Privilege on the Precipice: Perceived Racial Status Threats Lead White Americans to Oppose Welfare Programs

Privilege on the Precipice: Perceived Racial Status Threats Lead White Americans to Oppose... ABSTRACT Here, we integrate prior work to develop and test a theory of how perceived macro-level trends in racial standing shape whites’ views of welfare policy. We argue that when whites perceive threats to their relative advantage in the racial status hierarchy, their resentment of minorities increases. This increased resentment in turn leads whites to withdraw support for welfare programs when they perceive these programs to primarily benefit minorities. Analysis of American National Election Studies data and two survey-embedded experiments support this reasoning. In Study 1, we find that whites’ racial resentment increased beginning in 2008, the year of Barack Obama’s successful presidential candidacy and a major economic downturn, the latter a factor previously shown to amplify racial threat effects. At the same time, whites’ opposition to welfare increased relative to minorities’. In Study 2, we sought to better establish the causal effect of racial status threats. We found that experimentally presenting information suggesting that the white majority is rapidly declining increased whites’ opposition to welfare, and this effect was mediated by heightened racial resentment. Finally, in Study 3 we found that threatening whites’ sense of their economic advantage over minorities led whites to report greater opposition to welfare programs, but only if these programs were portrayed as primarily benefiting minorities, not if they were portrayed as benefiting whites. These findings suggest that whites’ perceptions that minorities’ standing is rising can produce periods of “welfare backlash” in which adoption of policies restricting or curtailing welfare programs is more likely. Introduction In recent decades, government programs providing low-income Americans with economic relief, particularly those programs commonly referred to as “welfare,” have been routinely attacked. From Ronald Reagan’s speeches describing the supposed excesses and luxurious lives of welfare recipients, to Bill Clinton’s vow to “end welfare as we know it,” few government programs have met with more vigorous and sustained criticism. This opposition is not restricted to political elites. The American public’s views of these programs have become increasingly polarized, and overall support for welfare has declined in recent years despite the hardship brought on by the Great Recession (Brooks and Manza 2013). What causes these programs to face such opposition? Here, we develop an integrated explanation of the role of racial prejudice in white Americans’ attitudes toward welfare programs that clarifies how larger social forces shape these sentiments. We argue that when whites perceive threats to their relative advantage in the racial status hierarchy, their resentment of minorities increases (Blumer 1958; Bobo 1999). This heightened resentment among whites in turn leads to greater opposition of welfare programs because these programs are perceived to primarily benefit racial minorities (Gilens 1999). Structural sources of perceived threat to white racial status include societal trends such as shifts in the majority status, political power, and economic advantages of dominant racial groups (Blalock 1967). Further, salient historical events seen as carrying symbolic significance for race relations, such as the election of the first black president (Parker, Sawyer, and Towler 2009), can create perceptions that whites’ privileges are precarious. Since scholarship finds that whites’ welfare attitudes are closely tied to prejudice against minorities (Fox 2004; Gilens 1996, 1999; Neubeck and Cazenave 2001), we propose that one source of these prejudicial attitudes is macro-level changes perceived to threaten whites’ status advantage. The present research thus clarifies how desires to maintain privileged racial status can influence white Americans’ views of anti-poverty policy. Below, we review prior explanations for Americans’ opposition to welfare, focusing on the extensive literature on the racial coding of welfare. We argue that social structural factors influence white Americans’ welfare attitudes and propose a model of whites’ attitudes synthesizing insights from historical, political, and social psychological literatures. We then test this explanation in a series of historical and experimental studies. First, we use nationally representative survey data to examine whether recent trends in Americans’ welfare attitudes are consistent with our claims that perceived threats to the racial status hierarchy should increase opposition to welfare among whites specifically. Next, we investigate whether experimentally induced threats to white Americans’ majority status or economic advantages over minorities increase whites’ opposition to welfare. We explore the intervening role of racial resentment and examine whether racial threats lead whites to oppose welfare programs only when they perceive them to disproportionately benefit minorities, but not when they perceive them to primarily benefit whites. Our research presents experimental evidence that whites’ welfare attitudes are affected by threats to the standing of whites and that these effects operate in part through heightened racial resentment. Further, we present evidence that this dynamic may help explain increased welfare opposition in the United States since 2008. Theoretically, these findings demonstrate that prior research linking whites’ racial resentment with welfare attitudes cannot be accounted for in terms of principled conservatism, without prejudice, as some have argued. They further suggest that perceptions of rising minority power, declines in whites’ relative economic status, and other perceived macro-level threats to whites’ racial status may produce periods of heightened anti-entitlement sentiment in which restrictions of the welfare state are more likely. Welfare Attitudes and the Contested Role of Racial Prejudice Understanding what factors influence Americans’ welfare attitudes is important because these attitudes impact policy at the state and national levels (Wright, Erikson, and McIver 1987) and because opposition to welfare programs has historically energized conservative movements (Lowndes 2008). Anti-poverty programs face a public opinion problem in “welfare backlash,” often mobilized by movements and elites (Reese 2005). The continuing degree of animosity toward welfare is all the more remarkable given the drastic cuts that have been made to these programs in the past two decades (e.g., Soss, Fording, and Schram 2011). Scholars have argued that racial politics help explain why Americans support social safety net programs less than the citizens of other industrialized, Western democracies (Quadagno 1994). Evidence suggests that welfare is strongly racially coded among Americans (Federico 2004; Gilens 1996; Neubeck and Cazenave 2001). Specifically among whites, opposition to welfare is associated with stereotypes of African Americans as lazy, unreliable, and lacking sexual morality (Gilens 1996; Hancock 2004). Because Americans tend to perceive welfare recipients as predominantly black (CBS 1996; Federico 2004), such stereotypes lead whites to view welfare recipients as undeserving of assistance (Gilens 1999). However, the degree to which opposition to welfare and other racialized policies reflects racial prejudice rather than conservative principles of small government and self-reliance continues to be debated (Huddy and Feldman 2009; Valentino and Brader 2011). Many scholars have claimed that the influence of racial prejudice on whites’ policy attitudes has been overstated (Gomez and Wilson 2006; Neblo 2009; Roth 1990; Sniderman and Carmines 1997; Sniderman, Crosby, and Howell 2000). For example, Sniderman and Carmines (1997) argue that racial prejudice is less important than moral arguments and policy specifics in determining whites’ support for welfare programs. Multivariate analyses correlating racial resentment or symbolic racism to policy attitudes have additionally been criticized on methodological grounds, with scholars arguing that these measures reflect conservative and individualist values rather than racial prejudice (Carney and Enos 2015; Huddy and Feldman 2009; Sniderman, Crosby, and Howell 2000). Because of the inherent uncertainty involved in drawing causal inferences from correlational data, much of the literature claiming the importance of racial resentment in whites’ opposition to welfare remains open to this critique (Sniderman and Carmines 1997; Sniderman, Crosby, and Howell 2000).1 Group Position Theory and Threats to Racial Hierarchy If racial prejudice does play a primary role in white Americans’ opposition to welfare programs, as a large though contested body of research suggests (e.g., Fox 2004; Gilens 1996, 1999; Neubeck and Cazenave 2001), what kinds of social processes and macro-level forces might lead whites to oppose these programs? Here, we emphasize the role that threats to whites’ privileged status play in sparking racial resentment. We argue that trends and events perceived as threatening the relative position of whites in the racial status hierarchy may prompt these feelings of threat, leading whites to increased opposition to welfare programs as a result of heightened racial resentment. Group position theory (Blumer 1958; Bobo 1999) proposes that larger social processes activate racial antipathy among members of higher-ranking racial groups. According to the theory, understandings of where racial groups stand in terms of social status and material rewards are historical products, formed mostly through discursive ideological activities of elites in which characterizations of different racial groups and the proper relationship between them are offered, debated, and refined. While a relatively stable sense of hierarchy need not always result, certain racial groups in particular social contexts—such as whites in the United States—often come to hold a “sense of group position,” or an expectation that their racial/ethnic group has a higher-status position than other groups and that this position is associated with greater access to economic, political, and social resources. A central claim of the theory is that prejudice primarily results from dominant group members’ desires to maintain this status advantage over other groups (Bobo 1999). Dominant group members react with racial prejudice when they sense that their group’s position—and its claim to the material or status advantages that should accrue to it—are under threat (Blumer 1958; Bobo 1999). Supporting this assertion, scholars have documented the importance of perceptions of competition and threat between racial groups in explaining prejudice (Craig and Richeson 2014a; Quillian 1995), in-group favoritism (Abascal 2015), and attitudes on explicitly racial policies like racial segregation (Bobo and Zubrinsky 1996) and affirmative action (Bobo 1998). Group position theory suggests that anti-welfare sentiment among whites should rise when there is a “felt challenge” (Blumer 1958) to whites’ group position and a corresponding increase in racial prejudice, and likewise abate when the relative advantage of whites is perceived as stable. The question, then, becomes one of identifying when the events of a particular historical period produce this sense of challenge. Prior scholarship suggests that reductions in political power, economic advantage, or population share among dominant group members can produce a sense of racial status threat (Blalock 1967), as may high-profile cultural or political events that indicate minority groups’ power or status is rising. In the contemporary American context, examples of such trends and events include the rising population share and electoral strength of racial minorities, and the election of the first nonwhite president (Parker, Sawyer, and Towler 2009). Finally, perceived threats to racial status are more likely to lead to resentment of racial minorities under conditions of macro-economic decline, as dominant group members perceive greater competition over resources and fear that minorities will overtake their group’s privileged access to these resources (Quillian 1995). In arguing that perceived threats to whites’ status prompt racial resentment, we treat group position theory and racial resentment as complementary accounts of white Americans’ racial attitudes, though these two conceptions of prejudice have long been seen as competing (e.g., Kinder and Sears 1981). Theorists of racial resentment characterize prejudice as resulting from early socialization experiences, such that prejudice is expected to be largely stable within a given individual after childhood. In contrast, group position theory characterizes prejudice as resulting from conflict between racial groups for material or symbolic resources, such that prejudice is seen as variable over time, triggered by perceptions of threat to dominant group status. Here, we argue that these two conceptions are compatible. We expect that racial resentment has both state- and trait-like properties, such that levels of resentment are shaped by socialization processes, have cross-temporal consistency, but can also rise and fall in response to novel information or experiences. In particular, following Blumer (1958) and Bobo (1999), we argue that social trends or events perceived to threaten the dominant group’s status will typically lead to increased resentment. Group Position and Welfare Attitudes As our review indicates, perceived threats to racial status have been shown to shape attitudes toward explicitly racial policies. Here, we propose that threats may also help explain whites’ attitudes toward policies, like welfare, that are formally race-blind, but perceived in racialized terms. However, prior work examining the relationship between racial threat and welfare attitudes shows mixed results. Using the common procedure of operationalizing threat as the proportion of minorities in the local context, some studies find that an increased proportion of blacks leads non-blacks to have more negative attitudes toward welfare (Johnson 2001; Luttmer 2001; Wright 1977), but other research does not find such effects (Oliver and Mendelberg 2000) or finds that the proportion of blacks only matters at some levels of analysis (Fox 2004). Further, this work has been largely observational, leaving open the possibility that unmeasured processes drove these effects. On a theoretical level, this prior work differs from our own in that it focuses on personal experiences of competition for resources. Here, however, we study the effects of symbolic threats to white Americans’ status in the larger society. These perceptions need not be a direct function of personal experience and instead can be socially constructed in the public sphere as elites discuss historical trends and events (Hopkins 2010). Further, others have noted that the emphasis on the proportion of black Americans in the local context has led scholars to leave unanalyzed historical and structural conditions that might lead to perceptions of racial threat (Brown 2013; Muller 2012; Soss and Bruch 2008). Finally, while we propose that threats to white racial status lead to welfare opposition as a result of heightened racial resentment, others have argued that racial threats lead whites to oppose welfare spending as a result of heightened in-group loyalty (e.g., Luttmer 2001). According to this argument, whites’ sense of racial group solidarity rises under conditions of threat, leading them to favor spending on programs benefiting whites rather than those seen as benefiting minorities. Similarly, O’Brien (2017) suggests that whites in diverse local contexts oppose progressive taxation and social spending not due to aversive feelings toward minorities, but because they are engaging in “opportunity hoarding,” or attempting to reserve scarce resources for members of their own racial groups. Here, we attempt to clarify the relationship between threat to racial status and white opposition to welfare by examining how macro-level trends and events can create perceptions of threat, and by investigating whether racial resentment or feelings of racial group identification drive the relationship between threat and welfare opposition. Figure 1 diagrams our explanation of the link between threats to racial status and welfare attitudes. We predict that social trends and events with symbolic importance for collective understandings of the relative status of racial groups—such as the election of the first black president and the rising minority population—threaten the perceived standing of whites, increasing their resentment of minorities (path a) and, as a result, their opposition to welfare programs (path b). In addition, the strength of these causal links is shaped by other factors. The effects of status threats on whites’ racial resentment should be particularly strong when they occur in conjunction with macro-economic downturns (path c), which can catalyze dominant group members’ sense of racial threat through heightened feelings of competition with subordinated groups for access to scarce resources. Finally, we expect that racial resentment prompted by racial threats will increase whites’ welfare opposition to the extent that individuals hold stereotypic views of welfare as largely benefiting minorities (path d). Where those stereotypic views are stronger, this link will be magnified, and where they are weaker or non-existent, the link will be reduced or mitigated entirely. Figure 1. View largeDownload slide Hypothesized effect of racial status threat in producing welfare opposition among whites. All arrows represent positive relationships Figure 1. View largeDownload slide Hypothesized effect of racial status threat in producing welfare opposition among whites. All arrows represent positive relationships Empirical Overview Based on the above, we argue that white Americans’ welfare attitudes are shaped by perceptions that significant macro-level trends and events threaten their racial status advantage. Further, we propose that this framework may help explain periods of “welfare backlash” that can impact the course of welfare state development. To analyze these claims, we first use American National Election Studies (ANES) data to examine whether our model fits the characteristics of Americans’ recent increased opposition to welfare. We expect that whites’ welfare opposition increased, relative to minorities’, beginning in 2008 because this period was charged with events likely to produce a “felt challenge” (Blumer 1958) to whites’ group position: the candidacy and election of President Obama and the beginning of the Great Recession. Next, we investigate experimentally whether perceiving that whites’ share of the US population is declining decreases support for welfare among white participants, further testing whether this effect is driven by heightened racial resentment. Finally, in Study 3 we compare the effects of threat on attitudes toward a welfare program benefiting minorities versus one that benefits whites, examining whether perceiving that whites’ economic advantages are declining leads whites to oppose welfare programs only if they believe the program primarily benefits minorities. In Study 3, we further assess whether the effects we find are due to whites’ resentment of minorities or due to heightened in-group loyalty. Our empirical strategy is designed to feature studies with a mix of internal and external validity to balance the value of demonstrating causality and real-world relevance of our findings. Study 1 relies on nationally representative survey data to offer an initial test of the historical relevance of our theory. Observational data, however, presents well-known barriers to causal inference. Studies 2 and 3 rely on non-representative samples, but—as experiments—allow us greater confidence in establishing causality. In addition, comparing responses to experimental stimuli is an unobtrusive method of examining how perceptions of decline in whites’ racial status might affect participants’ welfare views and racial resentment. Thus, these experiments are useful for examining an issue where direct self-reports of individuals’ motivations can be misleading, either because individuals lack insight on the social forces shaping their attitudes, or because social desirability concerns make them unwilling to report them reliably. Study 1: Racial Divergence in Americans’ Welfare Attitudes since 2008 We begin by analyzing ANES data to examine whether our theory is consistent with recent trends in welfare and racial attitudes in the general population. Based on our theoretical reasoning, we argue that the period beginning in 2008 has likely seen increased opposition to welfare among whites relative to minorities as a result of political and economic events perceived as threats to whites’ position in the racial status hierarchy. To be clear, we do not propose that our theory can account for all sources of historical variation in whites’ welfare attitudes, allowing us to predict absolute values of these attitudes over time. Instead, our predictions concern the racial gap in welfare attitudes as an indication of perceived racial threat felt among whites uniquely, net of other social factors. What trends or events occurring in the period beginning in 2008 might be perceived as threatening whites’ racial status? While whites continue to enjoy higher incomes, wealth, and representation in government than African Americans and Latinos, much public discourse about race in this period emphasized America’s increasing demographic diversity and the declining dominance of white Americans. Most prominently, the presidential candidacy and election of Barack Obama sparked pronouncements that the United States was entering a “post-racial” era (Tesler and Sears 2010). Further, media commentators discussed Obama’s election and re-election as resulting from the growing electoral power of racial minorities (e.g., Dougherty 2009), heightening the salience of the declining white majority. Finally, discussions of increases in demographic diversity and minorities’ political influence were coupled with experiences of economic hardship during the Great Recession. While macro-economic decline is likely to increase support for welfare among all Americans (e.g., Erikson, MacKuen, and Stimson 2002), research suggests that downturns are also likely to magnify racial threat effects (Quillian 1995). Therefore, we argue that this combination of factors increased whites’ racial resentment. While we expect that whites’ racial resentment should have risen during this time, previous research investigating changes in whites’ attitudes in the wake of Obama’s election has been mixed (Parker 2016). Some studies initially uncovered what was dubbed “The Obama Effect,” where whites showed decreased implicit bias and stereotyping of black Americans after Obama’s election (e.g., Bernstein, Young, and Claypool 2010), presumably due to exposure to Obama as a salient counter-stereotypical exemplar (Goldman and Mutz 2014). Other studies, however, have found that implicit and explicit racial attitudes remained stable (Schmidt and Nosek 2010), and yet others have found increased racial prejudice (Pasek et al. 2014). The literature therefore is divided about whether we would expect to see a rise in racial resentment beginning in 2008. Our theoretical reasoning, however, would suggest that perceptions of minorities’ increased political power during a period of macro-economic decline should lead to a sense of threat to whites’ group status. Therefore, we predict that whites’ racial resentment increased beginning in 2008, and that the racial gap between whites’ and minorities’ welfare attitudes has grown larger beginning in 2008, with whites increasing opposition to welfare relative to minorities. Method We analyze ANES data from 2000 to 2012 to investigate these claims.2 Table 1 describes the variables used in our analysis. Our primary independent variables are respondent’s race and dummy variables indicating the year the respondent was surveyed, with 2008 set as the baseline year. Table 1. Variables used in analysis of American National Election Studies Variable Item wording/Coding Welfare attitudes Should federal spending on welfare programs be increased, decreased, or kept about the same? 1 = increased; 2 = kept about the same; 3 = decreased Racial resentment Composite of four items asked on 1–5 scale where 1 = strongly agree and 5 = strongly disagree: “It’s really a matter of some people not trying hard enough; if blacks would only try harder they could be just as well off as whites.” (reverse-coded) “Generations of slavery and discrimination have created conditions that make it difficult for blacks to work their way out of the lower class.” “Over the past few years, blacks have gotten less than they deserve.” “Irish, Italians, Jewish, and many other minorities overcame prejudice and worked their way up. Blacks should do the same without any special favors.” (reverse-coded) Respondent’s race/ethnicity Dummy variables categorizing race/ethnicity as non-Hispanic white; black; or Hispanic, with respondents identifying as another race or ethnicity as baseline Survey year Dummy variables identifying survey year as 2000, 2004, or 2012, with the 2008 survey year as baseline Gender 0 = female; 1 = male Age Years Education Dummy variables categorizing education into five different categories: less than high school credential; high school credential; some college; bachelor’s degree; and advanced degree Employment status 1 = unemployed; 0 = else. Marital status 1 = married; 0 = else Class identification 1 = middle or upper class; 0 = else Income quintile Total household income coded in percentiles. 