1
Eating behavior in canteens of soccer clubs in
The Netherlands
How to stimulate healthy eating habits in canteens?
2
Eating behavior in canteens of soccer clubs in
The Netherlands
How to stimulate healthy eating habits in canteens?
Camiel Gerlach Department of Marketing Master thesis 30-06-2017 Korreweg 183b 9714AK Groningen 0681077800 c.c.gerlach@student.rug.nl S2376903
Supervisor: prof. dr. ir. K. Van Ittersum External supervisor: M. T. Van der Heide
3
Abstract
Obesity and overweight are worldwide problems and it also concerns to children in The Netherlands. There are two main ways to tackle overweight: more healthy eating habits and more physical
activity, often at a sports club. Soccer is in The Netherlands the most popular sport, also during childhood. Opposite to logic, the canteen of a soccer club is especially a place where unhealthy snacks and drinks are served to these children. There is often even no healthy food available. In The Netherlands, there is more and more consciousness about this strange fact; it is more or less the world upside down. However, not many clubs are willing to change its assortment, because they are afraid of a lack of success (i.e. decreasing profits and increasing efforts). Not much research has done to find out how children can be attracted to eat more healthy in the canteen of their club. When that is clear, most of the arguments clubs have to do not change the assortment, are refuted. This paper describes a field experiment that is performed to find out how to serve healthy food effectively to children. The use of celebrity endorsers and fun elements are marketing techniques that are often used to children, but it is never investigated in a setting as in a soccer club canteen. The results of a logistic regression suggest that both techniques are not effective to let children eat more healthy in such a setting. Several reasons for the insignificant relations are discussed in the last part of this paper.
4
Preface
I wrote this paper to graduate at the University of Groningen (Rijksuniversiteit Groningen, RUG), at the study Msc Marketing Management.
Special thanks to prof. dr. ir. K. (Koert) van Ittersum, professor of Marketing and Consumer Well-being at the University of Groningen. He guided me through the process of the Master thesis, gave feedback on several moments and made suggestions for the research set-up. He is the first
supervisor in this project.
The second person I want to thank is M. T. (Martine) Van der Heide, PhD Student at the University of Groningen. She also supported me with feedback, made suggestions through the process and is the second supervisor.
At last I want to thank the University of Groningen for the opportunity I received to graduate at Msc Marketing Management, after I finished the Bsc program Bedrijfskunde.
5
Table of Contents
1. INTRODUCTION ... 6 2. THEORETICAL BACKGROUND ... 9 2.1 Food advertising ... 10 2.1.1. Fun elements ... 10 2.1.2. Celebrity endorsement ... 12 2.2 Eating habits ... 14 3. METHODOLOGY ... 173.1 The experiment and design ... 17
3.1.1. Conditions ... 20
3.2 Validity ... 21
3.3 Control variables ... 21
3.3.1 Age ... 22
3.3.2 Liking the products ... 22
4. RESULTS ... 23
4.1 Liking the celebrity ... 26
5. DISCUSSION, LIMITATIONS AND FURTHER RESEARCH ... 279
5.1 Gender ... 299
5.2 Influence of the circumstances ... 30
5.2.1 Peer influence ... 3131
5.2.2 Chaotic environment ... 3131
5.3 Exposure to the conditions ... 3232
5.3.1 Celebrity endorsement ... 3232
5.3.2 Fun elements ... 3232
5.4 Healthy eating habits ... 33
6. REFERENCES ... 34
6
1. INTRODUCTION
Overweight and obesity in children is recognized as a worldwide problem (Livingstone, 2001). Much research shows that many children’s diets in the developed world are not as healthy as it should be (Brown et al, 2004). Also in The Netherlands the prevalence of obesity among children increased in the last decades. Worse still, it doubled or even tripled from 1980 to 1997, and again from 1997 to 2004 a two-or threefold increase was seen for almost all ages (Schokker et al, 2005). In 2015, over 12% of the children in The Netherlands from 4 to 16 years old were overweight (Volkgsgezondheid en Zorg, 2015).
While obesity in adulthood is associated with increased morbidity and mortality, the long-term effects associated with childhood overweight are less clear (Livingstone, 2001). That is because overweight in children does not necessarily lead to overweight in adulthood. The older a child with overweight or obesity is, the better it is a predictor for adult obesity (Serdula et al, 1993) and the related health issues. About a third of obese preschool children were obese as adults and about half of obese school-age children were obese as adults (Serdula et al, 1993). It is difficult to predict the development of a child’s health related to overweight in the future, because natural age-related physiological variations in body compositing during childhood make it difficult to distinguish between normal and excessive adiposity (Livingstone, 2001). While obesity is a well-established risk factor for chronic diseases such as heart disease, hypertension, stroke or diabetes in adults (Power et al, 1997), it is known that obesity in childhood also is associated with, for example, premature death from endogenous (self-inflicted injury) causes (Franks et al, 2010). It is important to pay attention to overweight related problems in childhood, because once confronted with overweight or obesity in adulthood, the change to achieve a healthy weight – and to maintain that weight – is very low (Stunkard and Penick, 1979).
There are two important determinants of overweight and related chronic diseases: unhealthy eating habits and physical inactivity (Centers for Disease Control and Prevention, 1997). In that case, children should eat more healthy and be physical active to decrease overweight and obese during childhood, but also in their adult life.
7
that “promotes over-consumption of calorie-dense, nutrient-poor foods” (Elliott, 2009). This seems to be the case in canteens of sport clubs too, which indicates that the revenues are more important to sport clubs than the healthiness of children is.
From the age of six, children can understand the consequences of healthy eating and start caring about which food is good for them (Atik and Ertekin, 2011; O’Dea, 2003). Obviously, that does not mean that they are not sensitive to advertising techniques. The consumption of unhealthy food often happens despite better knowledge. This is the case for both children and adults (Von Normann, 2009), and there are several techniques that marketers use to attract consumers to their food. Numerous studies from around the world have shown that the majority of adverts broadcast are related to food that can be described as unhealthy (Batada et al, 2008; Batrinou and Kanellou, 2009; Chapman et al, 2006; Jenkin et al, 2009; Kelly et al, 2010). In this research, we define food as
unhealthy when it is both energy dense and low in nutrients.
Advertisers use particular persuasive techniques to appeal to children, and one of that techniques is the use of celebrity endorsement (Boyland and Halford, 2013). Celebrity endorsement has become a popular approach in advertising in terms of gaining and keeping attention, and in creating favorable associations with the particular product or brand (Carrol, 2008). Celebrity endorsement is recognized as a potentially potent tool in communications, with celebrities viewed as more powerful than anonymous models (Carroll, 2008) and this technique is also often used with unhealthy products. The use of a celebrity to promote unhealthy food seems to be an effective way to attract children, provided that there is a match between the celebrity, the food and the target audience. The question that arises is why this technique is not used to promote healthy food, with the alarming development of increasing overweight in childhood in mind.
There has been a lot of research to the subject of eating habits and factors that may influence this besides marketing techniques, also with a focus on childhood. Nonetheless, findings of different studies about a particular factor that influences eating behavior sometimes contradict each other. Some of the inconsistencies in past findings might have to do with the foods being studied (Wansink et al, 2004).
