of sleep has a negative influence on childhood obesity because of the growth hormone secretion, or because sleep reduces the child´s exposure to factors in the environment which promote obesity, such as late night food consumption (Reilly et al., 2005),

(15)

research to date, including hunger and satiety, sex, age, body weight, impulsivity, and eating goals. From the literature, it appeared that little evidence has been found which supports the idea that these factors influence modeling. It was only found that women are more vulnerable to modeling effects on food intake (Cruwys et al., 2014). However, Cruwys et al. (2014) state that individual factors do not appear to be critical in explaining effects on eating behavior of consumers; social influence is a primary determinant.

2.3 Environmental Cues

An increasing number of people nowadays want to express healthy preferences, however there are many social and environmental barriers that have to be overcome (Hawkes et al., 2015). Evidence shows that the environments in which people develop their dietary behavior and make choices concerning food have a significant influence on their purchasing and consumption behavior (World Health Organization, 2015). As stated before, especially children are not aware of why they choose the foods they choose (Wansink, 2013). Therefore, they are very sensitive to their eating environments. Unfortunate environmental factors are seen as an important cause of the increase in childhood obesity (Ebbeling, Pawlak, & Ludwig, 2002).

(16)

In order to provide practioners in architecture and public health with a set of strategies to make canteen environments more suitable for healthy eating, Huang et al. (2013) developed a new tool, called “Healthy Eating Design Guidelines for School Architecture”. In these guidelines, multiple ways in which the environment may influence food choices in canteens are incorporated. The following subsections of this chapter will discuss multiple classes of environmental cues more thoroughly.

2.3.1 Food Environment

(17)

plates, bowls, and spoons as an indication of how much should be served and consumed (Wansink, 2004). In their experiment which especially focused on the difference between adults and children, Zampollo et al. (2011) found that children prefer plates with more elements, colours, and with entrees placed in the front of the plate. To improve nutritional choices, healthy defaults can also be incorporated. Peters et al. (2016) tested this phenomenon in Walt Disney World restaurants, and found that healthy defaults can shift food and beverage selection patterns of children toward healthier options. This is because the choice for an unhealthy default option became an opt-out decision, which makes the threshold for children to explicitly ask for an unhealthy default option high.

2.3.1.1 Pricing

(18)

2.3.2 Social Marketing

The use of marketing in promoting healthy foods can be seen as “social marketing”, which is the use of marketing to design and implement programs to promote socially positive behavior change (Grier & Bryant, 2005). Marketing of foods is important for children since food promotions are among their favorite types of promotion (Hastings et al., 2006). Children are interested in trying promoted foods which are, unfortunately, often unhealthy foods (Hastings et al., 2006). Food promotion influences children’s knowledge, food preferences, consumption and other diet-related behaviors and outcomes (Hastings et al., 2006). Although childhood obesity is caused by many factors, the marketing actions of food producers, stores, and restaurants can be seen as one of the key causes (Dubé et al., 2010). In this section, we will first discuss the use of games and sponsorship as examples of social marketing. Thereafter, the two forms of social marketing used in this research, branding and bundles, will be discussed.

2.3.2.1 Games

(19)

2.3.2.2 Sponsorship

A significant marketing tool related to the food choices made by children in soccer canteens is sponsorship (Kelly et al., 2012). For sports clubs in particular, sponsorship has influence on the members and their food choices (Kelly et al., 2010b; Maher et al., 2006). Therefore, it can be suggested that having sponsors who are linked to unhealthy foods and soft drinks may contribute to the increase in childhood obesity (Danylchuk & MacIntosh, 2009). Respondents in the research of Kelly et al. (2012) stated that they were most supportive of restricting the use of unhealthy food logos on children’s uniforms since, according to them, this had the biggest impact on unhealthy food choices made by children. Kelly et al. (2011) found that children have a high recall of food and beverage company sport sponsors, and that many children have positive attitudes towards these sponsors. They even found that children perceive sponsors as being kind, generous and cool. This may be a cause for concern since it may increase children’s preferences and consumption for the food the sponsor sells. Thus, sponsorship of children’s sports by unhealthy food companies may contribute to unhealthy food choices since sponsorship influences brand awareness and attitudes (Kelly et al., 2010b). In this way, sponsorship can also be used to promote healthy foods. Hence, sports clubs need to carefully choose their sponsors in order to increase healthy, and reduce unhealthy, eating behaviors.

2.3.2.3 Branding

(20)

study that children almost doubled their apple choice if a sticker with a familiar character, like Elmo, had been put on the fruit. Additionally, Robinson et al. (2007) showed that branding of foods and beverages influences the taste perceptions of young children. Zampollo et al. (2011) found that children prefer figurative designs, to causal designs, through arranging bacon as a smile, peas in a heart shape, and cupcakes with images. The study of Wansink, Shimizu, and Camps (2012) shows the positive effects of the use of identity in marketing by which children are encouraged to eat more healthy foods. They find that children can be primed to purchase healthier foods merely by thinking about what their favorite superhero, their role model, would eat: Parents asking their children “What would Batman eat?” In the research of Hoffman et al. (2010) they used cafeteria posters reflecting the fruit and vegetable of the day to increase the fruit and vegetable intake of children. The posters included cartoon characters, which were intended to serve as symbolic role models to students and to capture their attention while they were standing in the cafeteria line. Through the use of Disney characters on our bundle, we also aim to capture the attention of children in the soccer canteens.

