Once upon a time in my culture
-An experimental study into the persuasive effects of culturally
grounded narratives
Research Master Thesis Saar Hommes
Linguistics – Language & Cognition s1995650
November 2018 saarhommes@gmail.com
Supervisor& second reader: Damsterdiep 22B
Prof. Dr. J.C.J. Hoeks 9711SL Groningen
Summary
Obesity rates in the Netherlands have never been higher, and young people should learn to make healthier eating choices to prevent dangerous consequences in later life. Yet, persuading people to change, is not without challenges. One possible way of overcoming resistance seems to be the use of narratives. Especially narratives that elicit emotions (fear appeals), prove to be effective persuaders. Scholars also call for incorporating culture within health promotion materials since ethnic minorities often lack the relevant health care they deserve.
In a 2x3 design, two cultural versions of narratives were read. Participants (n=65, ages:14-16) either read (1) a narrative centered in the Dutch cultural background while being Dutch themselves (D-D), (2) a narrative tailored to the Turkish cultural background, while having another cultural background themselves (T-O), or (3) the Dutch narrative while having another cultural background (D-O).
Statistical analyses showed that cultural condition had a significant effect on behavioral intentions (F(58)≈4.843, p≈.0113 with h2≈.1431). Contrary to what was expected by earlier studies, people in the third (mismatch) condition scored significantly higher than people in the second (mismatch) condition. The cultural match condition (D-D) did not outperform mismatch conditions. No significant effects were found on attitudes (since a ceiling effect occurred), narrative persuasion, and intentions to talk about healthy eating and weight.
Results indicate that health promotion material does not need to be matched to the cultural background of the reader as it can persuade people who do not match the culture depicted in the narrative as well. The study also shows that transportation is the most important predictor of narrative persuasion and health practitioners should focus on ways to strengthen the transportation process. Finally, the study is one of the first to include teenage participants and recognizes that they are the ones that need to be convinced of a healthier life style in order to prevent major health risks in later stages of their lives.
Acknowledgements
First of all, I would like to thank prof. dr. John Hoeks for all his advice and guidance during my thesis work and throughout my Research Master. Also, I would like to thank dr. Gregory Mills for his efforts as a second reader of this thesis.
I started thinking about my thesis subject as early as in 2016. While meeting with prof. dr. Carel Jansen and dr. Joëlle Ooms for the ‘Narrative, Fear Appeals & Emotions’-project, I became persuaded by persuasive health communication.
During these meetings food was often consumed and differences between both professors became apparent. Whereas prof. dr. John Hoeks consumed his vegan lunch, prof. dr. Carel Jansen often ate a pita with shawarma. Therefore, discussions about healthy eating occurred not only since health communication research begins to recognize the problems related to diet, but were evoked spontaneously as well. At the same time, since I was involved in the beforementioned project on narratives, combining narrative theory with healthy eating seemed promising. Personally, I have always been fascinated by culture and its possible influence in daily life. Coming from a multicultural school in Lelystad, I was often surprised by the predominantly Dutch students in Groningen (especially since I did a Bachelors in Dutch) and how much their high school experiences differed from mine. Combining culture, healthy
eating and narratives was therefore not only a theoretically fruitful combination, but made
sense for me personally as well.
Finally, I would like to thank my friends and family, especially my mom since she is the first person I learned how to cook from and how to properly write a large paper such as this thesis (remember my “Werkstuk” on witches?), and Dominick for at least pretending to care about my research enough to ask (often difficult to answer) questions such as “please tell what your thesis is about in plain language” (I promise, by the way, that this sentence is the longest one of this thesis, so no worries). If I have scared you off, I hope you will at least read the summary of this thesis and nod politely when I continue to talk about research in the future as a PhD-candidate.
Table of Contents
1. Summary ………2
2. Acknowledgements ………3
3. Introduction………..……6
4. Theoretical framework………7-26
Chapter 1: Defining culture………..…7-13
1. Culture in health communication research 7-10
2. Why is culture important in health communication? 10-11 3. Does cultural adaptation of health materials work? 12
4. Chapter summary 12-13
Chapter 2: Cultural narratives……….……14-22
1. What are narratives? 14
2. Are narratives able to persuade people? 14-16
3. Narrative mechanisms 16-18
4. Narrative fear appeals 18-20
5. Culturally grounded narratives and persuasion 20-21
6. Do culturally grounded narratives work? 21-22
7. Chapter summary 22
Chapter 3: Healthy eating………23-26
1. What is healthy eating? 23
2. Why is healthy eating relevant? 23-24
3. How is healthy eating cultural? 24-26
4. Chapter summary 26
5. Method………...…..27-36
1. Operationalisation 27-30
3. Materials 31-34
4. Design & Procedure 35-36
5. Data processing 36
6. Results……….…….37-44
1. Narrative persuasion 37-40 2. Intentions 40-41 3. Attitudes 41-42 4. Explorative analyses 43-447. Conclusion & Discussion……….45-48
8. References………...……….49-53
Appendices………..………54-127
1. Appendix A: The Narratives 55-57
2. Appendix B: The Questionnaire 58-70
3. Appendix C: Letter to the schools 71
4. Appendix D: Letter to the parents 72-73 5. Appendix E: Letter of approval ethics committee 74
6. Appendix F: Introduction to the study 75
7. Appendix G: Word of thanks 76
8. Appendix H: E-mail to the winners 77 9. Appendix I: Statistical analyses in R 78-127
1. Introduction
Obesity rates in the Netherlands have never been higher, and these rates continue to climb (Lobstein, Baur & Uauy, 2004). Children born around 2000 might be the first generation ever to have a lower life expectancy rate than their parents due to problems linked with over-eating (Dubé, 2010). At the same time, cultural minorities are even more at risk of becoming overweight (Hubbard et al., 2016) and often do not receive relevant health care (Ujcic-Voortman et al., 2011). That is why health communication researchers suggest that culture should be accounted for when designing persuasive health materials (Kreuter et al., 2002).
One promising way of persuading people to live healthier lives is the use of narratives (Green & Brock, 2000; Moyer-Gusé, 2008; Nabi, 2010 & De Graaf, Sanders & Hoeken, 2016), especially narratives that elicit strong emotions such as fear appeals (Ooms, Jansen, Hommes & Hoeks, 2017). Incorporating both narrative persuasion techniques and accounting for culture when designing health materials could therefore be a fruitful combination and these culturally tailored narratives have been a recent subject in health communication research (e.g. Larkey & Hecht, 2010; Huang & Shen, 2016). Empirical evidence for the effects of culturally grounded materials however is scarce (Huang & Shen, 2016) and mixed (Koops van ‘t Jagt et al., 2016). The central question of this study is: “What effects do culturally grounded narratives about
healthy eating have on narrative persuasion, intentions and attitudes of teenagers (ages 14-16), who either match or mismatch the cultural background of the narratives?”.
