Rather being the victim or the culprit?
An experimental study on the effect of story perspective on the attitude towards and behavioural intention to diminish or stop mobile phone usage in traffic, mediated by identification, transportation
and emotion, and moderated by the degree of distressing imagery.
Yenny Luu 10573178 Master’s Thesis
Graduate School of Communication
Master’s Programme Communication Science – Persuasive Communication prof. dr. Bas van den Putte
2 Abstract
The objective of this study was to investigate the effect of a story presented from a victim’s compared to a culprit’s perspective on the attitude towards the message, attitude towards the provoked behavior and behavioral intention to diminish or stop mobile phone usage in traffic. In addition, the current study researched whether this effect was mediated by identification with the main character, transportation in the story and if the effect was moderated by the use of high compared to low distressing imagery. Participants (N = 172) completed an online survey that included a narrative in which the perspective of the main character was
manipulated (victim perspective vs. culprit perspective). Narratives also included distressing imagery in which the degree of distress was manipulated by presenting either low or high distressing images. The research findings showed that presenting a story in a victim
perspective had an influence on attitude toward the message, only when identification took place with the main character. No significant (mediation) effects were found of story perspective on attitude toward the message, attitude toward the behavior and behavioral intention. In addition, results indicated that the degree of distressing imagery did not moderate the effect of story perspective.
Keywords: story perspective, narrative persuasion, identification, transportation, emotion
3 Introduction
According to Stichting Wetenschappelijk Onderzoek Verkeersveiligheid (SWOV), a Dutch scientific research institute in the field of traffic safety, 62 percent of the Dutch
motorists have used their phone while driving at least once over a period of one year (SWOV, 2017). Although this behaviour might not seem very harmful, almost 80 percent of the traffic accidents and 65 percent of near-accidents are caused by forms of distractions or inattention in the three seconds prior the (near) accident (Neale et al., 2005). This is one of the reasons why numerous governments have decided to introduce (stricter) traffic rules concerning the use of mobile phones while driving.
The Dutch law, for example, forbids drivers of motor vehicles to hold a mobile phone and make a (handheld) call, send or read a text message while driving. However, although handsfree calling is allowed, it does not mean it is without risks due the loss of concentration. According to Ranney et al. (2000), the driver’s attention to the task of driving can be taken away by four different forms of distraction. The first distraction is known as visual distraction (e.g., looking at a phone screen instead of the road). The second type of distraction are
auditive distractions, such as responding to call and message signals that come from a mobile phone. The third form of distraction concerns the biomechanical distraction (e.g., creating text messages). And finally, the cognitive distraction occurs when the driver’s focus and attention is not on the driving task but is rather drifting away (e.g., being lost in thought during a conversation on the phone). Taken these forms and given examples of driver’s distraction into consideration, it is marked that the use of mobile phone touches all four distracting forms.
Given the risks and dangers of mobile phone use in car traffic, it is extremely
important to have a strong focus on prevention of this behavior in order to reduce the numbers of traffic accidents and fatalities. The use of narratives in health communication is a common strategy to stimulate certain desired health behaviors or to prevent or even stop certain
unwanted behaviors (De Graaf, Sanders & Hoeken, 2016; Moyer-Gusé, 2008). Because a story has a narrative structure, the recipient becomes involved in the story. As a result,
identification with the story and the characters is facilitated and the resistance to persuasion is reduced (Moyer-Gusé, 2008).
In the current study two specific story aspects will be examined that have an influence on the persuasion of narratives: the story perspective and images. The story perspective of a narrative is also known as the perspective in which the story is told. Common perspectives are
4 the first or third person, but the story perspective could also represent the main character’s role in the story. For example, a story could be told from a villain’s or hero’s point of view, or could be presented from a culprit’s or victim’s perspective. According to Hoeken, Kolthoff and Sanders (2016) an important mechanism of narrative persuasion is the identification with the story’s character. In their study two groups of participants had to read a story in which the perspective was manipulated by describing the events as experienced and narrated by the lawyer of a culprit and a widow, a heir of the victim. It seemed that the story perspective influenced identification with the character, which subsequently had a mediated effect on the participant’s attitude. The perspective of the story can therefore contribute in the identification process with specific characters and can contribute to the shifts in knowledge, attitudes and behavioral intentions (Murphy, Frank, Chatterjee, & Baezconde-Garbanati, 2013). Although the relationship between story perspective and identification has been examined in prior research, the effect of culpability of a character has not specifically been tested before. It is therefore interesting to find out whether readers would identify themselves more with the culprit or victim in the story, to give insight to future studies on the effects of story perspective.
Another possible feature that might have an effect on the persuasive power of narratives is the use of images. Previous studies have shown that health promotion messages including
distressing images can convince people to show healthier behavior (Witte & Allen, 2000; Tannenbaum et al., 2015). However, emotionally confronting people with the negative effects of their behavior can also lead to countereffects (Kok, 2016). According to Kok (2016) a positive behavior change takes place when there seems to be a high threat and a high self-efficacy, that is, the belief that you can do something yourself to prevent or stop the threat. Despite the fact that many health communication studies have examined the effect of
narrative persuasion and many scholars have focused on the effect of using distressing images (e.g., in fear appeal), little to no studies have examined the combination of storytelling and the use of distressing imagery in narratives to affect attitudes and behavioral change. Therefore, this study takes a first step into combining these different persuasion strategies with the aim to optimize the effect of health communication.
Although various meta-analytical studies showed inconsistent effects of narratives (Shen, Sheer & Li, 2015; Zebregs, Van den Putte, Neijens, & De Graaf, 2015), stories would be a promising intervention strategy for tackling unhealthy behaviors. By examining the effect of story perspective (in particular: victim vs. culprit perspective), combined with the power of
5 distressing imagery, the effect of narrative persuasion is examined from another angle of approach, with the aim to make it more successful. In this way, future research on narratives may have more positive effects and health communication can be optimized for
communicating and achieving health objectives.
By conducting this study, for instance, governments, health institutions and insurances know whether health campaigns need to have a victim or culprit story perspective and
whether this has to be combined with low or high distressing images, to raise awareness of the dangerous consequences of calling, texting or using mobile phone applications while driving and persuade motorists to not execute this behavior. Thus, besides contributing to closing the scientific gap and supporting scientific evidence for the effects of narratives, this study may also contribute to social purposes. It aims to gain more insight into the persuasive effects of narratives on attitudes, behavorial intention and the role of perspectives by attempting to answer the following research question:
‘What is the effect of story perspective (culprit vs. victim) on the attitude toward the message, attitude toward the behavior and behavioural intention to diminish or stop mobile phone usage in traffic? And will this effect be influenced by identification with the main character, transportation in the story and the use of distressing imagery (low vs. high)?’
