Beyond Changing Minds: Raising the Issue Importance of Expanding Legal Immigration

Abstract

How can public opinion change in a pro-immigration direction? Recent studies suggest that those who support immigration care less about it than those who oppose it, which may explain why lawmakers do not enact pro-immigration reforms even when voters are pro-immigration. To see if the personal issue importance of immigration can be changed, I conducted a probability-based, nationally representative US survey experiment (N = 3,450) exposing respondents to verifiable arguments about the broad national benefits of expanding legal immigration and the costs of not doing so. Using new measures of issue importance, my descriptive results show that only one-fifth of voters who prioritize the issue have a pro-immigration preference. Furthermore, while anti-immigration respondents prioritize policies regarding law enforcement and (reducing) future immigration, pro-immigration respondents prioritize (helping) immigrants already here. The experimental results confirm that the provided arguments raised immigration’s importance among pro-immigration voters but did not backfire by mobilizing anti-immigration voters. Contrary to expectations, the arguments increased pro-immigration policy preferences, but did not change voters’ subissue priorities within immigration or their willingness to sign a petition. Overall, the treatment was effective beyond changing minds by shifting stated issue positions and priorities in a pro-immigration direction. It can thus be used in a nontargeted information campaign to promote pro-immigration reforms.

Why do US lawmakers not pass pro-immigration reforms even when the public seems to support such reforms?¹ Recent studies indicate that one possible behavioral explanation is related to the much lower perception of immigration as a personally important issue (or “priority”²) among those voters who support it compared to those who oppose it (e.g., Kustov 2023a). As a result of this “immigration issue importance asymmetry,” politicians rarely run on a pro-immigration platform (Levy, Wright, and Citrin 2016) and the few pro-immigration advocates there are appear to prioritize aiding existing immigrant constituents over improving the system for future immigrants (e.g., Tichenor 2002). Unless pro-immigration advocates figure out a reliable way to raise immigration issue importance among sympathetic respondents only, they will always face this systematic disadvantage compared to their counterparts.

So, can voters’ issue priorities be changed systematically in a pro-immigration direction? To address this question, I field a large-scale (N = 3,450), nationally representative, and preregistered survey experiment of US adults. This study (1) provides better measures of immigration issue importance and then (2) tests whether exposing pro-immigration respondents to verifiable arguments about the substantial national benefits of expanding immigration to the US and the costs of retaining existing restrictions can increase their perceived importance of the issue. The experiment focuses especially on those respondents who are already pro-immigration, since these are the people most likely to be responsive to pro-immigration information and bring about related policy changes (see, e.g., Ivarsflaten and Sniderman 2022).

As an important descriptive contribution, I first corroborate immigration issue importance asymmetry and the widespread perception that immigration is important compared to other issues using a new measure of “issue-public” membership (Ryan and Ehlinger 2023). I then, for the first time, identify the differences between (predominantly liberal and Democratic) pro-immigration respondents and (predominantly conservative and Republican) anti-immigration respondents in the perceived importance of immigration subissues (from border security and treatment of immigrants to admission selection and numbers). These new results indicate that the asymmetry may be even greater than previously assumed: only one-fifth of voters who prioritize the issue have a pro-immigration preference. Furthermore, I show that, while anti-immigration respondents prioritize policies regarding law enforcement and (reducing) future immigration, pro-immigration respondents prioritize (helping) immigrants already here.

The experimental results confirm that short, verifiable narratives with new, relevant information about the broad benefits of increasing legal immigration can motivate pro-immigration respondents by increasing the personal importance of the issue to them. The results also corroborate that such narratives do not backfire by increasing immigration’s personal issue importance among anti-immigration respondents. Contrary to expectations, however, information did not shift people’s issue priorities within immigration or their willingness to sign a petition. At the same time, the information also unexpectedly increased pro-immigration preferences among all respondents. Overall, the proposed information treatment was effective above and beyond changing minds by reducing immigration issue importance asymmetry and shifting the public’s stated issue positions and priorities in a pro-immigration direction. It can thus be used in a general, nontargeted information campaign (Green et al. 2022) to promote the political priority of pro-immigration reforms to benefit both citizens and noncitizens alike. However, the extent to which such shifts in stated issue importance translate into policy change remains uncertain as informed respondents did not show a significantly increased willingness to sign a petition, suggesting a need for further research on the relationship between issue importance and political action.

The study draws on and contributes to several strands of immigration politics, political behavior, and persuasion literatures. Although there has been a lot of research on how populist politicians raise the issue importance of immigration issues among anti-immigration voters to improve their electoral fortunes (e.g., Hutter and Kriesi 2021), it is still unclear whether it is possible to increase the importance of immigration exclusively among the pro-immigration populations. While information interventions intended to flip people’s immigration issue positions have rarely been successful, the project assesses for the first time whether relevant information can also raise the importance of immigration and its expansion among the already sympathetic population. Given my results, this approach can also be fruitfully applied to other nonimmigration issues across countries where there is an evident disconnect between public preferences and policy outcomes, from gun control and abortion to climate change mitigation.

This work also contributes to a better understanding and measurement of personal issue importance and salience, “preference intensity,” and “attitude strength” beyond immigration (Cavaillé, Chen, and Van Der Straeten, forthcoming; Dennison 2019; Hill 2022; Howe and Krosnick 2017; Miller, Krosnick, and Fabrigar 2017; Ryan and Ehlinger 2023). It informs the literature on information experiments and their effectiveness, which so far has been primarily concerned with changing voters’ issue preferences rather than issue importance (Coppock 2022; Haaland, Roth, and Wohlfart 2023). By focusing on how weak attitudes can “crystallize” and become more important to voters (and potentially more susceptible to priming) in response to new information, my study also informs the long-standing debate about the nature of campaign effects (Lenz 2009; Tesler 2015).

