Since the realignment of the late 1960s, American political parties have become both more internally coherent and more polarized from each other on race, economics, religion, and other issues (Abramowitz and Saunders Reference Abramowitz and Saunders1998; Adams Reference Adams1997; Aldrich Reference Aldrich1995; Carmines and Stimson Reference Carmines and Stimson1989; Layman Reference Layman2001; McCarty, Poole, and Rosenthal Reference McCarty, Poole and Rosenthal1997; Noel Reference Noel2013; Rohde Reference Rohde1991). This is well known (Mason Reference Mason2018; Grossman and Hopkins Reference Grossman and Hopkins2016, Reference Grossman and Hopkins2015; Freeman Reference Freeman1986). We know less, however, about the implications of this partisan reordering and polarization for internal party conflict. Such a dispute seems to be happening in the current GOP, led by Donald Trump. Despite a candidacy that often departed from mainstream conservative positions on many issues, he won the 2016 Republican presidential nomination. However, many Republican elites have claimed that, whatever Trump’s goals, they are not commensurate with conservatism.
Even before he was elected, Trump’s hostility and bellicosity alarmed some Republicans, so much so that National Review, the conservative organ founded by the late William Buckley Jr., published a special issue in the winter of 2016, titled: “Conservatives Against Trump.” The roster of twenty-one “Never Trumpers” in that issue reads like a veritable “Who’s Who” of the conservative establishment.Footnote 1 Recent converts to the roster of Trump detractors include former administration officials like Rex Tillerson (Secretary of State), and General James Mattis (Secretary of Defense), as well as recently retired senators Jeff Flake (R-AZ), and Bob Corker (R-TN).
This opposition stands in stark contrast to much of the GOP conference. With few exceptions, congressional Republicans support their president, most voting in lock step with him.Footnote 2 Apparently, the base takes its cues from congressional Republicans, as Trump’s departures from conservative doctrine appear to do no harm to his standing among party activists and the mass public. Consider the following: Trump had an approval rating of 88% among conservatives after meeting with the presidents of North Korea and Russia (Bender Reference Bender2018). To the extent that Trump’s embrace of both leaders ran counter to long-standing preferences on the right (Reilly Reference Reilly2018), granting an audience to long-time adversaries, his poll numbers were nothing short of remarkable.
Trump’s support from many Republicans, even against the objections of well-respected conservative figures, fails to square neatly with conventional wisdom on partisan polarization. Most accounts of polarization highlight inter-party differences, something driven by the increasing alignment of partisan and ideological identities. More than ever before, Democrats are more liberal, Republicans are more conservative. This is accompanied by the increasing embeddedness of partisan identities (Levendusky Reference Levendusky2009; Abramowitz and Saunders Reference Abramowitz and Saunders2008; Fiorina, Abrams, and Pope Reference Fiorina, Abrams and Pope2005; Mason Reference Mason2015, Reference Mason2018). Often omitted from these explanations, however, is any discussion of conflict among co-partisans. What happens when members of a party are faced with competing messages from leaders within the party? If contemporary Republican activists and voters really are ideologically conservative, then we might expect them to align more closely with traditional conservative elites. After all, many in the conservative intelligentsia have abandoned Trump because they do not see him as conservative. Yet if this charge is true, what motivates Trump’s supporters? Are they less ideologically aligned than previously thought or do they see Trump as offering an alternative to conventional conservatism (and if so, what is this alternative)?
One explanation popular among many pundits is that economically disadvantaged voters, whose concerns about downward mobility drew them to Trump, were largely responsible for his victory (Cherlin Reference Cherlin2016). Yet sustained analysis of Trump’s supporters increasingly demonstrates that what distinguished them from other Republicans was not socioeconomic status: many of his supporters were relatively well off, as well as well educated (Manza and Crowley Reference Manza and Crowley2017; Major, Blodorn, and Blascovich Reference Major, Blodorn and Major Blascovich2016). On the surface, it was a propensity to view the world in terms of race and identity that drove Trump’s candidacy and electoral success (Sides, Tesler, and Vavreck Reference Sides, Tesler and Vavrek2018). However, a more penetrating analysis suggests that it is status threat that makes race and identity salient in the first place (Mutz Reference Mutz2018; MacWilliams Reference MacWilliams2016). Of course, the tendency to see politics through a prism of fear, anger, and anxiety is something with which the right frequently struggles (Lipset and Raab Reference Lipset and Raab1978; Parker and Barreto Reference Parker and Barreto2014).
It might be tempting to view Trump—and the basis for his support, as an unlikely or extreme outcome due to exigent circumstances. Indeed, some have argued that it was a product of Barack Obama’s election and subsequent rise of the Tea Party (Parker and Barreto Reference Parker and Barreto2014, chap. 1). Further, the Tea Party (which included elements of racism, nativism, and homophobia) galvanized the far-right within the Republican Party, pushed it further to the right on domestic and foreign policy issues, and ultimately paved the way for the candidacy of Donald Trump (Gervais and Morris Reference Gervais and Morris2018).
