Immigrants and healthcare

Social determinants of health, which include legal status, have been shown to effect both levels of healthcare access and health outcomes for individuals without legal status and their families in the United States (Adler and Rehkopf, Reference Adler and Rehkopf2008; Castañeda et al., Reference Castañeda, Holmes, Madrigal, Young, Beyeler and Quesada2015; Footracer, Reference Footracer2009; Hacker et al., Reference Hacker, Anies, Folb and Zallman2015; Khan et al., Reference Khan, Velazquez, O’Connor, Simon and De Groot2011; Martinez et al., Reference Martinez, Wu, Sandfort, Dodge, Carballo-Dieguez, Pinto, Rhodes, Moya and Chavez-Baray2015; Vargas Bustamante et al., Reference Vargas Bustamante, Fang, Garza, Carter-Pokras, Wallace, Rizzo and Ortega2012; Vargas and Ybarra, Reference Vargas and Ybarra2017; Zamora et al., Reference Zamora, Kaul, Kirchhoff, Gwilliam, Jimenez, Morreall, Montenegro, Kinney and Fluchel2016). Partly in acknowledgement of the role that legal status serves as a barrier to healthcare access and positive health outcomes, the Director of US Immigration and Customs Enforcement (ICE) issued the “Enforcement Actions at or Focused on Sensitive Locations” memorandum in 2011 (Morton, Reference Morton2011). This policy discouraged immigration enforcement actions in and around specific locations such as healthcare facilities and schools (Morton, Reference Morton2011). In 2021, the policy was updated and expanded (Mayorkas, Reference Mayorkas2021), again acknowledging the role that legal status can play in deterring actions such as accessing healthcare and demonstrating a commitment to keep hospitals and clinics free from the politics of immigration. The goal was to allow undocumented immigrants the ability to seek healthcare without fear of arrest and deportation (Mayorkas, Reference Mayorkas2021; Morton, Reference Morton2011).

The policy environment surrounding immigration has shifted significantly in recent years, particularly since the inauguration of the second Trump Administration. Following up on significant anti-immigrant rhetoric during the presidential campaign (Woolhandler et al., Reference Woolhandler, Himmelstein, Ahmed, Bailey, Bassett, Bird, Bor, Bor, Carrasquillo and Chowkwanyun2021), the Trump Administration has implemented a slew of policies that negatively affect immigrants, particularly those without legal status (Siegel et al., Reference Siegel, Natanson and Meckler2025; Trump, Reference Trump2025). These policy changes also included ending the sensitive locations policy via executive order. This follows instances from the first Trump Administration where immigrants were arrested when seeking care including during emergency care, ambulance transfers, and labour despite the policy still technically being in place (Sánchez and Chang, Reference Sánchez and Chang2025).

States have responded to the changing political environment based on predictable partisan dispositions (Fortiér, Reference Fortiér2025; Sánchez and Chang, Reference Sánchez and Chang2025). On the one hand, many Republican states have actively embraced the types of policy changes initiated by the Trump Administration. Indeed, two states (Texas and Florida) had already gone a step further and passed state-level policies which instruct healthcare institutions to gather information on the legal status of their patients (Sánchez and Chang, Reference Sánchez and Chang2025). The Florida legislation, SB1718, which was signed into law in 2023, requires hospitals that accept Medicaid to ask patients if they are lawfully present in the United States (Arranda and Molina, Reference Arranda and Molina2024). The Texas law, which was enacted in 2024, requires all hospitals in Texas to ask this same question, aggregate the data, and report it to the Texas Health and Human Services Commission (HHSC) beginning in January 2026 (Figueiredo, Reference Figueiredo2024).

While previous work has argued that the federal sensitive locations policy minimises the risk of documenting legal status in healthcare settings (Kim et al., Reference Kim, Molina and Saadi2019), the removal of this protection raises new questions about such documentation. And while the stated purpose of both bills is to determine the financial costs that hospitals incur from providing healthcare to individuals without legal status, previous research suggests that it is likely to deter immigrants – with and without legal status – from seeking needed medical care (Berlinger and Gusmano, Reference Berlinger and Gusmano2013; Blackburn and Sierra, Reference Blackburn and Sierra2021). However, not all states equally support these types of policy changes. As a result, the policies implemented in Florida and Texas differ substantially from states like California and Massachusetts which are advising healthcare providers to avoid documenting immigration status and encouraging healthcare providers to advertise that no information on legal status will be collected for patients (Fortiér, Reference Fortiér2025; Sánchez and Chang, Reference Sánchez and Chang2025). In these states, healthcare providers have also been advised that they are not required to assist immigration enforcement but that they also should not interfere with any investigations (Fortiér, Reference Fortiér2025).

