Sometimes stories sell: When are narrative appeals most likely to work?

Abstract

Research has demonstrated that narratives can be effective in eliciting attitude change, especially when recipients become transported into the narrative. In three studies, we addressed whether some people are predisposed to be influenced by narratives and whether narrative and rhetorical appeals are differentially effective for different people. In Study 1, participants read an experimental or a control narrative, and completed measures of attitudes, need for affect (NFA), need for cognition (NFC), transportation, and transportability. The results revealed that NFA and NFC were positively correlated with transportation and transportability. In Study 2, participants read either a narrative appeal or a rhetorical appeal about cervical cancer and completed a measure of attitudes and the individual difference constructs. Study 3 was a replication of Study 2 using a different topic (organ donation). In both studies, the results revealed a consistent pattern of correlations among the individual difference measures. Further, we found that although the narrative and rhetorical appeals were judged to be of equal efficacy, the persuasiveness of the narrative appeal differed as function of individual differences in NFA and NFC. The implications for narrative persuasion are discussed. Copyright © 2011 John Wiley & Sons, Ltd.

Narratives are an important part of everyday life, from parents reading bedtime fables to their children to people lining up to see the latest Hollywood blockbuster. That said, within the attitudes literature, research has devoted limited attention to the role of narratives as persuasive tools, instead focusing on the role of arguments in changing attitudes (see Maio & Haddock, 2007, 2010). Recently, however, researchers have started to address the effectiveness of narratives in eliciting attitude change, with promising results (e.g., Dal Cin, Zanna, & Fong, 2004; Green & Brock, 2000; Mazzocco, Green, Sasota, & Jones, 2010). The present research addressed two questions: (a) whether some people are predisposed to be influenced by narratives and (b) whether receptivity to narrative and rhetorical appeals differs across people.

Narrative Persuasion

Research has provided evidence that narratives can influence people's attitudes (Green & Brock, 2000; Marsh, Meade, & Roediger 2003; Mazzocco et al., 2010; Prentice, Gerrig, & Bailis, 1997). For example, Green and Brock (2000) found that readers who became involved in a story subsequently reported attitudes more congruent with the narrative's theme. In one study, Green and Brock (2000; Study 4) asked participants to read a short story about an Eskimo boy and his dog who become stranded on an ice floe, with the story intending to highlight the values of friendship and loyalty in difficult circumstances. The boy and his dog have no food or supplies, and during the night, both characters consider killing the other for food, but do not, through loyalty to each other. In the morning, the boy and dog are rescued. Green and Brock (2000) found that readers who became involved in the story subsequently attributed greater importance to the values of friendship and loyalty between humans and animals.

Green and Brock (2000) claim that transportation is the mechanism underlying this type of attitude change. As described by Gerrig (1993), transportation is a process by which a person engaging with a particular narrative becomes immersed into the narrative world. While engaged in the narrative, individuals become involved in events, are swept up in the emotional journey of the story, and then emerge at the end, changed by the trip. It is posited that when individuals are transported into a particular narrative, they can become persuaded by aspects of the narrative's content. For example, Green and Brock (2000) suggested that narratives may be particularly persuasive because highly transported readers lose access to some of their real world knowledge and are unable to counterargue information they encounter within the story. Evidence supports the assertion that the process of transportation reduces responses that contradict a story's central message (Escalas, 2007; Marsh & Fazio, 2006; Slater & Rouner, 2002). Indeed, Dal Cin et al. (2004) suggest that narratives can ‘slip under the radar,’ implying that individuals are often unaware of an attempt to change their attitude.

Dal Cin et al. (2004) also suggested that there are individual differences in the extent to which people generally become transported into narratives. They suggested that being transported into a particular narrative, as measured by Green and Brock's (2000) transportation scale, differs from an individual's general tendency to become transported into any narrative. To explore the relation between these concepts, Dal Cin et al. (2004) created a transportability scale and tested whether individuals high in transportability are more likely to become transported into narratives and more likely to be persuaded by messages embedded within a narrative. Supporting their argument, Dal Cin et al. (2004) found that transportation was correlated with transportability and that transportation predicted post-narrative attitudes after accounting for transportability and participants' pre-existing attitudes toward the narrative topic (see Mazzocco et al., 2010, for a comparable finding).

Transportation, Transportability, and Individual Difference Constructs

Although there is a growing body of evidence suggesting that narratives can elicit attitude change, and that transportation has a role in bringing about this change (see Green & Brock, 2005), limited attention has been paid to who may be most susceptible to transportation. One focus of this paper is a consideration of individual difference constructs, in addition to transportability, that may be associated with transportation. In this research, we focused on studying the relation of two individual difference measures with transportation, transportability, and receptivity to narrative and rhetorical appeals; need for affect (Maio & Esses, 2001) and need for cognition (Cacioppo & Petty, 1982; Cacioppo, Petty, Feinstein, & Jarvis, 1996).

