Do Anti-immigration Voters Care More? Documenting the Issue Importance Asymmetry of Immigration Attitudes

Abstract Why do politicians and policymakers not prioritize pro-immigration reforms, even when public opinion on the issue is positive? This research note examines one previously overlooked explanation related to the systematically greater importance of immigration as a political issue among those who oppose it relative to those who support it. To provide a comprehensive empirical assessment of how personal immigration issue importance is related to policy preferences, I use the best available cross-national and longitudinal surveys from multiple immigrant-receiving contexts. I find that compared to pro-immigration voters, anti-immigration voters feel stronger about the issue and are more likely to consider it as both personally and nationally important. This finding holds across virtually all observed countries, years, and alternative survey measures of immigration preferences and their importance. Overall, these results suggest that public attitudes toward immigration exhibit a substantial issue importance asymmetry that systematically advantages anti-immigration causes when the issue is more contextually salient.

In the summer of 2020, Gallup reported that for the first time since the organization started their polling on the issue in 1965, more Americans said they would prefer to see immigration increased rather than decreased (34 per cent versus 28 per cent) (Younis 2020). Many news outlets have subsequently reported this finding, which seemed to be especially hopeful in the midst of the global COVID-19 pandemic, with foreign origin. Although these news reports were generally optimistic, many commentators also wondered why these positive public attitudes have not translated into increased pressure on the US government to enact a comprehensive immigration reform. Indeed, why do politicians and policymakers rarely prioritize pro-immigration appeals and policies, even in contexts where public opinion on the issue is seemingly positive? 1 This research note empirically assesses one possible behavioral explanation of this puzzle related to the systematically greater perception of immigration as an important issue among those voters who oppose it (relative to those who support it). Although both anti-immigration and pro-immigration advocates can be extremely motivated by their cause (Kustov 2021), it is striking that while there have been many prominent single-issue political parties and politicians with an anti-immigration platform across the world, there have been none with a predominantly pro-immigration platform. However, despite the fact that any account of immigration politics must make an assumption about whether immigration preferences are equally important to all voters, there has been no empirical evidence on this question so far.
To remedy this omission, I provide the first comprehensive empirical assessment of the relationship between the personal issue importance (PII) of immigration to voters and their preferences on the issue across various high-income, immigrant-receiving democratic contexts. 2 First, I use multiple cross-sectional samples of the American National Election Studies (ANES) and the British Election Study (BES) for the baseline test of a potential positive correlation between anti-immigration preferences and the perception of immigration as the most important national problem or issue. After establishing the expected issue importance asymmetry in the major US and UK election surveys, I utilize the Transatlantic Trends Survey (TTS) and show that this asymmetry similarly extends to other major immigrant-receiving countries. I then use additional BES, Voter Study Group (VSG), Institute for the Study of Citizens and Politics (ISCAP), and Eurobarometer data to establish that this asymmetry holds regardless of the particular survey instrument. Finally, I compare the relationship between PII and preferences in immigration to that of other political issues and political interests in general using the Cooperative Election Study (CCES), indicating that the ideological patterns of the revealed asymmetry are rather idiosyncratic to immigration. In sum, I find that compared to those who support immigration, those who oppose immigration feel stronger about the issue and more likely to consider it as both personally and nationally important across virtually all observed countries, years, and alternative survey measures.
Overall, these results suggest that the pro-immigration preferences observed in public opinion surveys may often be not as strong as anti-immigration preferences in terms of people's behavioral or cognitive engagement with the issue and their willingness to vote based on it. In other words, even when the public support of pro-immigration policies is seemingly greater than or similar to that of anti-immigration policies in the raw poll numbers, it is likely the case that the anti-immigration side is still more politically motivated and influential. While the exact causes and consequences of this phenomenon are beyond the scope of this research note, these results also imply that public attitudes toward immigration exhibit a substantial issue importance asymmetry that systematically advantages anti-immigration causes when the issue is more contextually salient in a particular country or time.