The federal level policy change that removed protections from immigration enforcement in healthcare settings, combined with state-level policy that requires healthcare providers to actively collect information on legal status, represent a substantial departure from policy precedent (Kim et al., Reference Kim, Molina and Saadi2019; Lamneck et al., Reference Lamneck, Alvarez, Zaragoza, Rahimian, Trejo and Lebensohn2023; Sconyers and Tate, Reference Sconyers and Tate2016). Particularly, the state-level laws on immigration status reporting by healthcare facilities raise several questions about healthcare access, respect for persons, and justice that have implications far beyond the two states who have implemented this legislation (Lively, Reference Lively2024). And while healthcare providers across the United States are emphasising that individuals will not be turned away from care (Figueiredo, Reference Figueiredo2024; Sánchez and Chang, Reference Sánchez and Chang2025), there are reasons to believe that these policies will have a chilling effect on healthcare access for those without status and their families (Abraham, Reference Abraham2018). This particularly holds as there is also evidence that these policy changes have created a climate of fear and distrust among immigrant families (Arranda and Molina, Reference Arranda and Molina2024; Pillai et al., Reference Pillai, Artiga and Rae2025).

Research also indicates that increasingly restrictive laws and enforcement can have broad implications for non-citizens, their families, and their communities (Cruz Nichols et al., Reference Cruz Nichols, LeBrón and Pedraza2018; Nichols et al., Reference Nichols, LeBrón and Pedraza2018; Pillai et al., Reference Pillai, Artiga and Rae2025; Torche and Sirois, Reference Torche and Sirois2019), particular for healthcare (Abraham, Reference Abraham2018; Martinez et al., Reference Martinez, Wu, Sandfort, Dodge, Carballo-Dieguez, Pinto, Rhodes, Moya and Chavez-Baray2015; Philbin et al., Reference Philbin, Flake, Hatzenbuehler and Hirsch2018; Rhodes et al., Reference Rhodes, Mann, Simán, Song, Alonzo, Downs, Lawlor, Martinez, Sun and O’Brien2015; White et al., Reference White, Yeager, Menachemi and Scarinci2014). Importantly, there is evidence to suggest that punitive policies can have significantly negative effects on undocumented immigrants and their families even when not implemented or enforced (Torche and Sirois, Reference Torche and Sirois2019).

These policy changes also present a slew of problems for healthcare providers (Kim et al., Reference Kim, Molina and Saadi2019; Lamneck et al., Reference Lamneck, Alvarez, Zaragoza, Rahimian, Trejo and Lebensohn2023; Sconyers and Tate, Reference Sconyers and Tate2016). To be sure, there are some benefits associated with documenting legal status because it can guide treatment and allow for a broader recognition of social determinants of health (Kim et al., Reference Kim, Molina and Saadi2019). However, most outcomes of requiring documentation would likely be detrimental to the health and well-being of individuals. For one, documenting legal status may affect the care experience of individuals because patients may be offered lower quality of care from clinicians with anti-immigrant beliefs (Kim et al., Reference Kim, Molina and Saadi2019) and because of potential stigmatisation (Kim et al., Reference Kim, Molina and Saadi2019). At its worst, documenting this type of information may render individuals subject to detainment, as there is some evidence that medical staff may have notified immigration officials when encountering individuals without legal status (Kim et al., Reference Kim, Molina and Saadi2019). Moreover, immigration enforcement may ignore and violate confidentiality laws (Lamneck et al., Reference Lamneck, Alvarez, Zaragoza, Rahimian, Trejo and Lebensohn2023). Documentation also opens the door for future efforts by immigration enforcement initiatives to utilise the information to detain immigrants as has been observed in other countries (Kim et al., Reference Kim, Molina and Saadi2019). Importantly, there are also broader historic concerns about physicians and other healthcare providers acting as “agents of the state” (Spevick, Reference Spevick2002). Lastly, these changes raise questions about whether future iteration of these policies go beyond merely documenting immigration status or tracking the number of undocumented individuals served for healthcare providers. That is, could healthcare providers be required to report individuals without status to immigration enforcement or even actively assist immigration authorities?