Theoretical background
How we connect with a story’s character
In the past decade, scholars have examined the persuasiveness of narratives in health communication. As many studies have indicated the positive effects of narratives, the interest in elements of narratives (e.g., main character’s characteristics, behaviors or story outcome) has grown in order to better understand how narratives can influence individuals’ attitudes, and even behavior (De Graaf, 2014). An attitude is also known as an evaluation of people, objects, events, activities, and ideas and can range from extreme negative to extreme positive (Wood, 2000). Another predictor of bahavior, and an indicator of message effectiveness, is the attitude toward the message (Bhaduri, Ha-Brookshire, & Leshner, 2017; Mitchell & Olsen, 1981). Attitudes are found to precede behavior in the persuasion process and are often antecedents of behavior change (Hoyer, MacInnis, & Pieters, 2012; Petty & Cacioppo, 1986). Beside having a positive attitude towards the message and a positive attitude towards the
6 behavior, the behavioral intention is an important predictor of behavior (Ajzen, 1991). It is suggested that the stronger the intention to engage in certain behavior, the more likely it will be to execute the behavior.
One of the elements in narratives is the feeling to be more connected to a specific character than others in the story. Robinson and Knobloch-Westerwick (2017) suggest three mechanism that could explain how individuals connect to story characters: similarity, identification and likability. The first mentioned mechanism, similarity, can be explained by Bandura’s social cognitive theory (Bandura, 2004), which states that similarity between the audience and the model who performs the desired behavior is an important factor for
behavioral change. According to Cohen (2001) perceived similarity can also be defined as the belief an individual has of how similar he or she is to the presented character. However, what types of similarity are of importance to bring about the desired effect? Previous studies indicated that recipients would feel more similar to characters who have corresponding demographic characteristics as themselves, such as age, gender and race (e.g., Appiah, 2001; Austin, Roberts, & Nass, 1990). Other less personal characteristics, such as personality, behavioral tendencies or life experiences, are also factors that could influence the perceived similarity of readers or viewers on story characters (e.g., Murray, 1999).
Secondly, identification explains why we feel connected to the character in a narrative. Identification is a common discussed mechanism in narrative persuasion, as it can make an individual lose track of himself and temporarily experience the character’s perspective, beliefs, attitude and goals (Moyer-Gusé, 2008). This can consequently lead to a change in one’s own attitude, perceived vulnerability (Moyer-Gusé & Nabi, 2010) and even an increase in one’s self-efficacy to enact in a certain health behavior or to prevent unhealthy behaviors (Moyer-Gusé, Chung, & Jain, 2011).
Finally, Raney (2004), states that people are more likely to identify themselves to a sympathetic character. This was also proven by Tal-Or and Cohen’s study (2010), in which their research participants identified themselves more strongly with a sympathetic character than with an unsympathetic character. Another study manipulated the likability of the main character in the narrative, to see whether the attitude toward the message would be more positive after reading the version with the likable character compared to the less likable character (Hoeken & Sinkeldam, 2014). The results confirmed that participants who were in the likable condition indeed held more positive attitudes compared to those who were in the less likeable condition.
7 Assuming that there is somewhat of similarity between the participants and the
narrative’s main character, namely both hold their drivers license and are able to drive a car, the culpability perspective of the story will be manipulated in this study to see if we are more inclined to identify ourselves with a culprit or victim character in a narrative. Given the results of the previous study, it is expected that the recipient will identify himself less with the culprit compared to the victim, as the culprit is less sympathetic and likable than the victim. Therefore, the following hypothesis will be assumed:
H1: A story told from a victim perspective leads to a more positive of (a) attitude towards the message, b) attitude towards the provoked health behavior and c) behavioral intention than a story from a culprit perspective.
The process of processing narratives: identification and transportation
As stated before, many prior studies showed that narratives can influence the beliefs, attitudes, intention and even the behavior of their audience (e.g., Appel & Richter, 2007; Green & Brock, 2000; Moyer-Gusé & Nabi, 2010). But how do these effects come about? To understand the power of narratives, it is therefore important to examine the underlying
process of processing stories. A reason why narratives are able to bring out these effects is because of how we process narratives and their unique ability to reduce message
counterarguing (Moyer-Gusé, 2008; Slater & Rouner, 2002). According to the extended elaboration likelihood model (E-ELM) engagement in the dramatic elements of an entertainment program leads to a state of less critical but more immersive engagement (Shrum, 2004).
The E-ELM is an elaborated theory from its predecessor the elaboration likelihood model (ELM), which is a dual process theory by Petty and Cacioppo (1986) and states that the degree of attitude and behavioral change is related to the way a persuasive message is
processed. According to this theory there are two main routes to persuasion, namely, the central route and the peripheral route. The central route can be described as the route in which the provided information and arguments in the message are carefully and thoughtful
considered by the receiver, and therefore involves a high level of message elaboration and cognition. The peripheral route, on the other hand, requires less conscious thinking as the receiver relies more on heuristic cues when processing and elaborating a message. Hence, an attitude change via central processing results in a relatively more resistant and predictive behavior compared to an attitude change via peripheral processing. However, whether a
8 message is processed via the central or peripheral route depends on the motivation
(involvement) and ability of the receiver to elaborate on a message. If one of the two is impaired, it is predicted that the receiver will engage in peripheral processing.
Given the theory of the ELM, it could be said that when reading a narrative we are relying more on heuristic cues and processing via the peripheral route, rather than the central route. For example, when reading a novel you find yourself more empathizing with the occurring events and feelings of the main character, rather than scrutinizing and elaborating the story. E-ELM explains this by stressing two main components of engagement that should reduce counterarguing: identification and transportation. The first component, identification, describes how readers lose their own identity and perceive the identity of the story’s character as their own by being immersed into the character’s perspective (Cohen, 2001). When
identification with a character in the story takes place, recipients may imagine themselves being that character and do, think or feel something they normally would not (Moyer-Gusé, 2008). Identification, therefore, not only leads to evoking and experiencing emotions (e.g., empathy and affinity) with the specific character, but it also leads to adopting the same goals and plans of the character in the narrative (Zillmann, 2006).
Prior studies have proven the role of identification in the persuasion of narratives. De Graaf et al. (2009), for example, showed that their scale with items about taking the
character’s perspective and items about empathizing with the character significantly
correlated with story-consistent attitudes. However, correlations between identification and attitudes are unfortunately not enough to state that identification is a predictor of attitudes and it leads to a significant change. As De Graaf et al. (2012) stated, no causal effect of
identification on narrative persuasion were tested; it was therefore not possible to conclude whether this effect was caused by identification or the opposite relation. For example, when a reader holds similar attitudes as the story’s character, the similarity in attitudes may cause him to identify with the character instead of the other way around. Therefore, another study was conducted to examine the causal effect of identification on narrative persuasion (De Graaf et al., 2012). In their study, identification was manipulated by presenting a story from different character’s perspectives, with both characters having opposing goals. Beside finding positive effects of perspective on identification and attitudes that were consistent with the story, De Graaf et al. also found identification to mediate the effect of perspective on attitudes. These findings were also supported by the study of Hoeken and Sinkeldam (2013). To contribute to the further examination of identification as a mechanism of narrative persuasion, this study
9 will test whether identification mediates the effect of a manipulated culpability perspective in a narrative. The second hypothesis is therefore:
H2: The effect of story perspective on (a) attitude towards the message, (b) attitude towards the behavior, and (c) behavioral intention is mediated by transportation, and will be stronger as more identification takes place with the main character.