The need for affect (NFA) is conceptualized as a motivation to seek out and become involved in emotional situations, whereas the need for cognition is conceptualized as a motivation to engage in and enjoy effortful cognitive activities (Cacioppo & Petty, 1982; Maio & Esses, 2001; see Haddock, Maio, Arnold, & Huskinson, 2008, for a brief review of each construct). Previous research has demonstrated the relevance of these two important constructs in relation to attitude change processes (e.g., see Briñol & Petty, 2005; Maio & Haddock, 2010). For example, Haddock et al. (2008) tested whether individual differences in need for affect and need for cognition were associated with differential levels of attitude change resulting from affect-based and cognition-based appeals. These authors found that individual differences in need for affect influenced receptivity to an affect-based (but not a cognition-based) message, whereas individual differences in need for cognition influenced receptivity to a cognition-based (but not an affect-based) message. This line of research provides initial evidence about the joint importance of need for affect and need for cognition in assessing the relative success of different persuasive appeals. Building upon this research, the current studies considered how individual differences in need for affect and need for cognition influence how individuals respond to narrative versus rhetorical persuasive appeals.

As applied to the present context, the need for affect is likely to be related to the utility of narrative persuasive appeals. Research has demonstrated that individuals high in the need for affect are more likely to watch emotional movies and become involved in public emotional events compared with individuals low on this construct (Maio & Esses, 2001). This motivation to experience emotion should encourage readers to become engaged with a narrative, as narratives tend to be evocative and to induce emotion. Participants high in NFA should be more likely to become absorbed within a narrative and identify with the characters, as they are involved in the events within the story. This motivation to engage with emotional situations should enhance transportation and be associated with enhanced transportation and transportability.

Regarding the need for cognition, research has found that individuals high in NFC are more likely to carefully process persuasive arguments and respond differently to strong and weak arguments (Cacioppo et al., 1996). In the context of narrative persuasion, individuals high in NFC should be likely to engage with a narrative, as they are likely to process the text carefully, attempting to fully understand the language and investing cognitive effort in working through the twists and turns of the plot. This should enhance engagement with the narrative and enhance the transportation process (cf. Green & Brock, 2000). Similarly, NFC might also be associated with transportability, such that people who are high in NFC might be more transportable in general.

STUDY 1

In Study 1, participants read an experimental or control narrative, followed by a measure of attitude, the transportation scale and measures of transportability, NFA and NFC. This study addressed two questions. First, we tested whether transportability, NFA and NFC were correlated with transportation. Second, we sought to replicate the findings of Green and Brock (2000) who found that more transported participants reported attitudes more congruent with the narrative theme. With this in mind, we included a control narrative with very different content to the experimental narrative, to ensure that any attitude change resulting from transportation into the experimental narrative would be specific to the content of that narrative (see Green & Brock, 2000). These narratives were approximately the same length and were rated by participants as being equally transporting (t < 1).

METHOD

Participants

Ninety undergraduate psychology students (83 women, 7 men; mean age = 20 years) from a British university participated in return for course credit.

Materials: Short Stories

Experimental Narrative—Two Were Left (Cave, 1956)

This narrative tells the story of a young Eskimo boy (Noni) and his dog (Nimuk). The two protagonists become stranded on an ice floe with no food. In turn, each character considers killing and eating the other, but neither does. The pair is eventually rescued by a pilot. This story was included as the experimental narrative, as it makes salient the values of friendship and loyalty. This narrative has been used in previous research (Green & Brock, 2000; Experiment 4).

Control Narrative—Breakfast (Steinbeck, 1938)

This narrative describes a scene in which the story narrator comes across a small family preparing breakfast. This narrative is included as the control narrative, as it does not engage with the same types of issues as ‘Two were left.’

Materials: Dependent Measures

Attitude Measure (Green & Brock, 2000)

The attitude measure used in this study was developed by Green and Brock (2000) to measure attitudes specific to the content of the Two Were Left narrative. The measure is intended to assess participants' perceptions about the importance of close friendships and loyalty between close friends within the context of this survival story about a dog and his master. Sample items from the scale include ‘A person should lay down their life for their best friend’ and ‘A starving person may betray friends to obtain food.’ Participants indicated how much they agreed or disagreed with 10 statements, with zero indicating total disagreement and seven indicating total agreement. A high score represents a more positive attitude (see Table 1 for Cronbach α value).