We probe whether Trump—and by extension, his supporters—really do possess a distinct set of policy preferences. The existing analysis of Trump focuses on domestic political issues (Major, Blodorn, and Blascovich Reference Major, Blodorn and Major Blascovich2016; MacWilliams Reference MacWilliams2016; Pettigrew Reference Pettigrew2017; Sides, Tesler, and Vavreck Reference Sides, Tesler and Vavrek2018). We extend this line of inquiry to the underexplored realm of foreign policy, a place where we are less likely to observe a rift among Republicans. After all, the Republican Party is generally viewed as more active than the Democratic Party on national defense and foreign policy. Further, these issues are also more important evaluative criteria for Republican than for Democratic presidents (Abramson at al. 2007; Gadarian Reference Gadarian2010; Karol and Miguel Reference Karol and Miguel2007; Nincic and Hinckley Reference Nincic and Hinckley1991; Norpoth and Sidman Reference Norpoth and Sidman2007; Petrocik Reference Petrocik1996). If Trump and his supporters represent an idiosyncratic moment in American politics, then we ought not observe a deep rift on the right when it comes to foreign affairs (Bartels Reference Bartels2018). In contrast, if “Trumpism” is symptomatic of an ascension of a distinct faction in the Republican Party that is “changing conservatism into something that’s different,” to echo Frum (2017; also Stephens Reference Stephens2017). then we ought to observe differences in the foreign policy preferences of Trump’s supporters and establishment Republicans.
We pursue two questions in this article. First, do the foreign policy views of Trump’s supporters also differ from those of other Republicans? If so, what informs the way that Trump, and by extension his supporters, see the world? To answer these questions, we compare the foreign policy preferences of Trump supporters with those of other Republicans across two datasets. The American National Elections Study (ANES) allows us to analyze Republican primary voters, and a new survey, the 2016 State Convention Delegates Study (SCDS), provides information about Republican state convention delegates.Footnote 3 The two surveys share three nearly-identical foreign policy items, asking respondents whether the United States should stay at home (isolationism), allow refugees/immigrants from Muslim-majority countries, and engage in free trade agreements. We use these items to evaluate competing explanations for what motivates the foreign policy views of Trump supporters.
The first explanation extends the account of conservative ideological cohesion from the asymmetric polarization literature, which would suggest that we ought not to observe systematic differences between the foreign policy views of Trump supporters and those of other Republicans. The second draws on findings that domestic and foreign belief systems are linked (Rathbun Reference Rathbun2008). This permits us to assess whether the status threat explanation from existing research on Trump supporters’ domestic-policy attitudes represents an alternative to conservatism.
Our evaluation of the foreign policy preferences of Trump supporters begins with a theoretical discussion of status threat, and where it differs with more conventional approaches to the right. We then discuss the implications of these competing belief systems for foreign policy, presenting two competing hypotheses. This is followed by an empirical analysis of the attitudes of Republican primary voters and activists on isolationism, Islamic immigrants/refugees, and trade policy, in which we find support for the idea that Trump supporters are differentiated from other Republicans by status threat.
Ultimately, we make two main contributions. First, theoretically, we extend the explanatory power of status threat beyond the realm of domestic politics, allowing for the evaluation of intra-party differences. Second, ours is the first study that assesses how widely Trump’s foreign policy preferences are shared by the Republican base, both in the mass public, as well as among activists. In short, this study marks an important step in distinguishing Trump’s style of conservatism from that of his predecessors, and in detecting the growing presence of a faction within the Republican Party motivated by a status-threat style of conservatism.
Background and Theory
Why did many Republicans support Trump, especially if, as claimed by Never Trumpers, he was not conservative? For starters, he wanted to “Make America Great Again.” Research suggests that this slogan was not at all about improving the economic situations of Trump supporters. As noted earlier, income had a weak relationship with support for Trump (Sides, Tesler, and Vavreck Reference Sides, Tesler and Vavrek2018), something that casts doubt on the “working-class whites” hypothesis. Instead, it seems that Trump supporters were ultimately motivated by a sense that they were losing “their” country. This is consistent with a trend that predated Trump’s candidacy, in which a number of white Americans felt threatened by demographic change. Research suggests that status threat explains why, when primed with the fact that America will soon become a “majority-minority” country, whites tend to harbor more intolerant policy preferences (Craig and Richeson Reference Craig and Richeson.2014). Consistent with status threat, whites who believed they would soon be in the minority were more likely to support Trump (and anti-immigrant policies) than those who remained unthreatened by demographic change (Major, Blodorn, and Blascovich Reference Major, Blodorn and Major Blascovich2016).
Voters motivated by status threat fear that the America that they have come to know and love is slipping away. For them, America is being overtaken by values and groups that will ultimately ruin “their” country. The solution is a return the “good old days,” a time during which the dominance of their group (i.e., white, male, middle-class, Christian, native-born folks) was unchallenged (Devos and Banaji Reference Devos and Banaji2005; Smith Reference Smith1991). Unlike typical group-conflict models in which the competition is over scarce resources (Bobo and Hutchings 2004; LeVine and Campbell Reference Levine and Campbell1972), status threat is a symbolic concern about conflicting ways of life (Gusfield Reference Gusfield1963). Although a sense of status threat tends to push voters—especially weak partisans—towards the Republican Party (Craig and Richeson Reference Craig and Richeson.2014), it is very different from traditional conservatism. The policy preferences of status-threatened voters reflect efforts to maintain the prestige of their group, even if this results in policies that run counter to professed conservative principles. This can mean ignoring the rule of law, flouting democratic norms (e.g., freedom of the press and free speech), suppressing voting rights, accepting policies that require more government (such as racial profiling), and even the possibility of inciting violence (Lipset and Raab Reference Lipset and Raab1978; MacLean Reference MacLean1994; McVeigh Reference McVeigh2009; Parker and Barreto Reference Parker and Barreto2014).
