Introduction
U.S. foreign policy is perhaps the most powerful force in global politics. Scholars have referred to the United States as a “great power” (Buzan Reference Buzan2004), a “hegemon” (Cox Reference Cox2001), and an “empire” (Bacevich Reference Bacevich2004). The U.S., which has experienced only nineteen years of peace since 1789, deployed military forces to 138 countries—nearly 70 percent of the world in 2016 (Turse Reference Turse2017). That same year, the U.S. military budget, which passed 85-15 in the Senate and 277-177 in the House of Representatives, was $611 billion—by far the highest in the world. To enable this foreign policy, the U.S. maintains over 900 military facilities in 46 countries and territories (Lutz Reference Lutz2009). Despite the importance of U.S. foreign policy and the public’s role in it, research on U.S. public opinion has focused considerably more on understanding attitudes toward domestic issues than international ones.
Extant theories regarding attitude formation on issues of foreign policy among the U.S. public have identified several important factors: ideology and partisanship (Nincic and Ramos Reference Nincic and Ramos2010; Rosenau et al. Reference Rosenau, Earnest, Ferguson and Holsti2006), political polarization (Jeong and Quirk Reference Jeong and Quirk2019), political elites (Baum and Potter Reference Baum and Potter2008; Bullock Reference Bullock2011; Gilens and Page Reference Gilens and Page2014), the news media (Kull, Ramsay, and Lewis Reference Kull, Ramsay and Lewis2003), patriotism (McCleary, Nalls, and Williams Reference McCleary, Nalls and Williams2009), and ethnocentrism (Kam and Kinder Reference Kam and Kinder2007). However, existing scholarship on the U.S. public’s foreign policy attitudes has left under-examined what many scholars have identified as the central organizing feature of U.S. society and politics: racial attitudes (Devine Reference Devine1989; Hirschfeld Reference Hirschfeld1998; King and Smith Reference King and Smith2005).
While race and racism have been closely examined as explanations for attitudes on many issues of domestic policy and immigration, this paper seeks to expand the scope of our understanding of racialization to foreign policy. I examine data from the American National Election Survey (ANES) from 1986 to 2020 regarding 1) support for U.S. military interventions, 2) the perception that foreign nations or individuals are military threats, and 3) support for US defense spending. In all three areas, I find that racial resentment is strongly associated with support for militaristic policy options. These findings contribute to the literature debating the structure of U.S. opinion on foreign policy and provide quantitative support for the argument that racism has a significant relationship with the U.S. public’s willingness to support foreign interventions and pay the costs of its foreign policy. This research also expands scholarship on the racialization of white public opinion in American politics to foreign policy and national security, helping us to further understand the role of racial attitudes in U.S. politics. Finally, this research is also in conversation with recent work on the important role of racism and racial hierarchy in international relations (Freeman, Kim, and Lake Reference Freeman, Kim and Lake2022) through its consideration of the role of the public in U.S. foreign policy.
Public Opinion and Foreign Policy
While early scholarship describes an inattentive public with unstable attitudes (Almond Reference Almond1950; Lippmann Reference Lippmann1946; Markus and Converse Reference Markus and Converse1979), later work, much of it influenced by the U.S. Southeast Asia War (Aldrich et al. Reference Aldrich, Christopher Gelpi, Reifler and Sharp2006), indicates that this early work “may have seriously underestimated attitudinal stability” (Bardes and Oldendick Reference Bardes and Oldendick1978; Holsti Reference Holsti1992; Hurwitz and Peffley Reference Hurwitz and Peffley1987). When it comes to foreign policy, one school of thought argues that opinion is a “top-down” process driven by elites (Bullock Reference Bullock2011). A second body of research advocates a “bottom-up” process of attitude formation with the public “better equipped for espousing judgments in foreign affairs than many of our top-down models claim” (Kertzer and Zeitzoff Reference Kertzer and Zeitzoff2017, 543). In the absence of detailed knowledge and information, the public largely relies on heuristics to form their attitudes about policy—which may be especially pertinent when it comes to foreign policy (Zaller Reference Zaller1992).
Considerable skepticism remains regarding the impact of foreign policy attitudes on political behavior. Foreign policy is just one of several major factors in Presidential evaluations (Edwards, Mitchell, and Welch Reference Edwards, Mitchell and Welch1995). Even when the public expresses strong and coherent attitudes, they do not always get the foreign policy they vote for (Page and Bouton Reference Page and Bouton2006). Elite opinion correlates stronger with US foreign policy than the public’s (Gilens and Page Reference Gilens and Page2014), especially when the public has limited information (Bullock Reference Bullock2011). Some argue this “disconnect” undermines the democratic nature of policymaking (B. I. Page and Bouton Reference Page and Bouton2006). Research on U.S. military interventions suggests “the public constrains, but does not set” U.S. foreign policy (Sobel Reference Sobel2001, 5). However, research on public opinion regarding wars in Southeast Asia and Iraq reveals significant social and political consequences (DeGrasse Reference DeGrasse2016; Malamud and Wozniak Reference Malamud and Wozniak2012; Modell and Haggerty Reference Modell and Haggerty1991). Opinion about the treatment of suspected terrorists during the Global War on Terror (GWOT) strongly influenced domestic politics and policies (M. Baum and Groeling Reference Baum and Groeling2010; Fawn and Hinnebusch Reference Fawn and Hinnebusch2006; Grose and Oppenheimer Reference Grose and Oppenheimer2007).
