Abstract
During a period of rising partisan and racial polarization, how do partisanship, ideology, and racial antagonism influence attitudes toward reparations policies directed at the descendants of slaves? In this article, we use a wealth of public opinion data to examine trends in attitudes toward reparations and analyze the correlates of opposition to reparations proposals. We hypothesize that, given the ascendance over the past decade of a powerful racial justice movement and ensuing conservative backlash, racial attitudes should be particularly powerful in determining attitudes toward reparations. Using four original, nationally representative surveys, we show that negative racial attitudes play a central role in determining opposition to reparations, with effects that typically rival or exceed those of Republican partisanship or conservative ideology.
Marking one hundred years since the Tulsa Massacre of 1921—in which a White supremacist mob murdered as many as three hundred African Americans and destroyed a thriving economic and cultural center known as the Black Wall Street of Tulsa, Oklahoma—activists with the Black Lives Matter (BLM) movement called for reparations for African Americans as compensation for past and present racial injustices. 1 “Why Reparations?” the group asked. “[Because] Black people in the U.S. have been forced to grapple with structural discrimination— experiencing the highest rates of poverty, unemployment, low wages, health disparities, incarceration inequities, and so much more.” “Without reparations,” they continued, “the structural discrimination that purposely holds Black people back will continue to thrive … Reparations directly address the legacy of slavery, acts of violence like the Tulsa Massacre, and decade after decade of discrimination that we still deal with today” (Black Lives Matter 2021).
Arguments on behalf of reparations have a rich history in American politics, but in recent years they have received unprecedented public attention. The ascendance of the BLM movement during the 2010s, the publication of the New York Times’ 1619 Project in 2019, and the massive protests following the horrific murder of George Floyd in 2020 have forced millions of Americans to consider what is owed to African American citizens. In Congress, Democrats have pressed for legislation that would establish a national commission to study whether and how to implement reparations for African Americans (Kasperowicz 2023). Several local governments and institutions of higher learning have instituted reparation policies for members of their African American communities; and many others have debated or studied whether and how to implement reparations (Germain 2022; Hain and Mulcahy 2023). Arguably, political momentum on behalf of reparations is stronger than it has been in decades, if ever.
At the same time, however, the case for reparations has been the target of historic political, ideological, and racial backlash stemming from a broader racialized counterreaction to demands from communities of color for recognition and social justice (Jardina 2019; Miller 2021; Mason, Wronski, and Kane 2021). In Congress, Republicans have repeatedly blocked the legislation that would establish a panel to study reparations (Summers 2021). In some instances, Republicans have expressed their opposition to reparations in frankly racist terms. For example, speaking at a pro-Trump rally in October 2022, Republican Senator Tommy Tuberville of Alabama exclaimed, “[Democrats] want crime because they want to take over what you got… . They want reparations because they think the people that do the crime are owed that. Bullshit, they are not owed that” (quoted in Kim 2022).
In this intense partisan and racialized environment, understanding Americans’ attitudes toward reparations is of great importance. Today, scholars and pundits are expressing concern that extreme partisanship, ideological polarization, and racial animus are fostering a “political sectarianism” that is undermining Americans’ support for democratic norms and institutions (Finkel et al 2020; Mason 2018; Lieberman et al. 2019; Bartels 2020). Given that African Americans experience both contemporary discrimination (Pager and Shepherd 2008; Alexander 2012; Lett et al. 2021; Williams et al. 2019) and ongoing harms stemming from enslavement and Jim Crow (Coates 2014; Williams 2022; Williams, Logan, and Hardy 2021), understanding how partisanship, ideology, and racial attitudes affect attitudes toward reparations will shed light on the scope of Americans’ commitment to the key democratic proposition, enshrined in the Declaration of Independence, that “all people are created equal.”
In this article, we investigate both historic and contemporary patterns of mass opinion toward reparations. We first extend the pioneering work of Craemer (2009a), and trace public opinion on reparations from 2001 to 2023 in public opinion polls archived by the Roper Center. We find that overall opposition to reparations remains high, with disapproval strongest among White Americans (Dawson and Popoff 2004). Using more recent polls, we find preliminary evidence that White opponents of reparations ground their disapproval primarily in the beliefs that African Americans are undeserving and that achievement of formal racial equality makes reparations unnecessary.
To further investigate patterns of opposition to reparations, we turn to an examination of the determinants of opposition using four nationally representative surveys fielded between April 2021 and January 2023. We hypothesize that, given elite opposition to reparations publicly expressed by prominent Republican elected officials, Republican partisanship and conservative ideology are likely to be associated with increased opposition to reparations (Zaller 1992; Schaffner, MacWilliams, and Nteta 2018; Schaffner 2022). However, given the long associative history between negative racial attitudes and opposition to policies in which African Americans are the chief beneficiaries or advocates, we posit that negative racial attitudes will play the leading role in determining opposition to reparations (Tesler 2012, 2015; Filindra and Kaplan 2016; Cramer 2020; Jardina and Piston 2019).
