Introduction
Public opinion toward immigration is highly polarized across partisan lines in the United States. Democrats are consistently more supportive of liberal immigration policies than Republicans (Gravelle, Reference Gravelle2016; Ollerenshaw and Jardina, Reference Ollerenshaw and Jardina2023; Sides, Tesler, and Vavreck, Reference Sides, Tesler and Vavreck2018, 142). This study examines whether this partisan gap reflects voters’ assumptions about the partisan leanings of newly arriving migrants.
Evidence has consistently shown that immigrants in the United States are more likely to support and vote for the Democratic Party (Baldassare et al., Reference Baldassare, Bonner, Dykman and Lopes2018; Mayda, Peri, and Steingress, Reference Mayda, Peri and Steingress2016; Patten and Lopez, Reference Patten and Lopez2013; Pillai and Brodie, Reference Pillai and Brodie2023). We expect that most American voters – regardless of their own partisan identity – are aware of this fact, in part because it has featured prominently in media coverage of immigration. For example, mainstream media outlets have noted that naturalized citizens are a reliable constituency for the Democratic Party, and portrayed citizenship for migrants as an “electoral bonanza” for Democrats that would “cripple” Republican prospects at the ballot box (Schultheis, Reference Schultheis2013). This line of reasoning has also been deployed by far-right media figures like Tucker Carlson, who has claimed that the Democratic Party is seeking to bring about demographic change via mass migration “in order to win and maintain power” (Johnson, Reference Johnson2021). In part because of public statements like these, one-third of Americans hold the misperception that there is an intentional effort to “replace” native-born Americans with immigrants to bring about a permanent Democratic majority (Snow, Reference Snow2022).Footnote 1
Do such presumptions about the political leanings of immigrants influence individual-level support for immigration in the United States? We seek to answer this question by testing the claim that Americans support higher levels of legal immigration when they believe that migrants share their political preferences. Individuals may favor politically like-minded immigrants due to strategic, electoral, considerations. Existing scholarship finds that Americans are more supportive of electoral reforms (Biggers, Reference Biggers2019) and vaccination campaigns (Kane and Anson, Reference Kane and Anson2025) when voters are informed that these measures will improve their preferred party’s performance at the polls. Similar strategic thinking could increase support for immigration if voters expect that, once naturalized, migrants will support their party in future elections (Hawley, Reference Hawley2011, 405). Moreover, the preference for co-partisan migrants may arise for reasons of political homophily. Because individuals tend to evaluate people who share their political views more favorably (Huber and Malhotra, Reference Huber and Malhotra2017; Clifford et al., Reference Clifford, Simas and Suh2025), they may express more support for co-partisan immigrants. Whether it reflects strategic calculations, an inherent preference for like-minded people, or some combination of the two, we expect Democrats and Republicans will be more supportive of immigration when they believe that immigrants share their partisan preferences.
To test these assertions, we fielded a survey in the United States the week before the 2024 US Presidential election, based on a large national sample benchmarked to demographic quotas. Our survey included an experiment that manipulated information about immigrants’ partisan preferences. We find that (1) Americans overwhelmingly assume that immigrants support Democrats in elections and (2) shifting respondents’ perceptions of immigrants’ political views significantly impacts support for immigration. In brief, when voters are informed that immigrants are likely supporters of Donald Trump and Republican candidates, support for immigration falls among Democrats but increases among Republicans. Indeed, partisan differences on immigration disappear completely in our sample when voters expect immigrants to cast their ballot for Trump.
