STIMULATING FRUIT CONSUMPTION AMONG CHILDREN IN SOCCER CANTEENS

(1)

STIMULATING FRUIT CONSUMPTION AMONG CHILDREN

IN SOCCER CANTEENS

The influence of attractiveness of a fruit bowl in stimulating children to eat

more fruit

by

Wouter Luinge

(2)

STIMULATING FRUIT CONSUMPTION AMONG CHILDREN

IN SOCCER CANTEENS

The influence of attractiveness of a fruit bowl in stimulating children to eat

more fruit

by

Wouter Luinge

s2183463

University of Groningen Faculty of Economics and Business

MSc Marketing 16th of January 2017 J.C. Kapteynlaan 36 9714CR Groningen (06) 11347455 w.luinge@student.rug.nl / wouter.luinge@knvb.nl

(3)

Management summary

Among children the problem of overweight and obesity is rising. Also in the Netherlands, obesity is a growing issue. One of the contributors of this problem is the marketing of unhealthy products to children. One technique to make such products more appealing to children is by the use of cartoon characters or real-life celebrities. From research it appeared that such characters can successfully influence purchase intentions for (unhealthy) products. Therefore, this study aims to investigate if such characters can also be applied successfully for unpackaged fruit at the point-of-purchase, by depicting a role model on a fruit bowl. Also the effect of using a (nutritional) claim on the fruit bowl is investigated.

As the largest sports association in the Netherlands, the Koninklijke Nederlandse Voetbalbond (KNVB) wants to take its responsibility and giving a contribution to the overweight problem. Since the majority of the assortment of soccer canteens consists of unhealthy foods, the KNVB is interested in ways how to stimulate the consumption of more healthy foods. In an environment where a healthy lifestyle is stimulated by means of exercising, it seems remarkable that many unhealthy products are consumed in soccer canteens. As fruit is seen as food product that helps to recover after an effort, fruit is the healthy food that is focused on in this study.

Therefore, in commission by the KNVB, a field experiment is conducted in 8 different soccer canteens in Dutch province of Groningen. A between participants factorial design with different treatment conditions is conducted: a branded fruit bowl, a branded fruit bowl with (nutritional) claim, a branded fruit bowl with role model and a branded fruit bowl with role model and claim. To be able to also investigate the effect of a neutral, regular fruit bowl, there is also a neutral condition: a non-branded, regular fruit bowl. As such, in total there are 5 different conditions. By means of a survey among 223 children, children’s attitude towards the fruit bowl and intention to buy fruit from the fruit bowl were investigated.

(4)

since messages delivered by role models are proved to be very effective in influencing children.

From the results of this study, it is concluded that a branded fruit bowl with role model and a branded fruit bowl with claim, are somewhat more liked (and children have somewhat higher intentions to buy fruit from them) than a branded only fruit bowl or neutral (non-branded) fruit bowl. However, the differences are insignificant. Nevertheless, a branded fruit bowl with role model and claim is significantly more liked (and has significantly higher purchase intentions from children) than a neutral fruit bowl, but insignificantly to the branded only fruit bowl. Based on this study, canteens are advised to place a branded fruit bowl with role model and claim on their bar in order to stimulate the consumption of fruit among children, since this is the only fruit bowl that is significantly more liked than a neutral (non-branded) / regular fruit bowl and that has significantly higher intentions of children to buy fruit from it.

(5)

Preface

After graduating my first master, Business Administration: Strategic Innovation Management, in January 2016, I only had to write my thesis for the master Marketing Management. Since the thesis of my first master was very theoretical in nature, I had the desire to write a more practical thesis for the master Marketing. Therefore, my attention was caught when the

possibility to conduct master thesis research at the KNVB, was announced. Two months later, it was the first day to go the headquarter of the Dutch Soccer Association, the ‘Bondsgebouw’ at the KNVB campus at Zeist. For me, as a person not especially interested in soccer, an exciting moment. However, given the fact that the research would be focused on healthy food and I am a passionate athlete, it fitted like a glove. I am looking back on a great time the past five months, with as one of the highlights my live radio interview with RTV Noord to talk about my research in December 2016. Moreover, nice to report: at least 3 out of the 8 canteens already decided to continue the offering of fruit in their assortment.

Summarized, it was great to conduct research in commission by the KNVB. I would like to express my gratitude to the KNVB for funding my thesis research, and in particular to Peter Legters, project leader Corporate Social Responsibility at the KNVB, for supporting me on behalf of the KNVB during the research. Besides that, I would like to express my gratitude to my thesis supervisor, Prof. Dr. Ir. Koert van Ittersum, Professor of Marketing and Consumer Well-Being at the University of Groningen, for his support and feedback during the period of my thesis.

Wouter Luinge

(6)

1. Introduction

1.1 Background

The last couple of decades, the prevalence of overweight and obesity has risen (The Global BMI Mortality Collaboration, 2016). From recent research of the World Health Organization (WHO), it is estimated that 1-3 billion adults worldwide have overweight and 600 million are obese (WHO, 2015). Also in the Netherlands, problems of overweight and obese are faced. In 2015, 50% of the adults and 12% of the children were dealing with overweight (CBS, 2016). A child is overweight when he or she has a body mass index in the 85th-95th percentile (Korinek, Bartholomew, Jowers & Latimer, 2013). Since child overweight and obesity has grown the last decades, the weight of children has been developed as a growing concern. Overweight children face increased risk of developing type 2 diabetes, high blood pressure, heart- and psychological problems (Korinek, et al., 2013) (Ruxton, 2004). Furthermore, it is claimed that obese children have a low self-esteem and are often socially excluded and bullied (Schwartz & Puhl, 2003). On the longer term, there is a higher risk of cancer and cardiovascular diseases (Ruxton, 2004). It is claimed by experts that a generation of children evolves who possibly do not persist their parents (de Droog, Valkenburg & Buijzen, 2011). Therefore, there is a high need to solve the problem of obesity among children and to come with solutions to deal with this problem.

1.2 Food advertising to children

Nevertheless, there is a massive growth of food advertising targeted to children, mostly promoting energy-dense foods and soft drinks. In the Netherlands, 94% of child-directed food advertising is related to unhealthy products (Consumentenbond, 2011). As a result, children’s attitudes to, purchase intentions to and actual consumption of those products are positively influenced (Smits & Vandebosch, 2012). For example, 97% of children’s purchase requests consist of candy (Voedsel en Waren Autoriteit, 2006). As such, it is stated that one of the main causes of child obesity is due to advertisements targeted to children (de Droog et al., 2011).

(9)

offering large packages or consuming unhealthy foods. (Wansink, 2010) (Birch & Fisher, 1997). When environments offer calorie-rich, sugary and fat food, it is said that children’s preferences for such foods are enhanced (Birch & Fisher, 1997).