Across our four surveys, we find that negative racial attitudes have a stronger and more consistent influence on opposition to reparations than do either partisan identity or ideology. Additionally, using data from our November 2022 survey, we show that, among Whites, negative racial attitudes have a stronger influence on opposition toward reparations than do other measures of racial attitudes such as in-group identification, closeness toward African Americans, or beliefs about attainment of racial equality in American society. Given these realities, overcoming opposition grounded in negative racial attitudes will likely require a sustained, mass mobilization of both African American and allied activists in support of reparations policies.
TRENDS IN ATTITUDES TOWARD REPARATIONS, 2001–2023
Do Americans support reparations directed at the descendants of slaves? In answering this question, we first examine trends in public opinion toward reparations from 2001 to 2023, using survey data from the Roper Center’s iPoll database. The iPoll database is the largest archive of public opinion survey data, with more than eight hundred thousand questions and twenty-three thousand datasets that date back to the nation’s first surveys in the 1930s. Employing this unique and comprehensive database, we searched for the terms reparations and slavery to identify pertinent questions derived from both probability and nonprobability surveys. Following Thomas Craemer (2009a), we examine support for reparations in each question, while coding for the identity of the provider of reparations (the United States, the federal government, or corporations), the intended recipients (African American descendants of enslaved people or African Americans in general), the modality (cash payments, an apology, educational support, or other), and whether the injustice of slavery was mentioned as a rationale.
Even though differences in question wording make it difficult to make precise comparisons across surveys, we can conclude that since 2001, a substantial majority of Americans oppose reparations, regardless of the modality, provider, or the recipient. However, the results in table 1 also indicate that, consistent with Craemer’s work, support for reparations is stronger when corporations are the provider, African American descendants of enslaved people are the targeted recipient, and the modality is something other than direct cash assistance.
Surveys on Slavery Reparations, 2001–2023, Roper Center for Public Opinion Research
WHY DO WHITES OPPOSE REPARATIONS? INSIGHTS ON WHITES’ BELIEFS
Next, using three recent polls from the Roper iPoll database for which full datasets are available, and combining these with survey data from the UMass Amherst Poll, we examine differences in support for reparations in the form of cash payments between Whites and African Americans, respectively, in table 2.2
Comparison of White and African American Support for Reparations in Recent Surveys
As table 2 shows, differences are very stark in support for reparations in the form of cash payments between Whites and African Americans in recent years.3 Why are so many Whites resistant to reparations? Both recent ethnographic research on the perceptions of Whites (Cramer 2016; Hochschild 2018) and quantitative research on White attitudes toward African Americans (Kinder and Sears 1981; Schuman 1997; Kinder and Sanders 1996; Tesler 2016) provide guidance on likely answers to this question. Collectively, this work reveals that many Whites believe that racial inequality is now largely a thing of the past, and therefore resent African American demands for policies that recognize and attempt to ameliorate contemporary racial inequities.
To investigate whether similar patterns hold with respect to reparations, in our January 2023 survey we asked respondents who expressed opposition to reparations their reasons for doing so. We required respondents to select only one of several options: providing cash payments or benefits would be too expensive; descendants of slaves do not deserve cash payments; African Americans are treated equally in society today; it is impossible to place a monetary value on the impact of slavery; and it would be too difficult to administer a reparations program.
Respondents’ reasoning about their opposition to reparations is revealing. Whites who take issue with reparations do not do so primarily because of concerns about perceived costs (selected by only 5 percent of White opponents) or administrative challenges (selected by 16 percent of White opponents). Instead, the most popular reason for White opposition is the belief that African Americans are undeserving of reparations (chosen by 32 percent of White opponents). Another 23 percent of White opponents of reparations base their opposition in the belief that African Americans are already treated equally in American society. In short, our findings are consistent with recent work that presents denialism of contemporary racial inequality and resentment of African American demands as the largest attitudinal obstacles to ameliorative policies among Whites.
RACIAL POLARIZATION, AFRICAN AMERICAN ACTIVISM, AND ATTITUDES TOWARD REPARATIONS
Even though opinion surveys have consistently found majoritarian opposition to reparations directed at the descendants of slaves, it remains an open question as to what individual-level factors best explain opposition. Research has investigated the influence of socioeconomic precarity on opposition (Reichelmann and Hunt 2022; Woessner and Kelly-Woessner 2006), as well as how beliefs about the impact of reparations on race relations informs individuals’ views (Campo, Mastin, and Frazer 2004; Reichelmann and Hunt 2022). Still other work has drawn attention to how the order and wording of survey questions may influence attitudes (Dawson and Popoff 2004; Craemer 2009a).
However, most research on attitudes toward reparations has centered on the influence of racial identity and racial views in shaping public opinions. Michael Dawson and Rovana Popoff (2004) find evidence not only of a stark racial divide on the issue of reparations, but that a host of racial views, most notably skepticism that African Americans are disadvantaged in American society, are associated with reduced support for reparations (see also Mazzocco et al 2006; Torpey and Burkett 2010). Another important strand of scholarship has investigated how perceptions of closeness toward African Americans influence attitudes, finding that Whites with stronger feelings of closeness with African Americans—or “self–other overlap”—are more supportive of reparations (Craemer 2009a, 2009b, 2014). A third strand of research finds that in-group identity, specifically pride in one’s own racial identity, informs individuals’ attitudes on the issue (Hunt and Reichelmann 2019; Reichelmann and Hunt 2021; Reichelmann, Roos, and Hughes 2022).
