From Stories to Action: How Framing,Perspective,and Identifiability in Personal Narratives Influence Vaccination Decisions

Abstract

Purpose

We examined how different narrative aspects related to the COVID-19 pandemic influenced unvaccinated individuals’ willingness to vaccinate (WTV) against a future virus. We tested whether the stories focused on the perspective of the actor (who chose to vaccinate or not) versus the affected (affected by that decision), framing the outcome as death versus survival, and presenting an identified individual versus an unidentified group.

Methods

A total of 1,545 respondents read scenarios depicting individuals’ (actors’) decisions to either vaccinate against COVID-19 or refuse vaccination, alongside the framing of the consequences for the affected individuals: death versus survival. The protagonists were either identified by name and photo or described as a group of unidentified people. Participants reported their emotions, perceived risk from the virus and the vaccine, and their future WTV against a new virus. They also reported their past vaccination decisions.

Results

When the narrative focused on affected individuals, framing outcomes in terms of death increased WTV by heightening the perceived threat of the virus. Conversely, when the focus was on the actor, the lifesaving frame was more effective, especially when the actor was identified. A concrete case of someone vaccinated who saved others evoked positive emotions, boosting WTV.

Limitations

Our hypothetical scenarios and the cross-sectional methodology might limit understanding of the long-term effects.

Conclusions

Scenarios highlighting a person who died increase the perceived threat of the virus and enhance WTV. Conversely, information about a person who was vaccinated and saved others boosts positive emotions and increases WTV.

Implications

Public health campaigns can boost vaccination by sharing stories of vaccinated individuals who saved lives, evoking positive emotions. Highlighting the virus’s dangers can also raise the perceived threat and motivate uptake.

Highlights

• Variations in narratives influence unvaccinated individuals’ willingness to vaccinate.

• Emphasizing the death of those affected evokes greater threat perception of the virus, enhancing vaccine intent.

• Personal stories of vaccinated individuals saving others can boost positive emotions and vaccination willingness.

During the COVID-19 pandemic, people were exposed to emotionally charged stories of suffering: patients dying alone due to visitation bans,^1,2“long-haulers” struggling with persistent symptoms,^3,4 and severe complications such as blood clots or tachycardia in young, healthy individuals. ⁵ Whether depicting specific individuals or broader patient experiences, these stories circulated widely, shaping readers’ perceptions of the virus and their willingness to vaccinate (WTV). ⁶ The present research examines how such narratives can promote future WTV against a new virus and identifies narrative features that may increase vaccination rates by evoking emotions that guide judgment and behavior. Specifically, we focus on 3 key narrative features: framing (lives lost v. lives saved), identifiability (an identified individual v. an unidentified group), and perspective (whether the story is told from the actor’s or the affected individual’s point of view).

Although emotionally charged stories can strongly influence health-related attitudes and behaviors, research on the effectiveness of health-communication narratives in motivating preventive actions, including vaccination, has yielded mixed findings. Their affect appears to depend on message context, narrative features, and audience characteristics. ⁷ A meta-analysis found that narratives promoting detection and prevention behaviors were effective, whereas those encouraging cessation were not. ⁸ In the context of vaccination, narratives are particularly persuasive when presented under uncertainty, and the message presents only one side of the argument. ⁹ Unlike factual messages based on data and logic (e.g., vaccine-safety information ¹⁰ ), narratives engage audiences emotionally,^11,12 thereby enhancing vaccination intentions. For example, a mother’s story describing her son’s suffering from a vaccine-preventable disease she chose not to prevent was far more persuasive than accounts by parents who vaccinated or by medical experts. ¹³ These findings prompt questions about what drives this effect: fear of the disease, identification with a vividly described individual, or the focus on the actor’s (the mother’s) behavior.

Given the widespread use of narratives and personal stories during the COVID-19 pandemic, we empirically test which content features such narratives should include to encourage future WTV in the case of a new virus. We examine 3 characteristics (illustrated in the scenario above) that have been previously shown to influence persuasion in health-related behaviors, although their effects have varied. We ask the following: Should narratives emphasize the perspective of the person who was vaccinated (or refused to be) or that of those who were affected as a result (i.e., those who became infected or remained uninfected)? Should they highlight death or survival? The experience of a single individual or broader trends involving many unidentified people?