8
from home, and then it is difficult to control their eating behavior. For example, children are an important target group for fast-food restaurant and these restaurants are clustering around schools, just to make it easier for youth to visit them (Austin et al, 2005). This is just one example of the easy accessibility of unhealthy food, when children are away from their home situation.
It is interesting to investigate if the availability of healthy food at home, which may be seen as an indicator of the attitude of parents toward healthy eating, is not only a determinant of healthy food habits at home, but if it also contributes to healthy eating habits away from home.
In this paper, the focus is on the food part of a healthy life and not on the physical activity. The research question is the following: How can healthy eating habits in a sports canteen during childhood be stimulated? To study this, we set up a field experiment in the canteen of a (amateur-) soccer club. We offer children unhealthy products and more healthy alternatives in different conditions, where we try to find ways to stimulate them to choose the healthy options over the unhealthy alternatives.
9
2. THEORETICAL BACKGROUND
Caterers have been operating within a commercial context for a long time and therefore, they only provide the foods that they know that children will eat (Noble et al, 2000). This seems also be the case in canteens of sport clubs, because the board of the clubs are almost all convinced that there is no demand for healthy food and this leads to an overload of fried food offers in canteens (Van Kalmthout, 2015). With their idea about offering more healthy food, the board does not make an exception for children.
As mentioned, from the age of six, children can understand the consequences of healthy eating and start caring about which food is good for them (Atik and Ertekin, 2011; O’Dea, 2003). The question is to what extend they really understand what they eat, or what the explanation is of nutrition content messages. It seems that they have a limited view on the meaning of healthy and unhealthy food. Some children perceived healthiness as the absence of fat, while others saw it in terms of the presence of vitamins (Noble et al, 2000), while that is a perspective which is probably too narrow. When you want to eat healthy, avoiding fat or intake of some vitamins is not sufficient. And besides that, just because children can understand the consequences of (un)healthy eating, that does not mean that they behave according to that knowledge. The consumption of unhealthy food often happens despite better knowledge (Von Normann, 2009).
According to a number of studies (Noble et al, 2000), children’s food preferences have found a wide range of personal preferences, as well as regional and gender differences. However, foods such as chips, sausages, burgers and pizza were commonly mentioned as favorites. Besides that, foods which were freshly prepared rather than pre-cooked and reheated were better liked (Noble et al, 2000). It is necessary to mention that the findings were related to school-meals and not to snacks or meals in a sport canteen, but it is imaginable that this corresponds to a certain degree. According to a
research of GFK (2014), grilled cheese sandwiches, bread with snacks on it and fried snacks were the most consumed food in 2014. Besides that, visiting sport canteens in The Netherlands often happens in the weekend – especially in childhood – and during the weekend people, including children, eat different from weekdays (Thompson et al, 1986). In that case, it is even probable that the favorability of unhealthy food in sport canteens will be stronger. Around the whole world people tend to eat less healthy compared to the work-week (Haines et al, 2003; Chai et al, 2008).
Research has shown that food consumption patterns of children’s are predicted by their food preferences (Skinner et al, 2002), which seems to be logic. Hence, understanding children’s food preferences and the factors related to this preferences can be critical in the fight against
10
fat food over healthy food (Noble et al, 2000) and almost all the foods promoted in advertisements are typically the opposite of dietary recommendations (Hastings et al, 2003; Powell et al, 2007b). It is known that food promotion influences food preferences and consumption patterns (Cairns et al, 2009). In that way, it is interesting to investigate if and how advertising techniques can have influence on the food preferences of children.
2.1 Food advertising
Advertising and promotion of food to children is possible through several types of marketing channels, especially: television, in-school marketing, product placements, kids clubs, the Internet, toys, products with brand logos, and youth-targeted promotions, such as cross-selling and tie-ins (Story et al, 2004). Around the whole world, children are exposed to high volumes of food advertising on television and most of that advertisements has content that is the opposite from general
nutritional recommendations (Kelly et al, 2010). The endless promotion of junk foods and the food-related media messages are seen as two of the main ‘suspects’ when it comes to overweight and obesity (Elliott, 2008). There has been longstanding public concern over the potentially harmful effects of food promotion to children (Livingstone et al, 2006). Food advertising on television is dominated by breakfast cereals, confectionery, savory snacks, soft drinks (Hastings et al, 2003). Fast-food restaurants are increasing their part of television advertising for Fast-foods high in fat, sugar and salt (Hastings et al, 2003), so the domination of these unhealthy foods only increased. In the last years, it seems that this barely changed. Besides that, also the Internet (online marketing) plays an important role in the marketing of food to children (Moore et al, 2007), and the role of online marketing to acquire consumers is still increasing (Kireyev et al, 2016). Therefore, watching a lot television and be online active is not only a negative determinant of children’s physical activity, it also increases the exposure to food that is marked as unhealthy. At a local sports club both of them can be countered when healthy food intake becomes more ordinary than nowadays.
It is clear that changing eating habits of children at sports clubs is very hard. To achieve this, it is interesting to investigate the possibility of healthy food intake, without the children being aware of it. One of the major reasons for food marketers to focus on youth, is the purchasing influence we have on children (Story et al, 2004), so why isn’t this possible in the canteen of a sports club?
2.1.1. Fun elements
11
the development of ice cream as a typical example; flavors are no longer limited to vanilla, chocolate or strawberry, but have exciting names and related colors. Food is more and more positioned as “fun” and eating as “entertainment” (Elliott, 2009). This can also be used to make healthy food more attractive, but obviously it is not possible to make all the healthy food blue colored.
Another option than changing the food itself, is to make things that are related to particular food more fun. Elliott (2008) suggests that foods can become more fun through (1) direct claims or allusions to fun and play on the package, (2) the use of cartoon iconography, often supported by funky lettering, pointedly directed toward children, (3) tie-ins with children’s television programs, merchandise, or films (e.g. SpongeBob Squarepants, Barbie), (4) puzzles, competitions, or games unambiguously targeted at children; or (5) the foregrounding of strange shapes, unusual colors, and unconventional product names or flavors.
As one can see, there are different possibilities to make foods to fun foods, without changing the food itself. The packaging of products is an often used way to increase the fun perception of food. Unfortunately, most healthy food packages tends to communicate only to adults (Pires et al, 2011) and are not related to fun or entertainment. According to Young (2004), especially the purchase decision of children is influenced by packaging, more than in adulthood.
12
children decide about this under low involvement. I expect that posters with cues that promote healthy food can influence children’s purchase decision, without being aware of it. As mentioned, children are relatively easy influenced at food purchase decisions and, besides that, this decision occurs typically in low involvement.
H1: When healthy food is presented with fun elements, children in sport canteens tend to choose the healthy food over an unhealthy alternative.
2.1.2. Celebrity endorsement
The use of celebrities as part of marketing communications strategy is a common practice for major firms (Erdogan, 1999) and it is considered an effective promotional tool by marketers worldwide (Spry et al, 2009). That is mainly because celebrities can generate positive feelings from consumers and this should lead to affectively favorable responses toward brands (Till et al, 2008). Moreover, children naturally focus their attention on techniques like these, and emotional appeals do distract children from other aspects of adverts, so that for example nutritional disclaimers do not have a lot effect (Wicks et al, 2009). So when used to promote something like food, it is not surprisingly that it are almost always unhealthy products (Dixon et al, 2011).