2.3.2.4 Bundle

(21)

certain behavioral outcomes. Their results reveal the relationship between drink context and food consumption; soft drinks combined with calorie dense foods is seen as favorably, while soft drinks combined with vegetables is not. Additionally they found that vegetables are more popular among children when the drink consumed is water. Venkatesh and Kamakura (2003) also studied the effect of the use of a meal bundle, and discovered that food and drinks are valued more highly when offered in a bundle.

2.3.3 Menu Strategies

(22)

Hoefkens et al. (2012) showed the effectiveness of simplified point-of-purchase (POP) nutrition information on menu choices in canteens, especially for more motivated students and for those with a greater objective nutrition knowledge. Skov et al. (2012) found during their literature review that health or nutrient information at POP indeed showed some evidence of effect on food choice, however this is not consistent on what type of foods.

2.3.4 Social Influences

(23)

to imitate another person when the child likes that person (Bandura, 1977), and when the person has the same age or is slightly older (Brody & Stoneman, 1981). Barthomeuf, Rousset, and Droit-Volet (2008) found that the emotions expressed on other people’s faces affect the desire to eat certain foods; when the other eater expressed disgust, the desire of the participant to eat the particular food decreased. Furthermore, when the other expressed pleasure while eating disliked food, the desire of the participant to eat the disliked food increased. These results show the strength of the social context while eating.

2.3.5 Atmospherics

(24)

them to enjoy and to stay longer (Stroebele & De Castro, 2004). The effect of lighting is suggested to be particularly strong when dining with others (Wansink, 2004). Ambient odors can influence eating behavior through taste enhancement or suppression (Razin, 1982). Unpleasant smells are likely to decrease food intake and shorten the eating duration, perhaps by speeding satiation (Chandon & Wansink, 2010).

2.3.5.1 Distractions

The previous mentioned findings concerning atmospherics highlight the role of distractions on children’s food choices, taste perceptions, and consumption volume (Ogden et al., 2013; Wansink, 2004). Ogden et al. (2013) studied the effects of distraction in the form of driving, television watching and interaction with others. Their results show that people who are watching television consumed more than the ones driving or having social interaction. Additionally, distraction may disrupt the link between food intake and subsequent reductions in hunger. Therefore, consumers will still feel hungry and increase their consumption volume as a result of the distraction. This is called “mindless eating”: Ways in which environmental factors trigger eating without the consumer consciously processing it (Sobal & Wansink, 2007; Wansink, 2004). Television watching has not only been linked to childhood obesity because of distractions, but also because it decreases physical activity (Robinson, 2001; Thomson et al., 2008) and increases snacking behavior (Thomson et al. 2008).

2.4 The CAN Approach

(25)

(comparatively or absolutely), and more normal (perceived or actual)” (Wansink, 2015, p. 486). According to Wansink (2013), increasing the convenience, attractiveness and normativeness of healthy products is more effective for reducing childhood obesity than removing unhealthy products from the assortment. This is because when children really love an unhealthy product, removing that option usually backfires. It might prevent children from eating it today, but on another day children will increase their intake of this product (Wansink, 2013).

The previous discussed concept “nudging” is the behavioral science principle which is used in the CAN approach. Nudging could include ways to alter social or physical environments to make certain behaviors – in this research increasing healthy food choices among children – more likely. Examples of nudging are providing information about what others are doing (social norm feedback), altering the layout of buildings, changing ways in which food and drinks are served and presented (Thaler & Sunstein, 2008), or manipulating the assortment structure, food order in buffet line (Wansink & Hanks, 2013), shelf layout, and to place healthy food near the cash register of a canteen (Van Kleef, Otten, & Van Trijp, 2012). Van Kleef et al. (2012) discovered that increasing the prominence of healthy snacks by enlarging their availability (75% healthy versus 25% unhealthy products) increases the probability of a healthier food choice. This is an example of increasing the normativeness of the products.

(26)

2013), or by using a “Last chance for fruit” sign displayed next to the fruit bowl at the cash register (Hanks, Just, & Wansink, 2013).

Multiple studies at homes, grocery stores, restaurants, and schools have shown that the CAN approach is effective in increasing the healthiness of products chosen by parents, shoppers, restaurant goers, and students (Wansink, 2015). The advantage of this approach is that the products themselves don’t have to be changed; only their setting. When the CAN approach is used creatively and effectively, it can eventually lead to habitually healthier choices (Wansink, 2015), which is crucial in the improvement of overall food choices of children, their well-being, and the decrease in childhood obesity.

(27)

& Smith (2013), is by offering pre-sliced fruit. In their research in school cafeterias they discovered that sliced fruit is more appealing to children than whole fruit since it is easier and tidier to eat.

Hence, the way in which products are presented is crucial in influencing the choices children make in canteens. Through using the Disney themed bundle, we particularly focus on the ‘attractiveness’ aspect of the CAN approach. While children are doubting about what snack to choose in the soccer canteen after playing or watching a soccer match, the use of a Disney themed bundle may cause the attention of these children to be captured (Hoffman et al., 2010), and therefore may nudge them towards healthier eating.

