Identification as a Mechanism of Narrative Persuasion

Abstract

To provide a causal test of identification as a mechanism of narrative persuasion, this study uses the perspective from which a story is told to manipulate identification experimentally and test effects on attitudes. In experiment 1, 120 participants read a story that was told either from the perspective of one character or another character, with both characters having opposing goals. Results showed that perspective influenced identification and story consistency of attitudes. Moreover, identification with one of the characters mediated the effect of perspective on attitudes. In experiment 2, 200 participants read a different story that was told from one of two perspectives, with both characters having opposing opinions. Results showed that identification with both characters mediated the effect of perspective on attitudes. The results of these experiments indicate that identification can be a mechanism of narrative persuasion.

Keywords

narrative persuasion identification narrative engagement transportation perspective

Identification as a Mechanism of Narrative Persuasion

It is by now well-established that narratives can have effects on readers’ real-world beliefs and attitudes (e.g., Appel & Richter, 2007; Diekman, McDonald, & Gardner, 2000; Strange & Leung, 1999). This phenomenon has been termed “narrative persuasion” and has attracted research interest from various disciplines such as health communication (Green, 2006), entertainment-education (Morgan, Movius, & Cody, 2009; Moyer-Gusé & Nabi, 2010), and cultivation research (Busselle & Bilandzic, 2008). However, the exact mechanisms through which narratives exert persuasive influence are not clearly established. Generally, there is a consensus that the extent to which a reader is engrossed in a story or transported into a story world plays a role in generating narrative effects (see, for example, Green & Brock, 2000; Moyer-Gusé, 2008). Some research has shown that this narrative experience can indeed influence persuasive outcomes (Green, 2004; Green & Brock, 2000). However, recent research has shown that the experience of a narrative is multidimensional. (Busselle & Bilandzic, 2009). Some dimensions may be more important for persuasive effects to occur than others. Busselle and Bilandzic (2009), for instance, showed that their dimension of emotional engagement was correlated more strongly to story-consistent attitudes than the dimension of attentional focus. Establishing which dimensions of the narrative experience actually lead to persuasive effects will further expand our understanding of the mechanisms of narrative persuasion (see Busselle & Bilandzic, 2009; Green, 2006, p. 165).

In this article, we focus on the dimension of identification with story characters. Several scholars put forward the idea that identification is a mechanism through which narratives exert persuasive influence (e.g., Green, 2006; Slater & Rouner, 2002). Also, some associations have been found between identification and story-consistent attitudes (De Graaf, Hoeken, Sanders, & Beentjes, 2009; Iguarta, 2010). However, a causal test of the effect of identification on attitudes has not yet been provided. The correlational results found so far may also indicate that readers who already share attitudes with a character identify more with this character. To rule out this explanation, this article reports on two studies that manipulate identification experimentally and examine the mediating role of identification in narrative persuasion.

Narrative Engagement

Scholars have given different names to the experience that narratives typically generate. Green and Brock (2000) adopt the term transportation from Gerrig (1993), which refers to the experience of going into a narrative world. Green and Brock (2000, p. 701) define transportation as the focus of attention, emotion, and imagery on a story. Slater and Rouner (2002, p. 178) employ the label absorption which they define as “vicariously experiencing the characters’ emotions and personality.” Although there is some overlap between these definitions, they differ in important respects. On the one hand, Slater and Rouner’s (2002) definition focuses specifically on characters and shows correspondence to identification, which has been defined as adopting the perspective of a character and consequently experiencing empathic emotions (Cohen, 2001, p. 251). On the other hand, both Gerrig’s (1993) and Green and Brock’s (2000) definition of transportation does not refer to characters. Green and Brock (2000) do include emotion in transportation, but it is not specified whether this needs to be related to characters. Moyer-Gusé (2008) acknowledges this point when she notes that identification with characters goes beyond involvement with the narrative itself, or transportation. However, she also notes the overlap between identification and transportation (Moyer-Gusé, 2008, p. 410). In sum, it is hard to draw a clear line between transportation and identification.

Empirical research by Busselle and Bilandzic (2009) provides insight into the relations between these concepts. They use the term narrative engagement to refer to the experience of a narrative. Their studies show that distinct dimensions of this experience can be discerned. For instance, their factor analyses yielded a dimension of emotional engagement, which consisted of both items about emotion evoked by the narrative and items about empathic emotions for the characters. This dimension thus includes both the emotion aspect of transportation and the empathy aspect of identification, which shows that these two concepts overlap. Other aspects of transportation, such as attentional focus on the story, loaded on a different dimension, revealing that the attention aspect can be distinguished from identification. This study adopts the approach of Busselle and Bilandzic (2009) to see narrative engagement as a complex concept in which multiple dimensions can be discerned.

Identification

As mentioned above, identification is an experience in which readers adopt the perspective of a character and see the narrative events through the character’s eyes (Busselle & Bilandzic, 2008; Cohen, 2001). Oatley (1994, 1999) conceptualizes this experience as a reader adopting the character’s goals and plans. The reader then simulates or imagines the events that happen to the character and experiences empathy or emotions which are in consonance with the success or failure of these plans (Oatley, 1994, p. 69; Zillmann, 2006). In other words, readers imagine what it is like to be a character and can have the illusion of being a character (Cohen, 2001; Tan, 1994, p. 25).

Identification is proposed to be one of the mechanisms through which narratives can change attitudes (Green, 2006; Slater & Rouner, 2002). Through the connections forged by identification, implications of experiences and assertions of the character may shift a reader’s beliefs (Green, 2006). When readers simulate or imagine the events that happen to a character in their imagination, they may come to understand what it is like to experience the described events and thus their attitudes may become more consistent with this vicarious experience (Mar & Oatley, 2008, p. 182). Similarly, Bandura’s social cognitive theory (1986, 2002) posits that identification with role models facilitates observational learning, such that individuals acquire new ways of thinking and behaving by observing a model. Although this theory is not limited to narratives or the media, characters in the media can also function as role models for observational learning (Bandura, 2002).