In the theoretical framework, three topics will be explored. In the first chapter, it will be discussed what culture in health communication research entails. The topic of the second chapter is narratives. What are the effects of (culturally grounded) health narratives and which mechanisms are responsible for narrative persuadion? The central subject of the third chapter is healthy eating. Questions that are addressed include: why is healthy eating a relevant societal issue and how could healthy eating be connected with culture?
In the method, the experiment will be discussed. This 2x3 experiment compares two cultural versions of a health narrative in three cultural conditions (Dutch narrative- Dutch cultural background reader, Turkish narrative – other cultural background reader, Dutch narrative – other cultural background reader). Then, results of the experiment are presented and limitations and implications for existing theory and practice will be dealt with in the discussion. The final conclusion will provide the answer to the research question.
2. Theoretical Framework
Chapter 1: Defining culture
“
Language and culture are the framework through which humans experience, communicate, and understand reality.”
(Lev Vygotsky, 1968)
In this chapter, I will discuss how culture is defined within health communication research, how it can be measured, why culture is important and if culture-sensitive health material is indeed effective according to existing empirical evidence. At the end of this chapter, I will provide a short summary by highlighting the important issues discussed.
1. Culture in health communication research
Defining culture is not an easy task and views on what culture entails have changed over time. The work of Hofstede (1980) can be seen as a starting point for defining culture within (health) communication research. In his influential work ‘Culture’s consequences: international
differences in work-related values’ he defines culture as “the collective programming of the
mind which distinguishes the members of one human group from another” (Hofstede, 1980, p.25). A more recent definition of culture is given by Kreuter et al. (2002). They state that nowadays social scientists generally agree “[…] culture is learned, shared, and transmitted from one generation to the next, and it can be seen in a group’s values, norms, practices, systems of meaning, ways of life, and other social regularities” (p.133).
Kreuter et al. (2002) draw their definition from an extensive literature search, and although the definition they give is quite clear, the question remains how health communication researchers deal with implementing culture in health materials. In other words, to design health materials in such a way that they are “culturally appropriate”. The authors (2002) conclude that researchers use a wide variety of strategies to achieve this. They (2002) categorized these methods into five strategies: peripheral strategies, evidential strategies, linguistic strategies, constituent-involving strategies and sociocultural strategies. I will briefly discuss them.
Peripheral strategies try to match health promotion materials to characteristics of the target audience by the way they are packaged. For example, by using pictures, or fonts that might appeal to the target group. In evidential health promotion material, evidence of the specific target group is presented, such as “This year, 14.100 African Americans will be
diagnosed with colorectal cancer, and 6.800 will die from it.” (Kreuter et al.; 2002, p. 135). The third set of strategies, linguistic strategies, “seek to make health education programs and materials more accessible by providing them in the dominant or native language of the target group” (p.135). This could include translation of material or the use of urban language. The constituent-involving strategies make use of the target audience as the messenger. For example, by communicating the health promotion message through testimonials of the target group. The last set, the sociocultural strategies, discuss the health subject in a deeper context. This may sound a bit vague, but the authors (2002) give examples of using religion to convey the message (see section 2). And, although Kreuter et al. (2002) do not mention Hofstede (1980; 2010) here, I would argue that using his dimensions would also be considered a sociocultural strategy as this involves a deeper understanding of the culture than is immediately visible to the eye.
In sum, Hofstede (1980, 2010) suggests that culture can be distinguished based on four dimensions: power distance, individualism-collectivism, uncertainty avoidance and masculinity-femininity. He later added long-term versus short-term orientation, and indulgence versus restraint as fifth and sixth dimensions. And although there is criticism on these dimensions and their generalizability for all people within one culture (for an overview, see Jones, 2007), they are still used in health communication research today (for example Yaman et al., 2010 and Kim, 2017). In an overview of criticism on Hofstede’s work, Jones (2007) concludes that “Far more scholars belong on the pro-Hofstede team than don’t […]” and “Although not all of what Hofstede has said stands up to public enquiry, the majority of his findings, have weathered the storms of time, and will continue to guide multi-national practitioners into the ‘global’ future.” (p. 2). I will briefly explain the dimensions Hofstede invented here.
Power distance is the acceptance of inequality between people with power and people in less powerful positions. Or, as Hofstede et al. (2010, p.61) put it: “the extent to which less powerful members of institutions and organizations within a country expect and accept that power is distributed unequally”. For example, power distance is used to describe how civilians relate to the government. Germanic and English-speaking Western countries typically score lower on power distance (e.g. within these cultures, people strive for social equality) than Asian countries (e.g. within these cultures, people want to strengthen powerful positions).
The distinction between individualism-collectivism is based on the ‘self’ versus the ‘group’. In individualist societies, people depend more on themselves than others and are expected to take care of themselves. In collectivistic societies, people depend strongly on others (such as their families) and are expected to look after others. Collectivists also tend to care more about
what others think and are more susceptible of social norms than individualists (Kim, 2017). Western countries tend to be more individualistic than Asian countries (Hofstede et al., 2010). Hofstede et al. (2010, p.191) define uncertainty avoidance as “the extent to which the members of a culture feel threatened by ambiguous or unknown situations”. Cultures with a high uncertainty avoidance tend to have more formal rules, focus on stability and have a stronger belief in absolute truths, whereas for low uncertainty avoidance cultures, the opposite is true (Kim, 2017). East and Central European countries tend to score higher on uncertainty avoidance than Germanic-speaking countries (Hofstede et al., 2010).
The masculinity-femininity dimension is somewhat different than the other dimensions mentioned in that both masculinity and femininity are present within one culture or society. How many people prefer either masculinity over femininity, or the other way around, determines if the culture is categorized as masculine or feminine. In a masculine society, gender roles are traditional (e.g. men are tough and assertive, women are tender and modest). In a feminine society, gender roles are more fluid (Kim, 2017). Japan and Italy are examples of masculine societies, the Netherlands is characterized as a more feminine society (Hofstede et al., 2010).
The distinction between a long-term and short-term culture is based on how members handle their orientation towards time. In a long-term culture, people tend to focus more upon the future and its possible rewards, whereas members of a short-term culture are more focused on the past and present. In a long-term culture, which is prevalent in Western countries, perseverance and thrift are important values, traditions and preservations characterizes short-term orientation prevalent in Asian countries (Hofstede et al., 2010).
The sixth and final dimension is indulgence versus restraint. In a high indulgence society, people can freely fulfill their wishes and desires, without having to conform to strict social norms. Important values for indulgence societies are democracy, human rights and freedom of speech (Kim, 2017). Examples of cultures with high indulgence are in Western Europe, whereas Asian and Muslim countries are characterized as restrain societies (Hofstede et al., 2010).
As mentioned before, Hofstede’s dimensions form the foundation of a vast amount of cross-cultural research, including within health communication. Researchers form hypotheses based on these cultural dimensions, such as “low power distance enhances public service motivation” (Kim, 2017). It should be noted here that, although Hofstede’s work is very usable within health communication research, the studies using his dimensions usually focus upon comparing one country to another (as was intended by Hofstede). That being said, it is now also commonly
used for comparing subcultures within one and the same country (for example, Yaman et al., 2010).