2.4 Hypotheses

2.4.1 Disney Themed Bundle

(28)

As stated before, children are sensitive to food-related environmental factors in determining the attractiveness of a product. In the marketing of foods to young children often positive and colorful environments and packages with appealing characters are used since children find these more attractive (Johnson et al., 2007). Especially branding of foods through characters can improve the attractiveness of healthier food to children (Wansink, Just, & Payne, 2012). In this research we study the effect of using a colorful bundle with Disney characters (see ‘Appendix 2’), including an incentive, in nudging children towards healthier eating behavior. Particularly, we focus on the influence of the attitude of children towards the bundle and its contents. Characters have been used to market products for already more than hundred years (Garretson & Niedrich, 2004). The type of character we use in our research, Donald Duck, is a “celebrity spokes-character” (or “cartoon media character”): Originally created for other purposes such as animated movies, cartoon programs, and/or comic strips and subsequently licensed by brands or companies to appear in promotions (Garretson & Niedrich, 2004). Overall, Disney characters are seen as “licensed characters” (Chandon & Wansink, 2010), i.e. animated characters from television and movies, and “fictional characters” (Hoorn & Konijn, 2003). The role of fictional characters in performing educational tasks is growing since it appeared to be effective (Hoorn & Konijn, 2003).

(29)

(1996) have shown that the presence of characters on packaging can be effective in creating positive attitudes towards products and in increasing sales, especially when the character is well established, recognizable and familiar. When children recognize the characters which are associated to certain products, they are more likely to choose those products (Batada & Borzekowski, 2008). Moreover, familiarity is an important factor contributing to the ease of processing (Fennis & Stroebe, 2016). This relates to the hedonic fluency model which assumes that a higher ease of processing is perceived as pleasant and that this positive affect will be used as information in the evaluation of the stimulus (Fennis & Stroebe, 2010). The character we use, Donald Duck, is a familiar character for most children and therefore will be recognized easily. Through the familiarity of this character, it can have a positive influence on the purchase intention for the bundle since past research showed that familiar characters are preferred to unfamiliar ones (De Droog, Buijzen, & Valkenburg, 2012; Kotler, Schiffman, & Hanson, 2012; Roberto et al., 2010; Smits & Vandebosch, 2012). Familiar media characters appear to have a stronger influence on children’s preference, choices and intake of less healthy foods compared to healthy foods (Kraak & Story, 2014). However, since we do not indicate that the food and drink included in the bundle are healthier alternatives, this finding should not cause any problems. Ülger (2008) showed in his research that packages with Walt Disney characters play an effective role in the food preferences of preschoolers. The chocolate wafer with the packaging branded with Walt Disney characters was chosen 74% of the times compared to the chocolate wafer which was advertised on the television (26%).

(30)

perceptions. They let children taste five pairs of identical foods and beverages in packaging from McDonald’s and compared this with matched, unbranded packaging. They concluded that marketing should be regulated to young children, and that branding may be a useful strategy for improving the eating behaviors of young children. Furthermore, Roberto et al. (2010) found that children preferred the taste of foods that had branded packages with popular cartoon characters on it, instead of the same foods without characters on their packaging. However, this effect occurred most strongly for energy-dense, nutrient-poor foods. Therefore, our research will try to discover whether this does works for healthy foods in soccer canteens. De Droog, Valkenburg, and Buijzen (2010) even found that characters can increase children’s liking of and purchase intentions for fruit up to a level similar as candy.

(31)

(32)

- H1: The attitude of children towards the Disney themed bundle in the promotion of

healthy foods will positively influence the purchase intention for the Disney themed bundle among children in soccer canteens.

- H2a: The promotion of healthy foods through the Disney themed bundle will positively

influence the purchase of healthy foods.

- H2b: The promotion of healthy foods through the Disney themed bundle will negatively

influence the purchase of unhealthy foods. 2.4.2 Involvement in and Liking of Disney

Since this research is conducted in collaboration with Disney, the influence of involvement in Disney is interesting to investigate for both the researchers and Disney. We aim to discover whether and how involvement in Disney moderates the relationship between attitude towards the Disney themed bundle and purchase intentions for the bundle. What is known about involvement in this context will be discussed in this section.

Involvement is often considered by researchers to understand the effects of promotion and advertising on consumer behavior (Muehling, Laczniak, & Andrews, 1993). However, involvement as a construct is hard to define since it can be related to many different labels, such as product classes, messages, advertising, and decision-making (Muehling, Laczniak, & Andrews, 1993). For this research, we use the definition provided by Antil (1984, p. 204), who defines involvement as “the level of perceived personal importance and/or interest evoked by a stimulus (or stimuli) within a specific situation”.

(33)

a certain attitude towards the Disney themed bundle. More attention to the Disney themed bundle might lead to a higher purchase intention. Furthermore, involvement is a process that includes an emotional and a cognitive experience (Perse & Rubin, 1990). Emotional experience entails the feelings of an individual towards the stimulus, and cognitive experience means that one will have attention for the stimulus, make recognition cues based on the familiarity of the stimulus, and will elaborate on the stimulus (Perse & Rubin, 1990). Zinkhan and Muderrisoglu (1985) also found that there is a relationship between involvement and familiarity. Like discussed previously, Donald Duck is for most children a familiar character and this may in turn influence the involvement of children with the character and with Disney in general.

(34)

(35)

encourage the child to buy the Disney themed bundle with healthier alternatives when the child likes Disney or particularly Donald Duck.