There is some empirical evidence for the role of identification in narrative persuasion. For instance, Busselle and Bilandzic (2009) showed that their scale of emotional engagement which consisted for a large part of items about empathic emotions was correlated to story-consistent attitudes. Similarly, De Graaf et al. (2009) showed that a scale that included items about taking the perspective of a character as well as items about empathizing with the character was related significantly to story-consistent attitudes. However, in both these studies, it is unclear whether identification was the cause of the effects on the attitude. The opposite relation may hold as well; if participants hold similar attitudes as the character, this similarity in attitudes may cause them to identify with this character rather than the other way around. Several studies have shown that perceived similarity of a reader to a character can increase identification with the character (e.g., Andsager, Bemker, Choi, & Torwel, 2006; Eyal & Rubin, 2003). Therefore, correlations between identification and attitudes are not enough to conclude that identification leads to attitude change. Instead, a causal test of the effect of identification on persuasive outcomes is needed to establish whether identification can indeed function as a mechanism of narrative persuasion.

The Manipulation of Identification: The Role of Perspective

To establish a causal effect of identification on attitudes, identification needs to be manipulated. The definition of identification as adopting the perspective of a character (Cohen, 2001) suggests that the perspective from which a narrative is told may be a suitable means to manipulate identification. Adopting the perspective of a character on the narrative events in one’s imagination is a reader experience, whereas the perspective from which a narrative is told is a text characteristic. Perspective as a text characteristic refers to the physical and psychological point of perception which is presented in a story (Bal, 1997, p. 143). So, when the perspective of a story is positioned with a particular character—who will be referred to as the “perspectivizing character” in this article—this character’s vision on the events and his or her reactions, such as thoughts, are presented. The most clear-cut example of a perspectivizing character is the first-person narrator. In first-person narratives, readers have access to the perception and thoughts of a particular character, referred to with “I,” which likely facilitates readers to take the perspective of this character in their imagination (Oatley, 1999; Sanders & Redeker, 1996).

Although several authors posit that story perspective can influence identification (Cohen, 2001; Oatley, 1999), this has not yet been empirically established. However, there is some empirical evidence for the effect of perspective on related concepts. Andringa (1986) reported that participants who read a first-person narrative about a court session from the judge’s perspective reported more understanding for the judge than participants who read a first-person narrative from the defendant’s perspective. Likewise, Van Peer and Pander Maat (1996) found that participants reading a story about a marital disagreement, in which the perceptions and thoughts of the wife were presented, sympathized more with her than participants reading the same story told from the perspective of the husband. As sympathy and understanding for a character are closely related to identification (Cohen, 2001), these studies suggest that the perspective from which a story is told is a suitable means to manipulate identification. Therefore, the first hypothesis of this study is:

Hypothesis 1: Readers of a story told from the perspective of a given character will identify more with that character than readers of the same story told from the perspective of another character.

The next step is to look at persuasive outcomes. Because identification is specifically related to characters, attitudes will be assessed that are consistent with individual characters. This means that the attitude is related to a character in the story, for instance because a character states his or her opinion on an issue, or because the attitude is implied by the experience of the character (e.g., when a character pursues a goal). Following from the idea that the perspective of a character leads to identification with that character, it is also expected that perspective influences character-related attitudes. Thus, the second hypothesis of this study is:

Hypothesis 2: Readers of a story told from the perspective of a given character will have posttest attitudes more consistent with the attitudes related to that character than will readers of the same story told from the perspective of another character.

Finally, to investigate whether identification with a character functions as a mechanism of narrative persuasion, the third hypothesis of this study is:

Hypothesis 3: Identification with characters will mediate the effect of the perspective manipulation on posttest attitudes.

Experiment 1

Method

Materials

Two versions of a story were written about a job interview for a position of web designer. One version of the story is told from the perspective of the applicant, a young man in a wheelchair, the other from the perspective of a member of the selection committee, a young programmer. This topic was chosen because the characters imply different sides on issues related to disabled persons on the job market, such as positive discrimination. The story ends when the applicant leaves the interview. It remains unclear whether he gets the job or not. The manipulation of perspective is intended to increase identification with the character from whose perspective the story is told. Participants who read the story from the perspective of the applicant should identify more with the applicant than participants who read the story from the perspective of the programmer and vice versa.

Perspective was established by several means (see Table 1 for examples of the manipulation). First, the perspectivizing character is referred to with I, whereas other characters are referred to with the third-person pronoun (he) or a full noun phrase (e.g., the man). Second, perceptions and thoughts of the perspectivizing character are presented. Thoughts could not be inserted in exactly the same place of the story for both characters, because this would have resulted in unnatural stories, but there are an equal number of thoughts in both versions. Finally, a short sequence of events experienced by the perspectivizing character before the actual job interview are included in both versions to enable readers to identify with this character. The applicant is described as talking to his physiotherapist about his preparation for the interview, whereas the programmer is described as talking to the manager about which candidates to invite for an interview.¹

Table 1.

Example of Perspective Manipulation: A Part of the Job Interview in Both Versions

Perspectivizing character = applicant	Perspectivizing character = programmer
“Let’s move on to talk about the position. Why did you apply for this job?” the older man asks. [Thought 1] Again a question that is listed on every website about job application. “I applied because the job description fits well to my education and work experience. (. . .) That is why the job seems very suitable for me.” “Can you say what you like so much about website design?’ the younger man asks. [Thought 2] That is a question I did not prepare for. After thinking about it shortly, I say: “The creative part. I like making something completely new.”	Theo [the manager of the company] continues: “Let’s move on to talk about the position. Why did you apply for this job?” “I applied because the job description fits well to my education and work experience. (. . .) That is why the job seems very suitable for me.” [Thought 1] Again such a rehearsed answer. I try to ask more to get a more original answer. “Can you say what you like so much about website design?” The man looks at the wall shortly. “The creative part. I like making something completely new.” [Thought 2] Ok, so he can also give an answer that he did not prepare beforehand.

Perspectivizing character = applicant

Perspectivizing character = programmer

“Let’s move on to talk about the position. Why did you apply for this job?” the older man asks.

[Thought 1] Again a question that is listed on every website about job application.

“I applied because the job description fits well to my education and work experience.

(. . .) That is why the job seems very suitable for me.”

“Can you say what you like so much about website design?’ the younger man asks.

[Thought 2] That is a question I did not prepare for. After thinking about it shortly, I say:

“The creative part. I like making something completely new.”

Theo [the manager of the company] continues: “Let’s move on to talk about the position. Why did you apply for this job?”

“I applied because the job description fits well to my education and work experience. (. . .) That is why the job seems very suitable for me.”