Militarization is a process which “involves an intensification of the labor and resources allocated to military purposes, including the shaping of other institutions in synchrony with military goals” (Lutz Reference Lutz2002, 73). Furthermore, “[m]ilitarization is intimately connected not only to the obvious increase in the size of armies and resurgence of militant nationalisms and militant fundamentalisms but also to the less visible deformation of human potentials into the hierarchies of race, class, gender, and sexuality, and to the shaping of national histories in ways that glorify and legitimate military action” (Lutz Reference Lutz2002, 723; Linenthal and Engelhardt Reference Linenthal and Engelhardt1996; Bernstein Reference Bernstein1999). Recent work has found that U.S. public remains supportive of a U.S. foreign policy that has been increasingly militaristic since 2001 (Toft and Kushi Reference Toft and Kushi2023).
Support for militarized foreign policy has taken, and can take, different political forms depending on timing, context, and actors. One important example of this is the “America First” or so-called “isolationist” faction of U.S. politics which is pro-military and supports a strong national defense but opposes the use of U.S. troops outside its core interests. While not opposed to U.S. military power per se, many isolationists resent the use of U.S. power in service of international institutions like the United Nations or values such as global human rights. While this brand of militarism was recently rekindled by Donald Trump and the Tea Party political movement, it has roots in the John Birch Society, Theodore Roosevelt, and previous iterations of pro-military and pro-imperial political movements whose messages were explicitly or implicitly coded in the language of race and racism (Horne Reference Horne1999; Mills Reference Mills1997; Parker and Barreto Reference Parker and Barreto2013).
Racism, Militarism, and Public Opinion
There are two main theoretical approaches under which this paper examines the empirical connection between racism and militarism among the white U.S. public. First, racialization and “othering” of non-white peoples and cultures contributes to seeing non-whites as acceptable targets for violence, unsuitable for cooperation or negotiation. Foundational work on racism and white public opinion focused on the role of social group position (Blumer Reference Blumer1958). Further work expanded this sociological approach to racism, focusing on status in the formation of white opinion (L. D. Bobo Reference Bobo1999; L. Bobo and Kluegel Reference Bobo, Kluegel, Martin and Tuch1997). Other work highlighted the competition for resources as a motivation for white racism (L. Bobo and Hutchings Reference Bobo and Hutchings1996; L. Bobo, Kluegel, and Smith Reference Bobo, Kluegel, Smith, Tuch and Martin1996). While explicit racism was thought to be on the decline over time, prompting a move towards the “symbolic racism” set of measures (Kinder and Sanders Reference Kinder and Sanders1996), other work points to its continuing relevance (Tesler Reference Tesler2016; Virtanen and Huddy Reference Virtanen and Huddy1998). The racialization of elite and public attitudes around domestic issues such as welfare (Gilens Reference Gilens1996), crime, and policing (L. Bobo Reference Bobo1997; Green, Staerklé, and Sears Reference Green, Staerklé and Sears2006; Jefferson, Neuner, and Pasek Reference Jefferson, Neuner and Pasek2021), and affirmative action (Gilens, Sniderman, and Kuklinski Reference Gilens, Sniderman and Kuklinski1998) is well established in the literature (Gilens Reference Gilens1996; Henry and Sears Reference Henry and Sears2002; Hutchings and Valentino Reference Hutchings and Valentino2004). Immigration attitudes have also been shown to be highly racialized (Citrin et al. Reference Citrin, Green, Muste and Wong1997; Levy and Wright Reference Levy and Wright2020; Masuoka and Junn Reference Masuoka and Junn2013). Previous research indicates that white policy preferences are “race-coded” in the sense that racism is strongly associated with opposition to affirmative action, welfare, and other social programs (Konitzer et al. Reference Konitzer, Shanto Iyengar, Valentino and Duch2018, 15) for white individuals. Rather than receding over time, “whites have increasingly brought racial considerations to bear” on their vote choices in presidential elections (Enders and Scott Reference Enders and Scott2019, 2).
Research shows significant differences in foreign policy opinion between Black and white individuals (Green-Riley and Leber Reference Green-Riley and Leber2023; Kinder and Winter Reference Kinder and Winter2001). There is also strong evidence of differences in how Black and white individuals respond to racial cues—including opinion on the 2003 Iraq War (White Reference White2007). Additionally, racism, stereotypes, and other racial attitudes play a significant role in public opinion regarding China (D. Kim Reference Kim2022; D. G. Kim Reference Kim2024; Ralston Reference Ralston2023). By expanding the scope of this framework to foreign policy and national security, this study seeks to better understand how the U.S. public might project and export its views on race around the globe—and beyond. This approach anticipates that racist views among the U.S. public will correspond with support for militarized policies of aggression towards non-white nations, applying racial attitudes to evaluate the humanity and deservingness of the people affected by U.S. military intervention.
The second major theoretical framework this study draws upon are theories of colonialism and decolonization. Race is not only the “central organizing feature” of U.S. domestic politics (Hutchings and Valentino Reference Hutchings and Valentino2004), there is compelling evidence that it is also foundational to the organization of international politics (DeConde Reference DeConde1992). Those expressing racism in the U.S. not only see non-whites as threatening or subhuman, they feel entitled to the possession of their lands and resources. The “racial contract” of the U.S. includes “the privileging of those individuals designated as [w]hite/persons and the exploitation of those individuals designated as nonwhite/subpersons” in creation and maintenance of the domestic and international system (Mills Reference Mills1997, 33; Darwin Reference Darwin2009; Hemmer and Katzenstein Reference Hemmer and Katzenstein2002). Scholars have noted that the structure of the international system is undergirded by racialized competition—even in an era where the liberal international order promotes racial equality (Búzás Reference Búzás2021; Bustamante Reference Bustamante2024).