1.3 Cartoon character or celebrity endorsement

In targeting children, marketers often use (licensed) brand characters – ‘images of (animated) animals or people connected to a brand’, (de Droog, Buijzen & Valkenburg, 2012: 1068) - which are often displayed at the packaging of products, in advertisements of the product or at displays showing the product (Smits & Vandebosch, 2012). Brand spokes characters, which could be either cartoon characters or athletes or celebrities, represent, enounce and promote brands (Henry & Story, 2009). The main contribution of cartoon/celebrity endorsers is likeability, which has a positive effect on attitude towards the advertisement and the product or brand (Levin & Levin, 2010). As such, it can increase purchase intention (Phillips & Lee, 2005). Furthermore, it provides meaning to consumers due to the value it adds beyond the functional benefits of a product (Hosany, Prayag, Martin & Lee, 2013). Besides that, products or advertisements with cartoon/celebrity endorsers lead to a higher attention level among children (Neeley & Schumann, 2004).

However, most of these marketing messages promote unhealthy food (Roberto, Baik, Harris & Brownell, 2011), while those promoting more healthy alternatives belong to a large minority (Smits & Vandebosch, 2012). Therefore, the use of such spokes-characters are criticized by experts and consumer representative groups (Roberto et al., 2012), since it has been proven that the use of cartoon-characters are effective in encouraging children to consume unhealthy foods (de Droog, et al., 2011).

(10)

1.4 Problem statement and research questions

To be able to conduct a field study at the point of purchase, this study is done in commission by the largest sports association of the Netherlands: the Koninklijke Nederlandse Voetbalbond (KNVB): the Royal Dutch Soccer Association, which is the official national organization of soccer in the Netherlands. The KNVB consists of 1,2 million members and employs more than 500 people (KNVB, 2016). Since the KNVB aims to take its responsibility as largest Dutch sports association, it considers socially responsible projects as highly important. One of their social responsibility projects is ‘Lekker Bezig,’ in which they attempt to make sports canteens healthier by also offering healthier food (KNVB, 2016). The offering of more healthy alternatives in sports canteens seems necessary. From research of consumer behavior, it is stated that decisions made by consumers often are irrational and susceptible to biases. When applied to sports canteens, in their paper van Kleef, Otten & van Trijp (2012: 1073) state: ‘people that visit a canteen often are in a hurry and tend to be hungry, which make them more prone to the aforementioned biases. Typical checkout counter assortments including a variety of chocolate bars, candy and savory snacks speak to the desire of immediate pleasure rather than uncertain rewards such as achieving a healthy

body weight.’ Especially in sports canteens, located in an environment promoting a healthy

lifestyle by exercising, it seems remarkable that a large majority of the assortment consists of unhealthy food. As purchases in soccer canteens often are spontaneous and impulsive, especially unhealthy foods that catch attention are bought by children (Nijkamp, Dorkhom & Akihary, 2016). Therefore, it is interesting to investigate how healthier food can be promoted in order to trigger people to make more health-conscious choices in soccer canteens. As such, soccer canteens are considered to be an appropriate environment to use in this study, as soccer canteens are places where children frequently buy something (mostly unhealthy) to eat or drink after a match or training.

The main aim of this study is to test at the point of purchase, by means of a field

(11)

Based on the above considerations, the following problem statement can be formed: ‘what is the influence of using a spokes-character (either with or without a nutritional claim) on a fruit bowl in stimulating children to consume more fruit in soccer canteens?’ This problem statement can be translated in several research questions: ‘does a spokes-character on a (branded) fruit bowl leads to more positive attitudes towards that fruit bowl and to a higher intention to buy fruit from that fruit bowl, compared to a (branded) fruit bowl without role model?’ ‘Does a (nutritional) claim on a (branded) fruit bowl leads to more positive attitudes towards the fruit bowl and higher intentions to buy fruit from it, compared to a (branded) fruit bowl without claim?’ And ‘what is the influence of a claim stated by a role model on a fruit bowl on children’s attitude towards that fruit bowl and the intention to buy fruit from it?’ Besides that, the effect of age is tested in the relationship between the fruit bowls with role model and the attitude towards those fruit bowls and intention to buy fruit from them. This leads to the following research question: ‘does age moderate the effect of a (branded) fruit bowl with role model on the attitude towards that fruit bowl and the intention to buy fruit from it?’ Furthermore, the effect of health consciousness of a child in the relationship between the fruit bowls with a claim and the attitude towards those fruit bowls and intention to buy fruit from them is investigated, resulting in the following research question: ‘does health consciousness moderate the effect of a (branded) fruit bowl with claim on the attitude towards that fruit bowl and the intention to buy fruit from it?’ As depicting role models could be expensive, it is also investigated if a cheaper option works well: by solely depicting a strong well-liked brand name on a fruit bowl, leading to the following research question: ‘what is the influence of depicting a strong, well-liked brand name or logo on a fruit bowl on the attitudes and intentions of children compared to a neutral non-branded fruit bowl?’

1.5 Contribution of this study

(12)

lifestyle. Besides that, it is useful to investigate how unpackaged, unprocessed food can be made more attractive for children, since the marketing of fruit to consumers remained very static and hardly changed among the time (de Droog et al., 2011). Furthermore, little research is executed on the effects of sports celebrity endorsement by food brands on children. As the problem of obesity is of growing relevance, it is important to find ways to direct children to consume more healthy food. As such, this study also tries to make a contribution to combatting the obesity problem.

Secondly, the scientific contribution of this study is that it enriches to the literature of marketing to children. Although already research exist on the topic of spokes-characters in the stimulation of more healthy foods among children, much of them is executed on packaged healthy food (e.g. packaged fruit) and in an artificial setting rather than at the point of purchase. At the point of purchase, especially in sports canteens, purchases are made more spontaneous and impulsive, possibly leading to different outcomes.

1.6 Thesis outline

This paper is structured as follows. Firstly, a theoretical framework is provided about how the preferences of food among children are formed. Further, there is elaborated on the use of cartoon or celebrity endorsers in order to stimulate children to consume more healthy food. Based on this framework, hypotheses are derived. After that, the methodology is provided that elaborates on the research setting of this study, followed by the analysis and discussion of the results. Subsequently, managerial and theoretical implications are given. Lastly, limitations are given and directions for future research.