Lives Lost versus Lives Saved Framing and Underlying Mechanisms

Prospect Theory ¹⁴ posits that risk preferences and behaviors are influenced by the way choices are framed. There is strong evidence that people tend to be risk seeking when avoiding negative outcomes and losses and more risk averse when seeking gains.^15,16 In the context of health behaviors, framing effects demonstrate that the impact of gain and loss framing depends on perceived risk: gain-framed messages (e.g., emphasizing lives saved) are more effective in promoting behaviors perceived as safe, such as engaging in physical activity.^17,18 Conversely, loss-framed messages (e.g., emphasizing death) tend to be more persuasive in riskier contexts, such as illness detection. ¹⁹ However, research has yielded mixed results, with framing effectiveness often depending on context.^18,20 Specifically, in the domain of COVID-19 vaccinations, both gain and loss frames have shown inconclusive effects. There is some evidence that loss-framed messages may be more effective among hesitant individuals, often depending on the recipient’s level of risk aversion and trust in the source of the message.^21,22

Several studies have demonstrated how the framing of death versus lifesaving operates in vaccination narratives. Death-focused stories enhance message processing and vaccination intentions,^23,24 largely through emotional mechanisms. Negative feelings may prompt avoidance or be linked to a perceived viral threat, motivating protection, whereas positive feelings may promote safety and acceptance. ²⁵ Similarly, discrete emotions—such as distress, blame, and anger in loss contexts or relief and pride in lifesaving ones—may also shape behavior. ²⁶ Overall affect, however, often predicts choices better than discrete emotions do.^27–30 Accordingly, we examine both general affect and discrete emotions as mediators of WTV.

Death framing also increases perceived virus threat, which enhances WTV.^31–33 This process aligns with the Terror Management Health Model, ³⁴ which posits that mortality reminders heighten negative emotions, motivating protective behavior. ³⁵ However, they may also trigger defensiveness when autonomy is perceived as being threatened.^36,37 Thus, we expect unvaccinated individuals to be more influenced by life-loss than by lifesaving messages due to heightened negative emotions and perceived risk of the virus. However, this effect may vary depending on the narrative perspective—which can increase or reduce defensiveness—and on the target’s nature, specifically whether the message features an identified individual or the general case of an unidentified group.

The Identified Victim Effect

People are often more willing to help a single, identified victim than unidentified or “statistical” ones—a robust phenomenon labeled the identifiability effect.^38–42 This effect appears even when identifying details provide no substantive information (e.g., initials or a number) and is limited to single victims.^39,40 Identified individuals evoke stronger emotional responses—such as empathy or distress—than groups do, whose members are processed more abstractly.^43,44 Consequently, single identified victims attract more help than groups do, regardless of whether group members are identified.

Emotional reactions to identified victims, however, can be either positive or negative. When victims are perceived as innocent, identification elicits sympathy and a desire to help, but when they are perceived as responsible, it provokes anger and blame.^45,46 For instance, a man with AIDS may evoke compassion if infected through hemophilia but anger if described as a drug user. ⁴⁷ Thus, identifiability can increase or reduce caring depending on the emotions triggered.

Beyond donations, the identified-victim effect influences other decisions such as punishment,^45,46 medical resource allocation, ⁴⁸ and organ donation. ⁴⁹ In the latter, identifying the recipient of an organ increased support for donation, whereas identifying the deceased donor evoked thoughts of death and reduced willingness. Likewise, identifying a living donor increased willingness by eliciting positive emotions. Hence, the impact of identifiability depends on whether the situation emphasizes death or lifesaving and the emotions each evokes. Consistent with this reasoning, we hypothesized that identifying a protagonist in vaccination narratives would increase WTV only in inspiring, lifesaving contexts—when the identified person vaccinated and saved lives, not when they refused and caused death.

Finally, the effect also varies by perspective—whether the protagonist is the actor or the affected party (e.g., donor v. recipient, deceased v. survivor). For example, Kogut ⁴⁵ demonstrated that identifying a misbehaving student led to harsher punishment from the teacher’s perspective (due to negative emotions) but leniency from the student’s perspective (positive emotions).