In contrast to identification the second component of engagement is transportation, which describes that the recipient is not immerse himself in one character’s perspective, but rather immerses himself into (an event in) the story. Green and Brock (2000) define transportation as the degree to which the recipient is cognitively as well as affectively completely immersed in a narrative. As a result, attention should be fully engaged and emotional responses should occur consistently with the presented experience of the fictional events. Research has shown that after being exposed to texts, transportation can contribute to the change of attitudes and beliefs (Green & Brock, 2000). In addition, Slater and Rouner (2002) have also showed that transportation is associated with reducing counter-arguing, which led to an increased
persuasion. It was also assumed that transportation into a story would lead to an increase of identification with characters and by experiencing the experiences of those characters, changes in attitudes would occur (Tal-Or, Cohen, 2010).
According to E-ELM, identification and transportation are two important underlying components for narrative persuasion, and both can be seen as ways to explain how readers become engaged and involved with narratives. Although identification is specified as a strong connection to a character as reflecting yourself to the character, transportation focuses on the engagement in the story, but it does not specifically specifies to what a reader is engaged with in the story (e.g., suspense, degree of presented details of the story). It is therefore
questionable to which extent these two constructs are associated with each other, and whether or not they could exist without the other. Green et al. (2004), for instance, have tried to manipulate transportation by altering the amount of details in the narrative, and their results showed that presenting a more detailed text would lead to more transportation. However, an increase of detailed text also means an increase of the length of the text, which consequently not only would result in more transportation but could also increase the ability to identify with the main character. By conducting this study, it will therefore be examined whether we can ascribe the possible effect of story perspective to either identification or transportation, or the power of both mechanism together. In addition to hypothesis 2, we will therefore assume:
10 H3: The effect of story perspective on (a) attitude towards the message, (b) attitude towards the behavior, and (c) behavioral intention is mediated by transportation, and will be stronger as more transportation takes place within the story.
Emotion as mediator
Beside identification and transportation, this study also focuses on another mechanism of the persuasion process of narratives, namely emotion. As narratives typically address emotion-evoking events (e.g., a loss of someone special) and identification with the character takes place, good stories should evoke simultaneous emotional responses in audiences (Nabi & Green, 2014). According to Morgan, Movius and Cody (2009) emotional involvement in a narrative increases attention to the story elements, the amount of imagery viewers generate, and increases cognitive processes. Prior studies supported this and have found emotion to play an important role in explaining the persuasiveness of stories. Buselle and Bilandzic (2009), for instance, found a strong relationship between emotions evoked by stories and participants’ attitudes. However, it has been proven that evoked emotions not only lead to changes in attitudes. The study of Murphy, Frank, Chatterjee and Baezconde-Garbanati (2013), for example, showed that by presenting narratives with characters similar to recipients led to more transportation, more identification and stronger experienced emotions. This
consequently contributed to a shift in participants’ knowledge, attitudes and behavioral intention.
Green and Brock (2000) believe an increased emotional response is necessary for transportation into a narrative. However, another study found that subsequent behavior change was predicted by positive and negative emotional responses to a cancer-related storyline and that this effect was independent of the level of transportation with the narrative and
identification with the main character (Murphy et al., 2011). These results, therefore, suggest that despite emotion and transportation are related, they should be considered as distinct constructs and emotion should be examined as a mediator in itself.
According to a discrete emotion perspective, each emotion has its unique causes and consequences (Lench, Flores, & Bench, 2011). Although most studies look at responses differentiated on a hedonic level, such as positive vs. negative affect (Murphy, Frank, Chatterjee & Baezconde-Garbanati, 2013), this current study will take a focus on negative emotions only. This is because the current study is interested in whether the culpability of the main character in a narrative has an effect on attitudes and behavioral intention. With regard to culpability, it is self-evident that the narrative will present unfortunate events that evoke
11 negative emotions rather than positive emotions. Nabi (1999) discusses several negative emotions, that seem to have unique functions and have often been studied to specify the appraisal patterns and action tendencies that are associated with each of these emotions. The negative emotions concern: fear, anger, sadness, disgust, guilt/shame.
By including emotion into the current study, it will be examined whether emotion is an important condition for narrative persuasion or whether it is just a side effect of stories. In particular, the current study will test whether (negative) emotion mediates the effect of story perspective. Thus, the following hypothesis is assumed:
H3: The effect of story perspective on (a) the attitude towards the message, b) the attitude towards the desired behavior and c) the behavioral intention to decrease or stop mobiles smartphone usage in traffic is mediated by emotion.
Beyond verbal communication: distressing images as moderator
Health campaigns using mass-reach methods often include distressing images to their message, such as imagery of severe injuries, impaired or diseased organs, that visualize physical and emotional trauma (Slater, 1999). These images are used with the intention to draw attention to the messages (Baron, Logan, Lilly, Inman, & Brennan, 1994), to elicit emotional response and stimulate prevention or cessation of risky and unhealthy behavior (Timmers & Van der Wijst, 2007).
Applying distressing images to the health messages is common in the use of negative persuasion appeals, which are persuasive messages that display threats and attempt to arouse negative emotions such as fear or disgust (Witte & Allen, 2000). By evoking these negative emotions, distressing images can motivate people to take a particular action, and thus could lead to attitudinal, intentional and behavioral changes. In Witte and Allen’s meta-analyses, results indicated that the stronger the fear appeal, the more effective the message appeal is. Kok (2016) also states that a positive behavior change takes place when there seems to be a high threat. However, the high threat must be accompanied by a high self-efficacy, that is, the belief that you can do something yourself to prevent or stop the threat.
Despite most of the fear appeal studies concerning distressing images have been done in the field of health communication, the effects of distressing images on the persuasiveness of narratives have hardly been studied before. Although audiences normally might respond defensively to distressing images, and fear appeal in general (Brown & West, 2015), it is
12 H4 H2 H3 H2 H3
assumed that the cognitive, affective and behavioral responses of distressing images, will have an effect on attitudes and behavioral intention. This is because narratives, on their turn, can transport the recipient into the flow of the story, leading the recipient to process via the peripherical route. As a result, the recipient will not be inclined to process and elaborate all given information critically as the effect of narrative persuasion and the narrative
involvement. Thus, the final hypothesis is:
H5: The degree of distressing imagery moderates the effect of story perspective, that is, high (versus low) distressing imagery leads to (a) more positive attitude towards the message, b) more positive attitude towards the desired behavior and c) higher behavioral intention towards decreasing or stopping mobile smartphone usage in traffic.