Further work has expanded on how race and racism shape U.S. foreign policy decision making (Freeman Reference Freeman2023). Scholars describe the U.S. political project as a “racial project”—exemplified by its policy of settler-colonialism (Omi and Winant Reference Omi and Winant2014). The notion of “American Exceptionalism” and the idea that it should govern not only its own territory but lead the world toward liberty and democracy cannot be separated from white supremacy and ideas like the “White Man’s Burden” (Horne Reference Horne1999; Horsman Reference Horsman1981; Kipling Reference Kipling1899; Krenn Reference Krenn1999). Many argue that Western military interventions—often in the name of political or economic liberalization—are evidence of U.S. imperialism and white supremacy (DeConde Reference DeConde1992; Horne Reference Horne1999; Kushi Reference Kushi2024).
The United States justified its “Indian Wars” (1775–1924) on racial terms—often referring to its “Manifest Destiny” to expand its borders further West (Horne Reference Horne1999). Native American people were described in racial stereotypes that depicted them as generally inferior to whites but specifically violent and untrustworthy in ways that made coexistence impossible if it was even desirable at all (Hirschfelder and Molin Reference Hirschfelder and Molin2018). Conflicts to protect and promote the slave trade were also argued for politically and militarily on racial terms: not only in that they were in support of slavery as an economic policy but also that action against these people, including the Spanish, was appropriate or necessary because they were not seen as white (Maass Reference Maass2020). In this regard, race has also been central to the territorial expansion (or lack thereof) of the United States since 1898—where the incorporation of places like the Philippines (which was granted independence after World War 2) was seen as undesirable because it would add a considerable number of non-white citizens to the U.S. polity (Immerwahr Reference Immerwahr2019; Maass Reference Maass2020).
The connections between race and U.S. foreign policy also extend to East Asia (Horne Reference Horne1999; Krenn Reference Krenn1999). U.S. wars in Asia “contributed significantly to this view of Asian Americans and Asians as foreign enemies” (Lipsitz Reference Lipsitz2006, 71). U.S. rivalry with China and Japan, along with its colonization of the Philippines, contributed to restrictive immigration and racial discrimination against Asians in the U.S. (Espiritu Reference Espiritu2008; Horsman Reference Horsman1981; Okihiro Reference Okihiro2014). This dynamic suggests that “the rationalizations originally used to justify conquest of Native Americans eventually applied to Mexicans and Filipinos” (Horsman Reference Horsman1981; Lipsitz Reference Lipsitz2006, 72). Combined with U.S. military intervention in Vietnam, Cambodia, Laos, and Korea (among many others), China has long been a potential or actual adversary to U.S. foreign policy. This manifests itself contemporarily in U.S. public opinion regarding so-called “rise of China” and the perception that a more powerful and economically robust China is an economic threat, a military threat, or both.
When it comes to the racialization of Muslims, the U.S. public is more likely to label acts of violence acts of terrorism if they are carried out by Muslims (Huff and Kertzer Reference Huff and Kertzer2017). The perception that Muslims are “distinctly misogynistic” may drive support for intervention due to “gendered Orientalism” among Westerners (Terman Reference Terman2017, 1). Furthermore, “concerns over the incompatibility between Islam and U.S. values and norms is largely a ‘cover story’ to mask beliefs in one’s inherent racial superiority and Muslims’ inherent inferiority” (Lajevardi and Oskooii Reference Lajevardi and Oskooii2018, 28; d’Urso and Bonilla Reference d’Urso and Bonilla2023). Strong majorities of the U.S. public supported torture under the Bush Administration (Blauwkamp, Rowling, and Pettit Reference Blauwkamp, Rowling and Pettit2018), consistent with a broader pattern: the U.S. public often responds to foreign threats with belligerence (Davis and Silver Reference Davis and Silver2004; Holsti Reference Holsti2004; Huddy et al. Reference Huddy, Feldman, Taber and Lahav2005; Page and Shapiro Reference Page and Shapiro1992). Research on the U.S. war on terrorism notes how this conflict “reanimated a longer history of American statecraft,” which includes settler colonialism, territorial expansion, and Indian removal (Singh Reference Singh2019). Racism, war, and the connection between them have been present in U.S. politics since the Founding but manifest themselves in different ways depending on the geopolitical context.
The deep and longstanding connections between racism and militarism means that it is difficult to disentangle white racial attitudes from their views on foreign policy. Rather, this paper builds on the theory advanced by previous scholars that white racial attitudes and their ideas about foreign policy are co-constituted: white racial attitudes have been constructed, in part, around ideas about dominance on the international stage and using military aggression to accomplish and maintain it (Mills Reference Mills1997; Horne Reference Horne1999; Mutz Reference Mutz2018; Omi and Winant Reference Omi and Winant2014; Weaver and Prowse Reference Weaver and Prowse2020).
Hypotheses
Whites that express racism in foreign policy may be influenced by both the racialization and colonial vectors described above—with both often pointing their guns in the same direction. The subsequent analysis tests a set of three hypotheses to help explore the influence of both frameworks on the foreign policy views of the U.S. public:
Hypothesis 1: High levels of racial resentment lead to support for military intervention against states, nations, and groups racialized as non-white.
Primarily influenced by the literature on racialization, I expect members of the U.S. public to use race as a heuristic to influence their preferences related to U.S. military interventions, seeing violence against non-whites as justified or necessary. This hypothesis (H1)—that racial resentment leads to support for military intervention against non-whites—is examined on the 2012 ANES with a series of five questions where respondents are asked about their support for policy options to prevent Iran from acquiring a nuclear weapon. This hypothesis is also addressed on the 2016 ANES with support for intervention against ISIS in Iraq, and on the ANES cumulative survey where respondents are asked about their willingness to use force to solve problems internationally. I also evaluate this prediction using the 2012 and 2016 ANES which asks respondents about their support for torture against suspected terrorists.