Reducing Immigration Issue Importance Asymmetry

Background: Relatively Stable Preferences and Relatively Volatile Issue Importance

One of the most robust findings in the public opinion literature is that the majority of US voters are more sympathetic toward and prefer particular—generally skilled and culturally similar—immigrants from certain countries (Hainmueller and Hopkins 2014; Lee and Fiske 2006). Voters’ immigration preferences and the very meaning of “immigration” can also significantly vary across contexts depending on the type of immigrant who is salient in respondents’ minds (Hopkins 2010). As a result, it is possible to imagine how people’s responses to immigration questions can vary depending on the perceived characteristics of immigrants in a survey context and thus be changed via priming of those characteristics. However, persuasion efforts to change people’s immigration preferences can arguably only be considered successful to the extent that they are able to durably change people’s opinions toward the exact same policies or groups across contexts.

In this respect, despite recent increases in aggregate positivity toward immigration, there has been growing longitudinal evidence that immigration policy preferences are generally stable at the individual level, especially in the long term (Goldstein and Peters 2014; Hopkins, Sides, and Citrin 2019; Kustov, Laaker, and Reller 2021; Lancaster 2022; Maxwell 2019). After all, these preferences form early in life and reflect deep-seated psychological predispositions such as openness to experience and ethnocentrism (Dražanová 2022; Kinder and Kam 2010). Preference changes, especially regarding particular policies, can happen in the short term in response to new information or an external shock, but they tend to persist only among some—especially young—voters (Laaker 2024). To the extent that durable immigration preference changes exist, they thus happen gradually and mostly due to generational replacement, not persuasion (McLaren and Paterson 2020).

Voters have also been shown to be rather ignorant about immigration issues (Hopkins, Sides, and Citrin 2019; Lutz and Bitschnau 2023). But while various attempts to durably change people’s minds on immigration by correcting misperceptions or providing new information have generally not been successful,³ there is ample evidence that people routinely change their perceptions about the “national salience” of various issues due to political messaging and events (Dennison and Geddes 2019). In other words, immigration preferences are relatively stable but their perceived salience and importance to voters is relatively volatile.

Although persuasion or preference change brought about by providing information is certainly possible even for racialized issues (Kustov and Landgrave 2023; Santiago, Kustov, and Valenzuela 2023), it appears to be much harder for issues that voters perceive as important (Howe and Krosnick 2017; Vidigal and Jerit 2022), such as immigration in many countries today. At the same time, while being evidently more volatile than preferences, personal issue importance or the extent to which one actually cares about issues is still supposed to be a much more stable individual characteristic than merely thinking about issues or recognizing them as “salient” at a particular point of time in response to news or campaign priming (Moniz and Wlezien 2021; Ryan and Ehlinger 2023).

The literature on issue importance is thus arguably central to the understanding of public opinion and its role in politics. When individuals attach personal importance to a certain policy issue, it means that they care about it relatively more than about other issues, as evidenced by their greater cognitive and behavioral engagement with that issue. They tend to think more frequently and deeply about it, as well as seek, gather, and spread information about it (Boninger et al. 1995). As a result, they are also more likely to cast their vote based on the issue (Dennison 2019).

Crucially, however, issue importance is not just about voting. People who care about a certain issue are also more likely to engage in other costly actions in favor of their cause (Miller, Krosnick, and Fabrigar 2017). The people who care particularly intensely about issues constitute potentially influential “issue publics” even when they are not numerous (Ryan and Ehlinger 2023). As recently shown by Hill (2022), politicians have an electoral incentive to side with such intense minorities over passive majorities, assuming the issue importance asymmetry is sufficiently high. The failure of gun control reform in the US, despite very high public support (but low issue importance among supporters), is perhaps the most prominent example of such a dynamic.

Unfortunately, the research on immigration attitudes has been developing separately from this discussion (but see Dunaway, Branton, and Abrajano 2010). This is disconcerting given that people for whom a particular immigration issue is very important are likely to place great weight on it when deciding how to vote or are likely to express their views otherwise, which can be just as consequential. Issue importance can relate to such costly actions as volunteering for immigration organizations, donating money, attending meetings and protests, contacting public officials, advocating for the issue, or even running for office (Moniz and Wlezien 2021).

Motivation: Revealing Immigration Issue Importance Asymmetry

While some scholars have recently tried to bring issue importance to the center of immigration research (Dennison and Geddes 2019; Hatton 2021; Paul and Fitzgerald 2021), the link between immigration preferences and their perceived importance has not been explored much due to the lack of relevant data. One exception is the recent paper by Kustov (2023a), who has identified all such data from the publicly available representative surveys to provide an empirical assessment of this relationship. Overall, this research finds that those who oppose immigration are more likely to consider it as both personally and nationally important than those who support it, which holds across different years, countries, and measures.

An important implication of this revealed “immigration issue importance asymmetry” is that even though voters’ anti-immigration preferences may be in decline, political events about immigration should be more likely to engage those who oppose it. In other words, pro-immigration advocates are always at a systematic disadvantage compared to their opponents when they want to draw national attention to the issue. As of now, however, it is still unclear what causes this issue importance asymmetry, whether it is about immigration in general or some specific issue like undocumented immigration in particular, and whether the observed gap between pro-immigration and anti-immigration voters can be reduced.

Goal 1: Descriptively Explore the Possible Issue Importance Asymmetries within Immigration

Immigration is a complex policy domain with a variety of distinct “subissues” (for a review of existing classifications, see Helbling et al. 2017). For instance, Tichenor (2002) famously distinguishes between “immigration admission” and “immigrant rights” policies that have had distinct and changing political coalitions throughout US history. More broadly, in addition to major admission and selection policies that impact the number of future immigrants (or “flows”), the US and other governments regulate naturalization and other rights that impact existing immigrants (or “stocks”), as well as unauthorized immigration and border security (or “enforcement”).

While the results of this project should be relevant regardless of one’s beliefs about the objective primacy of immigration and its subissues, there is an extensive literature documenting the importance of immigration over other government policies (Clemens 2011) and the practical importance of addressing flows over stocks (Ruhs 2013) or enforcement (Massey, Durand, and Pren 2016). After all, most immigration policies arguably regulate the incorporation of noncitizens into the country and thus are ultimately contingent on the allowed number of legal immigrants.