Our results provide new insights into public opinion about immigration. Previous work recognizes that public support for immigration depends on the attributes of immigrants, though most prior work has focused on the economic and cultural features of immigrants (Bansak, Hainmueller, and Hangartner, Reference Bansak, Hainmueller and Hangartner2016; Hainmueller and Hiscox, Reference Hainmueller and Hiscox2010; Hainmueller and Hopkins, Reference Hainmueller and Hopkins2015; Newman and Malhotra, Reference Newman and Malhotra2019). This study contributes to an emerging body of scholarship that highlights the importance of immigrants’ political features. Existing scholarship has found that Americans are more supportive of internal migration and are more open to extending voting rights to foreign migrants when newcomers are viewed as co-partisans (Whitaker and Giersch, Reference Whitaker and Giersch2021; Alarian and Zonszein, Reference Alarian and Zonszein2025). Holland, Peters, and Zhou (Reference Holland, Peters and Zhou2024) find that Colombians favor migrants with similar ideological views to their own. Our findings add to our existing knowledge by showing that perceptions of immigrants’ partisan preferences impact support for foreign migration among Americans on both ends of the political spectrum.
Practically, this study suggests that if the American public’s assumptions about immigrants’ partisan allegiances move away from the status quo, underlying preferences for immigration will also change. More precisely, if immigrants come to be viewed as political “free agents” rather than reliable supporters of the Democratic Party, our findings imply that partisan differences on immigration will become less polarized, with support for immigration among Democrats falling and support among Republicans increasing. Considering evidence that Donald Trump outperformed expectations among immigrant communities (Frimpong and Sanchez, Reference Frimpong and Sanchez2024), we may be closer to that shift than many anticipate.
Empirical strategy
To assess how the perceived partisan preferences of immigrants affect Americans’ support for immigration, we embedded a preregistered experiment into a large nationwide survey. The survey was fielded in the United States from October 24 to November 1, 2024, just days before Americans headed to the polls to select their new president. Dynata recruited 3,000 subjects from their panel of respondents. Demographic quotas for gender, age, and education were included to obtain a sample that is broadly representative of the overall population. Appendix A describes the survey instrument and sample in greater detail.
Our main experimental treatment informed roughly half of our subjects that a particular immigrant group favored Donald Trump and the Republican Party in the 2024 election. We did not include a separate treatment where immigrants were described as likely supporters of Harris and the Democrats because we anticipated that most Americans would already expect this to be the case – an assumption that we also test below. To maintain factual accuracy, the experiment focuses on migrants from Vietnam and Venezuela as voters from these immigrant communities swung heavily toward Donald Trump in the 2020 presidential election (AAPI, 2020; Bahar, Reference Bahar2021; Daugherty et al., Reference Daugherty, Smiley, Ocasio and Wieder2020). The fact that these countries differ in many other ways, from their predominant ethnicity to their geopolitical relationship with the United States, helps ensure that our results generalize across these dimensions.
Our experiment uses a 2 × 2 factorial design. Half of the subjects were asked about immigration from Venezuela, and the other half about immigration from Vietnam. The second intervention, which constitutes the experimental manipulation of theoretical interest, provides information about the partisan preferences of these migrant groups. The exact phrasing of the question is below, with the text of the main treatment condition in italics.
In the last ten years, a significant number of immigrants into the United States have come from [Vietnam/Venezuela]. A majority of immigrants from [Vietnam/Venezuela] support Donald Trump and the Republican Party. Do you agree or disagree that the United States should permit higher levels of immigration from [Vietnam/Venezuela]?
Respondents were then offered a scale ranging from 0 to 10, where 0 means “strongly disagree” and 10 means “strongly agree.”
Our main treatment aims to increase the likelihood that respondents perceive immigrants as supporters of Donald Trump and the Republican Party. If individual-level support for immigration reflects purely principled considerations, then immigrants’ attributes, including their political views, should have no bearing on immigration preferences (Wright, Levy and Citrin, Reference Wright, Levy and Citrin2016). However, if voters are more likely to favor immigration when they believe that immigrants share their political identity, manipulating the perceived political affiliation of immigrant groups should influence support for immigration. We therefore test the following two hypotheses:
H1: Information that immigrants support Donald Trump increases support for immigration among Republicans.
H2: Information that immigrants support Donald Trump decreases support for immigration among Democrats.