Hypothesis 2: High levels of racial resentment lead to perceiving non-white nations as threats to national security.
The second hypothesis (H2) is that racial resentment is associated with the perception of foreign nations—especially those racialized as non-white—as threats to U.S. national security. While this hypothesis draws strongly on the racialization framework, the colonial framework may also play a significant role. I expect resentful whites not only to be more likely to perceive non-states as threats, I expect they will also be more likely to see more powerful countries as threats to U.S. dominance and sovereignty. This hypothesis is evaluated using the 2012 ANES and 2016 ANES questions about China as well as the 2020 ANES battery of questions which asks about the threat that six nations pose to the U.S.: China, Iran, Russia, Japan, Mexico, and Germany. This hypothesis is also tested through evaluations of terrorism in 2012 and 2016 where items ask respondents about threats posed by terrorism.
Hypothesis 3: High levels of racial resentment lead to support for the military and spending on it.
Drawing mainly on the colonial framework, Hypothesis 3 (H3) predicts that racial resentment leads to support for military spending and support for the military as an institution. Data from the ANES cumulative survey (which contains racial resentment measures from 1986 to 2020) includes several questions that address general support for the military, whether respondents believe the U.S. should spend more on the military and foreign aid, and more general questions about military policy. These items have been asked consistently over the course of the data analyzed—including the 2012, 2016, and 2020 surveys—and provide a robust set of data to test Hypothesis 3.
Data and Research Design
To assess these three hypotheses, I analyze data on racial resentment among self-identified non-Hispanic white individuals and militarized foreign policy across several recent U.S. presidential elections. The best available source of data for this test is the American National Election Survey (ANES)Footnote 1 because it is one of the only major surveys that includes measures of both racial resentment (the main independent variable in the analysis) and the major foreign policy items (the dependent variables in the analysis). The ANES has some limitations, and it is widely seen as the “gold standard” in public opinion research and contains foreign policy and racial attitude measures over long stretches of time. While boundaries of not just race and ethnicity, but who the U.S. perceives as its “enemies” on the international stage, change along racial, ethnic, religious, and other dimensions. Use of the ANES prevents true intertemporal analysis, the variation in items across the various surveys allows this study to assess the association (or lack thereof) between racial resentment and foreign policy opinion among the U.S. public across a wide spectrum of issues. This study will focus on the ANES Cumulative Study, covering 1986–2016Footnote 2 along with the 2012, 2016, and 2020 ANES surveys. Given the measures available in these datasets, I have disaggregated the analysis into three categories corresponding to the hypotheses outlined above.: H1) supporting military intervention as a policy response, H2) perceiving foreign actors as military threats, and H3) institutional support for the military. In all three cases, I expect racial resentment to be associated with increased levels of support for the militaristic option. I also include additional variables which have been demonstrated by previous research to influence foreign policy attitudes: income, partisanship, ideology, gender, ethnocentrism, authoritarianism, and social dominance orientation with each discussed briefly below.
Independent Variable
The key independent variable in this analysis is racial resentment (question wording in Appendix C). Previous research indicates that racial resentment, or so-called “symbolic racism:” captures racial attitudes, especially of white individuals, better than more overt expressions of racial attitudes (Reyna et al. Reference Reyna, Henry, Korfmacher and Tucker2006; Tesler Reference Tesler2016; Valentino and Sears Reference Valentino and Sears2005; Virtanen and Huddy Reference Virtanen and Huddy1998). Rather than being directed at one specific minority, racial resentment applies broadly to non-white minority groups. Therefore, I expect this general predisposition to influence attitudes regarding foreign policy affecting Black people around the world, but also applying to other non-white and racialized groups. For example, racial resentment has been shown to be a reliable measure of attitudes about Latinos (Abascal Reference Abascal2022; Ramirez and Peterson Reference Ramirez and Peterson2020), Asians (D. Kim Reference Kim2022), Native Americans (Foxworth and Boulding Reference Foxworth and Boulding2022), and Muslim Americans (Lajevardi and Oskooii Reference Lajevardi and Oskooii2018; d’Urso and Bonilla Reference d’Urso and Bonilla2023). Furthermore, the U.S. public applies broad generalizations about foreign countries to their foreign policy attitudes, such as assuming everyone in the Middle East is Muslim and/or Arab (Lajevardi and Oskooii Reference Lajevardi and Oskooii2018; d’Urso and Bonilla Reference d’Urso and Bonilla2023). These generalizations are key to perceptions of threat for foreign interventions, as well as assessments of incidents of terrorism (Avdan and Webb Reference Avdan and Webb2019)Footnote 3 .
Dependent Variables
The dependent variables in the analysis address support for militarized foreign policy. To demonstrate the wide-ranging impact of racialization I test my hypotheses across four major areas: 1) Iran, 2) China, 3) counterterrorism, and 4) support for the U.S. military. First, on the 2012 ANES, respondents were asked a series of five questions about the Iranian government’s pursuit of nuclear weapons and the U.S. response. Since the Islamic Revolution in 1979, the governments of the United States and Iran have had an adversarial relationship, to put it mildly. Despite this history, the 2015 Joint Comprehensive Plan of Action (JCPOA), signed by Barack Obama and ratified by the U.S. Senate, was a significant issue in the 2012 Presidential election and step toward rapprochement. However, following his election in 2016, Donald Trump quickly moved to abandon the U.S.’ treaty obligations, officially leaving the agreement in 2018. Respondents are asked for their support for the following range of policies: 1) “Iran is developing a nuclear weapon”; 2) “US should engage in direct diplomacy with Iran”; 3) “US should enact diplomatic sanctions on Iran”; 4) “US should bomb Iranian nuclear sites”; and 5) “US should invade Iran”. On the 2020 ANES, respondents are asked about the level of threat that Iran poses.