To the extent that status threat seeks, at best, to preserve existing social and economic arrangements, and at its worst to return to a time during which these arrangements were less egalitarian, this style of politics is typically associated with right-wing political movements (Lipset and Raab Reference Lipset and Raab1978; Mutz Reference Mutz2018; Parker and Barreto Reference Parker and Barreto2014; Federico and de Zavala Reference Federico and Golec de Zavala2018). Cleavages on the right between groups driven by status politics and more establishment-style conservatives are traceable to at least the early 1960s, a time during which conservatives began to part ways with the wing of the party preoccupied with maintaining white social prestige.
It began with the rise of the John Birch Society (JBS) in the late 1950s. Among other things, they thought President Eisenhower a communist sympathizer, and sought to have Chief Justice of the Supreme Court, Republican Earl Warren, removed on the basis of the Brown decision. The breach between these factions accelerated in the aftermath of Richard Nixon’s failed bid for the presidency in 1960. Alarmed at the progress of the Civil Rights movement, the JBS was chief among those who insisted that racial progress was engineered by the Soviet Union. Robert Welch, founder of the JBS, and his supporters, believed the Soviet Union sought to undermine America through the civil rights movement. Eventually, the father of postwar conservatism, William Buckley Jr., ejected Welch and the JBS from the conservative movement properly understood, but not before the latter played no small role in the nomination of Senator Barry Goldwater in 1964 (Kabaservice Reference Kabaservice2012; McGirr Reference McGirr2001). More recently, it was the Tea Party, and its fixation on the Obama presidency, that adhered to an approach consistent with the status threat model (Parker and Barreto Reference Parker and Barreto2014).
Conservatives, on the other hand, are not primarily animated by threat. They are willing to accept change so long as it means retaining social, economic, and political stability. Relative to those motivated by eliminating threats to status, conservatives typically reject racial profiling and suppressions of free speech, especially if such actions might result in violence. One way in which to view the observed differences between conservatives and people moved by status threat is that the former are more pragmatic, whereas the latter are more dogmatic. Whereas conservatives prize order and stability above all else (Allitt Reference Allitt2009; Jost et al. Reference Jost, Glaser, Kruglanski and Sulloway2003; Kirk Reference Kirk1953; Rossiter Reference Rossiter1982), people operating under status threat are more concerned with maintaining group prestige in a changing society (Hofstadter Reference Hofstadter1965; Lipset and Raab 1970).
In light of Rathbun’s (2007, 403) observation that “it makes little sense to study domestic and foreign policy in isolation,” we draw on tenets of status politics in domestic politics, and extend these to foreign policy. We have already illustrated, anecdotally, the disconnect between Trump’s approach to foreign policy and the conservative establishment. By and large, the establishment tends to hew close to realism. This amounts to an amoral, sober-minded view of dealing with the world, a view that prioritizes the safety and security of America and its interests. It is likely the case that threatening nuclear war with North Korea, and pulling American troops out of Syria (inviting Iran and Russia to have outsized influence in the region) fails to square with the preferences of mainstream conservatives, advocates of realism, and a more cautious foreign policy.
Again, the disagreement is anecdotal, limited to elites. It may well be the case that many party activists and Republicans in the mass public agree with Trump. This plays into the narrative that, as far as his base is concerned, he can do no wrong (Kay Reference Kay2017). In contrast, status politics (or status threat) turns on the belief that social change is subversive. These sentiments are often triggered by the perception, at least in the context of the United States, that the culturally dominant group’s social prestige is in decline; their place in America is no longer secure. Further, they believe that the very values that they hold dear, the ones that define their way of life, are under attack. To forestall this perceived decline, they seek to restrict access to power and privilege.
Hypotheses
Is there really a cleavage on the right between the foreign policy attitudes of Trump supporters and those of other Republicans? Second, if such a cleavage exists, how might we identify it? In what follows, we outline four hypotheses (two paired sets). One hypothesis in each set applies to Republican primary voters, and the other to State Convention Delegates. The first set of hypotheses extends the claim that there is a faction within the Republican Party that is informed by status threat when it comes to foreign affairs. The second set of hypotheses draws on literature documenting partisan ideological alignment, suggesting that the foreign policy attitudes of Trump’s supporters will be more consistent with conservatism than with status threat.
