The Racialization of Policies

Americans, especially white Americans, have strong cognitive narratives and stereotypes associated with race-based policies (Brader et al., Reference Brader, Valentino and Suhay2008; Gilens, Reference Gilens2009; Hancock, Reference Hancock2004). While deeply rooted psychological factors predisposing humans to group-based thinking contribute to these patterns, elite cues often activate these predispositions in intentional efforts to undermine support for certain policies by racializing them. White Americans tend to support policies seen as “universal” more than policies targeted to ameliorate racial inequalities, which are cast as privileging a racially minoritized group over their own (Hamilton, Reference Hamilton1977; Gillespie, Reference Gillespie2010, Reference Gillespie2012; McIlwain and Caliendo, Reference McIlwain and Caliendo2011, but see Stout Reference Stout2015, Reference Stout2020).

Racialization refers to the process by which issues and policies become linked—consciously or subconsciously—with racially minoritized groups in the eyes of the public. A canonical example is welfare policy. Elites worked to promote the idea that welfare programs only benefit “undeserving” and “lazy” Black Americans, and opposition grew as a result (Brown-Iannuzzi et al., Reference Brown-Iannuzzi2021; Gilens, Reference Gilens2009; Hancock, Reference Hancock2004; Soss et al., Reference Soss, Fording and Schram2011). Racialization of health care reform using the label of “Obamacare” is another example in which tying health reform to a Black president contributed to the activation of racial predispositions when expressing health policy attitudes and fomented public opposition (Tesler, Reference Tesler2016).

By linking policies with particular (minoritized) ethnoracial groups, politicians and other elites activate whites’ racial animus to depress support for such policies (Hutchings and Jardina, Reference Hutchings and Jardina2009; Mendelberg, Reference Mendelberg2001; Valenzuela and Reny, Reference Valenzuela and Reny2020). Survey experiments fielded during the 1990s and early 2000s found that subtle references to race were especially effective at mobilizing opposition to race-based policies. Implicit appeals, such as those that rely on code words or racial imagery, worked to cue whites’ hostile racial attitudes and activate their use in opinion formation. Recent work has demonstrated that whites have come to use racial attitudes frequently in evaluating policies regardless of how subtle or overt the racial cues may be (Christiani, Reference Christiani2021; Reny et al., Reference Reny, Valenzuela and Collingwood2020; Valentino et al., Reference Valentino, Neuner and Matthew Vandenbroek2018). As explicit racial appeals have become less stigmatized and more effective for those who wield them, racial attitudes appear to be constantly activated, especially among whites (Engelhardt, Reference Engelhardt2021; Valentino et al., Reference Valentino, Neuner and Matthew Vandenbroek2018).

This literature about racialization and the attitudes that follow from linking issues to minoritized groups suggests that any mention of race within a policy domain would immediately map onto longstanding racial narratives and undermine white Americans’ support for policies in that domain. Hearing about racial inequalities that emerged during the pandemic may simply prompt white people to draw on existing and pervasive narratives rooted in anti-Black stereotypes. Such a process would enable them to readily justify racial inequalities with rationalizations rooted in negative racial stereotypes about individual failings on the part of Black Americans, depressing support for race-based policies as a result.

This is how prompting attention to race, and especially racial differences, has been shown to operate time and again across a wide array of policy domains ranging from immigration and criminal justice to welfare and health policy. In fact, LaFleur Stephens-Dougan (Reference Stephens-Dougan2023) found this precise pattern with specific reference to the COVID-19 pandemic. She shows that simply highlighting how Black Americans suffered disproportionate COVID mortality reduced support for COVID mitigation efforts among whites with higher levels of racial prejudice, presumably because such efforts came to be seen as a burden that primarily benefited Black Americans. Although Stephens-Dougan (Reference Stephens-Dougan2023) did not aim to evaluate reactions to information about the structural origins of pandemic disparities as we do here, her findings suggest that white Americans with higher levels of racial prejudice readily deployed their existing racial narratives within this new policy domain. Similarly, white Republicans have been found to harbor higher levels of racial resentment (Engelhardt, Reference Engelhardt2021) and to have more exposure to racial stereotypes (Reese, Reference Reese2005; Stefancic and Delgado, Reference Stefancic and Delgado1996). When individuals receive new information, they are not taking it in as a blank slate or as individuals with specified knowledge about each individual policy item. Instead, part of the way people receive information is through the lens of their prior beliefs (Zaller, Reference Zaller1991, Reference Zaller1992). Thus, these preconceptions, exposures, and beliefs may motivate resistance to this new information, especially among white Republicans who are more likely to encounter and adhere to negative racial narratives.