Table 1. Study 1: Correlations among individual difference variables and attitudes

	NFA	NFC	Transportation	Transportability	Attitude
Note: NFA, need for affect; NFC, need for cognition. * significant at .05 level, 1-tailed; ** significant at .01 level, 1-tailed. Cronbach's α on diagonal.
NFA	0.90	0.26*	0.37**	0.47**	0.35**
NFC		0.92	0.31**	0.36**	0.23
Transportation			0.83	0.64**	0.33**
Transportability				0.87	0.16
Attitude					0.62
Scale means	1.02	0.75	4.26	5.98	4.57
Scale SD	0.84	1.27	0.82	0.97	0.70

Transportation Scale (Green & Brock, 2000)

This scale is comprised of 15 items that describe the effects of becoming transported into a specific narrative. A sample item is ‘I was mentally involved in the narrative while reading it.’ Participants responded to each item on a scale ranging from one (not at all) to seven (very much). A high score indicates that an individual is highly transported into the particular narrative (see Tables 1, 2, 3 for Cronbach α values in the individual studies).

Table 2. Study 2: Correlations among individual difference variables

	NFA	NFC	Transportation	Transportability
Note: NFA, need for affect; NFC, need for cognition. * significant at .10 level; ** significant at .05 level; *** significant at .01 level (1-tailed). Cronbach's α on diagonal.
NFA	0.89	0.23**	0.20*	0.17**
NFC		0.88	0.12	0.32***
Transportation			0.84	0.59***
Transportability				0.90
Scale mean	1.02	0.89	5.08	6.35
Scale SD	0.83	1.15	0.68	1.01

Table 3. Study 3: Correlations individual difference variables

	NFA	NFC	Transportation	Transportability
Note: NFA, need for affect; NFC, need for cognition. * significant at .05 level; ** significant at .01 level (1-tailed). Cronbach's α on diagonal.
NFA	0.87	0.24*	0.42**	0.39**
NFC		0.89	0.31*	0.30**
Transportation			0.86	0.51**
Transportability				0.93
Scale mean	1.09	0.56	4.66	5.88
Scale SD	0.73	1.16	0.91	1.26

Transportability Scale (Dal Cin et al., 2004)

The transportability scale was developed by Dal Cin et al. (2004) to measure individual differences in the extent that people are generally predisposed to transportation into a narrative. The scale contains 20 items, a sample item being ‘I am often emotionally affected by what I have read.’ Participants responded to each item on a scale ranging from one (not at all) to nine (very much). A high score indicates that an individual is highly transportable (see Tables 1, 2, 3 for Cronbach α values in the individual studies).

Need for Affect Scale (Maio & Esses, 2001)

This scale features 26 items that describe the way people feel about emotions and how they approach or avoid emotional situations. A sample item is ‘I feel that I need to experience strong emotions regularly.’ Participants responded to each item on a scale ranging from plus three (strongly agree) to minus three (strongly disagree). A high score indicates that an individual is high in NFA (see Tables 1, 2, 3 for Cronbach α values in the individual studies).

Need for Cognition Scale (Cacioppo & Petty, 1982)

This scale features 18 items about the kind of mental activities people like or dislike to engage in. A sample item from this scale is ‘I would prefer complex to simple problems.’ Participants responded to each item on a scale ranging from plus four (strongly agree) to minus four (strongly disagree). A high score indicates that an individual is high in NFC (see Tables 1, 2, 3 for Cronbach α values in the individual studies).

PROCEDURE

Participants completed the study individually. The experimenter explained that she was interested in how people engage with short stories when reading for pleasure. Participants were then randomly given the experimental or control narrative and told to read the story at their own pace and alert the experimenter when they had finished. When participants had read the narrative, they were given the attitude measure and transportation scale to complete. After completing this task, participants were given the transportability, NFA, and NFC scales (in a randomized order, along with other materials not relevant to the paper).1

RESULTS AND DISCUSSION

Correlational Analyses

We expected that NFA and NFC would be positively related to transportation and transportability. The correlations among these variables are presented in Table 1. These correlations are based on responses provided by participants in both the experimental and control narrative conditions. The only exception is for the attitudes measure, in which we present data from the experimental condition to assess how a topic-relevant narrative influences attitudes. As expected, participants who reported high scores on NFA or NFC were more transported into the narrative and more transportable in general.

Transportation and Condition as Predictors of Attitudes

Following Green and Brock (2000), it was hypothesized that participants who were more transported into the experimental narrative would be more likely to express attitudes congruent with the narrative. A regression analysis was conducted where condition (dummy coded), transportation, and their interaction were entered as predictors of attitudes (based upon Aiken & West, 1991). The analysis revealed a marginal main effect of condition (β = −.93, p = .10). Participants who read the experimental narrative tended to express more message-congruent attitudes than participants who read the control narrative. This effect was qualified by a marginally significant interaction (β = .99, p = .08). As expected, transportation was positively related to attitudes in the experimental condition (β = .33, p = .02), whereas there was no relation between transportation and condition in the control condition (β = −.04, ns). Among participants who read the experimental narrative, we also considered whether transportation mediated the relation between transportability and attitudes. We found that transportation partially mediated the marginal relation between transportability and attitudes (Sobel z = 1.83, p < .07). This is consistent with the results described by Dal Cin et al. (2004).