[Thought 1] Again such a rehearsed answer. I try to ask more to get a more original answer.

“Can you say what you like so much about website design?”

The man looks at the wall shortly.

“The creative part. I like making something completely new.”

[Thought 2] Ok, so he can also give an answer that he did not prepare beforehand.

Note: Insertions (e.g., [Thought 1]) and blank lines were not present in the original versions.

The two versions of the story were pretested to ensure that they were considered natural and suitable by the target participants. Five students were instructed to read both versions of the story critically and mark all elements in the story that they thought were unnatural or wrong, either stylistically or in content. With their suggestions, the two versions of the story were improved and adapted to the target group. The final versions of the story both consist of approximately 3,700 words. Readability was kept the same for both versions (Gunning Fog indexes: applicant: 8.55, programmer: 8.59).

Participants and procedure

The sample consisted of 120 students (79.8% female, 20.2% male) from a Dutch university who were recruited in several first-year classes. The majority studied at the Faculty of Arts (55.8%), whereas the rest were either law students (22.5%) or social sciences students (21.6 %). Age varied between 18 and 28, with an average of 19. The experiment was administered to groups with a maximum of 10 participants in a classroom setting. In every group, the conditions were randomly assigned. The experiment took 20 to 25 minutes to complete. After everyone in a group was finished, participants were debriefed and paid €5 for participation. One participant in the control group was excluded from the analysis because he missed an entire page of items. Other participants did not miss any items.

Design

The study used a between-subjects design with three conditions. One group (n = 40) read the version of the story told from the applicant’s perspective and a second group (n = 40) read the version of the story told from the programmer’s perspective. Both groups answered all questions after reading the story. A third group (n = 40) acted as a control group and answered the questions about attitudes before reading the story in order to establish a baseline of attitudes.

Questionnaire

The questionnaire consisted of questions about personal characteristics (gender, age), attitudes toward issues related to disabled persons on the job market, identification, additional aspects of narrative engagement, and perceived realism.

Attitudes were measured with eight 7-point rating scale items, ranging from completely disagree to completely agree. Five items were about affirmative action for disabled applicants (Cronbach’s α = .82), such as: “When candidates are equally suitable, disabled applicants should get preference over others.” Three items were about the autonomy of employers to reject disabled applicants, such as: “I think an employer should be able to reject an applicant because he or she is disabled” (Cronbach’s α = .73). These attitudes represent positions that are more in line with either the interest of the disabled applicant or that of the programmer.

Identification was measured with eight 7-point rating scale items for each character. Three items were about imagining events from the position of each character, for example, “During reading, I imagined what it would be like to be in the position of the applicant/programmer.” Two items were about experiencing empathy with each character, for example, “I empathized with the applicant/programmer.” And three items were about the illusion of being each character, for example, “In my imagination it was as if I was the applicant/programmer.” These items were formulated based on the conceptualization of identification given in the introduction; because existing scales do not capture all aspects (most notably “illusion of being”) and confound other concepts (see Busselle & Bilandzic, 2009, p. 342).

To be able to check whether other aspects of narrative engagement were (unintentionally) influenced by perspective, these aspects were measured as well. Eight 7-point Likert-type scales were about attention for the story, for example, “During reading I was fully concentrated on the story,” there were four items about imagery of the story, for example, “I had a vivid image of the events in the story”; four items were about emotion evoked by the story, for example, “The story affected me”; and seven items were about the feeling of going into the narrative world, for example, “When I was reading the story, it seemed as if I was there in my thoughts.” Some items were selected from scales to measure transportation (Green & Brock, 2000) or telepresence (Kim & Biocca, 1997).

An exploratory principal components factor analysis with oblimin rotation was run on all items (see appendix). Data of the control group were not included in this analysis because the attitude questions that were answered before reading the story may have made participants suspicious of the story’s persuasive intent. The eigenvalue rule indicated seven factors, but the scree plot showed that a five-factor solution was more appropriate.

Factor 1, “Being in Narrative World,” contained all items about imagery of the story and four items about the sensation of going into a narrative world (eight items, Cronbach’s α = .92). Factor 2, “Identification with Programmer,” contained all items about imagining events from the position of the programmer as well as empathy with the programmer and the illusion of being the programmer (eight items, Cronbach’s α = .92). Factor 3 contained all items about “Emotion” evoked by the story (four items, Cronbach’s α = .91). Factor 4 contained six items about “Attentional Focus” (six items, Cronbach’s α = .88). Factor 5 contained all items about “Identification with Applicant” (eight items, Cronbach’s α = .91). The remaining items were not included in any factor because they loaded too low on their primary factor or cross-loaded on a secondary factor.

Finally, perceived realism was measured, because several researchers assign an important role to perceived realism in narrative persuasion (e.g., Busselle, Ryabovolova & Wilson, 2004; Green, 2004). To assess whether the two story versions were perceived equally realistic, three 7-point Likert-type scales were used (Cronbach’s α = .79), such as “It is possible the events in the story really happened.”

Results

Effects of manipulation

With regard to Hypothesis 1 and Hypothesis 2, it was tested whether perspective had the intended effect of increasing identification with the perspectivizing character and whether attitudes were more consistent with the perspective of the story version. Additionally, it was checked whether the manipulation had unintentionally influenced perceived realism, or the engagement dimensions Attentional Focus, Being in Narrative World, and Emotion. Table 2 contains the mean scores and standard deviations for identification with both characters and attitudes that were implied by the story.

Table 2.

Means and Standard Deviations (Between Brackets) for Identification^a and Attitudes^b by Condition

	Perspectivizing character = applicant	Perspectivizing character = programmer
Identification with applicant	5.06 (1.10)^c	4.17 (1.09)^c
Identification with programmer	3.44 (1.23)^c	4.15 (1.20)^c
Attitude toward affirmative action for disabled people	3.54 (1.12)	3.33 (.91)
Attitude toward employers’ autonomy to reject disabled people	2.51 (1.02)^c	2.96 (.83)^c

1 = very low identification, 7 = very high identification.

1 = very negative attitude, 7 = very positive attitude.