If this is the process at work, then encountering information about racial disparities in the pandemic—even if the information emphasizes the structural roots of these inequalities—would increase white opposition to policies designed to ameliorate racial inequality. Moreover, this process would likely be most pronounced among white respondents with higher levels of racial resentment and among white Republicans.

Structural Racism and the American Dream

But perhaps the way we talk about race and racial inequalities makes a difference for people’s attitudes. Thus far, there is limited experimental evidence examining what happens when individual-level racial narratives are countered by structural accounts of racism and inequality. Can such an intervention interrupt the cognitive processes associated with racialization, which often depend on directing individualized blame toward members of minoritized groups? The theoretical framework below fleshes out such a possibility.

Despite the fact that many white Americans support policies designed to make racial inequality worse, they often hold onto egalitarian norms rooted in the myth of the American Dream (Bobo, Reference Bobo2001; Bonilla-Silva, Reference Bonilla-Silva2009; Hochschild, Reference Hochschild1996; Wolak and Peterson, Reference Wolak and Peterson2020). While people’s understandings of the “American Dream” are multifaceted and continuously in flux (Allen, Reference Allen2019; Hanson and White, Reference Hanson and White2011), here we define it as the belief that anyone can succeed in America as long as they work hard (Wolak and Peterson, Reference Wolak and Peterson2020). This is a widely held ideology throughout the population (Hanson and White, Reference Hanson and White2011; Hochschild, Reference Hochschild1996; Vasilogambros, Reference Vasilogambros2016; Wolak and Peterson, Reference Wolak and Peterson2020). Moreover, the tenets of this ideology are consequential for structuring political attitudes (Hochschild, Reference Hochschild1996; Wolak and Peterson, Reference Wolak and Peterson2020). While belief in the American Dream is widespread, Wolak and Peterson (Reference Wolak and Peterson2020) demonstrate that adherence to this idea varies over time and responds to economic and social change. As inequality increases, fewer Americans subscribe to the notion that hard work results in success. Inequality challenges the myth of the American Dream and functions to erode beliefs in its tenets. These findings suggest that inequality poses a challenge to meritocratic values.

White Americans often ease the cognitive dissonance associated with this contradiction by blaming individuals for societal inequalities rather than finding fault with society as a whole. The individualization of blame enables white people to rationalize patterns of racial inequality through stereotypes based on race. Thus, people may only become cognizant of the challenge systemic inequalities pose to the myth of the American Dream when they come to understand disadvantage as operating structurally—consistently harming certain groups of people in ways that result from the place they occupy in a social hierarchy rather than individual failings.

But structural understandings of racism do not arise naturally from typical American ideals and myths, and many white Americans avoid grappling with the realities of the racial inequalities that pervade American society. Indeed, white Americans rarely encounter information that describes racism as structural (Adams et al., Reference Adams2008; Lopez et al., Reference Lopez, Gurin and Nagda1998). Even when they do hear about racial inequalities, a desire to maintain a positive view of their own racial group can result in an unwillingness to accept this information (Bonam et al., Reference Bonam2019; Nelson et al., Reference Nelson, Adams and Salter2013). For instance, Kraus et al. (Reference Kraus2019) demonstrate that “willful ignorance” leads Americans to vastly underestimate the racial wealth gap because they want to see society as “fair, just, and merit-based” (906). When people are confronted with information that contradicts this perception, they are motivated to explain away racial inequality either as statistical noise or as justifiable (Kraus et al., Reference Kraus2019). Similarly, Feldman and Huddy (Reference Feldman and Huddy2018) find that whites are motivated to deny the existence of racial discrimination as a result of “racial prejudice, political ideology, and plain ignorance” (189). Mueller (Reference Mueller2017) identifies four “epistemic maneuvers” that white college students use to reestablish ignorance even when confronted with assignments that specifically engage with structural inequality, allowing them ultimately to maintain their commitment to colorblindness. These studies demonstrate the pervasiveness of Charles Mills (Reference Mills1997)’s “epistemology of ignorance,” which he argues is central to the maintenance of white supremacy (93).