Results
Manipulation check
Before turning to our main experimental findings, we first assess whether voters’ perceptions about immigrants’ partisan preferences align with our expectations and whether our treatment changed voters’ perceptions. Immediately following our main question about support for immigration, subjects were also asked whether they expected Venezuelan or Vietnamese migrants to vote for Kamala Harris or Donald Trump in the 2024 presidential election. As expected, respondents in the untreated group overwhelmingly assumed that migrants would support Harris (79.6%) compared to Trump (20.4%). Conversely, individuals in the treatment group were more likely to say that migrants from these countries would support Trump (61.5%) compared to Harris (38.5%), indicating that our treatment changed most subjects’ presumptions about immigrants’ political preferences. The difference between the two groups is statistically significant (p < 0.01). See full manipulation check results in Appendix B.Footnote 2
Main results
We next compare mean levels of support for higher levels of immigration among self-identified Democrats, Independents, and Republicans in the treated and untreated groups. Party identity was coded based on individuals’ self-reported party affiliations.Footnote 3 Figure 1 reports our sample means by party identity, alongside 95% confidence intervals. Here, we have combined the subject responses whether they were asked about immigration from Venezuela or Vietnam.
Figure 1. Mean level of support for immigration.
Note: Main bars and error bars report sample means and 95% confidence intervals, respectively, for each partisan group by treatment condition.
In the control condition, where respondents overwhelmingly assume that migrants support Democrats, support for increasing immigration varies greatly by party identity. Democrats are most supportive of more immigration (5.3) while Independents (4.2) and Republicans (3.0) are decreasingly supportive. Differences between the groups are statistically significant (p < 0.01). Figure 1 also reveals another notable pattern: in the treatment condition, the gap in immigration attitudes between Democrats and Republicans completely disappears. Compared to the control condition, where the gap between these partisan groups is 2.3 points, this is a striking change.
These patterns, displayed in Figure 1, also provide preliminary support for our main hypotheses.Footnote 4 Consistent with H1, support for immigration among Democrats is nearly a full point lower in the treatment condition compared to the control condition, and this difference is statistically significant (p < 0.01). Support for immigration among Republicans is more than a point higher in the treatment condition than in the control condition, and this difference is also statistically significant (p < 0.01). By contrast, Independents’ opinions about immigration are nearly identical in the treatment and control conditions.
Figure 2 displays estimated average treatment effects alongside 95% confidence intervals across three party identification categories, formally testing whether the differences observed in Figure 1 are statistically distinguishable from zero. These estimates are from an ordinary least squares regression model that contains interaction terms between a respondent’s party identity and whether they received the Trump treatment. The model also controls for several demographic variables that may be predictive of immigration attitudes. We account for a respondent’s age, gender, educational attainment, income level, and whether they identify as non-Latino White, the inclusion of which should help lead to a more precisely estimated treatment effect.Footnote 5
Figure 2. Effect of trump treatment on support for increased immigration.
Note: Circles and lines display average treatment effects and 95% confidence intervals, respectively, of the Trump treatment. Estimates based on a specification including demographic controls.
The results presented in Figure 2 follow the general pattern observed in Figure 1. Information that most migrants support Trump and the Republican Party reduces support for more immigration by 0.9 points among Democrats. On the other end of the spectrum, the Trump treatment increases support for immigration by 1.3 points among Republicans. For both groups, the treatment effects are statistically significant. Independents are the only group for which the Trump treatment does not meaningfully change attitudes toward immigration, indicating that citizens unattached to either party are indifferent about immigrants’ partisan preferences.
Additional analyses
Our sample is broadly representative of the US population, but it skews slightly female, more educated, and older than the overall population. To assess whether this impacts our results, we applied poststratification weights based on gender, age, and level of education. The results are consistent with our unweighted estimates (see Table C2).