Neeley and Schumann (2004) use the hierarchy of effects model to divide the influence of characters in stages. This model contains the following six stages of effects, including attention, recognition, association, liking, preference, and intention and choice. This means that after a child has attention for the promotion, recognizes the character, creates certain associations, and determines its level of liking for Donald Duck, this can lead to product preference, intention, and choice (Neeley & Schumann, 2004). Therefore, a higher level of involvement in Disney, and thus liking, might strengthen the relationship between attitude towards the bundle and purchase intention for the bundle. Furthermore, Neeley and Schumann (2004) found that character recognition and liking leads to product recognition and liking. This product liking can subsequently lead to an increase in purchase intention.

Children get to know and learn about characters through social contacts, media and marketing environments including their parents, television, movies, the Internet, and packaging of foods (Richert, Robb, & Smith, 2011). Thereafter they develop a certain level of liking for and “parasocial relationships” with their favorite characters which refers to emotionally infused friendships based on the attractiveness of characters and the message they convey (Bond & Calvert, 2014). This message can subsequently influence their eating behavior (Bond & Calvert, 2014). Since children are exposed to celebrity spokes-characters often through television, social media and movies, liking for specific characters will increase through mere repeated exposure. By mere exposure is meant “a condition which just makes the given stimulus accessible to the individual's perception” (Zajonc, 1968, p. 1). The more children are exposed to certain stimuli, the more they will like it.

(36)

argues that if cognitive responses, here in the form of involvement, are positive, a positive attitude or purchase intention is expected. Nelson, Duncan, and Frontczak (1985) state that involvement tends to improve the attitude of individuals towards advertisements and increase the probability that the advertised product will be purchased. This might also be the case when a certain promotion is used and individuals’ involvement for the particular promotion is high. Acuff and Reiher (1997) already stated earlier that the beliefs and desires of children concerning a product may be moderated by the child’s affective evaluation, liking, of the character. However, the study of Macklin (1986) did not find an influencing effect of a liked character on the choice of a specific product. Taking it more broadly, we will try to discover whether involvement in Disney does have a moderating effect on the relationship between attitude towards and purchase intention for the bundle. Since involvement in characters is positively influenced by interest (Auter & Davis, 1991; Bryant & Vorderer, 2006; Zaichkowsky, 1986), familiarity (Zinkhan & Muderrisoglu, 1985), similarity (Cohen, 2001; Giles, 2002; Hoorn & Konijn, 2003), and liking (Konijn & Hoorn, 2005), we expect that involvement in Disney will strengthen the relationship between attitude towards the Disney themed bundle and purchase intention for the bundle. The following hypothesis has been formulated:

- H3: The extent to which children are involved in Disney strengthens the effect of the

attitude of children towards the Disney themed bundle on the purchase intention for the Disney themed bundle.

2.5 Conceptual Model

(37)

Figure 1: Conceptual Model

3. METHODOLOGY

In this chapter, the methodology will be described. This includes the data collection method and study design, information about the participants, and the data analysis method.

3.1 Data Collection Method and Study Design

(38)

offered some healthier alternatives in their canteen. Therefore, we were able to measure the effect of the implementation of the Disney themed bundles without being biased by the newness of the products. However, because of the collaboration with Disney, some stricter guidelines needed to be satisfied and therefore we needed to search for a new and healthier drink to include in the bundle. One guideline was that Disney does not allow sweeteners in the drinks they promote. Therefore, we choose to include DubbelFrisss Free, a sweetener free drink which only includes water, fruit juice from concentrate, natural flavor, and carbonic acid (DubbelFrisss, n.d.). From the moment we decided to include this drink in the bundle, we made sure that the drink was available in the soccer canteen of VV Heino to reduce the bias of newness of the product. The snack included in the bundle was a mini bag of salt JIMMY’s Popcorn. JIMMY’s Popcorn contains less fat, less sugar, and less salt than regular popcorn and is also naturally rich in healthy fiber (JIMMY’s Popcorn, n.d.).

(39)

Two different quantitative methods were used to collect the data needed to answer the research questions, including a change in sales numbers method and a survey. These methods will be discussed more thoroughly in the following two paragraphs.

3.1.1 Change in Sales Numbers Method

The change in sales numbers method was used to be able to support our findings with real sales numbers, and in order to find insights concerning H2a and H2b, whether the promotion of healthy

foods through the Disney themed bundle would positively influence the purchase of healthy foods, and negatively influence the purchase of unhealthy foods. Since VV Heino did not have an automated system that kept track of the sales per product, we had to count the sales ourselves. Before the implementation of the Disney themed bundle, we counted the sales during two Saturdays (the 2nd and the 9th of April). Additionally, we counted the sales during two Saturdays (the 16th_{and the 23}th_{of April) after the implementation of the Disney themed bundle. In this}

way, we could later on compare the sales numbers before the implementation of the bundle with the sales numbers after the implementation. Our intention was to increase the sales of the healthier alternatives through the implementation of the Disney themed bundle, while the sales of the unhealthier, comparable foods (such as chips) and drinks (such as Capri Sun and AA) decrease. It is important that the sales of unhealthier, comparable foods decreased since the total caloric intake of children needed to decrease. When children would buy both the Disney themed bundle and still the unhealthy foods, their caloric intake would even be increased by the implementation of the bundle. Therefore, we needed to count the sales of both the healthier alternatives (both in the bundle and separate) and the sales of relatively unhealthy, comparable foods and drinks to be able to discover whether and to what extent there was a shift in sales. 3.1.2 Survey