Now that I have demonstrated what culture entails in health communication research and how health materials can be made culturally appropriate by using a set of strategies, the question remains how culture can be assessed. Health communication researchers distinguish cultures based on gender, age, ethnicity, race, education, religion and socio-economic status (Kim, 2016; Kreuter et al., 2002). Kreuter et al. (2002) however make a valid point when they say that “in practice, […], culture is more often assumed than assessed. For example, race and ethnicity are frequently used as proxies for a culture. Such variables may be important –even central- parts of a given culture but are not, in and of themselves, culture.” (p. 134). Betsch et al. (2016) seem to make the same point when they state that: “Culture is a collective sense of consciousness with both quantifiable and unquantifiable components that can audibly or silently reveal themselves through history and language. Culture is never static and is commonly reinforced through structures, even though those structures are not always palpable and visible as physical structures.” (p.814). According to Betsch et al. (2016), culture should be seen as something dynamic with components that are not always visible at the surface. However, to this date researchers have not come up with a practical solution on how exactly culture should be measured then. It goes beyond the scope of the current study to focus on this measurement problem, but it could be an exciting direction for future research (see more on this in the ‘Discussion’ section and review the ‘Method’ section for how culture was assessed in the current project).
2. Why is culture important in health communication?
Since health disparities between cultural groups are a problem (Kreuter et al., 2002; Hoeken & Korzilius, 2003; Larkey & Hecht, 2010; Huang & Shen, 2016), health communication research has called for the adaptation of health messages to culture (e.g. Kreuter et al., 2002). Culture shapes “the way we think, feel, and behave” (Huang & Shen, 2016, p.696), including the way we think and feel about health, and our health behaviors (Hoeken & Korzilius, 2003; Huang & Shen, 2016). Culture might even be essential for constructing an idea of what healthy is (Betsch et al., 2016). Kreuter et al. (2002) illustrate the influence of culture on health nicely by giving an example of the outcomes of a study by Klonoff & Landrine (1996). They (1996) found that African Americans are more likely to belief in the curing powers of prayer than other ethnic groups (p<.05). This belief has an effect on their exercise behavior; African Americans were significantly less likely to exercise on a regular basis than members of a cultural group that
were not religious (in chapter 3 I will explain the relationship between culture and healthy eating more in depth).
Because of the relationship between culture and health, researchers assume that the effect of a message aimed at changing health behavior will depend on the cultural background of the receiver. This is known as the cultural congruency hypothesis (Uskul, Sherman & Fitzgibbon, 2009). The common logic behind this hypothesis is that cultural differences between groups exist and adapting health materials to the cultural needs of the population is necessary since it will decrease disparities (Kreuter & McClure, 2004). Cultural minorities are often “less likely to receive relevant prevention and screening services” (Huang & Shen, 2016, p.694) compared to the dominant cultures. Resolving these disparities is therefore an important goal of health practitioners and health communication researchers in order to ensure health care for everyone within society.
Figure 1 shows a model of the place culture-sensitive health communication could have within society (Betsch et al., 2016). As is visible form the model, there is a close connection between medical sciences and health communication (or “behavioral, social, communication science”, as Betsch et al. (2016) call it). The model shows that health communication depends on medical sciences, and in order to reach the target population communication should be adapted to the cultural background of receivers.
Figure 1. The idealized process of culture-sensitive health communication as an evidence-informed
way of communicating evidence-based medical information, adapted to the cultural background of the message’s receiver. Reprinted from “Improving medical decision making and health promotion through culture-sensitive health communication: an agenda for science and practice,” by C. Betsch, R. Böhm, C.O. Airhihenbuwa, R. Butler, G.B. Chapman, N. Haase, B. Hermann, T. Igarashi, S. Kitayama, L. Korn, U. Nurm, B. Rohrmann, A.J. Rothman, S. Shavitt, J.A. Updegraff & A.K. Uskul, 2016, Medical Decision Making, 36, p. 815.
3. Does cultural adaptation of health materials work?
Although health communication researchers believe that cultural tailoring is an effective way of achieving attitude and behavior change, the empirical evidence supporting this claim is not overwhelming. In a recent systematic review, Huang & Shen (2016) created an overview of the persuasive effects from studies that examined whether culturally adapted materials about cancer outperform ‘plain’ materials that were not culturally tailored. The authors included 58 different research designs in their meta-analysis and conclude that “culturally tailored cancer messages had an overall small and significant influence on persuasion” with (r=.120, p <.001) (p.694). Factors that moderated the persuasive outcomes of the health materials were: cancer type, message design approach, message format, media channels, ethnicity, gender, and study design. There are several things to keep in mind when interpreting the conclusion of Huang & Shen (2016) that underline the significance of this thesis. First, Huang & Shen (2016) found that cancer type moderates the effects of the health materials, so it is unsure whether ‘healthy eating’ as a health subject would result in the same positive effects. Second, Huang & Shen (2016) only included 9 studies that incorporated a narrative design, while at the same time suggesting that narratives were especially effective in persuading people to change their health behavior (see more on this in Chapter 2 ‘Narratives’). Third, Huang & Shen (2016) might have evidence that culturally tailored messages outperform plain messages but do plain messages actually exist? Aren’t all messages using culture, if not intentional, unintentionally? And fourth, Huang & Shen (2016) did not prove whether culturally grounded messages could also be persuasive when someone did not match the cultural background of the health materials. In other words: only when including a match/mismatch design we can say with certainty that messages should be adapted to the cultural background of the receiver. Especially since the opposite might be just as plausible; receivers might find messages rooted in a different cultural background than their own just as, or even more, appealing. In the next chapter on narratives, this point is further explained while referring to work of Koops van ‘t Jagt et al. (2017).
4. Chapter summary: Defining culture
In this chapter, I discussed how culture can be defined within health communication research. Nowadays, researchers generally agree that culture is something dynamic, and the dimensions by Hofstede (1980; 2010) and colleagues continue to be important within health communication research. Health communication researchers suggest that applying culture in health materials is necessary since it will ensure that everyone within a multicultural society has a fair chance at being healthy, and several strategies are used in order to achieve the culturally appropriateness
of health materials. Although Huang & Shen (2016) systematically reviewed empirical studies on the effects of these culturally-sensitive health materials about cancer and found that these materials are more persuasive than the ‘untailored’ equivalent, there are some important considerations to be kept in mind when interpretation their conclusion.
Chapter 2: Cultural narratives
“
A narrative is like a room on whose walls a number of false doors have been painted; while within the narrative, we have many apparent choices of exit, but when the author leads us to one particular door, we know it is the right one because it opens.”
(John Updike, Introduction to The Best American Short Stories of 1984, 1984)
In the first chapter, I discussed what culture means in health communication research and how it relates to health materials. (Cultural) narratives will be the topic of the current chapter. First, I will explain what narratives are, which theories lie behind narrative persuasion and discuss empirical evidence for the persuasive effects of narratives in general. Then, I will explain what culturally tailored narratives are and how they fit in with general narrative persuasion theory. In the chapter summary, the important points of this chapter are summed up.