Indicates significant differences.

t tests revealed a significant effect of perspective on both Identification with Applicant (t(78) = 3.65, p < .001, η² = .15) and Identification with Programmer (t(78) = 2.61, p < .05, η² = .08). As expected, participants who had read the story from the applicant’s perspective identified more with the applicant than participants who had read the version from the programmer’s perspective, and vice versa for identification with the programmer. However, inspection of the means suggested an inequality between the characters. Therefore, we did a post-hoc paired-samples t test which showed that overall, Identification with Applicant (M = 4.61, SD = 1.18) was higher than Identification with Programmer (M = 3.79, SD = 1.26); (t(79) = 4.17, p < .001, η² = .18). In a 2 × 2 mixed ANOVA with perspective as between-subjects factor and character as within-subjects factor, the interaction was also significant (F (1,78) = 20.72, p < .001, η² = .21). Subsequent simple effects analysis showed that participants who had read the perspective of the applicant identified more with the applicant than with the programmer (p < .001), whereas participants who had read the perspective of the programmer identified equally with the applicant and the programmer (p = .94).

With regard to the additional engagement dimensions and perceived realism, t tests showed that, as intended, there were no effects of perspective on Attentional Focus (M_appl = 4.95, SD = 1.01; M_progr = 4.83, SD = 1.18; t(78) = 0.49. p = .62), on Being in Narrative World (M_appl = 5.58, SD = .62; M_progr = 5.36, SD = 1.02; t(78) = 1.21, p = .23), or on Perceived realism (M_appl = 6.25, SD = 0.63; M_progr = 6.07, SD = 0.72; t(78) = 1.21. p = .23). However, there was an unexpected effect of perspective on Emotion (t(78) = 2.38, p < .05, η² = .068). Readers of the applicant perspective (M = 4.15, SD = 1.26) reported more emotion than readers of the programmer perspective (M = 3.47, SD = 1.30).

No difference was obtained between perspectives for the attitude toward affirmative action for disabled people in job applications, t(78) = 0.89, p = .38. However, there was a difference for the attitude toward employers’ autonomy to reject disabled people in job applications, t(78) = 4.67, p < .05, η² = .057. Participants who had read the applicant perspective reported more negative attitudes than readers of the programmer perspective.

Mediation analysis

With regard to Hypothesis 3, mediation analyses using a bootstrapping procedure (Preacher & Hayes, 2008) were carried out to establish whether Identification with Applicant and Identification with Programmer were responsible for the effect of perspective on the attitude toward employers’ autonomy. Also, it was checked whether this effect should be ascribed to the unintentionally affected engagement dimension of Emotion. All mediation analyses used 5,000 bootstrap samples, as recommended by Hayes (2009).

Preacher and Hayes’ (2008) bootstrapping procedure estimates a 95% confidence interval for whether the indirect effect is significant. If this interval does not include 0, the indirect effect is significant. These analyses showed that the indirect effect of Identification with Applicant is significant (B = .19, SE = .10, 95% CI = [.0580, .4400]). Figure 1 provides descriptive information concerning the relations between perspective, Identification with Applicant, and the attitude toward employers’ autonomy. The indirect effects of Identification with Programmer (B = .03, SE = .08, 95% CI = [−.0928, .2027]) and Emotion (B = .03, SE = .07, 95% CI = [−.0966, .2140]) were not significant.

Figure 1.

Descriptive results of the relations between perspective, identification with applicant and the attitude toward employer’s autonomy to reject disabled applicants

Control group

Finally, the attitude scores of the control group were employed to assess the direction of the persuasive influence. Did the applicant version make attitudes more negative or did the programmer version make attitudes more positive compared to baseline attitudes? t tests with Bonferroni adjustments for multiple comparisons showed that the attitude toward affirmative action of the control group (M = 3.60, SD = 1.07) did not differ from both perspective versions (both p’s > .76). However, with respect to the attitude toward employer’s autonomy, participants in the control group held less negative attitudes (M = 3.23, SD = 1.20) compared to the participants in the applicant’s perspective (t(77) = 3.27, p < .01, η² = .097). The control group did not differ from the programmer perspective (p = .72)

Discussion

The results confirm Hypothesis 1 and Hypothesis 2. The character’s perspective from which the story was told led to more identification with this character and to a posttest attitude more consistent with the character. The applicant perspective made the attitude toward employers’ autonomy more negative, which is consistent with the applicant’s goal.

Hypothesis 3 can be confirmed for identification with one of the characters in the story. The mediation analysis showed that identification with the applicant mediated the effect of perspective on attitudes. However, identification with the programmer did not affect the attitude toward employers’ autonomy. This difference in impact may be explained by several differences between the characters, which are also reflected in the result that participants identified more with the applicant overall than with the programmer and that even participants who read the story from the programmer’s perspective identified equally with the applicant and the programmer. First, the position of the applicant is probably more familiar for the participants than the position of the programmer, because the participating students are more likely to have applied for a job in their lives, than to have been in a position of hiring someone. Also, being physically disabled, the applicant was in a position of greater dependence and vulnerability than the programmer: depending on the programmer’s power to decide about his application and vulnerable to potential prejudice about disabled applicants. To be able to strengthen conclusions about identification as a mechanism of narrative persuasion, a follow-up experiment was conducted that used a story in which the background and position of two opposing characters are more similar in position and mutual dependence.

Experiment 2

Method

Materials

A new story was written about two sisters whose mother has been in an irreversible coma for over a month. In the story, there is only one difference between the sisters, namely their opinion on what to do next. Both women want to do what is in their mother’s best interest, but they disagree about what this best interest is. One of the sisters thinks it may be best to consider euthanasia in this situation without any prospects, whereas the other sister does not want to consider euthanasia and thinks it is best to search for a good nursing home. They argue about what should be done, but the argument is not resolved in the story.

To manipulate identification, two versions of the first-person narrative were written.² In one version, the perspectivizing character is in favor of considering euthanasia for her permanently comatose mother, whereas the nonperspectivizing sister is against considering euthanasia. In the other version, these roles are reversed; the perspectivizing character is against considering euthanasia for her mother, whereas the other sister is in favor. This manipulation is intended to increase identification with the perspectivizing character, so that it can be tested whether identification leads to persuasion. Participants who read the story in which the perspectivizing character is in favor of considering euthanasia should identify more with this character than participants who read the story in which the perspectivizing character is against considering euthanasia and vice versa. Perspective was established as in Experiment 1 (see Table 3 for examples of the manipulations); that is, the perspectivizing character is referred to with I, so that the events are described from her point of view and her perceptions and thoughts are presented. Nevertheless, the same arguments for both positions were given in both versions. The rest of the two versions were also held constant; the perspectivizing character as well as the nonperspectivizing character performed the same actions irrespective of their opinion.¹

Table 3.