Each set of hypotheses applies separately to (1) activists and (2) primary voters. If status threat really has created division in the contemporary Republican Party, as we hypothesize, then we ought to observe evidence of this not only among party activists but also among party adherents in the mass public. Yet it is also possible that we might observe status threat differences only among one group. Activists tend to have more “constrained” ideological views and defined positions on more issues (Converse Reference Converse1964; Herrera Reference Herrera1992; McClosky, Hoffmann, and O’Hara Reference McClosky, Hoffman and O’Hara1960; Layman et al. Reference Layman, Carsey, Green, Herrera and Cooperman2010; Zaller Reference Zaller1992; Noel Reference Noel2013; Miller and Jennings Reference Miller and Kent Jennings1986; Campbell et al. Reference Campbell, Converse, Miller and Stokes1960; Stone and Abramowitz Reference Stone and Abramowitz1983; Conway and Feigert Reference Conway and Feigert1968; Miller and Jennings Reference Miller and Kent Jennings1986; McCann Reference McCann1995), and are sometimes regarded as more ideologically “purist” than party leaders and elected officials (Grossman and Hopkins Reference Grossman and Hopkins2015; La Raja and Shafner Reference La Raja and Shaffner2015). It would be reasonable to expect Trump supporters, even among activists, to possess more entrenched foreign policy views, and thus to be less likely to abandon the conservative position in favor of status threat. In contrast, members of the mass public are thought to be less ideologically constrained and less polarized, making them potentially more likely to update their foreign policy attitudes to mirror those of their preferred candidate (Lenz Reference Lenz2012; Jacobson Reference Jacobson, Bond and Fleisher2000; Zaller Reference Zaller1992). After all, existing research on the appeal of status threat to Trump supporters has focused on voters in the mass public, making it plausible that we might observe more variability in the foreign policy attitudes of Republican primary voters than in their activist counterparts.
H1a (activists): Given evidence of a growing divide on the right and the link between foreign policy and domestic attitudes, status threat will serve as a source of cleavage between Trump supporters and other Republican activists.
H1b (mass public): Given evidence of a growing divide on the right and the link between foreign policy and domestic attitudes, status threat will serve as a source of cleavage between Trump supporters and other Republican primary voters.
However, there are reasons to believe that when it comes to foreign affairs, there is no real distinction in the policy attitudes of voters on the right who are concerned about status threat, and those of other conservatives. As noted previously, a sizable body of literature suggests that partisans are both more polarized and more ideologically aligned than ever before, meaning that Republicans tend to be ideologically conservative (Levendusky Reference Levendusky2009; Abramowitz and Saunders Reference Abramowitz and Saunders2008; Fiorina, Abrams, and Pope Reference Fiorina, Abrams and Pope2005; Mason Reference Mason2018, Reference Mason2015; Bartels Reference Bartels2018). This tends to be the case particularly among activists, but to some extent also characterizes highly engaged voters (Converse Reference Converse1964; Herrera Reference Herrera1992; McClosky, Hoffmann, and O’Hara Reference McClosky, Hoffman and O’Hara1960; Layman et al. Reference Layman, Carsey, Green, Herrera and Cooperman2010; Zaller Reference Zaller1992; Noel Reference Noel2013; Miller and Jennings Reference Miller and Kent Jennings1986; Campbell et al. Reference Campbell, Converse, Miller and Stokes1960; Stone and Abramowitz Reference Stone and Abramowitz1983; Conway and Feigert Reference Conway and Feigert1968; McCann Reference McCann1995). It is thus possible that one or both groups of Trump supporters (convention delegates and primary voters) will fail to register substantially different foreign policy attitudes from other Republicans, and will instead continue to organize their views in keeping with conservatism.
H2a (activists): Given partisan ideological alignment, the foreign policy attitudes of activists who support Trump will not differ significantly from those of other Republican activists.
H2b (mass public): Given partisan ideological alignment, the foreign policy attitudes of primary voters who support Trump will not differ significantly from those of other Republican primary voters.
Examining the Foreign Policy Attitudes of Trump Supporters: Data
We evaluate whether the foreign policy attitudes of Trump supporters differ from those of other Republicans at both the mass and activist levels using the 2016 ANES and the 2016 State Convention Delegate Study. The two surveys contain three foreign policy items that are nearly identical, measuring respondents’ opinions on isolationism, banning refugees, and free trade agreements. These three variables, the latter two of which ought not necessarily elicit disagreement on the right given the general consensus among Republicans on such issues, allow us to assess whether Trump supporters are indeed holding different foreign policy attitudes than other Republicans, and whether any differences are consonant with status threat.
To examine the beliefs of Trump supporters in the mass public, we turn to the 2016 ANES. We constrained our analysis to those respondents who reported voting in the 2016 Republican primaries, which results in a total N=859. Respondents were asked to indicate whether they voted for Ted Cruz, John Kasich, Marco Rubio, or Donald Trump. This question will allow for the comparison of the foreign policy views of Trump primary supporters with those of other Republican primary voters.
Our activist data come from the 2016 State Convention Delegate Study (SCDS), which includes Republican State Convention delegates from Illinois (214), Utah (856), and Texas (1,347), for a total N=2,397.Footnote 4 Delegates were asked “Whom did you most prefer as your party’s nominee for president in 2016?” and asked to choose one of the following: Donald Trump, Ted Cruz, Marco Rubio, Jeb Bush, John Kasich, Ben Carson, Rand Paul, Chris Christie, or someone else (write-in). This question allows us to examine whether differences exist between the foreign policy views of Trump supporters and those of other Republican delegates, and if so, along what dimension. (Detailed information on the SCDS, including candidate choice by State, can be found in the online appendix).