When marketers use celebrity endorsement, popular and famous celebrities are either depicted on the product itself or are shown when using and approving the product. Children are targeted by this way by both television commercials and advertisements on the Internet (Calvert, 2008). These celebrities can be musician, athletes, actors, and so on. Even politicians are sometimes depicted on products or can play a role in a commercial. According to Misra et al (1990), it is very important that there is a certain relation between the celebrity and the brand or product. They call it the match-up between those two. For instance, Roger Federer is probably a more appreciate endorser for
13
can stimulate their fantasies about ambience, feelings and advantages to themselves that would result from using or consuming a product (Ross et al, 1984). Highlighting a national or local celebrity endorser can increase the fit with customers and can boost evaluations positively (Spry et al, 2011). Next to possible benefits of celebrity endorsement in marketing activities, there are, of course, also possible backwards. Erdogan (1999) named some benefits and related backwards in his review on this subject: Celebrity endorsement leads to increased attention, but may overshadow the brand. It supports polishing your image, but it can lead to public controversy. It can help with a successful brand introduction and/or repositioning, but there is also the change of an unintended change in image, loss of public recognition, and overexposure. There are several tactics to maximize the benefits and prevent the possible negative consequences described, but due to the context of this study, it is not relevant to discuss that in detail.
During this study, I expect that children will eat more healthy when the healthy is promoted by a famous Dutch soccer player. The manipulation of a product endorser had a greater impact under low than high involvement (Petty et al, 1983), and I earlier explained why one can assume that the children are low involved in this decision. I expect that it is not necessary to hold in mind that earlier research is mainly based on television advertising, because also when the advertisement is written (and not vividly seen), the attitude a potential consumer has towards the brand or product is partly predicted by the attitude toward the endorser (Silvera et al, 2003). When in the right settings, celebrity endorsement can positively influence children in their judgment about the credibility and expertise of the presenter and it can stimulate their fantasies and anticipations about ambience, feelings, and advantages to themselves that would result from using or consuming a product (Ross et al, 1984).
H2a: When healthy food is presented with a famous soccer player, children in sport canteens tend to choose the healthy food over an unhealthy alternative
14
H2b: Liking the celebrity strengthens the effect that presenting him with healthy food has on the preference of healthy food.
2.2 Eating habits
Most of the children, indicate that they eat what is available and allowable at home, at school, and at friends’ homes (O’Dea, 2003). This may suggest that unhealthy eating habits – and overweight or obesity as one of the consequences – are relatively easy to fight. This is supported by Powell et al (2007a), who suggest that the environment is playing an increasingly role in influencing people’s lifestyles and risks for developing obesity. The purchase decisions of parents may influence this a lot. If there is (almost) no unhealthy food at home to eat, children will automatically eat much more healthy. But it is not that simple. According to Lee (2012) it is also important what is available in the supermarket and what kind of restaurants are existing in the neighborhood. That is also the case for the availability of food in sport canteens. Children make purchase decisions by themselves, and besides that, just the presence of unhealthy options in the environment make children eat less healthy, although they do not eat that particular food.
It is not surprisingly that the parents have a key role to encourage healthy eating for children at an early age (Benton, 2003), but they do not always act as they should be. Parents may inadvertently promote excess weight gain in childhood by using inappropriate child-feeding behaviors (Clark et al, 2007). It suggests that it is difficult to act in the right way in this subject. Child-feeding practices that control what and how much children eat can affect their food preferences (Birch et al, 1997). An often used method of parents to let their children eat healthy is the using a food in the instrumental component of a contingency: “Drink your milk and then you can watch TV” (Birch et al, 1982). Perhaps for that moment that is an efficient way to work, but if it is an effective way to stimulate healthy eating in the long term is questionable. In fact, the results of that research indicate a negative shift in preference for the target food consumed instrumentally but no change in the preference for the related activity (Birch et al, 1982). In other words, the child still enjoys watching TV, but drinking milk is experienced even worse.
15
effects on food acceptance patterns, just as the provision of nutrition information to children (Scaglioni et al, 2008).
It is clear that it is very difficult to find exactly the right method to make children familiar and consciousness about the importance and tastiness of healthy food in such a way that it improves their health. Also setting the right example is not always sufficient way to motivate children eating healthy. In the study of Johannsen et al (2006), there is found no significant relationship between parents’ eating behaviors and children’s weight status. It is even questionable if there is one right way. As Scaglioni et al (2008) claims, it is a complex interplay of innate, familial and environmental factors. Following this, a good way to act for parents is not easy found and it depends on the circumstances.
The influence of parents on eating habits of children, has to be emphasized in this case. Healthy eating and healthy food are generally connected to family and less with friends and other social activities, and therefore, it is strongly recommended parents to introduce children to healthy nutritious food early on (Pires et al, 2011). But there seems to be no ideal way for parents to
motivate children to eat healthy, and every option has backwards. In this research, I try to find out if having healthy eating habits, influences the decision making process of children about food away from home. Sports canteens are places where most of the children buys unhealthy food (Van Kalmthout, 2015). Based on the above, children that have healthy eating habits probably are intrinsically motivated by that and not only by their parents. For that reason, one could expect that children with healthy eating habits will choose the healthy option more often than children with unhealthy eating habits.
H3a: Having healthy eating habits, children tend to choose the healthy food over an unhealthy alternative in sport canteens
H3b: Having healthy eating habits positively influence the effect the celebrity has on food choice in sport canteens
16
17
3. METHODOLOGY
The hypotheses derived from the literature research will be studied in a field experiment. Because it is really important that eating habits in childhood change, I do not investigate the relation with the variables towards attitudes or intentions to eat more healthy. Webb and Sheeran (2006) showed in a Meta-Analysis of Experimental Evidence that a medium-to-large change in intention often only lead to a small-to-medium change in behavior. Besides that, the goal of this study is to find out if children are influenced by celebrity endorsement or fun elements, without being aware of it. For that reason, terms like ‘healthy’ or ‘unhealthy’ are not emphasized or even mentioned during this experiment. In this study, actual behavior is measured.
3.1 The experiment and design
The experiment was conducted in Groningen, a city in the north of the Netherlands. The stand was set up in the canteen of amateur soccer club SV Lycurgus, located on sports complex “Sportpark West End” at the 6th and the 13th of May 2017. The target behavior was the choice between
unhealthy food (chips and coke) or a less unhealthy alternative (salted popcorn and ‘Dubbel Fris apple peach’). In table 1 one can see the content of energy, fat, sugar and salt of the products used in this experiment. Although the ‘healthy choice’ is not very healthy, it is way less unhealthy compared to the other ‘unhealthy choice’. Besides, when a child can choose between an unhealthy option and a healthy choice with, for example, a banana and a glass of water, I expect almost all of them will choose the unhealthy option and as a consequence, the findings of this survey become worthless.
Nutritional values per portion
Unhealthy choice Healthy choice
Coke (0,2l) Chips (40g) Dubbel Fris (0,2l) Salted popcorn (20g)
Energy 84 Kcal 222 Kcal 28 Kcal 65 Kcal
Fat 0,0 g 14 g 0,0 g 3,38 g
Sugar 21,6 g 0,2 g 6,4 g 0 g
Salt < 0,001 g 0,24 g 0,006 g < 0,001 g
Table 1 – content of the food that is used in this experiment
18
2011; O’Dea, 2003). When, for instance, a seven years old child wanted to participate, I did not refuse him or her. However, the data I will use in the next part of this paper is just the data from participants between 8 and 14 years old.