Second, to assess how racialization impacts white individuals’ views of China, I analyze the impact of racial resentment on 1) whether they perceive China as an economic threat to the United States (in 2012) and 2) if they perceive China as a military threat to the United States (2012, 2016, and 2020). This is one of the most important variables in the analysis given that it is the only survey items related to a specific country on foreign policy consistent across the 2012, 2016, and 2020 ANES studies. Given the sea change in U.S. rhetoric and policy regarding China in the last few presidential administrations, these findings are consequential for the present and future of U.S. competition with its closest rival internationally.
Third, I assess the impact of racial resentment on support for the torture of suspected terrorists and the perception of threat from terrorism. On the 2012 survey, respondents are asked about their support for torture, the likelihood of a potential attack, and their support for the war in Afghanistan. The 2016 ANES focuses on the threat of ISIS (the so-called “Islamic State”), asking about support for a more aggressive approach, including “boots on the ground” in Iraq and Syria. Respondents were also asked about their support for Syrian refugees emigrating to the United States.
Fourth and finally, I examine H3—the connection between racial resentment and support for the militarization of U.S. foreign and domestic policy. Using the ANES cumulative survey, I assess the connection between racial resentment and support for high levels of military spending, support for military intervention, and the perception that America’s position in the world is threatened. I expect racially resentful members of the U.S. public to support using even more revenue on the military. Data from the ANES cumulative survey provide the most robust data set to examine
Additional and Control Variables
Research shows that gender plays a significant role in foreign policy opinion, where respondents “deem men more competent at legislating around issues of national security,” reducing support for female presidential candidates (Lawless Reference Lawless2004, 479). Moreover, “increased threat perceptions lead men but not women to be more likely to support the use of torture” (Lizotte Reference Lizotte2017, 772). Additional research finds that “women are less likely to support the use or escalation of force, and they are more supportive of humanitarian and peacekeeping missions” (Eichenberg Reference Eichenberg2016, 147). Given evidence that policy attitudes can differ markedly between whites born in the United States and those residing here that were born outside the United States (Davis and Hendricks Reference Davis and Hendricks2007), a binary variable is included where 1=born outside the United States and 0=born in the United States. In the sample analyzed, approximately 10 percent of self-identified non-Hispanic whites were born outside the U.S. or its territories.
Significant segments of the U.S. public have authoritarian beliefs and “become susceptible to ‘authoritarian thinking’ when they perceive a grave threat to their safety” (Hetherington and Suhay Reference Hetherington and Suhay2011, 536). Social dominance orientation (SDO) has also been shown to increase the likelihood that individuals will support militaristic foreign policy (Alexander, Levin, and Henry Reference Alexander, Levin and Henry2005). SDO has been shown to be more heavily distributed among men (Pratto, Stallworth, and Sidanius Reference Pratto, Stallworth and Sidanius1997) and those of high-status groups, including those based on race and ethnicity (Pratto et al. Reference Pratto, Liu, Levin, Sidanius, Shih, Bachrach and Hegarty2000). Authoritarianism and SDO have also been shown to have a strong relationship with racism (Parker and Towler Reference Parker and Towler2019). This relationship is exemplified by research on “racial authoritarianism,” which connects the concepts of the racial contract which privileges whiteness with the often authoritarian and undemocratic practices undertaken by governments and individuals to maintain that order (Weaver and Prowse Reference Weaver and Prowse2020). The view that one’s own group is superior to others, commonly known as “ethnocentrism,” has been shown to influence foreign policy attitudes (Kinder and Kam Reference Kinder and Kam2010). Ethnocentrism is included as a control variable in all models presented in the analysis section except for Study 4 using the ANES cumulative survey where the item is not available. The ethnocentrism measure is created using an index of racial stereotypes about the ingroup and outgroups. In general, I find the impact of racial resentment to have a much more consistent and significant impact on foreign policy attitudes than ethnocentrism. Although it has some commonalities with ethnocentrism, authoritarianism, and SDO, white supremacy is a specific worldview involving a hierarchy of racial categories—and the enforcement of that hierarchy both domestically and internationally. This racial hierarchy has many purposes, key among them helping to create, maintain, and justify vast material inequality between whites and non-whites in the U.S. and around the world (Horne Reference Horne1999; Mills Reference Mills1997; Omi and Winant Reference Omi and Winant2014). White supremacy manifests itself in many ways, including in public opinion regarding anti-Black racism, discrimination, prejudice, and, most pertinently to this study, other forms of anti-Blackness such as “symbolic racism” and racial resentment (Bonilla-Silva Reference Bonilla-Silva2001). The creation and maintenance of white supremacy in the U.S. also relies on a legal hierarchy that privileges whiteness—not just protecting white property and wealth but leading to a state in which whiteness itself is seen as a valuable commodity (Harris Reference Harris1993). To address possible statistical collinearity and conceptual overlap, a correlation matrix is presented in Appendix B exploring the relationship between racism, authoritarianism/SDO, ethnocentrism, right-wing authoritarianism (RWA), isolationism, and patriotism (Sibley, Robertson, and Wilson Reference Sibley, Robertson and Wilson2006; Zakrisson Reference Zakrisson2005). While there is some overlap, there is strong evidence that these concepts are distinguishable both theoretically and empirically. Interest and attention to politics also may play a key role in foreign policy attitudes (Bullock Reference Bullock2011)—especially given that more information may allow individuals to be less reliant on the heuristic of race to form their opinionFootnote 4 . Finally, in line with previous research on public opinion discussed above, I include additional variables for liberal-conservative ideology, partisanship, income, education level, military veteran status, and age (Eichenberg Reference Eichenberg2016; Holsti Reference Holsti2004).