In review, Study 1 had two aims. First, we explored the relations among NFA, NFC, transportation, and transportability. Consistent with existing research, we expected that transportation and transportability would be positively correlated. Of greater interest, we expected that both NFA and NFC would be positively related to transportation and transportability. Second, we sought to replicate Green and Brock's (2000) findings that participants who were more transported into the narrative would be more persuaded than those who were less transported into the narrative.

Overall, the results were consistent with our expectations. First, we found that NFA and NFC were positively related to transportation and transportability. Both NFA and NFC tap motivations to approach certain types of situations. Individuals high in NFA are motivated to approach emotional situations and individuals high in NFC are motivated to approach situations that are cognitively challenging. In the context of narrative persuasion, one can speculate that these two motivations appear to encourage engagement with the narrative and lead to enhanced transportation. Second, consistent with Green and Brock (2000), participants who were more transported into the experimental narrative were more persuaded by its content than participants who were less transported.

STUDY 2

Study 2 had two aims. First, we sought to replicate the findings of Study 1 regarding the relations among the individual difference measures, this time using a different narrative topic of an applied nature. Second, we explored whether the persuasiveness of narrative and rhetorical appeals differs across individuals. To test these ideas, we developed two appeals that advocated screening for cervical cancer. The narrative appeal was adapted from an article in a magazine about a woman who suffered from cervical cancer. The rhetorical appeal was developed by taking the medical information from the narrative and presenting it in a clear, bullet-point form. Both appeals endorsed the same positive attitude toward cervical screening.

Based on the results of Study 1, we made a number of predictions concerning whether the two appeals might elicit differentially favorable attitudes as a function of individual differences in NFA and NFC. First, we anticipated that the narrative appeal would be more persuasive among high NFA individuals compared with low NFA individuals. The emotional nature of the narrative should lead individuals who are high in NFA to be more engaged with the narrative's content, which should result in greater attitude change. The rhetorical message, in contrast, is less emotionally evocative, and participants high in NFA should be less likely to become engaged with its content. Put differently, NFA should be positively associated with attitudes among participants who read the narrative appeal, with a null (or possibly opposite) effect among participants who read the rhetorical appeal.

When considering NFC, Study 1 found that NFC was positively related to transportation, suggesting that individuals high on this construct may be more likely to become transported into a narrative and perhaps be persuaded by it. It was less clear, however, whether a narrative appeal might be differentially effective in altering attitudes as a function of individual differences in NFC. It could be argued that the narrative appeal might be more effective among high (compared with low) NFC individuals, because high NFCs are likely to become more transported into the narrative. This engagement with the narrative could make a narrative appeal potentially more persuasive among high NFC individuals than among low NFC individuals.

METHOD

Participants

Ninety-four undergraduate female psychology students (mean age = 20 years) from a British university participated in this study for partial course credit.

Materials: Appeals

Narrative Appeal

This narrative was adapted from a feature in Glamour magazine (Lyle, 2006). It featured a girl who contracted cervical cancer and followed her progress through her treatment and palliative care. At the end of the narrative she dies from her condition. The narrative contained information about screening for cervical cancer and treatment of cervical cancer.2

Rhetorical Appeal

This appeal took the medical information from the narrative and put it into a format that was similar to health advocacy leaflets available from a doctor. The feature was headed with a short paragraph that introduced the information and included bullet points and short simple sentences. The narrative and rhetorical appeals were both four pages in length.

Materials: Dependent Measures

Attitude Measure

Participants completed eight items that assessed participants' perceived importance about regular cervical screening and whether they intended to have regular screening in the future. Sample items include ‘It is important for me to have regular cervical smear tests’ and ‘I intend to get regular smear tests in the future.’ Participants indicated their agreement on a nine-point scale, with high scores indicating high agreement. The items were combined to form a single index (α = .78).

Appeal Efficacy

A measure was developed to assess the perceived efficacy of the appeals. Participants rated how informative, useful, and effective the appeal was in highlighting the importance of regular smear tests. These responses were made on nine-point scales with high values indicating greater perceived informational value, usefulness, and effectiveness. These items were combined to form an index of appeal efficacy (α = .84).

Message Evaluation

Participants were asked to write down the thoughts they had about the good and bad qualities of the narrative. We then asked participants to go back through their responses and to indicate which thoughts they thought were positive, negative or neutral (i.e., participants served as their own raters). A message evaluation score was derived by subtracting the number of negative qualities listed by participants from the number of positive statements they provided. This measure was included to further assess the efficacy of the different appeals.

Materials: Individual Difference Constructs

Individuals completed the measures as described in Study 1. The transportation measure was completed only by participants in the narrative condition, as it was neither relevant nor appropriate to the rhetorical appeal.