In keeping with the focus on the importance of racial perceptions in shaping public opinion toward reparations, Kamri Hudgins and colleagues (2024, this issue), using both a large representative survey of the city of Detroit alongside a national nonrepresentative survey, find that controlling for a host of demographic and political factors, Latino, White, and African American respondents in Detroit who exhibit a strong awareness of the legacy of racism in accounting for the socioeconomic status of African Americans are more likely to support reparation policies to “counter the impact of slavery and discriminatory policies.” At the national level, the authors find that perceptions of the size of the racial wealth gap as well as views of governmental responsibility structure support and opposition to reparations respectively, particularly among White respondents.
Although this research makes important contributions to our understanding of how assessments of the role of race in society, racial identification, and perceptions of closeness to African Americans influence attitudes toward reparations, it also has limitations. First, many of these studies rely on nonrepresentative convenience samples, samples of single states, or samples of single localities, limiting their ability to generalize to the U.S. adult population (Woessner and Kelly-Woessner 2006; Mazzocco et al. 2006; Reichelmann and Hunt 2021, 2022; Reichelmann, Roos, and Hughes 2022; Hudgins et al. 2024). Second, many of these studies rely on relatively mature survey data (Dawson and Popoff 2004; Mazzocco et al. 2006; Craemer 2009a, 2009b, 2014; Torpey and Burkett 2010), meaning that they are not able to shed light on how either the dramatic intensification of polarization on the basis of partisan identity, ideology, and racial attitudes (Mason 2018; Mason and Wronski 2018; Finkel et al. 2020) or renewed attention to the issue of racial inequity (Tesler 2016; Parker 2022) of the past two decades may influence attitudes toward reparations.
Most important, this research has not closely examined the impact of negative racial attitudes on public opinion toward reparations. For decades, scholars of American politics have explored the impact of negative racial attitudes in shaping American public opinion, policy preferences, campaigns, and vote choice (for reviews, see Cramer 2020; Jardina and Piston 2019; Stephens-Dougan 2020; Tesler 2016). Both ethnographic work and quantitative public opinion research indicate that many Americans, and particularly Whites, perceive that African Americans, despite having achieved (what they view as) formal equality, make unreasonable demands on government and fail to conform to norms of individual responsibility and hard work (Hochschild 2018; Kinder and Sanders 1996; Cramer 2020). More recent work adds that many Americans deny that Whites have privileges in American society and perceive that racial problems are rare (DeSante and Smith 2020). We build on this work by integrating the study of attitudes toward reparations into the broader scholarly investigation of how negative racial attitudes are eroding support for basic democratic norms in the United States (Mason 2018; Schaffner, MacWilliams, and Nteta 2018; Appleby and Federico 2018; Reny, Collingwood, and Valenzuela 2019; Mason, Wronski, and Kane 2021; Bartels 2020; Schaffner 2020; Newman et al. 2021; Nteta et al. 2023).
Scholars have consistently found that negative racial attitudes structure Americans’, and in particular, White Americans’ responses to policies such as affirmative action, criminal justice reform, and even the payment of college athletes, policies in which African Americans are perceived to be the target or the beneficiary (Gilens 2009; Peffley and Hurwitz 2002; Tesler 2012; Wallsten et al. 2017). According to these studies, when Whites perceive that the policy concerns African Americans, this directly primes White racial attitudes, and these “top of the head” considerations are brought to bear when expressing their opinion (Zaller 1992; Mendelberg 2001; Tesler 2012, 2016).
The influence of racial attitudes on policies relating to African Americans—like reparations— is likely to be particularly powerful today. In recent years, the political salience of racial issues has grown appreciably (Parker 2016, 2022; Jardina 2019). Partisan and ideological elites have increasingly made racial issues central to their electoral campaigns and political messaging, leading to ever-stronger interconnections between partisan identification, ideology, racial attitudes, and policy preferences in the mass public (Mason 2018; Mason and Wronski 2018; Finkel et al. 2020; Westwood and Peterson 2020; Englehardt 2021). For example, scholars have demonstrated that Barack Obama’s status as the nation’s first black president polarized Americans on the basis of racial attitudes, fostered the spillover of racial attitudes into ostensibly nonracial domains, and led many Whites to link “old fashioned” racist views to their partisan attachments and policy preferences (Tesler and Sears 2010; Piston 2010; Tesler 2012; Tesler 2016; Valentino, Neuner, and Vandenbroek 2018; Jardina and Traugott 2019; Jardina 2021). Donald Trump, who made implicit and explicit racial appeals central to his political persona, further exacerbated these trends (Smith and King 2021). Trump’s racialized presidency helped establish a close relationship between racial attitudes and public attitudes toward him, with those with more negative and resentful attitudes expressing stronger support (Schaffner, MacWilliams, and Nteta 2018; Reny, Collingwood, and Valenzuela 2019; Mason, Wronski, and Kane 2021). Trump also helped normalize racism among Whites, making prejudiced attitudes and racist behavior more socially acceptable (Schaffner 2020; Newman et al 2021). However, particularly in the post-Trump era, other people of color may express anti-Black racism and opposition to pro-Black policies, especially when they feel their social and economic position is threatened (Pérez, Robertson, and Vicuña 2023).