Example of Perspective Manipulation: Some Parts of the Story in Both Versions

Perspectivizing character = in favor of considering euthanasia	Perspectivizing character = against considering euthanasia
[Part 1] Miriam [= name of sister] immediately concluded we should look for a nursing home, but could we really do that to our mother? Shouldn’t we consider whether she would want that herself?	[Part 1] I thought it was clear we had to start looking for a nursing home, but incomprehensibly, Miriam did not agree. She said we should consider whether she would want that herself.
[Part 2] I started: “Every time I look at her, I feel so extremely sorry for her, that she just has to lie there. And if I think that this would go on for years and years . . . It just seems to me so terrible for her.”	[Part 2] Miriam started: “Every time I look at her, I feel so extremely sorry for her, that she just has to lie there. And if I think that this would go on for years and years . . . It just seems to me so terrible for her.”
“But she is still our mother. She has always taken care of us and now she needs our care. We should search for the best nursing home there is.”	“But she is still our mother. She has always taken care of us and now she needs our care. We should search for the best nursing home there is.”
[Part 3] That afternoon at work I could not concentrate. I had to verify data that had already been entered, but my thoughts strayed again and again.	[Part 3] That afternoon at work I could not concentrate. I had to verify data that had already been entered, but my thoughts strayed again and again.

Note: Insertions (e.g., [Part 1]) and blank lines were not present in the original versions.

As in Experiment 1, the two versions of the story were pretested with members of the target population. Each version was read by seven students. With their suggestions, the two versions of the story were improved and adapted to the target group. After these adaptations, both versions of the story consisted of approximately 1,800 words. Readability was the same for both versions (Gunning Fog index: 7.54).

Participants and procedure

The sample consisted of 200 students (81.7% female, 19.3% male) enrolled in a first-year communication class at a Dutch university, who participated for course credit. Their ages varied between 17 and 28, with an average age of 20. The experiment was administered to groups of 12-20 participants in a classroom setting. In every group, the conditions were randomly assigned. The experiment took 15 to 20 minutes to complete. After everyone in a group was finished, participants were debriefed and any remaining questions were answered. Three participants were excluded from the analyses because they missed eight items or more. Other sporadic items that were not filled in, were set to missing.

Design

The study used a between-subjects design with three conditions. One group (n = 80) read the version of the story in which the perspectivizing character was in favor of considering euthanasia. Another group read the version of the story in which the perspectivizing character was against considering euthanasia (n = 79). Participants in these groups answered all questions after reading the story. Participants in the control group (n = 41) answered the questions about attitudes before reading the story in order to establish a baseline of attitudes.

Questionnaire

The questionnaire was similar to the one used in the first experiment, including questions about personal characteristics, attitudes toward issues implicated in the story, dimensions of narrative engagement, and perceived realism.

Attitudes were measured with six 7-point Likert-type scales. Three items were about considering euthanasia for a loved one who is in a permanent coma (Cronbach’s α = .72), for example, “If a loved one is in a coma he or she will never wake up from, I think it is good to consider euthanasia.” One item was about considering admission to a nursing home: “If a loved one is in a coma he or she will never wake up from, I think it is good to consider care in a nursing home.” Two items were about the importance of feelings in making decisions for a permanently comatose patient (Cronbach’s α = .72), for example, “If a loved one is in a coma he or she will never wake up from, I think you should predominantly follow your feelings when making decisions for that person.” These attitudes represented the different positions taken by the characters in the story.

To measure identification and the remaining dimensions of narrative engagement, a selection of items from Experiment 1 was used. To reduce the burden on participants, fewer items were used than in Experiment 1. Identification was measured with six 7-point rating scale items for both the character that was in favor of considering euthanasia (character FCE) and the character that was against considering euthanasia (character ACE). Attentional focus was also measured with six items. The experience of going into a narrative world, imagery of the story, and emotion evoked by the story were measured with four items each. To increase consistency, an indication of time such as “during reading” was added to every item about narrative engagement.

Factor analysis of the narrative engagement items resulted in the same clusters as in Experiment 1, with some items falling out because of double loading on attention and imagery. Scales were made for “Identification with Character FCE” (six items, Cronbach’s α = .91), “Identification with Character ACE” (six items, Cronbach’s α = .92), “Attentional Focus” (five items, Cronbach’s α = .80), “Being in Narrative World” (six items, Cronbach’s α = .88), and “Emotion” (four items, Cronbach’s α = .86). Finally, the items to measure perceived realism were the same as in Experiment 1 (three items, Cronbach’s α = .68).

Results

Effects of manipulation

With regard to Hypothesis 1 and Hypothesis 2, it was tested whether perspective increased identification with the perspectivizing character and whether the consistency of attitudes was likewise affected. Also, it was checked whether the manipulations had unintentionally influenced the remaining dimensions of narrative engagement and perceived realism. Table 4 contains the mean scores and standard deviations for identification with both characters and attitudes.

Table 4.

Means and Standard Deviations (Between Brackets) for Identification^a and Attitudes^b by Condition

	Perspectivizing character = in favor of considering euthanasia	Perspectivizing character = against considering euthanasia
Identification with character FCE	4.90 (1.09)^c	4.12 (1.30)^c
Identification with character ACE	3.45 (1,22)^c	4.62 (1.18)^c
Attitude toward considering euthanasia	5.75 (.81)^c	5.32 (1.28)^c
Attitude toward considering nursing home	3.96 (1.66)^c	4.49 (1.63)^c
Attitude toward prevalence of feeling	3.94 (1.39)	4.05 (1.30)

1 = very low identification, 7 = very high identification.

1 = very negative attitude, 7 = very positive attitude.