Similarly, although most people can be consistently classified as either pro-immigration or anti-immigration (Kustov, Laaker, and Reller 2021; Ruedin 2020), many also make meaningful, nuanced distinctions to describe their views. This includes differentiating between individual immigrants and various policies beyond the common “legal/illegal” dichotomy (Helbling, Maxwell, and Traunmüller 2024; Margalit and Solodoch 2022). Thus, they may attach different priorities to these policy views too.

According to recent evidence, most US adults exhibit a “stock premium” due to their stronger sense of moral obligation toward people who are already residing in the country (Margalit and Solodoch 2022). As a result, pro-immigration respondents may be expected to be relatively more supportive of stocks (compared to flows) while anti-immigration respondents may be expected to be relatively less opposed to stocks (compared to flows). In other words, in addition to the general asymmetry observed earlier, pro-immigration respondents may also be expected to prioritize supporting existing immigrants over future immigration while anti-immigration respondents may be expected to prioritize opposing future immigration over existing immigrants.⁴

Goal 2: Experimentally Explore the Malleability of Immigration Issue Importance Asymmetry

While the exploration of how individuals decide what issues to care about is still in its infancy, the literature usually singles out three main systematic individual-level sources related to material self-interest, group interest, and predispositions (Boninger et al. 1995; Howe and Krosnick 2017; Miller, Krosnick, and Fabrigar 2017). Notably, all of these factors are also used to explain policy preferences within issues. Consequently, although issue importance and preferences are both conceptually distinct components of individual attitudes, they may be more or less empirically related depending on the particular issue and political context (Delton, DeScioli, and Ryan 2020; Jennings and Wlezien 2015).

Why do those who oppose immigration care more? One possibility is related to the previously documented asymmetry in the organization of US parties and the respective media environments (e.g., Grossmann and Hopkins 2016). If (increasingly anti-immigration) Republicans are more homogeneous in terms of ideology and media attention, they may care about relatively fewer issues than (increasingly pro-immigration) Democrats, but more intensely. Given that the same asymmetry is present outside the US in countries with multiple parties and different institutions and media environments (Kustov 2023a), however, it is unlikely to be just a function of the US partisan context.⁵

Another much likelier possibility is that the perceived economic or cultural threat from immigration is simply more psychologically potent and mobilizing than perceived opportunities (Ivarsflaten and Sniderman 2022), which may in turn be an instance of the more general phenomena of loss aversion in politics (Alesina and Passarelli 2019; Baumeister et al. 2001). Consistent with this, research indicates that the asymmetry appears to be specific to immigration: anti-immigration respondents care more about immigration in particular, not politics in general (see also figure 1). Given that issue importance is a relative concept and the amount of time people have is limited, it is likely the case that pro-immigration respondents care more about other issues such as healthcare, the environment, gun control, abortion, and so forth (the details of which likely depend on a particular context).

Figure 1 Immigration Issue Importance Asymmetry

Notes: The graphs show responses to the pretreatment “issue public” question by pretreatment (a), immigration preference or (b), partisanship, based on the UAS survey. Left columns show any-issue public members; right columns show immigration issue public members. Bars are 95% CI.

What can convince such people to think of immigration and its expansion as more important than other issues? Those people who generally support immigration already do not consider it a threat. They think that it is a good thing and that it should be increased, they oppose deportations, and they believe that all people should be treated humanely regardless of their immigration status. But they do not necessarily see increasing immigration as an enormous or urgent opportunity to solve their country’s problems, such as the aging population, labor shortages, or other issues they may care about (especially compared to other policies). Quite equivalently, they do not see current restrictions as imposing an urgent threat by preventing millions of people—including their compatriots—from reuniting with their families and participating in mutually beneficial interactions, including employment, investments, and trade.

At the same time, many scholars document the dominant motivating role of “sociotropic” considerations or perceived national interest (both economic and noneconomic) over people’s self-interest or humanitarian concerns in shaping their political attitudes, including those about immigration (Hainmueller and Hopkins 2014; Kustov 2021). I thus argue that informing pro-immigration respondents about the substantial (non)economic national benefits of expanding immigration to the United States (and thus the urgency of relaxing existing costly restrictions to help solve the country’s other ongoing problems), which many may simply be not aware of, should be an effective way to raise the personal importance of the issue among this population.

I focus on pro-immigration respondents in particular because these are the people who should be more responsive to this information due to its ideological congruency (Kahan 2016). Relatedly, I also focus on relatively uncommon sociotropic arguments as opposed to common humanitarian arguments in favor of expanding immigration, because pro-immigration respondents are already more likely to be familiar with and agree with the latter than the former. While most research focuses on anti-immigration voters and their motivations, pro-immigration citizens are arguably the group that is most likely to bring about a pro-immigration political change (Ivarsflaten and Sniderman 2022). This is especially true if one assumes that immigration preferences are relatively stable while the personal importance of immigration to voters is relatively volatile.

Research Design and Hypotheses

Designing an Effective Information Treatment to Reduce Immigration Issue Importance Asymmetry

Many studies show that choosing a good way of framing immigration such that it appeals to other values that voters hold dear may constitute an effective messaging strategy for increasing immigration support (Dennison 2020; Gilliam 2010). In a competitive political environment where voters are constantly exposed to anti-immigration counterframes, however, such an approach would require constant repetition (and media dominance) to be effective. According to recent extensive reviews of persuasion experiments (Coppock 2022; Haaland, Roth, and Wohlfart 2023), such framing or priming interventions that simply make existing knowledge accessible are less effective and durable than information provision interventions that instead make new knowledge applicable.

Consequently, simply reminding pro-immigration respondents about the issue by priming it with a short text or other media would likely not be sufficient to change their issue priorities in a durable way. In line with this, the only survey experiment I am aware of in which the authors tried to increase the personal issue importance of immigration by showing fictitious news articles on the issue was not successful (Arias and Blair 2021). It appears that robustly changing personal issue importance is much harder than simply changing the “salience” or mere accessibility of a given issue in a certain context (Bizer and Krosnick 2001; Dennison and Geddes 2019).