We argue that the politics of reparations represents a microcosm of this broader partisan and racial dynamic. In recent years, African American activists, organizations, and intellectuals, along with their allies, have engaged in unprecedented mobilization to highlight the ongoing reality of racial injustice and thrust conversations about reparations into the political spotlight. Most important, the rise of the BLM movement in the 2010s drew public attention to the struggle for racial justice to a degree unprecedented since the 1960s (Dunivin et al. 2022). The horrific murder of George Floyd by White Minneapolis police officer Derek Chauvin in May 2020, captured in gruesome video, spurred what was quite possibly the largest social movement action in U.S. history (Buchanan, Bui, and Patel 2020). The New York Times’ widely publicized 1619 Project, which sought to reorient the narrative of American history around the reverberating impacts of slavery, also raised attention to the issue of reparations (Hannah-Jones 2021; Silverstein 2021).
However, this wave of antiracist protest and intellectual advocacy sparked a massive backlash among conservative politicians, activists, and movement organizations, which exploited simmering resentment toward African Americans in many sectors of American society (Cramer 2016; Hochschild 2018). During his presidency, Trump consistently denied the legitimacy of the BLM movement, calling it a “symbol of hate” and suggesting that BLM “race riots” were themselves the cause of violence (Liptak and Holmes 2020; Bump 2020). In numerous states, Republicans sought to enact legislation that would criminalize BLM protest activity or protect individuals who violently interfered with BLM protests (Quinton 2021). Similarly, widely circulated claims in conservative media that 1619 Project–inspired critical race theory (CRT) curricula were being taught in American elementary and secondary schools led to enactment of anti-CRT legislation in dozens of Republican-controlled states (Frey 2022).
Thus we strongly suspect that the question of whether reparations should be granted to African American descendants of enslaved people has been drawn into the vortex of partisan and racial polarization that is plaguing contemporary American politics. We expect partisanship, ideology, and racial attitudes to play important roles in structuring American opinion toward reparations with Republicans, conservatives, and Americans who hold conservative racial views more likely to express opposition to any and all forms of reparations for African Americans.
DATA AND METHODS
To test our expectations, we rely on three nationally representative surveys of American adults fielded by YouGov between April 2021 and January 2023, along with a fourth nationally representative survey fielded as a module of the 2022 Cooperative Election Study (CES).4 Our first survey was in the field from April 21 through April 23, 2021. The second was fielded from December 14 through December 20, 2021, and the third was in the field from January 5 through January 9, 2023. For each of these three surveys, we interviewed one thousand respondents, and in each survey propensity score weights were designed to ensure that our sample was representative of the adult population with respect to age, gender, race, ethnicity, education, ideology, and region. Our CES module was in the field from September 29 to November 8, 2022. We interviewed one thousand respondents, and like the other three surveys in our study, the CES module used propensity score weights to provide a representative sample of the adult population. Mindful that people of color may express racist attitudes—especially when they feel that their social position is under threat (Pérez, Robertson, and Vicuña 2023)—we conducted the main analyses presented in this article on all respondents.5
We use several variables to measure opposition to reparations. First, on the April 2021 and January 2023 surveys, we asked respondents, “Do you think the federal government should or should not make cash payments to the descendants of slaves?” We measured opposition on a 4-point scale ranging from definitely should to definitely should not.6 To provide further insights on public attitudes toward reparations (Campo, Mastin, and Frazer 2004; Craemer 2009a), on the December 2021 survey and in the 2022 CES module we asked whether the federal government should or should not provide various forms of reparations that have been examined in previous work and have been proposed in more recent discussions of reparations programs (Craemer 2009a, 2009b; Hunt and Reichelmann 2019; Reichelmann and Hunt 2022): “make cash payments to the descendants of slaves”; “apologize to the descendants of slaves”; “provide free college tuition to the descendants of slaves”; or “provide housing assistance to the descendants of slaves.” Again, we measured opposition on a 4-point scale ranging from definitely should to definitely should not.
Our first main independent variable of interest, strength of partisan identification, was measured using the traditional 7-point scale ranging from a strong Democrat to a strong Republican. We measured our second main independent variable, the respondent’s ideological identity, with the standard 5-point scale that ranges from very liberal to very conservative.