Research on health behavior highlights how persuasive narratives differ by perspective, whether told from the actor’s or the affected individual’s point of view. In smoking cessation, counseling emphasizing harm to others (affected) produced higher abstinence rates than counseling focused on self-harm (actor). ⁵⁰ Similarly, road-safety campaigns that centered on victims—such as naming distracted-driving laws after them ⁵¹ or requiring offenders to attend victim impact panels ²² —increased public support and reduced recidivism. During the COVID-19 pandemic, people were more likely to adopt protective behaviors when these were framed as protecting others rather than themselves. ⁵² Accordingly, we expect vaccination messages emphasizing harm to vulnerable others (affected perspective) to motivate unvaccinated individuals more than those stressing personal choice would (actor perspective).

Mortality-salient narratives may interact with perspective to shape persuasion. According to the Terror Management Health Model and reactance theory,^34,36,37 reminders of death can motivate health behavior when they elicit empathy but may backfire when they threaten autonomy. Thus, describing the death of an affected individual may foster shared concern, whereas highlighting an actor’s responsibility for others’ deaths may evoke blame and defensiveness.

Perspective may also interact with identifiability and outcome framing. When the affected person is identified, persuasion should be stronger—especially under death framing—since identifiability increases caring, mostly when the target is perceived as a “victim.” ⁴⁹ For the actor, identifiability may increase WTV primarily under lifesaving framing; namely, an identified actor who vaccinated and saved lives may inspire imitation. However, an identified actor who caused death may provoke distress and reactance.^34,36,37 This pattern parallels findings from organ-donation research, in which people unwilling to sign up as organ donors were more persuaded by stories of a living, identifiable donor who saved lives than by those of a deceased donor. ⁴⁹

The Present Study

We examined how the presentation of COVID-19 cases affects people’s WTV against future deadly viruses. We focused on individuals who had refused the COVID-19 vaccine, as previous studies show that they are significantly less likely to be vaccinated in the event of future pandemics. ² Moreover, in our recent research, we found that unvaccinated individuals are also less likely to vaccinate against the flu or engage in other preventative medical routines (significantly more than vaccinated individuals; authors, under review).

We designed narratives to reflect realistic scenarios of a pandemic. The actor was a young person who decided whether to vaccinate, while the affected individual was an older person who was protected or harmed as a result. Outcome framing varied between lifesaving (e.g., a vaccinated individual preventing others’ deaths) and death (e.g., an unvaccinated individual causing others’ deaths). Identifiability was manipulated by featuring either a single individual (the actor or the affected) presented with personal details (name, age, and a photo) or by describing an unidentified group condition in which participants read about “many young men” or “many elderly people.”

In what follows, references to “emotions” denote 2 operationalizations that we analyze in parallel: discrete emotions (e.g., distress, upset, blame, happiness, relief) and general valenced affect (composites of negative and positive emotions).

Drawing on the literature presented above, we derived the following hypotheses:

In the Affected Condition

• H1. We expect an overall framing effect such that unvaccinated individuals will be more influenced by life-loss messages (e.g., an elderly person who died from COVID-19) than by lifesaving messages (e.g., an elderly person who survived), as death framing increases perceived risk and emotional arousal.

• H2. This effect will be stronger when the deceased is identified, as identified targets evoke more intense emotional reactions that are likely to motivate action.

• H3. Death framing will increase discrete negative emotions and overall negative affect (and decrease positive) and will also increase perceived virus risk; these, in turn, will enhance WTV.

In the Actor Condition

• H4. Adding identifying information will increase WTV only in the inspiring, lifesaving context (i.e., identifying a person who vaccinated and saved lives), which is expected to evoke positive emotions. In contrast, identifying a person who did not vaccinate and caused death is expected to elicit negative emotions and psychological distance (reactance).

• H5. The interaction between framing and identifiability in the actor condition will be mediated by emotions: identifying a single person who vaccinated and saved lives will increase discrete positive emotions and overall positive affect, which in turn will enhance WTV.