2. Theoretical framework

(13)

2.1 From unhealthy to healthy food marketing to children

As there is a high need to cope with the problem of child obesity, it is important to make a shift from unhealthy to healthy marketing to children. However, on this moment unhealthy marketing to children – each process of distribution, promotion or sales of unhealthy foods - is the rule rather than exception (Folta, Goldberg, Economos, Bell & Meltzer (2006). As a result, children are exposed to many marketing activities pertaining unhealthy food, encouraging them to purchase more of those foods, which positively influences their food consumption (Molnar, Garcia, Boninger & Merrill, 2008). From research executed in the Netherlands, it is stated that 84% of the parents purchase products on requests of their children (Voedsel en Waren Autoriteit, 2006). Such requests of children for products are named ‘pester power,’ (Harris, Schwartz & Brownell, 2009). Furthermore, it is said that children have cognitive limitations, which make them more susceptible to advertisements (Hoyer, MacInnis & Pieters, 2013) (Molnar et al., 2008). As a result, among experts there is a raising concern about child-directed food promotions (Harris et al., 2009). For those reasons, in many countries there are implemented advertisements codes or sometimes advertisements targeted to children are even prohibited (de Droog et al., 2011)†. Important targets related to child advertising in order to prevent child obesity are: diminishing the amount to which children are exposed to marketing of unhealthy foods, as well as to reduce the persuasiveness of those marketing (Kraak & Story, 2015).

2.2 The essence of fruit

It is important to take into consideration that solely diminishing the consumption of unhealthy food cannot solve the obesity problem. Is it also stated that the consumption of healthy food must be promoted. One could think of the promotion of fruit and vegetables (de Droog et al., 2011). It has been proven that the consumption of fruit and vegetables, as a result of its low amount of calories and high water and fiber composition, supports a balance in weight (Korinek et al., 2013). As such, there is a relationship between a higher consumption of fruit and vegetables and a lower body weight (Lin & Morrison, 2002). In general, fruit contains a low amount of calories and a high amount of vitamin C, beta-carotene, potassium, fiber and

†_{Recently (Fall 2016) the ‘Federatie Nederlandse Levensmiddelen Industrie’ announced that during 2017, child} idols (e.g. Spongebob, Minions) will disappear from the packages of unhealthy food products in the

(14)

folate (Bezbaruah & Brunt, 2012). Furthermore, risks for chronic diseases are diminished (Korinek et al., 2013). For children in the age of 4-12 years, two pieces of fruit and 150 grams of vegetables are recommended (Zeinstra, Koelen, Kok & de Graaf, 2007). However, many children do not meet these guidelines regarding fruit and vegetables consumption. Therefore, it could be useful to promote the intake of fruit and vegetables by means of advertisements or other marketing cues to make those foods more attractive for children and in order to be able tackle the obesity problem. As such, is would be useful to gain more insights in how children process advertisements in order to effectively influence children in promoting fruit and vegetables.

2.3 Information processing of children

The ability of people to identify content that is commercial in nature and to comprehend on the underlying persuasive aim is referred to as ‘persuasion knowledge,’ (Waiguny, Nelson & Terlutter, 2014). The Persuasion Knowledge Model (PKM) describes that ‘individuals hold a general set of beliefs about how persuasion agents (e.g. advertisers) operate, including perceptions of agents’ goals and tactics, evaluation of the effectiveness or appropriateness of persuasion attempts, and self-reflection of a target’s own ability to cope with the attempts,’ (Waiguny et al., 2014: 258). When consumers identify the intention of marketers, they are able to critically investigate, discount and decline the marketing communication (Staiano & Calvert, 2012). It is said that people have a tension to have opposition to persuasive aims when they identify them (Waiguny et al., 2014).

(15)

(Dixon, Scully, Niven, Kelly, Chapman, Donovan, Martin, Baur, Crawford & Wakefield, 2013). Nevertheless, children from the age of 8 years are also influenced by advertisements. Although they are somewhat more cynical to advertisements, it is stated that there are still changes in development and metacognitive judgment. As such, among studies it is agreed that children do not have sufficient cognitive mechanisms to defend themselves against implicit persuasion (e.g. status symbols or peer pressure) (Dias & Agante, 2011) (Staiano & Calvert, 2012). Nevertheless, they evolve increased capabilities for hypothetical reasoning and they tend to rely less on judgments from adults about what is right or wrong. It is argued that children into this period are provided with mechanisms to defend themselves against excessive influence of advertisements (Ross, Campbell, Wright, Huston, Rice & Turk, 1984). They start to strengthen their analytical, logical and rational assessments, which enables them to be more self-reflective in the retaining and processing of information (Dias & Agante, 2011). As such, they understand that commercial messages can be biased (Miller, Seiders, Kenny & Walsh, 2011). Hence, they are able to combine, separate and transform actions and objects (McAlister & Cornwell, 2010). From the age of 13 to 16 years old, it is said that children achieve levels of processing messages compared to that of adults, in which they become critical to commercial aims. Nevertheless, this ability is still impeded by peer pressure and identity formation (Staiano & Calvert, 2012).

2.3.1 Theory of mind

(16)

intents is essential, since children frequently have to decode persuasive intentions that are implicit. Within the development of a theory of mind, one of the first skills children develop (between the age of 3 to 5 years) is the recognition that other persons possibly have false beliefs. Moreover, it is claimed that an important point in the development of a theory of mind is the understanding that other persons can have different beliefs and behave according to those beliefs (Lapierre, 2015). Although this skill is a necessary condition in understanding persuasion, it is not sufficient. The understanding of other’s mental states is denoted by first-order beliefs, while the understanding of second-first-order beliefs is defined as understanding about another’s mental state concerning another person’s beliefs and thoughts. This ability is seen as the critical condition in understanding the selling goals of marketing actions. When developed second-order beliefs, children are able to recognize that persuasive intents change their own beliefs (Lapierre, 2015).

2.3.2 Executive functioning

Another important concept is ‘executive functioning.’ McAlister & Cornwell (2010: 210) define this as follows: ‘executive functioning encompasses behavior planning, rule adherence, mental flexibility, and inhibition.’ Without developed this skill, children are not able to process information about objects appropriately, since they are not able to change their attention from one information cue to another. They only focus on the most salient information (McAllister & Cornwell, 2010). According to McAllister & Peterson (2006), executive functioning is developed between the ages of 3 to 5 years.

Notwithstanding, there are inconsistencies in literature concerning the age at which children are able to understand the persuasive goals of marketing. Some scholars claim that children of 8 years are able to understand the persuasive goal, while others suggest the appropriate age is over ten years or already at 6 years (Lapierre, 2015).

2.4 Forming of eating habits of children

(17)

2.4.1 Influence of parents

From research it appeared that parents have a substantial influence on the food preferences of children. It is argued that one of the first parental influences on a child’s experience with food and flavors is the decision to use breastfeed milk or formula-feed milk. Children from the latter category have experience with solely one flavor, while those in the former category have experience with different flavors, since they are exposed to multiple flavors that are transferred to the milk from the mother’s diet. It is said that this variety of experience with flavors can contribute to more acceptance of solid foods during the period of weanling. As such, breastfeed children have a higher initial acceptance of new foods compared to formula-fed children. Therefore, when children are exposed to a variety of foods early, they have a higher acceptance of new foods later (Birch & Fisher, 1997).