1. What are narratives?
Health promotion research has spent decades on finding ways to present health promotion messages that are most effective in convincing people of certain behavior. The use of narratives has shown promising results (e.g. De Graaf, Sanders & Hoeken, 2016). Reviewing several definitions used in literature to define a narrative, De Graaf, Sanders & Hoeken (2016) conclude that a narrative consists of at least the following: “a presentation of concrete event(s) experienced by specific character(s) in a setting” (p. 91). The authors refer to the character as an agent with specific goals and intentions and they characterize events as transitions of states which are causally connected. With ‘setting’ they mean the time and place the events take place in.
2. Are narratives able to persuade people?
In the first chapter, I discussed why culture could be an important contributor to health behavioural change. In the current chapter, I will discuss how narratives and fear appeals might convince people and thus bridge the gap between intentions to change and actual change. I will explain what I mean by “bridging the gap” here. In sum, health persuasion works as follows; people demonstrate a kind of health behaviour that is undesired since it effects people’s health negatively and health practitioners want to change that behaviour for the better by using a message. For example, health practitioners want to make people quit smoking (to prevent long
cancer), eat healthier (to prevent cardiovascular diseases) or perform a breast exam (to detect breast cancer early on). Unfortunately for health practitioners, they cannot force people to change their behaviour, so they try to persuade people by influencing their beliefs, attitudes and behavioural intentions through these messages. Here, it should be noted that theories behind health persuasion are all based on the premise that intentions, beliefs and/or attitudes people have about a certain health behaviour have an effect on the actual behaviour of people. In reality however, there seems to be an intention-behaviour gap. Sniehotte, Scholz & Schwarzer (2005) explain this gap by saying: “Although some people may develop an intention to change their health behaviour, they might not take any action.” (p. 143). People seem to encounter some sort of resistance towards the recommended behaviour. Health practitioners hope, by employing the narrative format to their message, and/or using the fear appeal strategy and/or accounting for culture, this intention-behaviour gap can be bridged, and this is exactly the aim of the current experiment.
Now, I will discuss some evidence for the effects of narrative persuasion. Narrative persuasion received much attention in health communication research, and indeed, research has shown that narratives are an effective tool in persuading people into healthier behavior (Strange & Leung, 1999; Diekman, McDonald & Gardner, 2000; Green & Brock, 2000; Slater, Rouner & Long, 2006; Appel & Richter, 2007; Morgan, Movius & Cody, 2009; Moyer-Gusé & Nabi, 2010). That is, as I discussed before, they have an influence on the intentions to change their behavior, and attitudes and beliefs that people have about a certain health behavior. From now on, I will just call these effects on behavior (as is common in health communication research). Narratives come in many forms, sizes and contexts. In order to investigate whether all these different types of narratives are able to persuade people into certain behavior, Shen, Sheer & Li (2015) conducted a meta-analysis on “the impact of narratives on persuasion in health communication” (p. 105). Their conclusion is that narratives had a small but significant positive influence on persuasion (r=.063, p<01), but that not all narratives were equally effective. Shen, Sheer & Li (2015) found that audio and video-narratives were effective, whereas print-narratives were not. Plus, print-narratives that promoted detection and prevention behaviors were significantly effective, but narratives that advocated cessation behaviors were not.
De Graaf, Sanders & Hoeken (2016) are in line with Shen, Sheer & Li (2015) when they conclude that narratives have a positive effect on persuasion. However, they also found a positive effect for print narratives and conclude that medium does not affect the effectiveness of narratives. Additionally, they (2016) distinguish several important moderators of these effects and found that the content, form and context of the narrative are important factors in
determining its success. Looking at the content of the narratives, the authors conclude that showing the healthy behavior rather than the unhealthy behavior was more effective for persuading people. It should be noted that this gain frame was only positively associated with intentions, not with attitudes and beliefs. Similarity of the surroundings of the characters with that of the reader was found to be effective, contrary to other forms of similarity. High expression of emotions in the narrative was also found to have a positive effect. For the form of the narrative, De Graaf, Sanders & Hoeken (2016) found that a first-person perspective was most effective. For the context, the authors conclude that presenting a narrative with a clear persuasive intent does not result in less effective narratives, contrary to what was assumed, for example by Moyer-Gusé (2008). She categorizes the messages she used in her research as “entertainment messages” and one of the key components of these entertainment messages is that they are not overtly persuasive. In other words, readers will not be made aware of the fact that the message is used to persuade them into different behavior. The entertainment messages have, as a main goal, to entertain people, not to persuade them. This, Moyer-Gusé (2008) theorizes, will help readers to be more accepting towards the message. But, as it turns out, De Graaf, Sanders & Hoeken (2016) found no support for this claim.
3. Narrative mechanisms
It has been established that narratives can influence the beliefs, attitudes and intentions of people, the question remains how narratives are able to do so. In other words, how do narratives work and what are the mechanisms underlying narrative persuasion? Narrative persuasion has been a heavily researched subject for quite some time (including Green, Brock & Kaufman, 2004; Moyer-Gusé, 2008; Murphy et al., 2013; Ooms et al., 2017; Jansen, Nederhoff & Ooms, in press) and research identifies the processes of transportation and identification as important mechanisms. Identification refers to the process of the reader identifying with the main character(s) within a narrative. Transportation deals with readers being transported into the world of the narrative. Transportation and identification are important since they ensure that readers are swept into the story and become emotionally involved with the story and its character(s) (Moyer-Gusé, 2008). This involvement with the story could help readers overcome possible resistance towards behavioral change and thus make the narrative effective.
This ‘involvement’, or ‘engagement’ with the story, seems to be key when talking about narrative persuasion. Yet, researchers do not agree on which elements of transportation and identification are especially relevant to achieve this, and if there are other processes that help achieve this involvement. Green & Brock (2000) for example, state that transportation consists
of attention (cognitive reactions while reading the story), emotions (feelings people have while reading the story) and imagery (imagination readers have while reading the story). Their definition and operationalization of the transportation is widely used among health communication researchers. For identification, most scholars use Cohen (2001). He (2001) suggests that identification consists of four dimensions: empathy (feeling the same emotions as the character), perspective (understanding the character’s perspective), motivation (sharing the goals of the character) and absorption (the degree to which self-awareness is lost while reading the narrative). Especially the last component Cohen (2001) identifies (absorption) demonstrates that it is not easy to distinguish between transportation and identification. One important consideration therefore is that identification always refers to the character of the story and transportation refers to the story as a whole.
Besides transportation and identification, other narrative processes have been identified. Moyer-Gusé (2008) talks about ‘entertainment features’. She includes transportation and identification as well as narrative structure, parasocial interaction (the relationship between the reader and the character(s) within the story), liking (positive evaluations the reader has of the character(s)), enjoyment (of the story) and perceived similarity (between reader and character(s)) (see Figure 2).