For the children, there were roughly three different ways to come in contact with me: 1) Almost every child within the range of 8 to 14 years old who visited the canteen was asked to participate and almost all reactions were positive. Besides, to prevent that the first condition only participants registered who did not play a match already, I stopped with actively asking them to participate in the early morning. 2) Some other children were participating after walking towards the table to ask what they could do there. 3) The overwhelming majority were attended to the experiment by friends or team mates. To prevent children participated more than once, I put a stamp on their hand as an evidence of participation. Besides their age, they had to fill in the team they played in, so that I can control twice if the participant is between 8 and 14 years old.
The way of participation was not complicated. Children were asked to fill in a questionnaire. The first question was related to their food choice, so that the other questions did not influence their choice. Both options were presented on two big posters on the table and also on the survey they saw the two different posters. Which posters they saw, was depending on the moment of participating, i.e., which condition was presented at that moment. After filling in which food combination they wanted to receive after finalize the questionnaire, they filled in the other questions. As explained, the surveys in each of the four posters were different in lay-out and can be seen in appendices 1A – 1E. Filling in this questionnaire, children reveal the relevant data to relate their behavior in the different conditions to the variables ‘healthy food habits at home’ and ‘liking of celebrity’, what can be related to their choice of food. In table 2 the questions that will provide this data are presented.
Variable Question Data
Healthy food habits at home
Which of the following food products do you consume at home (almost) each day?
Vegetables, soda, chips, water, candy, fruit
Liking of celebrity
What do you think of Arjen Robben? 1 (very stupid) to 5 (like him a lot) on a 5 points Likert scale.
Table 2 – questions on the questionnaire that participants fill in
As showed in the table, the selected celebrity is Arjen Robben. He is a world-famous Dutch soccer player and is currently playing at FC Bayern Munich in Germany and in the Dutch national team. He is seen as the best Dutch player nowadays and he has won a lot of trophies in his career so far.
19
and the celebrity (Misra et al, 1990). Most of the respondents lives in Groningen or in the close neighborhood. In this case, the celebrity is an idol to the respondents. For that reason, I choose him as a spokesperson instead of another celebrity.
Due to the literature described, it is important to know if a respondent likes the celebrity or not (Tantiseneepong et al, 2012). When a respondent does not like a celebrity, (s)he may not act like the way the celebrity suggests is the good way, i.e., the respondent is not triggered to eat the food that is related to the endorser and (s)he even may show contrary behavior. I assume that almost all the respondents like Arjen Robben. To prevent to draw unjustified conclusions, I will measure the degree the respondents individually like him.
During the day, there are different conditions in which the data is conducted. A distinction can be made between the control condition, in which nothing is manipulated, and three different conditions with the celebrity endorsement and/or fun elements. During a particular period the healthy food options are related to the celebrity, then the healthy food is related to fun elements, and after that it is related to both the celebrity and fun elements. The desired amount of respondents is
approximately 35 – 50 each condition, which means that at least 140 respondents are necessary. In this case, the more respondents the better. The study has a 2*2 between subject design, which means that each respondents is only participating in one of the described conditions. The design is showed in table 3:
Celebrity endorsement
Fun elements No Yes
No (1) Posters – Healthy food/unhealthy food
(2) Posters – Healthy food with celebrity endorsement/unhealthy food
Yes (3) Posters – Healthy food with fun elements/unhealthy food
(4) Posters – Healthy food with both fun elements and celebrity endorsement/unhealthy food
Table 3 – design of the study
As said, the experiment was run on two Saturdays, the 6th and the 13th of May. At the 6th of May,
from 9.00 am to approximately 10.00 am the first condition was displayed (38 participants). After that, the second condition was presented, until 11.00 am (35 participants). Then, until 12.00 am, the participants were exposed to the third condition (42 participants) and from 12.00 until the end of the morning (approximately 00.30 pm) the fourth condition was presented (15 participants). I did not determine these time slots before running the experiment, but just waited until there were enough participants (at least 35) who participated and prevented that I changed conditions in busy moments. That may lead to confusion at the participants or a less clear effect. The 13th of May started with the
20
displayed again for about an hour, to collect more data in that condition (11 participants). I ended up with the celebrity condition, to collect data of 11 more participants. After all, 184 participants filled in the questionnaire, from which 164 were used as input in the analysis.
Because of the design of this experiment, it is probable that one or more children were aware that they take part in a survey. However, due to the questionnaire they were not aware of the exact relations I studied. The questionnaire can be found in appendix 2A – 2B (example of the first condition). There were a lot of questions on that were not part of the conceptual model. Some of these data are used for control.
3.1.1. Conditions
The conditions are designed in such a way that there is no possibility that the two marketing
techniques (celebrity endorsement and fun elements) are intertwined. I did not want to get attention by the children because of an extraordinary stand with balloons, flyers, banners and other stuff, because that can negatively effects the reliability of the findings. These elements can be seen as fun elements, who are already presented in two of the four conditions. The two other conditions should totally be separated from those elements.
Cartoons are often used as fun elements (Elliott, 2008), but in this case that may lead to certain multicollinearity. It is not totally equal to a celebrity endorser as a real person, but cartoons may have the same impact on children while it is ‘somebody’ from television who they like. Because of that, the fun elements in this experiment are 1) a crazy smiley and 2) colored flags.
In the four conditions, children are exposed to two posters. One contains the healthy food and the other contains the unhealthy food. The poster with the unhealthy food is similar in each condition, while the poster with healthy food differs. In condition 1) I just present the participants a poster with pictures of the healthy food on a solid background. In conditions 2) and 3) the healthy food on the poster is combined with respectively Arjen Robben and with fun elements. In condition 4), I show them one poster with both Arjen Robben and the fun elements. It is important to note that the design of the posters is totally equal in the different conditions, except for the elements mentioned, i.e., the background color was ‘soft’ orange and the food was presented at exactly the same place on the posters, in the same size.
21
participated in the experiment under the fourth condition, where the celebrity and fun elements both were presented combined with the healthy food. In total, there participated 168 children in this experiment from who the data will be used in analyzing the relations.
3.2 Validity
Children at most of the clubs in The Netherlands are a realistic overview of the society. That is strengthen by the fact that there are a lot of children from other clubs (the away teams) that have the possibility to take part in our research. Hence, they are from different ages and they are living in different villages or cities, most of them in the north of The Netherlands. From the 168 participants, 120 played at SV Lycurgus and 41 played at other clubs. Most of the times in the close neighborhood, or in villages in the province Groningen or Drenthe. To complete, 7 participants do not play soccer at all. This indicates high external validity, i.e., the relationship that I will find can be safely be
generalized to other soccer canteens in the Netherlands. To improve generalizability one could argue to ask to the education level or income of parents, but in this case I think it is not appropriate to ask this.
The children are randomly assigned to the conditions, which means that they all have equal change of participate in the survey in a particular condition. The matches of the different age-groups are distributed around the whole morning. In that way, I prevent that there is an effect on the relation between the independent variable and the dependent variable due to differences in personal characteristics of the participants.
A disadvantage in field experiments in general is the difficulty to achieve high internal validity, which arises from alternative explanations of relations found in the experiment. In their review of 69 experiments published between 1974 and 2014, Cruwys et al (2015) found out that almost all of studies have found that there is a certain (limited) evidence for a moderating effect of age towards eating behavior, including food choice. For that reason, it is important to explain how the survey will deal with this potential difficulty.