We measured racial attitudes with several different items.7 Our first measure of racial attitudes, available on all four surveys, used items from the fear, institutionalized racism, and empathy (FIRE) scale (DeSante and Smith 2020). In our study, we used three items from the FIRE scale: “White people in the U.S. have certain advantages because of the color of their skin”; “Racial problems in the U.S. are rare, isolated situations”; and “I am angry that racism exists.” Christopher DeSante and Candice Smith (2020) demonstrate that the first two statements capture a respondent’s acknowledgment of the existence of racism in the United States, and that the third statement measures a respondent’s affective reaction to the problem of racism in American society. A scale of these three items is now a very common measure of racial attitudes in public opinion research in political science (Schaffner et al 2018; Algara and Hale 2019; Algara and Hale 2020; Benegal and Holman 2021; Schaffner 2022; Nteta et al 2023).8 On each survey, respondents indicated on a 5-point scale their level of agreement with each statement. On the April 2021 survey, the items have a scale reliability of 0.74; on the December 2021 survey, they have a scale reliability of 0.74; on the module of the 2022 CES the scale reliability is 0.71; and on the January 2023 survey, they have a scale reliability of 0.67.9
Although the FIRE index is an important measure of racial attitudes, we also advance research on the relationship between racial attitudes and opposition to reparations by exploring the potential influence of a broader range of attitudinal and behavioral measures relating to race (Cramer 2020). Racial resentment, unlike the FIRE index, has long been used to measure a respondent’s racial views of African Americans (Kinder and Sanders 1996; Kinder and Sears 1981; Tesler 2016). We measure racial resentment with the following four items: “Over the past few years, blacks have gotten less than they deserve”; “It’s really a matter of some people not trying hard enough; if blacks would only try harder they could be as well off as Whites”; “Generations of slavery and discrimination have created conditions that make it difficult for blacks to work their way out of the lower class”; and “Irish, Italian, Jewish and many other minorities overcame prejudice and worked their way up. Blacks should do the same without any special favors.” Respondents indicated their level of agreement using a 5-point scale (strongly agree to strongly disagree), and the scale reliability is 0.90.
We also examine in the 2022 module of the CES items that have been found in previous research to predict American opinion, and in particular White opinion, toward reparations. Following Ashley Reichelmann and her colleagues (Hunt and Reichelmann 2019; Reichelmann and Hunt 2021; Reichelmann, Roos, and Hughes 2022), we included a measure of a respondent’s in-group identification. Our measure borrows from the work of Ashley Jardina (2019), and asks respondents the following two questions: “How important is being (Insert R’s race) to your identity?” and “How important is it that (Insert R’s race)s work together to change laws that are unfair to (Insert R’s race)s?” Response options to these questions range from “Extremely important” to “Not all that important”, and the scale reliability of this index is 0.74.
To capture closeness to African Americans, a key focus in work by Craemer (2009a, 2009b, 2014), we used a proxy item that measures the assessment of the closeness of a respondent’s interactions with racial minorities that asks White respondents, “How distant or close are your interactions with people from racial minority backgrounds?” with response items ranging from very distant to very close.10 Finally, to measure a respondent’s perception of the achievement of racial equality, a factor identified by Dawson and Popoff (2004) as a key determinant of White opinion on reparations, we included a proxy measure that asked White respondents to indicate their level of agreement (measured from strongly agree to strongly disagree) with the following statement, “The U.S. has already made the changes necessary to give all races equal rights.” Because items that measured a respondent’s assessment of closeness with racial minorities were only asked of White respondents, we restrict our analysis of the impact of these factors to White respondents in the 2022 CES.
In all our surveys, we also measured the standard socioeconomic variables of sex, age, race, income, education, and employment status, using the question wordings and response options employed by the CES. Finally, because attitudes toward reparations may be influenced by religious views, we measured respondents’ religiosity with a 6-point indicator of the frequency of church attendance ranging from never to more than once a week.11 To ease interpretation of results, all variables were rescaled to range from 0 to 1. We estimated our statistical models using ordinary least squares (OLS) regression.12
RESULTS
We now present the results of our analyses, starting with models of opposition to cash payments, and continuing with models of opposition to various modalities of reparations.
Opposition to Federal Cash Payments to Descendants of Slaves
How do partisanship, ideology, and racial attitudes, respectively, influence public opinion toward reparations?13 Table 3 presents the results of OLS regression models of opposition to federal cash payments to descendants of slaves from our April 2021 and January 2023 surveys, respectively.
OLS Regression Models of Opposition to Cash Payments, April 2021 and January 2023 Surveys
The models indicate that Whites (b = 0.152, p < .001 in April 2021; b = 0.086, p < .001 in January 2023) and older Americans (b = 0.256, p < .001 in April 2021; b = 0.463, p < .001 in January 2023) are consistently more opposed to federal provision of cash payments, and that more religious Americans are consistently less opposed (b = -0.147, p < .001 in April 2021; b = -0.162, p < .001 in January 2023).14 Additionally, among respondents to the January 2023 survey, more affluent Americans express greater opposition to federal provision of cash payments than do Americans with lesser means (b = 0.158, p < .01).