Table 1 displays demographic information for both the Republican State Delegates from the SCDS and the Republican primary voters from the ANES. The two samples are highly similar. The average respondent in both samples is in their mid-50s, conservative ideologically (registering near 6 on a 7-point scale), between Independent and Weak Republican (5–6 out of 7), and white (90%). The differences between the two samples largely comport with what we know about activists—for example, the SCDS respondents are, on average, more educated and wealthier than the ANES respondents (Rosenstone and Hansen Reference Rosenstone and Hansen1993; Campbell et al. Reference Campbell, Converse, Miller and Stokes1960; Verba, Schlozman, and Brady Reference Verba, Lehman Schlozman and Brady1995). In addition, there are 16% fewer female respondents in the SCDS sample.
Summary statistics compared for Republican primary participants
Table 2 shows support for Republican primary candidates among ANES and SCDS respondents. In particular, the table shows support for the four candidates who received the most support from respondents—Cruz, Kasich, Rubio, and Trump—along with the percentage supporting other candidates. (The ANES does not list any other candidate choices. A fuller breakdown of the SCDS candidate choices and percentages is in the online appendix.) The activists and primary voters show slightly different trends in candidate support. The majority of ANES respondents—51.9%—indicated that they supported Donald Trump in the Republican primaries, followed by Cruz (18.9%), Kasich (13.3%), Rubio (9.9%), and other (6.1%). There was somewhat less consensus among the state delegates, a majority of whom did not agree on their favored candidate. In contrast to the ANES preference for Trump, a plurality of state delegates supported Cruz (40.9%), reflecting the large number of Texas convention delegates in the sample. Cruz was followed by Trump (16.9%), Kasich (11.0%), Rubio (7.5%), and other (6.4%).
GOP primary candidate preferences of ANES and SCDS respondents
Foreign Policy Dependent Variables
The ANES and SCDS contain three overlapping foreign policy questions. The first gauges respondents’ attitudes towards isolationism, the second towards immigrants/refugees from Islamic countries, and the third towards free trade agreements. Using these dependent variables, we evaluate whether the foreign policy preferences of Trump supporters differ from those of other Republicans, and whether these differences align with what we might expect from a conservatism of status threat rather than establishment conservatism or neoconservatism. Descriptive statistics for each of these variables can be found in the online appendix.
Isolationism
The first variable we consider measures attitudes toward isolationism. The ANES and SCDS versions of the question are worded identically, asking whether respondents agree that “This country would be better off if we just stayed home and did not concern ourselves with problems in other parts of the world.” The ANES variable has two substantive response categories, agree/disagree, which we coded as a binary variable where 1 = agree, and 0 = disagree. The SCDS version uses a 1–5 Likert scale, ranging from disagree strongly to agree strongly.
Although isolationism has frequently been associated with libertarianism (Holsti and Rosenau Reference Holsti and Rosenau1996; Keele Reference Keele2005), recent work suggests that isolationist sentiments are more closely related to “nationalist unilateralism,” or “militant internationalism” as a way of meeting threats to one’s country (Dueck Reference Dueck2006; Rathbun Reference Rathbun2013; Rathbun et al. Reference Rathbun, Kertzer, Reifler, Goren and Scotto2016). We thus might expect Trump supporters, if they are indeed motivated by status threat, to register more highly isolationist sentiments than other Republicans.
Islamic Immigrants/Refugees
The second dependent variable, regarding refugees and immigrants from Muslim-majority countries, more obviously relates to status threat. The question wording differs slightly between the ANES and SCDS, but both tap into the combination of anti-immigration and anti-Muslim sentiment that came to uniquely characterize early actions of the Trump administration. The ANES version of the variable asks respondents: “Should Syrian refugees be allowed to come into the U.S?” Responses range from favor a great deal (1), to oppose a great deal (7). The SCDS version more explicitly draws the link between threat and immigration, asking respondents if they agree that “People from countries in which Islamic terrorist organizations have a significant presence should be barred from entering the U.S. until we have a better system in place for identifying individuals who are terrorist threats.” Responses range from disagree strongly (1) to agree strongly (5). Both questions are thus coded such that higher values indicate more opposition to refugees/immigrants.
If the status threat model indeed distinguishes Trump supporters’ foreign policy views, then questions about allowing Syrian refugees into the United States and barring immigrants from Islamic countries should be areas in which a cleavage emerges among Republican respondents. We might expect that, even prior to Trump’s election and the travel ban, his supporters gravitated towards him because his threat-centric rhetoric resonated with them.
Trade
The final dependent variable evaluates another foreign policy issue that Trump has brought to the fore: free trade agreements with other countries. The ANES item asks respondents if they “Favor or oppose free trade agreements with other countries,” and how strongly. This yields seven response options, coded from strongly favor to strongly oppose. The SCDS version is nearly identical, saying: “There has been a lot of talk recently about the U.S. making free trade agreements with other countries,” and asking respondents where they would place themselves on a 1–7 scale ranging from strongly favoring an increase in trade to strongly opposing.
Inasmuch as the Republican Party has embraced free market principles, free trade is something that we might generally expect conservatives to favor. It is thus possible that there will be no cleavage on this issue among Republicans. Yet given Trump’s emphasis in his 2016 presidential campaign on “stupid” free trade agreements, especially those with Mexico and Canada, which he alleged destroyed American jobs, it is also possible that status politics, nationalism, or both could have driven divergent views of trade even among primary voters (Economist 2016).