1 = 0–16 percentile; 2 = 17–33 percentile; 3 = 34–67 percentile; 4 = 68–95 percentile; 5 = 96–100 percentile Income trajectory Would you say that you are better off or worse off than you were a year ago? 1 = worse off; 0 = else Political ideology 1 = extremely liberal; 2 = liberal; 3 = slightly liberal; 4 = moderate/middle of the road; 5 = slightly conservative; 6 = conservative; 7 = extremely conservative Variable Item wording/Coding Welfare attitudes Should federal spending on welfare programs be increased, decreased, or kept about the same? 1 = increased; 2 = kept about the same; 3 = decreased Racial resentment Composite of four items asked on 1–5 scale where 1 = strongly agree and 5 = strongly disagree: “It’s really a matter of some people not trying hard enough; if blacks would only try harder they could be just as well off as whites.” (reverse-coded) “Generations of slavery and discrimination have created conditions that make it difficult for blacks to work their way out of the lower class.” “Over the past few years, blacks have gotten less than they deserve.” “Irish, Italians, Jewish, and many other minorities overcame prejudice and worked their way up. Blacks should do the same without any special favors.” (reverse-coded) Respondent’s race/ethnicity Dummy variables categorizing race/ethnicity as non-Hispanic white; black; or Hispanic, with respondents identifying as another race or ethnicity as baseline Survey year Dummy variables identifying survey year as 2000, 2004, or 2012, with the 2008 survey year as baseline Gender 0 = female; 1 = male Age Years Education Dummy variables categorizing education into five different categories: less than high school credential; high school credential; some college; bachelor’s degree; and advanced degree Employment status 1 = unemployed; 0 = else. Marital status 1 = married; 0 = else Class identification 1 = middle or upper class; 0 = else Income quintile Total household income coded in percentiles. 1 = 0–16 percentile; 2 = 17–33 percentile; 3 = 34–67 percentile; 4 = 68–95 percentile; 5 = 96–100 percentile Income trajectory Would you say that you are better off or worse off than you were a year ago? 1 = worse off; 0 = else Political ideology 1 = extremely liberal; 2 = liberal; 3 = slightly liberal; 4 = moderate/middle of the road; 5 = slightly conservative; 6 = conservative; 7 = extremely conservative Table 1. Variables used in analysis of American National Election Studies Variable Item wording/Coding Welfare attitudes Should federal spending on welfare programs be increased, decreased, or kept about the same? 1 = increased; 2 = kept about the same; 3 = decreased Racial resentment Composite of four items asked on 1–5 scale where 1 = strongly agree and 5 = strongly disagree: “It’s really a matter of some people not trying hard enough; if blacks would only try harder they could be just as well off as whites.” (reverse-coded) “Generations of slavery and discrimination have created conditions that make it difficult for blacks to work their way out of the lower class.” “Over the past few years, blacks have gotten less than they deserve.” “Irish, Italians, Jewish, and many other minorities overcame prejudice and worked their way up. Blacks should do the same without any special favors.” (reverse-coded) Respondent’s race/ethnicity Dummy variables categorizing race/ethnicity as non-Hispanic white; black; or Hispanic, with respondents identifying as another race or ethnicity as baseline Survey year Dummy variables identifying survey year as 2000, 2004, or 2012, with the 2008 survey year as baseline Gender 0 = female; 1 = male Age Years Education Dummy variables categorizing education into five different categories: less than high school credential; high school credential; some college; bachelor’s degree; and advanced degree Employment status 1 = unemployed; 0 = else. Marital status 1 = married; 0 = else Class identification 1 = middle or upper class; 0 = else Income quintile Total household income coded in percentiles. 1 = 0–16 percentile; 2 = 17–33 percentile; 3 = 34–67 percentile; 4 = 68–95 percentile; 5 = 96–100 percentile Income trajectory Would you say that you are better off or worse off than you were a year ago? 1 = worse off; 0 = else Political ideology 1 = extremely liberal; 2 = liberal; 3 = slightly liberal; 4 = moderate/middle of the road; 5 = slightly conservative; 6 = conservative; 7 = extremely conservative Variable Item wording/Coding Welfare attitudes Should federal spending on welfare programs be increased, decreased, or kept about the same? 1 = increased; 2 = kept about the same; 3 = decreased Racial resentment Composite of four items asked on 1–5 scale where 1 = strongly agree and 5 = strongly disagree: “It’s really a matter of some people not trying hard enough; if blacks would only try harder they could be just as well off as whites.” (reverse-coded) “Generations of slavery and discrimination have created conditions that make it difficult for blacks to work their way out of the lower class.” “Over the past few years, blacks have gotten less than they deserve.” “Irish, Italians, Jewish, and many other minorities overcame prejudice and worked their way up. Blacks should do the same without any special favors.” (reverse-coded) Respondent’s race/ethnicity Dummy variables categorizing race/ethnicity as non-Hispanic white; black; or Hispanic, with respondents identifying as another race or ethnicity as baseline Survey year Dummy variables identifying survey year as 2000, 2004, or 2012, with the 2008 survey year as baseline Gender 0 = female; 1 = male Age Years Education Dummy variables categorizing education into five different categories: less than high school credential; high school credential; some college; bachelor’s degree; and advanced degree Employment status 1 = unemployed; 0 = else. Marital status 1 = married; 0 = else Class identification 1 = middle or upper class; 0 = else Income quintile Total household income coded in percentiles. 1 = 0–16 percentile; 2 = 17–33 percentile; 3 = 34–67 percentile; 4 = 68–95 percentile; 5 = 96–100 percentile Income trajectory Would you say that you are better off or worse off than you were a year ago? 1 = worse off; 0 = else Political ideology 1 = extremely liberal; 2 = liberal; 3 = slightly liberal; 4 = moderate/middle of the road; 5 = slightly conservative; 6 = conservative; 7 = extremely conservative Our first dependent variable is a composite scale of four racial resentment items (Cronbach’s α = 0.75) (e.g., “It’s really a matter of some people not trying hard enough; if blacks would only try harder they could be just as well off as whites”). Our second dependent variable is a single item gauging welfare attitudes, measured on a three-point scale, asking whether federal spending on welfare programs should be increased, kept about the same, or decreased. We include controls for demographic and socioeconomic characteristics and political ideology. All analyses use recommended weighting to account for complex sampling features (DeBell 2010). Results Descriptive statistics suggest that whites’ racial resentment rose beginning in 2008 and continued rising in 2012 (figure 2). We perform an ordinary least squares regression to examine whether the rise in whites’ racial resentment is statistically reliable, controlling for socioeconomic characteristics and political ideology (Supplementary Table 1). This model shows that whites’ racial resentment rose between 2004 and 2008 (β = –0.103, p = 0.034) as well as between 2008 and 2012 (β = 0.090, p = 0.014). This pattern is consistent with our reasoning that 2008 marked the beginning of a period of increased racial status threat among white Americans that prompted greater resentment of minorities. Figure 2. View largeDownload slide Average racial resentment from 2000 to 2012 among white respondents Figure 2. View largeDownload slide Average racial resentment from 2000 to 2012 among white respondents Next, descriptive statistics suggest that relative levels of welfare opposition among whites and minorities diverged beginning in 2008 (Supplementary Figure 1). We perform a series of ordered logistic regressions to examine whether this racial divergence is statistically reliable, controlling for other factors. We first estimate a model predicting welfare attitudes by survey year, respondent’s race, and control variables (model 1, table 2). This model shows that among all Americans, regardless of race, opposition to welfare remained stable between 2004 and 2008, and then rose in 2012 (β = 0.931, p < 0.001). To examine how these trends may have varied between whites and minorities, we estimate the model again, now including a term for the interactions between white race and survey year dummy variables (model 2, table 2). Table 2. Ordered logistic regression models predicting Americans’ welfare opposition: coefficients and linearized standard errors Model 1: Model 2: Without White × Survey wave interactions With White × Survey wave interactions White 0.009 0.189 (0.137) (0.178) Black −0.821*** −0.786*** (0.156) (0.158) Hispanic −0.477** −0.432** (0.158) (0.161) Male 0.211*** 0.212*** (0.055) (0.056) Age −0.004* −0.004* (0.002) (0.002) Less than high school credential −0.363** −0.364** (0.129) (0.129) High school credential −0.060 −0.062 (0.084) (0.084) Some college 0.174* 0.169* (0.076) (0.076) Advanced degree −0.366*** −0.365*** (0.089) (0.089) Unemployed −0.518*** −0.519*** (0.120) (0.120) Married 0.070 0.070 (0.063) (0.063) Middle- or upper-class identification 0.043 0.043 (0.063) (0.063) Perceived downward income trajectory 0.063 0.064 (0.042) (0.042) Income quintile 0.243*** 0.243*** (0.027) (0.027) Conservative 0.427*** 0.427*** (0.021) (0.021) 2000 survey wave 0.487*** 1.087*** (0.106) (0.226) 2004 survey wave 0.008 0.317† (0.094) (0.186) 2012 survey wave 0.931*** 0.992*** (0.099) (0.139) White × 2000 −0.780** (0.253) White × 2004 −0.438* (0.215) White × 2012 −0.105 (0.141) N 7,696 7,696 Model 1: Model 2: Without White × Survey wave interactions With White × Survey wave interactions White 0.009 0.189 (0.137) (0.178) Black −0.821*** −0.786*** (0.156) (0.158) Hispanic −0.477** −0.432** (0.158) (0.161) Male 0.211*** 0.212*** (0.055) (0.056) Age −0.004* −0.004* (0.002) (0.002) Less than high school credential −0.363** −0.364** (0.129) (0.129) High school credential −0.060 −0.062 (0.084) (0.084) Some college 0.174* 0.169* (0.076) (0.076) Advanced degree −0.366*** −0.365*** (0.089) (0.089) Unemployed −0.518*** −0.519*** (0.120) (0.120) Married 0.070 0.070 (0.063) (0.063) Middle- or upper-class identification 0.043 0.043 (0.063) (0.063) Perceived downward income trajectory 0.063 0.064 (0.042) (0.042) Income quintile 0.243*** 0.243*** (0.027) (0.027) Conservative 0.427*** 0.427*** (0.021) (0.021) 2000 survey wave 0.487*** 1.087*** (0.106) (0.226) 2004 survey wave 0.008 0.317† (0.094) (0.186) 2012 survey wave 0.931*** 0.992*** (0.099) (0.139) White × 2000 −0.780** (0.253) White × 2004 −0.438* (0.215) White × 2012 −0.105 (0.141) N 7,696 7,696 † p < 0.10 * p < 0.05 ** p < 0.01 *** p < 0.001 Table 2. Ordered logistic regression models predicting Americans’ welfare opposition: coefficients and linearized standard errors Model 1: Model 2: Without White × Survey wave interactions With White × Survey wave interactions White 0.009 0.189 (0.137) (0.178) Black −0.821*** −0.786*** (0.156) (0.158) Hispanic −0.477** −0.432** (0.158) (0.161) Male 0.211*** 0.212*** (0.055) (0.056) Age −0.004* −0.004* (0.002) (0.002) Less than high school credential −0.363** −0.364** (0.129) (0.129) High school credential −0.060 −0.062 (0.084) (0.084) Some college 0.174* 0.169* (0.076) (0.076) Advanced degree −0.366*** −0.365*** (0.089) (0.089) Unemployed −0.518*** −0.519*** (0.120) (0.120) Married 0.070 0.070 (0.063) (0.063) Middle- or upper-class identification 0.043 0.043 (0.063) (0.063) Perceived downward income trajectory 0.063 0.064 (0.042) (0.042) Income quintile 0.243*** 0.243*** (0.027) (0.027) Conservative 0.427*** 0.427*** (0.021) (0.021) 2000 survey wave 0.487*** 1.087*** (0.106) (0.226) 2004 survey wave 0.008 0.317† (0.094) (0.186) 2012 survey wave 0.931*** 0.992*** (0.099) (0.139) White × 2000 −0.780** (0.253) White × 2004 −0.438* (0.215) White × 2012 −0.105 (0.141) N 7,696 7,696 Model 1: Model 2: Without White × Survey wave interactions With White × Survey wave interactions White 0.009 0.189 (0.137) (0.178) Black −0.821*** −0.786*** (0.156) (0.158) Hispanic −0.477** −0.432** (0.158) (0.161) Male 0.211*** 0.212*** (0.055) (0.056) Age −0.004* −0.004* (0.002) (0.002) Less than high school credential −0.363** −0.364** (0.129) (0.129) High school credential −0.060 −0.062 (0.084) (0.084) Some college 0.174* 0.169* (0.076) (0.076) Advanced degree −0.366*** −0.365*** (0.089) (0.089) Unemployed −0.518*** −0.519*** (0.120) (0.120) Married 0.070 0.070 (0.063) (0.063) Middle- or upper-class identification 0.043 0.043 (0.063) (0.063) Perceived downward income trajectory 0.063 0.064 (0.042) (0.042) Income quintile 0.243*** 0.243*** (0.027) (0.027) Conservative 0.427*** 0.427*** (0.021) (0.021) 2000 survey wave 0.487*** 1.087*** (0.106) (0.226) 2004 survey wave 0.008 0.317† (0.094) (0.186) 2012 survey wave 0.931*** 0.992*** (0.099) (0.139) White × 2000 −0.780** (0.253) White × 2004 −0.438* (0.215) White × 2012 −0.105 (0.141) N 7,696 7,696 † p < 0.10 * p < 0.05 ** p < 0.01 *** p < 0.001 The interactive model shows that the overall appearance of stability between 2004 and 2008 masks distinct trends between whites and members of other racial groups. Whites’ and minorities’ attitudes diverged beginning in 2008, as evidenced by the significant interactions between the white race variable and the 2000 and 2004 dummy variables (β = –0.780, p = 0.002 and β = –0.438, p = 0.041, respectively). In models analyzing whites’ and minorities’ responses separately, minorities showed a significant decrease in welfare opposition between 2004 and 2008 at the onset of a severe economic downturn (β = 0.379, p = 0.037; Supplementary Table 2), while whites did not. Instead, whites’ opposition to welfare increased somewhat (though not significantly) between 2004 and 2008, creating an increased racial gap in welfare attitudes beginning in 2008. Returning to the interactive model, the insignificant interaction between the white race variable and the 2012 dummy variable indicates that the racial divergence in attitudes that began in 2008 persisted in 2012, when Americans overall expressed heightened opposition to welfare. This pattern is consistent with our reasoning that perceptions of threat to whites’ racial status should lead to an increased racial gap between whites’ and minorities’ welfare attitudes, with whites increasing opposition to welfare relative to minorities. Discussion Using representative survey data, we find that the period beginning in 2008 has seen changes in Americans’ attitudes toward welfare, and these changes differ by race. While minorities show more positive attitudes toward welfare in 2008, whites’ attitudes became somewhat more negative. When all Americans increased welfare opposition in 2012, this racial gap remained statistically unchanged. The result is that 2008 marks the beginning of a widened divergence among whites and minorities’ attitudes toward welfare. In addition, we find that whites’ racial resentment rose beginning in 2008. These findings are consistent with our claim that feelings of racial threat—particularly, the perception of increased political power among minorities during a period of economic recession—helped shape whites’ welfare attitudes in recent years. Why do we observe that minorities increase support for welfare in 2008? Theoretical traditions in political economy (e.g., macro polity theory) argue that publics typically respond to economic crises with increased support for government action to reduce economic insecurity (e.g., Erikson, MacKuen, and Stimson 2002; Page and Shapiro 1992). Here, we find that minorities showed this frequently observed increase in welfare support in 2008, at the onset of a major economic downturn, but whites did not. This pattern can be understood as an across-the-board increase in welfare support in response to economic crisis being offset among whites by their response to racial status threat. Additional analyses support this interpretation, suggesting that economic anxiety experienced by whites suppressed an increase in opposition to welfare between 2004 and 2008 (Supplementary Table 3). Before controlling for economic anxiety, whites’ welfare opposition remains stable from 2004 to 2008 (model 1). However, whites reported significantly greater economic anxiety from 2004 to 2008 (model 2), and economic anxiety is associated with decreased opposition to welfare among whites (model 3). Once we control for economic anxiety, we find that whites’ opposition to welfare increased between 2004 and 2008, though this increase is only marginally significant (model 4). These results are consistent with our argument that, absent the economic hardship brought on by the Great Recession, racial status threat would have acted to increase whites’ welfare opposition from 2004 to 2008. In addition, we find that this (suppressed) increased welfare opposition is partially mediated by racial resentment (Supplementary Table 4). However, analysis of cross-sectional survey data is limited in its capacity to test the causal claims of our model, leaving other plausible interpretations of our results. Other factors besides threat to white racial status occurring in American society in this period may have driven the results we observe, including the rise of anti-tax movements (Martin 2008), the proliferation of pro-austerity economic philosophies, and the passage of the American Recovery and Reinvestment Act of 2009.3 In addition, it is possible that the racial divergence in welfare attitudes is due to some factor that specifically increased minorities’ support for welfare. For example, welfare support may have risen among minorities because of increased confidence that welfare benefits would be more fairly allocated under a coming Obama administration. While the findings we present are consistent with our theoretical predictions, barriers to causal inference posed by cross-sectional survey data mean we cannot prove that our explanation of these patterns is correct or historically accurate. Thus, we turn to experiments to craft controlled tests of our theoretical claim that racial status threats increase opposition to welfare among whites by increasing racial resentment. Study 2: Threat to Whites’ Majority Status In Study 2, we experimentally test whether the perception that whites’ status as the majority racial/ethnic group in the United States is precarious leads to welfare opposition among whites. While majority status is not inevitably correlated with social standing, racial/ethnic groups with greater population share typically have greater political power in democracies and hold greater influence over dominant cultural values (Blalock 1967). Previous research has also found that whites perceive information suggesting their majority status is on the decline as a threat to group status (Craig and Richeson 2014a; Outten et al. 2012) and react to this information with increased favoritism toward other whites (Abascal 2015). Therefore, if our reasoning is correct, we would expect white participants who are led to believe whites are losing majority status to show increased opposition to welfare compared to white participants led to believe that whites continue to have a strong numerical advantage. Further, we test whether racial status threat is producing opposition to welfare as a result of increased racial resentment. In our first experimental test, we utilize a historical trend occurring at the time we conducted our research to highlight the consequences of the declining white majority for whites’ political power. Following the re-election of President Obama, media commentators emphasized that the declining proportion of white Americans meant the white voting bloc could no longer determine elections. Thus, we conducted a survey-based experiment in which participants were presented either with information emphasizing whites’ declining population share, or with information emphasizing the persistence of the white majority, and introductory text connected this information to ongoing discussions of a changing electorate. Our hypotheses concern white participants’ reactions to a threat to majority status. However, we also include minority participants in the research design as a comparison group. If perceptions that whites’ population share is declining lead to qualitatively different results among whites and minorities, this supports our claim that this information is functioning as a threat to whites’ racial group position, not a threat that impacts all Americans similarly. Method A total of 151 participants (74 women and 77 men) were recruited via Amazon’s Mechanical Turk (MTurk). MTurk is an online marketplace where more than 100,000 individuals complete short jobs posted by “requesters.” While a self-selected group, the MTurk participant pool is more demographically diverse than the college