Figure 2. Entertainment overcoming resistance model (EORM). Reprinted from: “Toward a Theory
of Entertainment Persuasion: Explaining the Persuasive Effects of Entertainment-Education Messages,” by E. Moyer-Gusé, 2008, Communication Theory, 18, p. 415.
The EORM theorizes that the different entertainment features have an effect (being ‘reducing’, ‘increasing’ or ‘changing’) on different forms of resistance people demonstrate. In turn, reducing this resistance will lead to “story consistent attitudes and behaviors”.
Cohen (2001) explains why liking, similarity and affinity are not components of identification, but measure different constructs since they call for the reader’s judgement of the character. In other words, rather than feeling similar as the character, it refers to feeling similar
to the character (for example: “I feel just as sad as X” versus “I could be friends with X”). In
practice however, many scholars measure identification and include notions of liking, similarity and affinity.
Buselle & Bilandzic (2009) try to create some order in how researchers measure narrative engagement and come up with four dimensions. They suggest that narrative engagement is built upon narrative understanding, attentional focus, emotional engagement and narrative presence. The first dimension, narrative understanding, deals with how difficult the readers find the story. Attentional focus refers to how distracted a person is while he reads the story. A reader is emotionally engaged when the narrative they read evokes certain emotions. In other words: “feeling for and with the characters” (Buselle & Bilandzic, 2009, p.341). The fourth and final dimension deals with the narrative presence a reader experiences while reading the story. It is “the sensation that one has left the actual world and entered the story” (Buselle & Bilandzic, 2009, p.341). Buselle & Bilandzic (2009) draw from theories proposed by Green & Brock (2000), Cohen (2001) and Moyer-Gusé (2008) (among others), rather than coming up with yet another measuring proposal, still a variety of theories are used to measure narrative engagement in practice.
4. Narrative fear appeals
Since emotions seem to play an important role in getting readers interested in the story (Graaf, Sanders & Hoeken, 2016; Ooms, Jansen, Hommes & Hoek, 2017), many researchers combine fear appeals with narratives (Ooms, Jansen, Hommes & Hoeks, 2017). I will briefly explain how fear appeals work and why they might be a valuable addition to the narrative.
Witte (1992; 1998) designed the Extended Parallel Process Model (EPPM) in order to be able to explain how fear appeals might be persuasive (see Figure 3).
Figure 3. Extended Parallel Process Model (EPPM) (Witte 1998). Reprinted from: “Developing
persuasive health campaign messages,” by C. Jansen, 2017. In: A. Black (Ed.), L. Paul (Ed.), O. Lund (Ed.) & S. Walker (Ed.), Gower handbook of information design, p. 6.
The reason for integrating fear appeals in a narrative structure is to enhance emotional engagement. The EPPM (Figure 3) predicts that: first, the message should contain self-efficacy, response efficacy, susceptibility and severity. Self-efficacy means that the person receiving the message has to feel like he/she is able to do something about the threat presented in the fear appeal. At the same time, for a fear appeal to work, receivers should experience enough fear (response efficacy). Also, receivers should feel like the threat presented is relevant for their own lives (susceptibility) and the threat should be severe enough to take seriously (severity). Only if these four components are met, receivers will either accept the message (danger-control process). If these components are off, receivers will reject the message (fear-control process). If the recipient feels that he/she is able to carry out the proposed method for preventing the threat (perceived threat) and that that method is effective (perceived efficacy) only then will the recipient enter the danger control process and thus agree with the advocated behaviour. However, if the recipient scores low on perceived efficacy and/or perceived threat, they will not fight the danger but avoid it. In other words: the fear appeal did not work.
Research only partly supports the EPPM proposed by Witte (1992, 1998). In an overview of sixty years of empirical evidence for fear appeals, Ruiter et al. (2014) did not find all relationships within the EPPM supported. The authors (2014) conclude that self-efficacy is especially important to include in a fear appeal message. When recipients are able to subtract a
measure to avoid the threat presented by a fear appeal, the fear appeal is more effective. Furthermore, Ruiter et al. (2014) found that if severity is too high, recipients might get defensive and demonstrate the fear-control process, the opposite of the goal of the fear appeal. To avoid this, the authors (2014) suggest to always incorporate susceptibility and an explanation of the measures people should take in order to avoid the presented threat. In another attempt to gain insight into the workings of fear appeals, Popova (2012) put the EPPM to the test and found no support for any of the claims made by the model. She is especially concerned with the individual differences that causes the fear appeal to work differently per person (Popova, 2012; Jansen, 2017). In this thesis, one of the factors that contributes to individual differences is investigated: the influence of culture on the outcome of narrative fear appeals.
After reviewing these different studies into narrative persuasion, I can conclude that (1) narratives, in their many forms, can sometimes be effective in persuading people into healthier behaviors, yet (2) that the effects of narratives as health materials are still inconsistent and effect sizes rather small. Incorporating culture within narratives could increase and strengthen these effects. In the next sections, I will discuss the combination of culture and narratives.
5. Culturally grounded narratives and persuasion
The call for incorporating culture in designing health promotion materials is made by many researchers (e.g. Kreuter et al., 2002; Uskul, Sherman, Fitzgibbon, 2009; Larkey & Hecht, 2010; Huang & Shen, 2016). Larkey & Hecht (2010) propose a model of how culturally appropriate narratives might work (see Figure 4).
Figure 4. A model of culture-centric narratives in health promotion. Reprinted from “A model of
effects of narrative as culture-centric health promotion,” by L.K. Larkey & M. Hecht, 2010, Journal
of Health Communication, 15, p. 122.
In the previous chapter I discussed how culture can be integrated within health materials in general, Figure 4 shows how a culturally grounded narrative might persuade people into healthier behavior. As is visible from the model, Larkey & Hecht (2010) hypothesize that culture influences identification, transportation and social proliferation. With that last term, they mean “uptake and diffusion of ideas or behaviors through discussion, rehearsal, and reciprocal support (caring, obligation and group action)” (p. 120). The fact that narratives could also have an influence on discussing the health behavior and that these discussions might also have a positive influence on behavioral change is also noted by Van den Putte et al. (2011). They (2011) suggest that “interpersonal communication” as they call it, can be of use to health practitioners in two ways. First, “to stimulate change through social interaction” and in a more indirect way “to further disseminate message content” (p. 470). Their study (n=1079 smokers) supported both functions of interpersonal communication for smoking cessation.
