Is narrative entertainment simply a form of recreation, or does it have meaningful effects on public opinion? In this blogpost, Bardia Rahmani, Beatrice Montano, Dylan Groves, and Don Green investigate how edutainment – educational or persuasive messages embedded in narrative entertainment – can shape audiences’ attitudes, beliefs and behaviours. A meta-analysis of 77 experimental studies shows that the effects are real, nudging people in the intended direction.
Narrative entertainment makes up a massive share not just of what people consume as media, but how they spend their free time. Recognizing the staggering global reach of feature films, television series, and radio dramas, policymakers have increasingly sought to design entertainment-education interventions (“edutainment”) that use stories to communicate public-interest messages to mass audiences.
Examples abound. After the wildly popular Mexican telenovela Ven Conmigo introduced a storyline in which workers enrolled in a government literacy program, real-life enrollments reportedly increased dramatically. The Nigerian series MTV Shuga, distributed through television, radio, and online streaming platforms, has reached tens of millions of young people across Africa with messages about HIV prevention. Films like Don’t Look Up have been viewed by hundreds of millions globally and, at least in principle, convey socially important ideas about climate change.
Yet despite the scale of these efforts – and growing investment in them – we still lack a clear sense of whether narrative entertainment is actually persuasive. We address this question by conducting the largest meta-analysis to date of experimental studies estimating the effects of narrative entertainment on audiences’ attitudes, beliefs, intentions, and behaviours. The 77 studies share a common design: participants are randomly exposed to narrative entertainment or to no message, an irrelevant message, or a non-narrative message. Beyond that shared approach, the studies vary widely. They span 12 countries across four continents, test the effects of television, film, and radio content on topics ranging from prejudice reduction to health to gender equality, and take place in settings from controlled laboratories to naturalistic field environments. By estimating average effects across studies, we determine whether narrative entertainment can serve as an effective vehicle for pro-social messages.
What we found: attitudes and beliefs
Across countries, issue areas, and message types, a consistent pattern emerges: stories can change what people feel, what they believe, and what they do.
We begin by estimating effects on attitudes – how people evaluate or feel toward certain people, things, or ideas. Attitudinal change is by far the most common outcome studied in the literature.
Narrative entertainment is often designed to reduce stigma toward particular social or ethnic groups. For example, the Canadian drama Little Mosque on the Prairie seeks to promote tolerance toward Muslims living in Canada. Other interventions tell stories about immigrants and migrant workers to increase public acceptance of these individuals, portray the experiences of people living with HIV to reduce stigma, or depict characters living with mental illness in ways that challenge negative stereotypes.
Figure 1: The geographic scope of the meta-analysis
Source: Rahmani et al, 2025
Entertainment might also shape policy attitudes. Researchers have examined, for instance, how dystopian fiction such as The Hunger Games film series might influence political views, including acceptance of radical political action. In another project, we designed a Swahili-language radio drama, Mikoko Yetu, to increase political support for environmental protection among villagers in rural Tanzania.
Looking across 161 experimental comparisons drawn from the 42 studies that explore attitudinal changes, we find sizable average effects of narrative entertainment on attitudes. Although the interventions and targeted attitudes differ widely, the average standardized effect is about 0.3, which is typically interpreted as a moderate effect.
Figure 2: Summary of treatment effects comparing narrative message vs no/placebo message
Source: Rahmani et al, 2025
The second outcome is beliefs, or what people think is objectively true about the world. Practitioners have used narrative entertainment to inform rural Indians about their rights under anti-poverty programs, improve health knowledge in Sub-Saharan Africa, and increase awareness of climate change in the United States. Entertainment can also unintentionally shape beliefs as a byproduct of storytelling. Researchers have asked, for example, whether the television show The X-Files increases endorsement of conspiracy theories.
Compared to attitudes, which are often relatively stable, beliefs should respond more readily to new information. That is broadly what we find. On average, narrative entertainment increases acceptance of the beliefs targeted by the intervention, with estimated effects that are slightly larger than those for attitudes.
For practitioners, however, the ultimate goal is not just influencing what people think, but changing what they actually do. We therefore examine behavioural intentions — what people say they plan to do in the future – such as getting vaccinated, obtaining a cancer screening, or volunteering for a cause. Some studies go further and measure real-world behaviour.
Here, too, the pattern is consistent. Narrative entertainment makes people more likely to say they will act in ways consistent with a story’s message. Crucially, it also leads to significant changes in actual behaviour on average. Nigerian teenagers who watched the television series MTV Shuga were twice as likely to get tested for HIV as those who did not watch. A radio drama designed to shift political culture in post-genocide Rwanda made listeners more willing to express dissent in community meetings rather than go along with the group. Other narrative interventions have successfully reduced unhealthy behaviours in the United States and Europe, including smoking, alcohol consumption, and unprotected sex.