Figure 1. Theoretical conceptual model and hypotheses. H4
H5
H1
Degree of distressing imagery: Low vs. High
Story perspective: Culprit vs. Victim
a) Attitude message
b) Attitude health behavior c) Behavioral intention
Transportation Identification
13 Method
Participants
A total of 172 participants took part in the online experiment and were recruited via the researcher’s personal network and social media (e.g., e-mail, Facebook, WhatsApp and LinkedIn). Those who had a reading time of less than 60 seconds were excluded from the data, as it was assumed they did not read the story thoroughly. Therefore, the remaining number of participants for this study was 159. Participants had an age range from 18 to 76, with a mean age of 35.30 years (SD = 15.16). More women (56.6%) than men (43.4%) participated in the experiment. The majority (62.9%) of the participants were highly educated and attended or finished university. The remaining percentage of the participants indicated senior secondary vocational education (23.9%), secondary education (11.3%) or other (1.9%) as their highest level of education. To be eligible to take part in the study, participants had to be able to read in Dutch, own a mobile (smartphone) device, be older than 18 years and hold their drivers license. Despite the fact that young people in the Netherlands are allowed to get their driving license from the age of 17, it is very unlikely that they use their mobile phone while driving as they must drive under the supervision of a coach until the age of 18.
Therefore, although participants of 17 years old did own a drivers license, they were excluded from this study.
Design and procedure
This study consisted of a 2 x 2 factorial design, with story perspective (2 levels, experimental factor: victim vs. culprit) as well as distressing imagery (2 levels, experimental factor: low vs. high) as between-subject variables. The participants were randomly assigned to one of the four conditions: (1) victim perspective – low distressing images (n = 38), (2) culprit perspective – low distressing images (n = 45), (3) victim perspective – high distressing
images (n = 33), or (4) culprit perspective – high distressing images (n = 43). The variables identification, transportation and emotions were measured as mediators, with the attitude toward the message, attitude toward the behavior and behavioral intention as dependent variables.
After the study has been approved by the ethical commission, participants were invited on several social media platforms to click on a link to the online survey tool Qualtrics
(http://www.qualtrics.com) where they could participate in the study. In order to participate in the experiment, participants had to read the informed consent and declare that they have
14 understood everything clearly. After providing the informed consent, they were randomly assigned to one of the conditions of the experiments and asked to individually read a story that has been drafted by the researcher. When finishing reading the story, participants were asked to fill out a questionnaire regarding the story and their demographic characteristics. Materials
Two versions of a narrative were developed about someone who gets involved in a serious traffic accident due to mobile phone usage while driving and insufficient attention to the road. The narrative was based on two true stories, which were told by a victim and a culprit of a severe car accident due to mobile phone usage while driving. Depending on the condition the participants have been assigned to, the perspective in which the story was manipulated by describing the events as experienced and narrated by the culprit of the serious traffic accident or the victim. Both versions were written in the first perspective and in both versions the main character was driving a car to work. Small alterations were made in the story to match the perspective of the victim’s or culprit’s character, resulting in a word count of 541 for the culprit version and 556 words for the victim version. In addition to the framed perspective of the story, two images were presented along the story to give a picture of the event or scene and manipulate the degree of distress. The images varied between high and low distressing images and were collected from the internet by searching for words such as
‘severe car accident’, ‘broken leg compound fracture’ and ‘leg in plaster’. In the high-distress images condition, the story was accompanied by photographic portrayals of severe traffic and health consequences. An image that was used showed a scenery from a severe traffic accident with employees of emergency services and a badly damaged car, where it should be obvious that the driver would also have been severely injured. Another used image showed a bloody, open fracture of the leg. Images for the low distress condition were collected by matching them thematically to a high-distressing image. For example, in the low distress condition the scenery of a severe car accident only showed the emergency services from a distance and the badly damaged car was not shown clearly, and the image of an open fracture of the leg was replaced by a picture of a leg in plaster. Depending on the conditions participants were either presented a story from a victim perspective with low distressing images (see Appendix 1a), a culprit perspective with low distressing images (see Appendix 1b), a victim perspective with high distressing images (see Appendix 1c) or a culprit perspective with high distressing images (see Appendix 1d).
15 Instrumentation
The questionnaire contained variables concerning the attitude toward the message, attitude toward the behavior, intention to use the smartphone while driving, identification with the main character, transportation, emotion and a manipulation check (see Appendix 2).
Attitude toward the message. Attitude toward the message was measured by five items on a 5-point semantic differential scale, which were based on the operationalizations of Biehal, Stephans and Curlo (1992) and Kim, Baek and Choi (2013). The items were based on the statement “I consider the story”, and were scored (1) bad to (5) good; (1) unpleasant to (5) pleasant; (1) negative to (5) positive; (1) unrealistic to (5) realistic; (1) boring to (5) interesting. Principle component analysis (PCA) with varimax rotation showed that the item “I consider the story (1) unpleasant to (5) pleasant” did not load with the other items. In addition, the Cronbach’s alpha of the five items was very low (Cronbach’s alpha = .56). The item was therefore deleted from the scale and the final scale used for analysis included four items (Eigenvalue = 1.95, explained variance = 48.71%). Although the Cronbach’s alpha increased after deleting one item, the items formed a moderate and questionable scale
(Cronbach’s alpha = .63). The means of the remaining four items were, however, used to form the construct attitude toward the message for subsequent analyses (M = 3.67, SD = .63).
Attitude toward the behavior. A composite scale to measure the attitude toward the behavior was created using five items, adapted from Nemme and White (2010) and Van Leeuwen, et al. (2017), “For me to use my smartphone while driving a car in the coming month would be”. Items were scored (1) bad to (5) good; (1) unwise to (5) wise; (1) unpleasant to (5) pleasant; (1) very dangerous to (5) hardly dangerous; (1) negative to (5) positive. Items were coded so that higher scores indicated a higher positive attitude toward [driving and mobile phone] use. The five items formed a reliable scale (Cronbach’s alpha = .94). Therefore, the mean of these items was used as a scale that covers the attitude toward the behavior (M = 1.43, SD = .82).
Behavioral intention. To measure the behavioral intention, two items were adapted from Nemme and White (2010): “I intend to drive without looking at my smartphone while driving a car in the coming month.”, and “It is likely that I will use my smartphone while driving in the coming month.”. Items were scored (1) strongly disagree to (5) strongly agree. Intention to use the smartphone while driving was also measured by using the scale by Banerjee and Greene (2012). This contained the question stem “How likely is it that you will
16 use your smartphone while driving soon?” with three different items: “Once or twice in the coming month.”, “Once in a while in the coming month.” and “Almost every day in the coming month.”. Items were scored (1) very unlikely to (5) very likely. Items were recoded in advance so that high scores corresponded to a high intention of smartphone usage while driving and were reliable (Cronbach’s alpha = .81, M = 1.6, SD = .85).
Identification with the main character. Identification with the main character was measured by using eight items of De Graaf et al. (2012). Examples of the presented items are: “When I had been reading for a while, it seemed as if I had become the main character in my thoughts.”, “I empathized with the main character”, “While I was reading, I pictured what it would be like for the main character to experience what was described.”, and “I put myself in the position of the main character.”. Items were scored (1) strongly disagree to (5) strongly agree. The items formed a reliable scale (Cronbach’s alpha = .92). The construct
identification with the main character was therefore formed by calculating the mean of these eight items (M = 3.14, SD = 1.06).