Nevertheless, the prominence of the Black Lives Matter (BLM) movement during the mid to late 2010s ^{Footnote 1} did push conversations about structural racism forward. Recent observational work has sought to understand whether this movement was able to shift attitudes, largely to mixed findings. On the one hand, some work has found that BLM protests increased perceptions that Black Americans experience discrimination, but only among politically liberal and lower-prejudice white Americans (Reny and Newman, Reference Reny and Newman2021). Somewhat similarly, Sawyer and Gampa (Reference Sawyer and Gampa2018) showed that BLM helped make white people less pro-white in their attitudes, and Engelhardt (Reference Engelhardt2021, Reference Engelhardt2023) found that racial attitudes are increasingly polarized by party among whites and that white Democrats are expressing lower levels of racial resentment. However, there is some doubt about the extent to which these changes can last. Chudy and Jefferson (Reference Chudy and Jefferson2021) found that support for Black Lives Matter faded quickly, and careful qualitative analyses from Smith (Reference Smith2023) demonstrate that colorblind ideologies are still dominant, even in the wake of the BLM movement’s attempt to advance a more structural account. Despite BLM’s attempt to shift the discourse away from racism being individual and toward it being more structural, the extent to which this message has been received appears mixed. This pattern may again point to the dominance of an epistemology of ignorance among white Americans (Mills, Reference Mills1997).

But this prior work on the effects of structural narratives has largely focused on Black Lives Matter, specifically by name, which is a term, movement, and diffuse organization that has been unfortunately demonized by segments of the mass media (Smith, Reference Smith2023). What remains unclear is the extent to which information concerning structural racism that is not connected to an already polarized topic like BLM might be able to shift policy attitudes.

We theorize that encountering information emphasizing the systemic origins of racial disparities—information that explicitly rejects racial inequalities as justifiable or random—could cause white people to become more supportive of policies that aim to redress these inequalities. While individual-level explanations of racial disparities do not create tension with the egalitarian ideals of the American Dream, structural understandings help expose the gap between these ideals and the reality of entrenched inequalities. If people are conscious of the ways racial disparities result from a fundamentally unequal playing field, then racial inequalities directly impugn the core myth of the American Dream. The resulting conflict between ideals and reality may prompt greater acceptance of policy efforts that aim to combat racial inequalities. Further intuition supporting this possibility comes from psychological theories suggesting that framing racism as individual, rather than emphasizing structural processes, has contributed to misunderstanding and curbed support for equitable ideas and policies (Payne et al., Reference Payne, Vuletich and Lundberg2017; Rucker and Richeson, Reference Rucker and Richeson2021). Therefore, promoting structural understandings of racial inequality may have the capacity to motivate support for policies designed to level the playing field.

While the underlying logic for this process is in place and, as discussed above, observational work has examined the effect of the Black Lives Matter movement specifically, experimental support remains sparse. In one experimental study, Adams et al. (Reference Adams2008) found that providing white college students intensive classroom instruction about structural racism motivated support for more egalitarian policies. In a recent piece, Fang and White (Reference Fang and White2022) showed that exposing white respondents to information about historical racial inequality can prompt a greater belief in structural Black-white inequality.

Thus, although the bulk of empirical evidence concerning discussions of racial disparities supports the racialization hypothesis, we have some findings to suggest that framing racial inequality as structural rather than individual has the potential to intervene in the process of racialization and motivate support for race-based policies. Under this view, white people may be willing to support race-based policies when the structural roots of racial inequality are emphasized. Such an intervention may disrupt typical processes of racialization and motivate support for racially egalitarian policies by emphasizing how America is not living up to the ideals of equality embedded in the myth of the American Dream—prompting recognition that policy interventions may be necessary to make this dream accessible for everyone.

Of course, there is heterogeneity in the extent to which white Americans conceive of racism as structural. As racial attitudes increasingly polarize (Engelhardt, Reference Engelhardt2021) and left-leaning social movements emphasize systemic racism (Sawyer and Gampa, Reference Sawyer and Gampa2018), white Democrats may be more likely to recognize the structural facets of racism than white Republicans. Segregated media environments that silo information about racism likely limit white Republicans’ exposure to evidence of structural racism and make them more inclined to view racial inequalities through an individualistic lens (Zell and Lesick, Reference Zell and Lesick2021). These patterns suggest that encountering information that calls attention to the systemic nature of racial inequality may vary across partisan groups with particularly pronounced effects among white Republicans.