Findings
Overall, I find strong support for all three of my hypotheses across the four iterations of the ANES (2012, 2016, 2020, and the cumulative survey using data from 1986 to 2020). Results for each are presented below. Plotted coefficients do not show covariates for the sake of brevity. Full regression models can be found in Appendix F.
Study 1: 2012
Study 1 uses OLS regression to analyze the relationship between racial resentment and U.S. militarism among whites using the 2012 ANES in three policy areas: 1) Iranian nuclear development, 2) China’s global power, and 3) counterterrorism. Panel 1 of Figure 1 presents the results for regression analysis of white public opinion on Iranian nuclear development across five variables, as discussed above. Plotted coefficients do not show covariates for the sake of brevity but the full regression tables can be found in Appendix F. The items asked on the 2012 ANES provide face validity to my hypotheses given that the items are discrete and simple in their formulation, and respondents with even a casual knowledge of global affairs can distinguish an aggressive military operation from diplomatic talks. Furthermore, rather than vignettes or hypothetical scenarios, respondents are confronted with real-world scenarios at a time when their opinions and willingness to vote based on them could have real consequences.
Racialization of Public Opinion on Iran Regression Coefficients (2012 ANES).
Racial resentment does not have a strong relationship with respondents’ belief that Iran was attempting to develop a nuclear weapon in 2012 or willingness to engage in diplomacy, but racially resentful whites were significantly more likely to support economic sanctions (p < 0.01). Perhaps most alarmingly, racial resentment among white individuals leads to support for bombing sites in Iran suspected of being involved in nuclear development (p < 0.001) and even for a full U.S. military invasion of Iran (p < 0.001). These results provide support for Hypothesis 1 which predicts a strong relationship between racism and support for military intervention and Hypothesis 2 which predicts a strong relationship between racism and threat perception.
In Figure 2, I extend my analysis to perceptions of China and U.S. counterterrorism. The results of this analysis provide strong support for the hypothesis that racial resentment leads to the perception that non-white foreign countries are threatening. White individuals reporting high levels of racial resentment were significantly more likely to believe that China poses a military threat (p < 0.001) and an economic threat (p < 0.001) to the United States. Furthermore, white individuals who are racially resentful are also much more likely to support the use of torture against suspected terrorists (p < 0.001). Higher levels of racial resentment also lead to the belief that a major terror attack is likely in the next year (p < 0.001) in the models analyzed—again indicating support for H2—the perception of threat from non-white individuals and nations. While the R2 scores in the models are lower than ideal, given the method of analysis and the complexity of the phenomena being analyzed the scores do not seem indicative of poor model fit or lack of predictive value (Ford Reference Ford2015). Across models, the lower R2 scores often accompany more abstract or complex issues and may indicate a lack of structure, knowledge, or strong opinions among the public on issues of foreign policyFootnote 5 .
Racialization of Public Opinion on China and the Global War on Terror: Regression Coefficients (2012 ANES).
Study 2: 2016
Study 2 turns to the 2016 ANES to analyze the connection between racial resentment and support for militaristic foreign policy related to China and counterterrorism. Results from Figure 3 provide strong support for Hypotheses 1 regarding intervention and Hypothesis 2 regarding threat perception. Model 1 returns to the question of whether China poses a military threat to the U.S. that was also asked in Study 1. Again, the results provide strong support for the hypothesis that racial resentment is leads to support for militaristic policy, where high levels of racial resentment lead to the belief that China is a military threat to the U.S. (p < 0.001). Racial resentment also leads to support for torture among U.S. whites in 2016 (p < 0.001), further supporting findings from the analysis above using the 2012 ANES. Model 3 addresses public opinion regarding whether U.S. should use force internationally to solve problems, finding that there is not a strong relationship between high levels of racial resentment and support for the use of force more generally in 2016 (p = 0.421). This supports the argument advanced in this paper that when it comes to use of force, racial resentment should be much more predictive when the potential adversary is racialized as non-white. Moving to Models 4 and 5, respondents who score highly on the racial resentment battery were also significantly more likely to believe that a terror attack was likely in the next year (p < 0.05), reinforcing results from Study 1, but not associated with respondent’s assessment of its likelihood (p = 0.152). Taken together, the results from Models 4 and 5 present an interesting comparison: racially resentful whites are more likely to report a fear of a major terror attack in the U.S. within the next year, but they are not more likely to expect one to happen. These findings suggest that the connection between racism and foreign policy attitudes is more about anxiety and fear of a non-white threat internationally than a strategic calculation that such threats are likely to manifest.
Racialization of Public Opinion on Foreign Policy: Regression Coefficients (2016 ANES).
Study 3: 2020
The 2020 ANES study includes a module that asks the same question about six different countries: are (China, Russia, Iran, Germany, Japan, or Mexico) threats to the United States? In this case, the items on the 2020 ANES asking respondents about their perception of threat were asked using a five-point scale—“not at all,” “a little,” “a moderate amount,” “a lot,” and “a great deal” of threat to the United States. The survey randomly selected the order in which the questions were asked to the respondent. The question text and summary of the mean value of the responses given are presented in Appendix C. This dataset provides an opportunity for comparison between countries which are perceived as white—Russia and Germany—along with countries perceived as non-white: Japan, China, Mexico, and Iran. This dataset also provides variation in the U.S. relationship with the nations the respondents are asked about: Germany, Japan, and Mexico are major U.S. partners internationally while Russia, China, and Iran are generally seen as having a more adversarial relationship with the U.S. globally. These six items are the main dependent variables in the analysis in Study 3, with racial resentment again serving as the main independent variable of interest. Summarized results of the OLS regression analysis are presented in Figure 4.
