Introduction

In recent years, millions have been compelled to seek refuge in neighboring countries due to crises in their countries of origin. Low- and middle-income countries now host over 40 per cent of the world’s migrant population and 83 per cent of refugees (UNHCR 2022). This has sparked concerns about these resource-constrained countries’ capacity to incorporate displaced individuals, leading to rapidly deteriorating attitudes.

Once anti-immigrant attitudes deteriorate, they are difficult to reverse, posing lasting barriers to immigrant integration and social cohesion. This challenge is acute in middle-income countries, where economic scarcity makes immigration’s economic impacts especially salient. In such contexts, competition for scarce state-subsidized goods and services can fuel conflict with host populations (Dancygier Reference Dancygier2010).

Colombia exemplifies this pattern. Almost seven million Venezuelans have fled their country’s crisis, and Colombia has become their primary destination, receiving over 2.9 million Venezuelan migrants (5 per cent of the population). Public opinion has turned sharply negative: Colombia experienced one of the largest declines worldwide in migrant acceptance, dropping from 61 per cent to 29 per cent during 2016–19. This pattern extends regionally, with Peru declining from 63 per cent to 36 per cent and Ecuador from 61 per cent to 35 per cent (Esipova et al. Reference Esipova, Ray and Pugliese2020).

Existing research highlights perspective-getting interventions – exposing host-nation members to immigrant personal narratives – as a promising path to attitude change. These interventions have proven durable in the United States and Europe (for example, Adida et al. Reference Adida, Lo and Platas2018; Kalla and Broockman Reference Kalla and Broockman2020; Simonovits et al. Reference Simonovits, Kezdi and Kardos2018; Williamson et al. Reference Williamson, Adida, Lo, Platas, Prather and Werfel2021) – where cultural and political concerns typically shape attitudes (Hainmueller and Hiscox Reference Hainmueller and Hiscox2010) – by embedding persuasive information within stories that require cognitive focus, reducing counterarguing (Batson and Ahmad Reference Batson and Ahmad2009; Bilandzic and Busselle Reference Bilandzic, Busselle, Dillard and Shen2013). Yet in resource-constrained host countries, where economic pressures are more salient, these strategies may require adaptation. In these contexts, narratives may need to confront economic concerns directly.

We develop our argument from research showing that both humanitarian and economic concerns shape attitudes towards displaced populations. Humanitarian considerations such as vulnerability and deservingness consistently predict more favorable attitudes (Alrababa’h et al. Reference Alrababa’h, Dillon, Williamson, Hainmueller, Hangartner and Weinstein2021; Bansak et al. Reference Bansak, Hainmueller and Hangartner2016). Economic factors operate through labor market competition – with skilled natives more supportive when they hold an advantage over immigrants (Mayda Reference Mayda2006) – or through fiscal pressures, particularly opposition to low-skilled migration due to welfare state demands (Hainmueller and Hiscox Reference Hainmueller and Hiscox2010; Valentino et al. Reference Valentino, Soroka, Iyengar, Aalberg, Duch, Fraile, Hahn, Hansen, Harell, Helbling, Jackman and Kobayashi2019). Dancygier (Reference Dancygier2010) suggests that conflict results from economic scarcity and competition for state-subsidized goods. This literature implies that narratives portraying humanitarian crises should improve affect towards immigrants. However, when economic impacts are highly salient, as in resource-constrained countries, humanitarian narratives alone may be insufficient to shift support for open immigration. Even when people feel empathy for migrants fleeing difficult circumstances, concerns about economic costs can prevent them from supporting policies that accept more immigrants.

We therefore ask whether narrative interventions need to address economic concerns – rather than only portraying humanitarian hardships – to shift support for open immigration in economically constrained contexts. We argue that narratives portraying migrant hardships may improve natives’ feelings towards immigrants, but to shift support for open immigration, they must address concerns about immigration’s economic costs, particularly regarding pressure on public services and reliance on state subsidies. By countering these concerns, stories reduce counterarguing and change attitudes.

We examine this argument using three preregistered survey experiments conducted in Colombia (n = 12,277) between 2021 and 2023. Colombia’s history as a migrant-sending country may foster warmer feelings towards migrants, but the dynamics we study reflect broader regional patterns where economic concerns dominate public opinion. This single-country study likely generalizes to other middle-income contexts where three conditions converge: (1) large-scale immigration creates salient economic concerns, (2) limited state capacity constrains the ability to offset costs, and (3) host communities experience tension between humanitarian and economic concerns. These conditions are prevalent in many middle-income countries throughout Latin America.