Later, when children move to a variety of foods, tastes that are natively preferred are sweetness and saltiness In contrast, sourness and bitterness are natively avoided tastes (Benton, 2004) (Birch & Fisher, 1997). Moreover, they have a natural tendency to associate the flavors of food to the context and consequences of eating (Birch & Fisher, 1997). Sweetness and sugar intake is seen as a source of energy, while bitterness is seen as a toxic taste. Therefore, children prefer sweet tastes and have an aversion to bitter tastes. Nevertheless, when becoming older the preference for sweetness diminishes (Benton, 2004). Furthermore, from research it appears that all children also have a natural tendency to have an aversion to new food and have a tension to disapprove it. Such aversion is present at ages from 18 to 24 months (Benton, 2004) and is called ‘neophobia,’ It is negatively correlated with the age of children (Bezbaruah & Brunt, 2012). Acceptance of new foods, with the exception of sweet and salty foods, does not arise directly, but occurs when children have repeated opportunities to consume such foods. As a result, the liking of those foods develops gradually and intake increases (Birch & Fisher, 1997) (Benton, 2004).

(18)

According to Benton (2004), parents often try to limit the availability of some foods, which could be for health reasons or which could be done to reward or punish a child. Food can be used as reward for consuming another (less liked) food. However, it is said that when children get a reward for eating a food they do not like, the preference for that food diminishes, while the rewarding food is increasingly preferred, leading to an undesired effect (Benton, 2004). Furthermore, when parents limit the access of certain food to their children, the liking of those foods by children is increased, which may have a positive impact on the consumption of such foods (Birch & Fisher, 1997).

2.4.2 The power of marketing in forming food preferences

Another factor that has influence on the food choice and preference of children is a child’s repeated exposure to marketing campaigns (Bezbaruah & Brunt, 2012) (Dixon et al., 2013). It is said that those campaigns increase awareness and familiarity with foods, diminishes neophobia and increases food consumption (Dias & Agante, 2011). Food preferences and choices are strongly reflected by children’s TV exposure experience. Advertisements on TV regarding food and beverages have an impact on the consumption of children (Dias & Agante, 2011). Furthermore, some foods are promoted to children, by persuading them to consume particular foods (Elliott, 2011). When children are repeatedly exposed to certain foods by marketing cues, it could be assumed that they increasingly like the food, an effect termed as ‘mere exposure effect,’ (Fennis & Stroebe, 2016).

From a study of Elliott (2011) it appears that food marketing targeted to children reshaped the more comprehensive perceptions of children about the meaning of food and what kinds of foods are meant for them. From this study it becomes clear that ‘children perceive food meant for them as categorically different than from regular or adult food,’ (p. 134). Children food is seen as a food category that is related to identity and/or age and not to certain eating moments as breakfast or dinner. Marketing has entirely composed the category of children food.

(19)

among children (Edwards & Hartwell, 2002). However, as children classify fruit and vegetables as food for others and junk food as food for them, this is alarming in an era where child obesity is a growing public issue (Elliott, 2011). Nevertheless, this does not mean that children only eat ‘children food’ (Elliott, 2011). For example, from a study of Phillipson & Jones (2008), it appeared that children from 6-12 years are able to understand that fruit and vegetables should be eaten everyday in order to be healthy. They are already aware of what foods are healthy and unhealthy (Smits & Vandebosch, 2012). Based on the above considerations, it can be concluded that marketing cues influences food preferences of children, as well as their purchase requests and actual consumption, which is related to the rising obesity problem among children (Gantz, Schwartz, Angelini & Ridout, 2007).

2.5 Advertising to children

According to Buijzen, Schuurman & Bomhof (2008), most of the food advertisements targeted to children relate to unhealthy foods. Many studies already pointed out the effects of unhealthy food advertising on the consumption patterns of children and to a lesser extent the effect of advertising on the consumption of more healthy foods. Therefore, the following section focuses on a way in which the intake of fruit can be stimulated among children, namely by the use of a cartoon character or celebrity endorser.

2.5.1 Nudging children towards fruit

(20)

From a study of O’Dougherty, Story & Stang (2006) it became clear that there is developed a large range of marketing attempts by producers of unhealthy food products to stimulate children to choose that particular product. Nevertheless, the marketing of fruit to consumers barely left the status quo. From de Droog et al., (2011) it becomes clear that popular marketing techniques (e.g. cartoon characters) could stimulate children’s preference for fruit. As such, cartoon or brand characters are increasingly used to promote the consumption of fruit and vegetables in order to increase the liking of those foods by children (de Droog et al., 2012).

2.5.2 Endorsement by cartoon characters or celebrities

To children, the use of a cartoon character on the marketing of products is highly appealing, since the appearance of the product looks more child-friendly (de Droog et al., 2011). It leads to increased attention to the product (Elliot & Brierley, 2012). Pictures of the character are placed on the packaging of the product, the distribution and on advertisements (Smits & Vandebosch, 2012). Approximately, 75% of child-directed food products use cartoon endorsers in order to appeal to children (de Droog et al., 2012). Also in-store marketing makes heavily use of cartoon endorsement (Chapman, Nicholas, Banovic & Supramaniam, 2006).

(21)

2.5.3 Why does it work: forming of parasocial relationships

When focusing on younger children, de Droog et al. (2011) identified two levels in which cartoon characters are attractive for them. On the first level, mainly a perceptual level, they state that young children, who have cognitive limitations, like cartoon characters due to their appearance. As such, they like characters due to the fact that they are mostly funny and are nicely colored. There is only focused on the most notable attributes owing to the fact that young children only have the competence to assess stimulus objects on a maximum of two dimensions. Therefore, they are not able to comprehend abstract or multiple stimuli. As a result, the things that are salient for them are simple.

The second level de Droog et al. (2011) identified is an emotional level. Children become more independent from their parents when they reach the age of 2 years. From that age they begin to admit and accept other persons, which can be humans as well as characters they see on television, on the Internet or in books. According to Hoffner (1996), special relationships can be developed with those characters, termed as ‘parasocial interaction:’ the feeling of a sincere social relationship with a character or personality encountered in the media. De Droog et al. (2012) state that when children are exposed to media for a certain period they learn more about the behavior and personality of the character. Furthermore, they learn about the characters through social, media and marketing environments, in which parents, television, food packages and movies have an enormous influence (Kraak & Story, 2015). From literature, it is stated that parasocial relationships share many characteristics with real relationships, which leads to the development of intense emotions for characters by children (de Droog et al., 2012). As a result, the character is perceived as a near friend and they have the desire to enter the social world of the character. By collecting products that are related to the character, children think they can achieve the goal to become part of the character’s world. The higher the presence of the character is the environment of the child, the higher the influence it has on the consumption pattern of the child (de Droog et al., 2011). When children have a high emotional appeal to a character, it is said that both the impact on the liking of the endorsed product is higher as well as the pester power, since children are more sensitive to information originated from friends or idols rather than from unknown persons (de Droog et al., 2011).