PROCEDURE

Participants completed the study individually. The experimenter explained that she was interested in how women used magazines and pamphlets as sources of health information. Participants were then given either the narrative or the rhetorical appeal. They were told to read the information at their own pace and alert the experimenter when they had finished. When participants had read the appeal, they completed the efficacy and message evaluation items and were given the attitude measure and transportation scale to complete (rhetorical condition participants did not complete the transportation scale; as its items are not appropriate to the nature of a rhetorical appeal). After completing this task, participants completed the transportability, NFA and NFC scales (in a randomized order, along with other materials not relevant to the paper).

RESULTS AND DISCUSSION

Preliminary Analyses

Analyses were performed on the appeal efficacy and message evaluation measures, to ensure that the appeals were perceived to be equally effective and strong.

Appeal Efficacy

A t-test assessed whether participants perceived the appeals to be equally effective. There was no difference between the narrative (M = 7.67) and rhetorical (M = 7.49) appeals (t < 1), indicating that participants perceived the appeals to be of equal efficacy. Appeal efficacy was not related to scores on the individual difference constructs.

Message Evaluation

Analyses were conducted on the number of positive (M_Narr = 2.02, M_Rhet = 2.00) and negative (M_Narr = 1.10, M_Rhet = 1.27) thoughts elicited by the appeals. There were no differences across conditions (ps > .30), suggesting that the messages were equally strong.3

Correlation Analyses

The pattern of correlations among NFA, NFC, transportation, and transportability are presented in Table 2. The correlations are based on responses provided by participants in both the narrative and rhetorical conditions, except for transportation, as this measure was completed only by participants in the narrative condition. Overall, the pattern of relations is consistent with what was found in Study 1.

Regression Analyses

Need for Affect

Condition (dummy coded), NFA, and their interaction were entered into a regression as predictors of attitude. The results revealed that the interaction was a marginally significant predictor of attitude (β = .32, p = .09). As seen in Figure 1, simple slopes analyses revealed that the narrative appeal was more influential among high NFA individuals compared with low NFA individuals (β = .23, p = .05). The impact of the rhetorical appeal did not vary as a function of individual differences in NFA (p > .20).4

Figure 1. Study 2: The impact of need for affect and appeal type on attitudes

Need for Cognition

Condition (dummy coded), NFC, and their interaction were entered into a regression as predictors of attitude. The results revealed that the interaction was a significant predictor of attitude (β = .26, p = .05). As seen in Figure 2, simple slopes analyses revealed that the narrative appeal was more influential among high NFC individuals compared with low NFC individuals (β = .12, p < .01). The impact of the rhetorical appeal did not vary as a function of individual differences in NFC (p > .20).

Figure 2. Study 2: The impact of need for cognition and appeal type on attitudes

The first aim of Study 2 was to replicate the findings from Study 1 regarding the relations among NFA, NFC, transportation, and transportability. On the whole, the pattern of relations was replicated. The second aim of this study was to assess whether the appeals were differentially effective across people. The results revealed that the individual difference variables showed evidence of different responsiveness to the narrative and rhetorical appeals. These effects were found despite the appeals being judged as equally effective. In the case of the narrative appeal, it was found that high NFA participants were more persuaded by this type of appeal relative to participants low in NFA. This was expected; as individuals high in NFA are more likely to approach emotional situations, they would have been drawn into the narrative appeal and become involved in the emotional events occurring. With respect to NFC, it was found that the narrative appeal was more influential among high NFC participants compared with low NFC participants. It is possible that NFC motivates people to become engaged with the narrative, which enhances transportation. Green and Brock (2000) suggest that transportation can lead to a reduction in counter-arguing. Thus, once transported, high NFC people may not highly scrutinize the narrative message, resulting in message-congruent attitude change.

STUDY 3

In Study 3, we sought to replicate and generalize the findings of Study 2 regarding differences in how participants responded to narrative and rhetorical appeals, using a different topic—organ donation. In this study, we developed a short narrative about a woman who decided to undergo a life-altering lung transplant. We also developed a rhetorical appeal that gave factual information about organ donation. Both appeals were equivalent in factual content, such that the relevant medical information from the rhetorical appeal appeared verbatim in the narrative. Participants read one of the appeals before completing a measure of attitudes towards organ donation (and British laws surrounding organ donation) and the individual difference measures. Consistent with the findings of Study 2, we expected that the effectiveness of the narrative appeal would differ as function of individual differences in NFA and NFC.

METHOD

Participants

Sixty undergraduate psychology students (52 women, 8 men; mean age 20 years) from a British university participated in return for course credit.