We evaluate perspective-getting interventions with two types of narratives. Building on evidence that humanitarian considerations promote positive attitudes (Alrababa’h et al. Reference Alrababa’h, Dillon, Williamson, Hainmueller, Hangartner and Weinstein2021; Bansak, et al. Reference Bansak, Hainmueller and Hangartner2016), we employ stories portraying migrant hardship related to home-country humanitarian crises, which we call ‘humanitarian narratives’. The second type, ‘economic narratives’, also depict hardship but address concerns about immigration’s economic costs – particularly fiscal burdens and pressure on public services – by emphasizing migrants’ contributions to the host country. We present the narratives in textual (Studies 1 and 2) and video (Study 3) formats, measuring effects immediately after the intervention (all studies) and two weeks later (Study 3).

Given these two narrative types and the theoretical expectations outlined above, we expect that both narratives will increase positive affect by portraying humanitarian hardship, but only economic narratives will also increase support for open immigration policies by directly addressing concerns that would otherwise prompt counterarguing.

Our findings reveal that both content and format are important for narrative effectiveness. Using meta-analysis estimates, we find that economic narratives have positive and statistically significant effects on both positive affect and open immigration policy support. In contrast, humanitarian narratives show positive effects but are only statistically significant for positive affect. While text-based narratives show modest effects, video presentations generate substantially stronger impacts. This format difference is particularly pronounced for humanitarian narratives, which show positive and statistically significant effects only when delivered in video. Economic narratives are effective whether delivered by text or video.

We also find differences across narrative types in the persistence of effects over time. When delivered in video format, both narratives generate immediate positive effects, but only the economic narratives’ effects persist two weeks after the intervention. Additionally, economic narratives specifically ameliorate immigration-related economic concerns, including perceptions of pressure on public services. Together, these findings suggest that economic narratives are more consistently effective across formats and time, possibly because they decrease concerns about immigration’s effects on the economy.

This paper contributes to the literature on attitudes towards migrants and migrant inclusion in several ways. First, it broadens the geographical scope of perspective-getting interventions beyond the high-income focus of most studies to middle-income contexts where economic constraints may alter how interventions work. We present evidence that narratives addressing economic costs are more consistently effective in shaping both affect towards immigrants and support for open immigration policies. This contributes to research on inclusionary interventions, where findings are often mixed: information correcting misperceptions yields null effects (for example, Hopkins et al. Reference Hopkins, Sides and Citrin2019), perspective-taking influences behavior but not attitudes (for example, Adida et al. Reference Adida, Lo and Platas2018), and information targeting immigrant-related concerns boosts affect but not policy support (for example, Williamson Reference Williamson2020). Few interventions increase both affect and policy support; those that do are perspective-getting strategies that address specific concerns by emphasizing migrant contributions (Kalla and Broockman Reference Kalla and Broockman2023), countering perceived security threats (Audette et al. Reference Audette, Horowitz and Michelitch2020), or embedding corrective information (Adida et al. Reference Adida, Lo, Platas, Prather and Williamson2025). Our results suggest this dual effectiveness occurs when narratives preemptively address concerns that would otherwise prompt counterarguing, making interventions more consistently effective across formats, time periods, and outcomes.

Second, our study complements research in developing countries on drivers of anti-immigrant attitudes by assessing interventions to shift attitudes rather than measuring drivers of attitudes, and by finding that interventions addressing immigration’s economic impact can be more consistently effective in reducing anti-immigrant attitudes. Some studies find that the economic impact of immigration does not drive exclusionary attitudes (see for example, Alrababa’h et al. Reference Alrababa’h, Dillon, Williamson, Hainmueller, Hangartner and Weinstein2021). Our findings may seem contradictory, but are actually complementary. In contexts with substantial international aid to support refugee populations, like Jordan receiving Syrian refugees, host communities may perceive immigration’s economic impact as less harmful or even benefit from aid spillovers (see for example, Zhou et al. Reference Zhou, Grossman and Ge2023), ameliorating exclusionary responses. In contexts without such aid, like Colombia, immigration’s economic impact remains salient, and interventions addressing it may be more effective at fostering support for immigration.

Third, our approach advances research by testing key methodological variations. We compare text and video delivery formats, assess persistence over two weeks rather than only immediate effects, and collaborate with migrant advocacy organizations to test stories already circulating in media campaigns, enhancing ecological validity. Fourth, our results suggest that weaker delivery methods (text) require narratives addressing core concerns, while stronger methods (video) can generate immediate effects through emotional engagement alone, though durability requires addressing these concerns.

Our findings have direct implications for advocacy organizations deploying these narratives across Latin America and similar contexts. Organizations can maximize impact by: (1) prioritizing economic narratives when resources are limited, (2) investing in video production when possible, and (3) developing sustained campaigns that build on economic narratives’ durability to achieve cumulative attitude change over time.