6. Do culturally grounded narratives work?
As discussed in chapter 1, Huang & Shen (2016) reviewed the persuasive effects of the cultural tailoring of cancer-communication materials. They conclude that “culturally tailored messages had an overall small and significant influence on persuasion (r=.120, p<.001)” (p.694). Additionally, they found that messages with a narrative structure were most effective. Unfortunately, the authors (2016) included only nine studies that used a narrative design, so one should be cautious when interpreting the results. Additionally, recent studies that were not included in the overview throw some doubt on the generality of their conclusion. Koops van ’t Jagt et al. (2017) for instance found that a narrative on diabetes that was tailored to the minority of Spanish-speaking individuals in the US with a low level of literacy, had the persuasive effect that was aimed for in the Netherlands too. This effect was found for both a low literacy group and a high literacy group. A possible explanation for this result is that the use of a narrative made cultural adaptation rather superfluous: every reader can be transported into a narrative world without necessarily sharing the cultural values of the protagonists. This supports the point I made in chapter 1 where I stated that “[…] the opposite might be just as plausible; receivers might find messages rooted in a different cultural background than their own just as, or even
more, appealing” (p.12). That is why, in the current project, a mismatch/match design is used in order to explore this point further.
7. Chapter summary: Cultural narratives
In this chapter, I began with discussing what narratives are and how they work. Mechanisms behind narrative persuasion that are especially relevant are transportation and identification, although scholars mention other processes as well. In order to enhance emotional engagement, fear appeals may be integrated in narratives. Huang & Shen (2016) conclude that culturally tailoring health materials about cancer positively influences persuasion. Yet, publication bias calls for replication of experiments and a recent study of Koops van ‘t Jagt et al. (2017) found that cultural adaptation of s a story could also enhance the persuasive effects of a cultural group that is not targeted in the story. Combining strategies of narrative persuasion, fear appeals and making health materials culturally appropriate could help to bridge the intention-behavior gap.
Chapter 3: Healthy eating
“
Food is not rational. Food is culture, habit, craving and identity.”
(Jonathan Safran Foer)Thus far, I have discussed what culture entails and how narrative persuasion works, in combination with the mechanisms behind fear appeals. Additionally, empirical evidence for narratives, fear appeals and culturally tailored materials were discussed. Yet, I have not discussed the health subject relevant to this study: healthy eating. In this chapter, I will explain (1) what healthy eating is, (2) why it is a significant subject and (3) how eating is cultural. The important points relevant to this chapter will be summed up in the chapter summary.
1. What is healthy eating?
Health communication scholars use different names when they refer to ‘healthy eating’, probably since healthy eating is associated with different health behaviors. For example, healthy eating could be linked to ‘fruit and vegetable intake’, ‘exercising behavior’, ‘reducing sugar intake’, ‘reducing salt intake’, ‘(healthy) dieting’, ‘reducing soda intake’, ‘weight loss’, ‘reducing fat intake’, ‘lowering carb-meals’, (…). This makes sense since a study by Winter Falk et al. (2001) showed that people (n=75) have up to seven definitions when they talk about healthy eating. They state that healthy eating means (1) eating low fat, (2) eating natural/unprocessed foods, (3) eating balanced, (4) eating to prevent diseases, (5) eating to maintain a nutrient balance, (6) eating to manage an existing disease and (7) eating to control weight. Consequently, advising people to simply “eat healthier” could lead to several outcomes.
As the current study takes place in the Netherlands, the definition of ‘het Voedingscentrum’
(the Food Centre) will be used as a guideline within this thesis. Het Voedingscentrum is a
government-sponsored institute that provides Dutch citizens with awareness about how to eat healthy. Their slogan is “honest about food” and their definition of healthy eating: “sustainable and safe food” (Voedingscentrum, 2018).
2. Why is healthy eating relevant?
Cancer and smoking behavior have received the most attention in health communication research. However, many researchers begin to recognize healthy eating as a possible health subject. The reason to focus on healthy eating as a health subject in health communication
research is twofold. A practical reason is that studies have found that unhealthy eating habits are related to different health issues (including diabetes, cardiovascular diseases, different types of cancer, sleep apnoea, gall bladder disease, osteoarthritis, endocrine disorders, social exclusion and depression) (De Wit et al., 2015; Ujcic-Voortman et al, 2011; Lobstein, Baur & Uauy, 2004). From a theoretical perspective, it is important to focus on healthy eating as well since the health subject of the message is known to moderate its effects (Huang & Shen, 2016).
I will expand on the societal importance of healthy eating some more. In 2017, 14% of teenagers in the Netherlands was overweight (Volksgezondheidenzorg.info, 2018). Unfortunately, obesity rates in the Netherlands are higher than ever, and this trend is expected to grow even more, especially in higher economically developed countries such as the Netherlands (Lobstein, Baur & Uauy, 2004). In fact, mainly because of problems related to obesity, children born around 2000 might be “the first generation ever to have shorter life expectancy than that of their parents” (Dubé, 2010). Additionally, Turkish migrants in the Netherlands are commonly more overweight than the ethnic Dutch population in the Netherlands (Ujcic-Voortman et al., 2011), and are therefore a high-risk group, making them of special interest for this study.
Additionally, having problems related to eating often already starts at a young age. Still, it remains a challenge to motivate young people into eating healthier (Lobstein, Baur & Uauy, 2004). Studies also found that “the food preference patterns learned in childhood may influence long-term patterns of dietary intake” and “learned food preferences could either promote or impede the preference for intake of nutritionally sound diets” (Haire-Joshu et al., 2004, p. 309). In their report on ‘Obesity in children and young people: a crisis in public health’, Lobstein, Baur & Uauy (2004) state that prevention is the only feasible option in tackling the problems linked to over-eating. Changing the way we eat at a young age, and finding a way to appeal to all audiences in the best way possible could therefore have a significant impact on the health of many people. It could also strengthen the link between health communication research and health communication practice, as health communication research is a societal driven field.
3. How is healthy eating cultural?
Culture has an influence on our health behavior. Kreuter et al. (2002) nicely demonstrate how religion might influence health behavior. They (2002) give an example of research done by Los Angeles, Klonoff & Landrine (1996). These authors (1996) found that “believers were significantly less likely than nonbelievers to exercise regularly (39% vs. 61%, p<.05) and indicated less desire than nonbelievers to be actively involved in their health care (p<.05).”
(Kreuter et al., 2002, p.138). Kreuter et al. (2002) also give an example of how culture is related to dietary behavior when they discuss a study done by Campbell et al. (1999). In this research (1999) 2.519 African Americans received either an expert orientated or a spiritually orientated dietary bulletin from North Caroline churches. Although content was the same for both bulletins, message trust was significantly higher for the group that received the spiritually oriented bulletins. Although this does demonstrate the fact that tailoring a message to culture could lead to higher beliefs, it is perhaps not that noteworthy since it might be expected of people who go to church and receive bulletins from that church to trust those bulletins more when they are spiritual (since you would expect spiritual content from the church). It would be interesting if the same results were found when bulletins were handed out by hospitals (or other non-religious institutes), or, as is the case in the current project, from neutral institutions such as schools.