(40)

using a survey is “ambiguity of language”: The meaning of words is often ambiguous, and subjects do not always fully understand the questions (Myers, & Newman, 2007). Therefore we tried to reduce the chance of ambiguity through making the questions as easy as possible to comprehend for young children. We waited with conducting surveys until the children had bought something in the soccer canteen in order to be sure we would not influence their purchase decision after conducting the survey, and in this way they could answer the question about whether they did or did not buy the Disney themed bundle, and their reason for this decision. Before the first Saturday on which we conducted the surveys, a short message was placed on the webpage of VV Heino to inform the parents of the potential participants about our research and the surveys. In this way, parents would not be surprised and worried about the fact that their children filled in a survey of which they did not know beforehand. Furthermore, when the parents were present in the canteen, we asked them for permission before conducting the survey. Almost in all cases there was one parent of one of the children in a small group present in the canteen who was watching multiple children. We then asked this parent for permission to also conduct the survey with these other children. When no parent was present, we explained our idea to the children and what we expected from them, and subsequently they could decide for themselves if they wanted to participate. When asking for permission from a parent and subsequently when talking to a child directly, we shortly explained that we were conducting an experiment in the canteen of VV Heino, and that we needed the children to fill in the survey to help us gather as many information as possible in order for our experiment to succeed. After a child agreed to participate, we showed the bundle to the children. In this way, the questions and answers related to the bundle would not be biased by the fact that they had not seen the bundle before filling in the survey. Subsequently, we helped the children with filling in the survey in order to reduce the chance of miscomprehension or missing data. See the picture in ‘Appendix 6’, where we are assisting a girls soccer team.

(41)

end questions. These questions are called fixed-alternative questions, which require the respondent to select from a predetermined set of responses (Malhotra, 2010). The three questions about why children play soccer, and why they did buy or did not buy the Disney themed bundle contained an option for the children to fill in their own reason when their reason was not available in one of the suggested answers. The survey started generally, with questions concerning demographics (age), soccer, and whether they already bought the Disney themed bundle once. The following questions subsequently became more focused through asking for their attitude towards the Disney themed bundle, food and beverages, Donald Duck, and Disney. Two questions were related to what the participant normally eats and drinks after a soccer match, or just in the soccer canteen without having played a match. Through these answers we could compare what the majority of our respondents normally eats and drinks with popcorn and DubbelFrisss Free. Our aim was to increase the healthiness of what children eat in the soccer canteen, therefore it would not be positive for our research when many children normally eat fruit and now choose to buy the bundle and thus eat popcorn instead of fruit. However, we expected that many children buy candy or crisps in the soccer canteen, and in that situation, popcorn is a healthier alternative.

(42)

point Likert scale ranging from ‘Not fun at all’/‘Not tasty at all /‘Not like to at all’/‘Never’ to ‘Great fun’/‘Very tasty’/‘Really like to’/‘Always’. The five point Likert scale was viewed using both smileys and one- or two words descriptions to decrease the chance of miscomprehension among the participants. Filling in the survey took the children around 2 to 5 minutes. When the children handed in the survey, they received a small reward in the form of a plastic straw (see ‘Appendix 8’) to increase involvement, to thank them, and in order to encourage other children to fill in our survey. Moreover, we wrote down two remaining demographics about the specific participant on their survey paper, including gender (a cross indicated a girl) and a guess of whether the participant has a normal weight, or is probably overweight (a circle indicated probably overweight). These remaining demographics were used when analyzing the results.

3.2 Participants

The study involved Dutch children aged four to fourteen years. These children were either soccer players from VV Heino or from competing clubs, or children who were watching the soccer match from the side of the field. The participants were approached in the canteen.

3.3 Data Analysis Method

In order to analyze the data in a quantitative manner, the predictive analytics software IBM SPSS Statistics 23 is used. In this software, data can be handled in order to receive descriptives about the respondents, and to discover relationships between the variables.

4. RESULTS

(43)

filling in the survey (see ‘Appendix 6’), no data was missing or invalid, and therefore there was no need for us to delete respondents from the data set.

On the first intervention Saturday four Disney themed bundles have been sold, and on the second intervention Saturday two bundles have been sold. Between these two days six bundles have been sold. Hence, in total, in the two weeks of the implementation of the bundle, twelve bundles have been sold.

4.1 Descriptive Statistics

4.1.1 Demographics

In ‘Appendix 9’ the SPSS output of all the descriptive statistics is presented. A total of 95 children, of which 65 boys (68.4 %) and 30 girls (31.6 %) participated in our research. The mean age of the participants was 9.34, the youngest participant was 4 years old, and the oldest was 14 years old. Since our participants are children from multiple ages, the sample can be considered as representative. ‘Figure 2’ shows the age distribution of our participants. It can be seen that most participants are ten years old (n = 19).

Figure 2: Age distribution participants

4.1.2 Soccer

The majority of participants were soccer players at clubs other than VV Heino (n = 59, 62.1 %). Furthermore, the main reason why our participants play soccer is because they enjoy it (fun)

(44)

(n = 67, 70.5 %). Other reasons were because it is healthy (n = 7, 7.4 %), because their friends also do it (n = 5, 5.3 %), or because their parents oblige them (n = 2, 2.1 %). ‘Figure 3’ visualizes the division of reasons why our participants play soccer. Since the whole research is conducted in the soccer canteen of VV Heino, we only conducted surveys with children from this soccer club, or from the same soccer region of VV Heino. This might make the sample less representative for all Dutch children. However, we do not expect big differences in attitudes towards the bundle and its contents among children from different regions in the Netherlands. Therefore, we still assume the sample to be representative for all Dutch children aged four to fourteen years.