Issue Importance Asymmetry of Immigration Attitudes
The literature on issue importance or salience is central to the understanding of any public attitudes toward government policies and their role in politics more generally. When individuals are said to attach personal importance to a certain policy issue, they are concerned or care about it, as manifested by their cognitive and behavioral engagement with that issue. This includes thinking frequently and deeply about it, gathering information about it, and using the issue as a basis for making voting and other political decisions (Miller, Krosnick, and Fabrigar 2017). Given that any engagement is costly for individuals in terms of their time and other limited resources, most scholars consider PII to be a relative (constant-sum) concept. Defined this way, many political outcomes, including those related to immigration, are significantly dependent on how much relative importance people attach to various issues (Dennison 2019). 2 I use the term "personal issue importance," as opposed to the related term "issue salience," throughout the text due to the greater ambiguity of the latter. While these terms are often used interchangeably at the individual level, some scholars rightly differentiate between them by defining salience as a broader function of (relatively stable) PII and (relatively volatile) political contexts, which may or may not frame a certain issue as a political problem at the moment (see Miller, Krosnick, and Fabrigar 2017;Moniz and Wlezien 2021). To avoid confusion, I thus use the term "contextual issue salience" to denote an objective relative importance of the issue as captured in the political discourse or media coverage of a particular country and time. In line with the previous literature, I also define "preferences" as any rankings derived from comparative evaluations of various policies and use "attitudes" as an umbrella concept for personal preferences and issue importance.
Unfortunately, until very recently, the literature on immigration attitudes has developed separately from this discussion. As a result, scholars have not seriously considered the theoretical implications and empirical reality concerning the interplay between immigration policy preferences and PII in determining electoral and other political outcomes. They have not identified when individuals are likely to develop their concerns about immigration issues and whether these concerns persist (independent of preferences). This is disconcerting given that any account of immigration politics that relates certain individual preferences or beliefs to other outcomes must make an assumption about the underlying (variation in) PII. After all, it is people for whom a particular policy issue is important who are likely to place great weight on it when deciding how to vote or meaningfully express their views otherwise, whether by donating money to relevant organizations, volunteering, contacting public officials, or even running for office themselves (Miller, Krosnick, and Fabrigar 2017;Moniz and Wlezien 2021).
How does PII (or preferences across issues) form and how is it different from the formation of comparative evaluations or policy preferences within issues? While the exploration of how individuals decide what issues to care about is still in its nascent stage, the literature usually singles out three major systematic individual-level sources related to material self-interest, group identity, and psychological predispositions (Miller, Krosnick, and Fabrigar 2017). Notably, all of these factors are also used to explain policy preferences within issues. Consequently, although relative PII 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 (see, for example, see Jennings and Wlezien 2015). 3 More recently, some scholars have tried to bring PII (and salience) to the center of immigration politics research. Hatton (2021), for instance, extensively shows that immigration issue importance is a meaningful construct that has distinct correlates with preferences at both individual and country levels. At the same time, Dennison and Geddes (2019) show that, at least at the aggregate level, political preferences are highly stable, while PII is fairly stable and national issue importance is fairly volatile. Kustov, Laaker, and Reller (2021) further confirm these patterns in individual-level panel data, showing that many of the prominent exogenous shocks that have often been thought to impact immigration preferences in the electorate (such as economic and refugee crises) are more likely to be changing the PII of immigration to individuals. In line with these results, Dennison (2020) also shows that it is increases in immigration issue importance (rather than preferences) that are positively related to populist voting at both aggregate and individual levels.
Due to the lack of relevant data, however, the relationship between immigration PII and preferences has not yet been systematically explored. To that end, this short article aims to identify the best available surveys and provide a comprehensive empirical assessment of this relationship across countries, years, and various PII measurements. Although the study does not stipulate a formal theoretically informed hypothesis, given the relative failure of immigration advocates to enact significant pro-immigration reforms, even in seemingly immigration-friendly electoral contexts, a reasonable prior expectation is: those who are generally opposed to immigration should be more likely to consider it an important issue than those who are generally supportive of immigration.

Data, Measurement, and Methods
Since an unambiguously effective survey measure of PII still remains elusive (Jennings and Wlezien 2011; Moniz and Wlezien 2021), 4 my goal was to identify all publicly available 3 It should be noted that PII asymmetries should, by definition, be present in so-called "valence" issues like unemployment or economic growth (as opposed to "position" issues like immigration or abortion), in which the political disagreement is mostly about the importance of a certain issue for government policy (see, for example, Green 2007). 4 The most common survey instrument to measure personal PII are the so-called "most important problem" (MIP) or the nearly identical "most important issue" (MII) questions, which ask respondents to self-report one or a few important political (quasi-)representative survey datasets with the greatest diversity of relevant self-reported measures on immigration preferences and PII. To minimize measurement error, in selecting the data, I gave a particular priority to higher-quality surveys with more valid items, larger or more numerous samples, and longitudinal as opposed to just cross-sectional designs. I also prioritized surveys with items that specifically highlight personal issue importance (as opposed to just perceptions of "most important national problem") and clearly defined immigration policy preferences (as opposed to just general feeling or beliefs about immigration and immigrants). Since, I need to have some relevant variation in both attitude dimensions and sufficient statistical power to examine the relationship between PII and preferences, I only considered democratic survey contexts with at least 40 individuals (that is, normally 4 per cent of the sample) reporting that they consider immigration an important issue.