Overall

• H6. Taken together, these predictions imply a 3-way interaction among perspective, framing, and identifiability. Specifically, the combination of an identified affected individual under a death frame and an identified actor under a lifesaving frame is expected to produce the highest WTV.

Vaccinated Individuals

• H7. Among vaccinated individuals, the effects of framing, perspective, and identifiability are expected to be similar but weaker overall, given their already high WTV in the future. Due to space limitations, results for vaccinated individuals are reported in the supplementary online materials (SOM).

We tested these predictions in a large-scale experiment in which participants read scenarios involving elderly individuals who either survived or died from COVID-19. In all versions of the scenarios, the survival or death of the elderly was attributed to the vaccination decision of their young grandchildren.

In the actor perspective conditions, we manipulated the identifiability of the grandchildren by either presenting a specific, identifiable individual named David, along with a generic picture of a young man, or referring to them in a general, unidentified manner (as young people or elderly people). In the affected perspective condition, we manipulated the identifiability of the elderly victim or survivor by either depicting a single elderly person named David, along with a generic picture of an old man, or by presenting them in an unidentified generalized fashion (as elderly people). Across both conditions, we tested how outcome framing (death versus survival) and identifiability affected participants’ willingness to receive a vaccine against a future deadly virus and their perception of risks associated with both the vaccine and the virus. Thus, the study design was a 2 (Perspective: Actor versus Affected) × 2 (Framing: Death versus Survival) × 2 (Identifiability: David versus an Unidentified Group) between-subjects experiment.

Methods

Participants and Procedure

The data were collected in February 2024, 9 mo after the federal COVID-19 public health emergency was declared over and respondents were not under any direct threat. Our target population was unvaccinated individuals. To determine the necessary sample size, we conducted a power analysis using G*Power (version 3.1) ⁵³ to detect a small to medium effect size (f = 0.20) with 90% power and α = 0.05 in a 2 × 2 × 2 analysis of variance (ANOVA). The analysis indicated that we required at least 265 unvaccinated participants. Based on data from the National Institute of Health indicating that approximately 17% of US citizens are vaccinated against COVID-19, we estimated an overall sample size of N = 1,545 participants. The sample comprised 54.4% females, with a mean age of 39.91 y (SD = 13.01). All participants resided in the United States and were recruited through Prolific Academic in exchange for a $1 participation fee. In total, 285 participants (58% female, M_age = 39.6, SD = 11.11) were unvaccinated, and 1,258 (53.5% female, M_age = 40.01, SD = 13.40) were vaccinated.

Participants were randomly presented with different versions of the same scenario describing decisions about COVID-19 vaccinations and the outcomes of such decisions. The identifiability and death framing manipulations are described in parentheses.

Participants in the actor perspective condition read about a young people [David, a 28- -old man] whose grandfather is alive [dead] today because the young people [David] received [refused] a COVID-19 vaccine while assisting them during the pandemic. This decision spared many elderly people [David’s grandfather] from infection, ultimately saving their lives [leading to infection and death].

In the affected perspective condition, participants read about elderly people [David, a 78-y-old man] who are [is] alive [dead] today because their [his] grandchildren received [refused] a COVID-19 vaccine while assisting them [him] during the pandemic. This decision spared many elderly people [David] from infection, ultimately saving their lives [leading to infection and death].

Next, all participants read some information about scientists discussing the inevitability of a future virus and anticipating that a vaccine will be made readily available to protect people. After reading this passage, participants rated their WTV (1 = extremely unlikely, 7 = extremely likely) against the future virus. This question was followed by a series of items asking participants to rate the extent to which they felt various emotions: sadness, anger, distress, upset, and blame were averaged (α = 0.91) to create a measure of “negative emotions” and happiness and relief (α = 0.91) to create a measure of “positive emotions.” All answers were given on a 1 = not at all to 7 = very much scale. Finally, participants rated their risk perceptions of the virus and the new vaccine by indicating how much each risk is expected to threaten their lives and the lives of their close ones (2 items for each risk), on a scale from 1 = no chance at all to 7 = very high chance (both α = 0.91). At the end of the study, they reported whether they were vaccinated against COVID-19 and completed basic demographic information.