Materials: Stimulus Materials

Narrative Appeal

This appeal was developed around the case of a woman, Sarah, who had cystic fibrosis, a genetic condition that causes severe respiratory problems. The story details how Sarah needed a lung transplant, and how the transplant changed her life. Embedded in the story were facts about organ donation in Britain, for example how many people need transplants every year (9000) versus how many transplant operations are actually performed (3000). The narrative endorsed donating organs and a new opt out system of organ donation in which people would automatically donate their organs unless they opted out of the scheme.

Rhetorical Appeal

The rhetorical appeal included all of the factual information that was embedded in the narrative appeal, excluding the basic story elements of the narrative.

Materials: Dependent Measures

Attitude Measure

Participants answered seven statements (e.g., ‘I would like to put my name on the organ donation register’) tapping their attitudes towards organ donation. Responses were made on nine-point scales ranging from 1 (do not agree) to 9 (agree very strongly). The items were combined to form a single index (α = .78).

Additional Measures

Participants completed measures of appeal efficacy (α = .78) and thought-listing, and the transportation (participants in the narrative condition only), transportability, NFA and NFC scales as in Study 2.

PROCEDURE

Participants completed the study individually. They were given either the narrative or the rhetorical appeal and told to read the information at their own pace and alert the experimenter when they had finished. After participants had read the appeal, they completed the efficacy and message evaluation items and were given the attitude measure and transportation scale to complete (as in Study 2, rhetorical appeal condition participants did not complete the transportation scale). Following completion of this task, participants completed the transportability, NFC, and NFA scales (in a randomized order).

RESULTS AND DISCUSSION

Preliminary Analyses

Analyses were performed on the appeal efficacy and message evaluation measures, to ensure that the narrative and rhetorical appeals were perceived to be equally strong.

Appeal Efficacy

A t-test assessed whether participants perceived the narrative and rhetorical appeals to be equally effective. There was no difference between the narrative (M = 7.37) and rhetorical (M = 7.02) appeals (t < 1), indicating that participants perceived the appeals to be of equal efficacy. Appeal efficacy was not related to scores on the individual difference constructs.

Message Evaluation

Analyses were conducted on the number of positive (M_Narr = 2.00, M_Rhet = 1.83) and negative (M_Narr = 0.90, M_Rhet = 0.86) thoughts elicited by the appeals. There were no differences between conditions (ps > .30), suggesting that the messages were equally strong.5

Correlation Analyses

The pattern of correlations among NFA, NFC, transportation, and transportability are presented in Table 3. As in Study 2, these data are based on responses provided by participants in both the narrative and rhetorical conditions, except for transportation, as this measure was completed only by participants in the narrative condition. Overall, the pattern of relations is consistent with what was found in Studies 1 and 2—both NFA and NFC were positively correlated with both transportation and transportability. Participants who were high on these variables were more transported in the narrative condition, and more transportable overall.

Regression Analyses

Need for Affect

Condition (dummy coded), NFA, and their interaction were entered into a regression as predictors of attitude. The results revealed that the interaction was a significant predictor of attitude (β = .58, p < .05). As seen in Figure 3, simple slopes analyses revealed that the narrative appeal was more influential among high NFA individuals compared with low NFA individuals (β = .57, p < .05). In contrast, the impact of the rhetorical appeal did not vary as a function of individual differences in NFA (p > .15).

Figure 3. Study 3: The impact of need for affect and appeal type on attitudes

Need for Cognition

Condition (dummy coded), NFC, and their interaction were entered into a regression as predictors of attitude. The results revealed that the interaction was a significant predictor of attitude (β = .36, p < .05). As seen in Figure 4, simple slopes analyses revealed that the narrative appeal was more influential among high NFC individuals compared with low NFC individuals (β = .21, p = .07). In this study, the rhetorical appeal was more influential among low NFC individuals compared with high NFC individuals (β = −.39, p < .05).

Figure 4. Study 3: The impact of need for cognition and appeal type on attitudes

The results of Study 3 supported the findings of the previous studies in two important ways. First, it replicated Studies 1 and 2 with respect the relations among NFA, NFC, transportation, and transportability. Second, this study revealed convergent patterns regarding the effectiveness of the narrative and rhetorical appeals across people. Once again, the narrative appeal was more effective among high NFAs compared with low NFAs. Regarding NFC, this study replicated Study 2 in showing that the narrative appeal was differentially effective as a function of individual differences in NFC (though the effect was smaller in this study). Interestingly, the rhetorical appeal was more effective among low NFCs compared with high NFCs. We suggest that while both appeals were perceived by participants to be objectively effective, the topics of illness and death, salient in both appeals, may have been personally threatening to participants and may have provoked resistance to the persuasive attempt.

GENERAL DISCUSSION

The present research was designed to address novel issues relevant to narrative persuasion. First, it explored the relations among relevant individual difference constructs with transportation and transportability. Second, it investigated whether narrative and rhetorical appeals vary in effectiveness as a function of individual differences in NFA and NFC.