3.3 Control variables
22
3.3.1 Age
To prevent that age will have an unrehearsed effect on the findings in this research, I limit the respondents to only those children that are within the range of 8 to 14 years old. Younger boys were more influenced by the glamour of an endorser and were more reliant on his advice than older boys were during the survey of Ross et al (1984). That was because they were more naïve. They accepted the endorsement as a valid source of product information more than older boys did. Due to the Elaboration Likelihood Model of Petty et al (1983), the more naïve people are, the less they critically evaluate the advertisement and the greater the change that they will consider the claims that are made by the endorser as the truth, or that they respond favorable to the suggestions that are made. More recently, the research of Livingstone et al (2006) gave the same findings: younger children are more influenced when advertising is based on celebrities than older boys are. Hence, there seems to be a certain difference in influencing the food choice of younger children on the one hand, and older children on the other hand. In this survey, the younger children are invited to participate because they are probably more influenced by the marketing techniques I use in this experiment.
3.3.2 Liking the products
23
4. RESULTS
In this chapter, I elaborate which analyses are used to come up with conclusions. Also the results of the relations investigated are showed in this chapter.
In table 4 and figure 2, one can find the choices made by the participants under each of the
conditions. As said in the prior part, the participants who signed that they did not like more than one of the products are excluded from the analysis.
Food choice Control (%) Celebrity (%) Fun (%) Celebrity + Fun (%)
Unhealthy choice 29 (74.4) 23 (65.7) 26 (72.2) 31 (72.1)
Healthy choice 10 (25.6) 12 (34.3) 10 (27.8) 12 (27.9)
Total participants 39 (100) 35 (100) 36 (100) 43 (100)
Table 4 – overview of the food choice of the participants in the different conditions
Figure 2 – overview of the food choice of the participants in the different conditions
As one can see, the differences between the different conditions are not very big, but maybe the celebrity condition deviate some from the other three. To analyze this, I will perform a logistic regression. I do not perform a linear regression, because the probabilities of the choice of the food (DV) cannot be higher than 1 or lower than 0, and also, the values of the DV are dichotomous (i.e. a child chose the healthy or the unhealthy option). In this case, I investigate how children can be influenced in their food choice (DV). In other words, are children sensitive to the marketing techniques used, and will they choose healthy food over unhealthy when the healthy food is presented with a celebrity and/or fun elements?
Mathematical, the logistic regression can be written in the following equation:
0 5 10 15 20 25 30 35
Control Celebrity Fun Celebrity + Fun
24
In this formula, the B0 is the constant, B1 – Bn is the logistic regression coefficients and the x is the independent variable.
To translate the variables “Eating habits at home” and “Liking Robben”, a dummy variable is made for both of the variables, based on a median split. In the case of the former, there were six indicators, from which three indicates a healthy food pattern (vegetables, fruit and water). These indicators had a value of 1. There were also three were indicators of unhealthy food habits (candy, chips and soda). These indicators were valued with -1, so that the mean of each respondent indicates the degree to which a person eats healthy or unhealthy. In table 5 and figure 3 one can find an overview of the data that is collected by this question. The dummy variable is made with the value 0 for children that eat very unhealthy (-3) to a bit healthy (1), which are in total 78 children. The value 1 is assigned to children who eat healthy or really healthy (2 and 3), which are 75 respondents.
Food Do you consume this
(almost) each day?
Food Do you consume this (almost) each day?
No Yes No Yes
Vegetables (%) 36 (23.5) 117 (76.5) Soda (%) 92 (60.1) 61 (39.9)
Water (%) 30 (19.6) 123 (80.4) Chips (%) 126 (82.4) 27 (17.6)
Fruit (%) 34 (22.2) 119 (77.8) Candy (%) 73 (47.7) 80 (52.3)
Total 100 359 Total 291 168
Table 5 – overview of the eating habits of the participants
Figure 3 – overview of eating habits of the participants – which food do you consume (almost) each day?
The variable “Liking Robben” was also rescaled into a dummy variable, which was based on the answer children gave on a 5 points Likert scale. As in line with the expectation, children in general
0 20 40 60 80 100 120 140
Vegetables Water Fruit Soda Chips Candy
25
like Robben, so the values were relatively high. One can see the frequencies in detail in table 6. The split was made between the value 4 (like him) and 5 (like him very much). That meansthe
moderation effect investigates whether there is a difference of the influence of the exposure to Robben for children who likes him very much versus the other children.
Condition 1 – very
stupid
2 – stupid
3 – not stupid, but do not like him
4 – like him 5 – like him very much Condition 2 – celebrity 0 2 7 14 12 Condition 4 -celebrity + fun 0 2 10 15 16 Total (%) 0 (0) 4 (5.1) 17 (21.8) 29 (37.2) 28 (35.1)
Table 6 – overview of the meaning of participants about Arjen Robben
In table 7 one can find the results of the model that is described in the theoretical part of this paper. The data is structured in a way that a negative relation suggest a higher probability on a healthy choice of the participant. A positive relation indicates the probability of a unhealthy choice is higher.
Variable B S.E. Wald Df Sig. Exp (B)
Celebrity condition .004 .694 0 1 .995 1.004
Fun condition -.144 .538 .072 1 .789 .866
Celebrity * fun .656 .755 .754 1 .385 1.926
Liking Robben .485 .628 .598 1 .440 1.625
Liking Robben * Celebrity -1.080 .815 1.759 1 .185 .339
Eating Habits .321 .546 .345 1 .557 1.378
Eating Habits * Celebrity -.273 .748 .133 1 .716 .761
Age .098 .124 .628 1 .428 1.103
Constant -.922 1.407 .043 1 .836 .747
Table 7 – output of the logistic regression
Unfortunately, these findings does not correlate with the expectations formed in the theoretical part of this survey. None of the relations
investigated seems to be significant (<.05), which means that all the hypotheses, without
exception, should be rejected. As one can see in figure 4, the condition with the celebrity is motivating children to eat more unhealthy, although it is just a very little difference. The fun elements leads to a bigger change in the choice Figure 4 – probability of food choice (0=healthy, 1=unhealthy) of children, in favor of the healthy food.
26
condition, one can assume that it is probably not reliable and therefore, it is not included in figure 4 and 5.
The interaction term Liking Robben * Celebrity condition is the closest to significant. With this specific interaction term, one can analyze whether the influence of the celebrity condition on the food choice of children is different when someone likes Arjen Robben or not. Because of the structure of the data, I analyze the difference between children who like him very much versus the other children. When this relation was significance, combined with the other insignificant results, hypothesis 2b was justified. In other words, one could conclude that children that like Arjen Robben very much tend to choose the healthy option over the unhealthy alternative when the former one is promoted with Arjen Robben. The other children may do so, but much less. The value of .339 (Exp B) indicates that they chose much more healthier, and that it is not a small difference.