More to the point, we find that partisanship, ideology, and especially racial attitudes are each closely related to opposition to federal provision of cash payments on both surveys. In a reflection of partisan polarization on the issue of reparations, we find that Republican partisan identification is statistically (p < .001) and substantively associated with greater opposition to federal provision of cash payments on both surveys (b = .177 and b = 0.189 in April 2021 and January 2023, respectively). Similarly, conservative ideology is associated with increased opposition at the p < .001 level on both surveys. This is a substantively notable effect: moving from very liberal to very conservative ideological self-identification is associated with a b = 0.289 increase in opposition in April 2021 and a b = 0.225 increase in January 2023, holding the other variables at their mean values.
However, the most striking result from the models in Table 3 is the very powerful influence of racial attitudes on opposition to federal reparations. More negative racial attitudes are associated with increased opposition to federal provision of cash payments at the < .001 level on both surveys. Moreover, the estimated effect is very large—with coefficient estimates of b = 0.483 in April 2021 and b = 0.444 in January 2023. Notably, the estimated effect of racial attitudes on opposition in both models is much larger than that of either partisanship or ideology, pointing to the central role of negative racial attitudes in determining opposition to reparations in the contemporary era.15
Opposition to Various Modalities of Reparations
In Table 4 we use data from our December 2021 survey to model opposition to a broader menu of reparations policies: cash payments, an apology for enslavement, free college tuition, and housing assistance.
OLS Regression Models of Opposition to Various Reparations Policies, December 2021 Survey
The results in Table 4 are generally consistent with those presented in Table 3, indicating that the patterns in public opinion we have observed are systematic to a wide range of reparations policies, and not just to cash payments. Whites, older Americans, and those with more income are consistently and significantly more opposed to each of these reparations policies; while those with more education, and more religious Americans, are consistently and significantly less opposed to each. More importantly, we again find that partisanship, ideology, and racial attitudes are closely associated with opposition to each of these policies.
Party identification is associated with opposition to cash payments (b = 0.089, p < .05), an apology (b = 0.188, p < .001), free college tuition (b = 0.129, p < .01), and housing assistance (0.079, p < .10), indicating that Republicans oppose each of these policies more strongly than Democrats. Meanwhile, ideology is associated with opposition to cash payments (b = 0.317, p < .001), an apology (b = 0.189, p < .001), free college tuition (b = 0.306, p < .001), and housing assistance (b = 0.244, p < .001). These results indicate that very conservative individuals are much more opposed to these forms of reparations than are very liberal individuals.
However, these effects are dwarfed by the influence of racial attitudes. The effect of the FIRE Index is statistically significant at the p < .001 level in each of the four models, and the substantive effect of this variable is very large, ranging from 0.553 to 0.793, holding the other variables at their mean levels. These findings provide further support for our view that negative racial attitudes are central to Americans’ attitudes toward reparations in contemporary American politics.
Examining the Effects of Competing Measures
While our results strongly suggest that racial attitudes, as measured by the FIRE Index, play a central role in predicting opposition to reparations, it is important to note that these models do not take into account other racial attitudes that may influence mass opinion on reparations. To explore the role that alternative racial attitudes may play in influencing opinion on reparations, we turn to an analysis of a module of the 2022 Cooperative Election Study.
Table 5 replicates the models presented above while adding the proxies for each of the three alternative measures of racial attitudes that have been found to predict opinion on reparations in previous work (recall, however, that this analysis is limited to Whites). We find that older Americans express greater opposition to reparations programs, particularly those involving an apology (b = .342, p < .05), tuition assistance (b = .202, p < .01), and housing assistance (b = .216, p < .01). Those with higher incomes are significantly more likely to oppose housing assistance (b = .120, p < .05) while those who have higher levels of educational attainment are more likely to support reparations aimed at providing housing assistance for the descendants of slaves (b = -. 084, p < .05). In line with our previous analysis, Republicans, more so than Democrats, express opposition to cash payments (b = .179, p < .01), an apology (b = .207, p < .01), tuition assistance (b = .215, p < .001), and housing assistance (b = .213, p < .001). Interestingly, unlike our earlier analysis, we find that neither ideology or religiosity emerges as a significant determinant of White opinion on reparations across each type of reparations policy.
OLS Regression Models of Opposition to Reparations Policies Among White Respondents, 2022 Cooperative Election Study
Do racial attitudes remain the strongest predictor of attitudes toward reparations among White Americans when controlling for racial group identification, closeness to people of color, and perceptions of racial equality? Here, we find little evidence that in 2022 the strength of identification with a respondent’s racial identity plays a role in structuring opinion toward any of the four reparations policies currently on the agenda. Similarly, we discover that assessments of the closeness to people of color do not appear to influence a White respondent’s policy preferences concerning reparations. However, we do find that our proxy that taps a respondent’s perception of the achievement of racial equality does consistently predict opposition to cash payments (b = .253, p < .001), an apology (b = .205, p < .01), tuition assistance (b = .277, p < .001), and housing assistance (b = .214, p < .001). Nonetheless, we again find that the strongest and most consistent predictor of White opinion is negative racial attitudes as measured by the FIRE index. Whites who hold the most virulent views on race, relative to Whites who exhibit more accepting racial attitudes, are more likely to oppose cash payments (b = .281, p < .001), an apology (b = .423, p < .001), college tuition assistance (b = .333, p < .001), and housing assistance (b = .398, p < .001).