Figure 3: Reason why participants play soccer

4.1.3 Bundle

Unfortunately, the majority of respondents did not buy the bundle (n = 88, 92.6 %). The main reason why these children did not buy the bundle was that they had not seen it (n = 73, 76.8 %). Five respondents (5.3 %) answered that their parents prohibited them to buy the bundle. Only four respondents (4.2 %) indicated that bundle did not look attractive. Lastly, two respondents (2.1 %) did not buy the bundle because nobody else bought the bundle. Reasons why the seven respondents did buy the bundle was because they indicated to like Donald Duck (n = 3, 3.2 %), the food in the bundle (n = 3, 3.2 %), or the bundle itself (n = 1, 1.1 %).

14 67 7 2 5 0 10 20 30 40 50 60 70 80 Doesn´t play soccer

Fun Healthy Parents Friends

N u m b er of r es p on d en ts

(45)

4.1.4 Food Consumption

We found that the majority of the participants eats crisps (n = 28, 29.5 %), candy (n = 25, 26.3 %), or bread (n = 21, 22.1 %) after playing a soccer match (or when they are in the canteen without playing soccer). Additionally, the majority of participants drinks sports drink (n = 25, 26.3 %), lemonade (n = 25, 26.3 %), water (n = 25, 26.3 %), or soda (n = 10, 10.5 %) after playing a soccer match. See ‘Figure 4’ and ‘Figure 5’ for a visual representation of our findings concerning food consumption.

Figure 4: What children normally eat after a soccer match

Figure 5: What children normally drink after playing a soccer match 4 25 2 28 9 21 3 3 0 5 10 15 20 25 30 Not applicable Candy Candy bar

Crisps Cookies Bread Fruit Other

(46)

4.1.5 Attitudes and Purchase Intention

The average attitudes and the average purchase intention of our respondents appeared to be high, see ‘Table 1’.

Table 1: Descriptives attitudes and purchase intention

Variable Mean Min. Max. Std. Dev.

Attitude towards bundle 4.19 1 5 .803

Attitude towards key ring 4.38 1 5 .774

Attitude towards popcorn 3.93 1 5 1.355

Attitude towards DubbelFrisss Free 4.06 1 5 1.050

Purchase intention bundle 4.08 1 5 .953

Attitude towards Donald Duck 4.34 1 5 .895 Frequency reading Donald Duck 3.49 1 5 1.465

Attitude towards Disney 4.00 1 5 1.130

4.2 Factor Analysis

(47)

communalities of all variables should be higher than 0.4. In ‘Appendix 10’ we can see that all communalities are higher than 0.4. We can conclude that the factor analysis is appropriate.

To be able to decide on the number of factors to continue with in our analysis, we conducted the Principal Component Analysis (PCA) (see ‘Appendix 10’). Firstly, we looked at the eigenvalues of the factors, which shows the total variance explained by each factor (Malhotra, 2010). The eigenvalue of the factors we will continue with should be more than 1. This is the case when we combine the three variables into one factor. Second, we look at the total explained variance, which should be over 60%. Third, each factor has to explain over 5%. The SPSS output (‘Appendix 10’) shows us that all three criteria are satisfied. Finally, to get a visual impression of the number of factors to choose, we created a scree plot (see ‘Appendix 11’). A scree at point two means one factor is the best choice.Taking into account the previous information, we decide to continue our analysis with one factor.

Finally, we will test the reliability of the factors through using the Cronbach’s Alpha in order to measure internal consistency. The Cronbach’s Alpha has to have a minimum value of 0.6 for the factors to be reliable. Since the Cronbach’s Alpha is higher than 0.6 (α = .668) (see ‘Appendix 12’), we can conclude that the factor analysis is reliable. Additionally we tested whether the Cronbach’s Alpha would increase when one of the three variables would be deleted from the factor. We can conclude that the Cronbach’s Alpha does not decrease when one variable is deleted, therefore we will combine all three variables into one factor. We named the factor ‘InvolvementDisney’ and labeled it ‘Involvement in Disney’. The mean of this factor appeared to be 3.94, with a minimum of 1 and a maximum of 5 (MInvolvementinDisney = 3.94, SD =

.920) (see ‘Appendix 12’).

(48)

comparing the sales numbers of before the intervention and after the intervention. The correlations test, and both analyses will be performed in the following sections.

4.3 Correlations

(49)

4.4 Multicollinearity

It is important to check the independent variables for multicollinearity. Multicollinearity can be defined as a state of very high intercorrelations among independent variables (Malhotra, 2010). This leads to difficulties in determining what causes the dependent variable, in our study purchase intention for the bundle, to change. This in turn makes it hard to distinguish the effects of independent variables in a regression. To analyze whether multicollinearity is present between the dependent variables, variance inflation factors (VIF) need to be created in SPSS. When the VIF value is over 4, there is moderate multicollinearity, and when the VIF value is over 10, it means that there is strong multicollinearity present. When the VIF value is under 4, it can be concluded that the model is not biased and that we are subsequently able to distinguish what causes changes in the dependent variable. We will check this for every regression.

4.5 Hypotheses Testing – Survey

The data obtained by the survey is used to test H1, whether the attitude of children towards the

Disney themed bundle in the promotion of healthy foods will positively influence the purchase intention for the bundle,and H3, whether the effect of the attitude of children towards the bundle

(50)

can use its value to analyze the attitudes of our participants. Thus, in our research we will regard the differences between the five choices as equal. This was also assumed by Baker, Hardyk, and Petrinovich (1966) and Labovitz (1967), who have empirically found that it matters little if an ordinal scale is treated as an interval scale.