(22)

assumed resemblance between endorser and receiver, familiarity results from increased knowledge from the endorser by means of frequent exposure and likeability develops when affective feelings are developed for the endorser as a result of their presence or behavior (Byrne et al., 2003). As a result, the receiver wants to develop a relationship with the endorser and imitates the endorser.

To summarize, affective attachments are developed by both children and adults when they encounter a familiar character or personality, which may lead to identification with the character or personality (Hoffner, 1996). One can say that a person develops the desire to be like or to imitate the behavior of the character or personality. From a study of Fernie (1981) it can be concluded that boys want to be like certain characters or personalities due to the character’s physical or social abilities. A way to achieve this is explained by the theory of parasocial relationships. From this theory it is stated that positive elaborate affective responses by children are evoked when a familiar character is combined with a product (de Droog et al., 2012).

2.5.4 Why does it work: an alternative explanation – theory of processing fluency

(23)

Children develop mental schemas about cartoon characters or celebrity or celebrity endorsers when they frequently encounter those characters or persons (e.g. in the media). As such, they become familiar for them. Processing fluency is increased when they encounter the character again due to the developed mental schemas. Children that see an advertisement or product that is accompanied by a familiar character will evoke positive automatic affective responses to the product due to the ease of processing the character (de Droog et al., 2012). Therefore, according to the hedonic fluency model, an increased ease of processing a familiar character triggers the experience of a pleasant feeling. This positive feeling is used as information when the stimulus (product that is endorsed) is evaluated (Fennis & Stroebe, 2016).

2.5.5 Cartoon character / celebrity endorsement of food products

(24)

2.5.6 Use of claims

Another powerful mechanism to induce children’s to choose a particular food is by the use of claims, e.g. ‘this food is good for your health,’ (Dixon et al., 2013). From the study of Dixon et al. (2013) it is concluded that pre-adolescents are tended to choose energy-dense and nutrient-poor products when nutrient content claims were made. This is in contrast to many studies advocating that nutrient content claims are informational appeals that are targeted to adults. However, a reason why it also works well with children can be explained by the fact that children interpret nutritional information as facts, since they have a lack of skepticism to marketing techniques (Wilcox et al., 2004) (Miller et al., 2011). As such, this can enhance preference for a particular product (Dixon et al, 2013). Especially for children from the age of 8 years, it can work positively since they have an understanding of what healthy eating means and are able to relate this to what they should eat (Edwards & Hartwell, 2002) (Zeinstra et al., 2007). Furthermore, children in this age have the competence to use a well-thought evaluation process and have the competence to look for useful information aiming to make a decision, which means that they able to comprehend and to act on claims on packages or in advertisements (Miller et al., 2011). Especially for children interested in health, such claims may work effectively since the information is easily accessible, which has a positive effect on information usage (Miller et al., 2011). However, also contrary findings are present as founded by the majority of studies, meaning that claims or statements on packages or advertisements leads to unhealthier choices among children, since children interpret a nutritional claim about a health product as negative: the claim leads to a belief that healthy food is not tasteful, which makes it less attractive. It was found that this was true for all types of nutritional claims: nutrient-content claims, health claims and general claims (Miller, 2011). Besides that, children perceive claims as not relevant for them, since they generally have a low health motivation. Therefore, Miller et al. (2011) recommend the need for external environmental cues that emphasize the essence of health at the point of purchase.

(25)

influencing the behavior of children (Phillipson & Jones, 2008), which could be explained by the fact that children are likely to identify themselves with the sports star (assuming that the live of the sports star is transferable and attainable for themselves). As such, more confidence is assigned to the sports star than optimal judgment would allow (Ross et al., 1984). Therefore, the sports star could attribute positive attitudes to the product they endorse, in which children believe that possession or use of the product lead to advantages for themselves (Ross et al., 1984). However, there must be some association between the endorser and the product, in order to have more effect. If there is some association present in the mind of children between endorser and the product, the endorser is seen as a more reliable expert compared to those who do not have a link with the product (Ross et al., 1984). From a study by Phillipson & Jones (2008), it appeared that famous sportspeople are perceived by children and teenagers to have a healthy lifestyle and to eat healthy foods as fruit and vegetables. As such, it can be assumed that there is some link between healthy food and sports stars, which could make them credible endorsers for healthy food products. It is already proved that sports stars promoting healthy foods stimulate the health of children in a positive way, as children assess the merits of some food or products based on authority of adults (Philipson & Jones, 2008). Based on a study by Cialdini, Kallgren & Reno (1991) about norm activation, it can be said that norms are more readily activated if there are also signs present that other people show behavior that is in congruence to the norm. As such, when a claim (representing a sort of norm, e.g. eating more healthy food) is combined with a celebrity endorser, an explicit link is made between the norm (e.g. eating more healthy food) and the endorser, which is more likely to lead to desired behavior by children, namely following the norm of eating more fruit (‘if my idol it says it is good, it must be good’).

2.6 The strength of strong brand names

(26)

From the age of 8 years, children already have a substantial understanding of the brands that are important for them (Lapierre, 2015). From this age, they have brand symbolism understanding, meaning that they have a comprehension of the meaning that is ascribed to a brand name. As such, they have an understanding of the manner in which user qualities and information about products accompanied by a brand are symbolized by a brand name (McAlister & Cornwell, 2010).

A study by Robinson, Borzekowski, Matheson and Kraemer (2007) evidenced that the preference of children for milk and carrots increased when the logo of McDonald’s was depicted on the package. Levin & Levin (2010) explain this finding by using balance theory. A theory stating that people have a preference to have balance, meaning that people have a tendency to perceive positively those attitude objects that are connected to something that is associated with a positive feeling. Therefore, when a positive feeling to a brand name moves to a product accompanying the brand name, balance is accomplished (Levin & Levin, 2010). From the study of Levin & Levin (2010) it appeared that the stimulation of healthy eating among children is better achieved when a product is accompanied by a familiar brand name (from a food brand) compared to an unfamiliar brand name.