Treatment effects

There is evidence in the literature that the framing of political information has significant effects on attitudes (Kahneman and Tversky, Reference Kahneman and Tversky1979; Norris, Reference Norris2021; Tversky and Kahneman, Reference Tversky and Kahneman1981). With regards to immigrants and immigration policy, framing becomes more important because most Americans hold positive views of immigrants even though most Americans also state that they support politicians who produce negative messages about immigrants (Asbury-Kimmel, Reference Asbury-Kimmel2023; Gallup, 2025). Importantly, there is evidence that Americans are not fully aware of the experiences of individuals without legals status and their role and contributions to US society and economy (Artiga, Reference Artiga2024; Bryant, Reference Bryant2024; Davis et al., Reference Davis, Guzman and Sifre2024). For example, Americans tend to overestimate the number of undocumented immigrants while underestimating their economic contributions (Davis et al., Reference Davis, Guzman and Sifre2024). Moreover, perceptions that individuals without legal status burden public services are widespread (Artiga, Reference Artiga2024; Davis et al., Reference Davis, Guzman and Sifre2024). Economic, cultural, and political frames have all been shown to have an effect on Americans’ attitudes towards immigration policy (Ayasli, Reference Ayasli2024; Kuntz et al., Reference Kuntz, Davidov and Semyonov2017; Mayda, Reference Mayda2006; Pardos-Prado and Xena, Reference Pardos-Prado and Xena2019). However, economic factors often overshadow other factors when Americans’ form attitudes about immigrants (Aalberg et al., Reference Aalberg, Iyengar and Messing2012). Additionally, previous research suggests that Americans are generally less supportive of government action aimed at protecting the rights of individuals without legal status in the United States (Voss et al., Reference Voss, Silva and Bloemraad2020), but even so, they show decreasing levels of support for immigration policy that prevents those without legal status from accessing necessary healthcare (Blackburn and Haeder, Reference Blackburn and Haeder2024). Given these facts, providing individuals with object facts and data about undocumented immigrants may decrease support for policies detrimental to them and their families. Hence, we hypothesise the following:

We note that this effect may be particular strong in cases that rely on negative framing (as we do in our treatments), which has been shown to be more powerful and salient than positive framing (Avdagic and Savage, Reference Avdagic and Savage2021; Baumeister et al., Reference Baumeister, Bratslavsky, Finkenauer and Vohs2001; Rozin and Royzman, Reference Rozin and Royzman2001; Soroka, Reference Soroka2014).

At the same time, it seems likely that respondents will be more concerned about potential policy effects that may affect themselves (ie. their community) as compared to those solely focused on the policy effects of the individuals without status. Indeed, there is a large literature highlighting the importance of personal policy benefits (Bechtel and Liesch, Reference Bechtel and Liesch2020; Hansford and Gomez, Reference Hansford and Gomez2015; Kiewiet and Lewis-Beck, Reference Kiewiet and Lewis-Beck2011; Kinder and Kiewiet, Reference Kinder and Kiewiet1981). Put differently, respondents will likely be less supportive of policies that are closer to them (ie. more likely to affect them) as compared to those that are more distant (Arnold, Reference Arnold1992; Chong et al., Reference Chong, Citrin and Conley2001; Holbrook et al., Reference Holbrook, Sterrett, Johnson and Krysan2016). Hence:

Administrative tasks versus active immigration enforcement

While we expect some general effects across all policies, there may be important differences based on the nature of the policy and what is expected of the healthcare provider. Specifically, we expect that not all policies will elicit similar responses from respondents because they differ in the type of action required by healthcare providers as well as how closely involved these healthcare providers get with actual immigration enforcement. As noted above, all five policies affect healthcare providers because, to one degree or another, they move them outside of their care-providing role and associates them with immigration policy and enforcement (Kim et al., Reference Kim, Molina and Saadi2019; Lamneck et al., Reference Lamneck, Alvarez, Zaragoza, Rahimian, Trejo and Lebensohn2023; Sconyers and Tate, Reference Sconyers and Tate2016). The varying degrees of moving healthcare providers into roles that may make them “agents of the state” (Spevick, Reference Spevick2002) seem likely to affect public attitudes.