Need for Affect, Need for Cognition, Transportation, and Transportability

The results of the research shed light on the role of individual difference constructs in response to narrative persuasion. Across the studies, NFA and NFC were positively related with both transportation and transportability. Individuals high in NFA are more likely to approach emotional situations and thus should be more likely to approach and engage with the emotional nature of narrative appeals. Individuals reading the narrative appeal are then more likely to become emotionally involved with the characters and the events occurring within the narrative. When they have finished reading the narrative, they are likely to be left with an impression of the narrative that is in some way relevant to them. This impression would be linked to any attitude formed during the transportation process.

Individuals high in NFC are motivated to process information thoroughly. This motivation to approach situations that are cognitively involving would result in individuals high on this construct being more likely to engage with the narrative, rather than passively reading it. However, once engaged with the narrative, these individuals become transported and swept along with the events. Transportation is conceptualized as a process that reduces responses that would contradict the central themes of a narrative, and so individuals, once transported, may become less likely to counterargue messages. Thus, NFC could enhance transportation by engagement with the narrative and through the motivation to more thoroughly understand the narrative, but the transportation process may ultimately reduce counter-arguing. As discussed, this idea is at present speculative but does provide an exciting avenue for future research.

Narrative versus Rhetorical Appeals

Across two different health issues, the results of Studies 2 and 3 revealed that individual differences in NFA were associated with receptivity to a narrative appeal, with high NFA individuals reporting more appeal-congruent attitudes in response to this type of persuasive communication. This suggests that participants who are motivated to approach emotional situations were more likely to engage with the narrative appeal compared to those low on this construct. Participants who are more likely to engage with the narrative message are more likely to become transported into it and persuaded by it. When they have finished reading the narrative, they are left with an emotional memory or impression of the narrative that is in some way relevant to them. This emotional impression would be linked to any attitude formed during the transportation process. At present, the findings concerning NFA and attitude in the narrative message condition hint at this process.

Individual differences in NFC were also associated with receptivity to the narrative appeal. High NFC participants were more likely to be persuaded by the narrative appeal than individuals low in NFC. Participants high in NFC would be motivated to approach a narrative appeal and engage with it to a greater extent than individuals low on this construct. Further, we suggest that once these individuals engage with the narrative, they become transported into the narrative and, consistent with Transportation-Imagery theory (Green & Brock, 2000), might be less likely to counterargue or criticize the messages embedded within the narrative. Alternatively, the narrative appeal might have been perceived as more complex by high NFC individuals compared with low NFC individuals; this is relevant as complexity is a variable that has been associated with enhanced persuasion among high NFC individuals (see Wheeler, Petty, & Bizer, 2005). It is surprising that NFC was negatively associated with attitude change in the rhetorical conditions in Study 3 (in light of the positive evaluation given to the rhetorical message). We suggest that the health topics (cancer and organ donation) used in Studies 2 and 3 may have been threatening to participants, as both relate to death and serious illness. Although participants may have perceived the appeals to be objectively persuasive, this threatening information have led high NFC individuals to become less receptive to (or even resistant to) the rhetorical message on a personal level. It would be useful to examine the above concepts using a less threatening health related topic to further clarify this effect.

We have proposed that NFA and NFC influence engagement with the narrative and influence transportation. This would imply that once an individual is transported into the narrative, this process should elicit attitude change. In the individual studies, we did not find evidence for mediation of the individual differences via transportation. This might be attributable to subtle nature of the effect. That said, to further explore this issue, we combined the Sobel test p-values from the individual studies among narrative condition participants and performed a meta-analysis using the Stouffer Method (see Rosenthal, 1991). Across the three studies, we combined the probabilities that the separate effects of the individual difference constructs on attitude were mediated by transportation. The result of this analysis for NFA was significant (p < .01), suggesting that an effect of NFA is partially mediated by transportation. For NFC, the combined analysis effect yielded a significant result (p < .05), suggesting that an effect of NFC was also partially mediated by transportation. That said, across both individual difference constructs, it is possible that additional factors such as self-schema consistency (Wheeler et al., 2005), message complexity (See, Petty, & Evans, 2009), and enhanced attention (see Haddock et al., 2008) might play a role in understanding the mechanism behind the observed effects. These are issues that should be addressed in future research.

It is worth noting the links between the current research and work that has considered how individual differences relate to the effectiveness of different types of persuasive appeals. Research on matching effects has explored how factors related to the content, structure, and function of attitudes impact responsiveness to appeals that vary in the degree to which they are based in affect or cognition (see e.g., Fabrigar & Petty, 1999; Haddock et al., 2008, Mayer & Tormala, 2010) or the attitude function highlighted by the appeal (e.g., Petty & Wegener, 1998; Snyder & DeBono, 1985). The current research provides novel evidence regarding conditions under which different appeals might elicit enhanced persuasion among different individuals.

Where Next?