Results

As found in previous research, people who were vaccinated against COVID-19 are much more willing to vaccinate in the future than those who avoided the vaccine. In our study, the 2 groups significantly differ, t(1, 541) = 34.18, P < 0.001, with vaccinated individuals expressing significantly higher WTV (M = 5.77, SD = 1.51) compared with unvaccinated ones (M = 2.33, SD = 1.63). As noted earlier, since people who are already vaccinated demonstrate very high WTV in the future, here we focus on individuals who were not vaccinated during the pandemic. However, we will discuss the main results observed among vaccinated individuals and refer to the SOM for details.

Analyses of Unvaccinated Participants

Willingness to Vaccinate

We first tested our hypothesis regarding the 2 × 2 × 2 interaction (H6). An ANOVA on WTV revealed the predicted significant 3-way interaction between perspective, identifiability, and the outcome framing of the vaccination decision, F(1, 277) = 4.94, P = 0.027, η² = 0.18 (see Figure 1). To further examine our distinct hypotheses in the actor and affected conditions, we next analyzed each perspective separately. Results of a 2 × 2 ANOVA on WTV in the affected perspective condition revealed only a main effect for framing, F(1, 132) = 4.50, P = 0.036, η² = 0.03. Supporting H1, participants were more WTV in the death framing (M = 2.62, SD = 1.72) than in the lifesaving framing (M = 2.03, SD = 1.45). However, we did not find support for H2, as no significant interaction was found between framing and identifiability. In the actor perspective condition, a significant interaction between framing and identifiability, F(1, 145) = 7.28, P = 0.008, η² = 0.05, supported H4. Specifically, in line with our prediction, simple effect analysis revealed no significant effect for identifiability (P = 0.59) in the death-framing condition. However, in the lifesaving condition, identifiability significantly encouraged greater WTV (M = 3.16, SD = 1.97 and M = 1.95, SD = 1.40, for the identified and unidentified conditions, respectively, P < 0.001).

Figure 1

Graphical illustration of the interaction between target, identifiability, and outcome framing. Error bars represent standard errors.

Mediation Analyses

Next, we explored the possible role of emotions in explaining the interaction effect found in the actor perspective condition on WTV (H5). We conducted a moderated mediation analysis (PROCESS macro for SPSS, model 7, with 95% bias-corrected bootstrap confidence intervals (CIs), based on 5,000 iterations). ⁵⁴ The predictor was framing, identifiability was the moderator, and the mediators were positive and negative emotions. There was a significant interaction between framing and identifiability, showing that identifying a single person who saved a life boosted positive emotions, b = 1.70, SE = 0.56, P = 0.003, 95% CI [0.577, 2.814]. The indirect effect of positive emotions was also significant, b = 0.49, SE = 0.22, 95% CI [0.142, 0.986], indicating that higher positive emotions increased WTV. The effect of framing was no longer significant, P = 0.31, indicating full mediation. The finding supports our hypothesis (H5) and aligns with prior research on organ donations. ⁴⁹ The interaction between framing and identifiability on negative emotions was also significant, b = −1.40, SE = 0.50, 95% CI [−2.374, −0.417], meaning that negative emotions were reduced when the scenario identified a single person who saved a life. However, the indirect effect was not significant, 95% CI [−0.517, 0.015]. Analyzing each emotion separately revealed similar results. Specifically, both happy and relief significantly and fully mediated the interaction between framing and identifiability (index of moderated mediation for happy = 0.445, SE = 0.229, 95% CI [0.079, 0.970], and for relief = 0.609, SE = 0.271, 95% CI [0.117, 1.213]), while every single emotion (except for anger) was significantly predicted by the interaction between framing and identifiability (indicating their combined effect on emotions). However, none of the distinct emotions significantly mediated the interaction between framing and identifiability on WTV.