3. Hypothesis development

As pointed out in the previous section, nudging children towards healthier food choice can be reached by making use of cartoon or celebrity endorsers. As the focus of this study is on sports canteens, in this study it is tested if the use of a sports star as celebrity endorser on a fruit bowl can trigger children in their choice to eat fruit, rather than more unhealthy food in a sports canteen.

3.1. In commission by the KNVB

(27)

3.2 The intervention

In this study, it is investigated how the consumption of unprocessed and unpackaged healthy food (fruits) can be best promoted among children in soccer canteens by manipulating the design (appearance) of a fruit bowl. From literature it became clear that the use of a cartoon or celebrity character (e.g. in the form of a sports star) could work effectively to trigger children to consume more healthy food as fruits. Also strong brand names can work effectively to achieve this goal. All of these insights taken together, this study will make use of four different conditions to stimulate children to eat more fruit in soccer canteens: a branded fruit bowl, a branded fruit bowl with a famous soccer player, a KNVB-branded fruit bowl with a claim about the product and a KNVB-KNVB-branded fruit bowl with famous soccer player making a claim about the product he endorses. The KNVB is considered to be a strong, well-liked brand. On recommendation of the KNVB, Wesley Sneijder is used as the endorser since he is a popular soccer player among children. Also a neutral condition is used: a neutral / regular (non-branded) fruit bowl, in order to investigate the attitudes, intentions and sales under a neutral (often cheaper) situation.

3.3 Hypotheses

Based on prior theory and research that strong brands, cartoon or celebrity characters can work well in order to stimulate the consumption of fruits among children, in this section several hypotheses are developed.

3.3.1 Branded vs. non-branded fruit bowl

Based on balance theory, it is expected that balance is reached when a positive attitude towards a brand name results in a positive attitude towards the product that is accompanied by the brand name. As such, a strong, well-liked brand transfers its likeability to the product that is accompanied by the brand. Therefore, this also leads to a higher purchase intention for healthy foods (Levin & Levin, 2010). Based on the above considerations, the following hypothesis can be derived:

(28)

3.3.2 Role model vs. no role model

Based on the theory of forming parasocial relationships with cartoon or celebrity endorsers and the theory of processing fluency, it can be said that people experience pleasant feelings by encountering a famous person. Subsequently, this positive feeling is transferred to the product that is accompanied by the celebrity endorser (further in this study referred to as: role model). This leads to an increased likeability of the product and to an increased preference to or choice for the product (de Droog et al., 2011) (de Droog et al., 2012) (Smits & Vandebosch, 2012). Therefore, the following hypothesis can be derived:

H2a: A branded fruit bowl with the inclusion of a role model leads to a more positive attitude towards the fruit bowl and to a higher intention to buy fruit from that fruit bowl than a branded fruit bowl without a role model.

3.3.3 The moderating effect of age

Age of a child is expected to have influence on the attitude towards the fruit bowl with role model and as such, influence on purchase intention or consumption of fruit. The liking of a cartoon or celebrity endorser as role model goes hand in hand with age. Younger children have a tension to prefer non-human figures, while older children have a tension to prefer more edgy endorsers (Smits & Vandebosch, 2012). Furthermore, celebrity endorsers are especially liked by teenagers and adolescents in helping them to form their identity (Harris, Brownell & Bargh, 2009). Based on this consideration, it can be said that age is expected to have a moderating effect, in the sense that a higher age is expected to have a more positive attitude towards the fruit bowl with a human celebrity endorser (role model), which also lead to a higher purchase intention for or consumption of fruit, i.e. a higher age leads to a stronger relationship.

H2b: Age strengthens the positive relationship between a fruit bowl with role model and the attitude towards the fruit bowl and intention to buy fruit from the fruit bowl.

3.3.4 Claim vs. no claim

(29)

making a specific claim about a product or food will be very powerful in influencing the behavior of children: a powerful link, delivered by the endorser, is made between the product and its benefits (Phillipson & Jones, 2011). Therefore, when a celebrity endorser is accompanied by a claim, an explicit link is created between the endorser and the norm / desired behavior (to increase the consumption of fruit). Based on Cialdini et al. (1991) is can be said that when a person sees others also show a particular behavior, norms are more readily activated. As such, it is expected that including a claim next to a celebrity endorser leads to a more positive attitude towards the fruit bowl or higher purchase intention for fruit or consumption of fruit, than solely depicting a celebrity endorser on a fruit bowl, since in the case of the latter there lacks an explicit link between the endorser and the desired outcome (i.e. the norm of consuming more fruit). However, when a claim is made without endorser, children are less likely to choose a healthy product since a claim about a healthy product leads to the belief that the food is not tasteful, and which leads to avoidance of healthy products due to the fact that it is identified by the claim as healthy or good (Miller et al., 2011). Furthermore, children perceive themselves as invulnerable, which makes their health and health information less relevant for them (Miller et al., 2011).

From these considerations, the following is hypothesized:

H3a: A branded fruit bowl with a nutritional claim leads to a less positive attitude towards the fruit bowl and a lower intention to buy fruit from the fruit bowl than a branded fruit bowl without a nutritional claim.

H3b: A branded fruit bowl with a role model and a nutritional claim leads to a more positive attitude towards the fruit bowl and a higher intention to buy fruit from the fruit bowl than a branded fruit bowl with a celebrity endorser but without nutritional claim.

3.3.5 The moderating effect of health consciousness

(30)

H2a H2b H3a H3b H4a H4b Role model Claim

Attitude towards the fruit bowl Purchase intention fruit Consumption unhealthy food Consumption of fruit + _ + _ ₊ _ Age + Health-consciousness _ + +

H4a: A high health consciousness weakens the negative relationship between a branded fruit bowl with a claim and attitude towards the fruit bowl and intention to buy fruit from the fruit bowl.

H4b: A high health consciousness strengthens the positive relationship between a branded fruit bowl with a role model making a claim and attitude towards the fruit bowl and intention to buy fruit from the fruit bowl.

3.3.6 Conceptual model

Based on the developed hypotheses, the following conceptual model can be derived (figure 1):

^ Figure 1: Conceptual model of this study

(31)

fruit in the canteen is solely bought by children, which may possibly lead to biases in sales data. For those reasons, it is more interesting to focus in the hypotheses on attitudes towards the fruit bowl and intention to buy fruit from it. According to the theory of reasoned action, behavior (consumption of fruit) is a function of behavioral intention (purchase intention for fruit). Intention, in turn, is determined by a person’s attitude towards something (attitude towards the fruit bowl) (Bagozzi, Wong, Abe & Bergami, 2000). A more positive attitude towards something is expected to have a positive relationship with purchase intention, which in turn is positively related to actual behavior (actually buying the product). As such, attitudes and intentions may be a proper predictor of behavior. Additional analyses about the actual sales of fruit are executed to see if attitudes towards the different conditions and intentions to buy fruit from it are in compliance with actual consumption of fruit.