Indicates significant differences.

t tests showed that there was a main effect of perspective on both Identification with Character FCE (t(154) = 4.10, p < .001, η² = .10) and Identification with Character ACE (t(154) = 6.08, p < .001, η² = .20). As expected, participants who had read the version of the story in which the perspectivizing character is in favor of considering euthanasia identified more with this character than participants who had read the version of the story in which the perspectivizing character is against considering euthanasia, and vice versa for Identification with Character ACE. Additionally, a paired-samples t test showed that Identification with Character FCE (M = 4.51, SD = 1.26) was higher overall than Identification with Character ACE (M = 4.03, SD = 1.34; t(155) = 3.51, p =.001, η² = .074). In a 2 × 2 mixed ANOVA with perspective as between-subjects factor and character as within-subjects factor, the interaction was also significant (F(1,154) = 77.85, p < .001, η² = .34). Subsequent simple effects analysis showed that participants who had read the perspective of the character in favor of considering euthanasia identified more with this character FCE than with the character ACE that was the antagonist in this version (p < .001). This effect was reversed for the other perspective; participants who had read the perspective of the character against considering euthanasia identified more with this character ACE than with the character FCE that was the antagonist in this version (p < .01). Thus, participants identified more with the perspectivizing character than with the antagonizing character, regardless of the opinion of the characters.

With regard to the remaining dimensions of narrative engagement and perceived realism, t tests showed that there were no effects of perspective on Attentional Focus, Being in Narrative World, and Emotion (all p’s > .65). There was also no effect of perspective on perceived realism (t(154) = 1.21, p = .23).

With regard to attitudes, t tests showed that there was an effect of perspective on both the attitude toward considering euthanasia (t(154) = 2.52, p < .05, η² = .040) and the attitude toward considering admission to a nursing home if a loved one is in a permanent coma (t(154) = 1.99, p < .05, η² = .025). As expected, participants who had read the perspective of the sister who is in favor of considering euthanasia held a more positive attitude toward considering euthanasia and a more negative attitude toward considering admission to a nursing home than participants who had read the perspective of the sister who is against considering euthanasia and wants to consider admission to a nursing home. For the attitude toward the prevalence of feeling over reason in making decisions for chronically comatose loved ones, there was no main effect of perspective (p = .61).

Mediation analysis

With regard to Hypothesis 3, it was established by mediation analyses using a bootstrapping procedure (Preacher & Hayes, 2008) whether Identification with Character FCE and Identification with Character ACE were responsible for the effects of perspective on attitudes. All mediation analyses used 5,000 bootstrap samples, as recommended by Hayes (2009). Results showed that the indirect effect of Identification with Character FCE on the attitude toward considering euthanasia was significant (B = −. 17, SE = .08, 95% CI = [−.3632, −.0447]). The indirect effect of Identification with Character FCE on the attitude toward considering a nursing home was not significant (B = −. 10, SE = .11, 95% CI = [−.0976, .3287]). For Identification with Character ACE, the indirect effect on both the attitude toward considering euthanasia (B = −.29, SE = .10, 95% CI = [−.5270, −.1321]), and the attitude toward considering a nursing home (B = −. 46, SE = .18, 95% CI = [.1504, .8431]) were significant. Relevant descriptive results are given in Figure 2.

Figure 2.

Descriptive results of the relations between perspective, identification with both characters and the attitude toward considering euthanasia and the attitude toward considering admission to a nursing home when a loved one is in a permanent coma

Control group

Finally, the control group was taken into account to examine the direction of the persuasive impact. t tests with Bonferroni adjustments were carried out to compare the control group to both perspectives. t tests showed that the control group did not differ from any of the perspective conditions on any of the attitudes (all p’s > .089).³

Discussion

The results show that all hypotheses are confirmed. Participants, who read the story that was told from the perspective of the character in favor of considering euthanasia, identified more strongly with that character and held a more favorable posttest attitude toward considering euthanasia. Participants who read the story that was told from the perspective of the character against considering euthanasia, identified more strongly with that character and held both a more negative posttest attitude toward considering euthanasia and a more favorable attitude toward considering a nursing home. In both cases, identification mediated the effects of perspective on attitude.

Identification with the character that is in favor of considering euthanasia did not mediate the attitude toward considering a nursing home. This is probably due to the fact that this character did not oppose considering a nursing home, but rather thought they should consider euthanasia as well. The other character did strongly oppose considering euthanasia and thought a nursing home was the best option. This explains why identification with the character that was against considering euthanasia made the attitude toward considering euthanasia more negative as well as the attitude toward considering a nursing home more positive.

General Discussion

The Role of Identification in Narrative Persuasion

The main aim of this study was to investigate the role of identification with story characters in narrative persuasion. The results of the studies show that identification can indeed be a mediating mechanism of narrative persuasion. Previous studies showed correlations between the level of identification and persuasive outcomes. (e.g., Busselle & Bilandzic, 2009; De Graaf et al., 2009; Iguarta, 2010). However, these correlations could also be explained by the fact that people are more likely to identify with characters holding attitudes similar to their own. The results of the second experiment, in which the only difference between the characters was their opinion about considering euthanasia, show the validity of this reasoning. The paired-samples t test showed that overall, participants’ level of identification with the character in favor of considering euthanasia was higher. The control group’s scores on attitudes indicated that this character’s attitude was more in line with the participants’ prior attitude. However, the perspective manipulation proved to override the impact of attitude similarity. Participants identified more strongly with the perspectivizing character than with the antagonizing character regardless of the opinion of the characters, and subsequently, participants adapted their attitudes accordingly. This study thus provides evidence for the relation between attitude similarity and identification, while at the same time establishing a causal relation between perspective, identification, and narrative persuasion.

Furthermore, the results show that identification can both reinforce and attenuate existing attitudes. In the first experiment, identification with the applicant made attitudes that were already negative, and thus consistent with the applicant, even more negative. This result indicates that identification with a character that shows attitude consistency can make already existing attitudes stronger. In this experiment, identification with the programmer, who had inconsistent attitudes, did not mediate persuasive effects. However, this may have been due to several differences in characteristics of the applicant and programmer. Therefore, in the second experiment, highly similar characters were used. In this experiment, again identification with the character that had consistent attitudes (character in favor) mediated persuasive effects, showing reinforcement of attitudes. Additionally, identification with the character that had inconsistent attitudes (character against) mediated persuasive effects. When identifying with this character, participants’ attitudes toward considering euthanasia become less positive, and thus shifted toward the attitude of the character. This result shows that identification with a character that shows attitude inconsistency can also attenuate readers’ existing attitudes.