4.1 Liking the celebrity
As earlier mentioned, a high match-up between the participants and the celebrity is necessary for the desired effect, when using the celebrity as a marketing technique (Misra et al, 1990). The data confirmed that most of the children liked Arjen Robben. In that context, one can conclude that there is a high match-up between the participants and the celebrity. However, due to the fact that almost all the participating children liked him to a certain degree, the moderating effect of “Liking the celebrity” to the relation between the conditions and the food choice is not very reliable. The
27
Another interesting option is analyze only the children that like Robben. In this case, only the children who ticked the ‘4 – like him’ or the ‘5 – like him very much’ box are used as input for a logistic regression. As a consequence, the number of respondents that are considered, are reduced to 118. In this case, as in the analyses before, I analyze the variables as displayed in table 8. The difference is that the variable Liking Robben is removed, just as the interaction term Liking Robben * Celebrity. Obviously, it is given that all the children in this analysis like Robben.
Variable B S.E. Wald Df Sig. Exp (B)
Celebrity condition -1.299 .771 2.418 1 .120 .301
Fun condition -.594 .602 .975 1 .323 .552
Celebrity * fun 1.691 .876 3.724 1 .054 5.426
Eating Habits .205 .595 .119 1 .730 1.228
Celebrity * Eating Habits .027 .843 0.001 1 .974 1.027
Age .162 .138 1.375 1 .341 1.176
Constant -.503 1.597 .099 1 .753 .605
Table 8 – output of the logistic regression
The results due to the celebrity condition are more directed to the way one could expect. In detail, the celebrity conditions stimulates the children to choose the healthy option over the unhealthy one, and the result is way closer to a significant level (.120) compared to the initial model (.995). But still, this is no significant evidence for the influencing effect that Robben has on the children. The Celebrity * Fun
interaction term is even more close to
significant (.054). Actually, one can consider this
Figure 5 – probability of food choice (0=healthy, 1=unhealthy) as marginal significant. Strangely, adding the fun elements next to the influence of Robben, leads to an opposite effect as expected. Next to that, the effect is extremely strong. Children tend to choose the unhealthy food over the healthy alternatives in this case.
To find a possible explanation for this result, it should be useful to analyze the distribution of data that is collected during the particular condition different ways. The data in this condition is collected on the 6th of May from 12.00 am to 0.30 pm and on the 13th of May from 9.30 am to 11.00 am.
28
29
5. DISCUSSION, LIMITATIONS AND FURTHER RESEARCH
As one can see in the previous chapter, the hypothesis should be rejected, which is contrary to the theoretical part of this research. In this part of the paper the possible reasons will be discussed. Next to that, some limitations and, based on these limitations, possible future research directions are addressed.
First of all one has to realize that the setting of the experiment is some different compared to a normal situation. In fact, there are two differences: (1) children do not pay with money, but they have to fil in a questionnaire to get some food and (2) they got it at a table in the corner of the building (i.e., not at the bar where they usually buy food). These two differences might have influence, but I assume that this changed the choice of the children not very much.
During the experiment, as said, there is more data collected than is used in the logistic analysis in the previous part. On the one hand it is done because of ‘hiding’ the specific relations investigated in this research, as an alternative kind of a cover story. On the other hand, this data can be analyzed in addition to the data that is used to estimate the model of the conceptual framework. When analyzing the other data in combination with some data that is already investigated, the variable “Gender” turned out to be an important one.
5.1 Gender
30
Condition Healthy choice Unhealthy choice Total
Control 6 6 12
Celebrity 1 6 7
Fun 2 1 3
Celebrity + fun 6 2 8
Total 15 15 30
Table 9 – overview of the food choice of the female respondents
This finding is in line with many research papers that are suggesting that there are significant differences in the reaction of men/boys and women/girls on marketing techniques (Darley and Smith, 1995; Meyers-Levy and Sternthal, 1991; Wolin and Korgaonkar, 2003). Next to that, during the research they performed, Oliver and Thelen (1996) found out that girls reported more eating-related concerns than boys regarding to peer influence. In the current experiment, children (both boys and girls) often participated with friends at the same time, which may indicate that they were influenced by peers and which may explain the difference between boys and girls in their food choice.
Unfortunately it is not reliable to take the collected data into a regression analysis, because the 30 female participants were assigned to the four different conditions. Because the other relations in the regression are adjusted by the gender, the sample size is too small to run a logistic regression
including the gender variable. To illustrate with an example, the condition with fun-elements contains only three female participants. It is not appropriate to draw conclusions based on this sample size, but it may be an interesting option to investigate further in the near future. The difference between girls and boys is also a possible explanation of the unexpected result founded in the last analysis of the results part of this paper, regarding to the fourth condition of the experiment. Only four out of the 31 respondents was female. Three of them chose the healthy option, where only five out of the 27 considered male respondents chose the healthy food. The unexpected result is in line with the findings that female respondents in general preferred the healthy option over the unhealthy one.
5.2 Influence of the circumstances
31
5.2.1 Peer influence
The canteen especially is a busy place with a lot of environmental incentives. As already mentioned, girls are more influenced by their peers (friends for instance) than boys are (Oliver and Thelen, 1996), but also boys are influenced by others (Salvy et al, 2012). It is questionable whether this influenced the results of this experiment or not. Based on a research paper of Birch (1980) that conclusion may be justified. She grouped a child who preferred vegetable A to B with three peers with the opposite preference. After all the children chose their favorite vegetable one day, the child with the deviated preference adjusted it to the same preference as the others. Similar to my experiment, Birch (1980) investigated food preference with just two options available. Next to that, one can assume that the children who participated in the current experiment together are friends, family or team mates, i.e., they see each other much more times than only during the experiment. That reduces the difference with the research of Birch (1980), where the peer influence is measured in a setting children see each other more than once. One can certainly not rule out the possibility that peers had influence on the food choice of the children in the current setting. For that reason, the influence of the different conditions children were exposed to, may be lower than expected.
However, this research was about investigating the eating behavior of children in sports canteen. In a real setting, there is also the presence of peers when visiting the canteen and buying something to eat or drink. It is because of that no option to prevent peer influence in these kinds of experiments. For that reason, it is not very helpful to investigate whether and to which degree there is peer influence in such a setting. A more valuable research direction is to investigate how children can be influenced to eat more healthy, taking into account that they do not always base their food choice on their own preference.
5.2.2 Chaotic environment
32
5.3 Exposure to the conditions
It is clear that the exposure of the posters to the children in general did not lead to a more healthy food choice. Based on the two factors discussed above, one could argue that this has to do with environmental factors. However, to find an answer on the question how children in a sports canteen can be stimulated to eat more healthy, one has to take into account that these circumstances are given. In that case, it has value to find out how the conditions that children faced in this experiment have more influence. It seems that, given the circumstances, the way of exposure through the posters in appendices 1A – 1E was not strong enough to influence the participants. This can be caused by different options. One could suggest that the use of posters is not the right technique to influence children in their choice of the food they want. Although the choice of the content of the posters is well supported in the theoretical part of this paper (the effectiveness of celebrity endorsement and the use of fun elements in marketing to children is proved), it is questionable whether the effectiveness of the way these elements are exposed to the children can be increased.
5.3.1 Celebrity endorsement
As Miciak and Shanklin (1994) found out, many commercials using celebrity endorsers do not match to the expectations of the advertisers. In this experiment, this is more or less also the case.