Campbell et al. (1999) demonstrated that culture could have an influence on health behavior. The same associating is established for eating behavior and culture. Many studies focus on the impact of culture on eating disorders (undereating), and try to explain for example, why white women are more prone to undereating than African American women (Rubin, Fitts & Becker, 2003). Although these studies confirm that there is a link between eating and culture, for the current project, the relationship between culture and over-eating is more relevant (as the aim of the narratives used is to lose weight, not to gain weight). I will focus on an article by Hubbard and collogues (2016) in some detail as (1) they give a nice overview of what research has said about the link between culture and eating thusfar and (2) performed an experiment that evaluates the reliability and validity of the Culturally-based Communication about Health, Eating and Food (CHEF) Scale to measure “the extent to which individuals’ culture, as they perceive it, influences perceptions of food-related health messages” (p. 399), which is of direct relevance to the current project.
Hubbard et al. (2016) give two possible explanations for the problematic relationship between overeating and culture. They state that in some cultures, over-eating is common (for example in Hispanic or African-American communities. In these communities, traditional foods are also high in fat and sodium, and therefore contrast dietary recommendations. Additionally, the authors suggest that culture also influences understanding of food and eating, confirming research on white women and their body image versus African American women and the way they view their bodies (Rubin, Fitts & Becker, 2003).
As I discussed in chapter 2, self-efficacy is an important predictor of effective outcomes of a fear appeal aimed at changing health behavior (Ruiter et al., 2014). Hubbard et al. (2016)
talk about ‘dietary self-efficacy’ and state that people who believe that they are able to eat healthily have a high dietary self-efficacy and consequently eat healthier than people with low dietary self-efficacy. The same goes for support people experience when making dietary decisions (the higher the support, the more influenced the choices). This support can either promote or demote healthy eating. In turn, dietary self-efficacy and support are related to culture as well (Hubbard et al., 2016).
Now that I have discussed the reasons Hubbard et al. (2016) mention for developing the CHEF-scale, I will discuss the results. Factor analyses showed five possible factors of influence on perception of messages about eating; (1) connection (the connection people felt to the foods of their culture e.g. “Cooking culturally representative foods connects me to my culture” (p. 402)), (2) authority (the perceptions that people have about health providers and how much they know about their cultural foods e.g. “My health care provider is probably familiar with the foods that are representative of my culture” (p. 402)), (3) unhealthy food perceptions and (4) healthy food perceptions (people’s perception that others might find their cultural foods (un)healthy e.g. “Teachers/educators probably consider the foods representative of my culture to be (un)healthy” (p. 402)) and (5) social value (the importance of cultural foods in social relationships e.g. “It is important that my romantic partners eat the foods that are representative of my culture” (p. 402)). Race/ethnicity had a significant effect on all factors except authority. So, the authors (2016) conclude that “Negative interactions or ignoring culture in these discussions [around obesity] might lead to increased race-related stress and related health problems”. In other words, messages aimed at changing health behavior already receive resistance (Moyer-Gusé, 2008), but changing eating habits in certain cultural groups might be an even bigger challenge.
4. Chapter summary: Healthy eating
In this chapter, I demonstrated that (1) healthy eating has no a clear-cut definition and research therefore investigates different aspects of healthy eating, (2) healthy eating is a very relevant persuasive health communication research topic, from a scientific and societal perspective, and (3) culture and healthy eating are related and this could negatively or positively influence the message recipient.
3. Method
3.1 Operationalisation
The central research question is “What effects do culturally grounded narratives about healthy eating have on narrative persuasion, intentions and attitudes of teenagers, who either match or mismatch the cultural background of the narratives?”. In order to answer this question, an experiment was designed using two versions of the same story (see Appendix A and “3.3 Materials”). Since evidence on culture as a moderator is mixed, no hypotheses were formulated. A questionnaire (Appendix B) was designed in order to measure the relevant variables. What follows is an overview of these variables and the way in which they are operationalized. The questionnaire consists of 71 questions about the narratives and 9 personal questions. Unless noted otherwise, the narrative-related questions use a Likert-scale from 1 (completely disagree) to 7 (completely agree) to avoid insensitive measure and forced choice (Choi & Pak, 2005). The internal consistency of the variables is also discussed, using Cronbach’s alpha.
3.1.1 Narrative mechanisms: transportation & identification
Since research generally agrees upon at least transportation and identification as important measures for narrative persuasion (Ooms et al., 2017; Moyer-Gusé, 2008), these variables were measured. Two validated questionnaires for transportation that are used most often are the Transportation Scale (Green & Brock, 2000) and the Narrative Engagement Scale (Buselle & Bilandzic, 2009). These methods were used for measuring transportation within this thesis as well. Two items were excluded due to irrelevance; since the narrative only consisted of one page, “The program/story created a new world, and then that world suddenly disappeared when the program/story ended” was not used (Buselle & Bilandzic, 2009, p. 337). Furthermore, (2) questions about how participant related the events to their everyday life were also excluded since this would only make sense if participants lost their father as well/were overweight (as the main character talks about being overweight and loss of his father). Green & Brock (2000) suggest that transportation consists of three components: attention (“while reading the story, I could imagine the events happening easily”) (Q1, Q2, Q4, Q15), emotions (“I felt sad, while reading the story”) (Q8, Q9, Q10) and imagery (Q16) (“I could imagine the events of the story as a movie before my eyes”). Q15 was negatively formulated (“While reading the story, I noticed my mind often wandered”) and was reversed for the Cronbach’s alpha analysis, which was satisfactory (.7).
For identification, items by Tal-Or & Cohen (2010) were used with Q12, Q14 (“I could be friends with Rick/Yassin), Q11, Q13 (“I look a lot like Rick/Yassin”) and Q5, Q6, Q7 (“I understand how Rick/Yassin must be feeling”). Cronbach’s alfa was satisfactory (.7).
3.1.2 Credibility of the story
To check whether or not participants found the story credible, they were asked whether the story they read was “over the top/not over the top” (Q17), “not credible, credible” (Q18), “unclear/clear” (Q19), “not understandable, understandable” (Q20). Cronbach’s alpha was good (.77).
3.1.3 Fear appeal: Self-efficacy, response efficacy, susceptibility & severity + perceived efficacy & perceived threat
The EPPM (Witte, 1998) proposes that a fear appeal consists of self-efficacy/perceived threat (Q32, Q33, Q36, Q43, Q47) (“I am able to lose weight”), response efficacy/perceived efficacy (Q26, Q27, Q28, Q29, Q30, Q31) (“Losing weight is a good way to prevent a heart attack when you’re too heavy”), susceptibility (Q24, Q25) (“There is a chance I will become too heavy”) and severity (Q21, Q22, Q23) (“I think that being overweight is really bad”). Items used were based on Witte, Cameron, McKeon & Berkowitz (1996) from their Risk Behavior Diagnosis Scale, as these items are still most commonly used in health communication research to measure fear appeal-constructs.
Cronbach’s alpha for self-efficacy/perceived threat was satisfactory (.68). Dropping Q43 (“Drinking less soda is easy for me”) would lead to a Cronbach’s alpha of .71, but as .68 is still satisfactory, no item was dropped. Cronbach’s alpha for response efficacy/perceived efficacy was good (.76) and Cronbach’s alpha for susceptibility was excellent (.9). For severity, Cronbach’s alpha was satisfactory (.67) and since dropping of a question did not yield in a higher alpha, no questions were dropped.