The overall data/measurement summary is given in Table A1 in the Online Appendix. For ease of interpretation, in the main text, I mostly highlight the results based on the binary transformations of survey items: share of respondents who consider immigration an important issue among those who want to decrease/restrict and increase/ease immigration. All variables were standardized from 0 to 1 (so that 1 indicates either important or anti-immigration issue preferences). 5

Analysis and Results
I start my baseline analysis with an examination of ANES, which arguably provides the best quality survey data with consistent items on both immigration policy preferences (measured as preferred immigration levels) and issue importance (measured as the somewhat imperfect national MIP). As can be seen from Figure 1, anti-immigration US voters are generally more likely to think of it as important, and this pattern is especially apparent during times of high contextual salience of immigration in the country. In the aftermath of Donald Trump's win in 2016, for instance, those who wanted to decrease immigration were much more likely to consider it the most important problem facing their country (27 per cent versus 16 per cent). At the same time, in 2012, when immigration was less salient nationally, the same asymmetry was 4 per cent versus 2 per cent (the difference was similar but not statistically significant in 2008). When one considers the subsequent post-election ANES survey in 2020-shortly after Gallup declared the net positivity of public attitudes for the first time-the comparable numbers were 17 per cent versus 12 per cent. In other words, despite the fact that more people now say they have pro-immigration preferences, they are relatively less engaged than those who have anti-immigration preferences.
Another way to look at it is to count the percentage of ANES respondents who support or oppose immigration only among those who also consider it the most important national issue, which arguably provides a better metric of the potential pro-immigration or anti-immigration pressure groups in US public opinion. This way, even in the United States of 2020-one of the most pro-immigration preference contexts recorded in polling history-there were possibly issues they think their country is facing at the moment (usually in an open-ended form). These measures have received substantial criticism due to the common conflation of the importance of a certain issue with the degree to which it is currently a (negative) "problem" or an "issue" by respondents (Wlezien 2005). Moreover, national MII and MIP items have also been shown to be better at revealing respondents' guesses about what other people think is important in their country, rather than measuring PII (Jennings and Wlezien 2011). Still, however imperfect, these items can be used to measure voters' issue attentiveness and their priorities for government action relative to all other issues (Jennings and Wlezien 2015). A major alternative survey instrument to measure PII is a direct question that asks respondents to self-report whether, and to what extent, a certain policy issue is important to them personally. While this addresses the main limitation of MII/ MIP items, it is less suited to measure relative importance because respondents can report that all issues are equally and highly important to them (Moniz and Wlezien 2021). 5 For a more complete assessment based on the indices comprised of multiple variables, and including those who are neutral regarding immigration (a closed interval of 0.25-0.75), with no change in substantive results, see the Online Appendix. The results are also robust to the inclusion of basic control variables, including general political interest, age, gender, education, urban residence, immigration status, and ideology (not shown). fewer engaged pro-immigration than anti-immigration voters (34 per cent × 12 per cent ≈ 4 per cent versus 28 per cent × 17 per cent ≈ 5 per cent of all respondents). For comparison, during the time of high contextual salience of immigration in 2016, the numbers were approximately 2 per cent versus 12 per cent of all respondents, indicating a very large public opinion skew in favor of the anti-immigration cause.
Importantly, however, these stark differences are not just about the 2016 election of Donald Trump. The contextual salience of immigration may come and go (and it may be hard to detect the asymmetry when only a few people mention immigration as an important issue). Nonetheless, the relative asymmetry in the explicitly personal self-reported issue importance of immigration appears to have been stable throughout the last decade, as captured using alternative measures in the US VSG or ISCAP data (see Figures A1 and A4 in the Online Appendix).
I then move to the analysis of BES, which consistently measures issue importance (as MII) and various immigration preferences (though as different items, depending on the particular survey). Unlike the relatively immigration-friendly US context, the UK has exhibited much higher degrees of negativity toward immigration than any other country over the last two decades, as well as perhaps its highest contextual salience. For instance, in March 2015, or approximately one year before the Brexit referendum, 73 per cent of the British public wanted to decrease immigration, while only 6 per cent wanted to increase it. Meanwhile, a staggering 26 per cent thought it was the single most important issue facing their country (in July 2015, this number rose to 35 per cent). The question, however, is whether issue preferences and importance are empirically related to each other. As can be seen from Figure 2, they indeed very much are-those who oppose immigration are consistently much more likely to consider it important than those who support it, and these differences are even higher than in the US (from 27 per cent versus 7 per cent in 2005, to 48 per cent versus 6 per cent in 2015, and then back to 23 per cent versus 2 per cent in 2017).