student samples often employed in experimental research (Buhrmester, Kwang, and Gosling 2011), and results of experiments conducted on MTurk are similar to results obtained through an online platform using population-based sampling (Weinberg, Freese, and McElhattan 2014). Participants were US residents between the ages of 18 and 75 (M = 33.4 years, SD = 11.5). Most participants identified as white (75 percent), while the other 25 percent were split about evenly between those identifying as Asian, African American, and Latino. Respondents’ median household income was between $40,000 and $49,999, and 55.0 percent had at least a college degree. Participants answered a series of demographic questions, and then were asked to answer questions about one of two charts describing trends in the population share of different racial/ethnic groups in the United States, ostensibly as part of a study of quantitative reasoning and social opinions. Both graphs presented information adapted from Census Bureau population projections (US Census Bureau 2012). Introductory text noted that the last presidential election had sparked discussion of changes in the demographic make-up of the country, and so “it is important to know the results of cutting-edge research.” Participants were randomly assigned to one of two conditions meant to highlight different aspects of these trends. Participants in the “Majority Salient” condition were shown population data from 2000 to 2020, when white Americans’ population share is expected to remain about 60 percent of the population (figure 3a). In contrast, participants in the “Decline Salient” condition were shown projections from 1960 to 2060, when whites’ population share is expected to fall to about 40 percent. This longer time scale was intended to communicate a sharp, stable decline in white population share. We also included a line for the share of “total non-white” Americans intended to underscore the growth in the minority population (figure 3b). The graphs were accompanied by captions describing the trends in words (e.g., “The majority of Americans will be nonwhite in about 25 years”), reinforcing the experimental manipulation. Participants were then quizzed regarding the information; one of these questions assessed whether the experimental manipulation was effective. Figure 3. View largeDownload slide a: Chart presented in majority salient condition. b. Chart presented in decline salient condition Sources: US Census Bureau, 2011, and University of Virginia Population Study Center, 2012. Figure 3. View largeDownload slide a: Chart presented in majority salient condition. b. Chart presented in decline salient condition Sources: US Census Bureau, 2011, and University of Virginia Population Study Center, 2012. Next, participants’ welfare attitudes were measured in two ways. First, participants were told to imagine they were on a Congressional committee charged with cutting $500 million from the federal budget. They were given a list of nine spending areas including “Temporary Assistance for Needy Families (Welfare)” and asked to indicate how much they would cut from each area. Second, participants indicated their agreement with two statements adapted from the General Social Survey: “We are spending too much money on welfare” (reverse-coded) and “Public assistance is necessary to ensure fairness in our society.” These items were averaged to create a composite scale of welfare attitudes (Cronbach’s α = 0.77), centered at zero, ranging from −3 to 3. Next, racial resentment was measured using a standard scale including the above items from the ANES and four additional items (Henry and Sears 2002). These items were averaged to form a composite (Cronbach’s α = 0.87), centered at zero, ranging from −3 to 3. Finally, participants were asked other measures of political beliefs to examine alternative explanations, thanked, and debriefed. Results Four participants were excluded from analysis because they answered the manipulation check item incorrectly, leaving a total of 147 participants for analysis.4 We first tested the effect of the population trend information on participants’ welfare support and racial resentment (table 3). White participants assigned to the Decline Salient condition reported significantly greater opposition to welfare than those assigned to the Majority Salient condition. While white participants who were told that whites continue to be the largest single ethnic group in the United States proposed cutting $28 million from federal welfare spending, those told that whites’ population share is substantially declining proposed cutting $51 million. In addition, whites in the Decline Salient condition reported significantly greater opposition to welfare and higher levels of racial resentment on survey measures. Table 3. Effects of information highlighting whites’ majority population share or decline of white population share on white and minority participants’ welfare attitudes Majority Salient Decline Salient t Mean (SD) Mean (SD) White participants Money cut from welfare (in millions of dollars) 28.42 (51.41) 50.78 (59.36) 2.13* Composite Welfare Attitudes scale (points on 6-point scale) 0.55 (1.40) −0.19 (1.70) −2.50* Symbolic Racism (points on 6-point scale) −0.47 (1.03) 0.07 (1.30) 2.40* Minority participants Money cut from welfare (in millions of dollars) 52.81 (77.44) 57.25 (116.41) 0.13 Composite Welfare Attitudes scale (points on 6-point scale) 0.34 (1.56) 0.63 (1.60) 0.53 Symbolic Racism scale (points on 6-point scale) −0.09 (1.27) −0.78 (1.29) −1.56 Majority Salient Decline Salient t Mean (SD) Mean (SD) White participants Money cut from welfare (in millions of dollars) 28.42 (51.41) 50.78 (59.36) 2.13* Composite Welfare Attitudes scale (points on 6-point scale) 0.55 (1.40) −0.19 (1.70) −2.50* Symbolic Racism (points on 6-point scale) −0.47 (1.03) 0.07 (1.30) 2.40* Minority participants Money cut from welfare (in millions of dollars) 52.81 (77.44) 57.25 (116.41) 0.13 Composite Welfare Attitudes scale (points on 6-point scale) 0.34 (1.56) 0.63 (1.60) 0.53 Symbolic Racism scale (points on 6-point scale) −0.09 (1.27) −0.78 (1.29) −1.56 * p < 0.05 Table 3. Effects of information highlighting whites’ majority population share or decline of white population share on white and minority participants’ welfare attitudes Majority Salient Decline Salient t Mean (SD) Mean (SD) White participants Money cut from welfare (in millions of dollars) 28.42 (51.41) 50.78 (59.36) 2.13* Composite Welfare Attitudes scale (points on 6-point scale) 0.55 (1.40) −0.19 (1.70) −2.50* Symbolic Racism (points on 6-point scale) −0.47 (1.03) 0.07 (1.30) 2.40* Minority participants Money cut from welfare (in millions of dollars) 52.81 (77.44) 57.25 (116.41) 0.13 Composite Welfare Attitudes scale (points on 6-point scale) 0.34 (1.56) 0.63 (1.60) 0.53 Symbolic Racism scale (points on 6-point scale) −0.09 (1.27) −0.78 (1.29) −1.56 Majority Salient Decline Salient t Mean (SD) Mean (SD) White participants Money cut from welfare (in millions of dollars) 28.42 (51.41) 50.78 (59.36) 2.13* Composite Welfare Attitudes scale (points on 6-point scale) 0.55 (1.40) −0.19 (1.70) −2.50* Symbolic Racism (points on 6-point scale) −0.47 (1.03) 0.07 (1.30) 2.40* Minority participants Money cut from welfare (in millions of dollars) 52.81 (77.44) 57.25 (116.41) 0.13 Composite Welfare Attitudes scale (points on 6-point scale) 0.34 (1.56) 0.63 (1.60) 0.53 Symbolic Racism scale (points on 6-point scale) −0.09 (1.27) −0.78 (1.29) −1.56 * p < 0.05 By contrast, minority participants proposed cutting roughly the same amount of money from welfare across conditions and showed opposite trends from whites on survey measures. These trends were not statistically significant, though due to the small number of minority participants, we have low statistical power to detect significance. For our purposes it is most important to note that the trends we observed among minorities were qualitatively different from those we found among whites. We further predicted that whites’ greater opposition to welfare in the Decline Salient condition would be driven by heightened racial resentment. To test this claim, we conducted statistical mediation analyses (Baron and Kenny 1986) for both of our measures of welfare opposition among white participants. Results of mediation analyses using the budget-allocation measure and the survey measure of welfare support are given in figure 4. Results for both measures indicate that racial resentment partially mediated the effect of experimental condition. These results support our prediction that whites who were told that whites’ population share is declining reported greater opposition to welfare as a result of heightened racial resentment. Figure 4. View largeDownload slide a: Results of mediation analysis for funding-cut measure. b. Results of mediation analysis for composite welfare measure Figure 4. View largeDownload slide a: Results of mediation analysis for funding-cut measure. b. Results of mediation analysis for composite welfare measure Finally, to examine alternative explanations, we tested whether white participants in the Decline Salient condition expressed more conservative opinions on non-racial political issues (support for the Republican Party or defense spending), since evidence of social instability might lead respondents to more conservative opinions generally (Craig and Richeson 2014b; Jost et al. 2003). We found no significant effects of condition for these measures. Discussion In this study, we found that white participants responded to information suggesting that whites are losing their majority status—and hence their power as a voting bloc is declining—with increased racial resentment, which led to greater opposition to welfare. This mediation result is particularly noteworthy because it provides evidence that the link between whites’ racial resentment and welfare attitudes is not spuriously driven by individualist or conservative principles, since white participants expressed greater racial resentment under conditions of threat to racial status. Further, this effect was unique to whites, and we found no evidence that whites reported more conservative opinions on non-racial issues. We have to this point relied on self-report measures and statistical mediation analysis to establish the role of racial resentment. However, socially desirable response bias can affect the reliability of these measures (Bertrand and Mullainathan 2001). Also, statistical mediation analysis is subject to omitted variables bias, particularly if unmeasured mediators of an effect are correlated with the claimed mediator (Green, Ha, and Bullock 2010). Therefore, these correlational methods leave open the possibility that we have incorrectly identified the causal pathway through which racial status threats lead whites to greater welfare opposition. Thus, in our next study, we sought to establish experimentally whether the observed effects of status threat on whites’ welfare attitudes were driven by racial resentment. Study 3: Entitlement Programs Benefiting Whites versus Minorities As highlighted in figure 1, we expect that racial threats should stimulate anti-welfare sentiment among whites to the extent that they perceive these programs to primarily benefit racial-ethnic minorities (path d). We have to this point assumed that participants typically perceive welfare as disproportionately benefiting minorities, consistent with past research (Federico 2004; Gilens 1999). However, a more rigorous approach to evaluating our argument would involve systematically manipulating both whites’ sense of racial status threat and whether a given welfare program primarily benefits whites or minorities. If our reasoning is correct, we would only expect threats to increase whites’ opposition to programs believed to primarily benefit minorities, not programs believed to benefit whites. To this end, in Study 3 we sought to experimentally manipulate participants’ perceptions of whether specific federal welfare programs benefited whites or minorities by presenting them with information about the demographic breakdown of program beneficiaries. We also tested a new form of racial status threat, randomly varying whether participants were told the racial income gap had declined or grown. If threatened whites show heightened opposition to a welfare program that benefits minorities, but not to a program that benefits whites, it would be convincing evidence that racial status threats lead whites to oppose welfare as a result of heightened racial antipathy. Further, in this study we examine whether the effects we have so far attributed to racial resentment may be better explained by increased racial group identification and “opportunity hoarding” (Luttmer 2001; O’Brien 2017). If the effects of racial status threat are primarily attributable to heightened in-group identification, we would expect the strongest effects of our economic threat manipulation on whites’ support for a program disproportionately benefiting whites. However, if racial resentment sparked by racial threat is a primary mechanism, we would expect the strongest effects for whites’ support for a program disproportionately benefiting minorities. Additionally, we measure participants’ racial identification following the racial threat manipulation to examine whether heightened identification among whites mediates any observed effects. Method A total of 250 participants (129 men and 121 women) were recruited via MTurk. Participants were US residents between the ages of 18 and 68 (M = 35.6 years, SD = 12.7). Participants mainly identified as white (80.4 percent); 8.8 percent of participants identified as African American, 6.4 percent as Asian, and 4.4 percent as Latino. Respondents’ median household income was between $40,000 and $49,999, and 53.8 percent had at least a college degree. Participants answered a series of demographic questions. Then they were asked to answer questions about one of two charts describing income trends of different racial/ethnic groups during the last recession, ostensibly as part of a study of quantitative reasoning and social opinions. Participants were randomly assigned to one of two conditions. Participants assigned to the “Gap Expanding” condition saw information suggesting that while white Americans’ income suffered at the onset of recession, they had returned to previous levels by 2010. Meanwhile, black and Latino Americans’ incomes declined, such that the racial income gap had widened (figure 5a). Participants assigned to the “Gap Closing” condition were shown data indicating that whites’ incomes had fallen steadily during the recession while the incomes of other groups stayed about the same, such that the racial income gap had shrunk significantly (figure 5b). The graphs were accompanied by captions describing the trends in words (e.g., “From 2007 to 2010, the median (or typical) household income for white Americans dropped…”), reinforcing the information. Figure 5. View largeDownload slide a: Chart presented in gap expanding condition. b. Chart presented in gap closing condition Figure 5. View largeDownload slide a: Chart presented in gap expanding condition. b. Chart presented in gap closing condition Next, participants were presented with information about the Temporary Assistance for Needy Families (TANF) program and the unemployment insurance program, ostensibly to provide context to inform participants’ evaluations. In these program descriptions, the racial composition of program beneficiaries was varied such that each participant evaluated one program that primarily benefited whites and another that benefited minorities. Half the participants were randomly assigned to see information that TANF beneficiaries are “…mostly African American (40 percent) or Latino (30 percent)…” and unemployment insurance beneficiaries are “…mostly white (77 percent)….”; the other half of participants saw the opposite information, that is, that unemployment insurance mostly benefits minorities, and TANF mostly benefits whites. Information on beneficiaries’ age, gender, and race was presented graphically to increase the impact of the manipulation. The order of program presentation was randomly counterbalanced. Following each program description, participants indicated their agreement on seven-point scales with four statements measuring support for the program and whether they saw it as a good use of government resources (e.g., “I support the [unemployment insurance/TANF (Temporary Assistance for Needy Families)] program”). These items were averaged to create a composite measure of program support (Cronbach’s α = 0.90) centered at 2, ranging from −1 to 5. Next, participants’ racial identification was measured with the “identity” and “private regard” subscales of a standard measure of collective self-esteem (Luhtanen and Crocker 1992). Participants indicated their agreement on seven-point scales with eight questions measuring how strongly they identified with their racial/ethnic group and how positively they felt about their racial/ethnic group (e.g., “My racial/ethnic group is an important reflection of who I am”). These items were averaged to create a composite (Cronbach’s α = 0.83), centered at 0, ranging from −3 to 3. Following recent research (Berinsky, Margolis, and Sances 2013), we used a more rigorous approach than in our previous study to identify participants who were sufficiently attentive to the study. Rather than relying on one survey item as we did in Study 2, here we included five screening questions throughout the study. Participants were coded as inattentive if they answered more than three questions incorrectly. We gauged suspicion regarding our stimulus materials by asking participants what they thought this study was about. Finally, participants were thanked and debriefed. Results Twelve participants were dropped from analysis because they reported suspicion of the study’s purpose, and two participants were dropped for being insufficiently attentive to the study, leaving a total of 236 participants for analysis.5 We model participants’ program support using a random-effects model where the independent variables are assignment to the Gap Expanding or Gap Closing condition (between-subjects factor) and whether the program was described as benefiting whites or minorities (within-subjects factor). We also include a dummy variable for whether the program was TANF or unemployment insurance. We first model whites’ program support (model 1 in table 4). These participants reported significantly less support for TANF than for unemployment insurance net of condition. We found no significant main effects of exposure to the Gap Closing condition or the racial composition of program beneficiaries. There was, however, a significant, negative interaction between assignment to the Gap Closing condition and the race of beneficiaries (p = 0.04), such that whites showed uniquely low support for programs that benefited minorities if they had been told that the white income advantage is closing. Table 4. Random-effects models predicting white and minority participants’ support for a social safety net program by experimental conditions: coefficients and standard errors Whites’ program support Minorities’ program support (1) (2) Program benefits minorities 0.130 0.258 (0.115) (−0.068) Gap Closing condition −0.057 0.380 (0.216) (0.410) Gap Closing condition × Program benefits minorities −0.338* −0.294 (0.165) (0.369) Program was TANF −0.305*** −0.054 (0.083) (0.185) Intercept 2.977*** 2.785*** (0.155) (0.299) N 189 47 Within-subjects R2 0.09 0.05 Between-subjects R2 0.01 0.01 Whites’ program support Minorities’ program support (1) (2) Program benefits minorities 0.130 0.258 (0.115) (−0.068) Gap Closing condition −0.057 0.380 (0.216) (0.410) Gap Closing condition × Program benefits minorities −0.338* −0.294 (0.165) (0.369) Program was TANF −0.305*** −0.054 (0.083) (0.185) Intercept 2.977*** 2.785*** (0.155) (0.299) N 189 47 Within-subjects R2 0.09 0.05 Between-subjects R2 0.01 0.01 Note: Reference categories are “Program benefits whites” and “Gap Expanding” conditions; * p < 0.05 ** p < 0.01 ***p < 0.001. Table 4. Random-effects models predicting white and minority participants’ support for a social safety net program by experimental conditions: coefficients and standard errors Whites’ program support Minorities’ program support (1) (2) Program benefits minorities 0.130 0.258 (0.115) (−0.068) Gap Closing condition −0.057 0.380 (0.216) (0.410) Gap Closing condition × Program benefits minorities −0.338* −0.294 (0.165) (0.369) Program was TANF −0.305*** −0.054 (0.083) (0.185) Intercept 2.977*** 2.785*** (0.155) (0.299) N 189 47 Within-subjects R2 0.09 0.05 Between-subjects R2 0.01 0.01 Whites’ program support Minorities’ program support (1) (2) Program benefits minorities 0.130 0.258 (0.115) (−0.068) Gap Closing condition −0.057 0.380 (0.216) (0.410) Gap Closing condition × Program benefits minorities −0.338* −0.294 (0.165) (0.369) Program was TANF −0.305*** −0.054 (0.083) (0.185) Intercept 2.977*** 2.785*** (0.155) (0.299) N 189 47 Within-subjects R2 0.09 0.05 Between-subjects R2 0.01 0.01 Note: Reference categories are “Program benefits whites” and “Gap Expanding” conditions; * p < 0.05 ** p < 0.01 ***p < 0.001. When evaluating programs they believed primarily benefited whites, white participants who were told the racial income gap was shrinking did not differ much in their program support from white participants who were told the income gap was widening. However, when the program primarily benefited African Americans and Latinos, whites in the Gap Closing condition reported less support than in the Gap Expanding condition (figure 6a). Figure 6. View largeDownload slide a: Means of whites’ support for social safety net programs, by experimental condition. b: Means of minorities’ support for social safety net programs, by experimental condition Figure 6. View largeDownload slide a: Means of whites’ support for social safety net programs, by experimental condition. b: Means of minorities’ support for social safety net programs, by experimental condition This pattern suggests that the significant interaction found in model 1 is driven by decreased support for a welfare program benefiting minorities, not increased support for a program benefiting whites. In addition, we found no significant effect of the Gap Closing condition on white participants’ racial identification (p > 0.58), and racial identification did not moderate the interaction between the Gap Closing condition and race of program beneficiaries (p > 0.95) (results available upon request). Next, we turn to minority participants to examine whether these results are unique to whites’ threat response. Again we found no significant main effects of our experimental manipulations (model 2 in table 4). We found a negative, though insignificant (p = 0.43), interaction of exposure to the Gap Closing condition and the race of program beneficiaries. Minorities supported a welfare program more if they were in the Gap Closing condition and the program in question benefited whites (figure 6b). While statistically insignificant, this interactional pattern may have been significant with a larger sample of minority participants. If this finding proved reliable, it would be open to several explanations. Minorities might anticipate whites’ negative responses to status threats, increasing their support for programs that benefit whites as a gesture of appeasement. Alternately, minorities (like Americans generally) may be more