Independent Variables
The core question in this analysis is whether the responses of Trump supporters to our three foreign policy dependent variables deviate substantively from those of other Republicans. We predict responses to these dependent variables based on Trump support, ideology, and demographic factors including education, income, age, gender, and being white. Other model specifications,Footnote 5 as well as further information on the coding of these variables can be found in the online appendix.
Trump Support
As explained earlier, both the SCDS and the ANES contain an item that allows us to distinguish Trump supporters from other Republicans. The SCDS item asked state delegates who they most preferred as their party’s nominee for president in 2016. The ANES version was similar, asking respondents who voted in the Republican primaries which candidate they chose (candidate preferences are compared in table 2). In our models, we use the nominee preference variable from the SCDS and the primary choice variable from the ANES to create indicator variables for Trump support, coded 1 if the respondent voted for Trump in the primary, and 0 if they voted for any other Republican.
The Trump support variable is our focus throughout the models. We hypothesized that Trump supports differ from other Republicans due to the former’s affinity to status threat. Although we cannot measure status threat directly, we can use the Trump support variable as a sort of proxy. If we find an effect for Trump support even when controlling for factors like conservative ideology, strong national identity, ethnocentrism, and so forth, then it might be reasonable to conjecture that the Trump support variable is standing in for a different style of conservatism.
Ideology
It is entirely possible that any differences in foreign policy attitudes between Trump supporters and other Republicans can be accounted for by ideology. In other words, it might be the case that Trump supporters are simply more ideologically extreme than are more establishment Republicans. Both the SCDS and the ANES include an ideology scale ranging from extremely liberal (1) to extremely conservative (7). As the mean for both samples, shown in Table 1, was between 5 and 6 on this scale, with few to no respondents choosing a response less than 4, we refer to this variable as “conservatism”. Because our analysis is already restricted to Republican primary voters, we do not include party identification in our final models.
Other Controls
We also include several standard demographic control variables in our models. These variables are summarized in table 1, and are explained briefly here. The first is education. In the SCDS, this is categorical, ranging from high school or less (1) to post-graduate training (5). We recoded the ANES variable to mirror the SCDS categories, ranging from high school or less (1), to post-graduate education (5). We also control for income, which has six categories in the SCDS, and is continuous in the ANES. The third control is age, continuous in both. To account for gender, we include a binary variable coded 1 for female. Finally, we include a binary variable coded 1 if a respondent is not white, and 0 if a respondent is white.
Results and Analysis
We analyze these data using a series of models. As all but one of our dependent variables is ordinal, we use ordered logistic regression for the majority of our analysis. One dependent variable, the isolationism item on the ANES, is binary, in which instance we use logistic regression (more detailed information on variable coding as well as on other model specifications can be found in the online appendix). The results from these regressions for the ANES can be found in table 3, and for the SCDS in table 4. To evaluate the effect of Trump support more directly, we plot the marginal effects of supporting versus not supporting Trump for each of our dependent variables (when all other respondent characteristics are held at their means). These are shown in figures 1–6.
ANES full models
***p < 0.05, ** p < 0.01, *** p < 0.001
Standard errors in parentheses
SCDS full models
***p < 0.05, ** p < 0.01, *** p < 0.001
Standard errors in parentheses
Effect of Trump support on isolationism (ANES)
Effect of Trump support on isolationism (SCDS)
Effect of Trump support on opposition to Syrian refugees (ANES)
Effect of Trump support on opposition to Islamic immigrants (SCDS)
Effect of Trump support on opposition to free trade agreements (ANES)
Effect of Trump support on opposition to free trade agreements (SCDS)
Predicting Isolationist Attitudes
Recall that the first foreign policy variable measures Republicans’ attitudes towards isolationism through agreement with the statement: “This country would be better off if we just stayed home and did not concern ourselves with problems in other parts of the world” (the wording is identical in both surveys). Among ANES respondents, Trump supporters are somewhat more likely to register isolationist sentiments than are other Republicans. In the logistic regression (table 3), the only predictors of isolationist attitudes that reach conventional levels of statistical significance are Trump primary support (status threat), which positively predicts isolationist attitudes, and the indicator for female, which predicts a (substantively small) decrease in isolationist attitudes. Figure 1 shows the marginal effects of being a Trump supporter on isolationist attitudes. Supporting Trump increases the average respondent’s likelihood of agreeing that the United States should stay at home by 13% (from 16% for non-Trump supporters to 29% for Trump supporters). These results support our hypothesis that Trump supporters’ opinions on isolationism are distinct from those of other Republicans.
We observe a similar trend among SCDS respondents. Again, Trump support is a substantively large predictor of isolationist attitudes. The only other positive and statistically significant predictor of isolationist attitudes is ideology, although the substantive effect is half that of Trump support. In addition, income and age have statistically significant but substantively small negative effects. Figure 2 allows for a comparison of the marginal effects of Trump support across all five levels of the SCDS version of this dependent variable. The difference between Trump supporters and other Republicans is not decisive across all levels of the dependent variable—the two groups were equally likely to choose options two or three (disagree or neither). Differences existed on the probability of choosing option one, with non-Trump supporters being about 10% more likely to strongly disagree that the United States. should stay at home. On options four and five, Trump supporters were 7% more likely to agree the United States. should stay at home, and 3% more likely to strongly agree. These results indicate differences—in particular at the extremes—between Republican activists who supported Trump (more comfortable supporting isolationism) and those who did not (more likely to strongly disagree with isolationism).