It is true that particularly striking, tragic events or stories—such as the famous 2015 photograph of a young Syrian child, Alan Kurdi, who drowned trying to reach Greece—may motivate sympathetic people to help or even demand political change. Unfortunately, these empathic effects are often short-lived, and simply reminding people about those tragedies may not prove very effective (Slovic et al. 2017). Furthermore, since the costs of immigration restrictions usually play out in the form of missed opportunities (e.g., a business that was not started because the jobs could not be filled), it is hard to find convincing stories in favor of expanding immigration. In fact, I am not aware of any recent events that made US voters demand more immigration.

Among possible information interventions, verifiable narratives—defined as generalizable selective depictions of reality with causal arguments (Dennison 2021)—that can generate new knowledge in favor of increasing immigration (Cattaneo and Grieco 2021; Facchini, Margalit, and Nakata 2022) and its importance should also be preferable to fact-checking approaches that simply attempt to correct people’s misperceptions about current immigrants and their alleged harms (Abascal, Huang, and Tran 2021; Grigorieff, Roth, and Ubfal 2020; Hopkins, Sides, and Citrin 2019).⁶ At the same time, those narratives that align with respondents’ sociotropic concerns by emphasizing benefits to citizens should be more effective than those that appeal to humanitarian concerns, even among pro-immigration adults (Dennison 2021; Gest 2022; Kustov 2021; Voelkel et al. 2022).

Finally, given the potential relevance of loss aversion and (the lack of) urgency in explaining the issue importance gap between pro-immigration and anti-immigration adults, pro-immigration “threat” narratives that directly address these concerns should also be more effective. In other words, providing explicit (counterfactual) information on the enormous economic and noneconomic costs of current immigration restrictions to the United States now, in addition to the benefits of increasing immigration in the future, might be informative to pro-immigration voters even if the former is arguably implied by the latter.⁷ Importantly, to the extent that such information could change people’s minds, it should work by generating new knowledge or updating people’s relevant empirical beliefs about the importance of immigration and its expansion.

What are the particular arguments about benefits that pro-immigration voters—who already believe that immigration is generally good—can find novel and thus potentially persuasive? According to my content analysis of over a hundred randomly selected immigration news stories from the last decade in major US media outlets (see appendix E), various appeals to the national benefits of immigration are quite common among those arguing in favor of it. However, pro-immigration stories predominantly focus on the arguments about the contributions and the rights of those immigrants who are already here. They also commonly address counterarguments against the frequently invoked harms of immigration (e.g., “immigrants lower wages”) and caution against decreasing legal immigration. But they rarely make an explicit, affirmative argument about the benefits of increasing legal immigration or address the costs of current immigration restrictions. In contrast, anti-immigration stories usually invoke a sense of urgent national threat coming from both current and potential immigrants with an explicit call to reduce future immigration as the main objective.

While the relevance of particular immigration benefits may vary across individuals and contexts, it is likely the case that mentioning specific reasons why increasing immigration is urgent or how it can help other issues pro-immigration voters may care about—like boosting the occupational mobility of native workers—should be an effective strategy (Dennison 2020).

Hypotheses

Based on the discussions above, I specify the following main hypothesis about the effect of information on immigration’s personal issue importance among pro-immigration respondents:⁸

H1 : Receiving relevant information about the national benefits of increasing immigration will increase the personal issue importance of immigration among pro-immigration respondents.

I also specify two related hypotheses focusing on shifting priorities within the broad domain of immigration policies and related behaviors:

H2 : Receiving relevant information about the national benefits of increasing immigration will increase the perceived issue importance of subissues related to immigration flows among pro-immigration respondents (relative to subissues related to immigration stocks).

H3 : Receiving relevant information about the national benefits of increasing immigration will increase the behavioral manifestations of the personal issue importance of immigration among pro-immigration respondents.

Although the study is primarily designed to test the effects of relevant information on immigration issue importance among pro-immigration respondents, there are arguably at least two additional hypotheses about the information effects among non-pro-immigration respondents that are worth considering explicitly. First, while the backlash against counterattitudinal information is rare in terms of preferences (Coppock 2022), it is possible that the information about immigration benefits can increase immigration issue importance among some anti-immigration respondents by simply reminding them about the issue they dislike. However, it is also possible that, after learning about the benefits of immigration, others may decrease the importance of opposing immigration or even decrease their opposition entirely. After all, there is some evidence that information about national benefits can change people’s minds on the issue (e.g., Facchini, Margalit, and Nakata 2022). Given that people tend to dislike counterattitudinal information even when they find it persuasive (Coppock 2022), most anti-immigration respondents and those currently indifferent regarding the issue will likely not find the provided pro-immigration information motivating or change their perceived importance of the issue. Consequently, I also specify the following auxiliary hypotheses regarding non-pro-immigration respondents:

H4 : Receiving relevant information about the national benefits of increasing immigration will not increase the personal issue importance of immigration among non-pro-immigration respondents.

H5 : Receiving relevant information about the national benefits of increasing immigration will not increase pro-immigration preferences (among all respondents).

Although my study makes a case that providing relevant information can change people’s issue priorities (if not their minds), null findings in a high-quality sample will be informative in light of the growing evidence of immigration attitude stability. Given that immigration issue importance asymmetry is present across contexts, it is possible that providing information cannot easily reduce the issue importance gap between pro-immigration and anti-immigration voters.

Finally, it is important to address the potential ethical issues of the proposed study. Even though my survey simply aims to provide new verifiable information to its participants, one may reasonably wonder about the normative considerations of shifting voter priorities. Given that people have limited time and resources, increasing immigration issue importance by definition should come at the expense of some other political issues (or nonpolitical activities for that matter). While reasonable people may disagree on the desirability of these changes (depending on the particular trade-offs involved), the study presents minimal risks to individual respondents since their participation is voluntary and there is no deception involved in any aspect of the study.

