Framing Inspirational Content: Narrative Effects on Attributions and Helping

Media coverage often construes stories ofmisfortune as inspirational accounts of individuals overcoming challenges. These reports fail to address the systemic issues that have predisposed these individuals to their current situation, and may have unintended consequences when it comes to the ability to collectively address these failings as a society. The current research examines how audiences are affected by inspirational narrative framings by comparing responses to a narrative that has inspirational coverage of a social challenge to one that includes direct acknowledgement of the larger systemic failings. Participants (N = 495) were randomly assigned to 1) read an inspirational story about a boy saving up to buy a wheelchair for his friend, 2) read a version of the story that emphasized the need for increased disability funding/services, or 3) a no-story control group. Both story conditions raised readers’ willingness to help peoplewith disabilities. Importantly, emphasizing social responsibility shifted readers’ perceptions: readers of the social responsibility story were less likely to believe an individual with a disability was responsible for paying for their medical devices, believed that some collective measures would have higher efficacy, and viewed the situation as less fair. Even though individuals in the social responsibility condition found the story less enjoyable, they were equally transported into it compared to the inspirational version, and were equally likely to want to share the story with others. Our results offer clear guidelines for media practitioners covering individual struggles and systemic issues within society.


Introduction
People often enjoy reading and sharing inspiring stories in which a deserving person receives help. For example, a young boy who sells pumpkins to raise money for a service dog may be considered an inspirational story (Hein, 2018), a girl selling lemonade for her cancer treatments may warm readers' hearts (WNDU, 2018), and a sense of community may be evoked when hardware employees build and donate a walker for a child with a rare muscular condition (Sanchez, 2019). However, these stories do not necessarily address the larger societal issues that these individuals are struggling with, such as a lack of affordable healthcare in the United States or problems with insurance coverage. Such inspirational narratives may prompt prosocial action such as giving to a GoFundMe campaign, but they may also shift attributions of responsibility toward individuals or communities and away from policy-related changes.
A challenge for stories that attempt to raise awareness of broader societal problems is that they may be more likely to evoke negative emotions from readers. Readers may feel sad or angry if they are reminded that the challenges faced by the story characters are widespread (e.g., that many people are not getting the assistive devices that they need). Narratives that are overwhelmingly unpleasant, such as stories of large numbers of people in need, can lead readers to engage in emotion regulation to buffer against negative feelings. However, regulating emotions can lead to a reduction in empathy, and thus a reduced willingness to help (Cameron & Payne, 2011;Shaw, Batson, & Todd, 1994;Small, Loewenstein, & Slovic, 2007). This effect has been termed the 'collapse of compassion' (Slovic, 2007). Therefore, if the goal is to encourage prosocial behavior in a reader, it is important to provide positive, hopeful content within a narrative in order to reduce the likelihood of the reader engaging in mood management processes.
Yet, some research suggests that stories do not need to remain purely focused on the positive in order to be effective. As noted above, eudaimonic media can create positive effects through mixed affect. Additionally, research on restorative narratives suggests that stories of suffering that also focus on recovery can have positive prosocial effects. Restorative narratives are stories of recovery from trauma which share negative experiences while highlighting the strength and meaningful progression of the individual (Tenore, 2015). Fitzgerald, Paravati, Green, Moore, and Qian (2020) found that a restorative narrative about a woman suffering from a rare disease, compared to a negative version of the same story, led to more positive and prosocial outcomes. In particular, the restorative narrative evoked more positive emotions and a greater desire to read and share the story with others as compared to the negative version. The restorative narrative also increased the willingness to help through these emotions.
Given that emotional experiences are an important component of what makes narratives persuasive, and more emotionally evocative narratives are often more transporting (e.g., Appel & Richter, 2010;Nabi & Green, 2015), one would expect that a story with a heartfelt message might persuade readers to change their attitudes in favor of helping others. However, the way in which these stories are framed may be valuable in some situations, but may not be motivating readers toward the most effective ways of creating widespread or lasting change. These emotional and empathic approaches tend to focus reader attention on an individual character. Thus, they may lead readers to wanting to directly help the individual identified by the narrative, rather than change the larger, systemic issues that led to the problem initially. Therefore, it is important to consider the framing of the issue within the narrative.