Celebrities are usually used to build a brand or to create brand equity (Erdogan, 1999; Till, 1998). That may indicate that it is not effective to use this technique in a one-time buy. Rice et al (2012) confirmed this by suggesting that repetition is an important tool to make the use of celebrity as a marketing tool effective. Repetition ensures a higher probability of recognizing the brand and celebrity connection (Costanzo and Goodnight, 2005). These findings may explain why the celebrity endorsement had no influence on the choices of children in the current experiment, while in this experiment only a one-time buy is measured. When a soccer club want to experiment with using a celebrity, it is interesting to run an experiment over more than one or two weeks, but for example over several months. Due to repetition, children can make a positive association between the celebrity and the food that is offered in such an experiment.
5.3.2 Fun elements
33
was presented on a poster and on the questionnaire, in the same space as the healthy food option. That is slightly different from the usual way to use this technique, because it is not on the food or on the package of the food itself (Elliot 2008). That may reduce the effect it has and is a possible
explanation of the negligible effect the fun elements had on the food choice of children. However, in this research it was not possible to change the products or the package of the products itself, so I had to find out other ways to use the fun elements.
5.4 Healthy eating habits
34
6. REFERENCES
Aronson, Elliot, Wilson, Timothy D. and Brewer, Marilynn B. (1998), “Experimentation in social psychology”, The handbook of social psychology, 1-2 (4), 99 – 142
Atik, Deniz., and Ertekin, Zeynep O (2011), “Children’s perception of food and healthy eating: dynamics behind their food preferences”, International Journal of Consumer Studies, 37 (1), 59 – 65 Austin, S. Bryn, Melly, Steven J., Sanchez, Brisa N., Patel, Aarti, Buka, Stephen, and Gortmaker, Steven L. (2005), “Clustering of Fast-Food Restaurants Around Schools: A Novel Application of Spatial
Statistics to the Study of Food Environments”, American Journal of Public Health, 95 (9), 1575 – 1581 Batada, Ameena, Seitz, Maia D., Wootan, Margo G. and Story, Mary (2008), “Nine out of 10 food advertisements shown during Saturday morning children’s television programming are for foods high in fat, sodium, or added sugars, or low in nutrients”, Journal of the American Dietetic Association, 108, 673 – 678
Batrinou, Anthimia, M. and Kanellou, Anastassia (2009), “Healthy food options and advertising in Greece”, Nutrition & Food Science, 39, 511 – 519
Benton, D. (2004), “Role of parents in the determination of the food preferences of children and the development of obesity”, International Journal of Obesity, 28, 858 – 869
Birch, Leann L. (1980), “Effects of Peer Models’ Food hoices and Eating Behaviors on Preschoolers’ Food Preferences”, Child Development, 51 (2), 489 – 496
---, Birch, David, Marline, Diane W. and Kramer, Laurie (1982), “Effects of Instrumental Consumption on Children’s Food Preference”, Journal for Intake Research, 3, 125 – 134
---, Fisher, Jennifer O. (1997), “Development of Eating Behaviors Among Children and Adolescents”, The American Acadamy of Pediatrics, 101, 539 – 549
---, ---, Grimm-Thomas, K., Markey, C.N., Sawyer, R., and Johnson, S.L. (2001), “Confirmatory factor analysis of the Child Feeding Questionnaire: a measure of parental attitudes, beliefs and practices about child feeding and obesity proneness”, Appetite, 36 (3), 201 – 210
Boush, David M., Friestad, Marian, and Rose, Gregory, M. (1994), “Adolescent Skepticism toward TV Advertising and Knowledge of Advertiser Tactics”, Journal of Consumer Research, 21, 165 - 175 Boyland, Emma, J., and Halford, Jason C.G. (2013), “Television advertising and branding. Effects on eating behavior and food preferences in children”, Appetite, 62, 236 – 241
Brown, Rachael, and Ogden, Jane (2004), “Children’s eating attitudes and behavior: a study of the modelling and control theories of parental influence”, Health Educational Journal, 19 (3), 261 – 271 Calvert, Sandra L. (2008), “Children as consumers: Advertising and marketing”, The future of
Children, 18 (1), 205 – 234
Carrol, Angela (2008), “Brand communications in fashion categories using celebrity endorsement”, Journal of Brand Management, 17 (2), 146 – 158
Cairns, G., Angus, K., and Hastings, G. (2009), “The Extent, Nature and Effects of Food Promotion to Children: A Review of the Evidence to December 2008”, World Health Organization
35
Chai, H.J., Hong, H., Kim, H.S., Lee, J.S., and Yu, C.H. (2008), “A Study of Food and Nutrient Intakes of Weekday and Weekend Among High School Boys in Seoul”, Korean Journal of Nutrition, 41 (6), 539 – 549
Chapman, Kathy, Nicholas, Penny, Banovic, Debbie, and Supramaniam, Rajah (2006), “The extent and nature of food promotion directed to children in Australian supermarkets”, Health Promotion
International, 21, 331 – 339
Clark, H.R., Goyder, E., Bissel, P., Blank, L., and Peters, J. (2007), “How do parents’ child-feeding behavior influence child weight? Implications for childhood obesity policy”, Journal of Public Health, 29 (2), 132 – 141
Coon, K.A., and Tucker, K.L. (2002), “Television and children’s consumption patterns. A review of the literature”, Minerva pediatrics, 54, 423 – 436
Costanzo, Paul J. and Goodnight, Janelle E. (2005), “Celebrity Endorsements: Matching Celebrity and Endorsed Brand in Magazine Advertisements”, Journal of Promotion Management, 11 (4), 49 – 62 Cruwys, Tegan, Bevelander, Kirsten E., Hermans, Roel C. J. (2015), “Social modeling of eating: A review of when and why social influence affects food intake and choice”, Appetite, 86, 3 – 18 Darley, William K. and Smith, Robert E. (1995), “Gender Differences in Information Processing Strategies: An Empirical Test of the Selectivity Model in Advertising Response”, Journal of Advertising, 24, (41 – 56)
Dixon, Helen, Scully, Maree, Wakefield, Melanie, Kelly, Bridget, Chapman, Kathy, and Donovan, Robert (2011), “Parent’s responses to nutrient claims and sports celebrity endorsements on energy-dense and nutrient-poor foods: an experimental study”, Public Health Nutrition, 14 (6), 1071 – 1079 Elliott, Charlene D. (2008), “Marketing Fun Foods: A Profile and Analysis of Supermarket Food Messages Targeted at Children”, Canadian Public Policy, 34 (2), 259 – 173.