Transportation. To measure transportation, 13 items were used of the transportation scale by Green and Brock (2000), with responses reported on a 5-point scale ranging from (1) strongly disagree to (5) strongly agree. Some presented items were: “While I was reading the story, I could easily imagine the events in it taking place.”, “My thoughts wandered off while reading the story.”, “While I was reading the narrative, activity going on in the room around me was on my mind.”, and “I could picture myself in the scene of the events described in the narrative.”. Items were recoded in advance so that high scores corresponded to a high degree of transportation. These thirteen items formed a reliable scale (Cronbach’s alpha = .83). By taking the means of these items, the construct transportation was formed (M = 3.32, SD = .67).
Emotion. Emotions evoked by presenting the narrative in different story perspectives were measured by eight items on a range of (1) strongly disagree to (5) strongly agree to express how the story made them feel (Murphy et al., 2013). The presented items were: “The story made me feel happy/angry/sad/disgusted/surprised/scared”. The item “The story made me feel happy” was recoded in advance, so that high scores of all items corresponded to a high degree of negative emotions. All items formed a reliable scale (Cronbach’s alpha = .72), and were computed to form the construct emotion (M = 3.45, SD = .68).
17 Manipulation check. A manipulation check has been performed to check whether the stimulus material had the desired effect. To see if the participants perceived the main
character as more culpable in the culprit condition than the main character in the victim condition, two statements have been asked: “The main character is responsible for the
accident.”, and “The protagonist could not do anything about the consequences of the event in the story.” (Hoeken & Sinkeldam, 2014). The items formed a reliable scale and were
computed to the variable character culpability (Cronbach’s alpha = .79, M = 3.43, SD = 1.65). In addition, the degree of the distressing imagery in both conditions were measured and checked by the following statements: “I found the images in the story distressing.”, “I found it annoying to look at the images.”, and “The images evoked negative feelings in me.”. All the statements were scored on a range of (1) strongly disagree to (5) strongly agree. The items formed a reliable scale (Cronbach’s alpha = .87). The two items were therefore computed to the variable distressing imagery (M = 2.79, SD = 1.22).
Demographical characteristics. Participants were also asked to answer questions regarding their gender (male/female/other), age, highest level of education and whether they were in possession of a driving license (yes/no). The variable highest level of education was recoded (into low/medium/high educated), in order to make the distribution of the groups relatively more equal for analyses. Participants who did not meet the research requirements, such as having a driving license, were excluded from the data.
Data analyses
The collected data were analysed by using several analysis in the statistical program SPSS. For testing hypotheses 1a, 1b and 1c, hypotheses 2a, 2b, 2c, hypotheses 3a, 3b and 3c, and hypotheses 4a, 4b and 4c, PROCESS was used. Testing hypothesis 5a, 5b, and 5c was done by conducting a factorial analysis, ANOVA. Lastly, a Chi-squared test was conducted for testing the demographic characteristics age, gender and highest level of education.
18 Results
Randomization check
Randomization checks were conducted for the variables gender, age and education level, to check whether these variables were distributed evenly across the four conditions. Results indicated that experimental conditions did not differ in age, F (3, 158) = .62, p = .601, gender (χ² (3) = 1.38, p = .711), nor in level of education (χ² (12) = 9.64, p = .648). Thus, it can be stated that randomization across conditions was successful.
Manipulation check
To test whether the used stimuli in the experiments were perceived as intended, and for further analyses, the variable condition_culpability was created by merging the conditions on the basis of story perspective. Which means, the victim condition was formed by merging the conditions victim perspective – low distressing images (n = 38) and victim perspective – high distressing images (n = 33), and the culprit condition was formed by combining the condition culprit perspective – low distressing images (n = 45) and culprit perspective – high distressing images (n = 43). In addition, the variable condition_distressingimagery was created by merging the conditions on the degree of distressing imagery. The low distressing condition was formed by combining participants in the victim perspective – low distressing condition (n = 38) and culprit perspective – low distressing condition (n = 45) and the high distressing condition was formed by merging the victim perspective – high distressing condition (n = 33) and the culprit perspective – high distressing condition (n = 43).
Independent-samples t-tests showed a significant difference in the perceived
culpability of the main character, t(99.72) = -10.21, p < .001, 95% CI[-2.59, -1.75], and the perceived distress of the images between the conditions, t(142.02) = -4.73, p < .001, 95% CI[-1.23, -.51]. Participants in the culprit-perspective condition perceived the main character in the story as more culpable and responsible for the traffic accident than the participants in the victim-perspective condition. Moreover, the participants in the low-distressing condition reported to perceive the images as lower distressing than the participants in the
high-distressing condition (see Table 1). Therefore, it can be concluded that the manipulations have succeeded.
19 Table 1
Descriptive statistics for perceived culpability and perceived distress per condition
Perceived culpability main character
Perceived distress
M SD n M SD n
Story perspective Victim 2.23 1.63 71 . . .
Culprit 4.40 .84 88 . . .
Distressing imagery Low . . . 2.38 1.00 83
High . . . 3.25 1.28 76
Control variables
Several statistical analyses were conducted to test whether the control variables (age, gender and level of education) significantly correlated with or differed from the mediators and dependent variables. Pearson’s correlation analyses showed that age significantly correlated with emotion (r = .21, p = .008) and attitude toward message (r = .32, p < .001; see Appendix
3, Table 1). This indicated that as age rises, more negative emotions are evoked and positivity of the attitude toward the message increases. Thus, age was included as covariate in further analyses.
An independent samples t-tests was used to examine whether gender significantly differed from the mediators and dependent variables. Levene’s tests were found to be significant for gender on emotion (F(113.70) = 10.94, p = .001) and behavioral intention (F(133.25) = 4.19, p = .042). As can be seen in Table 2, gender significantly differed from transportation (t (157) = -2.84, p = .005, [-.51, -.09]) and behavioral intention (t (133.25) = 2.40, p = .015, 95% CI [.06, .60]). Gender was therefore included as covariate in further analyses.
20 Table 2
Descriptives and results for t-tests between gender on mediators and dependent variables
Males (n = 69) Females (n = 90) t(157) p 95% CI mean (SD) mean (SD) Mediators Identification 2.95 1.05 3.28 1.05 -1.93 .055 [-.66, .01] Transportation 3.15 .69 3.45 .63 -2.84** .005 [-.51, -.09] Emotion 3.36 .81 3.52 .55 -1.45 .149 [-.39, .06] Dependent variables
Attitude toward the message 3.66 .64 3.67 .64 -1.16 .872 [-.22, .18] Attitude toward the behavior 1.48 .72 1.40 .89 .65 .516 [-.17, .35] Behavioral intention 1.82 .91 1.49 .77 2.40* .015 [.06, .60] Note. N = 159. ** p < .01.
Wilk’s Λ showed that there was a significant difference between the levels of education on the mediators and dependent variables, Λ = .86, F(12, 300) = 1.92, p = .032. Univariate tests indicated that the level of education significantly differed from attitude towards the message (F(2, 156) = 7.09, p = .001, ηp² = .083; see Appendix 3, Table 2). Tukey post-hoc analysis showed that medium-educated participants had a significant higher positive attitude towards the message compared to the high educated participants, mean difference = .377, p = .001. Thus, education level was included as covariate in further analyses.