Do these effects last?
Studies differ in how long they wait between exposing respondents to a narrative message and measuring outcomes. About half measure outcomes shortly after exposure, while the rest wait weeks, months, or even a year. Some studies also use repeated measurements to track how respondents change over time. This variation in timing allows us to examine the persistence of narrative persuasion.
Effects tend to be somewhat smaller weeks or months after exposure than immediately afterward, but they often remain positive and statistically significant, and the difference over time is not itself statistically significant. In some cases, effects persist for as long as a year. This finding suggests that messages embedded in stories can remain influential well beyond the moment of viewing or listening.
One final point is worth emphasising. Narrative entertainment does not produce wholesale transformations in what people think or do. Across outcomes, the effects are real but moderate. That the literature does not report implausibly large impacts is reassuring. A single film, television show, or radio serial will not reshape society on its own. But like other forms of communication, narrative entertainment reliably moves people in the intended direction. And when outcomes involve things like HIV and cancer testing, reductions in prejudice toward out-groups, or quitting smoking, even moderate changes can meaningfully affect people’s lives and livelihoods. Perhaps continuous and repeated exposure to similar messages leads to even greater attitude and behaviour change – a possibility future research should explore.
What we did not find
We also examined studies that conduct head-to-head tests of narrative and non-narrative messages. Instead of comparing people exposed to a story with those who received no information (or an unrelated message), these studies compare treated people to those who receive the exact same information in a straightforward, non-narrative format.
These experiments address a different question from the ones above. Rather than asking whether narrative entertainment can persuade, they ask whether storytelling is more persuasive than simply presenting the same information directly. In other words, is there something uniquely powerful about narrative as a form of communication?
Interestingly, we find no conclusive evidence that narratives are more persuasive than non-narrative messages. While narratives appear slightly more persuasive, the difference is neither significant nor substantial.
Taken at face value, these non-results tell us something about the nature of narrative persuasion. A large body of theoretical work maintains that narrative entertainment has a special ability to persuade. For instance, one school of thought holds that, while people normally counter-argue against information that contradicts their existing beliefs, entertainment places them in a mental state in which they evaluate messages less critically. Another suggests that stories are especially persuasive because viewers and listeners tend to model their behaviours and attitudes off of that of appealing protagonists (or seek to do the opposite of antagonists). If these theories were true, we would probably expect narrative messages to have bigger effects than equivalent non-narrative ones. The fact that we don’t might suggest that narrative messages persuade people in just the same way that other media messages do: by introducing them to new information or making arguments that they find compelling.
Yet we should not necessarily conclude that narrative entertainment is without its advantages. One possibility is that narratives matter not because they are particularly persuasive, but because they are especially interesting, which makes people more willing to actually consume them in the first place. In other words, putting information in a narrative format might increase the likelihood that people actually sit down to watch or listen to it. The ability to generate interest may be especially important in a media-saturated world where so much competes for people’s attention.
Here, we run up against the bounds of existing evidence. Most studies measure the effects of narrative messages conditional on exposure. Much less is known about whether people are more likely to choose narrative content over non-narrative content conveying the same information. Understanding what media content people choose to consume, rather than merely the effects of such content, is an important frontier for future research.
Implications for policymakers
One of the more striking findings in our analysis is that narrative entertainment found ‘in the wild’ – commercially produced movies, television shows, and serials – is no more persuasive than edutainment interventions deliberately designed by policymakers to change attitudes and behaviours. We find essentially no difference in average persuasive effects between these two kinds of media. This finding suggests that policymakers and practitioners can design narrative interventions that are just as compelling as the entertainment people encounter in everyday life.
As the rest of our results show, these interventions generally work. Across a wide range of outcomes and domains, edutainment interventions accomplish what they set out to do on average. They shift attitudes, beliefs, and behaviours in the intended direction. Because these interventions are delivered through mass media formats, they are also inherently scalable and cost-effective. And in a crowded media environment in which people may ignore straightforward policy messaging or public service announcements, narrative entertainment may have a special ability to grab and hold the attention of large and diverse audiences. Policymakers would be wise to deliver pro-social messages through the kinds of media people already choose to consume.
You can read the full article here.
Bardia Rahmani is an Academy Scholar at the Harvard Academy for International and Area Studies at Harvard University
Beatrice Montano is a PhD Candidate in the Department of Political Science at Columbia University
Dylan W. Groves is Assistant Professor of Political Science (Government & Law) at Lafayette College
Donald P. Green is the J.W. Burgess Professor of Political Science at Columbia University