Data and Measurement

The study is based on a probability-based, nationally representative US survey experiment (N = 3,450) administered by University of Southern California’s Understanding America Study (UAS) in June 2023 (Alattar, Messel, and Rogofsky 2018; Kustov 2024).⁹ The survey took eight minutes to complete on average. Prior to treatment, the subjects were first asked about a set of standard sociodemographic questions, issues they care about in an open-ended way (Ryan and Ehlinger 2023), and their immigration preferences (to determine pro-immigration voters), among other political questions (to avoid priming and demand effects). Respondents were then randomly exposed to one of the information treatments with an encouragement to read it carefully and answer a substantive question to encourage attentiveness (or no treatment). Finally, respondents completed a set of self-reported questions about immigration (sub)issue importance, repeated posttreatment preference items, a behavioral item on petition signing, and manipulation checks (for specific procedures and items, see the appendix).

As for the experimental setup, 50% of the respondents were exposed to one of the five texts about the benefits of expanding legal immigration to the US (combined treatment group); the remaining 50% of respondents were in one of the four placebo groups or the pure control group, with 10% in each (combined control group). For all hypotheses, the treatment effects were estimated across all treatment texts pulled together compared to all placebo and control groups pulled together.

The approximately two-hundred-word information treatment(s) were based on recent policy reports and editorials that communicate the significant benefits of expanding migration to the US through an accessible, verifiable, and nonjudgmental narrative.¹⁰ Given the recent call for “stimuli sampling” as an important complement to “participant sampling” to improve external validity (Clifford, Leeper, and Rainey 2023; Gigerenzer 2022), respondents in the treatment group were randomly assigned to one of the five conceptually similar but distinct narratives (see table 1 for one example).¹¹

Table 1 An Example of a Treatment Narrative about the Broad National Benefits of Immigration

With some variation in the language and arguments used, all of these texts convey that (1) increasing legal immigration is beneficial to the US and its citizens. They also explain that this implies that (2) current restrictions are costly economically or threatening otherwise, and that (3) increasing immigration or relaxing restrictions can help to solve other important problems. Allowing more people to come to the US legally should significantly help the country by growing the economy, filling labor shortages, increasing innovation, boosting native incomes and careers, or improving international standing. At the same time, the current immigration restrictions in place hurt Americans, and retaining them would prevent natives from reuniting with their families and stop willing immigrants from contributing to the US. Rather than changing people’s minds from opposing to supporting immigration, these pretested information treatments are designed to change the personal importance of increasing immigration among those who already support it.

To avoid demand effects, none of the treatment texts explicitly said that immigration was more important than other issues. The placebo treatments further helped to account for potential concerns about priming effects and social desirability bias related to pro-immigration respondents adjusting their issue importance responses in response to any (pro-)immigration information or arguments about other policies and their benefits.

Given the random assignment, to test my first three hypotheses (H1–H3), I simply compared the mean values for relevant issue importance indices between the combined treatment and the combined control and placebo groups using a standard difference-in-means estimator among the pro-immigration respondents. To test H4, I similarly compared the issue importance index between the combined treatment and the combined control and placebo groups among the non-pro-immigration respondents. To test H5, I maximized statistical power by using a “pre-post” experimental design (Clifford, Sheagley, and Piston 2021) and comparing the posttreatment preferences between the pulled treatment and control groups after statistically controlling for pretreatment preferences among all respondents in linear regression.¹²

Since there could arguably be several possible “nonspecific effects” on issue importance behind my treatment informing people about the benefits of increasing immigration, I followed the recently suggested “agnostic approach” to experimental controls by Porter and Velez (2022) and averaged over multiple placebo treatments alongside the pure control. Most importantly, these nonspecific effects could include increased perceptions of immigration importance due to simple priming or invoking any positive arguments in favor of the issue and its decreased perceived importance due to discussing any other policy issues and their benefits. To address these issues, 10% of the respondents were exposed to a text mentioning policy-neutral facts about immigration (placebo group 1); 10% of the respondents were exposed to a text mentioning common superficial arguments in favor of immigration, such as “immigrants fill jobs that natives do not want” or that “they deserve our compassion” (placebo group 2); 10% of the respondents were exposed to a text about the national benefits of reducing healthcare costs (placebo group 3); 10% of the respondents were exposed to a text about the national benefits of building more housing (placebo group 4); and 10% of the respondents were exposed to no text (pure control group). Although none of these placebo treatments were expected to have any significant effects compared to the pure control, they consisted of a combination of texts that could either slightly attenuate or strengthen the effect of the intervention when included in the control group, and they were deliberately designed to reflect realistic information alternatives that people might encounter in their lives.

No attention checks or other filters were used to screen the participants in the target sample. All hypotheses and specifications were preregistered prior to data collection. For a summary of hypotheses and methods, see table A1. For power analysis, see appendix C. For a summary of all results using preregistered ordinary least squares (OLS) specifications, see table A2. For alternative exploratory specifications, see table A3. In testing hypotheses, I also adjusted for multiple comparisons using the Holm-Bonferroni correction (see table A4). For the discussion of ethics, see appendix D.

Measuring Immigration Preferences, Issue Importance, and Their Behavioral Manifestations

To minimize measurement error, the survey included multiple previously validated preference and issue importance items averaged as 0–1 indices. To measure immigration preferences (and categorize respondents as pro-immigration or anti-immigration), I used three standard four-point items gauging people’s support for increasing/decreasing immigration flows, making immigration easier/harder, and relaxing/tightening immigration laws.

To date, there is no universally accepted way to measure personal issue importance (Moniz and Wlezien 2021).¹³ Fortunately, unlike previous studies mostly based on the existing data created for other purposes, I did not have to rely on the common “what is the most important problem facing the country today?” question or similar items. Such traditional methods have a notable limitation: they can confuse immigration’s immediate accessibility in people’s minds due to recent news exposure or any other priming stimuli with long-term concern for the issue. Additionally, equating the recognition of an issue as a “problem” with genuine care about it can be misleading (Ryan and Ehlinger 2023; Wlezien 2005). This is especially true for pro-immigration voters who might be reluctant to label “immigration” as a problem even when they care about it.