likely to see whites as members of the “deserving poor,” and thus increase support for programs supporting whites when whites appear to need them. For our purposes, it is sufficient to note that minority participants responded in a different manner from white participants, consistent with our claim that information on narrowing income differences was experienced as a racial threat by whites. Discussion In this study, we provide evidence supporting the final link in our logic: that racial threats lead to anti-welfare sentiment among whites because they perceive such programs to mostly benefit minorities. Where a welfare program was portrayed as primarily benefiting whites, threatened white participants reported almost identical support for welfare as unthreatened white participants. These findings provide discerning support for our claim that whites’ opposition to welfare following racial threat is due to increased racial resentment. The within-subjects design of Study 3 further strengthens our claims that racial status threats uniquely increase whites’ opposition to programs benefiting minorities. Rather than randomly assigning some participants to evaluate a welfare program that benefited whites and other participants to evaluate a program that benefited minorities, all participants rated both types of programs. Thus, the same threatened white participants who expressed opposition to a program benefiting minorities on average expressed greater support for a program benefiting whites, within the same survey setting. While white Americans often engage in impression management to appear color-blind, such efforts did not overcome the effect of racial status threat on participants’ willingness to voice heightened opposition to programs that benefit minorities. General Discussion Our findings provide consistent support for our claim that white Americans’ welfare attitudes are shaped by concerns about the status of their racial group in American society. First, we found that whites’ and minorities’ welfare attitudes diverged in 2008, the year of the candidacy and election of President Obama and the financial crisis, and whites’ racial resentment rose during this time as well. Next, we found that white Americans who saw a demographic report emphasizing the decline of the white majority tended thereafter to voice greater opposition to welfare, and this effect was partially mediated by increased racial resentment. In our final study, we found that information threatening the white economic advantage resulted in increased opposition to welfare programs when whites perceived those programs to primarily benefit minorities, but did not affect support for programs portrayed as benefiting whites. These findings implicate racial status threats as a causal factor shaping whites’ opposition to welfare. Results of our studies are consistent with our model of the role of racial prejudice in whites’ attitudes toward welfare, more so than rival theoretical explanations. For example, a critic could argue that the threats we studied led respondents toward a more diffusely conservative political orientation because they indicated social instability (Jost et al., 2003), not because they threatened the racial status hierarchy. However, in Studies 1 and 2 we found no effects of threat on measures of non-racial attitudes. Further, this explanation cannot account for our findings that threat effects were unique to white respondents, nor why whites’ support for social welfare programs following threats depended critically on the perceived race of program beneficiaries. Nevertheless, the current research has limitations we hope future research will address. While our experimental evidence provides strong support for a causal link between racial status threats and welfare opposition, these experiments are potentially limited in their generalizability due to the use of convenience samples for practical reasons. Ideally, our experimental findings would be replicated using a nationally representative sample. Further, our focus here on whites’ attitudes means our work offers limited insights on how indications that the racial status hierarchy is changing might affect minorities’ attitudes toward welfare. Further empirical and theoretical work considering how concerns about racial group status may affect minorities’ welfare attitudes is a promising avenue for future research. Finally, we would welcome exploration of the sociodemographic and contextual factors that may affect whites’ perceptions of, or responses to, racial status threat. This variation may prove useful in explaining established patterns in public opinion and political mobilization around a number of racialized policy issues. Our research provides a novel theoretical synthesis of the role of prejudice in affecting white Americans’ attitudes toward welfare and offers a perspective for understanding a current, consequential trend in public opinion. We provide evidence that racial resentment rises in response to macro-historical trends that threaten whites’ standing in the racial status hierarchy, particularly in eras of economic decline. We demonstrate that racial resentment rather than in-group identification drives the relationship between racial threat and opposition to welfare. Further, because public attitudes partially drive developments in anti-poverty policy, these findings suggest that perceptions of rising minority power, declines in whites’ relative socioeconomic status, or other perceived macro-level threats to whites’ racial status may provoke adoption of more restrictive welfare regimes. Previous scholars have noted that minorities’ civil rights victories, increasing migration of nonwhites into state jurisdictions, and episodes of racial conflict have spurred periods of anti-welfare sentiment among whites and state policies to roll back social safety net programs (Brown 2013; Quadagno 1994; Reese 2005; Soss, Fording, and Schram 2011). The current research specifies why and how such periods of “welfare backlash” may occur, making adoption of policies that restrict or curtail welfare state development more likely. In addition, our findings demonstrate that the documented link between whites’ racial resentment and opposition to welfare programs cannot be explained in terms of principled conservatism, with no role for racial prejudice. We present a model of racial threat that moves beyond the standard operationalization as proportion of minorities in the local context, a contribution earlier scholars have noted is needed in this literature (Brown 2013; Muller 2012; Soss and Bruch 2008) and which may prove relevant to other domains of racialized public policy. We demonstrate that the relationship between racial resentment and welfare opposition remains robust even in the post-welfare reform era, when benefits have been reduced significantly and are subject to stringent sanctions (Soss, Fording, and Schram 2011). Finally, by examining how salient events may lead to perceptions of racial status threat among whites, we provide evidence for a social psychological mechanism promoting the durability of racial inequality through individuals’ political responses. Because status rank is hierarchical and zero-sum, any increases in economic or political power of lower-status groups can be interpreted as a threat to the relative standing of dominant group members. Thus, any progress toward equality may provoke resentment on the part of dominant group members, who may react politically in ways that undermine or even reverse progress to racial equality. In the case of American social welfare programs, this further implies that evidence of increased racial equality could exacerbate overall economic inequality. As whites attempt to undermine racial progress they see as threatening their group’s status, they increase opposition to programs intended to benefit poorer members of all racial groups. Footnotes 1 While differences exist in the conceptualization of symbolic racism and racial resentment, they have been operationalized similarly over time, such that most analysts treat them as essentially interchangeable concepts (e.g., Sears and Henry 2005). 2 We begin our study period in 2000 because this was the first year ANES administered questions on welfare attitudes and racial resentment after the passage of the Personal Responsibility and Work Opportunity Act in 1996. Results of analyses including earlier survey waves are substantively identical to those presented here. 3 Brooks and Manza (2013) argue that in the face of economic disruption, partisan identifiers looked for cues on how to interpret the situation from partisan sources, leading to greater opposition to welfare among Republicans especially. This explanation would be incompatible with our own if Republicans of all racial/ethnic groups increased welfare opposition, suggesting the effect is not due to whites’ perceptions of racial threat. However, if white Republican increased welfare opposition as a result of heightened racial resentment, this would suggest that both mechanisms help explain increased welfare opposition. Additional analyses support the latter, complementary account. Race and ideology appear to interact in producing opposition to welfare, such that white conservatives show increased racial resentment and welfare opposition in the post-2008 period. 4 All predicted effects remained significant (p < 0.05) if these participants were retained in analysis. 5 All predicted effects remained significant (p < 0.05) if inattentive participants were retained in analysis. About the Authors Rachel Wetts is a PhD candidate in Sociology at the University of California, Berkeley. Her research interests include social psychology, American political culture, and organizational sociology. Ongoing projects use experimental and quantitative methods to investigate how political elites activate and harness feelings of racial prejudice, and how cultural and organizational processes affect which conceptions of climate change become dominant in mainstream media. Robb Willer is a Professor of Sociology at Stanford University. His research interests include politics, morality, cooperation, and hierarchy. His recent research has focused on sources of political division and mechanisms for creating political consensus. His research has appeared in American Sociological Review, American Journal of Sociology, Social Forces, Annual Review of Sociology, Journal of Personality and Social Psychology, Proceedings of the National Academy of Sciences, and elsewhere. Supplementary Material Supplementary material is available at Social Forces online. References Abascal , Maria . 2015 . “ Us and Them: Black-White Relations in the Wake of Hispanic Population Growth .” American Sociological Review 80 ( 4 ): 789 – 813 . Google Scholar CrossRef Search ADS Baron , Reuben , and David Kenny . 1986 . “ The Moderator–Mediator Variable Distinction in Social Psychological Research: Conceptual, Strategic, and Statistical Considerations .” Journal of Personality and Social Psychology 51 ( 6 ): 1173 – 82 . Google Scholar CrossRef Search ADS Berinsky , Adam , Michele Margolis , and Michael Sances . 2013 . “ Separating the Shirkers from the Workers? Making Sure Respondents Pay Attention on Self‐Administered Surveys .” American Journal of Political Science 58 ( 3 ): 739 – 53 . Google Scholar CrossRef Search ADS Bernstein , Michael , Steven Young , and Heather Claypool . 2010 . “ Is Obama’s Win a Gain for Blacks? Changes in Implicit Racial Prejudice Following the 2008 Election .” Social Psychology 41 ( 3 ): 147 – 51 . Google Scholar CrossRef Search ADS Bertrand , Marianne , and Sendhil Mullainathan . 2001 . “ Do People Mean What They Say? Implications for Subjective Survey Data .” American Economic Review 91 ( 2 ): 67 – 72 . Google Scholar CrossRef Search ADS Blalock , Hubert . 1967 . Toward a Theory of Minority-Group Relations . New York : Wiley . Blumer , Herbert . 1958 . “ Race Prejudice as a Sense of Group Position .” Pacific Sociological Review 1 ( 1 ): 3 – 7 . Google Scholar CrossRef Search ADS Bobo , Lawrence . 1998 . “ Race, Interests, and Beliefs About Affirmative Action: Unanswered Questions and New Directions .” American Behavioral Scientist 41 ( 7 ): 985 – 1003 . Google Scholar CrossRef Search ADS ——— . 1999 . “ Prejudice as Group Position: Microfoundations of a Sociological Approach to Racism and Race Relations .” Journal of Social Issues 55 ( 3 ): 445 – 72 . Google Scholar CrossRef Search ADS Bobo , Lawrence and Camille Zubrinsky . 1996 . “ Attitudes on Residential Integration: Perceived Status Differences, Mere In-Group Preference, or Racial Prejudice? ” Social Forces 74 ( 3 ): 883 – 909 . Google Scholar CrossRef Search ADS Brooks , Clem , and Jeff Manza . 2013 . “ A Broken Public? Americans’ Responses to the Great Recession .” American Sociological Review 78 ( 5 ): 727 – 48 . Google Scholar CrossRef Search ADS Brown , Hana . 2013 . “ Racialized Conflict and Policy Spillover Effects: The Role of Race in the Contemporary US Welfare State .” American Journal of Sociology 119 ( 2 ): 394 – 443 . Google Scholar CrossRef Search ADS Buhrmester , Michael , Tracy Kwang , and Samuel Gosling . 2011 . “ Amazon’s Mechanical Turk A New Source of Inexpensive, Yet High-Quality, Data? ” Perspectives on Psychological Science 6 ( 1 ): 3 – 5 . Google Scholar CrossRef Search ADS Carney , Riley , and Ryan Enos . 2015 . “Conservatism, Just World Belief, and Racism.” Paper presented at the Annual Meeting of the Midwest Political Science Association, Chicago, IL, April 16–19. CBS . 1996 . CBS News Survey, October. Craig , Maureen , and Jennifer Richeson . 2014 a. “ More Diverse Yet Less Tolerant? How the Increasingly Diverse Racial Landscape Affects White Americans’ Racial Attitudes .” Personality and Social Psychology Bulletin 40 : 750 – 61 . Google Scholar CrossRef Search ADS ——— . 2014 b. “ On the Precipice of a ‘Majority-Minority’ America: Perceived Status Threat from the Racial Demographic Shift Affects White Americans’ Political Ideology .” Psychological Science 25 : 1189 – 97 . Google Scholar CrossRef Search ADS DeBell , Matthew . 2010 . “How to Analyze ANES Survey Data.” ANES Technical Report Series No. nes012492. Palo Alto, CA, and Ann Arbor: Stanford University and the University of Michigan. Dougherty , Conor . 2009 . “Minority Turnout Was Critical to Obama’s Election, Data Show.” Wall Street Journal, July 21. http://online.wsj.com/articles/SB124813588601666995. Erikson , Robert , Michael MacKuen , and James Stimson . 2002 . The Macro Polity . Cambridge : Cambridge University Press . Federico , Christopher . 2004 . “ When Do Welfare Attitudes Become Racialized? The Paradoxical Effects of Education .” American Journal of Political Science 48 ( 2 ): 374 – 91 . Google Scholar CrossRef Search ADS Fox , Cybelle . 2004 . “ The Changing Color of Welfare? How Whites’ Attitudes toward Latinos Influence Support for Welfare .” American Journal of Sociology 110 ( 3 ): 580 – 625 . Google Scholar CrossRef Search ADS Gilens , Martin . 1996 . “ ‘Race Coding’ and White Opposition to Welfare .” American Political Science Review 90 ( 3 ): 593 – 604 . Google Scholar CrossRef Search ADS ——— . 1999 . Why Americans Hate Welfare: Race, Media, and the Politics of Antipoverty Policy . Chicago : University of Chicago Press . Goldman , Seth , and Diana Mutz . 2014 . The Obama Effect: How the 2008 Campaign Changed White Racial Attitudes . New York : Russell Sage . Gomez , Brad , and J. Matthew Wilson . 2006 . “ Rethinking Symbolic Racism: Evidence of Attribution Bias .” Journal of Politics 68 ( 3 ): 611 – 25 . Google Scholar CrossRef Search ADS Green , Donald , Shang Ha , and John Bullock . 2010 . “ Enough Already about ‘Black Box’ Experiments: Studying Mediation Is More Difficult Than Most Scholars Suppose .” Annals of the American Academy of Political and Social Science 628 ( 1 ): 200 – 208 . Google Scholar CrossRef Search ADS Hancock , Ange-Marie . 2004 . The Politics of Disgust: The Public Identity of the Welfare Queen . New York : NYU Press . Henry , Patrick , and David Sears . 2002 . “ The Symbolic Racism 2000 Scale .” Political Psychology 23 ( 2 ): 253 – 83 . Google Scholar CrossRef Search ADS Hopkins , Daniel . 2010 . “ Politicized Places: Explaining Where and When Immigrants Provoke Local Opposition .” American Political Science Review 104 ( 1 ): 40 – 60 . Google Scholar CrossRef Search ADS Huddy , Leonie , and Stanley , Feldman . 2009 . “ On Assessing the Political Effects of Racial Prejudice .” Annual Review of Political Science 12 : 423 – 47 . Google Scholar CrossRef Search ADS Johnson , Martin . 2001 . “ The Impact of Social Diversity and Racial Attitudes on Social Welfare Policy .” State Politics & Policy Quarterly 1 ( 1 ): 27 – 49 . Google Scholar CrossRef Search ADS Jost , John , Jack Glaser , Arie Kruglanski , and Frank Sulloway . 2003 . “ Political Conservatism as Motivated Social Cognition .” Psychological Bulletin 129 ( 3 ): 339 – 75 . Google Scholar CrossRef Search ADS Kinder , Donald , and David Sears . 1981 . “ Prejudice and Politics: Symbolic Racism Versus Racial Threats to the Good Life .” Journal of Personality and Social Psychology 40 ( 3 ): 414 – 31 . Google Scholar CrossRef Search ADS Lowndes , Joseph . 2008 . From the New Deal to the New Right: Race and the Southern Origins of Modern Conservatism . New Haven, CT : Yale University Press . Luhtanen , Riia , and Jennifer Crocker . 1992 . “ A Collective Self-Esteem Scale: Self-Evaluation of One’s Social Identity .” Personality and Social Psychology Bulletin 18 ( 3 ): 302 – 18 . Google Scholar CrossRef Search ADS Luttmer , Erzo . 2001 . “ Group Loyalty and the Taste for Redistribution .” Journal of Political Economy 109 ( 3 ): 500 – 28 . Google Scholar CrossRef Search ADS Martin , Isaac William . 2008 . The Permanent Tax Revolt: How the Property Tax Transformed American Politics . Stanford, CA : Stanford University Press . Muller , Christopher . 2012 . “ Northward Migration and the Rise of Racial Disparity in American Incarceration, 1880–1950 .” American Journal of Sociology 118 ( 2 ): 281 – 326 . Google Scholar CrossRef Search ADS Neblo , Michael . 2009 . “ Three-Fifths a Racist: A Typology for Analyzing Public Opinion about Race .” Political Behavior 31 ( 1 ): 31 – 51 . Google Scholar CrossRef Search ADS Neubeck , Kenneth , and Noel Cazenave . 2001 . Welfare Racism: Playing the Race Card Against America’s Poor . New York : Routledge . O’Brien , Rourke . 2017 . “ Redistribution and the New Fiscal Sociology: Race and the Progressivity of State and Local Taxes .” American Journal of Sociology 122 ( 4 ): 1015 – 49 . Google Scholar CrossRef Search ADS Oliver , J. Eric , and Tali Mendelberg . 2000 . “ Reconsidering the Environmental Determinants of White Racial Attitudes .” American Journal of Political Science 44 ( 3 ): 574 – 89 . Google Scholar CrossRef Search ADS Outten , H. Robert , Michael Schmitt , Daniel Miller , and Amber Garcia . 2012 . “ Feeling Threatened About the Future: Whites’ Emotional Reactions to Anticipated Ethnic Demographic Changes .” Personality and Social Psychology Bulletin 38 ( 1 ): 14 – 25 . Google Scholar CrossRef Search ADS Page , Benjamin , and Robert Shapiro . 1992 . The Rational Public: Fifty Years of Trends in Americans’ Policy Preferences . Chicago : University of Chicago Press . Google Scholar CrossRef Search ADS Parker , Christopher . 2016 . “ Race and Politics in the Age of Obama .” Annual Review of Sociology 42 : 217 – 30 . Google Scholar CrossRef Search ADS Parker , Christopher , Mark Sawyer , and Christopher Towler . 2009 . “ A Black Man in the White House? The Role of Racism and Patriotism in the 2008 Presidential Election .” Du Bois Review 6 ( 1 ): 193 – 217 . Google Scholar CrossRef Search ADS Pasek , Josh , Tobias Stark , Jon Krosnick , Trevor Tompson , and B. Keith Payne . 2014 . “ Attitudes toward Blacks in the Obama Era: Changing Distributions and Impacts on Job Approval and Electoral Choice, 2008–2012 .” Public Opinion Quarterly 78 ( S1 ): 276 – 302 . Google Scholar CrossRef Search ADS Quadagno , Jill . 1994 . The Color of Welfare: How Racism Undermined the War on Poverty . New York : Oxford University Press . Quillian , Lincoln . 1995 . “ Prejudice as a Response to Perceived Group Threat: Population Composition and Anti-Immigrant and Racial Prejudice in Europe .” American Sociological Review 60 ( 4 ): 586 – 611 . Google Scholar CrossRef Search ADS Reese , Ellen . 2005 . Backlash Against Welfare Mothers: Past and Present . Berkeley : University of California Press . Roth , Byron . 1990 . “ Social Psychology’s Racism .” Public Interest 98 : 26 – 36 . Schmidt , Kathleen , and Brian Nosek . 2010 . “ Implicit (and Explicit) Racial Attitudes Barely Changed during Barack Obama’s Presidential Campaign and Early Presidency .” Journal of Experimental Social Psychology 46 ( 2 ): 308 – 14 . Google Scholar CrossRef Search ADS Sears , David , and Patrick Henry . 2005 . “ Over Thirty Years Later: A Contemporary Look at Symbolic Racism .” Advances in Experimental Social Psychology 37 : 95 – 150 . Google Scholar CrossRef Search ADS Sniderman , Paul , and Edward Carmines . 1997 . Reaching Beyond Race . Cambridge, MA : Harvard University Press . Sniderman , Paul , Gretchen Crosby , and William Howell . 2000 . “The Politics of Race.” In Racialized Politics: The Debate About Racism in America , edited by David Sears , Jim Sidanius , and Lawrence Bobo , 236–79 . Chicago : University of Chicago Press . Soss , Joe , and Sarah Bruch . 2008 . “Marginalization Matters: Rethinking Race in the Analysis of State Politics and Policy.” Paper presented at the Annual Meeting of the American Political Science Association, Boston, MA, August 27–30. Soss , Joe , Richard Fording , and Sanford Schram . 2011 . Disciplining the Poor: Neoliberal Paternalism and the Persistent Power of Race . Chicago : University of Chicago Press . Google Scholar CrossRef Search ADS Tesler , Michael , and David Sears . 2010 . Obama’s Race: The 2008 Election and the Dream of a Post-Racial America . Chicago : University of Chicago Press . Google Scholar CrossRef Search ADS US Census Bureau, Population Division . 2012 . “Table 6. Percent Distribution of the Projected Population by Race and Hispanic Origin for the United States: 2015 to 2060.” NP2012-T6. Valentino , Nicholas , and Ted Brader . 2011 . “ The Sword’s Other Edge: Perceptions of Discrimination and Racial Policy Opinion after Obama .” Public Opinion Quarterly 75 ( 2 ): 201 – 26 . Google Scholar CrossRef Search ADS Weinberg , Jill , Jeremy Freese , and David McElhattan . 2014 . “ Comparing Data Characteristics and Results of an Online Factorial Survey Between a Population-Based and Crowdsource-Recruited Sample .” Sociological Science 1 : 292 – 310 . Google Scholar CrossRef Search ADS Wright , Gerald . 1977 . “ Contextual Models of Electoral Behavior: The Southern Wallace Vote .” American Political Science Review 71 ( 2 ): 497 – 508 . Google Scholar CrossRef Search ADS Wright , Gerald , Robert Erikson , and John McIver . 1987 . “ Public Opinion and Policy Liberalism in the American States .” American Journal of Political Science 31 ( 4 ): 980 – 1001 . Google Scholar CrossRef Search ADS © The Author(s) 2018. Published by Oxford University Press on behalf of the University of North Carolina at Chapel Hill. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Social Forces Oxford University Press

Privilege on the Precipice: Perceived Racial Status Threats Lead White Americans to Oppose Welfare Programs

Social Forces , Volume Advance Article – May 31, 2018

Loading next page...