Specifically, we believe that respondents may make substantial distinction between tasks that most would consider mostly administrative in nature as compared to those with require them to become more active agents in the identification and detention of undocumented immigrations (Kim et al., Reference Kim, Molina and Saadi2019; Lamneck et al., Reference Lamneck, Alvarez, Zaragoza, Rahimian, Trejo and Lebensohn2023; Sconyers and Tate, Reference Sconyers and Tate2016). Administrative tasks like tracking and documenting are often considered to mere routine acts of record-keeping (Kim et al., Reference Kim, Molina and Saadi2019). These administrative acts allow for a substantial amount of distance between the administrative act and the potential active involvement in immigration enforcement. Moreover, administrative tasks are much more in line with behaviour and occupational expectations for healthcare workers as well as their social identity as providers of care. Lastly, administrative acts also offer an important semblance of neutrality and procedural justice where as active involvement may elicit a number of moral and ethical questions and minimise concerns about breaking the trust between doctor and patient (Ornelas-Dorian et al., Reference Ornelas-Dorian, Torres, Sun, Aleman, Cordova, Orue, Taira, Anderson and Rodriguez2021). That is, given the accepted role of healthcare providers as healers (Kim et al., Reference Kim, Molina and Saadi2019), respondents may likely be more comfortable supporting policies that appears as avoiding a direct involvement of healthcare providers in immigration enforcement tasks.

In short, respondents may be more supportive of policies that only require healthcare providers to fulfil administrative tasks like asking patients’ immigration status or marking the status on charts and bills as compared to those focused on reporting individuals or even actively assisting enforcement. We thus hypothesise the following:

Given that of the five policies, the requirement to assist immigrant enforcement serves as the clearest example of active involvement (as compared to administrative compliance) we expect that the results should be particularly stark in this case.

Partisanship & presidential vote choice

Given the increasing levels of polarisation and partisanship in the United States (Druckman et al., Reference Druckman, Klar, Krupnikov, Levendusky and Ryan2021a, Reference Druckman, Klar, Krupnikov, Levendusky and Ryan2021b; Van Bavel et al., Reference Van Bavel, Gadarian, Knowles and Ruggeri2024; Zingher, Reference Zingher2022), legal status is not only relevant as a social determinant of health, however. Political determinants of health, which result in unequal health outcomes as the result of deliberate political actions, is also relevant to healthcare access and outcomes for immigrants in the United States. Specifically, public health deteriorates as political polarisation increases and individuals become less likely to trust experts and more likely to trust party officials for health information (Haeder and Yackee, Reference Haeder and Yackee2025; SteelFisher et al., Reference SteelFisher, Findling, Caporello, Lubell, Vidoloff Melville, Lane, Boyea, Schafer and Ben-Porath2023; Van Bavel et al., Reference Van Bavel, Gadarian, Knowles and Ruggeri2024). Added to this context, decisions made by these political leaders then shape the conditions that make health more or less accessible to some portions of the population (Dawes, Reference Dawes2020). Political polarisation and the political determinants of health that can become exacerbated as a result, are particularly salient for those without legal status in the United States, a topic that has historically been highly controversial and only become more so in recent years (Anderson and Turgeon, Reference Anderson, Turgeon, Anderson and Turgeon2022; Blackburn and Haeder, Reference Blackburn and Haeder2024; Haeder and Moynihan, Reference Haeder and Moynihan2025; Kraut, Reference Kraut1995).

Political determinants of health manifesting through increased political polarisation can affect opinion formation and fluidity on issues that are highly polarised, such as immigration. Previous research argues that political parties are critical to policy-decision making among citizens (Leeper and Slothuus, Reference Leeper and Slothuus2014) and the elites of those parties are responsible for framing and influencing the opinions of party members on particular political issues (Druckman et al., Reference Druckman, Klar, Krupnikov, Levendusky and Ryan2021a). This deference to party elites contributes to polarisation on issues that may have traditionally been considered apolitical (Druckman et al., Reference Druckman, Klar, Krupnikov, Levendusky and Ryan2021b) and partisan issues become increasingly polarised as individuals start to villainise the people and opinions of the opposing party (Bankert, Reference Bankert2024). This literature suggests that American views on immigration policy are particularly vulnerable to the framing of elites and extreme polarization.