Instead, I used a novel, previously validated “issue-public” question (Ryan and Ehlinger 2023) as a “strong” pretreatment measure of personal issue importance (of immigration). This two-part question is particularly well suited to revealing any substantively important issue importance asymmetry since it explicitly asks respondents to report if they have any long-standing engagement with any political issues (without prompting them about “immigration” or “problems”). The item reads as follows: “Some people have a political issue that they care about more than most other issues. They might think about the issue a lot. They might pay particular attention to news about that issue, even when it’s not making national news. They might focus on what political candidates say about that issue, and decide who to vote for on the basis of that issue. Or they might just care about the issue a lot. Is there an issue like that for you?” (“Yes” or “No”; if respondents say “Yes,” they are asked a follow-up question: “In just a few words, what issue or two do you care about?”)

Despite its ability to capture how much people care about issues, the issue-public membership question may not be ideal as a posttreatment measure of personal issue importance since it explicitly asks people about their long-standing engagement with issues in the past, which cannot feasibly change during the survey. As my main posttreatment outcome of interest, I thus used a more standard battery of three four-point items gauging whether respondents believe immigration issues are more important to the US than other political issues, to what extent immigration issues are important to them personally, and how strongly they feel about immigration issues.¹⁴ I also included a quasi-behavioral issue importance item in which respondents could express their willingness to sign a petition in line with their preferences.

To measure perceived subissue importance, I asked respondents to select up to three (neutrally worded) issues that they believe are the most important to address and then simply calculated the number of mentions by issue or issue category (e.g., “flows” or “stocks”). The classification of immigration subissues is based on the original, previously pretested battery item derived from the widely used Immigration Policies in Comparison (IMPIC) policy coding (Helbling et al. 2017).

Analysis and Results

In line with the previous evidence on the relatively low prevalence of political interest in the public, only about 49% of respondents answered affirmatively to the first issue-public question asking if there was any political issue at all they thought a lot about, considered more important than other issues, focused on when mentioned by politicians, paid particular attention to in the news, or based their vote upon. Among those respondents, 13% then specified they cared particularly about immigration or an immigration-related issue like “securing our border” or “treating immigrants with respect” (unprompted). While this may not seem like a big number, there was only a single other issue chosen by more respondents at the time of the survey in June 2023 (abortion, chosen by 16% of issue-public members). Similar to recent benchmark surveys, 35% and 40% of respondents then also reported consistently pro-immigration and anti-immigration preferences (pretreatment).

Descriptive Findings 1: Pro-Immigration Voters Care Less about Immigration

Overall, based on this new issue-public membership measure, the results confirm the existence and qualify the substantial size of “issue importance asymmetry” between pro-immigration and anti-immigration groups (see figure 1a). While immigration has been consistently one of the most important policy issues to US voters, only four (±1.5%, given a 95% CI) pro-immigration respondents said they care about it. At the same time, 9% of anti-immigration respondents could be identified as such “immigration issue-public members.” In other words, only approximately 20% of the immigration issue-public members or voters who truly care about the issue have a pro-immigration preference. This is particularly notable since pro-immigration voters are generally more, not less, politically active (51% are able to identify at least one issue they care a lot about compared to 43% of anti-immigration voters).

As can be seen from figure 1b, while Republicans and Democrats care equally about politics in general, Republicans care more about immigration in particular. Importantly, however, the immigration importance asymmetry is not just a function of partisanship. The same asymmetry is present across all partisan groups: anti-immigration Republicans and Democrats care more about the issue (14% and 6%) than pro-immigration Republicans and Democrats (6% and 4%).¹⁵

Similarly, although less dramatically, the (posttreatment) immigration issue importance index has been higher for anti-immigration than pro-immigration voters in the pure control group (0.63 versus 0.59). In particular, anti-immigration voters are more likely to say that immigration is very important to them personally, that it is more important than other issues, and that they feel strongly about it (67% versus 60% on average).

Interestingly, however, despite these results and the findings of previous research (Kustov 2023a), pro-immigration and anti-immigration voters were more or less equally willing to sign a petition to express their views regarding the issue (66±3% among all respondents in the control group). While this requires further investigation, it is possible that pro-immigration respondents were more likely to sign any political petition regardless of the issue due to their greater political engagement.

Table A3 presents the possible demographic differences of interest. As expected, foreign-born respondents were much more likely to perceive immigration as an important issue, regardless of their immigration preference. Additionally, older women who support immigration and older men who oppose it were more likely to consider the issue important. Interestingly, having a college degree was not a significant factor, while high-income pro-immigration respondents were less likely to perceive immigration as important.

Descriptive Findings 2: Pro-Immigration Voters Care about Helping Immigrants, Anti-Immigration Voters Care about Preventing Future Immigration

What do people mean exactly when they say they care about immigration? The novel descriptive results on specific immigration “subissues” indicate that, while anti-immigration respondents prioritize policies about (reducing) future immigration flows (and law enforcement), pro-immigration respondents prioritize helping immigrants who are already here (see figure 2). In particular, anti-immigration respondents cited “allowed numbers of immigrants each year” and “skill and English requirements for new immigrants,” alongside the more expected “illegal immigration and deportations” and “wall construction along the US–Mexico border,” as the most important immigration issues for the government to address. The most important immigration issues among pro-immigration respondents were, as expected, about the “time and difficulty of acquiring US citizenship,” “legal status of children brought here illegally,” and the “treatment of immigrants by natives,” alongside “the process of applying for asylum in the US.”

Figure 2 Immigration Subissue Importance Asymmetry

Notes: The graphs show the importance of immigration subissues by (pretreatment) immigration preference or partisanship based on the UAS survey. Bars are 95% CI.