Framing
News stories can take a variety of approaches in covering current events. Framing is the process by which the mass media define and construct issues by emphasizing certain dimensions to the exclusions of others (Gamson, 1992). For instance, previous research on attributions of responsibility for societal problems has examined episodic versus thematic framing (e.g., Iyengar, 1990Iyengar, , 1991. 'Episodic' framing focuses on the experience of an individual, such as a story of a person who experienced unemployment or lived through a natural disaster (Shen, Ahern, & Baker, 2014). In contrast, 'thematic' news frames are those that focus on more abstract information, with wide coverage of issues or trends over time, such as the causes of unemployment or the scope of the devastation caused by a natural disaster.
Previous studies on framing effects have further found that different frames may influence how people view a given problem (e.g., see Iyengar, 1990Iyengar, , 1991Matthes, 2009). Iyengar (1990) found that when evaluating the issue of poverty, episodic frames engendered a stronger sense of individual responsibility (e.g., readers focused on the choices that a person made that led them into poverty or actions they could take to get out of poverty), while thematic frames engendered a stronger sense of governmental or social responsibility (e.g., readers believed that government or social programs should address poverty).
Our research extends this work by focusing on two different types of episodic frames: frames that maintain a focus on the individual, and frames that highlight the potential responsibility of other organizations (e.g., insurance companies or government). This is different from most previous approaches because it makes simple framing changes within a story, rather than comparing two completely different types of coverage on an issue as a whole. Where other work in framing might compare an individual-focused story to coverage of the wider issue that omits or minimizes individual experiences, we maintain an individual focus while manipulating whether attention is called to the wider issues at play. This approach also fits well with journalism practice, because in the current reporting environment, purely thematically framed stories (e.g., an in-depth report on different approaches to healthcare policy) that focus on a social issue may be relatively rare. Both individual or systemic frames might be incorporated through contextualization; that is, because news reporting is typically driven by current events, individually-focused or systemicallyfocused information may be included as part of the context of a story about a specific event.
It is thus possible that news organizations could retain the human interest element of episodic framing, but nonetheless explicitly draw attention to the broader societal issues that are implicated by the story. Our research examines whether adding these elements changes the type of helping that individuals are willing to provide (individual-focused versus community action), as well as the psychological processes that occur when individuals read the story. Therefore, we explore whether minor changes in focus can change readers' perceptions, with important theoretical and practical implications.

Narrative Engagement
Narratives are also particularly influential when individuals are transported into the narrative, experiencing a state of cognitive and emotional immersion (Green & Brock, 2000;van Laer, Ruyter, Visconti, & Wetzels, 2014). Higher levels of experienced transportation have been associated with greater attitude, belief, and behavior changes, as readers are more likely to like and identify with characters and have stronger emotional experiences (Green & Clark, 2013). Because the main storyline remains the same across the different framings, we expect that transportation will be equivalent across the two conditions. However, directly highlighting broader social issues is somewhat uncommon in 'inspirational' or 'good news' stories, so we examine whether this violation of expectations affects reader engagement.
Similarly, recent research has drawn distinctions between enjoyment and appreciation of stories, where enjoyment is a more purely pleasurable response (i.e., resulting from hedonic media experiences) and appreciation involves more thoughtful feelings of meaningfulness or being moved by the story (i.e., resulting from eudaimonic media experience; Oliver & Bartsch, 2010). We test whether the stories differ on these dimensions. It is plausible that inspirational stories are more enjoy-able due to their focus on the good outcome for the individual, whereas social-focused stories may prompt greater thoughtfulness or appreciation.
Media outlets are also often interested in the reach of their stories: how willing readers are to share those stories or return to the outlet (e.g., the media website). We examine whether the different framings affect these types of future engagement.
Finally, research on helping suggests that individuals are most willing to help when they perceive a person as being in need of help, and when they care about the person's welfare (Batson, 1987(Batson, , 1991. These perceptions prompt empathic concern, which can lead to helping. We test whether the two versions of the story differ in the perceptions of the main character.