To examine H3, which predicted that, in the affected condition, emotions would mediate the effect of framing on WTV, we conducted a simple mediation analysis with framing as the predictor, positive and negative emotions as mediators, and WTV as the dependent variable. Supporting H3, death framing decreased positive emotions, b = 1.96, SE = 0.27, 95% CI [1.441, 2.490] and increased negative emotions, b = −1.25, SE = 0.24, 95% CI [−1.720, −0.774.]. Both the indirect effects of positive, b = 0.34, SE = 0.14, 95% CI [0.084, 0.627] and negative emotions, b = −0.44, SE = 0.17, 95% CI [−0.816, −0.161] were significant. Simply put, highlighting death decreased positive emotions and increased negative emotions, which in turn increased WTV. Importantly, the framing effect was no longer significant (P = 0.14), showing full mediation. Moreover, each of the single emotions (positive and negative) significantly replicates the above meditation, except for guilt, which did not reach significance.

Risk Perceptions

As expected, unvaccinated people perceived the risk of the vaccine (M = 4.65, SD = 1.71) as greater than the risk from the virus (M = 3.72, SD = 1.45; t[135] = 4.37, P < 0.001, d = 0.37). Moreover, replicating previous research,^1,16,17 a linear regression analysis revealed that the 2 risks significantly predict WTV, F(2, 282) = 60.84, R² = 0.30, P < 0.001. Perceived risk from the virus predicted greater WTV (t = 8.06, β = −0.38, B = 0.40, P < 0.001), while perceived risk from the vaccine predicted lower WTV (t = −7.53, β = −0.37, B = −0.37, P < 0.001).

We next examined the effect of our manipulation on both risk perceptions. An ANOVA on perceived risk from the virus revealed a significant effect for framing, F(1, 277) = 12.6, P < 0.001, such that death framing (M = 3.97, SD = 1.65) increased the perceived risk compared with lifesaving (M = 3.29, SD = 1.53). This main effect was qualified by a significant 2-way interaction with perspective, F(1, 277) = 4.36, P = 0.038, such that the death frame increased perceived risk from the virus in only the affected perspective condition (P < 0.001) but not the actor perspective condition (P = 0.29). An ANOVA on perceived risk from the vaccine revealed no significant results, suggesting that unvaccinated individuals’ negative perceptions of the vaccine are less susceptible to change than their perceptions of the virus.

Supporting H4b, the results of a mediation analysis in the affected perspective condition with framing as the predictor, perceived risk from the virus as the mediator, and WTV as the dependent variable revealed significant results. The indirect effect of perceived risk from the virus was significant, b = 0.14, SE = −0.46, 95% CI [−0.76, −0.21], while framing was no longer significant 95% CI [−0.64, 0.39], P = 0.63. This shows that, in line with our prediction, death framing increased perceptions of the risk from the virus, in turn increasing WTV.

The analyses of vaccinated participants appear in the SOM. Consistent with H7, the results showed patterns similar to those observed among unvaccinated individuals, although the effects were weaker and, in some cases, did not reach statistical significance.

Discussion

Increasing vaccination rates during a pandemic is critical for achieving herd immunity and saving lives. Yet persuading those who previously refused vaccination remains particularly challenging. We examined how scenarios depicting individuals who chose to vaccinate or refuse vaccination—and the resulting consequences for others—affect future vaccination intentions, especially among previously unvaccinated individuals.

Our study focused on 3 factors known to influence behavior in other domains, such as donation,^39,40 health,^14,18,49,32 and punishment^43,45: the protagonist’s perspective (actor v. affected), identifiability (specific individual v. unidentified groups), and outcome framing (death v. survival).

When the focus was on affected individuals, a loss-of-life framing was more persuasive than a lifesaving one, as it evoked stronger threat perceptions and negative emotions. In contrast, when attention centered on the actor—the person whose decision influenced others—the lifesaving framing was more effective, particularly when the actor was identified. An identified actor who vaccinated and saved lives elicited positive emotions that increased WTV among unvaccinated participants. This pattern aligns with findings in organ-donation research, ⁴⁹ in which identifying a living donor has been shown to enhance positive emotions and increase willingness to donate.

Conversely, stories of unvaccinated individuals causing deaths may have been too distressing, triggering defensive responses such as denial or avoidance^34,37,52,55 and reducing compliance. Emphasizing lifesaving outcomes may instead provide comfort and motivation to act. ⁴⁹ Identified protagonists further amplify these positive reactions by offering a concrete emotional target. ²⁴ Future research is needed to directly examine the psychological mechanisms underlying this pattern, including defense mechanisms such as avoidance or reactance.