To provide further confidence in our argument that negative racial attitudes drive opposition to reparations, in online appendix 8 we replace the FIRE index with the racial resentment scale (Kinder and Sanders 1996).16 Using this alternative specification, we again find that negative racial attitudes more powerfully predict opposition to reparations than do measures of White identity, racial closeness, or beliefs about achievement of racial equality.
DISCUSSION
In this study, we have argued that the subject of reparations has been pulled into the vortex of partisan and racial sectarianism that is afflicting all areas of contemporary American politics. Although this conflict has deep partisan and racialized roots, the increasing salience and divisiveness of racialized matters in American politics makes it likely that attitudes toward reparations are powerfully influenced by racial attitudes. Thus, we hypothesized that, even though Republican partisanship and conservative ideology likely increase opposition to reparations, negative racial attitudes likely have an even larger impact.
We use four original, nationally representative surveys of American adults fielded between April 2021 and January 2023 to investigate the respective influences of partisanship, ideology, and racial attitudes on opposition toward reparations. Across the four studies, we find that though Republican partisanship and conservative ideology are usually associated with increased opposition to reparations, negative racial attitudes emerge as the strongest and most consistent predictor of opposition. Additionally, using data from the 2022 Cooperative Election Study, we show that among Whites, negative racial attitudes exert a more powerful influence on opposition to reparations than other kinds of racial attitudes, such as in-group racial identification, closeness toward African Americans, or perceptions of the state of racial equality. We conclude that negative racial attitudes constitute the largest single obstacle to public support for reparations to African Americans in American politics.
Our work underscores that the politics of reparations cannot be separated from the systemic processes of partisan, ideological, and especially racial attitude polarization that are currently structuring all aspects of politics in the United States today. Enactment of federal reparations thus faces substantial obstacles. In this context, both the history of civil rights struggle in the United States and recent American experience suggest that the most promising— though challenging—path to reparations involves massive popular mobilization on behalf of this issue. Previous civil rights milestones such as the Civil Rights Act of 1964 and the Voting Rights Act of 1965 were accomplished in significant part in response to widespread protests by an intensely mobilized African American civil rights movement that gained the sympathy, and sometimes direct participation, of White Americans (Garrow 1978; Andrews and Gaby 2015). More recently, research has shown that the BLM movement, which during its peak in the summer of 2020 gained extensive involvement from sympathetic Whites (Buchanan, Bui, and Patel 2020), has produced significant changes in public beliefs and rhetoric about policing around the nation, as well as changes in police behavior at the local level (Olzak 2021; Shuman et al 2022; Campbell 2023).
However, although reparations have played a notable part in BLM rhetoric, it is yet unclear whether reparations will move to the center of the agenda of either major institutional civil rights groups or of grassroots African American activists (Smith and King 2021). This is understandable given the many crises—from racialized voter suppression to the end of affirmative action—competing for the attention of civil rights activists today. Nonetheless, it seems clear that overcoming the extraordinary partisan and racial obstacles facing reparations will require an unprecedented mobilization of ordinary Americans—both African Americans and their progressive allies.
FOOTNOTES
↵1. Of course, the call for reparations is not new, and can be traced back to the years immediately following the Civil War, when African American activists petitioned for land grants for formerly enslaved people (Berry and Blassingame 1982; Brophy 2006). More recently, numerous scholars, activists, and journalists have made eloquent appeals on behalf of reparations (Allen 1998; Fullinwider 2000; Robinson 2000; Feagin 2000; Coates 2014; Darity and Mullen 2020).
↵2. We examined attitudes toward reparations in the form of cash payments because relatively recent survey questions tended to ask about this form of support.
↵3. Differences in support among Whites and African Americans between the ABC/WaPo surveys and UMass Poll surveys are likely attributable to differences in question wordings and response options across the two sets of studies. In the ABC/WaPo surveys, the survey question is “Do you think the federal government should or should not pay money to Black Americans whose ancestors were slaves as compensation for that slavery?” To this, there are only two possible response options: should pay or should not pay. In the UMass Poll surveys, the survey question is “Do you think the federal government should or should not make cash payments to the descendants of slaves?” The response options are definitely should, probably should, probably should not, and definitely should not.
↵4. Stephen Ansolabehere and Brian Schaffner (2014) show that carefully designed surveys fielded using an opt-in online panel like that used by YouGov produce estimates that are as accurate as a telephone survey.
↵5. In online appendix 1 and our discussion of the CES results, we replicated these analyses while limiting our sample to Whites. We find that the effect of the FIRE Index, our measure of negative racial attitudes, on opposition to reparations policies among Whites is very strong, typically dwarfing the respective effects of partisanship or ideology. See https://www.rsfjournal.org/content/10/3/30/tab-supplemental.