We also explore whether the Trump treatment effect was moderated by the migrants’ designated country of origin. Figure D1 in the Appendix shows that people respond similarly to the partisan affiliation of immigrants irrespective of whether they were assigned to the Venezuela group or the Vietnam group. Interestingly, all three partisan groups prefer Vietnamese over Venezuelan immigrants (see Table D1), which is consistent with previous work showing that Americans prefer Asian over Latino immigrants (Timberlake and Williams, Reference Timberlake and Williams2012) while also favoring immigrants from countries that have better political relations with their home country (Wimmer et al., Reference Wimmer, Bonikowski, Crabtree, Fu, Golder and Tsutsui2025). Despite these baseline differences, the Trump treatment exerts a similar effect across the two origin countries for Democrats. For Republicans, the Trump treatment effect is somewhat stronger for those assigned the Vietnamese immigrant condition compared to the Venezuelan condition. The weaker effect for those in the latter group may reflect stronger negative priors about immigrants from Venezuela. Overall, this suggests that while the partisan preferences of immigrants are likely to impact Americans’ attitudes about immigration in a directionally consistent manner regardless of the immigrants’ country of origin, the magnitude of this effect will vary across different source countries.
We also examine whether the main findings are contingent on various respondent attributes. The basic pattern holds for both female and male respondents, younger and older ones, rich and poor individuals, those with low and high education, as well as White and non-White Americans (see Figure E1).
Another potential challenge is determining whether it is individuals’ partisan identities, as opposed to their ideology, that determines how they respond to the treatment. Table F1 shows that the moderation effects of ideology and partisanship are directionally the same, which is unsurprising given the strong correlation between these two variables. Notably, however, ideology has much less explanatory power than partisanship, providing some evidence that peoples’ partisan identities have a larger impact in determining their responses to the information in the treatment. In addition, when we split the sample into three ideological groups (left, center, and right), we find that partisanship moderates the effect of the treatment within each of these ideological camps. These patterns increase our confidence that individuals’ partisan identities moderate how they respond to the treatment.
Finally, if voters favor immigrants for their strategic electoral value, it seems reasonable to expect that individuals with strong partisan attachments will react more strongly to the treatment than those with weaker attachments. Thus, we also test whether there are differences between “strong” and “weak” partisans. To do so, we compare five groups of voters, distinguishing Independents from respondents who identify as “strong” Republicans or Democrats, and those who claim to be “not very strong” partisans.Footnote 6 The results, presented in Figure 3 (see Table C3 for full results), show that the effects of our treatment are magnified for the strongest partisans.
Figure 3. Comparing strong and weak partisans.
Note: Circles and lines display average treatment effects and 95% confidence intervals, respectively, of the Trump treatment. Estimates based on a specification including demographic controls.
The treatment effect is 1.1 points for strong Democrats compared to 0.6 for weak Democrats, and this difference is marginally significant (p < 0.1). Similarly, we observe a stronger treatment effect among strong Republicans (1.7) than among weak Republicans (0.96), and the difference between these two groups is statistically significant (p < 0.05). This indicates that as the strength of voters’ partisan attachments increases, their sensitivity to immigrants’ political leanings increases as well.
Conclusion
In the United States, immigrants are overwhelmingly portrayed and perceived as future Democratic voters. Our sample aligns with this narrative, where eight out of ten untreated subjects believed that migrants would support Kamala Harris over Donald Trump in the 2024 Presidential election. We demonstrate that Americans’ perceptions of immigrants’ partisan leanings influence their attitudes toward immigration. When Americans are informed that migrants are likely to support Donald Trump, Republicans become more supportive of immigration while support among Democrats retrenches. In fact, partisan differences between the two sides of the political spectrum vanish when conventional wisdom about the perceived partisan identity of immigrants is flipped on its head. Importantly, our main result holds across two distinctive immigrant groups and among all major demographic categories in our sample. Moreover, the effect of the treatment is largest among strong partisans of both parties. In sum, the degree to which Americans support higher levels of immigration depends heavily on whether they expect newcomers to favor the same party that they do.