 
/lp/ou_press/privilege-on-the-precipice-perceived-racial-status-threats-lead-white-ZWXNSg0aju
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press on behalf of the University of North Carolina at Chapel Hill. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com.
ISSN
0037-7732
eISSN
1534-7605
D.O.I.
10.1093/sf/soy046
Publisher site
See Article on Publisher Site

Abstract

ABSTRACT Here, we integrate prior work to develop and test a theory of how perceived macro-level trends in racial standing shape whites’ views of welfare policy. We argue that when whites perceive threats to their relative advantage in the racial status hierarchy, their resentment of minorities increases. This increased resentment in turn leads whites to withdraw support for welfare programs when they perceive these programs to primarily benefit minorities. Analysis of American National Election Studies data and two survey-embedded experiments support this reasoning. In Study 1, we find that whites’ racial resentment increased beginning in 2008, the year of Barack Obama’s successful presidential candidacy and a major economic downturn, the latter a factor previously shown to amplify racial threat effects. At the same time, whites’ opposition to welfare increased relative to minorities’. In Study 2, we sought to better establish the causal effect of racial status threats. We found that experimentally presenting information suggesting that the white majority is rapidly declining increased whites’ opposition to welfare, and this effect was mediated by heightened racial resentment. Finally, in Study 3 we found that threatening whites’ sense of their economic advantage over minorities led whites to report greater opposition to welfare programs, but only if these programs were portrayed as primarily benefiting minorities, not if they were portrayed as benefiting whites. These findings suggest that whites’ perceptions that minorities’ standing is rising can produce periods of “welfare backlash” in which adoption of policies restricting or curtailing welfare programs is more likely. Introduction In recent decades, government programs providing low-income Americans with economic relief, particularly those programs commonly referred to as “welfare,” have been routinely attacked. From Ronald Reagan’s speeches describing the supposed excesses and luxurious lives of welfare recipients, to Bill Clinton’s vow to “end welfare as we know it,” few government programs have met with more vigorous and sustained criticism. This opposition is not restricted to political elites. The American public’s views of these programs have become increasingly polarized, and overall support for welfare has declined in recent years despite the hardship brought on by the Great Recession (Brooks and Manza 2013). What causes these programs to face such opposition? Here, we develop an integrated explanation of the role of racial prejudice in white Americans’ attitudes toward welfare programs that clarifies how larger social forces shape these sentiments. We argue that when whites perceive threats to their relative advantage in the racial status hierarchy, their resentment of minorities increases (Blumer 1958; Bobo 1999). This heightened resentment among whites in turn leads to greater opposition of welfare programs because these programs are perceived to primarily benefit racial minorities (Gilens 1999). Structural sources of perceived threat to white racial status include societal trends such as shifts in the majority status, political power, and economic advantages of dominant racial groups (Blalock 1967). Further, salient historical events seen as carrying symbolic significance for race relations, such as the election of the first black president (Parker, Sawyer, and Towler 2009), can create perceptions that whites’ privileges are precarious. Since scholarship finds that whites’ welfare attitudes are closely tied to prejudice against minorities (Fox 2004; Gilens 1996, 1999; Neubeck and Cazenave 2001), we propose that one source of these prejudicial attitudes is macro-level changes perceived to threaten whites’ status advantage. The present research thus clarifies how desires to maintain privileged racial status can influence white Americans’ views of anti-poverty policy. Below, we review prior explanations for Americans’ opposition to welfare, focusing on the extensive literature on the racial coding of welfare. We argue that social structural factors influence white Americans’ welfare attitudes and propose a model of whites’ attitudes synthesizing insights from historical, political, and social psychological literatures. We then test this explanation in a series of historical and experimental studies. First, we use nationally representative survey data to examine whether recent trends in Americans’ welfare attitudes are consistent with our claims that perceived threats to the racial status hierarchy should increase opposition to welfare among whites specifically. Next, we investigate whether experimentally induced threats to white Americans’ majority status or economic advantages over minorities increase whites’ opposition to welfare. We explore the intervening role of racial resentment and examine whether racial threats lead whites to oppose welfare programs only when they perceive them to disproportionately benefit minorities, but not when they perceive them to primarily benefit whites. Our research presents experimental evidence that whites’ welfare attitudes are affected by threats to the standing of whites and that these effects operate in part through heightened racial resentment. Further, we present evidence that this dynamic may help explain increased welfare opposition in the United States since 2008. Theoretically, these findings demonstrate that prior research linking whites’ racial resentment with welfare attitudes cannot be accounted for in terms of principled conservatism, without prejudice, as some have argued. They further suggest that perceptions of rising minority power, declines in whites’ relative economic status, and other perceived macro-level threats to whites’ racial status may produce periods of heightened anti-entitlement sentiment in which restrictions of the welfare state are more likely. Welfare Attitudes and the Contested Role of Racial Prejudice Understanding what factors influence Americans’ welfare attitudes is important because these attitudes impact policy at the state and national levels (Wright, Erikson, and McIver 1987) and because opposition to welfare programs has historically energized conservative movements (Lowndes 2008). Anti-poverty programs face a public opinion problem in “welfare backlash,” often mobilized by movements and elites (Reese 2005). The continuing degree of animosity toward welfare is all the more remarkable given the drastic cuts that have been made to these programs in the past two decades (e.g., Soss, Fording, and Schram 2011). Scholars have argued that racial politics help explain why Americans support social safety net programs less than the citizens of other industrialized, Western democracies (Quadagno 1994). Evidence suggests that welfare is strongly racially coded among Americans (Federico 2004; Gilens 1996; Neubeck and Cazenave 2001). Specifically among whites, opposition to welfare is associated with stereotypes of African Americans as lazy, unreliable, and lacking sexual morality (Gilens 1996; Hancock 2004). Because Americans tend to perceive welfare recipients as predominantly black (CBS 1996; Federico 2004), such stereotypes lead whites to view welfare recipients as undeserving of assistance (Gilens 1999). However, the degree to which opposition to welfare and other racialized policies reflects racial prejudice rather than conservative principles of small government and self-reliance continues to be debated (Huddy and Feldman 2009; Valentino and Brader 2011). Many scholars have claimed that the influence of racial prejudice on whites’ policy attitudes has been overstated (Gomez and Wilson 2006; Neblo 2009; Roth 1990; Sniderman and Carmines 1997; Sniderman, Crosby, and Howell 2000). For example, Sniderman and Carmines (1997) argue that racial prejudice is less important than moral arguments and policy specifics in determining whites’ support for welfare programs. Multivariate analyses correlating racial resentment or symbolic racism to policy attitudes have additionally been criticized on methodological grounds, with scholars arguing that these measures reflect conservative and individualist values rather than racial prejudice (Carney and Enos 2015; Huddy and Feldman 2009; Sniderman, Crosby, and Howell 2000). Because of the inherent uncertainty involved in drawing causal inferences from correlational data, much of the literature claiming the importance of racial resentment in whites’ opposition to welfare remains open to this critique (Sniderman and Carmines 1997; Sniderman, Crosby, and Howell 2000).1 Group Position Theory and Threats to Racial Hierarchy If racial prejudice does play a primary role in white Americans’ opposition to welfare programs, as a large though contested body of research suggests (e.g., Fox 2004; Gilens 1996, 1999; Neubeck and Cazenave 2001), what kinds of social processes and macro-level forces might lead whites to oppose these programs? Here, we emphasize the role that threats to whites’ privileged status play in sparking racial resentment. We argue that trends and events perceived as threatening the relative position of whites in the racial status hierarchy may prompt these feelings of threat, leading whites to increased opposition to welfare programs as a result of heightened racial resentment. Group position theory (Blumer 1958; Bobo 1999) proposes that larger social processes activate racial antipathy among members of higher-ranking racial groups. According to the theory, understandings of where racial groups stand in terms of social status and material rewards are historical products, formed mostly through discursive ideological activities of elites in which characterizations of different racial groups and the proper relationship between them are offered, debated, and refined. While a relatively stable sense of hierarchy need not always result, certain racial groups in particular social contexts—such as whites in the United States—often come to hold a “sense of group position,” or an expectation that their racial/ethnic group has a higher-status position than other groups and that this position is associated with greater access to economic, political, and social resources. A central claim of the theory is that prejudice primarily results from dominant group members’ desires to maintain this status advantage over other groups (Bobo 1999). Dominant group members react with racial prejudice when they sense that their group’s position—and its claim to the material or status advantages that should accrue to it—are under threat (Blumer 1958; Bobo 1999). Supporting this assertion, scholars have documented the importance of perceptions of competition and threat between racial groups in explaining prejudice (Craig and Richeson 2014a; Quillian 1995), in-group favoritism (Abascal 2015), and attitudes on explicitly racial policies like racial segregation (Bobo and Zubrinsky 1996) and affirmative action (Bobo 1998). Group position theory suggests that anti-welfare sentiment among whites should rise when there is a “felt challenge” (Blumer 1958) to whites’ group position and a corresponding increase in racial prejudice, and likewise abate when the relative advantage of whites is perceived as stable. The question, then, becomes one of identifying when the events of a particular historical period produce this sense of challenge. Prior scholarship suggests that reductions in political power, economic advantage, or population share among dominant group members can produce a sense of racial status threat (Blalock 1967), as may high-profile cultural or political events that indicate minority groups’ power or status is rising. In the contemporary American context, examples of such trends and events include the rising population share and electoral strength of racial minorities, and the election of the first nonwhite president (Parker, Sawyer, and Towler 2009). Finally, perceived threats to racial status are more likely to lead to resentment of racial minorities under conditions of macro-economic decline, as dominant group members perceive greater competition over resources and fear that minorities will overtake their group’s privileged access to these resources (Quillian 1995). In arguing that perceived threats to whites’ status prompt racial resentment, we treat group position theory and racial resentment as complementary accounts of white Americans’ racial attitudes, though these two conceptions of prejudice have long been seen as competing (e.g., Kinder and Sears 1981). Theorists of racial resentment characterize prejudice as resulting from early socialization experiences, such that prejudice is expected to be largely stable within a given individual after childhood. In contrast, group position theory characterizes prejudice as resulting from conflict between racial groups for material or symbolic resources, such that prejudice is seen as variable over time, triggered by perceptions of threat to dominant group status. Here, we argue that these two conceptions are compatible. We expect that racial resentment has both state- and trait-like properties, such that levels of resentment are shaped by socialization processes, have cross-temporal consistency, but can also rise and fall in response to novel information or experiences. In particular, following Blumer (1958) and Bobo (1999), we argue that social trends or events perceived to threaten the dominant group’s status will typically lead to increased resentment. Group Position and Welfare Attitudes As our review indicates, perceived threats to racial status have been shown to shape attitudes toward explicitly racial policies. Here, we propose that threats may also help explain whites’ attitudes toward policies, like welfare, that are formally race-blind, but perceived in racialized terms. However, prior work examining the relationship between racial threat and welfare attitudes shows mixed results. Using the common procedure of operationalizing threat as the proportion of minorities in the local context, some studies find that an increased proportion of blacks leads non-blacks to have more negative attitudes toward welfare (Johnson 2001; Luttmer 2001; Wright 1977), but other research does not find such effects (Oliver and Mendelberg 2000) or finds that the proportion of blacks only matters at some levels of analysis (Fox 2004). Further, this work has been largely observational, leaving open the possibility that unmeasured processes drove these effects. On a theoretical level, this prior work differs from our own in that it focuses on personal experiences of competition for resources. Here, however, we study the effects of symbolic threats to white Americans’ status in the larger society. These perceptions need not be a direct function of personal experience and instead can be socially constructed in the public sphere as elites discuss historical trends and events (Hopkins 2010). Further, others have noted that the emphasis on the proportion of black Americans in the local context has led scholars to leave unanalyzed historical and structural conditions that might lead to perceptions of racial threat (Brown 2013; Muller 2012; Soss and Bruch 2008). Finally, while we propose that threats to white racial status lead to welfare opposition as a result of heightened racial resentment, others have argued that racial threats lead whites to oppose welfare spending as a result of heightened in-group loyalty (e.g., Luttmer 2001). According to this argument, whites’ sense of racial group solidarity rises under conditions of threat, leading them to favor spending on programs benefiting whites rather than those seen as benefiting minorities. Similarly, O’Brien (2017) suggests that whites in diverse local contexts oppose progressive taxation and social spending not due to aversive feelings toward minorities, but because they are engaging in “opportunity hoarding,” or attempting to reserve scarce resources for members of their own racial groups. Here, we attempt to clarify the relationship between threat to racial status and white opposition to welfare by examining how macro-level trends and events can create perceptions of threat, and by investigating whether racial resentment or feelings of racial group identification drive the relationship between threat and welfare opposition. Figure 1 diagrams our explanation of the link between threats to racial status and welfare attitudes. We predict that social trends and events with symbolic importance for collective understandings of the relative status of racial groups—such as the election of the first black president and the rising minority population—threaten the perceived standing of whites, increasing their resentment of minorities (path a) and, as a result, their opposition to welfare programs (path b). In addition, the strength of these causal links is shaped by other factors. The effects of status threats on whites’ racial resentment should be particularly strong when they occur in conjunction with macro-economic downturns (path c), which can catalyze dominant group members’ sense of racial threat through heightened feelings of competition with subordinated groups for access to scarce resources. Finally, we expect that racial resentment prompted by racial threats will increase whites’ welfare opposition to the extent that individuals hold stereotypic views of welfare as largely benefiting minorities (path d). Where those stereotypic views are stronger, this link will be magnified, and where they are weaker or non-existent, the link will be reduced or mitigated entirely. Figure 1. View largeDownload slide Hypothesized effect of racial status threat in producing welfare opposition among whites. All arrows represent positive relationships Figure 1. View largeDownload slide Hypothesized effect of racial status threat in producing welfare opposition among whites. All arrows represent positive relationships Empirical Overview Based on the above, we argue that white Americans’ welfare attitudes are shaped by perceptions that significant macro-level trends and events threaten their racial status advantage. Further, we propose that this framework may help explain periods of “welfare backlash” that can impact the course of welfare state development. To analyze these claims, we first use American National Election Studies (ANES) data to examine whether our model fits the characteristics of Americans’ recent increased opposition to welfare. We expect that whites’ welfare opposition increased, relative to minorities’, beginning in 2008 because this period was charged with events likely to produce a “felt challenge” (Blumer 1958) to whites’ group position: the candidacy and election of President Obama and the beginning of the Great Recession. Next, we investigate experimentally whether perceiving that whites’ share of the US population is declining decreases support for welfare among white participants, further testing whether this effect is driven by heightened racial resentment. Finally, in Study 3 we compare the effects of threat on attitudes toward a welfare program benefiting minorities versus one that benefits whites, examining whether perceiving that whites’ economic advantages are declining leads whites to oppose welfare programs only if they believe the program primarily benefits minorities. In Study 3, we further assess whether the effects we find are due to whites’ resentment of minorities or due to heightened in-group loyalty. Our empirical strategy is designed to feature studies with a mix of internal and external validity to balance the value of demonstrating causality and real-world relevance of our findings. Study 1 relies on nationally representative survey data to offer an initial test of the historical relevance of our theory. Observational data, however, presents well-known barriers to causal inference. Studies 2 and 3 rely on non-representative samples, but—as experiments—allow us greater confidence in establishing causality. In addition, comparing responses to experimental stimuli is an unobtrusive method of examining how perceptions of decline in whites’ racial status might affect participants’ welfare views and racial resentment. Thus, these experiments are useful for examining an issue where direct self-reports of individuals’ motivations can be misleading, either because individuals lack insight on the social forces shaping their attitudes, or because social desirability concerns make them unwilling to report them reliably. Study 1: Racial Divergence in Americans’ Welfare Attitudes since 2008 We begin by analyzing ANES data to examine whether our theory is consistent with recent trends in welfare and racial attitudes in the general population. Based on our theoretical reasoning, we argue that the period beginning in 2008 has likely seen increased opposition to welfare among whites relative to minorities as a result of political and economic events perceived as threats to whites’ position in the racial status hierarchy. To be clear, we do not propose that our theory can account for all sources of historical variation in whites’ welfare attitudes, allowing us to predict absolute values of these attitudes over time. Instead, our predictions concern the racial gap in welfare attitudes as an indication of perceived racial threat felt among whites uniquely, net of other social factors. What trends or events occurring in the period beginning in 2008 might be perceived as threatening whites’ racial status? While whites continue to enjoy higher incomes, wealth, and representation in government than African Americans and Latinos, much public discourse about race in this period emphasized America’s increasing demographic diversity and the declining dominance of white Americans. Most prominently, the presidential candidacy and election of Barack Obama sparked pronouncements that the United States was entering a “post-racial” era (Tesler and Sears 2010). Further, media commentators discussed Obama’s election and re-election as resulting from the growing electoral power of racial minorities (e.g., Dougherty 2009), heightening the salience of the declining white majority. Finally, discussions of increases in demographic diversity and minorities’ political influence were coupled with experiences of economic hardship during the Great Recession. While macro-economic decline is likely to increase support for welfare among all Americans (e.g., Erikson, MacKuen, and Stimson 2002), research suggests that downturns are also likely to magnify racial threat effects (Quillian 1995). Therefore, we argue that this combination of factors increased whites’ racial resentment. While we expect that whites’ racial resentment should have risen during this time, previous research investigating changes in whites’ attitudes in the wake of Obama’s election has been mixed (Parker 2016). Some studies initially uncovered what was dubbed “The Obama Effect,” where whites showed decreased implicit bias and stereotyping of black Americans after Obama’s election (e.g., Bernstein, Young, and Claypool 2010), presumably due to exposure to Obama as a salient counter-stereotypical exemplar (Goldman and Mutz 2014). Other studies, however, have found that implicit and explicit racial attitudes remained stable (Schmidt and Nosek 2010), and yet others have found increased racial prejudice (Pasek et al. 2014). The literature therefore is divided about whether we would expect to see a rise in racial resentment beginning in 2008. Our theoretical reasoning, however, would suggest that perceptions of minorities’ increased political power during a period of macro-economic decline should lead to a sense of threat to whites’ group status. Therefore, we predict that whites’ racial resentment increased beginning in 2008, and that the racial gap between whites’ and minorities’ welfare attitudes has grown larger beginning in 2008, with whites increasing opposition to welfare relative to minorities. Method We analyze ANES data from 2000 to 2012 to investigate these claims.2 Table 1 describes the variables used in our analysis. Our primary independent variables are respondent’s race and dummy variables indicating the year the respondent was surveyed, with 2008 set as the baseline year. Table 1. Variables used in analysis of American National Election Studies Variable Item wording/Coding Welfare attitudes Should federal spending on welfare programs be increased, decreased, or kept about the same? 1 = increased; 2 = kept about the same; 3 = decreased Racial resentment Composite of four items asked on 1–5 scale where 1 = strongly agree and 5 = strongly disagree: “It’s really a matter of some people not trying hard enough; if blacks would only try harder they could be just as well off as whites.” (reverse-coded) “Generations of slavery and discrimination have created conditions that make it difficult for blacks to work their way out of the lower class.” “Over the past few years, blacks have gotten less than they deserve.” “Irish, Italians, Jewish, and many other minorities overcame prejudice and worked their way up. Blacks should do the same without any special favors.” (reverse-coded) Respondent’s race/ethnicity Dummy variables categorizing race/ethnicity as non-Hispanic white; black; or Hispanic, with respondents identifying as another race or ethnicity as baseline Survey year Dummy variables identifying survey year as 2000, 2004, or 2012, with the 2008 survey year as baseline Gender 0 = female; 1 = male Age Years Education Dummy variables categorizing education into five different categories: less than high school credential; high school credential; some college; bachelor’s degree; and advanced degree Employment status 1 = unemployed; 0 = else. Marital status 1 = married; 0 = else Class identification 1 = middle or upper class; 0 = else Income quintile Total household income coded in percentiles. 