Hypotheses testing
To test hypotheses 1, 2, 3 and 4, Model 4 was used in the PROCESS macro (Hayes, 2013). The models with 5000 bootstrap samples had condition culpability as independent variable and, depending on the hypothesis, attitude toward message, attitude toward behavior or behavioral intention as dependent variable.
Hypothesis 1 predicted that participants who read a story from a victim’s perspective would have more positive (a) attitudes towards the message, b) attitude towards the provoked health behavior and c) behavioral intention than participants who read a story from the perspective of a culprit. Results indicated that there was no significant effect found of story perspective on the attitude towards the message, attitude toward the behavior and behavioral intention (see Table 3). Hypothesis 1a, 1b and 1c were therefore rejected.
21
b = -.45, p = .007 b = .16, p = .009
Direct effect, b = .10, p = .273 Indirect effect, b = -.07, 95% CI[-.17, -.01]
b = -.45, p = .007 b = .01, p = .894
Direct effect, b = .01, p = .969 Indirect effect, b = -.01, 95% CI[-.07, .08] Table 3
Overview results of story perspective on dependent variables
b SE p 95% CI
Attitude toward the message .10 .09 .273 [-.08, .28]
Attitude toward the behavior .01 .14 .969 [-.27, .28]
Behavioral intention -.04 .14 .753 [-.32, .23]
Note. N = 159. Attitude toward the message controlled for age and education level, behavioral intention was controlled for gender. 95% CI is confidence interval.
For testing hypotheses 2a, 2b and 2c, results of the mediation analyses indicated a significant indirect effect of story perspective on people’s attitude toward the message through
identification with the main character (see Figure 2). The results imply that a story presented in the culprit perspective had more negative attitude toward the message than when presented in the victim perspective, through the mediation effect of identification. Hypothesis 2a was therefore supported.
Figure 2. Significant and non-significant results of mediation analysis on H2a. Note. Attitude toward message was controlled for age and education level.
With respect to H2b and H2c, the model revealed that story perspective did have a negative significant effect on identification (b = -.45; b = -.45). However, no significant indirect effect was found between story perspective and attitude toward behavior, nor story perspective and behavioral intention (see Figure 3 and Figure 4). Thus, hypotheses 2b and 2c were rejected.
Figure 3. Significant and non-significant results of mediation analysis on H2b. Story perspective: Victim vs. Culprit Attitude toward the message Identification Story perspective: Victim vs. Culprit Attitude toward the behavior Identification
22
b = -.45, p = .007 b = -.04, p = .627
Direct effect, b = -.04, p = .754 Indirect effect, b = .02, 95% CI[-.06, .11]
b = -.18, p = .09 b = .20, p = .033
Direct effect, b = .10, p = .273 Indirect effect, b = -.04, 95% CI[-.11, .01]
b = -.18, p = .091 b = -.07, p = .609
Direct effect, b = .01, p = .969 Indirect effect, b = .01, 95% CI[-.04, .10]
Figure 4. Significant and non-significant results of mediation analysis on H2c. Note. Behavioral intention was controlled for gender.
Hypotheses 3a, 3b and 3c proposed a mediation effect of transportation between story perspective and attitude toward message, attitude toward behavior and behavioral intention. Results showed that although transportation was a significant predictor of attitude toward message (b = .20), no mediaton effect was found on the relationship between story perspective and attitude toward message (see Figure 5). Therefore, hypothesis 3a was not supported.
Figure 5. Significant and non-significant results of mediation analysis on H3a.
Note. Attitude toward message was controlled for age and education level, transportation was controlled for gender.
With respect to H3b and 3b, the model showed similar results. There was no significant indirect effect found for the relationship between story perspective and attitude toward behavior, nor between story perspective and behavioral intention through
transportation (see Figure 6 and Figure 7). Hypothesis 3b and 3c were therefore rejected.
Figure 6. Significant and non-significant results of mediation analysis on H3b. Note. Transportation was controlled for gender.
Story perspective: Victim vs. Culprit Behavioral intention Identification Story perspective: Victim vs. Culprit Attitude toward the message Transportation Story perspective: Victim vs. Culprit Attitude toward the behavior Transportation
23
b = -.18, p = .091 b = .11, p = .445
Direct effect, b = -.04 , p = .754 Indirect effect, b = -.02, 95% CI[-.10, .03]
b = -.31, p = .003 b = -.09, p = .282
Direct effect, b = .10, p = .273 Indirect effect, b = .03, 95% CI[-.03, .09]
b = -.31, p = .003 b = -.01, p = .932
Direct effect, b = .01, p = .969 Indirect effect, b = -.00, 95% CI[-.07, .05]
Figure 7. Significant and non-significant results of mediation analysis on H3c. Note. Behavioral intention and transportation were controlled for gender.
For testing hypotheses 4, emotion was used as mediation variable. With respect to H4a, 4b and 4c, similar results indicated that story perspective was a significant predictor of emotion (b = -.31). Presenting a story in the culprit’s perspective had a negative effect on the evoked emotions of the participants, compared to a story from a victim’s perspective.
However, no significant mediation was found on the relationship between story perspective and attitude toward message, story perspective and attitude toward behavior, and story perspective and behavioral intention (see Figure 8, Figure 9 and Figure 10) Consequently, hypotheses 4a, 4b and 4c were not supported.
Figure 8. Significant and non-significant results of mediation analysis on H4a.
Note. Attitude toward the message was controlled for age and education level, emotion was controlled for age.
Figure 9. Significant and non-significant results of mediation analysis on H4b. Note. Emotion was controlled for age.
Story perspective: Victim vs. Culprit Behavioral intention Transportation Story perspective: Victim vs. Culprit Attitude toward the message Emotion Story perspective: Victim vs. Culprit Attitude toward the behavior Emotion
24
b = -.31, p = .003 b = -.05, p = .45
Direct effect, b = -.04 , p = .754 Indirect effect, b = .02, 95% CI[-.07, .09]
Figure 10. Significant and non-significant results of mediation analysis on H4c.
Note. Behavioral intention was controlled for gender and emotion was controlled for age.
H5 predicted that the degree of distressing imagery would moderate the effect of story perspective. More precisely stated, H5 predicted that high (versus low) distressing imagery would lead to (a) more negative attitude towards the message, b) more positive attitude towards the desired behavior and c) higher behavioral intention towards decreasing or stopping mobile smartphone usage in traffic. To test hypothesis H5, a two-factorial
multivariate analysis of covariance (MANCOVA) was used, where age, gender and education level were included as covariates.
No correlations between the dependent variables were found, indicating that
multicollinearity was not of concern. Box’s M test indicated that homogeneity of variance-covariance was violated (p = .046). However, Levene’s tests showed that equal variances were assumed for story perspective on attitude toward message (F(3, 155) = 1.17, p = .324), attitude toward behavior (F(3, 155) = 1.80, p = .149) and behavioral intention (F(3, 155) = 1.29, p = .279; for descriptives of moderation test see Appendix 3, Table 3). As can be seen in Table 4, no significant interaction effects were found between story perspective and attitude toward message, story perspective and attitude toward behavior, nor story perspective and behavioral intention, through the moderation of degree of distressing imagery. This indicated that a high degree of distressing imagery did not influence the effect of story perspective on the dependent variables of the current study. Therefore, H5a, 5b and 5c were rejected.