Types of Help
We differentiate between individual-focused ways of helping, which provide assistance to a single individual, and collective or social ways of helping, which attempt to change broader systems or policies. For example, donating to a fundraiser or volunteering to help individuals with disabilities would be an individual-focused way of helping, whereas advocating for social change by signing a petition, contributing to a group working to change policies, or contacting government officials to advocate for a particular policy would be collective or social way of helping. In the current study, we focused on the United States healthcare context, where affordable healthcare is less readily available than in some other countries. We suggest that the emphasis of the story may affect the type of helping that individuals are willing to give. Furthermore, the stories might also change individuals' perceptions of the efficacy of that helping (how likely the help is to make a difference). That is, if readers see that individualfocused behaviors (such as donating a wheelchair) are successful, they may perceive these acts as more effective than collective acts that might not reach individuals. Conversely, if readers understand that individual acts are not enough to help everyone in need or even to completely fill one individual's needs, they may be more likely to seek collective action.

Hypotheses
We draw on framing and narrative persuasion research described above to propose the following hypotheses and research questions: H1: Participants in the inspirational condition will attribute more responsibility for helping individuals with disabilities to the individual and less to insurance companies or government programs, compared to participants in the systemic issue condition.
H2: Participants in the inspirational condition will be more likely to help in individual-focused ways such as donating to a particular individual, whereas participants in the systemic issue condition will be more likely to help through civic action such as contacting their representatives or donating to organizations that promote policies helping individuals with disabilities.
H3: Participants in the inspirational condition will view individual-focused methods of helping as being more effective (e.g., will have higher response efficacy), whereas individuals in the systemic issue condition will view civic/social actions as having higher response efficacy. RQ1: Does either story condition increase helping intentions compared to a no-story control condition? RQ2: Will the inspirational or systemic issue frame lead to a) greater perceptions of the importance of the issue of helping individuals with disabilities or b) influence perceptions of spending on disability services (is the amount spent on disability services too much, too little, or about the right amount)?
RQ3: Will individuals in either condition experience a) greater transportation, b) more enjoyment/appreciation, c) more positive and meaningful emotions, or d) greater moral elevation outcomes? RQ4: Will individuals in either condition be more likely to have future engagement with the story (sharing with others, reading similar stories)?
RQ5: Does either story condition affect the perceptions of the fairness of the wheelchair recipient's situation, his likability, concern for his welfare, or his need for help?

Method
We randomly assigned participants to read one of two stories (inspirational [n = 175] or systemic issue [n = 173]) or to a no-story control condition (n = 147). Participants in the control condition completed only measures related to helping, efficacy, and issue importance in addition to demographic and exposure control items.

Participants
We recruited participants living in the United States through Amazon's Mechanical Turk in two rounds of data collection, approximately one week apart. A total of 555 participants completed the study, 290 in the first round and 265 in the second. However, 34 participants failed a set of attention checks, and 26 participants provided blank responses, leaving a final sample of 495 participants (n male = 258, 52.1%, n female = 231, 46.8% n Transgender/Other/Omitted = 5, 1%; n Caucasian = 372, 75.2%, n Asian/Pacific Islander = 55, 11.1%, n Black/African American = 46, 9.3%, n Hispanic/Latino = 31, 6.3%, all other races < 1.6%; M Age = 40.4, SD Age = 12.5). Due to a survey programming error, the responsibility, support for funding, and ways of helping items were only included in the second round of data collection.

Narratives
Participants in both conditions read a story about a high school student, Theo, who saved money for two years to buy an electric wheelchair for his friend, Matt. The story was adapted from actual news coverage of the occasion (Klausner, 2019;Patterson, 2019). In the inspirational version (609 words), the story focused on the boys' friendship and the selflessness of the gift.
In the systemic issue version (655 words), the story highlighted the broader need for insurance coverage or disability programs to provide essential equipment such as wheelchairs, and for greater change in terms of governmental policy. The primary change between the conditions is whether during the course of the story, the individual heroism of the friend is highlighted by writer and the people interviewed, or whether these quotes and observations focus on the need for stronger social services. These points were made in six places in the story. For example, in the inspirational version, the story began: "A high school senior showed how far he was willing to go for his disabled best friend," whereas in the systemic issue version, the sentence read: "A high school senior showed he was willing to step up when insurance companies and government programs fell short for his disabled best friend." Full versions of the stories and questionnaires are available through the Open Science Framework (http://bit.ly/DarkNarrOSF). Table 1 reports the means and standard deviations for all variables by condition.