1 = 0–16 percentile; 2 = 17–33 percentile; 3 = 34–67 percentile; 4 = 68–95 percentile; 5 = 96–100 percentile Income trajectory Would you say that you are better off or worse off than you were a year ago? 1 = worse off; 0 = else Political ideology 1 = extremely liberal; 2 = liberal; 3 = slightly liberal; 4 = moderate/middle of the road; 5 = slightly conservative; 6 = conservative; 7 = extremely conservative Variable Item wording/Coding Welfare attitudes Should federal spending on welfare programs be increased, decreased, or kept about the same? 1 = increased; 2 = kept about the same; 3 = decreased Racial resentment Composite of four items asked on 1–5 scale where 1 = strongly agree and 5 = strongly disagree: “It’s really a matter of some people not trying hard enough; if blacks would only try harder they could be just as well off as whites.” (reverse-coded) “Generations of slavery and discrimination have created conditions that make it difficult for blacks to work their way out of the lower class.” “Over the past few years, blacks have gotten less than they deserve.” “Irish, Italians, Jewish, and many other minorities overcame prejudice and worked their way up. Blacks should do the same without any special favors.” (reverse-coded) Respondent’s race/ethnicity Dummy variables categorizing race/ethnicity as non-Hispanic white; black; or Hispanic, with respondents identifying as another race or ethnicity as baseline Survey year Dummy variables identifying survey year as 2000, 2004, or 2012, with the 2008 survey year as baseline Gender 0 = female; 1 = male Age Years Education Dummy variables categorizing education into five different categories: less than high school credential; high school credential; some college; bachelor’s degree; and advanced degree Employment status 1 = unemployed; 0 = else. Marital status 1 = married; 0 = else Class identification 1 = middle or upper class; 0 = else Income quintile Total household income coded in percentiles. 1 = 0–16 percentile; 2 = 17–33 percentile; 3 = 34–67 percentile; 4 = 68–95 percentile; 5 = 96–100 percentile Income trajectory Would you say that you are better off or worse off than you were a year ago? 1 = worse off; 0 = else Political ideology 1 = extremely liberal; 2 = liberal; 3 = slightly liberal; 4 = moderate/middle of the road; 5 = slightly conservative; 6 = conservative; 7 = extremely conservative Table 1. Variables used in analysis of American National Election Studies Variable Item wording/Coding Welfare attitudes Should federal spending on welfare programs be increased, decreased, or kept about the same? 1 = increased; 2 = kept about the same; 3 = decreased Racial resentment Composite of four items asked on 1–5 scale where 1 = strongly agree and 5 = strongly disagree: “It’s really a matter of some people not trying hard enough; if blacks would only try harder they could be just as well off as whites.” (reverse-coded) “Generations of slavery and discrimination have created conditions that make it difficult for blacks to work their way out of the lower class.” “Over the past few years, blacks have gotten less than they deserve.” “Irish, Italians, Jewish, and many other minorities overcame prejudice and worked their way up. Blacks should do the same without any special favors.” (reverse-coded) Respondent’s race/ethnicity Dummy variables categorizing race/ethnicity as non-Hispanic white; black; or Hispanic, with respondents identifying as another race or ethnicity as baseline Survey year Dummy variables identifying survey year as 2000, 2004, or 2012, with the 2008 survey year as baseline Gender 0 = female; 1 = male Age Years Education Dummy variables categorizing education into five different categories: less than high school credential; high school credential; some college; bachelor’s degree; and advanced degree Employment status 1 = unemployed; 0 = else. Marital status 1 = married; 0 = else Class identification 1 = middle or upper class; 0 = else Income quintile Total household income coded in percentiles. 1 = 0–16 percentile; 2 = 17–33 percentile; 3 = 34–67 percentile; 4 = 68–95 percentile; 5 = 96–100 percentile Income trajectory Would you say that you are better off or worse off than you were a year ago? 1 = worse off; 0 = else Political ideology 1 = extremely liberal; 2 = liberal; 3 = slightly liberal; 4 = moderate/middle of the road; 5 = slightly conservative; 6 = conservative; 7 = extremely conservative Variable Item wording/Coding Welfare attitudes Should federal spending on welfare programs be increased, decreased, or kept about the same? 1 = increased; 2 = kept about the same; 3 = decreased Racial resentment Composite of four items asked on 1–5 scale where 1 = strongly agree and 5 = strongly disagree: “It’s really a matter of some people not trying hard enough; if blacks would only try harder they could be just as well off as whites.” (reverse-coded) “Generations of slavery and discrimination have created conditions that make it difficult for blacks to work their way out of the lower class.” “Over the past few years, blacks have gotten less than they deserve.” “Irish, Italians, Jewish, and many other minorities overcame prejudice and worked their way up. Blacks should do the same without any special favors.” (reverse-coded) Respondent’s race/ethnicity Dummy variables categorizing race/ethnicity as non-Hispanic white; black; or Hispanic, with respondents identifying as another race or ethnicity as baseline Survey year Dummy variables identifying survey year as 2000, 2004, or 2012, with the 2008 survey year as baseline Gender 0 = female; 1 = male Age Years Education Dummy variables categorizing education into five different categories: less than high school credential; high school credential; some college; bachelor’s degree; and advanced degree Employment status 1 = unemployed; 0 = else. Marital status 1 = married; 0 = else Class identification 1 = middle or upper class; 0 = else Income quintile Total household income coded in percentiles. 1 = 0–16 percentile; 2 = 17–33 percentile; 3 = 34–67 percentile; 4 = 68–95 percentile; 5 = 96–100 percentile Income trajectory Would you say that you are better off or worse off than you were a year ago? 1 = worse off; 0 = else Political ideology 1 = extremely liberal; 2 = liberal; 3 = slightly liberal; 4 = moderate/middle of the road; 5 = slightly conservative; 6 = conservative; 7 = extremely conservative Our first dependent variable is a composite scale of four racial resentment items (Cronbach’s α = 0.75) (e.g., “It’s really a matter of some people not trying hard enough; if blacks would only try harder they could be just as well off as whites”). Our second dependent variable is a single item gauging welfare attitudes, measured on a three-point scale, asking whether federal spending on welfare programs should be increased, kept about the same, or decreased. We include controls for demographic and socioeconomic characteristics and political ideology. All analyses use recommended weighting to account for complex sampling features (DeBell 2010). Results Descriptive statistics suggest that whites’ racial resentment rose beginning in 2008 and continued rising in 2012 (figure 2). We perform an ordinary least squares regression to examine whether the rise in whites’ racial resentment is statistically reliable, controlling for socioeconomic characteristics and political ideology (Supplementary Table 1). This model shows that whites’ racial resentment rose between 2004 and 2008 (β = –0.103, p = 0.034) as well as between 2008 and 2012 (β = 0.090, p = 0.014). This pattern is consistent with our reasoning that 2008 marked the beginning of a period of increased racial status threat among white Americans that prompted greater resentment of minorities. Figure 2. View largeDownload slide Average racial resentment from 2000 to 2012 among white respondents Figure 2. View largeDownload slide Average racial resentment from 2000 to 2012 among white respondents Next, descriptive statistics suggest that relative levels of welfare opposition among whites and minorities diverged beginning in 2008 (Supplementary Figure 1). We perform a series of ordered logistic regressions to examine whether this racial divergence is statistically reliable, controlling for other factors. We first estimate a model predicting welfare attitudes by survey year, respondent’s race, and control variables (model 1, table 2). This model shows that among all Americans, regardless of race, opposition to welfare remained stable between 2004 and 2008, and then rose in 2012 (β = 0.931, p < 0.001). To examine how these trends may have varied between whites and minorities, we estimate the model again, now including a term for the interactions between white race and survey year dummy variables (model 2, table 2). Table 2. Ordered logistic regression models predicting Americans’ welfare opposition: coefficients and linearized standard errors Model 1: Model 2: Without White × Survey wave interactions With White × Survey wave interactions White 0.009 0.189 (0.137) (0.178) Black −0.821*** −0.786*** (0.156) (0.158) Hispanic −0.477** −0.432** (0.158) (0.161) Male 0.211*** 0.212*** (0.055) (0.056) Age −0.004* −0.004* (0.002) (0.002) Less than high school credential −0.363** −0.364** (0.129) (0.129) High school credential −0.060 −0.062 (0.084) (0.084) Some college 0.174* 0.169* (0.076) (0.076) Advanced degree −0.366*** −0.365*** (0.089) (0.089) Unemployed −0.518*** −0.519*** (0.120) (0.120) Married 0.070 0.070 (0.063) (0.063) Middle- or upper-class identification 0.043 0.043 (0.063) (0.063) Perceived downward income trajectory 0.063 0.064 (0.042) (0.042) Income quintile 0.243*** 0.243*** (0.027) (0.027) Conservative 0.427*** 0.427*** (0.021) (0.021) 2000 survey wave 0.487*** 1.087*** (0.106) (0.226) 2004 survey wave 0.008 0.317† (0.094) (0.186) 2012 survey wave 0.931*** 0.992*** (0.099) (0.139) White × 2000 −0.780** (0.253) White × 2004 −0.438* (0.215) White × 2012 −0.105 (0.141) N 7,696 7,696 Model 1: Model 2: Without White × Survey wave interactions With White × Survey wave interactions White 0.009 0.189 (0.137) (0.178) Black −0.821*** −0.786*** (0.156) (0.158) Hispanic −0.477** −0.432** (0.158) (0.161) Male 0.211*** 0.212*** (0.055) (0.056) Age −0.004* −0.004* (0.002) (0.002) Less than high school credential −0.363** −0.364** (0.129) (0.129) High school credential −0.060 −0.062 (0.084) (0.084) Some college 0.174* 0.169* (0.076) (0.076) Advanced degree −0.366*** −0.365*** (0.089) (0.089) Unemployed −0.518*** −0.519*** (0.120) (0.120) Married 0.070 0.070 (0.063) (0.063) Middle- or upper-class identification 0.043 0.043 (0.063) (0.063) Perceived downward income trajectory 0.063 0.064 (0.042) (0.042) Income quintile 0.243*** 0.243*** (0.027) (0.027) Conservative 0.427*** 0.427*** (0.021) (0.021) 2000 survey wave 0.487*** 1.087*** (0.106) (0.226) 2004 survey wave 0.008 0.317† (0.094) (0.186) 2012 survey wave 0.931*** 0.992*** (0.099) (0.139) White × 2000 −0.780** (0.253) White × 2004 −0.438* (0.215) White × 2012 −0.105 (0.141) N 7,696 7,696 † p < 0.10 * p < 0.05 ** p < 0.01 *** p < 0.001 Table 2. Ordered logistic regression models predicting Americans’ welfare opposition: coefficients and linearized standard errors Model 1: Model 2: Without White × Survey wave interactions With White × Survey wave interactions White 0.009 0.189 (0.137) (0.178) Black −0.821*** −0.786*** (0.156) (0.158) Hispanic −0.477** −0.432** (0.158) (0.161) Male 0.211*** 0.212*** (0.055) (0.056) Age −0.004* −0.004* (0.002) (0.002) Less than high school credential −0.363** −0.364** (0.129) (0.129) High school credential −0.060 −0.062 (0.084) (0.084) Some college 0.174* 0.169* (0.076) (0.076) Advanced degree −0.366*** −0.365*** (0.089) (0.089) Unemployed −0.518*** −0.519*** (0.120) (0.120) Married 0.070 0.070 (0.063) (0.063) Middle- or upper-class identification 0.043 0.043 (0.063) (0.063) Perceived downward income trajectory 0.063 0.064 (0.042) (0.042) Income quintile 0.243*** 0.243*** (0.027) (0.027) Conservative 0.427*** 0.427*** (0.021) (0.021) 2000 survey wave 0.487*** 1.087*** (0.106) (0.226) 2004 survey wave 0.008 0.317† (0.094) (0.186) 2012 survey wave 0.931*** 0.992*** (0.099) (0.139) White × 2000 −0.780** (0.253) White × 2004 −0.438* (0.215) White × 2012 −0.105 (0.141) N 7,696 7,696 Model 1: Model 2: Without White × Survey wave interactions With White × Survey wave interactions White 0.009 0.189 (0.137) (0.178) Black −0.821*** −0.786*** (0.156) (0.158) Hispanic −0.477** −0.432** (0.158) (0.161) Male 0.211*** 0.212*** (0.055) (0.056) Age −0.004* −0.004* (0.002) (0.002) Less than high school credential −0.363** −0.364** (0.129) (0.129) High school credential −0.060 −0.062 (0.084) (0.084) Some college 0.174* 0.169* (0.076) (0.076) Advanced degree −0.366*** −0.365*** (0.089) (0.089) Unemployed −0.518*** −0.519*** (0.120) (0.120) Married 0.070 0.070 (0.063) (0.063) Middle- or upper-class identification 0.043 0.043 (0.063) (0.063) Perceived downward income trajectory 0.063 0.064 (0.042) (0.042) Income quintile 0.243*** 0.243*** (0.027) (0.027) Conservative 0.427*** 0.427*** (0.021) (0.021) 2000 survey wave 0.487*** 1.087*** (0.106) (0.226) 2004 survey wave 0.008 0.317† (0.094) (0.186) 2012 survey wave 0.931*** 0.992*** (0.099) (0.139) White × 2000 −0.780** (0.253) White × 2004 −0.438* (0.215) White × 2012 −0.105 (0.141) N 7,696 7,696 † p < 0.10 * p < 0.05 ** p < 0.01 *** p < 0.001 The interactive model shows that the overall appearance of stability between 2004 and 2008 masks distinct trends between whites and members of other racial groups. Whites’ and minorities’ attitudes diverged beginning in 2008, as evidenced by the significant interactions between the white race variable and the 2000 and 2004 dummy variables (β = –0.780, p = 0.002 and β = –0.438, p = 0.041, respectively). In models analyzing whites’ and minorities’ responses separately, minorities showed a significant decrease in welfare opposition between 2004 and 2008 at the onset of a severe economic downturn (β = 0.379, p = 0.037; Supplementary Table 2), while whites did not. Instead, whites’ opposition to welfare increased somewhat (though not significantly) between 2004 and 2008, creating an increased racial gap in welfare attitudes beginning in 2008. Returning to the interactive model, the insignificant interaction between the white race variable and the 2012 dummy variable indicates that the racial divergence in attitudes that began in 2008 persisted in 2012, when Americans overall expressed heightened opposition to welfare. This pattern is consistent with our reasoning that perceptions of threat to whites’ racial status should lead to an increased racial gap between whites’ and minorities’ welfare attitudes, with whites increasing opposition to welfare relative to minorities. Discussion Using representative survey data, we find that the period beginning in 2008 has seen changes in Americans’ attitudes toward welfare, and these changes differ by race. While minorities show more positive attitudes toward welfare in 2008, whites’ attitudes became somewhat more negative. When all Americans increased welfare opposition in 2012, this racial gap remained statistically unchanged. The result is that 2008 marks the beginning of a widened divergence among whites and minorities’ attitudes toward welfare. In addition, we find that whites’ racial resentment rose beginning in 2008. These findings are consistent with our claim that feelings of racial threat—particularly, the perception of increased political power among minorities during a period of economic recession—helped shape whites’ welfare attitudes in recent years. Why do we observe that minorities increase support for welfare in 2008? Theoretical traditions in political economy (e.g., macro polity theory) argue that publics typically respond to economic crises with increased support for government action to reduce economic insecurity (e.g., Erikson, MacKuen, and Stimson 2002; Page and Shapiro 1992). Here, we find that minorities showed this frequently observed increase in welfare support in 2008, at the onset of a major economic downturn, but whites did not. This pattern can be understood as an across-the-board increase in welfare support in response to economic crisis being offset among whites by their response to racial status threat. Additional analyses support this interpretation, suggesting that economic anxiety experienced by whites suppressed an increase in opposition to welfare between 2004 and 2008 (Supplementary Table 3). Before controlling for economic anxiety, whites’ welfare opposition remains stable from 2004 to 2008 (model 1). However, whites reported significantly greater economic anxiety from 2004 to 2008 (model 2), and economic anxiety is associated with decreased opposition to welfare among whites (model 3). Once we control for economic anxiety, we find that whites’ opposition to welfare increased between 2004 and 2008, though this increase is only marginally significant (model 4). These results are consistent with our argument that, absent the economic hardship brought on by the Great Recession, racial status threat would have acted to increase whites’ welfare opposition from 2004 to 2008. In addition, we find that this (suppressed) increased welfare opposition is partially mediated by racial resentment (Supplementary Table 4). However, analysis of cross-sectional survey data is limited in its capacity to test the causal claims of our model, leaving other plausible interpretations of our results. Other factors besides threat to white racial status occurring in American society in this period may have driven the results we observe, including the rise of anti-tax movements (Martin 2008), the proliferation of pro-austerity economic philosophies, and the passage of the American Recovery and Reinvestment Act of 2009.3 In addition, it is possible that the racial divergence in welfare attitudes is due to some factor that specifically increased minorities’ support for welfare. For example, welfare support may have risen among minorities because of increased confidence that welfare benefits would be more fairly allocated under a coming Obama administration. While the findings we present are consistent with our theoretical predictions, barriers to causal inference posed by cross-sectional survey data mean we cannot prove that our explanation of these patterns is correct or historically accurate. Thus, we turn to experiments to craft controlled tests of our theoretical claim that racial status threats increase opposition to welfare among whites by increasing racial resentment. Study 2: Threat to Whites’ Majority Status In Study 2, we experimentally test whether the perception that whites’ status as the majority racial/ethnic group in the United States is precarious leads to welfare opposition among whites. While majority status is not inevitably correlated with social standing, racial/ethnic groups with greater population share typically have greater political power in democracies and hold greater influence over dominant cultural values (Blalock 1967). Previous research has also found that whites perceive information suggesting their majority status is on the decline as a threat to group status (Craig and Richeson 2014a; Outten et al. 2012) and react to this information with increased favoritism toward other whites (Abascal 2015). Therefore, if our reasoning is correct, we would expect white participants who are led to believe whites are losing majority status to show increased opposition to welfare compared to white participants led to believe that whites continue to have a strong numerical advantage. Further, we test whether racial status threat is producing opposition to welfare as a result of increased racial resentment. In our first experimental test, we utilize a historical trend occurring at the time we conducted our research to highlight the consequences of the declining white majority for whites’ political power. Following the re-election of President Obama, media commentators emphasized that the declining proportion of white Americans meant the white voting bloc could no longer determine elections. Thus, we conducted a survey-based experiment in which participants were presented either with information emphasizing whites’ declining population share, or with information emphasizing the persistence of the white majority, and introductory text connected this information to ongoing discussions of a changing electorate. Our hypotheses concern white participants’ reactions to a threat to majority status. However, we also include minority participants in the research design as a comparison group. If perceptions that whites’ population share is declining lead to qualitatively different results among whites and minorities, this supports our claim that this information is functioning as a threat to whites’ racial group position, not a threat that impacts all Americans similarly. Method A total of 151 participants (74 women and 77 men) were recruited via Amazon’s Mechanical Turk (MTurk). MTurk is an online marketplace where more than 100,000 individuals