Comparisons of the experimental conditions to the control group showed that only one of the story versions significantly changed attitudes from baseline. Even though the consistent, significant differences between the two experimental conditions indicate that there were effects on attitudes, the lack of differences from the control group indicates that these effects were modest. In the second experiment, one perspective shifted the participants’ attitude somewhat in one direction and the other perspective shifted the attitude somewhat in the other direction. As the stories were quite short and not full books or movies, it is unsurprising that the attitudinal effects were modest. Nonetheless, effects were detectable even though short stories were used, which suggests that longer narratives may even have stronger effects. Also, modest effects of single short stories may accumulate if individuals read or view several stories with the same message over time (see Bilandzic & Busselle, 2008).

Although the effects were modest, the finding of this study that both experiments showed persuasive effects of the stories mediated by identification is remarkable because the stories differed in the way in which attitudes were implied. In the first story, attitudes are implied by the characters’ goals. The fact that the applicant is disabled and wants to get the job implies that he opposes policies which would inhibit this goal, such as employers’ rights to reject applicants because they are disabled. This suggests that identification can change attitudes when these are related to the character’s goal, which is in line with Oatley’s (1994, 1999) conceptualization of identification. In the second story, by contrast, the characters state their attitudes directly. The sisters tell each other whether they are in favor of or against considering euthanasia and this position becomes apparent in their thoughts as well. When opinions are stated this clearly, readers who identify with a character can simply shift their own attitudes toward the attitude of the character. The result that identification with story characters led to attitudes more consistent with this character in both experiments thus indicates a robust effect, as both experiments used stories in which attitudes were embedded quite differently.

Limitations and Future Research

In this study, only two stories were used that both had an open ending, which may have influenced persuasion as well. A positive or negative outcome of certain actions can imply an attitude toward these actions. Possibly, identification with characters is less influential when such an outcome is included in the story events. Even though an open ending is a common feature of stories (see Bruner, 1991; Van Dijk, 1975), future research should examine whether identification plays a similar role in persuasion by stories with a specific outcome. Also, the majority of participants of the study were female. Women may more readily identify with characters than men. However, prior research has not found consistent gender differences in narrative persuasion (Green & Brock, 2000, p. 704). It would be interesting for further research to systematically investigate differences between men and women in identifying with characters and being persuaded by narratives. Finally, the potential mediating role of dimensions of narrative engagement that were not focused on in this study, such as the feeling of being in a narrative world, should be addressed in future research. The results of this study show that the dimension of identification with a character can be a mechanism of narrative persuasion.

Footnotes

Appendix

Table A1.

Principal Components Factor Analysis with Oblimin Rotation of Narrative Engagement for Experiment 1 (n = 80)

	1	2	3	4	5
I had a vivid image of the events in the story (IMA, TS)	.883
I pictured the described events (IMA)	.878
While I was reading the story, I visualized the events that took place in it (IMA)	.772
During reading, I had the feeling as if I was present at the events in the story (GNW)	.770
While I was reading the story, I was in the world of the story in my imagination (GNW, TP)	.763
During reading, I saw before me what was described in the story (IMA)	.756
During reading, it was as if I was present in the spaces that were described (GNW)	.678
When I was reading the story, it seemed as if I was there in my thoughts (GNW)	.621
My attention was focused on the events that were described in the story (ATT)	.488			.394
During reading, it felt as if I made a journey to the story world (GNW)	.406		.381
In my imagination the story created another world that I visited (GNW, TP)	.395
As the story progressed, I had the feeling I was being pulled into the story world more and more (GNW)	.372		.303
I had the feeling I went through what the programmer went through (AIP)		.839
In my imagination it was as if I was the programmer (AIP)		.828
I put myself in the position of the programmer (POP)		.817
When I had been reading for a while, it seemed as if I had become the programmer in my thoughts (AIP)		.800
During reading, I imagined what it would be like to be in the position of the programmer (POP)		.795
I empathized with the programmer (EMP)		.777
While I was reading, I pictured what it would be like for the programmer to experience what was described (POP)		.702
I felt for the programmer (EMP)		.698
I found the story moving (EMO)			.816
The story stirred emotions in me (EMO)			.777
Because of the story, feelings arose in me (EMO)			.760
The story affected me (EMO, TS)			.690
While I was reading the story, I forgot my daily affairs (ATT)				.882
During reading, I was fully concentrated on the story (ATT)				.791
When I read the story, my thoughts were only with the story (ATT)				.725
During reading, I was hardly aware of the space around me (ATT, TS)				.707
I did not really notice things that happened around me (ATT)				.677
During reading, I did not think for a while about the things that had been on my mind lately (ATT)				.632
During reading, my attention was fully captured by the story (ATT)	.354		.344	.465
When I had been reading for a while, it seemed as if I had become the applicant in my thoughts (AIA)					.845
I empathized with the applicant (EMA)					.830
In my imagination, it was as if I was the applicant (AIA)					.808
I had the feeling I went through what the applicant went through (AIA)					.778
I put myself in the position of the applicant (POA)					.729
During reading, I imagined what it would be like to be in the position of the applicant (POA)					.719
While I was reading, I pictured what it would be like for the applicant to experience what was described (POA)					.676
I felt for the applicant (EMA)					.648
% of variance explained (before rotation)	28.30	14.80	11.78	6.78	4.34
Sum of squared loadings	8.14	6.32	5.75	5.63	6.37

Note: Factor loadings <.30 are not reported. ATT = item about attention focused on the story, EMO = item about emotional reactions evoked by the story, IMA = item about imagery of the story, GNW = item about the sensation of going into a narrative world, EMA/EMP = item about empathy with the applicant/programmer, AIA/AIP = item about the illusion of being the applicant/programmer, POA/POP = item about experiencing the story from the position of the applicant/programmer, TS = item based on Transportation Scale (Green & Brock, 2000), and TP = item based on Telepresence Scale (Kim & Biocca, 1997).

Acknowledgements

The authors would like to thank Donna Rouner, Gisela Redeker, Jan van den Bulck and two anonymous reviewers for their helpful comments on earlier versions of this article. Also, the authors would like to thank Christian Burgers for his help with translating the materials.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Notes

Bios

Anneke de Graaf is an assistant professor at the Amsterdam School of Communication Research (ASCoR) at the University of Amsterdam.

Hans Hoeken is a full professor at the Department of Business Communication at the Radboud University Nijmegen.

José Sanders is an associate professor at the Department of Business Communication at the Radboud University Nijmegen.