--- (2009), “Healthy Food Looks Serious: How Children Interpret Packaged Food Products”, Canadian Journal of Communication, 34, 359 – 380
Erdogan, B. Zafer (1999), “Celebrity Endorsement: A Literature Review”, Journal of Marketing Management, 15, 291 – 314
FC Groningen (2011), “Arjen Robben Ambassadeur ‘FC Groningen in de Maatschappij’“,
https://www.fcgroningen.nl/nieuws/arjen-robben-ambassadeur-fc-groningen-in-de-maatschappij/2, retrieved in April 2017
Franks, Paul W., Hanson, Robert L., Knowler, William C., Sievers, Maurice L., Bennett, Peter H., and Looker, Helen C. (2010), “Childhood Obesity, Other Cardiovascular Risk Factors, and Premature Death”, The New England Journal of Medicine, 362 (6), 485 – 493
Genschow, Oliver, Reutner, Leonie, and Wänke, Michaela (2012), “The Color Red Reducees Snack Food and Soft Drink Intake”, Appetite, 58 (2), 699 – 702
GFK, Convenant Gezond Gewicht (2015), “De Gezonde Sportkantine: Onderzoek onder koplopers sportkantines”,
https://jongerenopgezondgewicht.nl/userfiles/De%20gezonde%20sportkantine%20(rapportage%20 maart%202015).pdf retrieved in Mach 2017
36
Hastings, G., Stad, M., McDermott, L., Forsyth, A., MacKintosh, A.M., Rayner, M., Godfrey, C., Caraher, M., and Angus, K. (2003), “Review of Research on the Effects of Food Promotion to Children”, Glasgow, Scotland: Centre for Social Marketing
Hunter, B.T. (2002), “Marketing Foods to Kids: Using Fun to Sell”, Consumers’ Research Magazine, 85 (3), 16 – 19
Jenkin, Gabrielle, Wilson, Nick, and Hermanson, Nicole (2009), “Identifying ‘unhealthy’ food advertising on television. A case study applying the UK Nutrient Profile Model”, Public Health Nutrition, 12, 614 – 623
Johannsen, Darcy L., Johannsen, Neil M., and Specker, Bonny L. (2006), “Influence of Parents’ Eating Behaviors and Child Feeding Practices on Children’s Weight Status”, Obesity, 14 (3), 431 – 439 Jonker, Laura (2015), “Trends in de populariteit van het Nederlandse voetbal en de ontwikkeling van onze unieke voetbalinfrastructuur”, http://www.sportenstrategie.nl/nieuws/31628/ (KNVB),
retrieved in June 2017
Kalmthout, van, Janine (2015), “Gezonde sportkantine”, Het Mulier Instituut
Kelly, Bridget., Halford, Jason C.G., Boyland, Emma J., Chapman, Kathy, Bautista-Castana, Inmaculada, Berg, Christina, Caroli, Margherita, Cook, Brian, Coutinho, Janine, G., Effertz, Tobias, Grammatikaki, Evangelia, Keller, Kathleen, Leung, Raymond, Manios, Yannis, Monteiro, Renata, Pedley, Claire, Prell, Hillevi, Raine, Kim, Recine, Elisabetta, Serra-Majem, Lluis, Singh, Sonia, and Summerbell, Carolyn (2010), “Television food advertising to children. A global perspective”, American Journal of Public Health, 100, 1730 – 1736
Kireyev, Pavel, Pauwels, Koen, and Gupta, Sunil (2016), “Do Display ads influence search? Attribution and Dynamics in Online Advertising”, International Journal of Research in Marketing, 33 (3), 475 – 490
Lee, Helen (2012), “The role of local food availability in explaining obesity risk among young school-aged children”, Social Science & Medicine, 74, 1193 – 1203
Lee, Jung-Gyo, and Thorson, Esther (2008), “The impact of celebrity-product incongruence on the effectiveness of product endorsement”, Journal of Advertising Research, 48 (3), 433 – 449 Livingstone, Barbara (2001), “Childhood obesity in Europe: a growing concern”, Public Health Nutrition, 4 (1A), 109 – 116
Livingstone, Sonia and Helsper, Ellen J. (2006), “Does advertising Literacy Mediate the Effects of Advertising on Children? A Critical Examination of Two Linked Research Literatures in Relation to Obesity and Food Choice”, Journal of Communication, 56 (3), 560 – 584
Martin, James H (1996), “Is the athlete’s sport important when picking an athlete to endorse a nonsport product?”, Journal of Consumer Marketing, 13 (6), 28 – 43
Mehta, Kaye, Phillips, Clare, Ward, Paul, Coveney, John, Handsley, Elizabeth, and Carter, Patricia (2012), “Marketing foods to children through product packaging: prolific, unhealthy and misleading”, Public Health Nutrition, 15 (9), 1763 – 1770
Meyers-Levy, Joan and Sternthal, Brian (1991), “Gender differences in the Use of Message Cues and Judgments”, Journal of Marketing Research, 28 (1), 84 – 96
37
Misra, Shekhar, and Beatty, Sharon .E. (1990), “Celebrity Spokesperson and Brand Congruence, An Assessment of Recall and Affect”, Journal of Business Research, 21, 159 – 173
Noble, Carel, Corney, Michael, Eves, Anita, Kipps, Michael, and Lumbers, Margaret (2000), “Food choice and school meals, primary schoolchildren’s perceptions of the healthiness of food and the nutritional implications of food choices”, International Journal of Hospitality Management, 19 (4), 413 – 432
O’Dea, Jennifer A. (2003), “Why do kids eat healthful food? Perceived benefits and barriers to healthful eating and physical activity among children and adolescents”, Journal of the American Dietetic Association, 103 (4), 497 – 501
Oliver, Krista K. and Thelen, Mark H. (1996), “Children’s Perceptions of Peer Influence on Eating Concerns”, Behavior Therapy, 27, (25 – 39)
Petty, Richard E., Cacioppo, John T., and Schumann, David (1983), “Central and Peripheral Routes to Advertising Effectiveness: The Moderation Role of Involvement”, Journal of Consumer Research, 10 (2), 135 – 146
Pires, Carla, and Agante, Luisa (2011), “Encouraging Children to eat more healthily: The influence of packaging”, Journal of Consumer Behaviour, 10 (3), 161 – 168
Powell, Lisa M., Slater, Sandy, Mirtcheva, Donka, Bao, Yanjun, Chaloupka, Frank J. (2007a), “Food store availability and neighborhood characteristics in the United States”, Preventive Medicine, 44 (3), 189 – 195
---, Szczypka, Glen, Chaloupka, Frank J., and Braunschweig, Carol L. (2007b), “Nutritional Content of Television Food Advertisements Seen by Children and Adolescents in the United States”, American Academy of Pediatrics, 120 (3), 576 – 583
Power, C., Lake, J.K., and Cole, T.J. (1997), “Review: Measurement and long-term health risks of child and adolescent fatness”, Journal of Obesity, 21, 507 – 526
Rice, Dan H., Kelting, Katie, and Lutz, Richard J. (2012), “Multiple endorsers and multiple
endorsements: The influence of message repetition, source congruence and involvement on brand attitudes”, Journal of Consumer Psychology, 22, 249 – 259
Roberto, Christina A., Baik, Jenny, Harris, Jennifer L., and Brownell, Kelly D. (2010), “Influenced Characters on Children’s Taste and Snack Preferences”, American Academy of Pediatrics, 126 (1), 88 – 93
Ross, Rhonda P., Campbell, Toni, Wright, John C., Huston, Aletha C., Rice, Mabel L., and Turk, Peter (1984), “When celebrities talk, children listen: An experimental analysis of children’s responses to TV ads with celebrity endorsement”, Journal of Applied Developmental Psychology, 5 (3), 185 – 202 Salvy, Sarah-Jeanne, Haye, de la, Kayla, Bowker, Julie C., and Hermans, Roel C. J. (2012) “Influence of peers and friends on children’s and adolescents’ eating and activity behaviors”, Physiology &
Behavior, 106 (3), 369 – 378
Samu, Sridhar, and Wymer, Walter (2009), “The effect of fit and dominance in cause marketing communications”, Journal of Business Research, 62 (4), 432 – 440
Scaglioni, Silvia, Salvioni, Michela, and Galimberti, Cinzia (2008), “Influence of parental attitudes in the development of children eating behavior”, British Journal of Nutrition, 99 (1), 22 – 25