The results of the present research have several important implications. First, our findings suggest that narratives may be more effective in persuading some people than others. Despite both types of appeal being rated as equally effective, participants high in NFA and NFC were more persuaded by the narratives. This suggests that the structure of the message, narrative or rhetorical, is another variable that should be taken into account during tailoring messages for particular groups of people (see Mayer & Tormala, 2010).

Second, the results of Studies 2 and 3 suggest that narratives may be particularly useful in the field of health promotion (see Fagerlin, Wang, & Ubel, 2005; Kreuter et al., 2007). The potential of narratives to subvert resistance would be particularly appealing to health communicators, given that information about risky health behaviors is often met with defensiveness or resistance (De Wit, Das, & Vet, 2008). Further, when information about illnesses is complex, it may be better to present this information within a narrative that vividly illustrates the consequences of having an illness and the importance of preventative measures. For example, Green (2006) suggested that narratives may be particularly useful in the promotion of cancer related information, by allowing people to get a better understanding of procedures involved in both initial screening and subsequent treatment. Narratives have also been found to be effective for certain groups (high sensation seekers) when employed for anti-drug public service announcements (Stephenson, 2002, 2003) and promotion of condom use (Dal Cin, Zanna & Elton-Marshall, 2006; Fisher, Fisher, Misovich, Kimble & Malloy, 1996).

The present findings suggest several areas for future research. First, our data suggest that the effectiveness of narrative and rhetorical appeals differs across individuals. One fruitful line of research would be to design narrative and rhetorical appeals tailored to match the individual difference constructs examined in this research. For example, narratives that are complex and include plot twists may be particularly engaging for individuals high in NFC. Similarly, the strong relations found among NFA, transportation, and transportability suggest an important emotional component to narrative persuasion. It may be useful to investigate the influence of constructs such as empathy and affect intensity (Davis, 1994; Larsen & Diener, 1987; Mazzocco et al., 2010). Particularly emotional narratives may be especially engaging for people high in affect intensity and empathy.

It would also be interesting to consider the topic of various persuasive appeals. Research on matching effects in persuasion suggests that strong appeals are more persuasive when their content matches the basis of the recipient's attitude (see Haddock et al., 2008; Maio & Haddock, 2007, 2010; Mayer & Tormala, 2010; Petty & Wegener, 1998). In the context of the present research, some topics are more cognitive than others, and rhetorical appeals may be more appropriate as persuasive messages (e.g., proposed changes to government policy on financial regulations). On the other hand, some issues may be more emotive and may naturally suit the narrative form (e.g., euthanasia). People who are drawn to narrative appeals may be less persuaded about topics that do not naturally lend themselves to the narrative form, whereas people who are more drawn to rhetorical appeals may be less persuaded on issues that do not lend themselves to rhetoric. If this is the case it poses a challenge in identifying which issues are more cognitive and which issues are more affective (see Breckler & Wiggins, 1989). It would then be interesting to try to construct effective persuasive arguments for people who are naturally drawn to cognitive or affective appeals that do not match these tendencies.

Finally, it is also possible that different types of media may be differently effective as persuasive tools for different groups of individuals. For example, Green, Strange, and Brock (2002) argued that written narratives are effective as they allow the individual to use their own imagination to become completely immersed in the narrative. Written narratives that can present a cognitive challenge may be particularly persuasive for high NFC individuals. On the other hand, visual media such as film may prove to be more persuasive for some individuals because in addition to the basic narrative, film has at its disposal many additional tools (see Green et al., 2008). For example, music may provide cues for mood and emotion manipulation, and good visual design may be particularly evocative of a given situation, which may be particularly persuasive for high NFA individuals. A comparison of different types of media may show that some types of media may be particularly effective, or that some types of media are more effective for certain issues, and people, than others.

ACKNOWLEDGEMENTS

This research was supported by an Economic and Social Research Council PhD fellowship to the first author. We thank Pablo Briñol, Greg Maio, Mark Zanna, and two anonymous reviewers for their comments on earlier drafts of this paper.

• 1

  Because the individual difference measures were assessed after the manipulation, we tested whether mean scores on these measures differed across appeal conditions. They did not (all ts <1). Similarly, there were no significant effects in subsequent studies.

• 2

  This narrative was moderately transporting (M = 5.08) and significantly more transporting than the narratives used in Study 1, F (2, 146) = 20.21, p < .01. This is not surprising, given the topic's relevance to our sample of female undergraduates.

• 3

  Across conditions, the number of positive and negative thoughts did not correlate with the individual difference measures, except for a positive correlation between transportability and negative thoughts in the rhetorical condition, r = .39, p = .02.

• 4

  All figures were constructed using values one standard deviation above and below the mean for the individual difference constructs used.

• 5

  Across conditions, the number of positive and negative thoughts did not correlate with the individual difference measures.