Predicting Attitudes on Refugees and Immigrants
The second set of dependent variables assesses attitudes towards immigrants from Islamic countries. Although worded slightly differently, both items allow us to probe Republicans’ attitudes towards immigrants from a region that Trump not only targeted in his campaign rhetoric, but later made the object of a travel ban.
The results support our hypothesis that status threat animates Trump supporters. Among ANES Republican primary voters, Trump support (status threat) is both a statistically significant and substantively large predictor of opposition to Syrian refugees. Although ideology is an important predictor, its substantive effect is one-third that of Trump support (the indicator for female also has a small but statistically significant effect). Figure 3 clarifies the impact of supporting Trump on the average respondent’s probability of choosing each response option on the Syrian refugee variable. Although Trump supporters and non-Trump supporters were similarly unlikely to favor refugees, they differed in the extent of their opposition. Non-Trump supporters were 13% more likely than Trump supporters to neither favor nor oppose (response 4) and 38% more likely to oppose refugees a little (response 5), and 7% more likely to moderately oppose refugees (response 6). In contrast, Trump supporters were substantially more likely—27%—to register the strongest possible opposition to refugees (response 7). In short, while Republicans were unlikely to favor refugees in general, Trump supporters were decidedly more ardent in their opposition (in keeping with what we might expect from those motivated by status threat).
This trend continues among SCDS respondents. In the ordered logit, several variables are important predictors of opposition to Syrian refugees. The chief among these is again Trump support (status threat), followed by ideology (though only half as large as the coefficient for Trump support), age (to a much smaller extent), and being non-white (in this sample, non-whites are largely Texas delegates who described themselves as Hispanic or Latino), which predicts a surprising amount of opposition to the immigrants in question. In addition, education has important and somewhat sizable muting effect on opposition to immigrants, as might be expected. Figure 4 clarifies the impact of Trump support on immigration attitudes, holding all other variables at their means. As with the ANES, non-Trump supporters were much less likely to register extreme opposition to immigrants (option 5). Non-Trump supporters were more likely to choose every other option, even those that registered the friendliest attitudes towards immigrants (unlike in the ANES). They were 5% more likely to choose option 1, 11% more likely to choose option 2, 7% more likely to choose the middle option (3), and 5% more likely to choose option 4. In contrast, being a Trump supporter increased one’s likelihood of registering the harshest attitudes (option 5) by 29%. As with the Republican primary voters, Republican state convention delegates who supported Trump had markedly more hostile attitudes towards immigrants from Islamic countries.
Predicting Attitudes on Free Trade Agreements
We turn finally to the variable on which we might expect to see the least difference among our Republican samples: free trade. This is an item on which conservatives ought not necessarily register much disagreement, and one on which many may not have strongly held opinions. For ANES respondents, the only important predictor of opposition to free trade in the ordered logit was Trump support. The only other variables that reached conventional levels of statistical significance was education, with higher levels of education predicting a sizeable decrease in favorability to free trade. Figure 5 allows us to examine the extent of the difference between Trump supporters and other Republicans. As one might expect, most respondents did not seem to have clear opinions on free trade, with the majority choosing option 4 (“neither”). Slight differences existed on four of the remaining response options. Non-Trump supporters were 4% more likely to strongly favor (option 1) and 7% more likely to favor (option 2) free trade agreements. In contrast, Trump supporters were 5% more likely to oppose (option 6) and 5% more likely to strongly oppose (option 7) such agreements. These differences are much smaller than those for the other two foreign policy variables, indicating that this is not a topic on which Republican primary voters substantively differed, regardless of their preferred candidate.
Among state delegates, Trump support is by far the most important predictor of opposition to free trade in the ordered logit, followed by the indicator for female, then ideology, and age (though to a much smaller extent). As in the ANES, more education reduced one’s propensity to oppose trade agreements, as did having higher income. To assess whether Trump supporters differed from other Republicans, even when controlling for these other variables, we turn to figure 6. Unlike the primary voters in the ANES, the activists in the SCDS generally held more defined opinions on the topic, with “neither” (option 4) being a generally less popular response option. Beyond that, non-Trump supporters were 7% more likely to strongly favor such agreements (option 1), 9% more likely to favor them (option 2), and 4% more likely to slightly favor them (option 3). Trump supporters were more opposed to free trade, as the regression coefficient indicated, being 4% more likely to slightly oppose such agreements (option 5), 9% more likely to outright oppose them (option 6), and 8% more likely to strongly oppose them (option 7). The difference between Trump supporters and other Republicans are more pronounced among activists than among primary voters, as we might expect. Although they are smaller than those for the other variables considered, a clear trend seems to exist, with non-Trump supporters systematically approving of free trade, and Trump supporters doing the opposite.
Discussion and Conclusion
We began this paper considering what, if any, relationship existed between Trump’s foreign policy beliefs and those of his supporters both among activists and the mass public, and whether the preferences of Trump’s supporters differed from those of other Republicans. More to the point, we examined whether the preferences of Trump’s supporters could be characterized by status threat, beyond conservatism. As a practical matter, this entailed limiting our analysis to the right, something accomplished by assessing Republican voters in the 2016 primaries and Republican delegates in 2016 state conventions. We hypothesized that the preferences of Trump supporters would depart from those of the Republicans who supported other candidates. Observing any daylight on the right, especially when it comes to foreign affairs, would be surprising given the attention spent on the issue by the political right. Yet this is exactly what we found.