However, it is important to acknowledge that some immigration subissues do not easily fit the distinction between flows and stocks, and that there is a considerable amount of nuance to these findings. For instance, contrary to expectations, pro-immigration respondents still considered “availability of temporary work visas” and “red tape and delays in processing immigration forms” as more important than anti-immigration respondents, while anti-immigration respondents prioritized “immigrants’ access to government benefits” more than pro-immigration respondents. Interestingly, while the partisan differences between Republicans and Democrats in immigration priorities largely mirrored the breakdown in figure 2, they were smaller across all subissues. Since subissue importance and partisanship are posttreatment variables, I replicate the same results exactly based on the pure control subsample (see figure A2). Among the possible demographic differences of interest (see table A3), only educated respondents were more likely to perceive immigration flows to be more important that other stock subissues.

Experimental Findings 1: Relevant Information Increases Immigration Issue Importance

In line with my main preregistered hypotheses and specifications (H1 and H4), the experimental results show that providing verifiable narratives informing respondents about the national benefits of increasing immigration to the US is effective. In particular, it increases immigration’s stated issue importance among pro-immigration adults (but not non-pro-immigration adults) and thus reduces the immigration issue importance asymmetry observed earlier (see figure 3).

Figure 3 Information Effects on Immigration Issue Importance

Notes: The graphs show the treatment effects on the personal importance of immigration among pro-immigration (a, H1) and non-pro-immigration respondents (b, H4) based on the UAS survey. Bars are 95(84)% CI.

After reading about the broad benefits of increasing immigration and the costs of not doing so, pro-immigration respondents were significantly more likely to report that immigration is important (~0.04 on a 0–1 index scale, or a Cohen’s d of 0.16). Substantively, this amounts to, on average, 9±5 percentage points (or 16%) more pro-immigration respondents considering immigration to be personally important, believing it is more important than other issues, and feeling strongly about it (given the baseline of 62%). Additional exploratory analyses indicate that the effects are similar or stronger (~0.08 on a 0–1 index scale, or a Cohen’s d of 0.33) with the pure control as a baseline, and that the effects are not driven by any one particular treatment text.

At the same time, in line with H4, the treatment does not affect the personal issue importance of the issue among non-pro-immigration voters. This holds true regardless of specification. Of course, the lack of statistical significance or the failure to reject the null hypothesis does not necessarily imply that the effect is zero.¹⁶ Additional exploratory analyses indicate that the substantively similar practical null findings are true for the subgroup of anti-immigration respondents too (excluding neutral or ambivalent respondents).¹⁷

Finally, a simple linear model of the outcome index as a function of the interaction between the treatment and the pretreatment preference indicates that the effects are significantly stronger among pro-immigration respondents (by 0.05 points on a 0–1 scale, p < 0.01). Importantly, this is comparable to the size of the immigration issue importance asymmetry between pro-immigration and anti-immigration respondents observed in the control group.

Experimental Findings 2: Relevant Information Does Not Change Immigration Subissue Priorities

Apart from the general importance of immigration to respondents, I also prespecified two additional possible outcomes of interests, including people’s stated importance of flows-related issues as opposed to stock-related issues within immigration, as well as their possible behavioral response in the form of signing a petition expressing their views on the issue.

As can be seen from figure A3a, contrary to expectations, the provided information does not impact the relative importance of immigration flows among pro-immigration respondents. In the preregistered specification, the effects are almost precisely estimated at zero. This is also true for a number of alternative specifications and operationalizations of the subissue importance index, including looking at the pure control or the simple count of flows-related issues (not shown).

In figure A3b, I test whether the treatment can increase a quasi-behavioral manifestation of issue importance related to the willingness of respondents to sign a petition sharing their views with members of Congress (H3). While the treatment coefficient is in the expected direction, it is not statistically significant at the conventional level. Equivalence tests further confirm the practical null effects of both H2 and H3 given the bounds equivalent to a Cohen’s d of ±0.15.

Experimental Findings 3: Relevant Information Also Increases Pro-Immigration Preferences

Finally, I also included a pre-post measure of people’s immigration policy preferences. Contrary to my expectations and much of the literature, the information had significant effects on not just people’s empirical beliefs about immigration but also their normative beliefs about what the government should do about it (see figure 4a). In particular, after reading about the provided narratives, respondents were significantly more likely to report pro-immigration preferences (~0.04 on a 0–1 index scale or a Cohen’s d of 0.2). Substantively, this amounts to, on average, 8±2 percentage points (or 17%) more respondents saying that immigration should be increased, that the process should be easier, and that the existing regulations should be relaxed (given the baseline of 46%).

Figure 4 Information Effects on (Weighted) Immigration Preferences

Notes: The graphs show the treatment effects on pro-immigration preferences (a, H5) weighted by immigration issue importance (b) based on the UAS survey. All estimates statistically control for pretreatment preferences. Bars are 95(84)% CI.

Additional Exploratory Tests

Additional exploratory specifications using the pure control only, excluding statistical controls for pre-treatment outcomes, or including demographic controls and survey weights, all confirm my main results (see table A3). My further analysis of manipulation checks indicates that, at least to some extent, the treatment changed immigration issue preferences and importance by providing novel information and changing people’s relevant beliefs about the issue. After seeing one of the treatment narratives, respondents were more likely to agree with the factual statements that the average US citizen would be better off with more immigration, that current restrictions are harmful, and that increasing immigration can help other important problems (see figure A4).

Given the unexpectedly strong and positive treatment effects on pro-immigration preferences, I also calculated an additional outcome of interest related to people’s immigration preferences weighted by how much it is personally important to them (so that both pro-immigration and anti-immigration preferences are amplified when people say they care about immigration). For ease of interpretation, I recoded the variable to vary from −1 (strongest anti-immigration preference) to +1 (strongest pro-immigration preference). As indicated in figure 4b, the information treatment on weighted preferences is strongly positive, and it is close to the sum of its effects on both issue importance and preferences. It also evidently shifts the mean weighted US opinion on immigration from slightly anti-immigration, as observed in previous research, to close to neutral. The analysis of possible subgroup effects by pretreatment preference shows that the treatment had a similar average treatment effect on preferences among pro-immigration and non-pro-immigration respondents (as well as Democrats and Republicans). Contrary to some prior research indicating that the issue publics may be less likely to change their mind (Vidigal and Jerit 2022), there were also no heterogeneous treatment effects by pretreatment issue importance.