Measures
Willingness to Help was assessed by asking participants to rate how much they would like to participate in a series of eight items with two individual-focused helping examples, including "Donate to a fundraiser for an electric wheelchair for a specific person" (M = 4.43, SD = 1.67) and six social helping examples, including "Contribute my time and energy to help people with disabilities" (M = 4.71, SD = 1.64, = .89) on a scale from 1 (very unlikely) to 7 (very likely).
Response Efficacy was assessed in a separate series of questions, wherein participants were also asked to rate how likely each of the helping options above would be to make a difference in the life of a person with a disability (1 very unlikely to 7 very likely; M = 4.38, SD = 1.33, = .93).
Attribution of Responsibility utilized three questions adapted from Springer and Harwood (2015). These items asked: "How much responsibility should individuals with disabilities/insurance companies/government programs have for paying for essential equipment such as wheelchairs?" Responses were given on a 5-point Likert scale ranging from 1 (none) to 5 (a lot). Support for Disability Funding asked: "Are we spending too much, too little, or about the right amount on programs to help people with disabilities in the United States?" Responses were given on a three-point scale (M = 2.75, SD = .47).
Ways of Helping asked participants to consider the differences between aid that focuses on an individual in need and aid that creates changes in policy. Participants indicated which type of helping they prefer on a scale ranging from 1 (definitely prefer individual helping) to 5 (definitely prefer working for policy change). They then indicated how effective they perceive each approach to be on a scale from 1 (not at all) to 5 (very much).
Issue Importance had participants indicate their perceived importance of 10 issues related to healthcare and employment, including "Guaranteed paid sick time" and "Universal health insurance" on a scale from 1 (not at all important) to 10 (very important). They also indicated which three of these 10 items they considered to be most important.
Perceptions of Wheelchair Recipient utilized four items asking participants to rate their perception of Matt's need (how great his need was prior to receiving the wheelchair), his welfare, the fairness of his situation, and his likability. Responses were given on scales ranging from 1 (not at all) to 7 (very much). Due to a Qualtrics error, the question about Matt's welfare was answered on a 1-6 scale during the first round of data collection. Therefore, results for this item are reported using a standardized score (Z-score) rather than the 1-7 scale.
Continued Engagement included five items about how much participants would like to read more stories like the one they just read and how likely they would be to share, recommend, and tell others about the story with others on a scale from 1 (not at all) to 7 (very much).
Transportation was measured with Appel, Gnambs, Richter, and  transportation scale shortform. Participants responded to six items on a 7-point scale (1 not at all to 7 very much). An example item is: "I was mentally involved in the narrative while reading it" (M = 4.35, SD = 1.50, = .91).
Moral Elevation Outcomes were assessed by combining two subscales from previous research (Aquino et al., 2011). This measure is frequently used to assess moral elevation by asking about the cognitive outcomes that result from moral elevation. The views of humanity subscale consisted of six items, such as "The actions of most people are admirable." Participants responded on a scale from 1 (strongly disagree) to 5 (strongly agree; M = 4.11, SD = .67, = .87). The desire to be a better person subscale consisted of five items, including "Be a better person." Participants were asked how often they were having those thoughts on a scale from 1 (not at all) to 5 (very much; M = 3.94, SD = 1.04, = .94; total combined = .91).
Recall and Attention checks asked participants to select the main idea of the story they read from four choices (including "I did not read a story" for the control condition).
Demographics and Issue Exposure control items had participants report basic demographic information. Additionally, a set of three items asked about participants' previous exposure to the disability issues discussed throughout the study. These items asked whether participants or someone close to them uses a wheelchair (n yes = 45, 9.1%) or has a disability (n yes = 153, 30.9%). The third item asked participants how much attention they typically pay to issues affecting individuals with disabilities, on a scale of 1 (none) to 7 (a lot; M = 4.35, SD = 1.67).

Results
Participants who failed the narrative attention check (i.e., those in the story conditions who did not respond with "a high school student getting a new wheelchair," those in the control condition who did not respond with "I did not read a story," and those with mainly blank survey responses including the narrative attentive check) were excluded from analyses. We conducted a series of ANOVAs to test our hypotheses, followed by Tukey post-hoc tests to compare between conditions.