In order to analyze the influences from our hypotheses, we will use linear regression. Regression analysis is a “powerful and flexible procedure for analyzing associative relationships between a metric dependent variable and one or more independent variables” (Malhotra, 2010, p. 568). According to Leliveld and Blaauwendraad (2013) and Malhotra (2010), there are multiple assumptions for regression. The first is that the sample needs to be based on independent observations. The second assumption entails that there should be a linear relationship between the dependent and the independent variable. Lastly, the residuals should be approximately normally distributed. To test this last assumption, we created a histogram and normal P-P plot of regression standardized residual of our dependent variable Purchase intention in SPSS, see ‘Appendix 15’. We included Attitude towards Disney themed bundle, Attitude towards key ring, Attitude towards DubbelFrisss, and Attitude towards popcorn as independent variables. Taking into account the histogram and P-P plot, it seems that a deviation in normality is present. However, the fact that we have an n which is quite high (n = 95), and that we have a small number of independent variables causes the central limit theorem to ensure that the confidence intervals and the p-values are good approximations, despite the violation of the normality assumption (Leliveld & Blaauwendraad, 2013).

(51)

4.5.1 Main Effect

In order to analyze the influence of the attitude of children towards the Disney themed bundle on purchase intention for the bundle (H1), we are interested in the effect of the attitude of

children towards the Disney themed bundle. We also incorporated the contents of the bundle (key ring, popcorn, and DubbelFrisss Free) in our analysis in order to also analyze those influences on purchase intentions for the bundle. We conducted a multiple regression of Attitude towards Disney themed bundle, Attitude towards key ring, Attitude towards popcorn, and Attitude towards DubbelFrisss, on Purchase intention Disney themed bundle. In ‘Appendix 16’ we can see that the regression was significant, R2 = 0.606, F (5, 89) = 27.37, p = .000. Multicollinearity is not a problem since the VIF values of all variables are lower than 4. Age is included as a covariate, and proves to have a significant influence on Purchase intention Disney themed bundle, B = -.118, p = .001. From the output it appears that Attitude towards Disney themed bundle positively influences Purchase intention Disney themed bundle, B = .555, p = .000. Thus, we can already conclude that H1 is supported.

Concerning the contents of the bundle, the output shows us that only Attitude towards popcorn significantly influences Purchase intention Disney themed bundle (B = .156, p = .001). Attitude towards key ring (B = .097, p = .260) and Attitude towards DubbelFrisss (B = .038, p = .539) both do not significantly influence the purchase intention. To compare the strengths of the effects of the independent variables in the model, we look at the standardized Beta coefficients. The Beta coefficients suggest that the influence of the attitude of children towards the bundle (Beta = .468) is stronger than the influence of the attitude of children towards popcorn (Beta = .222). The relationship between attitude towards the bundle and purchase intention, and attitude towards popcorn and purchase intention are visualized in ‘Figure 6’.

(52)

and DubbelFrisss Free do not. The influence of the attitude of children towards the bundle is the strongest effect. ‘Table 2’ gives an overview of the results of the multiple regression.

Figure 6: Relationship between attitudes and purchase intention for the bundle

Table 2: Results from multiple regression main effect

* = Significant

Dependent variable: Purchase intention

4.5.1.1 Comparing Groups

Another way in which the influence of the attitudes of children can be analyzed, is by using an independent samples t-test. We compared the average attitudes, purchase intentions, and involvement in Disney of the group of children who did buy the bundle with those of the group of children who did not buy the bundle. In order to discover whether the two groups differed significantly, we conducted an independent samples t-test of attitude towards the bundle, the key ring, popcorn, DubbelFrisss Free, purchase intention, and involvement in Disney, with Purchase as grouping variable. From the SPSS output (see ‘Appendix 17’), it appeared that the independent samples t-test was not significant for attitude towards the bundle (t (93) = 1.31, p

Variable B Beta Sig. VIF

Age -.118 -.299 .001* 1.624

(53)

= .192), attitude towards the key ring (t (93) = .18, p = .861), attitude towards DubbelFrisss Free (t (93) = -1.29, p = .200), and purchase intention (t (93) = .99, p = .323). However, the average means of the children in the two groups did significantly differ in the attitudes towards popcorn (t (93) = 2.22, p = .029) and involvement in Disney (t (93) = 2.05, p = .043). This indicates that children who did buy the bundle differ significantly from children who did not buy the bundle in their attitude towards popcorn (Mdidbuy = 5.00, Mdidnotbuy = 3.84) and in the

extent to which they are involved in Disney (Mdidbuy = 4.62, Mdidnotbuy = 3.89). An overview of

the means per variable of both groups is visualized in ‘Figure 7’and presented in ‘Table 3’. Figure 7: Differences in means per variable between participants who did and did not buy

Table 3: Differences in means per variable between participants who did and did not buy

Variable Purchase Mean Std. Dev. Sig.