Story perspective: Victim vs. Culprit
Behavioral intention Emotion
25 Table 4
Results of two-factor ANOVAs of story perspective x distressing imagery on attitude toward message, attitude toward behavior and behavioral intention
Outcome SS df MS F p η²
Attitude toward the message
Story perspective .01 1 .01 .03 .869 .000
Distressing imagery .07 1 .07 .18 .669 .001
Story perspective * distressing imagery
.12 1 .12 .322 .571 .002
Error 54.40 152 .36
Total
Attitude toward the behavior
2199.31 159
Story perspective .03 1 .03 .04 .833 .000
Distressing imagery .99 1 .99 1.46 .228 .010
Story perspective * distressing imagery .05 1 .05 .07 .789 .000 Error 103.24 152 .68 Total Behavioral intention 432.52 159 Story perspective .02 1 .02 .02 .881 .000 Distressing imagery 1.74 1 1.74 2.49 .117 .016
Story perspective * distressing imagery
.13 1 .13 .19 .668 .001
Error 106.19 152 .70
Total 537.08 159
Note. All results were measured on a significant level of p = .05. Attitude toward the message was controlled for age and education level, behavioral intention was controlled for gender.
26 Discussion
The aim of this study was to investigate the effect of a story presented from a victim’s compared to a culprit’s perspective on the attitude towards the message, attitude towards the provoked behavior and behavioral intention to diminish or stop mobile phone usage in traffic. In addition, the current study researched whether this effect was mediated by identification with the main character, transportation in the story and if the effect was moderated by the use of high compared to low distressing imagery.
Firstly, it was predicted that reading a story from the perspective of a victim,
compared to a culprit perspective, would lead to a more positive attitude towards the message, attitude towards the provoked health behavior and the behavioral intention. Results, however, showed that H1a, 1b and 1c were not supported and that presenting the story in different perspectives did not lead to a significant effect. This contradictive finding might be explained by E-ELM, which states that identification and transportation are the two main components of narrative engagement and persuasion (Slater & Rouner, 2002). It is therefore a possibility that the effect of story perspective did not occur, as identification and/or transportation are
mediators and thus conditional in the relationship between story perspective and the dependent variables. Hence, the mediating effects of identification and transportation were assumed and tested for Hypotheses 2 and 3.
In extension of Hypothesis 1, H2a, 2b and 2c predicted that the effect of story perspective was mediated by identification with the main character. Results indicated that there was a significant effect of story perspective on the attitude towards the message: presenting a narrative in a culprit perspective led to a more negative attitude toward the message compared to presenting a story from a victim’s perspective, but only when
identification took place with the main character. Hypotheses 2a was therefore supported and can be confirmed. However, although previous findings and theories have stated that
identification leads to a change in attitudes and behavioral intention, H2b and H2c were found not to be significant. The effect of story perspective did not apply to the attitude towards the behavior and behavioral intention to stop or diminish the usage of mobile phone usage.
According to Bandura’s social cognitive theory (2004), identification with the
character takes place when there is similarity between the reader and the story’s character. By presenting the narrative’s character with the same demographic and personal characteristics as the audience, identification and consequently behavioral change could be achieved. The used narratives in this study, however, were not tailored to the demographic and personal
27 the main character were not communicated in the story. For future research it is therefore recommended to tailor the demographical and personal characteristics to participants.
Beside identification, previous studies have stressed transportation to be an important component of narratives as well. Therefore, it was predicted that transportation into the story would lead to more positive attitudes toward the message (H3a), the behavior (H3b) and a higher behavioral intention (H3c) to diminish mobile phone use behind the stirring wheel. Despite the theories and prior findings, however, Hypotheses 3a, 3b and 3c were not supported. Although results did indicate that transportation was a significant predictor of attitude toward the message, no significant mediation effect of transportation was found on the effect of story perspective. This insignificant effect might be due the length of the short story. According to Green et al. (2004) transportation can be stimulated and increased by the degree of details in the text. As the narrative only contained about 550 words, the presented events and feelings experienced by the main character in the story might not be expressed and transferred effectively. To stimulate transportation in future research, the length and
immersion of the story could be extended by providing a more detailed narrative.
Another mediator that was included in this study, and has often been investigated in prior studies, is emotion. Results showed that story perspective did have an effect on emotion, and indicated that a victim perspective would lead to more evoked emotion. However, there were no mediation effects found for story perspective on attitude toward the message, attitude toward the behavior and behavioral intention through emotion. The results of this study did not support findings of prior researchers and H3a, H3b and H3c must therefore be rejected. Although Murphy et al. (2011) indicated emotion to be an independent mediator for behavioral change, Green and Brock (2002) suggest there is a relationship between
transportation and emotion. It is believed that an increased emotional response could lead to transportation, but more transportation into a story could also lead to more emotional
response. As the results of this study have not found transportation be a significant mediator, this might also be an explanation why there an effect of emotion did not take place. Future research might therefore examine the relationship of these two constructs and explorer the effect story perspective on attitudinal and behavioral changes, through the mediation of emotion and transportation.
Lastly, this study predicted that the degree of distressing imagery (low versus high) would moderate the effect of story perspective on the attitude towards the message (H5a), attitude towards the behavior (H5b), and the behavioral intention (H5c). Results, however, found that the degree distressing imagery was not a significant moderator of story perspective
28 and H5a, 5b and 5c must therefore be rejected. According to Witte and Allen (2000) stronger fear appeals produce stronger outcomes than weaker fear appeals. However, although the manipulation check has found that participants perceived the images in the high distress condition as more distressing than the images in the low distress condition, the degree of perceived distress could differ from one another. The contradictive results of this study can also be appointed to the fact that the used narrative in the experiment, did not contain a strong efficacy message to give the participants a strong belief they can prevent or stop the threat. Although at the end of the narrative, the main character did give his opinion on how people should not use their mobile phones while driving, this can not be considered as a clear and strong efficacy message. While the effect of story perspective was found through
identification, future studies on the combination of story perspective and distressing images are therefore recommended to include a strong efficacy message.
Limitations
This study contains a number of limitations. The first mentioned limitation concerns the nature of this study, which effects the generalizability of the results. As this research was based on a convenience sampling, the participants were acquired by their convenient
availability and the proximity to the researcher and therefore mostly consists of high educated respondents.
As the experiment was conducted online, another limitation is the fact that the researcher was not able to control the environment of the respondents in which they participated in the experiment. Moreover, the participants were able to stop the online experiment and continue on another time or day and this could have a negative effect on the outcomes or even cancel out the effect the intended effects of the stimuli. To evoke the
desired effects of narratives, such as emotions or persuasion, it is of importance that the reader is fully transported into the story and identifies himself with the character (Green & Brock, 2000; Slater & Rouner, 2002). However, when the reader drops out of the story and continues reading later in time, the process of narrative persuasion might be interrupted and might not be effective anymore. The measured effects on the dependent variables could therefore also be unpure. This also applies when there is a time gap between reading the narrative and answering the questionnaire. Hence future research might combine laboratory and field to exclude these unwanted effects.