Finally, the analysis of separate text-specific effects indicates that the findings are likely not driven by any particular treatment or placebo narrative (see figures A5–A7). While there is some variation in the point estimates across treatment and placebo texts, with the common talking-points condition (placebo 2) occasionally showing effects similar to the treatment conditions, there is no clear pattern across all specifications. It is important to highlight, however, that these exploratory analyses were not preregistered and are underpowered. Given this lack of consistency and power, the observed variations may be due to random chance rather than any systematic mechanism. Future research with larger sample sizes and alternative treatments could help to investigate the potential nuances in how different types of information affect issue importance.

Discussion and Conclusion

Even when Americans support immigration in the polls, lawmakers are hesitant to enact significant pro-immigration reforms. Research shows that to the extent these positive attitudes are real, they are likely not meaningful in terms of people’s engagement with the issue. In fact, when one considers personal issue importance, there have always been more Americans who oppose immigration and consider it more important than otherwise.

Using a new issue-public measure of personal issue importance, the descriptive results provide an important qualification to the sheer scope of the previously identified asymmetry of immigration attitudes that systematically disadvantages the pro-immigration side beyond the partisan dynamic. The novel detailed results on immigration subissues further reveal for the first time that pro-immigration and anti-immigration voters often prioritize very distinct problems and policies within the issue. Most generally, it appears that pro-immigration voters care more about helping existing immigrant populations while anti-immigration voters care more about reducing future immigrant populations and enforcing existing restrictions. I also show that these patterns are largely mirrored in mainstream media where, despite the abundance of explicitly pro-immigration stories, stakeholders rarely make a strong case for expanding legal immigration.

These descriptive findings indicate that pro-immigration advocates will continue to face a systematic disadvantage compared to their counterparts unless they figure out a reliable method to raise the importance of immigration among sympathetic voters only. The experimental part of this study aimed to address this concern by testing the potential effectiveness of information campaigns that clearly communicate the positive implications of increasing immigration to the United States and the threats of not doing so.

Based on a large-scale representative trial, the experimental results confirm that providing such relevant information through a short, verifiable, nonjudgmental narrative can raise the perceived issue importance of immigration among pro-immigration voters and thus reduce the observed issue asymmetry. At the same time, the results indicate that motivationally incongruent information does not change the personal issue importance of immigration among anti-immigration adults. This is an important finding since it implies that the proposed pro-immigration intervention is unlikely to backfire by unintentionally mobilizing those who oppose immigration.

However, it is important to acknowledge that not all prespecified tests have panned out as planned. On the one hand, contrary to my expectations, the provided information was ineffective at either changing people’s priorities within immigration or encouraging more petitions. These precise null results imply that it is likely more difficult to shift those outcomes compared to a general personal issue importance or to measure these outcomes properly (or perhaps a combination of both). For instance, the fact that more people stated their willingness to sign a petition than said that immigration is important to them indicates that this was not perceived as a particularly costly behavior (or that it was no more “real” than a related stated attitude).

On the other hand, contrary to my expectations and a growing literature on the stability of immigration preferences and their robustness to new information, reading these narratives did make people significantly more pro-immigration across a diverse group of voters. There can be a few possible reasons why persuasion effects have been observed, at least in the short term. First, it is possible, though unlikely (Clifford, Sheagley, and Piston 2021), that the observed effects are an artifact due to the use of a pre-post design and the related demand effects. What is more likely is that persuasion happened as a real “side” effect of experimental design. While the presented narratives were designed to appeal to pro-immigration voters and increase how much they care about the issue, they all also make explicit nonpartisan arguments to increase immigration, which some skeptical voters may find persuasive (and more so than simple immigration fact-checking commonly used in previous research). Combined with the positive effects on issue importance among sympathetic voters, this implies that the treatment can reliably shift the “weighted” pro-immigration preferences in the electorate, which is arguably a more meaningful indicator of public opinion and voter influence than the policy preference responses only.

Nonetheless, it is also important to replicate these results using alternative treatments across different contexts. While the observed effect sizes are rather small, they are in line with comparable treatments in the existing literature. They are also arguably realistic given the nature of the treatment and the setting, a half-page text with which most respondents engage for no more than a few minutes as part of an online survey. It is possible that repeated exposure to a set of related narratives about the benefits of expanding immigration, especially if it is decisively framed in terms of the threat of not doing so, would produce more change. It is also important to acknowledge the potential limitations and the difficulty of changing people’s priorities within issues and their behavioral manifestations.

Future research can test various alternative ways to boost the (sub)issue importance of immigration in the US and/or across other immigrant-receiving countries. It can also explore whether the effects observed here are long lasting in a longitudinal survey and whether alternative treatments can do a better job at changing people’s priorities within issues and related behaviors. Finally, future research can consider how the more ideologically congruent information about nonimmigration issues (e.g., gun rights) can decrease the relative personal issue importance of immigration among anti-immigration respondents (or even pro-immigration respondents). Given that the treatment selection in all of these cases has to be based on distinct issue-specific and context-specific considerations, it was beyond the scope (and the sample constraints) of this study.

All in all, the project descriptively and experimentally examined an important yet largely overlooked reason why the positive public opinion on immigration has not translated into pro-immigration political change. However, it is important not to overstate the potential consequences of the immigration issue importance asymmetry and its possible change for policy outcomes. Public opinion is only one policy input among many. Due to increasing partisan polarization and congressional gridlock, for instance, even the much higher issue importance among the larger number of pro-immigration voters may not straightforwardly result in the desired policy change. Still, to the extent the observed issue importance asymmetry is at least in part a result of not sufficiently reliable information campaigning in favor of increasing legal immigration, my results suggest that the benefits of such campaigns in terms of shifting voters’ priorities have a potential to outweigh the possible costs even in the currently polarized US context.