complete short jobs posted by “requesters.” While a self-selected group, the MTurk participant pool is more demographically diverse than the college student samples often employed in experimental research (Buhrmester, Kwang, and Gosling 2011), and results of experiments conducted on MTurk are similar to results obtained through an online platform using population-based sampling (Weinberg, Freese, and McElhattan 2014). Participants were US residents between the ages of 18 and 75 (M = 33.4 years, SD = 11.5). Most participants identified as white (75 percent), while the other 25 percent were split about evenly between those identifying as Asian, African American, and Latino. Respondents’ median household income was between $40,000 and $49,999, and 55.0 percent had at least a college degree. Participants answered a series of demographic questions, and then were asked to answer questions about one of two charts describing trends in the population share of different racial/ethnic groups in the United States, ostensibly as part of a study of quantitative reasoning and social opinions. Both graphs presented information adapted from Census Bureau population projections (US Census Bureau 2012). Introductory text noted that the last presidential election had sparked discussion of changes in the demographic make-up of the country, and so “it is important to know the results of cutting-edge research.” Participants were randomly assigned to one of two conditions meant to highlight different aspects of these trends. Participants in the “Majority Salient” condition were shown population data from 2000 to 2020, when white Americans’ population share is expected to remain about 60 percent of the population (figure 3a). In contrast, participants in the “Decline Salient” condition were shown projections from 1960 to 2060, when whites’ population share is expected to fall to about 40 percent. This longer time scale was intended to communicate a sharp, stable decline in white population share. We also included a line for the share of “total non-white” Americans intended to underscore the growth in the minority population (figure 3b). The graphs were accompanied by captions describing the trends in words (e.g., “The majority of Americans will be nonwhite in about 25 years”), reinforcing the experimental manipulation. Participants were then quizzed regarding the information; one of these questions assessed whether the experimental manipulation was effective. Figure 3. View largeDownload slide a: Chart presented in majority salient condition. b. Chart presented in decline salient condition Sources: US Census Bureau, 2011, and University of Virginia Population Study Center, 2012. Figure 3. View largeDownload slide a: Chart presented in majority salient condition. b. Chart presented in decline salient condition Sources: US Census Bureau, 2011, and University of Virginia Population Study Center, 2012. Next, participants’ welfare attitudes were measured in two ways. First, participants were told to imagine they were on a Congressional committee charged with cutting $500 million from the federal budget. They were given a list of nine spending areas including “Temporary Assistance for Needy Families (Welfare)” and asked to indicate how much they would cut from each area. Second, participants indicated their agreement with two statements adapted from the General Social Survey: “We are spending too much money on welfare” (reverse-coded) and “Public assistance is necessary to ensure fairness in our society.” These items were averaged to create a composite scale of welfare attitudes (Cronbach’s α = 0.77), centered at zero, ranging from −3 to 3. Next, racial resentment was measured using a standard scale including the above items from the ANES and four additional items (Henry and Sears 2002). These items were averaged to form a composite (Cronbach’s α = 0.87), centered at zero, ranging from −3 to 3. Finally, participants were asked other measures of political beliefs to examine alternative explanations, thanked, and debriefed. Results Four participants were excluded from analysis because they answered the manipulation check item incorrectly, leaving a total of 147 participants for analysis.4 We first tested the effect of the population trend information on participants’ welfare support and racial resentment (table 3). White participants assigned to the Decline Salient condition reported significantly greater opposition to welfare than those assigned to the Majority Salient condition. While white participants who were told that whites continue to be the largest single ethnic group in the United States proposed cutting $28 million from federal welfare spending, those told that whites’ population share is substantially declining proposed cutting $51 million. In addition, whites in the Decline Salient condition reported significantly greater opposition to welfare and higher levels of racial resentment on survey measures. Table 3. Effects of information highlighting whites’ majority population share or decline of white population share on white and minority participants’ welfare attitudes Majority Salient Decline Salient t Mean (SD) Mean (SD) White participants Money cut from welfare (in millions of dollars) 28.42 (51.41) 50.78 (59.36) 2.13* Composite Welfare Attitudes scale (points on 6-point scale) 0.55 (1.40) −0.19 (1.70) −2.50* Symbolic Racism (points on 6-point scale) −0.47 (1.03) 0.07 (1.30) 2.40* Minority participants Money cut from welfare (in millions of dollars) 52.81 (77.44) 57.25 (116.41) 0.13 Composite Welfare Attitudes scale (points on 6-point scale) 0.34 (1.56) 0.63 (1.60) 0.53 Symbolic Racism scale (points on 6-point scale) −0.09 (1.27) −0.78 (1.29) −1.56 Majority Salient Decline Salient t Mean (SD) Mean (SD) White participants Money cut from welfare (in millions of dollars) 28.42 (51.41) 50.78 (59.36) 2.13* Composite Welfare Attitudes scale (points on 6-point scale) 0.55 (1.40) −0.19 (1.70) −2.50* Symbolic Racism (points on 6-point scale) −0.47 (1.03) 0.07 (1.30) 2.40* Minority participants Money cut from welfare (in millions of dollars) 52.81 (77.44) 57.25 (116.41) 0.13 Composite Welfare Attitudes scale (points on 6-point scale) 0.34 (1.56) 0.63 (1.60) 0.53 Symbolic Racism scale (points on 6-point scale) −0.09 (1.27) −0.78 (1.29) −1.56 * p < 0.05 Table 3. Effects of information highlighting whites’ majority population share or decline of white population share on white and minority participants’ welfare attitudes Majority Salient Decline Salient t Mean (SD) Mean (SD) White participants Money cut from welfare (in millions of dollars) 28.42 (51.41) 50.78 (59.36) 2.13* Composite Welfare Attitudes scale (points on 6-point scale) 0.55 (1.40) −0.19 (1.70) −2.50* Symbolic Racism (points on 6-point scale) −0.47 (1.03) 0.07 (1.30) 2.40* Minority participants Money cut from welfare (in millions of dollars) 52.81 (77.44) 57.25 (116.41) 0.13 Composite Welfare Attitudes scale (points on 6-point scale) 0.34 (1.56) 0.63 (1.60) 0.53 Symbolic Racism scale (points on 6-point scale) −0.09 (1.27) −0.78 (1.29) −1.56 Majority Salient Decline Salient t Mean (SD) Mean (SD) White participants Money cut from welfare (in millions of dollars) 28.42 (51.41) 50.78 (59.36) 2.13* Composite Welfare Attitudes scale (points on 6-point scale) 0.55 (1.40) −0.19 (1.70) −2.50* Symbolic Racism (points on 6-point scale) −0.47 (1.03) 0.07 (1.30) 2.40* Minority participants Money cut from welfare (in millions of dollars) 52.81 (77.44) 57.25 (116.41) 0.13 Composite Welfare Attitudes scale (points on 6-point scale) 0.34 (1.56) 0.63 (1.60) 0.53 Symbolic Racism scale (points on 6-point scale) −0.09 (1.27) −0.78 (1.29) −1.56 * p < 0.05 By contrast, minority participants proposed cutting roughly the same amount of money from welfare across conditions and showed opposite trends from whites on survey measures. These trends were not statistically significant, though due to the small number of minority participants, we have low statistical power to detect significance. For our purposes it is most important to note that the trends we observed among minorities were qualitatively different from those we found among whites. We further predicted that whites’ greater opposition to welfare in the Decline Salient condition would be driven by heightened racial resentment. To test this claim, we conducted statistical mediation analyses (Baron and Kenny 1986) for both of our measures of welfare opposition among white participants. Results of mediation analyses using the budget-allocation measure and the survey measure of welfare support are given in figure 4. Results for both measures indicate that racial resentment partially mediated the effect of experimental condition. These results support our prediction that whites who were told that whites’ population share is declining reported greater opposition to welfare as a result of heightened racial resentment. Figure 4. View largeDownload slide a: Results of mediation analysis for funding-cut measure. b. Results of mediation analysis for composite welfare measure Figure 4. View largeDownload slide a: Results of mediation analysis for funding-cut measure. b. Results of mediation analysis for composite welfare measure Finally, to examine alternative explanations, we tested whether white participants in the Decline Salient condition expressed more conservative opinions on non-racial political issues (support for the Republican Party or defense spending), since evidence of social instability might lead respondents to more conservative opinions generally (Craig and Richeson 2014b; Jost et al. 2003). We found no significant effects of condition for these measures. Discussion In this study, we found that white participants responded to information suggesting that whites are losing their majority status—and hence their power as a voting bloc is declining—with increased racial resentment, which led to greater opposition to welfare. This mediation result is particularly noteworthy because it provides evidence that the link between whites’ racial resentment and welfare attitudes is not spuriously driven by individualist or conservative principles, since white participants expressed greater racial resentment under conditions of threat to racial status. Further, this effect was unique to whites, and we found no evidence that whites reported more conservative opinions on non-racial issues. We have to this point relied on self-report measures and statistical mediation analysis to establish the role of racial resentment. However, socially desirable response bias can affect the reliability of these measures (Bertrand and Mullainathan 2001). Also, statistical mediation analysis is subject to omitted variables bias, particularly if unmeasured mediators of an effect are correlated with the claimed mediator (Green, Ha, and Bullock 2010). Therefore, these correlational methods leave open the possibility that we have incorrectly identified the causal pathway through which racial status threats lead whites to greater welfare opposition. Thus, in our next study, we sought to establish experimentally whether the observed effects of status threat on whites’ welfare attitudes were driven by racial resentment. Study 3: Entitlement Programs Benefiting Whites versus Minorities As highlighted in figure 1, we expect that racial threats should stimulate anti-welfare sentiment among whites to the extent that they perceive these programs to primarily benefit racial-ethnic minorities (path d). We have to this point assumed that participants typically perceive welfare as disproportionately benefiting minorities, consistent with past research (Federico 2004; Gilens 1999). However, a more rigorous approach to evaluating our argument would involve systematically manipulating both whites’ sense of racial status threat and whether a given welfare program primarily benefits whites or minorities. If our reasoning is correct, we would only expect threats to increase whites’ opposition to programs believed to primarily benefit minorities, not programs believed to benefit whites. To this end, in Study 3 we sought to experimentally manipulate participants’ perceptions of whether specific federal welfare programs benefited whites or minorities by presenting them with information about the demographic breakdown of program beneficiaries. We also tested a new form of racial status threat, randomly varying whether participants were told the racial income gap had declined or grown. If threatened whites show heightened opposition to a welfare program that benefits minorities, but not to a program that benefits whites, it would be convincing evidence that racial status threats lead whites to oppose welfare as a result of heightened racial antipathy. Further, in this study we examine whether the effects we have so far attributed to racial resentment may be better explained by increased racial group identification and “opportunity hoarding” (Luttmer 2001; O’Brien 2017). If the effects of racial status threat are primarily attributable to heightened in-group identification, we would expect the strongest effects of our economic threat manipulation on whites’ support for a program disproportionately benefiting whites. However, if racial resentment sparked by racial threat is a primary mechanism, we would expect the strongest effects for whites’ support for a program disproportionately benefiting minorities. Additionally, we measure participants’ racial identification following the racial threat manipulation to examine whether heightened identification among whites mediates any observed effects. Method A total of 250 participants (129 men and 121 women) were recruited via MTurk. Participants were US residents between the ages of 18 and 68 (M = 35.6 years, SD = 12.7). Participants mainly identified as white (80.4 percent); 8.8 percent of participants identified as African American, 6.4 percent as Asian, and 4.4 percent as Latino. Respondents’ median household income was between $40,000 and $49,999, and 53.8 percent had at least a college degree. Participants answered a series of demographic questions. Then they were asked to answer questions about one of two charts describing income trends of different racial/ethnic groups during the last recession, ostensibly as part of a study of quantitative reasoning and social opinions. Participants were randomly assigned to one of two conditions. Participants assigned to the “Gap Expanding” condition saw information suggesting that while white Americans’ income suffered at the onset of recession, they had returned to previous levels by 2010. Meanwhile, black and Latino Americans’ incomes declined, such that the racial income gap had widened (figure 5a). Participants assigned to the “Gap Closing” condition were shown data indicating that whites’ incomes had fallen steadily during the recession while the incomes of other groups stayed about the same, such that the racial income gap had shrunk significantly (figure 5b). The graphs were accompanied by captions describing the trends in words (e.g., “From 2007 to 2010, the median (or typical) household income for white Americans dropped…”), reinforcing the information. Figure 5. View largeDownload slide a: Chart presented in gap expanding condition. b. Chart presented in gap closing condition Figure 5. View largeDownload slide a: Chart presented in gap expanding condition. b. Chart presented in gap closing condition Next, participants were presented with information about the Temporary Assistance for Needy Families (TANF) program and the unemployment insurance program, ostensibly to provide context to inform participants’ evaluations. In these program descriptions, the racial composition of program beneficiaries was varied such that each participant evaluated one program that primarily benefited whites and another that benefited minorities. Half the participants were randomly assigned to see information that TANF beneficiaries are “…mostly African American (40 percent) or Latino (30 percent)…” and unemployment insurance beneficiaries are “…mostly white (77 percent)….”; the other half of participants saw the opposite information, that is, that unemployment insurance mostly benefits minorities, and TANF mostly benefits whites. Information on beneficiaries’ age, gender, and race was presented graphically to increase the impact of the manipulation. The order of program presentation was randomly counterbalanced. Following each program description, participants indicated their agreement on seven-point scales with four statements measuring support for the program and whether they saw it as a good use of government resources (e.g., “I support the [unemployment insurance/TANF (Temporary Assistance for Needy Families)] program”). These items were averaged to create a composite measure of program support (Cronbach’s α = 0.90) centered at 2, ranging from −1 to 5. Next, participants’ racial identification was measured with the “identity” and “private regard” subscales of a standard measure of collective self-esteem (Luhtanen and Crocker 1992). Participants indicated their agreement on seven-point scales with eight questions measuring how strongly they identified with their racial/ethnic group and how positively they felt about their racial/ethnic group (e.g., “My racial/ethnic group is an important reflection of who I am”). These items were averaged to create a composite (Cronbach’s α = 0.83), centered at 0, ranging from −3 to 3. Following recent research (Berinsky, Margolis, and Sances 2013), we used a more rigorous approach than in our previous study to identify participants who were sufficiently attentive to the study. Rather than relying on one survey item as we did in Study 2, here we included five screening questions throughout the study. Participants were coded as inattentive if they answered more than three questions incorrectly. We gauged suspicion regarding our stimulus materials by asking participants what they thought this study was about. Finally, participants were thanked and debriefed. Results Twelve participants were dropped from analysis because they reported suspicion of the study’s purpose, and two participants were dropped for being insufficiently attentive to the study, leaving a total of 236 participants for analysis.5 We model participants’ program support using a random-effects model where the independent variables are assignment to the Gap Expanding or Gap Closing condition (between-subjects factor) and whether the program was described as benefiting whites or minorities (within-subjects factor). We also include a dummy variable for whether the program was TANF or unemployment insurance. We first model whites’ program support (model 1 in table 4). These participants reported significantly less support for TANF than for unemployment insurance net of condition. We found no significant main effects of exposure to the Gap Closing condition or the racial composition of program beneficiaries. There was, however, a significant, negative interaction between assignment to the Gap Closing condition and the race of beneficiaries (p = 0.04), such that whites showed uniquely low support for programs that benefited minorities if they had been told that the white income advantage is closing. Table 4. Random-effects models predicting white and minority participants’ support for a social safety net program by experimental conditions: coefficients and standard errors Whites’ program support Minorities’ program support (1) (2) Program benefits minorities 0.130 0.258 (0.115) (−0.068) Gap Closing condition −0.057 0.380 (0.216) (0.410) Gap Closing condition × Program benefits minorities −0.338* −0.294 (0.165) (0.369) Program was TANF −0.305*** −0.054 (0.083) (0.185) Intercept 2.977*** 2.785*** (0.155) (0.299) N 189 47 Within-subjects R2 0.09 0.05 Between-subjects R2 0.01 0.01 Whites’ program support Minorities’ program support (1) (2) Program benefits minorities 0.130 0.258 (0.115) (−0.068) Gap Closing condition −0.057 0.380 (0.216) (0.410) Gap Closing condition × Program benefits minorities −0.338* −0.294 (0.165) (0.369) Program was TANF −0.305*** −0.054 (0.083) (0.185) Intercept 2.977*** 2.785*** (0.155) (0.299) N 189 47 Within-subjects R2 0.09 0.05 Between-subjects R2 0.01 0.01 Note: Reference categories are “Program benefits whites” and “Gap Expanding” conditions; * p < 0.05 ** p < 0.01 ***p < 0.001. Table 4. Random-effects models predicting white and minority participants’ support for a social safety net program by experimental conditions: coefficients and standard errors Whites’ program support Minorities’ program support (1) (2) Program benefits minorities 0.130 0.258 (0.115) (−0.068) Gap Closing condition −0.057 0.380 (0.216) (0.410) Gap Closing condition × Program benefits minorities −0.338* −0.294 (0.165) (0.369) Program was TANF −0.305*** −0.054 (0.083) (0.185) Intercept 2.977*** 2.785*** (0.155) (0.299) N 189 47 Within-subjects R2 0.09 0.05 Between-subjects R2 0.01 0.01 Whites’ program support Minorities’ program support (1) (2) Program benefits minorities 0.130 0.258 (0.115) (−0.068) Gap Closing condition −0.057 0.380 (0.216) (0.410) Gap Closing condition × Program benefits minorities −0.338* −0.294 (0.165) (0.369) Program was TANF −0.305*** −0.054 (0.083) (0.185) Intercept 2.977*** 2.785*** (0.155) (0.299) N 189 47 Within-subjects R2 0.09 0.05 Between-subjects R2 0.01 0.01 Note: Reference categories are “Program benefits whites” and “Gap Expanding” conditions; * p < 0.05 ** p < 0.01 ***p < 0.001. When evaluating programs they believed primarily benefited whites, white participants who were told the racial income gap was shrinking did not differ much in their program support from white participants who were told the income gap was widening. However, when the program primarily benefited African Americans and Latinos, whites in the Gap Closing condition reported less support than in the Gap Expanding condition (figure 6a). Figure 6. View largeDownload slide a: Means of whites’ support for social safety net programs, by experimental condition. b: Means of minorities’ support for social safety net programs, by experimental condition Figure 6. View largeDownload slide a: Means of whites’ support for social safety net programs, by experimental condition. b: Means of minorities’ support for social safety net programs, by experimental condition This pattern suggests that the significant interaction found in model 1 is driven by decreased support for a welfare program benefiting minorities, not increased support for a program benefiting whites. In addition, we found no significant effect of the Gap Closing condition on white participants’ racial identification (p > 0.58), and racial identification did not moderate the interaction between the Gap Closing condition and race of program beneficiaries (p > 0.95) (results available upon request). Next, we turn to minority participants to examine whether these results are unique to whites’ threat response. Again we found no significant main effects of our experimental manipulations (model 2 in table 4). We found a negative, though insignificant (p = 0.43), interaction of exposure to the Gap Closing condition and the race of program beneficiaries. Minorities supported a welfare program more if they were in the Gap Closing condition and the program in question benefited whites (figure 6b). While statistically insignificant, this interactional pattern may have been significant with a larger sample of minority participants. If this finding proved reliable, it would be open to several explanations. Minorities might anticipate whites’ negative responses to status threats, increasing their support