Data and Methods

To evaluate the above expectations, we turn to a survey experiment fielded with the firm Lucid in the summer of 2021. ^{Footnote 2} The survey itself was funded and designed to test another research question, ^{Footnote 3} which is the subject of a different manuscript. We decided to include race-based policy-dependent variables on the survey as well—all of which we analyze here. ^{Footnote 4} In order to meet the stipulations of the granting agency that funded the survey, we had to preregister the project early on—and thus, the preregistration did not include the secondary dependent variables that were added later and are the subject of this manuscript. As this survey experiment was designed for another purpose, we encourage readers to interpret the analyses presented here as largely exploratory. In that spirit, the theory presented above outlined competing expectations that could explain either positive or negative effects of information about racial inequality on support for race-based policies. Despite these considerations, we believe that this survey experiment provides us with a unique opportunity to start to unpack how structural understandings of racism may affect whites’ attitudes about race-based public policies.

The experimental treatments exposed participants to news stories emphasizing the structural origins of racial inequalities that emerged during the pandemic. We then asked respondents to reflect on those stories and report their support for a range of race-based policies. The experiment included three versions of the treatment conditions as well as a control group. These four conditions are summarized below and fully reported in the appendix.

In total, 4,113 white respondents completed the survey. ^{Footnote 5} Respondents were recruited to match national distributions on age, gender, education, and region. Descriptive statistics are in the appendix. Respondents first answered a series of pre-treatment questions about demographics, ideology, and racial attitudes. Then they completed two attention checks, but they were not terminated if they failed. ^{Footnote 6} After the attention checks, respondents were randomly assigned to read one of four news stories, representing three versions of the treatment condition and the control. These stories were fabricated, but they mirror real news articles and report accurate information.

The control condition reported on the pandemic and the roll-out of the COVID-19 vaccine without mention of racial disparities. The treatment conditions mirrored the control but added information about racial disparities—that Black communities were hit particularly hard by the pandemic and that they were less likely to have access to vaccines. Respondents were explicitly informed that this fact was not a result of individual choices or failings, but instead that disparities were related to other forms of structural oppression faced by the Black community. Respondents were told that Black Americans are “just as willing to get the vaccine as whites” and that difficulties “getting time off work, finding transportation to distant clinic sites, and barriers to using online scheduling portals” were producing the racial differences.

Further, respondents in the treatment conditions read that these disparities were not particular to COVID, but instead rooted in structural inequalities that disadvantage Black communities. The treatment informed respondents that: “These experiences of racial gaps in health outcomes are not unique to the coronavirus pandemic. Instead, they are the result of longstanding inequalities in American society. Blacks are far less likely to have access to health care than whites, and they tend to receive worse care and die earlier from the same diseases. These kinds of inequalities put black Americans at increased risk of illness and contribute to their disproportionate deaths from COVID-19.” As such, treated respondents not only read about racial disparities in the pandemic but also about how these gaps were the result of structural inequality. Because the treatment text was designed to emphasize how these systemic inequalities contributed to unequal life chances for Black Americans, it directly contradicts the meritocratic myth of the American Dream and has the potential to undermine white Americans’ use of individualistic narratives to explain away these disparities.

The first treatment condition (the inequality condition) stopped with this information about the structural origins of racial inequalities. The second and third treatment conditions built on the first, but added further information about policy responses to the pandemic. The second treatment (the equity treatment) talked about a fabricated policy designed to ameliorate racial gaps in the pandemic through a program that aimed to promote greater equity in vaccine access. The third treatment condition (the reproducing inequality treatment) mirrored the equity treatment, but instead of reporting on a policy designed to increase vaccine equity, it reported on a policy that would increase access to the vaccine but not in an equitable way, likely leading to deeper racial disparities in access. These additional treatments were designed primarily to test hypotheses pertaining to other dependent variables which we analyze in another paper, and the first inequality treatment condition provides the cleanest test of the competing hypotheses that we lay out above—as it is solely focused on reporting about structural inequality without additional policy information. We nevertheless present the results for the second and third treatment conditions separately, both for reasons of transparency and because it is possible that these conditions may produce similar directional effects as the first treatment, but with different magnitudes. To be precise, the second treatment condition, which emphasizes a government policy that will likely be successful in redressing some of the inequalities discussed, may produce stronger effects—as this treatment not only reports on structural inequality but also on government policies that may successfully address such inequalities. Whether the racialization or structural hypothesis are at work, this should be the strongest treatment. The third treatment condition reports on structural inequality, but also mentions a failed government policy that will probably make inequality worse. As such, this third condition may result in the weakest effects on policy attitudes, as the government is painted as incapable of redressing inequality.