↵6. We note that, because our survey question mentions the federal government as the provider of reparations and cash payments as the modality, overall opposition may be higher than if we suggested a nongovernmental provider or a noncash modality (Campo et al 2004; Craemer 2009a); moreover, it is possible that the relationship between the main independent variables of interest and opposition may be different than if our question wording had mentioned alternative providers or modalities. However, our question wording indicates that the targets of reparations are the “descendants of slaves,” which, as Craemer (2009a) shows, are a relatively popular recipient group relative to African Americans in general.
↵7. In modeling the relationship between racial attitudes and opinions toward race-related policies (like reparations) with observational survey data, one potential obstacle to inference is that these may all be manifest indicators of the same underlying (latent) construct. If this is the case, we would expect there to be a strong relationship between racial attitudes and opinions toward reparations; but this relationship would be tautological, rather than predictive in a meaningful way. In online appendix 2, we undertake an exploratory factor analysis to assess whether our measures of racial attitudes and our measures of opposition to reparations policies represent the same latent construct. We find that our measures of racial attitudes load on different factors than do our measures of opposition to reparations policies and conclude that our models are not tautological.
↵8. In online appendix 3, we use exploratory factor analysis to assess whether the three variables comprising the FIRE Index load on a single underlying factor on each of the surveys in our study. We find that they do. Given that the items also scale reliably, this provides empirical evidence in favor of our approach combining these items into a single scale.
↵9. In online appendix 4, we also examine the independent effect of each of the three variables comprising the FIRE index on opposition to reparations. We find that each of these variables are positive, statistically significant, and substantively strong predictors of opposition to reparations policies, though across the surveys the item “Whites Have Advantages” appears to have the largest effect of the three items.
↵10. Our measure of closeness to racial minorities is an explicit measure, and is thus quite different from Craemer’s implicit measure of closeness to African Americans. We discuss the implications of, and possible objections to, our measure of closeness to racial minorities in online appendix 5.
↵11. Because the statistical models presented in the article use observational data, there is always the possibility that coefficient estimates may be attributable in significant part to modeling decisions. To help address this problem, we used the MRobust program (Young and Holsteen 2017) to assess the impact of one critical modeling decision—the inclusion of control variables—on the coefficient estimate for our main variable of interest, the FIRE Index. The MRobust program allows users to assess the influence of the inclusion of covariates on inferences by estimating every possible model including combinations of the covariates and reporting on how the inclusion and exclusion of (combinations of) control variables affect the magnitude, direction, and statistical significance of the main variable of interest. In online appendix 6, we use MRobust to examine the effect of our inclusion of covariates on each of the models presented in the article. We find convincing evidence that our preferred estimates of the effect of the FIRE index on opposition to reparations are not unduly biased by our preferred model specifications.
↵12. In online appendix 7, we replicate our results using ordered logistic regression instead of OLS regression and our substantive results do not change.
↵13. In online appendix 9, we investigate whether patterns of relationships between racial attitudes and opposition to reparations are similar to patterns of relationships between racial attitudes and other racialized policies.
↵14. In online appendix 10, we explore why religiosity appears to be associated with increased support for reparations.
↵15. One possible objection to this inference is that, because the FIRE index includes a battery of items, it may contain less measurement error and therefore yield larger and more precise estimates than our measures of party identification and ideological orientation, respectively. To address this issue, in online appendix 4 we reestimate the models in table 3, swapping in each of the three component measures of the FIRE index for the full scale. We find that each of the three components of the FIRE index is a statistically and substantively significant predictor of opposition to federal provision of cash payments, and that the effect of the item “Whites Have Advantages” is consistently larger than that of either party identification or ideology.
↵16. See online appendix (https://www.rsfjournal.org/content/10/3/30/tab-supplemental).
- © 2024 Russell Sage Foundation. Rhodes, Jesse H., Tatishe M. Nteta, Lilliauna Hopkins, and Gregory Wall. 2024. “Why Reparations? Race and Public Opinion Toward Reparations.” RSF: The Russell Sage Foundation Journal of the Social Sciences 10(3): 30–48. https://doi.org/10.7758/RSF.2024.10.3.02. Direct correspondence to: Jesse H. Rhodes, at jrhodes{at}polsci.umass.edu, 502 Thompson Tower, 200 Hicks Way, University of Massachusetts Amherst, Amherst, MA, 01003, United States; Tatishe M. Nteta, at nteta{at}polsci.umass.edu, 524 Thompson Tower, 200 Hicks Way, University of Massachusetts Amherst, Amherst, MA 01003, United States; Lilliauna Hopkins, at lkhopkins{at}umass.edu, Thompson Lowrise 16, 200 Hicks Way, University of Massachusetts Amherst, Amherst, MA 01003, United States; Gregory Wall, at gwall{at}umass.edu, 516 Thompson Tower, 200 Hicks Way, University of Massachusetts Amherst, Amherst, MA 01003, United States.
Open Access Policy: RSF: The Russell Sage Foundation Journal of the Social Sciences is an open access journal. This article is published under a Creative Commons Attribution-NonCommercial-NoDerivs 3.0 Unported License.
REFERENCES
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.