Johannes W. J. Beentjes is a full professor at the Amsterdam School of Communication Research (ASCoR) at the University of Amsterdam.

References

Andringa

(1986). Perspektivierung und Perspektivenuebernahme [Perspectivization and taking over of perspective]. Spiel, 5(1), 135-146.

Andsager

J. L.

Bemker

Choi

H. L.

Torwel

(2006). Perceived similarity of exemplar traits and behavior: Effects on message evaluation. Communication Research, 33(1), 3-18.

Appel

Richter

(2007). Persuasive effects of fictional narratives increase over time. Media Psychology, 10(1), 113-134.

Bal

(1997). Narratology, introduction to the theory of narrative (2nd ed.). Toronto, Ontario, Canada: University of Toronto Press.

Bandura

(1986). Social foundations of thought and action: A social cognitive theory. Englewood Cliffs, NJ: Prentice-Hall.

Bandura

(2002). Social cognitive theory of mass communication. In Bryant

Zillmann

(Eds.), Media effects: Advances in theory and research (2nd ed., pp. 121-153). Mahwah, NJ: Lawrence Erlbaum.

Bilandzic

Busselle

(2008). Transportation and transportability in the cultivation of genre-consistent attitudes and estimates. Journal of Communication, 58, 508-529.

Bruner

(1991). The narrative construction of reality. Critical Inquiry, 18(1), 1-21.

Busselle

Bilandzic

(2008). Fictionality and perceived realism in experiencing stories: A model of narrative comprehension and engagement. Communication Theory, 18, 255-280.

10.

Busselle

Bilandzic

(2009). Measuring narrative engagement. Media Psychology, 12, 321-347.

11.

Busselle

Ryabovolova

Wilson

(2004). Ruining a good story: Cultivation, perceived realism and narrative. Communications—The European Journal of Communication Research, 29, 365-378.

12.

Cohen

(2001). Defining identification: A theoretical look at the identification of audiences with media characters. Mass Communication and Society, 4, 245-264.

13.

De Graaf

Hoeken

Sanders

Beentjes

(2009). The role of dimensions of narrative engagement in narrative persuasion. Communications—The European Journal of Communication Research, 34, 385-405.

14.

Diekman

A. B.

McDonald

Gardner

W. L.

(2000). Love means never having to be careful: The relationship between reading romance novels and safe sex behavior. Psychology of Women Quarterly, 24, 179-188.

15.

Eyal

Rubin

A. M.

(2003). Viewer aggression and homophily, identification, and parasocial relationships with television characters. Journal of Broadcasting and Electronic Media, 47, 77-98.

16.

Gerrig

R. J.

(1993). Experiencing narrative worlds. New Haven, CT: Yale University Press.

17.

Green

M. C.

(2004). Transportation into narrative worlds: The role of prior knowledge and perceived realism. Discourse Processes, 38, 247-266.

18.

Green

M. C.

(2006). Narratives and cancer communication. Journal of Communication, 56, S163-S183.

19.

Green

M. C.

Brock

T. C.

(2000). The role of transportation in the persuasiveness of public narratives. Journal of Personality and Social Psychology, 79, 701-721.

20.

Hayes

(2009). Beyond Baron and Kenny: Statistical mediation analysis in the new millennium. Communication Monographs, 76, 408-420.

21.

Iguarta

J. J.

(2010). Identification with characters and narrative persuasion through fictional feature films. Communications—The European Journal of Communication Research, 35, 347-373.

22.

Kim

Biocca

(1997). Telepresence via television: Two dimensions of telepresence may have different connections to memory and persuasion. Journal of Computer-Mediated Communication, 3(2). Retrieved from http://onlinelibrary.wiley.com/doi/10.1111/j.1083-6101.1997.tb00073.x/full

23.

Mar

R. A.

Oatley

(2008). The function of fiction is the abstraction and simulation of social experience. Perspectives on psychological science, 3, 173-192.

24.

Morgan

S. E.

Movius

Cody

M. J.

(2009). The power of narratives: The effect of entertainment television organ donation storylines on the attitudes, knowledge, and behavior of donors and nondonors. Journal of Communication, 59, 135-151.

25.

Moyer-Gusé

(2008). Toward a theory of entertainment persuasion: Explaining the persuasive effects of entertainment-education messages. Communication Theory, 18, 407-425.

26.

Moyer-Gusé

Nabi

R. L.

(2010), Explaining the Effects of Narrative in an Entertainment Television Program: Overcoming Resistance to Persuasion. Human Communication Research, 36, 26–52.

27.

Oatley

(1994). A taxonomy of the emotions in literary response and a theory of identification in fictional narrative. Poetics, 23(1), 53-74.

28.

Oatley

(1999). Meeting of minds: Dialogue, sympathy and identification in reading fiction. Poetics, 26, 493-454.

29.

Preacher

Hayes

(2008). Asymptotic and resampling strategies for assessing and comparing indirect effects in multiple mediator models. Behavior Research Methods, 40, 879-891.

30.

Sanders

Redeker

(1996). Speech and thought in narrative discourse. In Fauconnier

Sweetser

(Eds.) Spaces, worlds and grammar (pp. 290-317). Chicago, IL: University of Chicago Press.

31.

Slater

M. D.

Rouner

(2002). Entertainment-education and elaboration likelihood: Understanding the processing of narrative persuasion. Communication Theory, 12, 173-191.

32.

Strange

J. J.

Leung

C. C.

(1999). How anecdotal accounts in news and in fiction can influence judgments of a social problem’s urgency, causes, and cures. Personality and Social Psychology Bulletin, 25, 436-449.

33.

Tan

E. S. H.

(1994). Film-induced affect as a witness emotion. Poetics, 23(1), 7-32.

34.

van Dijk

T. A.

(1975). Action, action description and narrative. New Literary History, 6, 273-294.

35.

van Peer

Pander Maat

(2001). Narrative perspective and the interpretation of characters’ motives. Language and Literature, 10, 229-241.

36.

Zillmann

(1994). Mechanisms of emotional involvement with drama. Poetics, 23(1), 33-51.

37.

Zillmann

(2006). Empathy: Affective reactivity to other’s emotional experiences. In Bryant

Vorderer

(Eds.), Psychology of entertainment (pp. 151-181). New York, NY: Lawrence Erlbaum.