After reading the news story, respondents reflected on what they read by writing two to five sentences about the feelings or thoughts the article prompted for them. In our survey, respondents engaged meaningfully, writing surprisingly frank reflections based on the information they had read as well as their own experiences with the pandemic. Nearly 80% of respondents produced meaningful responses to this prompt and almost everyone wrote something—an especially high rate considering that we did not force a response to this (or any) item on the survey and that we did not exclude respondents based on pre-treatment attention checks (Condon and Wichowsky, Reference Condon and Wichowsky2020). After presenting the quantitative results, we draw on these responses as a source of qualitative insights into participants’ reactions to the information they encountered.

Following this reflection, respondents were asked to report a variety of social and political attitudes, including their support for race-based public policies. The first two race-policy-dependent variables are directly related to the specific policy domain discussed in the news story, while the next three are more general race-based policy items. The first question asked respondents the extent to which they agreed that: “The government should implement policies that reduce racial inequalities in health care in our country” (6-point scale). The second asked respondents: “Some have suggested that the government should make more efforts to guarantee vaccine access for black people. Do you approve or disapprove of this policy proposal?” Respondents reported their approval or disapproval on a 4-point scale. Next, participants answered three items drawn from the Cooperative Election Study (CES) asking whether they favored or opposed three policies designed to “deal with the problems of poverty and unemployment among black Americans.” The policies were as follows: (1) government giving business and industry special tax breaks for location in black neighborhoods; (2) spending more money on black schools; and (3) providing scholarships for black students who maintain good grades. ^{Footnote 7} Respondents indicated the extent to which they favored or opposed each of these policies on 4-point scales. The purpose of including these more general race-based public policies was to tease out whether effects of learning about racial inequalities in one policy domain would spill over and affect attitudes in others. ^{Footnote 8} While none of these policy items would fully address structural racial inequality, just as no single policy produced it, these items tap support for race-based public policies that are often discussed as potential strategies for addressing racial inequality and in ways that are more direct than universal programs.

Information about Structural Racism Increases Support for Race-based Policies

To test the competing theoretical expectations laid out above, we regress this series of dependent variables on indicators for each experimental condition. ^{Footnote 9} Table 1 reports the results. These analyses use only respondents who passed at least one attention check (about 70% of the sample), but the findings are robust to analyzing the full sample (reported in the appendix). We also note that balance tests confirm that randomization was successful and manipulation checks demonstrate that respondents read and understood the information from the treatment conditions (results in the appendix).

Table 1. Effects of reading about structural inequality on support for race-based policies

Note: ${{\rm{\;}}^{{\rm{***}}}}p$ < .001, ${{\rm{\;}}^{{\rm{**}}}}p$ < .01, ${{\rm{\;}}^{\rm{*}}}p$ < .05. Dependent variables are coded so that higher values indicate more support for race-based policies. Results from OLS regressions predicting support for the dependent variable. Includes white respondents who passed at least one attention check.

Table 1 demonstrates that reading a story about racial inequalities during the pandemic—specifically one describing these disparities as rooted in structural disadvantages—increases whites’ support for race-based policies designed to address racial inequality. Respondents in all three treatment conditions encountered information about the structural origins of racial inequalities in the pandemic, and they were more likely to favor government efforts to address racial disparities than those in the control. Further, the results in the last three columns demonstrate that these effects are not confined to policies that relate specifically to the pandemic or healthcare. Rather, we see spillover to other policy domains.

Across the three treatment conditions and five dependent variables, fifteen relevant coefficients were estimated, ten of which are statistically significant at or above the 95% confidence level. Respondents who read either of the first two treatments emphasizing the structural origins of racial inequalities during the pandemic were more likely to support a variety of affirmative-action-like programs. They were not only more supportive of policies aiming to reduce health disparities and facilitate vaccine access for Black people but also of government tax breaks for businesses in Black neighborhoods and of spending more money on Black schools. Those in the first treatment condition were also more favorable toward providing scholarships to Black students, and those in the reproducing inequality position (T3) were more supportive of government tax breaks to businesses in Black neighborhoods.

Across every dependent variable, the reproducing inequality treatment (T3) moves respondents’ commitments to race-based policies the least. Although the effects of this treatment are not statistically distinguishable from those of the first and second treatments, only one coefficient achieves conventional levels of statistical significance, relative to the control condition. Given that this treatment reports on a new policy that does not aim to address racial disparities, but is instead colorblind in design, this version of the treatment essentially emphasizes policy solutions that follow a logic contrary to the race-based policies that are the subject of the dependent variables here. Thus, while respondents may be motivated to address racial inequality after reading this article, the focus on a colorblind policy solution may make them less likely to favor government policies as effective vehicles.