4. Methodology

In this chapter, there is elaborated on the methodology and data collection of this study. In order to be able to test the hypotheses and to test under which condition attitudes are most positive and the consumption of fruit is highest, a field study is conducted and a survey is executed. More details about the field study, the survey and the way of data collection are provided in this section.

4.1 Research method

4.1.1 Experimental setting: a field experiment

(32)

4.1.2 Study design

Several conditions are tested which differ in how the fruit is presented (i.e. the appearance of a fruit bowl). The fruit bowls are designed by a professional design agency. This study consists of four different treatment conditions: a branded fruit bowl (KNVB branded: a fruit bowl depicting the KNVB logo), a branded fruit bowl with a nutritional claim (KNVB branded with claim ‘Eating fruit gives energy and leads to a better recovery after a match or training,’) a branded fruit bowl with celebrity/sports endorser (role model) (KNVB branded with Wesley Sneijder from the Dutch National Soccer Team) and lastly, a branded fruit bowl with role model and with a nutritional claim (KNVB branded with Wesley Sneijder with the claim ‘Eating fruit gives energy and helps to recover after a match or training’). There is chosen for the nutritional claim ‘Eating fruit gives energy and leads to a better recovery after a match or training,’ since fruit contributes to a human’s energy-supply and recovery of the human body (Voedingscentrum, 2016). Additionally, in order to be able to investigate the effect of a neutral fruit bowl, there is also a neutral condition: a neutral, regular (non-branded) fruit bowl presenting the fruit. For a picture of the different fruit bowls in the canteen: see appendix 1.

The four manipulated (branded) conditions are included in a between participants factorial design where the factors are role model vs. no role model, claim vs. no claim. As such, the effects of multiple independent variables are measured at various levels and there are interactions between the variables (Malhotra, 2010). Every participant participates in only one condition.

All of the conditions are executed in two different canteens. To convince canteens to participate, they were offered each ten tickets for a match of the National Dutch Soccer Team: The Netherlands vs. Italy on the 28th of March 2017. However, due to some recruitment problems, the neutral fruit bowl condition and the KNVB-branded fruit bowl condition, each is executed in only one canteen. As such, a total of eight different canteens participated, located in the province of Groningen. All of the soccer canteens were not selling fruit before the intervention. See appendix 2 for the canteens and the assigned treatment condition.

(33)

the selling of fruit in the canteen. Each canteen sold bananas, apples and tangerines. The canteens were strictly instructed how to place the fruit bowl and how to present the fruit in order to increase the reliability of the study, to control for extraneous influences as much as possible and to avoid biases (Blumberg, Cooper & Schindler, 2011). Furthermore, to exert some more control, a whatsapp-group was created in which each canteen is being member of. On every Saturday, the canteen was asked to take a picture of the setting (i.e. the fruit bowl on the bar) to share into the group. If needed, corrections were given in how to position the fruit bowl. Another way is which the reliability of this study is guaranteed is by executing the intervention on several Saturdays rather than one day (if possible) and by using (if possible) two canteens per conditions (van Aken, Berends & van der Bij, 2012). An additional advantage of an experimental design is random assignment, meaning that every participant has an equal likelihood to be assigned to one particular treatment of the experiment. To put it differently: they cannot choose in which treatment condition they want to participate (Aronson et al., 1998).

The survey is targeted to children in the age span of 7-14 years old visiting the canteen on either the 26th of November or the 3rd of December. Due to match schedules, for some canteens it was better to distribute and collect the surveys on the 26th of November, and for others it was better on the 3rd of December. Besides that, sales data is collected during all of the three intervention days until 1 o’clock in the afternoon (after that time it is expected that children have left the canteen, since their matches are in the morning). These data are used to examine the percentage of all sales contributed by fruits.

4.2 Data collection

4.2.1 Survey

A survey (in Dutch) is created in order to get the majority of the needed data. The survey is used to measure the control variables, the dependent variables ‘attitude towards the fruit bowl’ and ‘purchase intention,’ and the moderators ‘health consciousness’ and ‘age’

(34)

represents a very negative feeling or opinion and a very happy smiley represents a very positive feeling or opinion. It is emphasized that there are no good or wrong answers and that filling in the survey takes not much time. Furthermore, to guarantee the respondent has seen the fruit bowl, the student that provided the survey, made the respondent aware of the fruit bowl on the bar. Also in the questionnaire, there is referred to the fruit bowl on the bar.

After permission by parents or by people responsible for the child, the survey was filled in by both children that participated in a soccer match as well as children who did not. Children did fill in the survey by themselves and if needed support was provided by the student or by parents. The survey (both in Dutch and English) can be found appendix 3.

4.2.2 Sales data

Also the sales of goods sold in the canteen are measured during the three Saturdays of the intervention until 1 o’clock in the afternoon. The collection of this sales data was done by the canteens themselves. On beforehand, a sales list was created with pre-determined categories of food and drinks. The canteen staff was asked to add a dash (‘turfje’) for every good sold to the corresponding category. As such, after 1 o’clock in the afternoon, the canteen staff could count every dash to determine the amount of goods sold per category. The sales list was based on a list created by marketing research agency GFK about food categories frequently sold in sports canteens (GFK, 2016): see appendix 4.

4.3 Measurements

4.3.1 Independent variables – the experimental conditions

(35)

4.3.2 Dependent variable – attitude towards the fruit bowl

In order to measure the attitude towards the fruit bowl, a 5-point likert scale is used, where the respondent is asked to give a rating on the question ‘what do you think of the fruit bowl with fruit?’ A rating of 1 represents a very negative attitude ‘very stupid’, while a rating of 5 represents a very positive attitude ‘very nice.’ The measurement is based on de Droog et al., (2011: 85), who measured ‘liking of a snack’ with the following question ‘how much do you like this snack?’ With this question, the input for all of the hypotheses is obtained.

4.3.3 Moderators – age and health consciousness

The first moderator, age, is measured by an open question asking how old the child is in years. It is used to provide input for hypothesis 2b. In order to obtain insights in how much importance a child attaches to a healthy lifestyle, the moderator ‘health consciousness’ is created. Items to measure this construct are slightly adapted from May & Hoffmann (2012: 326). In their study, they measured the health consciousness of consumers by asking respondents to give a rating on the following statements: ‘I reflect about my health a lot,’ ‘I am very self conscious about my health,’ ‘I am generally attentive to my inner feelings about my health’ and ‘I am constantly examining my health.’ For the aim of this study, the items are slightly adapted in order to be more appealing to children. The last item ‘I am constantly examining my health’ is skipped, since this seems not very relevant for children. As such, the following items are used to measure ‘health consciousness’ of children: ‘To me, to live healthy is important,’ ‘I try to live healthy,’ and ‘I spend a lot to healthfulness.’ The children were asked to rate these statements on a 5-point likert scale with a rating of 1 representing ‘totally disagree’ and 5 representing ‘totally agree.’ Since children from the age of 7 have an understanding of the term health (Edwards & Hartwell, 2002) (Phillipson & Jones, 2008), it is expected they are able to answer these questions. This question is used to provide input for hypotheses 4a and 4b.