Given that the UK and especially the US are among the biggest and oldest immigrant destinations, it may also be important to consider some of the newer immigrant-receiving contexts. To that end, I use the cross-national TTS data from 2014 (see also the additional analysis of the Eurobarometer data from 2019 in Figure 3a). According to Figure 4, the previously uncovered asymmetry is further evident in every other immigrant-receiving country in the dataset, from France and Germany, to Denmark and Sweden. The only possible exceptions are Hungary and Romania-countries with very low immigration (at least until very recently)-where the differences in issue salience between pro-immigration and anti-immigration voters are not statistically significant. Note: The bars represent 95 per cent confidence intervals. For variable descriptions, see Table A1 in the Online Appendix. Source: Based on ANES data.
In line with the prominent critique by Wlezien (2005), however, one may reasonably wonder whether this result is merely an artifact of the particular MII/MIP question. If these survey items indeed conflate the importance and "problemness" of immigration in the minds of many respondents, it is possible that the preceding results simply show that those who oppose immigration are more likely to think of it as a (negative) "problem," which is nearly tautological.
Luckily, in addition to such MII/MIP items, the BES and a few other surveys also provide alternative operationalizations of PII that arguably do not have this limitation, allowing for the direct comparison of various survey measurements. The BES 2015, for instance, also asked their respondents: "how strongly do you feel about [immigration]?" (see Figure 2). Quite similarly, the US VSG asked: "how important are the following issues [such as immigration] to you?" (see Figure A1 in the Online Appendix). As can be seen from these graphs (comparing BES 2015and BES 2015ALT and VSG 2011and VSG 2011, the same asymmetry of immigration attitudes is present regardless of the particular operationalization of PII. While the baseline rate of PII is understandably much higher in these alternative multiple-choice items (for example, respondents could pick that all political issues are very important to them), those who oppose immigration are still much more likely to report that the issue is important to them, and this basic finding holds across all samples in these datasets. 6 As a compromise between these two imperfect operationalizations, Eurobarometer-one of the major cross-national surveys in the European Union-also asks about the most important issues facing the country and the respondents personally. As Figure 3 shows, anti-immigration voters are, on average, two to three times more likely to think of it as both a more nationally and a more personally important issue. Interestingly, perceived national concerns about immigration are, on average, much more pronounced than self-reported personal concerns across all countries. 7 Overall, regardless of the particular operationalization, anti-immigration voters are, on average, more likely to say that immigration is among the most important issues. Notes: The bars represent 95 per cent confidence intervals. ALT stands for an alternative operationalization of PII; p stands for Internet Panel survey. For variable descriptions, see Table A1 in the Online Appendix. Source: Based on BES data.

6
As for more quasi-behavioral measures, the 2018 US ISCAP survey-which specifically asks people to report what issues are important to their vote-shows a similar asymmetric pattern (see Figure A4 in the Online Appendix). At the same time, my original quasi-representative UK sample (available upon request) indicates that anti-immigration voters report a greater willingness to sign a petition to share their views with Parliament (72 per cent versus 63 per cent). 7 This is in line with the idea that the national MII item is an imperfect function of PII combined with other factors, such as people's perceptions about what others consider important in a political context (Moniz and Wlezien 2021). In particular, the individual-level correlation between the two Eurobarometer measures is 0.2-0.3, depending on the particular country, such that the majorities of respondents who consider immigration personally important also report that they believe it is nationally important (but not vice versa). Notes: The bars represent 95 per cent confidence intervals. For variable descriptions, see Table A1 in the Online Appendix. Source: Based on 2019 Eurobarometer 91.5 data.
One may also wonder whether this observed asymmetry is unique to immigration or also present in other issues. While the detailed examination of this question is beyond this article's scope, the US CCES fruitfully asks their respondents about a number of issues and their importance. According to Figure A2 in the Online Appendix, while PII asymmetries are present in some other "position" issues, the particular ideological patterns of this asymmetry are rather idiosyncratic to immigration, where the more conservative, restrictive side reports higher PII. Abortion is perhaps the most similar issue to immigration in terms of the observed asymmetry and its ideological associations, though the gap between "pro-life" and "pro-choice" sides is much smaller than between anti-immigration and pro-immigration sides. At the same time, more liberal supporters of fewer restrictions on gay marriage or more restrictions on guns, for instance, are more likely to think of the respective issues as important than their more conservative counterparts. 8 Furthermore, unlike most other issues, immigration appears to be distinct in terms of the issue "moderates" reporting higher PII than the pro-immigration side (see also Figure 2).