Attribution of Responsibility
H1 predicted that participants in the inspirational condition would attribute more responsibility to the individual for helping people with disabilities, and less responsibility to insurance companies or government programs. In line with this prediction, those who read the inspirational story rated individuals with disabilities as being significantly more responsible for paying for essential equipment such as wheelchairs compared to those who read the systemic issue story (post hoc p = .016), and marginally more than those in the no-story control condition (post hoc p = .072), overall F(2, 246) = 3.20, p = .042, 2 p = .02. However, we found no significant differences between conditions for how much responsibility should be attributed to insurance companies or government programs (see Table 1 for all means and standard deviations). Thus, H1 was partially supported.

Ways of Helping
H2 predicted that participants in the inspirational condition would be more likely to help in individual-focused ways, whereas participants in the systemic issue condition would be more likely to help through civic engage-ment such as contacting representatives or donating to disability organizations. There were no significant differences between story conditions for either way of helping. Therefore, H2 was not supported. However, reading either story led to a greater likelihood to help in general compared to the no-story control condition, both in individual-focused ways, F(2, 492) = 4.16, p = .016, 2 p = .02, and civic ways, F(2, 492) = 3.39, p = .035, 2 p = .01. A Tukey post hoc analysis revealed that, specifically, those who read the systemic issue story were significantly more likely to help in both individual-focused and civic ways compared to the control (post hoc p = .02 individual; p = .03 civic). Those who read the inspirational story were also marginally more likely to help than those in the control, but only in individual-focused ways (post hoc p = .05).

Response Efficacy
H3 predicted differences between conditions in perceived response efficacy, where participants in the inspirational condition would view individual-focused methods of helping as being more effective and those in the systemic issue condition would view civic actions as being more effective. There were no significant differences between conditions for ratings of the overall effectiveness of the two approaches: individual solutions were rated similarly effective regardless of the three conditions, F(2, 246) = 2.17, p = .116. Similarly, policy advocacy was rated as being similarly effective across all three conditions, F(2, 246) = 1.12, p = .327. However, we also examined efficacy ratings for specific ways of helping and found that, compared to the no-story control, those in the systemic issue condition perceived greater efficacy for signing a petition for insurance companies to cover the cost of electric wheelchairs,

Issue Importance
RQ1 asked whether the inspirational or systemic issue frame would influence perceptions of the importance of health policies which helped individuals with disabilities. However, results revealed no significant differences between any of the three conditions on how important participants rated the requirement for health insurance policies to cover essential medical equipment, F(2, 490) = .17, p = .844, or importance for universal health insurance, F(2, 490) = .16, p = .848. RQ1 also asked whether the story frame would influence perceptions of spending on disability services. However, results revealed no significant differences between conditions on perceptions of spending, F(2, 245) = .91, p = .40.
We examined additional variables related to the story conditions specifically.

Story Variables
RQ3 asked whether story conditions would differ on a series of narrative engagement and outcome variables, including transportation (RQ3a), enjoyment/ appreciation (RQ3b), emotional responses (RQ3c), and moral elevation outcomes (RQ3d).
Third, results revealed a significant difference for how enjoyable participants rated their stories, F(1, 346) = 9.61, p = .002, 2 p = .03, but not for how much participants appreciated them, F(1, 346) = .49, p = .483. Thus, those who read the inspirational story rated their story as being more enjoyable compared to those who read and rated the systemic issues story (see Table 1).

Continued Engagement
RQ4 asked whether story conditions would differ on intentions to share the story with others or read similar stories. However, there were no significant differences in the future story engagement activities including reading similar stories or sharing the story with others, F(1, 346) = .51, p = .478.

Supplementary Analyses of Political Orientation
Because conservatives and liberals in the United States may have different views on healthcare issues, we conducted supplementary analyses in which we tested our hypotheses and research questions while control-ling for political orientation (measured by the question: "How would you describe your political orientation?" with the response options of 1 = Very conservative, 2 = Conservative, 3 = Moderate, 4 = Liberal, and 5 = Very liberal). The results did not change when controlling for this variable and we therefore do not detail them here. A table comparing these two sets of analyses is available in our repository on the Open Science Framework (http://bit.ly/DarkNarrOSF).