In the current hyper-partisan political environment, partisanship and vote choice serve as crucial determinants of policy positions (Campbell et al., Reference Campbell, Converse, Miller and Stokes1960; Downs, Reference Downs1957; Mondak, Reference Mondak1993; Schaffner and Streb, Reference Schaffner and Streb2002; Zaller, Reference Zaller1992). Additionally, Zingher (Reference Zingher2022) argues that much of the hyper-partisan environment is driven by political elites. Within this framing, voting is driven by these elites seeking to mobilise their most dedicated and extreme voters (Zingher, Reference Zingher2022). Therefore, the political messaging for everything from singular policy issues to overall presidential approval rates is driven largely by the partisan framing of elites and the support this drives from the most extreme members of the voter base provides little incentive for elites to moderate their messaging (Harrison, Reference Harrison2016; Robison, Reference Robison2023). There is substantial evidence that this is the case for Americans’ views of immigrants (Alamillo et al., Reference Alamillo, Haynes and Madrid2019; Chenane, Reference Chenane2022; Dzordzormenyoh and Boateng, Reference Dzordzormenyoh and Boateng2023; Knoll et al., Reference Knoll, Redlawsk and Sanborn2011; Merolla et al., Reference Merolla, Ramakrishnan and Haynes2013; Reich and Mendoza, Reference Reich and Mendoza2008; Sagir and Mockabee, Reference Sagir and Mockabee2023). This effect has increased over time (Baker and Edmonds, Reference Baker and Edmonds2021). We expect this to particularly hold true for such a salient issue as policies affecting individuals without legal status (Blackburn and Haeder, Reference Blackburn and Haeder2024). Moreover, issues related to health policy have proven particularly partisan (Bussing et al., Reference Bussing, Patton, Roberts and Treul2020; Haeder and Chattopadhyay, Reference Haeder and Chattopadhyay2022; Haeder and Moynihan, Reference Haeder and Moynihan2023; Jacobs and Mettler, Reference Jacobs and Mettler2020; Oberlander, Reference Oberlander2020; Sances and Clinton, Reference Sances and Clinton2021; Wang, Reference Wang2022). Unsurprisingly, conservatives and Republicans have been identified as more anti-immigrant than liberals and Democrats in a number of different circumstances (Cadena Jr., Reference Cadena2023; Dzordzormenyoh and Boateng, Reference Dzordzormenyoh and Boateng2023; Willnat et al., Reference Willnat, Ogan and Shi2023). These effects may be particularly strong based not only on traditional partisanship but also whether a respondent voted for President Trump in the most recent presidential election given the policy positions of the candidates (Alamillo et al., Reference Alamillo, Haynes and Madrid2019; Sagir and Mockabee, Reference Sagir and Mockabee2023; Woolhandler et al., Reference Woolhandler, Himmelstein, Ahmed, Bailey, Bassett, Bird, Bor, Bor, Carrasquillo and Chowkwanyun2021). We hence hypothesise that:

Personal connection to individuals without legal status

There is limited research on the way that personal connections with undocumented individuals influence attitudes towards policies related to health, but there is more substantial literature examining how personal connections influence attitudes towards immigration policy more broadly. These studies find that having a personal connection to an individual who lacks legal status or who has been deported, increases the salience of immigration policies and political engagement (Roman et al., Reference Roman, Walker and Barreto2022; Sanchez et al., Reference Sanchez, Vargas, Walker and Ybarra2015). However, there is also evidence that personal connections play an important role in shaping attitudes about public policies, particularly as they are related to health. Personal connection may offer a better understanding of the policy effects on individuals, and they may encourage a more empathetic outlook (Addison and Thorpe, Reference Addison and Thorpe2004; Corrigan et al., Reference Corrigan, Kuwabara and O’Shaughnessy2009; Sattler et al., Reference Sattler, Escande, Racine and Göritz2017). For example, research by Grogan and Park (Reference Grogan and Park2017, Reference Grogan and Park2018) indicates that a connection to Medicaid increased support for the programme. At the same time, those with connections have also been found to be less tolerant of administrative burdens that reduce access (Haeder and Moynihan, Reference Haeder and Moynihan2024; Halling et al., Reference Halling, Herd and Moynihan2022). Other research suggest that personal connection may affect attitudes about other health issues (Haeder et al., Reference Haeder, Herd and Moynihan2025; Haeder et al., Reference Haeder, Sylvester and Callaghan2021; Sylvester et al., Reference Sylvester, Haeder and Callaghan2022). Lastly, personal connections are likely to provide individuals with direct information about the hardships experienced by undocumented individuals while reducing the effect of misinformation about them. We hence hypothesise the following:

Residence in border states

The literature also indicates that there are notable differences in attitudes towards immigrants based on geographic location within the United States. Studies have found that living in proximity to immigrant groups affects attitudes towards immigrants, though the directionality of this effect is mixed (Abrajano & Hajnal, Reference Abrajano and Hajnal2015; Berinsky et al., Reference Berinsky, Karpowitz, Peng, Rodden and Wong2023; Chan et al., Reference Chan, Raychaudhuri and Valenzuela2023; Hopkins, Reference Hopkins2010). Several studies found living in geographic locations with higher numbers of immigrants and their families leads to an increase in pro-immigrant attitudes (Berinsky et al., Reference Berinsky, Karpowitz, Peng, Rodden and Wong2023; Chan et al., Reference Chan, Raychaudhuri and Valenzuela2023), while other studies have found that living near immigrant populations is associated with negative attitudes of immigrants (Abrajano & Hajnal, Reference Abrajano and Hajnal2015; Campbell et al., Reference Campbell, Wong and Citrin2006; Enos, Reference Enos2016; Ha, Reference Ha2010; Hopkins, Reference Hopkins2010). For non-Hispanic Whites, living near Hispanic immigrants makes those individuals more likely to hold negative stereotypes (Ha, Reference Ha2010). However, there is no consensus in the literature about how living in geographic proximity to immigrants affects support for restrictive immigration policies. Some studies have found that it influences views of immigration policies (Campbell et al., Reference Campbell, Wong and Citrin2006; Newman et al., Reference Newman, Velez, Hartman and Bankert2015), while others have shown it has little effect (Hopkins et al., Reference Hopkins, Sides and Citrin2019; Swire-Thompson et al., Reference Swire-Thompson, Ecker, Lewandowsky and Berinsky2020). However, sudden changes in the number of immigrants in a population or community increases hostile views of immigrants (Hopkins, Reference Hopkins2010). Based on the literature related to geographic location and attitudes towards immigrants and immigrant policy as well as the substantial increase in undocumented immigrants along the southern border of the United States (CA, NM, AZ, TX), it seems plausible that attitudes of individuals living in border states may differ from those in other geographic areas. However, as the literature is mixed in this regard, we do not have a clear hypothesis.

Data

To assess public attitudes about requirements for healthcare providers related to individuals who lack legal status, we programmed an original survey in Qualtrics and relied on Lucid to recruit respondents. Lucid hosts a large, online, double opt-in panel and offers respondents various incentives to complete surveys. Lucid was compensated $1.50 per completed response. Data generated from Lucid’s pool has been validated and is generally considered to be of high quality (Coppock and McClellan, Reference Coppock and McClellan2019; Kennedy et al., Reference Kennedy, Mercer, Keeter, Hatley, McGeeney and Gimenez2016; Stagnaro et al., Reference Stagnaro, Druckman, Berinsky, Arechar, Willer and Rand2024). It has been frequently used in work related to social science, health, and politics (Stagnaro et al., Reference Stagnaro, Druckman, Berinsky, Arechar, Willer and Rand2024). They survey was fielded from 7 March to 26 March 2025. Overall, 6948 US adults opted into the survey (18 years or older) and 6651 consented to take the survey. Of these, 6049 completed the survey. Lucid has implemented a pre-survey attention screener to ensure data quality. Lucid relies on quota sampling to ensure a sample that approaches national population benchmarks on gender, race, income, and education. We further improved fit by utilising weights based on the US Census Current Population Survey (for further details see Appendix Exhibits 1 & 2). This study was declared exempt by the appropriate Institutional Review Board.

Survey design and dependent variables

Methods

The data for the analyses were generated from a survey experiment where respondents were randomly assigned to one of six treatment groups or a control group (Appendix 2). As a result, we estimated a series of weighted least square models (WLS), one for each of our five outcomes (policies), that included an indicator for the treatment group utilising the aforementioned survey weights to make comparisons across treatments. For our subgroup analyses, we interacted the treatment indicator with indicators for partisanship, presidential vote, residence in a border state, or personal connection to a person without legal status, respectively. We then relied on comparisons of predicted means to assess statistical differences using mlincom in Stata (Cameron and Trivedi, Reference Cameron and Trivedi2010; Long and Freese, Reference Long and Freese2014). We note that WLS estimation over ordered logit models to facilitate comparisons, interpretation, and presentation of results, as is a common approach. We considered a p-value lower than 0.05 as statistically significant throughout our analyses.