How Much of a Threat Does (country) Pose to the United States? Regression Coefficients (2020 ANES).
The results from Figure 4 present a clear pattern and provide support for my second hypothesis regarding the perception of non-white nations as threatening: racial resentment has a significant and positive relationship with the perception that the non-white nations—Iran, China, Mexico, and Japan—are threatening to the United States. The findings regarding China in the 2020 ANES support those in Studies 1 and 2 from the 2012 and 2016 where racial resentment has a strong relationship with the perception of China as a threat to the United States (p < 0.001). The relationship between high levels of racial resentment and the perception that Iran is a threat to the U.S. (p < 0.001) reinforces the findings from the 2012 ANES above, where racial resentment was associated with support for aggressive military action against Iran. Despite Japan being a major ally of the United States, higher levels of racial resentment lead to the belief that they are a potential threat (p < 0.001). Resentful whites are also much more likely to see USMCA trade partner Mexico a potential threat to the United States (p < 0.001). Racial resentment has a much smaller effect on the perception that Germany poses a threat to U.S.—to the point of substantive insignificance. For Russia, racial resentment is not predictive of the perception that Russia is a threat. In both the German and Russian case in 2020, the role of Donald Trump cannot be ignored. Previous research has found racial resentment to be a strong predictor of support for Donald Trump (Sides, Tesler, and Vavreck Reference Sides, Tesler and Vavreck2018), while further work finds that support for Trump is more driven by outgroup hostilities toward Black, Asian, and Muslim than support for culturally or religiously motivated policy goals (Marsh Reference Marsh2021). During his presidency Trump feuded with German Chancellor Angela Merkel and repeatedly spoke of his admiration for Russian President Vladimir Putin’s leadership. While this analysis is somewhat limited in its ability to disentangle completely support for Trump, high levels of racial resentment, sexism, and support for aggressive foreign policy, to further explore this relationship I perform an analysis of similar regression equations using multinomial logit as the method of analysis. Results are provided in Appendix D. Another possibility is the complicated history of racialization, “othering,” and the Cold War, where the Soviet Union was often depicted as outside of Europe, closer to Asia, and subject to many familiar Orientalist stereotypes—even as the USSR often applied many of these same attitudes toward their imperial periphery and minority populations (Bassin Reference Bassin1991; Klein Reference Klein2003; Neumann Reference Neumann1998). For opinion on China, Iran, Mexico, Japan, and Russia, racial resentment had a stronger relationship with foreign policy attitudes than ethnocentrism, while for Germany ethnocentrism was a stronger predictor of threat perception than racial attitudes. While the analysis here supports findings from previous research that ethnocentrism is an important factor in the foreign policy attitudes of white individuals, analysis the 2012, 2016, and 2020 ANES suggests that racial attitudes have an even more significant relationship. Overall, the results from Study 3 provide strong evidence that racial resentment has a significant impact on opinion even while controlling for alternative explanations.
In addition to the item battery above, there were several more items worth exploring on the 2020 ANES. Figure 5 shows results of the relationship between racial resentment and selected control variables. High levels of racial resentment lead to the perception that U.S. status in the world was improving (p < 0.001), which makes sense considering how strongly racial resentment and Trump support have been shown to be related in the literature. This can be seen represented in the model by the strong and positive correlation between GOP partisanship and the perception that the U.S. is in a better position under Trump than Obama. Racial resentment also leads to the worry of a major terror attack in 2020, as it was in the 2016 and 2012 studies above. Racial resentment did not have a significant relationship with the belief that using military force is a good way to solve problems internationally in general—reinforcing findings from Study 2. Full results of the OLS model are presented in Appendix Table F.
Racialization of Public Opinion on Foreign Policy: Regression Coefficients (2020 ANES).
Racial resentment had a strong, negative relationship with support for U.S. membership in the international diplomatic institutions like the United Nations (p < 0.001) and North Atlantic Treaty Organization (p < 0.001) for white individuals on the 2020 ANES. These findings strongly support my hypothesis that racial resentment leads to opposition to diplomacy. Finally, racial resentment negatively predicts the belief that Russia would seek to interfere in the 2020 U.S. presidential elections. Partisanship had a stronger relationship with this belief, however, again indicating the influence of personal support for Trump on foreign policy attitudes present in 2020. Overall, the findings from Study 3 examining the 2020 ANES provide strong support for all three of my hypotheses. Hypothesis 1—that racial resentment is associated with support for militaristic foreign policy—is supported by the findings displayed in Figure 5 relating to the use of military force. Hypothesis 2—that racial resentment leads to the perception of foreign, non-white nations as threatening, is supported across the battery of items in Figure 4. And hypothesis 3, that racial resentment leads to support for the military, is supported by the findings in Figure 5 regarding U.S. status in the world under Trump’s presidency.
Study 4: 1986-2020
The racial resentment battery was asked beginning with the 1986 ANES and included nearly every year since. Including this dataset in the analysis allows me to explore several additional questions regarding the connection between racism and militarism. For example, previous literature has suggested that the election of Barack Obama leads to the racialization of many domestic policies in the U.S. (Tesler Reference Tesler2016), and the election of Donald Trump may have continued this trend (Sides, Tesler, and Vavreck Reference Sides, Tesler and Vavreck2018). One limitation of this dataset, however, is that it lacks the robust array of control variables included on each survey. Control variables for age, income, education, partisanship, political ideology, interest and attention to politics, isolationism, and gender are identical to the items on the 2012, 2016, and 2020 surveys, used in Studies 1, 2, and 3, respectively. However, only political interest (not attention) is available in the Cumulative survey. The Cumulative survey also lacks measures for religious identity—for which I have substituted church attendanceFootnote 6 —and patriotism, for which I could not identify a reliable substitute. Finally, measures for egalitarianism and authoritarianism were also absent from the Cumulative survey in frequencies that allowed for robust analysis. Overall, while the cumulative survey used in Survey 3 offers insight over a long period of time, it is correspondingly limited in the depth of questions in the data set, despite including most of the major control variables included in Study 1 and Study 2. In many ways, the three datasets are similar enough to allow for comparison, but unique enough to offer something important to each study. Together, the three datasets complement each other in providing support for my hypotheses regarding the connection between racism and militarism among the U.S. public.