Attitude towards bundle Did buy 4.57 .535 .192 Did not buy 4.16 .815

Attitude towards key ring Did buy 4.43 .787 .861 Did not buy 4.38 .778

Attitude towards popcorn Did buy 5.00 .000 .029* Did not buy 3.84 1.372

Attitude towards DubbelFrisss Free Did buy 3.57 1.397 .200 Did not buy 4.10 1.018

Purchase intention bundle Did buy 4.43 .535 .323 Did not buy 4.06 .975

Involvement in Disney Did buy 4.62 .488 .043* Did not buy 3.89 .926

* = Significant

Grouping variable: Purchase

(54)

4.5.2 Moderating Effect

A moderator is a variable which affects the direction and/or strength of the relation between an independent and a dependent variable (Baron & Kenny, 1986). The moderator in our study, involvement in Disney, is a quantitative variable (Baron & Kenny, 1986). In order to test whether involvement in Disney has a moderating effect on the relationship between the attitude of children towards the Disney themed bundle and the purchase intention for the bundle (H3),

we have to show that the nature of the relationship changes when the value of the moderator (involvement in Disney) changes. We will do this by including the interaction effect in the model and subsequently check whether the interaction effect is significant. First, we will mean center the variables which will be included in the interaction variable in order to avoid potentially problematic high multicollinearity. We made an interaction variable of the mean centered versions of Attitude towards Disney themed bundle and Involvement in Disney, and started its name with “MC”. To discover whether a moderating effect is present, we conducted a linear regression in SPSS. The output (see ‘Appendix 18’) shows us that the VIF value of all variables is lower than 4, and therefore multicollinearity is not a problem. Again, Age is included as a covariate, and proves to have a significant influence on Purchase intention Disney themed bundle, B = -.132, p = .000.

In the first table of the SPSS output in ‘Appendix 18’ it can be seen that both Model 1 (F (3, 91) = 45.98, p = .000) and Model 2 are significant (F (4, 90) = 34.26, p = .000). In deciding whether Model 2 accounts for more variance than Model 1, we look at the R2 change. The second table shows that Model 2 with the interaction effect accounted for slightly more variance but no significance compared to Model 1 with only the single variables (R2_{change =}

(55)

However, the third table of ‘Appendix 18’ shows that involvement in Disney does significantly correlate with the purchase intention for the Disney themed bundle (B = .253, p = .001). Therefore, we can conclude that involvement in Disney seems to have a direct influence on purchase intention, but no moderating effect on the relationship between attitude towards the bundle and purchase intention for the bundle. The relationship between involvement in Disney and purchase intention for the bundle is visualized in ‘Figure 8’, and an overview of the results from the linear regression with the interaction effect is presented in ‘Table 4’.

Figure 8: Relationship between involvement in Disney and purchase intention

Table 4: Results from linear regression moderating effect

* = Significant

Dependent variable: Purchase intention

4.6 Hypotheses Testing – Sales Numbers

In order to analyze hypotheses 2a and 2b, whether the promotion of healthy foods through the Disney themed bundle would positively influence the purchase of healthy foods and negatively influence the purchase of unhealthy foods in the soccer canteen, we compared the sales numbers

Variable B Sig. VIF

Age -.132 .000* 1.602

Attitude towards bundle .461 .000* 2.193

Involvement in Disney .253 .001* 1.186

(56)

of before the implementation (two Saturdays) with the sales numbers after the implementation of the Disney themed bundle (two Saturdays). We assumed that, on all four Saturdays, the same number of people has been in the canteen of VV Heino. In ‘Appendix 19’ the sales numbers of the four Saturdays in April are presented. In order to manage and compare the sales numbers of the four weeks more easily, we created categories for both the drinks and the snacks. We divided the drinks into categories according to the number of calories the drink contained per 100 milliliter, whether they contain sweeteners, and two separate categories for DubbelFrisss Free and water. ‘Appendix 20’ presents the different categories and their corresponding drinks, including the number of calories per 100 milliliter. ‘Figure 9’ visualizes the average sales numbers of each category of drinks before and after the implementation of the bundle.

Figure 9: Sales numbers drinks before and after the implementation

Next, we divided the snacks into categories according to type of snack, including cookies, candy, single candy, candy bar, crisps, and popcorn. These categories can also be found in ‘Appendix 20’, including the number of calories per piece. Subsequently, we summed up the sales per category per week, see ‘Appendix 21’. ‘Figure 10’ visualizes the average sales numbers of each category of snacks before and after the implementation of the Disney themed

THE DISNEY THEMED BUNDLE

THE DISNEY THEMED BUNDLE

The influence of using a Disney themed bundle in nudging

children in soccer canteens towards healthier eating

THE DISNEY THEMED BUNDLE

The influence of using a Disney themed bundle in nudging

children in soccer canteens towards healthier eating

ABSTRACT

EXECUTIVE SUMMARY

PREFACE

TABLE OF CONTENTS

1. INTRODUCTION

1.1

Children

1.2

Soccer Canteens

1.3

KNVB

1.4

Disney

1.5

Summary

1.6

Research Questions

1.7

Thesis Outline

2. LITERATURE REVIEW

2.1

Healthy Eating Behavior

2.2

Individual Factors

2.3

Environmental Cues

2.4

The CAN Approach

2.4

Hypotheses

2.5

Conceptual Model

3. METHODOLOGY

3.1

Data Collection Method and Study Design

3.2

Participants

3.3

Data Analysis Method

4. RESULTS

4.1

Descriptive Statistics

4.2

Factor Analysis

4.3

Correlations

4.4

Multicollinearity

4.5

Hypotheses Testing – Survey

4.6

Hypotheses Testing – Sales Numbers