Consistent with prior research, perceptions of virus and vaccine risk predicted WTV. ¹⁷ Although death framing heightened perceived virus risk, it did not affect perceived vaccine risk, which remained high among unvaccinated individuals even in the lifesaving condition. This pattern suggests that negative views of the vaccine are more resistant to change than perceptions of the virus are. Hence, emphasizing the virus’s risks may be a more effective strategy to increase WTV. Unsurprisingly, vaccinated individuals expressed an overall strong WTV in the future and were therefore less affected by how the narratives were presented, showing similar yet weaker effects.

Our research contributes to the literature on health communication and vaccination decision making by integrating insights from framing effects, the identified-victim effect, and persuasion strategies. We examined how these factors interact and differ depending on the protagonist’s perspective (actor v. affected). Our finding that an identified vaccinated actor who saves lives increases future vaccination intentions aligns with the identified-victim effect, showing that personal, identifiable stories evoke positive emotions that promote health behavior. Conversely, emphasizing death—especially from the perspective of affected individuals—heightens negative emotions and perceived risk, encouraging action. These results demonstrate how distinct framings trigger different emotional pathways that shape decision making and how narrative elements can be tailored to enhance message effectiveness in emotionally charged health contexts.

Beyond vaccination, our work extends the identified-victim literature by showing that identifiability can both increase and decrease willingness to act, depending on the perceiver’s perspective and emotional response. When identifiability fosters empathy, it promotes prosocial action^26,39,40; when it evokes negativity, it may backfire.^42,43,45 Unlike prior research, which focused on identified victims as the direct target of the decision, we show that identified individuals can also motivate action by illustrating a broader message, either encouraging similar behavior or raising awareness. ⁴⁹

We also offer practical implications for public health campaigns aiming to increase vaccination rates. Messages featuring vaccinated individuals—especially identifiable ones who successfully save lives—can boost positive emotions and future vaccination intentions, particularly among unvaccinated populations. In addition, framing messages around the heightened risks of the virus, such as emphasizing the danger and death associated with a pandemic, can strengthen perceived threat and motivate preventive actions. Public health officials and policy makers should carefully consider narrative content, emphasizing both emotional resonance and concrete, identifiable examples to maximize impact. Personalized stories that highlight positive outcomes of vaccination may be especially effective in driving behavioral change, supplementing traditional informational strategies with emotional engagement to combat vaccine hesitancy.

Limitations and Future Directions

Despite these contributions, some limitations should be noted. First, the use of hypothetical scenarios may not fully capture the complexity of real-world decision making, where emotional and environmental factors interact dynamically. Future research could test these narratives in more ecologically valid settings or field experiments to examine actual vaccination behaviors. Furthermore, our scenarios featured a young actor and an elderly person as the affected individual. Future research could explore how the protagonist’s age affects vaccination decisions and whether people are more sensitive to the death of a younger person.

Second, while our manipulation targeted specific narrative features, other message variables—such as source credibility, message intensity, or cultural relevance—may also influence outcomes and warrant further exploration. Regarding the role of specific emotions, we observed similar patterns when analyzing distinct emotions and general affect. However, our study included only 7 discrete emotions, and other emotions, such as fear, may further illuminate the mechanisms underlying the observed effects. Future research should therefore examine additional emotions more directly.

Finally, the cross-sectional nature of our data precludes understanding long-term effects; subsequent studies should investigate whether narrative effects persist over time and how they interact with ongoing public health efforts.

Supplemental Material

sj-docx-1-mdm-10.1177_0272989X251409811 – Supplemental material for From Stories to Action: How Framing, Perspective, and Identifiability in Personal Narratives Influence Vaccination Decisions

Supplemental material, sj-docx-1-mdm-10.1177_0272989X251409811 for From Stories to Action: How Framing, Perspective, and Identifiability in Personal Narratives Influence Vaccination Decisions by Tehila Kogut, Andrea Pittarello and Paul Slovic in Medical Decision Making