Figure 6 analyzes five OLS regression models differing only in the dependent variable analyzed. Model 1 focuses on foreign aid and looks at support for cutting it. Model 2 examines opinion on the military budget—should defense spending be raised? Model 3 analyzes the policy of isolationism—that the U.S. should stay out of the rest of the world’s business. Model 4 assesses the connection between racial resentment and support for the military using a feeling thermometer. Finally, Model 5 looks at a question that is only available on limited surveys but directly addresses the central claim of the article—does the respondent support using military force to solve problems in the world?
Racialization of Foreign Policy Attitudes: Regression Coefficients (ANES Cumulative Survey 1986−2020).
Model 1 presents analysis of respondents’ support for U.S. foreign aid. They were given three options for the question on foreign aid: increase, decrease, or stay the same. The results provide clear support for the hypothesis, indicating that racially resentful respondents were significantly more likely to support cutting foreign aid (p < 0.01). Turning to Model 2, racial resentment also leads to support for increasing military spending (p < 0.001). Turning to Model 3, those reporting high levels of racial resentment are significantly more likely to report (p < 0.001) the idea that the U.S. would be better off if it kept itself out of global affairs. While this belief is rooted in U.S. tradition dating back to Washington’s farewell address, in the modern era, it manifests more often as putting U.S. interests above those of the rest of the world—a rejection of internationalism. While further analysis is necessary to examine this relationship, this pattern is consistent with the “America First” foreign policy offered by President Donald Trump and has roots dating back to the John Birch society, as discussed above. Respondents with lower education and income levels, younger respondents, and those who went to church less often were also more likely to support isolationism (p < 0.001).
Model 4 supports the analysis above, finding that high levels of racial resentment lead to positive feelings about the military (p < 0.001). The results from Model 5 provide further support for my hypothesis: racial resentment also predicts support for the belief that the U.S. should use military force to solve problems in the world (p < 0.001). This finding corroborates the findings from Study 4 above looking at the 2020 ANES. Across Models 1-5, lower levels of education predict support for the militaristic option among white individuals from 1986 to 2016. While the remaining control variables achieved significance in some models, analysis did not establish a consistent pattern.
Overall, the results from Study 4 provide additional evidence for my hypotheses as well as further support for the findings from Studies 1, 2, and 3. Racial resentment predicts support for putting U.S. interests and sovereignty above internationalism, increasing the defense budget, supporting the use of military force by the U.S. to solve problems in the world, and cutting foreign aid. While each result is notable on its own, taken together they show a clear pattern establishing the strong connection between high levels of racial resentment and support for militaristic foreign policy among white individuals from 1986 to 2016.
Conclusion and Discussion
Analysis of the ANES from 2012, 2016, 2020, and the cumulative study from 1986 to 2020 provides strong evidence for this study’s three main hypotheses. My first hypothesis, that racial resentment leads to support for military interventions against non-white actors, is supported by findings from the 2012 ANES regarding Iran, where scoring highly on the resentment scale leads to rejecting diplomacy, supporting airstrikes, and even a military invasion. It also supported by findings from the 2012, 2016, and cumulative survey regarding use of force and in the 2012 and 2016 surveys regarding terrorismFootnote 7 . My second hypothesis—that racial resentment has a strong relationship with international threat perception—is supported with evidence that racial resentment leads to perceiving China as a threat from the 2012, 2016, and 2020 ANES. Fear of terrorist attacks and support for the use of torture against suspected terrorists also provide support for this hypothesis. The hypothesis regarding the connection between threat perception is exemplified by findings from the 2020 ANES, where racial resentment had a strong, positive relationship with perceiving China, Iran, and even Mexico and Japan as military threats. My third hypotheses—that racial resentment predicts support for the military - is also well-supported with results regarding defense spending, the use of military force, and support for the military overall. The perception that U.S. power is in on the decline is also predicted by racial resentment on the cumulative survey, supporting previous work which connects white racial attitudes to international status (Mutz Reference Mutz2018).
In the era of Donald Trump and a return to “America First” politics, the findings of this study should motivate scholars to leverage experimental methods and time-series data to isolate the mechanisms that drive the associations between racism and militarism. Furthermore, future work is necessary to investigate the dynamics at play in racial, ethnic, generational, and religious cohorts. The findings in this study are limited to those identifying as white, leaving additional work needed to better understand the foreign policy opinion of the many other racial and ethnic groups present in U.S. politics. Future work could expand the findings regarding how Islamophobia informs U.S. public opinion—especially relating to Iran—and how it is related to racial resentment. Another limitation of this study is its use of measures—like the racial resentment scale—asked consistently on multiple iterations of major surveys of white public opinion. Future studies might apply either new or more robust measures of racism or measures specific to the racial or ethnic groups involved in the policy being considered and their specific histories of racialization in the U.S. and internationally. The U.S. public has shown the capacity to both find foreign policy salient and influence foreign policy electorally. However, additional work is needed to understand more about under what conditions or in what this occurs.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/rep.2024.30