Discussion
Reading stories about an individual being helped raised readers' willingness to help. Importantly, emphasizing information about social responsibility shifted readers' perceptions: Compared to the inspirational condition, readers of the social responsibility story were less likely to believe an individual with a disability was responsible for paying for their medical devices, believed that some collective actions would have higher efficacy, and viewed the situation as less fair. Compared to the no-story control, the inspirational story increased individual attributions of responsibility and only marginally increased willingness to help, while the social responsibility version significantly increased willingness to help and also increased the perceived efficacy of some specific helping items (signing petitions to advocate for better insurance coverage and government funding). Thus, both types of stories appear to provide some motivation for prosocial action compared to control, but the inspirational story also carried the potential downside of increasing the perception that individuals with disabilities should be responsible for paying for their essential medical devices themselves. This speaks to Landmann's (2021) discussion that there is a light and a dark side to eudaimonic emotions; such complex emotional experiences can motivate audiences toward prosocial behavior, but they can also manipulate perceptions of responsibility around the issues discussed. Although our study focused only on immediate effects, this heightened perception of individual responsibility may lead to decreased support for systemic action in the long term, a potential unintended negative consequence of reading 'feel good' or inspirational stories, compared to not reading this type of material.
These findings have theoretical importance because they suggest ways to shift reader focus toward collective action even in the context of an episodic news report or story; telling the story of an individual does not necessarily have to promote individual responsibility attributions. Additionally, the focus on collective or systemic action has been relatively neglected in studies of prosocial media effects, which tend to focus on individual action. We hope that our approach can serve as a springboard to broader consideration of these issues within the field.
Even though individuals in the social responsibility condition found the story less enjoyable, they were equally transported into it compared to the inspirational version, and were equally likely to want to share the story with others. Therefore, focusing on social responsibility information in stories of individual helping appears to have benefits for encouraging collectively-oriented prosocial behavior, and does not have significant costs in terms of story engagement or promotion of individualfocused helping.
However, the difference in enjoyment may explain why such 'inspirational' coverage of these situations are so prevalent, as journalists may wish to invoke more positive emotions in their readers. There were otherwise no differences in reader response. Journalists might consider alternative ways to evoke positive emotions in stories that focus on systemic issues, perhaps by mentioning organizations that are making progress in addressing these issues on a larger scale.
One challenge with creating socially-focused helping is that it is often less clear what behaviors individuals should take. An individually-focused story may have a clear path to helping: donate to that individual. However, the steps that individuals would take to change insurance company policies, support government programs, and so on are often much less obvious. A direction for future research may be to test ways to provide effective 'calls to action' for broader social change.

Limitations
The current study used only one set of stories; future studies should test whether these effects hold with other stories and topics. Additionally, some of the differences between the conditions came from characters themselves stating the need for broader social change; in real interviews or news stories, individuals may not spontaneously raise these issues.

Conclusion
These results offer clear guidelines for media practitioners who are balancing audience engagement while covering societal issues. It is our hope that this work can encourage reporting that highlights the larger systemic failures within our society and in coverage of individual struggles when appropriate. By focusing on isolated instances of 'heartwarming' assistance from friends and neighbors, we fail to ask questions about why these individuals are suffering in the first place. While it may provide hope to see a good friend save up to gift an electric wheelchair, these stories ignore the greater issues in society and do little to motivate a response toward greater change for the many other, anonymous individuals who are suffering. Our results show that it is possible to engage with the larger systemic issues in a way that still highlights individual perseverance while offering readers an outlet to engage in helping behaviors (collectively and individually) and also increasing their perceived efficacy of those behaviors. In the words of Pope Francis (2020): "It is an act of charity to assist someone suffering, but it is also an act of charity, even if we do not know that person, to work to change the social conditions that caused his or her suffering." Kaitlin Fitzgerald (MA, University at Buffalo, SUNY) is a PhD Candidate at the University at Buffalo-State University of New York. Her research examines the cognitive and emotional processes involved in narrative influence and engagement, especially within the context of entertainment media. Her research has a particular focus on how certain narrative experiences can foster long-term, prosocial outcomes such as empathy and helping, moral sensitivity, and well-being.
Elaine Paravati received her PhD in Social-Personality Psychology from the University at Buffalo-State University of New York and is a Visiting Assistant Professor in the Department of Psychology at Hamilton College. Her research interests include the need to belong, restorative narratives, gender and personality differences in social need fulfillment, and the influence of technology on social interaction. She is passionate about writing, teaching, reading, and cats.