Despite our clear topline findings, our study is limited in several ways that open the door for future research. For example, because our study focuses on Americans’ attitudes toward legal immigration, we cannot definitively say whether our results would travel to voters’ views of other aspects of immigration policy, such as attitudes toward unauthorized migration and border security.Footnote 7 Furthermore, we did not include a question gauging individuals’ support for overall immigration levels in the United States, instead asking only about support for more immigration from two specific countries. Given that respondents in our survey respond similarly across these two very different countries, it seems likely that our results will apply to other countries as well as to attitudes about overall immigration levels more generally. However, additional research is needed to confirm whether this is the case. In addition, we are unable to conclusively determine why Americans favor co-partisan immigrants. Future research should examine whether political homophily, electoral concerns, or some combination of the two mechanisms impact why immigrants’ political leanings shape Americans’ support for immigration.
We also acknowledge that our experimental results could potentially be impacted by “information equivalence” problems (Dafoe et al., Reference Dafoe, Zhang and Caughey2018). Information about immigrants’ partisan identities might change respondents’ perceptions of other immigrant attributes, such as their skill level or English fluency, which could in turn impact their preference for immigration. While we cannot fully rule out this possibility, it seems unlikely that this can explain our main results. Previous work establishes that Democrats and Republicans both prefer immigrants with higher levels of education and English fluency (Hainmueller and Hopkins, Reference Hainmueller and Hopkins2015), and we find that they both share a preference for Vietnamese over Venezuelan immigrants (see Table D1). Hence, if survey respondents were really reacting to these other, nonpartisan, attributes of immigrants, we would expect the treatment to have a similar impact on the immigration preferences of Democrats and Republicans, not the opposite effect that we observe.
One implication of our findings is that Americans’ attitudes on immigration may begin to depolarize if migrants come to be viewed as politically up for grabs rather than being perceived of as a reliable Democratic constituency. This may already be happening. Initial analysis of the 2024 election indicates that Trump made substantial gains among the roughly 10% of the electorate who are immigrant voters, helping Republicans secure the White House (Frimpong and Sanchez, Reference Frimpong and Sanchez2024). A statistical report in one major media outlet found that the higher the foreign-born share of a US county’s population, the more likely it was to swing to Trump in the election, leading to the conclusion that “the people changing their minds were immigrants themselves” rather than native-born residents (The Economist, 2024, emphasis added). Whether these trends continue will depend on a range of factors, including how immigrant communities react to the Trump administration’s aggressive crackdown on unauthorized immigration.
For now, at least, the national narrative on migrants’ partisan leanings may be starting to change. For example, in one postelection podcast, New York Times columnist Lydia Polgreen opined: “I think that this idea that immigrants automatically just kind of filter into the Democratic coalition, I think, has just been absolutely and completely disproved” (Colette et al., Reference Colette, Douthat, Lozada and Polgreen2024). Elsewhere, a Time opinion piece argued the election showed that “Democrats are increasingly struggling to connect with immigrant…voters” (Touchan, Reference Touchan2024). Whether such elite rhetoric continues, and whether it changes mass perceptions about migrant political views, is an open question. If so, and more Americans learn that immigrants are no longer steadfast Democrats, differences in immigration preferences across party lines should attenuate.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/XPS.2025.10013.
Data availability
The data, code, and any additional materials required to replicate all analyses in this article are available in McDowell and Steinberg (Reference McDowell and Steinberg2025) at The Journal of Experimental Political Science Dataverse within the Harvard Dataverse Network, at: https://doi.org/10.7910/DVN/IBYKME.
Competing interests
The authors have no conflicts of interest to state.
Ethics statement
Project approved by Johns Hopkins University IRB (HIRB00021176) under the title: “American Public Opinion on Tariffs, Currencies, and Immigration.”
Our research adheres to APSA’s Principles and Guidance for Human Subjects Research.
Open access