3.1.4 Intentions
Several behavioral intentions were measured. Intentions to lose weight in general (Q51, Q52, Q53) (“I intend to start losing weight in the following four weeks”) were measured. Fishbein & Yzer (2003) suggest that it is best to measure specific behaviors as well since “the most effective interventions will be those directed at changing specific behaviors” (p.168). That is why the specific behaviors to change (1) intake of candy/snacks (Q35, Q39) (“I want to eat less unhealthy foods in the following four weeks”), intake of fruit (Q37, Q38) (“I plan on taking fruit with me for a snack in the following four weeks”), intake of soda (Q41, Q42) (“I plan on quitting drinking soda in the following four weeks”) and exercising behavior (Q45, Q46) (“In the following four weeks I plan on exercising intensely for at least thirty minutes per week”).
Since participants were high school students, specific eating behavior was aimed at changing what students could eat at school and buy in the canteen, rather than changing their evening meals for example (as teenagers usually do not cook but one of their parents does). Cronbach’s alpha showed that these intentions could be grouped (.78). Dropping one or more items did not yield in a higher Cronbach’s alpha, so no item was dropped.
3.1.5 Attitudes
Attitudes towards losing weight were measured using three questions: “losing weight while being overweight is …” “unwise/wise” (Q48), “bad/good” (Q49) and “not important/important” (Q50). Cronbach’s alpha was excellent (.87). These semantic differential items were derived from De Hoog, Stroebe & de Wit (2008).
3.1.6 Talking about it
Van den Putte, Yzer, Soutwell, de Bruin & Willemsen (2011) suggest that talking about a recommended health behavior could serve as an indirect influence on cessation or change of that health behavior (p.470). That is why intentions to talk about the health behavior were also measured. Intentions to talk about weight and healthy eating with friends (Q68, Q69) (“I plan on talking about my weight with friends in the following four weeks”) and family (Q66, Q67) “I plan on talking about healthy eating with family (mother/father/brothers/sisters) in the following four weeks” had an excellent Cronbach’s alpha (.9).
Since culture might have an influence on social behavior (see Theoretical Framework -Chapter 1), the ease, frequency and liking of talking to family and friends were also measured because the ease, frequency and attitude towards talking with family and friends could be moderated by culture. The ease of talking to friends (Q60, Q61) “I think it’s easy to talk about my weight with friends” and family (Q54, Q55) “I think it’s easy to talk about healthy eating with my family (mother/father/brothers/sisters)” about weight and healthy eating had a good Cronbach’s alpha (.82). The frequency of talking to a friend about healthy eating and weight was consistent (Cronbach’s alpha=.83), but the frequency of talking to a family member about these subjects was inconsistent (.52). In other words, how much you talk to your family and friends about healthy eating and weight seems not to be related, so these intentions were separated in “frequency talking to friend” and “frequency talking to family”. How much you like to talk about healthy eating and weight is the same for family members (Cronbach’s alpha= .75), but how much you like talking to a friend about these subjects is not related (Cronbach’s alpha=.47). It was therefore decided to only include the “ease of talking” in further statistical analyses.
3.1.7 (Dis)liking the health subject
How much the participants (dis)liked the health subject was measured using two questions: Q70 (“When I encounter something about being overweight, I would rather not hear or see about it”) and Q71 (“I do not want to read anything about being overweight”). Cronbach’s alpha was good for these items (.77).
3.1.8 Behavior
Self-reported behavior was measured for behavior in general (Q34) “I almost never eat something unhealthy”, soda intake (Q40) “How often do you drink soda? I drink soda … never/very often” and exercising behavior (Q44) “I exercise every week intensely for thirty minutes”. Cronbach’s alpha was very low (.041), suggesting that these are three completely different health behaviors and should be treated as such.
3.2 Participants & schools
3.2.1 Characteristics of participants
In total, 65 teenagers participated in the study. Table 1 shows some characteristics of these participants.
Table 1. Characteristics of participants (n=65) based on their age, gender, cultural background and whether or not they had relatives that were overweight.
Gender Age Cultural Background Overweight relatives F= 27, M=38 14-16 (mean=15.32) Dutch=44, Other=21 No=46, yes=19
Preknowledge Perception weight Being overweight Thinking about a diet No= 7,
Yes=56, NA=2
“exactly right”=44 “little too heavy”=10 “little too skinny”=9 “much too heavy”=2
No=52, Yes=11, NA=2 No=53, Yes=11, NA=1
For ‘gender’, participants were asked whether they thought of themselves as a ‘boy’, ‘girl’, or ‘other’ (“Personal Questions (PQ) -Q1”). For ‘age’, participants could be “13 years or younger”, “14 years”, “15 years” or “16 years or older” (PQ- Q2). Cultural background was scored as Dutch when both “nationality” was “Dutch” and “other language” was scored as a “no” (PQ- Q3–Q4). If participants answered “do you speak another language?” with “yes, Gronings/Grunnegs”, this was also scored as the participant being Dutch since ‘Gronings/Grunnegs’ is a language/dialect spoken in the Netherlands. The same goes for when
participants would state “Frisian” as a second language. Participants were asked whether they had ‘overweight relatives’ with “is one of your family members overweight (mother/father/brothers/sisters)?” (PQ-Q7). Participants were also asked “Did you know that being overweight could lead to a heart attack before reading this story? (yes/no)” (PQ-Q9) testing the participants ‘pre-knowledge’ of the health subject. The ‘perception of weight’ the participants had of themselves was tested asking “What do you think of your own weight? (way too skinny/a little too skinny/exactly right/a little too heavy/much too heavy)” (PQ-Q7). Whether they were actually overweight was measured by computing BMI’s based on their length (PQ-Q5) and weight (PQ-Q6). BMI’s were then compared to the table used by the Dutch ‘Voedingscentrum’ (2010) in order to determine whether or not the participants were considered ‘being overweight’. However, some participants though of themselves of being “a little too heavy” or “much too heavy”, when in fact they had a perfectly normal BMI. The variable ‘thinking about a diet’ includes participants who viewed themselves as needing a diet, regardless of whether or not they actually needed it.
3.2.2 Schools
Participants were recruited from two different schools, the Ubbo Emmius school at Stadskanaal (province of Groningen) (two classes, n=19, n=21) and the Helen Parkhurst at Almere (province of Flevoland) (one class, n=25). All classes were at the VMBO/MAVO level and the high school students were in their third or last year of high school education.
3.3 Materials
3.3.1 List of materials
Several materials were used in order to carry out the study. What follows is a list of the materials used and where they can be found in the Appendix. All materials in the experiment were in Dutch. In this Method section however, questionnaire items and narratives are translated in English.
- The narratives (Appendix A) - The questionnaire (Appendix B) - Letter to the school (Appendix C) - Letter to the parents (Appendix D)
- Letter of approval from the ethics committee (Appendix E) - Introduction to the study (verbal) (Appendix F)