for programs that benefit whites as a gesture of appeasement. Alternately, minorities (like Americans generally) may be more likely to see whites as members of the “deserving poor,” and thus increase support for programs supporting whites when whites appear to need them. For our purposes, it is sufficient to note that minority participants responded in a different manner from white participants, consistent with our claim that information on narrowing income differences was experienced as a racial threat by whites. Discussion In this study, we provide evidence supporting the final link in our logic: that racial threats lead to anti-welfare sentiment among whites because they perceive such programs to mostly benefit minorities. Where a welfare program was portrayed as primarily benefiting whites, threatened white participants reported almost identical support for welfare as unthreatened white participants. These findings provide discerning support for our claim that whites’ opposition to welfare following racial threat is due to increased racial resentment. The within-subjects design of Study 3 further strengthens our claims that racial status threats uniquely increase whites’ opposition to programs benefiting minorities. Rather than randomly assigning some participants to evaluate a welfare program that benefited whites and other participants to evaluate a program that benefited minorities, all participants rated both types of programs. Thus, the same threatened white participants who expressed opposition to a program benefiting minorities on average expressed greater support for a program benefiting whites, within the same survey setting. While white Americans often engage in impression management to appear color-blind, such efforts did not overcome the effect of racial status threat on participants’ willingness to voice heightened opposition to programs that benefit minorities. General Discussion Our findings provide consistent support for our claim that white Americans’ welfare attitudes are shaped by concerns about the status of their racial group in American society. First, we found that whites’ and minorities’ welfare attitudes diverged in 2008, the year of the candidacy and election of President Obama and the financial crisis, and whites’ racial resentment rose during this time as well. Next, we found that white Americans who saw a demographic report emphasizing the decline of the white majority tended thereafter to voice greater opposition to welfare, and this effect was partially mediated by increased racial resentment. In our final study, we found that information threatening the white economic advantage resulted in increased opposition to welfare programs when whites perceived those programs to primarily benefit minorities, but did not affect support for programs portrayed as benefiting whites. These findings implicate racial status threats as a causal factor shaping whites’ opposition to welfare. Results of our studies are consistent with our model of the role of racial prejudice in whites’ attitudes toward welfare, more so than rival theoretical explanations. For example, a critic could argue that the threats we studied led respondents toward a more diffusely conservative political orientation because they indicated social instability (Jost et al., 2003), not because they threatened the racial status hierarchy. However, in Studies 1 and 2 we found no effects of threat on measures of non-racial attitudes. Further, this explanation cannot account for our findings that threat effects were unique to white respondents, nor why whites’ support for social welfare programs following threats depended critically on the perceived race of program beneficiaries. Nevertheless, the current research has limitations we hope future research will address. While our experimental evidence provides strong support for a causal link between racial status threats and welfare opposition, these experiments are potentially limited in their generalizability due to the use of convenience samples for practical reasons. Ideally, our experimental findings would be replicated using a nationally representative sample. Further, our focus here on whites’ attitudes means our work offers limited insights on how indications that the racial status hierarchy is changing might affect minorities’ attitudes toward welfare. Further empirical and theoretical work considering how concerns about racial group status may affect minorities’ welfare attitudes is a promising avenue for future research. Finally, we would welcome exploration of the sociodemographic and contextual factors that may affect whites’ perceptions of, or responses to, racial status threat. This variation may prove useful in explaining established patterns in public opinion and political mobilization around a number of racialized policy issues. Our research provides a novel theoretical synthesis of the role of prejudice in affecting white Americans’ attitudes toward welfare and offers a perspective for understanding a current, consequential trend in public opinion. We provide evidence that racial resentment rises in response to macro-historical trends that threaten whites’ standing in the racial status hierarchy, particularly in eras of economic decline. We demonstrate that racial resentment rather than in-group identification drives the relationship between racial threat and opposition to welfare. Further, because public attitudes partially drive developments in anti-poverty policy, these findings suggest that perceptions of rising minority power, declines in whites’ relative socioeconomic status, or other perceived macro-level threats to whites’ racial status may provoke adoption of more restrictive welfare regimes. Previous scholars have noted that minorities’ civil rights victories, increasing migration of nonwhites into state jurisdictions, and episodes of racial conflict have spurred periods of anti-welfare sentiment among whites and state policies to roll back social safety net programs (Brown 2013; Quadagno 1994; Reese 2005; Soss, Fording, and Schram 2011). The current research specifies why and how such periods of “welfare backlash” may occur, making adoption of policies that restrict or curtail welfare state development more likely. In addition, our findings demonstrate that the documented link between whites’ racial resentment and opposition to welfare programs cannot be explained in terms of principled conservatism, with no role for racial prejudice. We present a model of racial threat that moves beyond the standard operationalization as proportion of minorities in the local context, a contribution earlier scholars have noted is needed in this literature (Brown 2013; Muller 2012; Soss and Bruch 2008) and which may prove relevant to other domains of racialized public policy. We demonstrate that the relationship between racial resentment and welfare opposition remains robust even in the post-welfare reform era, when benefits have been reduced significantly and are subject to stringent sanctions (Soss, Fording, and Schram 2011). Finally, by examining how salient events may lead to perceptions of racial status threat among whites, we provide evidence for a social psychological mechanism promoting the durability of racial inequality through individuals’ political responses. Because status rank is hierarchical and zero-sum, any increases in economic or political power of lower-status groups can be interpreted as a threat to the relative standing of dominant group members. Thus, any progress toward equality may provoke resentment on the part of dominant group members, who may react politically in ways that undermine or even reverse progress to racial equality. In the case of American social welfare programs, this further implies that evidence of increased racial equality could exacerbate overall economic inequality. As whites attempt to undermine racial progress they see as threatening their group’s status, they increase opposition to programs intended to benefit poorer members of all racial groups. Footnotes 1 While differences exist in the conceptualization of symbolic racism and racial resentment, they have been operationalized similarly over time, such that most analysts treat them as essentially interchangeable concepts (e.g., Sears and Henry 2005). 2 We begin our study period in 2000 because this was the first year ANES administered questions on welfare attitudes and racial resentment after the passage of the Personal Responsibility and Work Opportunity Act in 1996. Results of analyses including earlier survey waves are substantively identical to those presented here. 3 Brooks and Manza (2013) argue that in the face of economic disruption, partisan identifiers looked for cues on how to interpret the situation from partisan sources, leading to greater opposition to welfare among Republicans especially. This explanation would be incompatible with our own if Republicans of all racial/ethnic groups increased welfare opposition, suggesting the effect is not due to whites’ perceptions of racial threat. However, if white Republican increased welfare opposition as a result of heightened racial resentment, this would suggest that both mechanisms help explain increased welfare opposition. Additional analyses support the latter, complementary account. Race and ideology appear to interact in producing opposition to welfare, such that white conservatives show increased racial resentment and welfare opposition in the post-2008 period. 4 All predicted effects remained significant (p < 0.05) if these participants were retained in analysis. 5 All predicted effects remained significant (p < 0.05) if inattentive participants were retained in analysis. About the Authors Rachel Wetts is a PhD candidate in Sociology at the University of California, Berkeley. Her research interests include social psychology, American political culture, and organizational sociology. Ongoing projects use experimental and quantitative methods to investigate how political elites activate and harness feelings of racial prejudice, and how cultural and organizational processes affect which conceptions of climate change become dominant in mainstream media. Robb Willer is a Professor of Sociology at Stanford University. His research interests include politics, morality, cooperation, and hierarchy. His recent research has focused on sources of political division and mechanisms for creating political consensus. His research has appeared in American Sociological Review, American Journal of Sociology, Social Forces, Annual Review of Sociology, Journal of Personality and Social Psychology, Proceedings of the National Academy of Sciences, and elsewhere. Supplementary Material Supplementary material is available at Social Forces online. References Abascal , Maria . 2015 . “ Us and Them: Black-White Relations in the Wake of Hispanic Population Growth .” American Sociological Review 80 ( 4 ): 789 – 813 . Google Scholar CrossRef Search ADS Baron , Reuben , and David Kenny . 1986 . “ The Moderator–Mediator Variable Distinction in Social Psychological Research: Conceptual, Strategic, and Statistical Considerations .” Journal of Personality and Social Psychology 51 ( 6 ): 1173 – 82 . Google Scholar CrossRef Search ADS Berinsky , Adam , Michele Margolis , and Michael Sances . 2013 . “ Separating the Shirkers from the Workers? Making Sure Respondents Pay Attention on Self‐Administered Surveys .” American Journal of Political Science 58 ( 3 ): 739 – 53 . Google Scholar CrossRef Search ADS Bernstein , Michael , Steven Young , and Heather Claypool . 2010 . “ Is Obama’s Win a Gain for Blacks? Changes in Implicit Racial Prejudice Following the 2008 Election .” Social Psychology 41 ( 3 ): 147 – 51 . Google Scholar CrossRef Search ADS Bertrand , Marianne , and Sendhil Mullainathan . 2001 . “ Do People Mean What They Say? Implications for Subjective Survey Data .” American Economic Review 91 ( 2 ): 67 – 72 . Google Scholar CrossRef Search ADS Blalock , Hubert . 1967 . Toward a Theory of Minority-Group Relations . New York : Wiley . Blumer , Herbert . 1958 . “ Race Prejudice as a Sense of Group Position .” Pacific Sociological Review 1 ( 1 ): 3 – 7 . Google Scholar CrossRef Search ADS Bobo , Lawrence . 1998 . “ Race, Interests, and Beliefs About Affirmative Action: Unanswered Questions and New Directions .” American Behavioral Scientist 41 ( 7 ): 985 – 1003 . Google Scholar CrossRef Search ADS ——— . 1999 . “ Prejudice as Group Position: Microfoundations of a Sociological Approach to Racism and Race Relations .” Journal of Social Issues 55 ( 3 ): 445 – 72 . Google Scholar CrossRef Search ADS Bobo , Lawrence and Camille Zubrinsky . 1996 . “ Attitudes on Residential Integration: Perceived Status Differences, Mere In-Group Preference, or Racial Prejudice? ” Social Forces 74 ( 3 ): 883 – 909 . Google Scholar CrossRef Search ADS Brooks , Clem , and Jeff Manza . 2013 . “ A Broken Public? Americans’ Responses to the Great Recession .” American Sociological Review 78 ( 5 ): 727 – 48 . Google Scholar CrossRef Search ADS Brown , Hana . 2013 . “ Racialized Conflict and Policy Spillover Effects: The Role of Race in the Contemporary US Welfare State .” American Journal of Sociology 119 ( 2 ): 394 – 443 . Google Scholar CrossRef Search ADS Buhrmester , Michael , Tracy Kwang , and Samuel Gosling . 2011 . “ Amazon’s Mechanical Turk A New Source of Inexpensive, Yet High-Quality, Data? ” Perspectives on Psychological Science 6 ( 1 ): 3 – 5 . Google Scholar CrossRef Search ADS Carney , Riley , and Ryan Enos . 2015 . “Conservatism, Just World Belief, and Racism.” Paper presented at the Annual Meeting of the Midwest Political Science Association, Chicago, IL, April 16–19. CBS . 1996 . CBS News Survey, October. Craig , Maureen , and Jennifer Richeson . 2014 a. “ More Diverse Yet Less Tolerant? How the Increasingly Diverse Racial Landscape Affects White Americans’ Racial Attitudes .” Personality and Social Psychology Bulletin 40 : 750 – 61 . Google Scholar CrossRef Search ADS ——— . 2014 b. “ On the Precipice of a ‘Majority-Minority’ America: Perceived Status Threat from the Racial Demographic Shift Affects White Americans’ Political Ideology .” Psychological Science 25 : 1189 – 97 . Google Scholar CrossRef Search ADS DeBell , Matthew . 2010 . “How to Analyze ANES Survey Data.” ANES Technical Report Series No. nes012492. Palo Alto, CA, and Ann Arbor: Stanford University and the University of Michigan. Dougherty , Conor . 2009 . “Minority Turnout Was Critical to Obama’s Election, Data Show.” Wall Street Journal, July 21. http://online.wsj.com/articles/SB124813588601666995. Erikson , Robert , Michael MacKuen , and James Stimson . 2002 . The Macro Polity . Cambridge : Cambridge University Press . Federico , Christopher . 2004 . “ When Do Welfare Attitudes Become Racialized? The Paradoxical Effects of Education .” American Journal of Political Science 48 ( 2 ): 374 – 91 . Google Scholar CrossRef Search ADS Fox , Cybelle . 2004 . “ The Changing Color of Welfare? How Whites’ Attitudes toward Latinos Influence Support for Welfare .” American Journal of Sociology 110 ( 3 ): 580 – 625 . Google Scholar CrossRef Search ADS Gilens , Martin . 1996 . “ ‘Race Coding’ and White Opposition to Welfare .” American Political Science Review 90 ( 3 ): 593 – 604 . Google Scholar CrossRef Search ADS ——— . 1999 . Why Americans Hate Welfare: Race, Media, and the Politics of Antipoverty Policy . Chicago : University of Chicago Press . Goldman , Seth , and Diana Mutz . 2014 . The Obama Effect: How the 2008 Campaign Changed White Racial Attitudes . New York : Russell Sage . Gomez , Brad , and J. Matthew Wilson . 2006 . “ Rethinking Symbolic Racism: Evidence of Attribution Bias .” Journal of Politics 68 ( 3 ): 611 – 25 . Google Scholar CrossRef Search ADS Green , Donald , Shang Ha , and John Bullock . 2010 . “ Enough Already about ‘Black Box’ Experiments: Studying Mediation Is More Difficult Than Most Scholars Suppose .” Annals of the American Academy of Political and Social Science 628 ( 1 ): 200 – 208 . Google Scholar CrossRef Search ADS Hancock , Ange-Marie . 2004 . The Politics of Disgust: The Public Identity of the Welfare Queen . New York : NYU Press . Henry , Patrick , and David Sears . 2002 . “ The Symbolic Racism 2000 Scale .” Political Psychology 23 ( 2 ): 253 – 83 . Google Scholar CrossRef Search ADS Hopkins , Daniel . 2010 . “ Politicized Places: Explaining Where and When Immigrants Provoke Local Opposition .” American Political Science Review 104 ( 1 ): 40 – 60 . Google Scholar CrossRef Search ADS Huddy , Leonie , and Stanley , Feldman . 2009 . “ On Assessing the Political Effects of Racial Prejudice .” Annual Review of Political Science 12 : 423 – 47 . Google Scholar CrossRef Search ADS Johnson , Martin . 2001 . “ The Impact of Social Diversity and Racial Attitudes on Social Welfare Policy .” State Politics & Policy Quarterly 1 ( 1 ): 27 – 49 . Google Scholar CrossRef Search ADS Jost , John , Jack Glaser , Arie Kruglanski , and Frank Sulloway . 2003 . “ Political Conservatism as Motivated Social Cognition .” Psychological Bulletin 129 ( 3 ): 339 – 75 . Google Scholar CrossRef Search ADS Kinder , Donald , and David Sears . 1981 . “ Prejudice and Politics: Symbolic Racism Versus Racial Threats to the Good Life .” Journal of Personality and Social Psychology 40 ( 3 ): 414 – 31 . Google Scholar CrossRef Search ADS Lowndes , Joseph . 2008 . From the New Deal to the New Right: Race and the Southern Origins of Modern Conservatism . New Haven, CT : Yale University Press . Luhtanen , Riia , and Jennifer Crocker . 1992 . “ A Collective Self-Esteem Scale: Self-Evaluation of One’s Social Identity .” Personality and Social Psychology Bulletin 18 ( 3 ): 302 – 18 . Google Scholar CrossRef Search ADS Luttmer , Erzo . 2001 . “ Group Loyalty and the Taste for Redistribution .” Journal of Political Economy 109 ( 3 ): 500 – 28 . Google Scholar CrossRef Search ADS Martin , Isaac William . 2008 . The Permanent Tax Revolt: How the Property Tax Transformed American Politics . Stanford, CA : Stanford University Press . Muller , Christopher . 2012 . “ Northward Migration and the Rise of Racial Disparity in American Incarceration, 1880–1950 .” American Journal of Sociology 118 ( 2 ): 281 – 326 . Google Scholar CrossRef Search ADS Neblo , Michael . 2009 . “ Three-Fifths a Racist: A Typology for Analyzing Public Opinion about Race .” Political Behavior 31 ( 1 ): 31 – 51 . Google Scholar CrossRef Search ADS Neubeck , Kenneth , and Noel Cazenave . 2001 . Welfare Racism: Playing the Race Card Against America’s Poor . New York : Routledge . O’Brien , Rourke . 2017 . “ Redistribution and the New Fiscal Sociology: Race and the Progressivity of State and Local Taxes .” American Journal of Sociology 122 ( 4 ): 1015 – 49 . Google Scholar CrossRef Search ADS Oliver , J. Eric , and Tali Mendelberg . 2000 . “ Reconsidering the Environmental Determinants of White Racial Attitudes .” American Journal of Political Science 44 ( 3 ): 574 – 89 . Google Scholar CrossRef Search ADS Outten , H. Robert , Michael Schmitt , Daniel Miller , and Amber Garcia . 2012 . “ Feeling Threatened About the Future: Whites’ Emotional Reactions to Anticipated Ethnic Demographic Changes .” Personality and Social Psychology Bulletin 38 ( 1 ): 14 – 25 . Google Scholar CrossRef Search ADS Page , Benjamin , and Robert Shapiro . 1992 . The Rational Public: Fifty Years of Trends in Americans’ Policy Preferences . Chicago : University of Chicago Press . Google Scholar CrossRef Search ADS Parker , Christopher . 2016 . “ Race and Politics in the Age of Obama .” Annual Review of Sociology 42 : 217 – 30 . Google Scholar CrossRef Search ADS Parker , Christopher , Mark Sawyer , and Christopher Towler . 2009 . “ A Black Man in the White House? The Role of Racism and Patriotism in the 2008 Presidential Election .” Du Bois Review 6 ( 1 ): 193 – 217 . Google Scholar CrossRef Search ADS Pasek , Josh , Tobias Stark , Jon Krosnick , Trevor Tompson , and B. Keith Payne . 2014 . “ Attitudes toward Blacks in the Obama Era: Changing Distributions and Impacts on Job Approval and Electoral Choice, 2008–2012 .” Public Opinion Quarterly 78 ( S1 ): 276 – 302 . Google Scholar CrossRef Search ADS Quadagno , Jill . 1994 . The Color of Welfare: How Racism Undermined the War on Poverty . New York : Oxford University Press . Quillian , Lincoln . 1995 . “ Prejudice as a Response to Perceived Group Threat: Population Composition and Anti-Immigrant and Racial Prejudice in Europe .” American Sociological Review 60 ( 4 ): 586 – 611 . Google Scholar CrossRef Search ADS Reese , Ellen . 2005 . Backlash Against Welfare Mothers: Past and Present . Berkeley : University of California Press . Roth , Byron . 1990 . “ Social Psychology’s Racism .” Public Interest 98 : 26 – 36 . Schmidt , Kathleen , and Brian Nosek . 2010 . “ Implicit (and Explicit) Racial Attitudes Barely Changed during Barack Obama’s Presidential Campaign and Early Presidency .” Journal of Experimental Social Psychology 46 ( 2 ): 308 – 14 . Google Scholar CrossRef Search ADS Sears , David , and Patrick Henry . 2005 . “ Over Thirty Years Later: A Contemporary Look at Symbolic Racism .” Advances in Experimental Social Psychology 37 : 95 – 150 . Google Scholar CrossRef Search ADS Sniderman , Paul , and Edward Carmines . 1997 . Reaching Beyond Race . Cambridge, MA : Harvard University Press . Sniderman , Paul , Gretchen Crosby , and William Howell . 2000 . “The Politics of Race.” In Racialized Politics: The Debate About Racism in America , edited by David Sears , Jim Sidanius , and Lawrence Bobo , 236–79 . Chicago : University of Chicago Press . Soss , Joe , and Sarah Bruch . 2008 . “Marginalization Matters: Rethinking Race in the Analysis of State Politics and Policy.” Paper presented at the Annual Meeting of the American Political Science Association, Boston, MA, August 27–30. Soss , Joe , Richard Fording , and Sanford Schram . 2011 . Disciplining the Poor: Neoliberal Paternalism and the Persistent Power of Race . Chicago : University of Chicago Press . Google Scholar CrossRef Search ADS Tesler , Michael , and David Sears . 2010 . Obama’s Race: The 2008 Election and the Dream of a Post-Racial America . Chicago : University of Chicago Press . Google Scholar CrossRef Search ADS US Census Bureau, Population Division . 2012 . “Table 6. Percent Distribution of the Projected Population by Race and Hispanic Origin for the United States: 2015 to 2060.” NP2012-T6. Valentino , Nicholas , and Ted Brader . 2011 . “ The Sword’s Other Edge: Perceptions of Discrimination and Racial Policy Opinion after Obama .” Public Opinion Quarterly 75 ( 2 ): 201 – 26 . Google Scholar CrossRef Search ADS Weinberg , Jill , Jeremy Freese , and David McElhattan . 2014 . “ Comparing Data Characteristics and Results of an Online Factorial Survey Between a Population-Based and Crowdsource-Recruited Sample .” Sociological Science 1 : 292 – 310 . Google Scholar CrossRef Search ADS Wright , Gerald . 1977 . “ Contextual Models of Electoral Behavior: The Southern Wallace Vote .” American Political Science Review 71 ( 2 ): 497 – 508 . Google Scholar CrossRef Search ADS Wright , Gerald , Robert Erikson , and John McIver . 1987 . “ Public Opinion and Policy Liberalism in the American States .” American Journal of Political Science 31 ( 4 ): 980 – 1001 . Google Scholar CrossRef Search ADS © The Author(s) 2018. Published by Oxford University Press on behalf of the University of North Carolina at Chapel Hill. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

Social ForcesOxford University Press

Published: May 31, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off