Finally, a related question is whether the revealed PII differences between anti-and pro-immigration voters may stem from their more general differences in political interest. As shown in Figure A3 in the Online Appendix, this is not the case-those who are interested in politics are more or less equally represented among both groups (as well as supporters and opponents of most other issues for that matter). In other words, compared to pro-immigration voters, anti-immigration voters care more about immigration in particular, not politics in general. 9

Discussion
Despite the fact that the US and some other major immigrant-receiving Western countries have now been experiencing the most pro-immigration public opinion recorded in their history, lawmakers are still hesitant to enact significant pro-immigration reforms and politicians do not seem to be eager to make strong pro-immigration appeals. Institutional complications aside, this short article documents an important yet largely overlooked behavioral reason why the positive public opinion on immigration has not translated into political changes: the fact that those who want less immigration generally care much more about the issue than those who want more immigration. However, while it has been a long-time complaint from pro-immigration activists that their Notes: The bars represent 95 per cent confidence intervals. For variable descriptions, see Table A1 in the Online Appendix. Source: Based on the TTS data.

8
As can be expected (see Footnote 4), the CCES data also indicate PII asymmetries in the case of more "valence-like" issues, such as regulating the environment or punishing crime (not shown). 9 Given that PII is a relative concept, it must be the case that those voters who happen to be pro-immigration care more about other non-immigration issues (the specifics of which likely depend on a particular context).
anti-immigration counterparts are more numerous and engaged, there has been no systematic empirical evidence on this question so far.
It is a reasonable expectation that perceptions of issue importance and particular policy preferences held by an individual are both distinct and independent of each other. Indeed, in theory, voters can be personally passionate about any particular political issue, whether it is conservative or liberal, something symbolic or material, or meant to upheld or change the status quo. As this article extensively documents, however, in practice, these two dimensions can be more or less overlapping with each other.
What are the implications of the documented issue importance asymmetry for immigration politics and public opinion? The first implication is that when one takes PII into account, the United States and other immigrant-receiving countries have likely experienced no discernible relative increase of people who want more immigration for whom this is an important issue. At the same time, given the rise the contextual salience of immigration over recent decades, the absolute number of engaged people who want to decrease immigration may have actually increased. More generally, while there are many contexts where pluralities oppose immigration and consider it their main political priority, there is no known political context in which a plurality of voters wanted to increase immigration and thought it was more important than other issues.
The second, related implication is that if immigration preferences and personal interests in the issue are rather stable, most political events that increase the contextual salience of the issue should be statistically more likely to engage those who oppose than those who support immigration-even when the former is a shrinking segment of the population. In turn, this helps explain why the contextual salience of immigration has been associated with a populist and antiimmigration vote (Dennison and Geddes 2019).
In the end, this article suggests that people's disagreements over the importance of immigration as a political issue may be as important as their disagreements over the merits of particular policies in determining which groups and political coalitions eventually get their way. Of course, it is important not to overstate the potential consequences of the immigration issue importance asymmetry for policy outcomes. As emphasized earlier, public opinion is only one policy input among many. Since it can also be indicated by the contentious issue of gun control in the United States, for instance, higher issue salience among the proponents of a certain cause does not straightforwardly result in any related policy change.
This article is not without limitations. First, while the descriptive finding of the asymmetry across high-income, immigrant-receiving democracies in the 21st century is uncontroversial, future research can explore the development of immigration attitudes and the causal mechanisms of the relationship in more detail. For example, do voters come to be interested in immigration first and then develop their preferences, or vice versa? It is also important to provide (quasi-) experimental evidence of whether the exogenous changes in contextual immigration salience (for example, due to unrelated political events of various valence) may have similar asymmetric effects in making voters care more or less about immigration.
Secondly, it is likely that the relationship between preferences and PII can vary depending on political (and even partisan) contexts, which has not been explored here. For instance, is there something distinct about immigration that makes it inherently more of a "political issue" among those who oppose it (for example, due the role of widespread ethnocentric predispositions), or is it just a function of the studied context (for example, outside of frontier and other regions where the major political challenge is, in fact, to attract more immigrants)?
Finally, future studies can examine the qualitative underpinnings of the attitudinal issue importance asymmetry, including its behavioral manifestations and the possible "sub-issue" differences in the priorities within the domain of immigration. For instance, it is possible that while pro-immigration actors are relatively more (behaviorally or cognitively) concerned with helping existing immigrants, or "stocks," anti-immigration groups and voters care more about preventing future immigration, or "flows" (see, for example, Margalit and Solodoch 2021).