4.3.4 Dependent variable – purchase intention for fruit

(36)

request intent’ using the following question: ‘Would you pick out this one the next time you are in the supermarket?’ With this question, the input for all of the hypotheses is obtained.

4.3.5 Dependent variables – actual consumption of fruit (and other food items)

The dependent variables ‘actual consumption of fruit’ and ‘actual consumption of unhealthy food’ are measured using the sales data from the three Saturdays. An important assumption made here is that purchase of food by a child, is consumed by the child in the canteen. As such, consumption is equivalent to sales. The actual consumption of a food category (e.g. fruit) (for categories see appendix 4) is measured by percentage of total sales contributed by that particular category (e.g. fruit). To overcome biases due to size of the canteen, percentages are used rather than absolute quantities.

4.3.6 Control variables

Also control variables are included in the study, namely the age and gender of the respondent and attitude towards the role model (used at the fruit bowl at some conditions), attitude towards the claim (used at the fruit bowl in some conditions), attitude towards healthy food and drinks in the canteen and attitude towards fruit. All of these variables, except gender, are measured on a likert scale, where 1 represents a very negative attitude and 5 represents a very positive attitude.

4.3.7 Background variables

Also some background questions are included in the survey, which may be useful in order to get more insights into the consumption patterns in the canteens. For example, what food and drinks the child bought today in the canteen (if any) and if they chose it by themselves or by their parents. It is expected that socially desired answers are minimal, since actual consumption is asked rather than what is normally bought in the canteen. It is expected that this latter option result in more socially desirable answers. Also there is asked for reasons for (not) buying fruit.

4.4 Plan of analysis

(37)

since attitudes are a strong predictor of intentions (MacInnis et al., 2013). If this is the case, they are combined into one variable attitude – intentions.

In order to be able to test hypothesis 1, an independent samples t-test seems most appropriate, since there is one independent variable with two levels / populations (the neutral, non-branded fruit bowl and the branded fruit bowl). (Malhotra, 2010). The dependent variables ‘attitude towards the fruit bowl’ and ‘intention to buy fruit from the fruit bowl’ are ordinal scaled. However, likert scales can be treated as interval data (Malhotra, 2010). When including control variables, an ANOVA must be performed to be able to include control variables in the model. When ‘attitude towards the fruit bowl’ and ‘intention to buy fruit from the fruit bowl’ do not correlate significantly, separate independent samples t-test are performed with the dependent variables for the model without control variables, for the model with control variables, a Multivariate Analysis of Variance (MANOVA) is performed, since then there are two dependent variables.

To test hypothesis 2a, 2b, 3a, 3b, 4a and 4b, an Univariate Analysis of Variance is used (when ‘attitude towards the fruit bowl’ and ‘intention to buy fruit from the fruit bowl’ correlate significantly and can be treated as a combined variables), since there are multiple independent factorial independent variables (celebrity endorser (role model) vs. no celebrity endorser, claim vs. no claim, lower age vs. higher age and lower health consciousness vs. higher health consciousness) and the dependent variable, the combination of ‘attitude towards the fruit bowl’ and ‘intention to buy fruit from the fruit bowl) is metric. Besides main effects, also interaction effects are relevant (no role model*claim, role model*claim, role model*age, claim*health consciousness and role model*claim*health consciousness). When the two dependent variables do not correlate significantly and cannot be combined into one dependent variable, a MANOVA is performed to be able to test these hypotheses, since there are two dependent variables then.

However, first for the moderator ‘health consciousness,’ a factor analysis is performed to examine if the three underlying dimensions measuring the ‘health consciousness construct’ measure the same. If they measure the same, a reliability analysis is performed to examine if it is allowed to combine the three dimensions into a single dimension. If it is allowed, they are combined in one variable: ‘health consciousness.’

(38)

5. Analyses and results

This chapter presents the results of this study. All of the analyses are executed with SPSS. The surveys are manually processed in SPSS. In order to get a valid dataset to analyze the data, the dataset is prepared. This is done in compliance with the steps as prescribed by Leliveld et al. (2012) about preparing a dataset. A first step is labeling the variables. In the questionnaire all possible response items were already coded. There was assigned a number to each response item in case of fixed items questions. To give an example: for the question: ‘would you buy fruit from the fruit bowl (again)?’ a value of 1 was assigned to ‘Definitely not,’ 2= ‘No, not’, 3= ‘I don’t know,’ 4= ‘Yes,’ and 5= ‘Definitely yes.’ A second step it to check for missing or invalid data. In that case, respondents are deleted. Lastly, the assumptions for performing an ANOVA are checked (see section 5.5.1).

5.1 Descriptive statistics

Firstly, some descriptive statistics are provided. More information is given about the respondents in the sample, e.g. about the age and gender. Furthermore, a closer look is given to the consumption of food and drinks in the canteen and reasons for (not) buying fruit.

5.1.1 Sample

STIMULATING FRUIT CONSUMPTION AMONG CHILDREN IN SOCCER CANTEENS

STIMULATING FRUIT CONSUMPTION AMONG CHILDREN

IN SOCCER CANTEENS

The influence of attractiveness of a fruit bowl in stimulating children to eat

more fruit

by

Wouter Luinge

STIMULATING FRUIT CONSUMPTION AMONG CHILDREN

IN SOCCER CANTEENS

The influence of attractiveness of a fruit bowl in stimulating children to eat

more fruit

by

Wouter Luinge

s2183463

Management summary

Preface

Table of contents

1. Introduction

1.1 Background

1.2 Food advertising to children

1.3 Cartoon character or celebrity endorsement

1.4 Problem statement and research questions

1.5 Contribution of this study

1.6 Thesis outline

2. Theoretical framework

2.1 From unhealthy to healthy food marketing to children

2.2 The essence of fruit

2.3 Information processing of children

2.4 Forming of eating habits of children

2.5 Advertising to children

2.6 The strength of strong brand names

3. Hypothesis development

3.1. In commission by the KNVB

3.2 The intervention

3.3 Hypotheses

4. Methodology

4.1 Research method

4.2 Data collection

4.3 Measurements

4.4 Plan of analysis

5. Analyses and results

5.1 Descriptive statistics