29 Conclusion
Reported results of this study demonstrate how the perspective of a story, mediated by identification with the main character, can influence the attitude toward the message. This study provides additional insights into existing knowledge about narrative persuasion, as desired and undesired study outcomes contribute to what is known about the effects of narrative persuasion.
As Hoeken and Sinkeldam (2014) pleaded, it is of importance to fully understand how narratives can evoke persuasive effects in order to understand when and why stories can lead to positive effects. By manipulating and investigating the underlying mechanisms of
narratives in future research, such as story perspectives, we are able to better understand with which story perspectives or character characteristics we are more inclined to identify with. These insights can be used and examined further in future research, to develop new or optimize existing persuasion strategies that could be implemented into promising health interventions.
30 References
Ajzen, I. (1991). The theory of planned behavior. Organizational behavior and human decision processes, 50(2), 179-211.
Appel, M., & Richter, T. (2007). Persuasive effects of fictional narratives increase over time. Media Psychology, 10(1), 113-134.
Appiah, O. (2001). Ethnic identification on adolescents' evaluations of advertisements. Journal of Advertising Research, 41(5), 7-22.
Austin, E. W., Roberts, D. F., & Nass, C. I. (1990). Influences of family communication on children's television-interpretation processes. Communication research, 17(4), 545-564.
Bandura, A. (2004). Health promotion by social cognitive means. Health education & behavior, 31(2), 143-164.
Baron, R., Logan, H., Lilly, J., Inman, M., & Brennan, M. (1994). Negative emotion and message processing. Journal of Experimental Social Psychology, 30, 181-181. Bhaduri, G., Ha-Brookshire, J. E., & Leshner, G. (2017). Too Good To Be True? Effect of
Consumers’ Brand Schemas on Apparel Brands’ Fair Labor Marketing Messages. Clothing and Textiles Research Journal, 35(3), 187-203.
Brown, S. L., & West, C. (2015). Sequencing the threat and recommendation components of persuasive messages differentially improves the effectiveness of high‐and low‐ distressing imagery in an anti‐alcohol message in students. British journal of health psychology, 20(2), 324-340.
Busselle, R., & Bilandzic, H. (2009). Measuring narrative engagement. Media Psychology, 12(4), 321-347.
Cohen, J. (2001). Defining identification: A theoretical look at the identification of audiences with media characters. Mass communication & society, 4(3), 245-264.
De Graaf, A. (2014). The Effectiveness of Adaptation of the Protagonist in Narrative Impact: Similarity Influences Health Beliefs Through Self‐Referencing. Human
31 De Graaf, A., Hoeken, H., Sanders, J., & Beentjes, H. (2009). The role of dimensions of
narrative engagement in narrative persuasion. Communications—The European Journal of Communication Research, 34, 385-405.
De Graaf, A., Hoeken, H., Sanders, J., & Beentjes, J. W. (2012). Identification as a mechanism of narrative persuasion. Communication Research, 39(6), 802-823. De Graaf, A., Sanders, J. M., & Hoeken, H. (2016). Characteristics of narrative interventions
and health effects: A review of the content, form, and context of narratives in health-related narrative persuasion research.
Green, M. C., & Brock, T. C. (2000). The role of transportation in the persuasiveness of public narratives. Journal of personality and social psychology, 79(5), 701.
Green, M. C., Brock, T. C., & Kaufman, G. F. (2004). Understanding media enjoyment: The role of transportation into narrative worlds. Communication Theory, 14(4), 311-327. Hoeken, H., Kolthoff, M., & Sanders, J. (2016). Story perspective and character similarity as
drivers of identification and narrative persuasion. Human Communication Research, 42(2), 292-311.
Hoeken, H., & Sinkeldam, J. (2013). De rol van emoties in narratieve overtuiging. Tijdschrift voor Taalbeheersing, 35(3), 226-236.
Hoeken, H., & Sinkeldam, J. (2014). The role of identification and perception of just outcome in evoking emotions in narrative persuasion. Journal of Communication, 64(5), 935-955.
Hoyer, W. D., Macinnis, D. J., & Pieters, R. (2012). Consumer Behavior, South-Western, Cengage Learning. International edition.
Kok, G. J. (2016). Social psychology applied: Politics, theories, and the future. Valedictorian lecture, Maastricht: Maastricht University. See
http://interventionmapping.com/wpcontent/uploads/2016/04/Kok-2016-valedictory-lecture.pdf
Lench, H. C., Flores, S. A., & Bench, S. W. (2011). Discrete emotions predict changes in cognition, judgment, experience, behavior, and physiology: a meta-analysis of experimental emotion elicitations. Psychological bulletin, 137(5), 834.
32 Mitchell, A. A. & Olson J. C. (1981). Are product attribute beliefs the only mediator of
advertising effects on brand attitude? Journal of Marketing Research, 18, 318-332. Morgan, S. E., Movius, L., & Cody, M. J. (2009). The power of narratives: The effect of
entertainment television organ donation storylines on the attitudes, knowledge, and behaviors of donors and nondonors. Journal of Communication, 59(1), 135-151. Moyer‐Gusé, E. (2008). Toward a theory of entertainment persuasion: Explaining the
persuasive effects of entertainment‐education messages. Communication Theory, 18(3), 407-425.
Moyer-Gusé, E., Chung, A. H., & Jain, P. (2011). Identification with characters and
discussion of taboo topics after exposure to an entertainment narrative about sexual health. Journal of Communication, 61(3), 387-406.
Moyer‐Gusé, E., & Nabi, R. L. (2010). Explaining the effects of narrative in an entertainment television program: Overcoming resistance to persuasion. Human Communication Research, 36(1), 26-52.
Murphy, S. T., Frank, L. B., Chatterjee, J. S., & Baezconde‐Garbanati, L. (2013). Narrative versus nonnarrative: The role of identification, transportation, and emotion in reducing health disparities. Journal of Communication, 63(1), 116-137. Murphy, S. T., Frank, L. B., Moran, M. B., & Patnoe‐Woodley, P. (2011). Involved,
transported, or emotional? Exploring the determinants of change in knowledge, attitudes, and behavior in entertainment‐education. Journal of
Communication, 61(3), 407-431.
Murray, S. (1999). Saving our so-called lives: Girl fandom, adolescent subjectivity, and My So-Called Life. In M. Kinder (Ed.), Kids’ media culture (pp. 221–235). Durham, NC: Duke University Press.
Nabi, R. L. (1999). A cognitive‐functional model for the effects of discrete negative emotions on information processing, attitude change, and recall. Communication theory, 9(3), 292-320.
Nabi, R., & Green, M. (2014). The Role of a Narrative's Emotional Flow in Promoting Persuasive Outcomes. Media Psychology, 18(2), 1-26.