Our positions (hypotheses 1a and 1b) were supported in evaluating our first two foreign policy items. Trump supporters, as status-threatened conservatives, were more likely to take isolationist positions than were other Republicans, and were more likely to adamantly oppose allowing Syrian refugees or immigrants from Islamic countries into the United States, and even more likely to oppose free trade agreements. The results for our third variable, free trade agreements, were more in line with hypothesis 1a than with 1b. Although Trump support was a significant predictor of opposition to free trade in our models, the substantive difference among Republican primary voters was quite small, with most of these respondents indicating no real position on free trade. A clearer trend existed among the state delegates, more in keeping with our first hypothesis as it pertains to activists.
On the issue of isolationism, Trump support predicted more isolationist attitudes among both activists and members of the mass public. The key is that even on an issue that should generate very little disagreement among Republicans—isolationism—Trump’s supporters were more likely to choose an isolationist posture than are their Republican counterparts, even when controlling for these other explanations. Inasmuch as isolationism can be seen as a response to dealing with threats, this is consonant with the hypotheses advanced in which we outlined reasons for differences among Republicans.
The effect of Trump support, as status threat, was even more pronounced on the issues of refugees and immigration, respectively. Although ideology also predicted opposition to Syrian refugees among ANES voters (exhibited by the propensity of all Republicans in the sample to oppose refugees to some extent), Trump supporters were substantially more likely to register the highest level of opposition to Syrian refugees. This is likely because the grounds on which conservatives refuse to welcome the refugees are different from those driven by status threat. Conservatives likely reject the prospect of receiving Syrian refugees out of a fear of terrorism (Holbrook et al. Reference Holbrook, López-Rodríguez, Fessler, Vázquez and Gómez2017), whereas status-threatened Republicans are more likely driven to reject immigrants from Islamic countries for nativist reasons, i.e. because they are not “real Americans” (Parker and Barreto Reference Parker and Barreto2014).
The results reveal an even more pronounced pattern among the state convention delegates. Activists who supported Trump were again far more likely to register the higher level of opposition to Islamic immigrants, whereas other Republicans were actually more likely to favor the immigrants in question. That support for Trump registers a direct effect on both the Syrian refugee and the Islamic immigrant items suggest the authenticity of status threat as motivation, beyond conservatism. In the present context, this means that for Trumps supporters, people from Syria and Islamic countries represent a threat to the “dominant” culture. This comports with other work on movements driven by status threat and their feelings about “illegals” (Parker and Barreto Reference Parker and Barreto2014).
The difference between Trump supporters and other Republicans on free trade was less apparent, supporting our first hypothesis for activists, but much less so for those in the mass public. Although primary voters who supported Trump were slightly more likely to oppose free trade, the attitudes of ANES voters on this topic were relatively unstructured on the whole. This is in keeping with literature that suggests that voters tend to have less defined positions on “hard” policy issues (Carmines and Stimson 1980; Pollock, Lillie, and Vittes Reference Pollock, Lilie and Elliot Vittes1993), and because trade is a topic conservatives historically worry about only insofar as it may result in war (Copeland Reference Copeland1996). The activists possessed clearer views on trade, with Trump supporters deviating from other Republicans in their opposition. This is what we might expect from those motivated by status threat, given Trump’s persistent linkage between trade and threats to American prestige, from job losses to national humiliation (Blair Center 2015).
These findings not only shed important light on the foreign policy of Trump’s supporters, but they also carry larger implications for scholarship on the Republican Party and conservatism. Had we only observed a difference between Trump supporters and other Republicans among primary voters, then we could attribute these findings to lower levels of ideological alignment among those in the mass public. But observing a distinct pattern of preferences consistent with status threat among both Republican voters and party activists—one that mirrors that of the Tea Party and aligns with the statements of Trump—suggests that both Trump’s foreign policy preferences and his election lends itself to an easier explanation than one may have initially thought: the perception that their way of life is under threat.
Status threat is nothing new to the American right. The main differences are that it seems to have moved from the periphery of right-wing thought into the mainstream. As recent scholarship suggests, the Tea Party era brought this status threat to greater prominence, paving the way for Trump’s nomination and election. Our results are in keeping with a larger trend on the right, wherein a fringe element has become a clear faction in the Republican coalition (see Blum Reference Blum2018; forthcoming). In particular, it seems possible that this faction is distinguished by eschewing conservatism in favor of status threat, a style different enough that it undergirds a rift in foreign policy preferences, something on which Republicans historically agree.
Supplementary Materials
A. State Convention Delegate Study (SCDS)
i. Methodology
ii. Summary Statistics
iii. Summary of SCDS Responses on Dependent Variables
iv. Additional SCDS Model Specifications
v. SCDS Variables and Coding
B. American National Elections Study (ANES)
i. Summary of ANES Responses on Dependent Variables
ii. Additional ANES Model Specifications
iii. ANES Variables and Coding
Appendix Tables 1-11
Appendix Figures 1-18
To view supplementary material for this article, please visit https://doi.org/10.1017/S1537592719000999