To illustrate the treatment effects, Figure 1 plots the predicted level of support for government efforts to reduce racial disparities in health outcomes (the first dependent variable) by experimental condition. This visualization drives home the points above: white respondents who read about the systemic roots of racial inequality were more supportive of government policies aimed at addressing these inequalities. As is common in experimental studies, the size of the effects is relatively small. Compared to the control condition, receiving information about structural inequality increases support for government efforts to reduce racial disparities in health outcomes by about .20–.30 points on a 6-point scale. While these are not huge effect sizes, the treatments were relatively mild compared to real-world encounters that individuals may have with this information, as well as compared to previous experimental studies that focused on stronger treatments like college courses. Similarly, as effects from survey experiments tend to fade over time, it is unlikely that such effects would persist over months or years. Nevertheless, the fact that a brief news article is able to increase support for race-based policies indicates a connection between encountering information about the structural nature of racial inequalities and support for public policies designed to redress these inequalities.

Figure 1. Support for government efforts to reduce racial disparities in health outcomes

Note: Based on the first model in Table 1 which predicts agreement that government should implement policies to reduce racial inequalities in health care. Includes only white respondents who passed at least one attention check.

These aggregate results appear to support the structural hypothesis, rather than the racialization hypothesis. When race is injected into the news article in a way that emphasizes the structural nature of racial inequalities, respondents become more supportive of government efforts to address inequality, not less supportive as the racialization hypothesis would expect.

But perhaps we have not yet given the racialization hypothesis its best shot. Recall that the racialization hypothesis also anticipates that bringing up race (as the three treatments do, compared to the control) would activate whites’ racial attitudes for use in judgment formation. If the racialization hypothesis is correct, we might find heterogeneity such that people with particularly hostile racial attitudes would become less supportive of race-based policies when exposed to the treatment. In other words, even if the overall treatment effects are positive, the effect may work in the opposite direction for people with strong anti-Black attitudes. To test this possibility, we interact the treatment indicators with the pre-treatment measure of racial resentment. Negative interactions between the treatment and racial resentment would provide evidence in line with the racialization hypothesis. We present the results from regressing the same five dependent variables on this interaction in Table 2.

Table 2. Effects of inequality on support for race-based policies, by racial resentment

Note: ${{\rm{\;}}^{{\rm{***}}}}p$ < .001, ${{\rm{\;}}^{{\rm{**}}}}p$ < .01, ${{\rm{\;}}^{\rm{*}}}p$ < .05. Dependent variables are coded so that higher values indicate more support for race-based policies. Results from OLS regressions predicting support for the dependent variable. Includes white respondents who passed at least one attention check.

Support for the racialization hypothesis in these models is sparse. The racialization hypothesis would anticipate that any treatment that brings up race, as all three do here, should activate racial attitudes in a way that encourages people to translate racist views into less support for race-based policies. But this is not what we observe. Of the fifteen relevant interaction coefficients in Table 2, only three are statistically significant, and these are signed in the opposite direction from what was anticipated by the racialization hypothesis. Neither the treatment effects alone nor these interaction effects with racial resentment align with the empirical patterns anticipated by the racialization hypothesis. It seems that encounters with information that specifically emphasizes the structural nature of racial inequalities operate differently than conventional expectations concerning racialization.

But we have more work to do before concluding that promoting systemic rather than individualistic understandings of racial disparities really works to advance support for race-based policies. To scrutinize the structural hypothesis further, we take two additional steps. First, we evaluate whether partisanship moderates the relationship between confronting information about structural inequality and support for race-based policies in the way the structural hypothesis expects. Second, we examine our qualitative data to consider whether people’s responses to the treatment evidence thought processes are in line with this hypothesis.

We turn now to evaluating the extent to which there are partisan differences in the effects of confronting information about structural racism on support for race-based policies. ^{Footnote 10} Recall that the structural hypothesis expects white Republicans, who have likely not encountered or adopted a structural understanding of racism, to be especially prone to this “new” information, as findings from Fang and White (Reference Fang and White2022) also suggest. The racialization hypothesis, on the other hand, would anticipate that white Republicans would become more opposed to race-based policies after any exposure to information cueing race.

Table 3 reports models that test these expectations by regressing the key dependent variables on an interaction between the treatment condition and an indicator for whether the respondent is a Republican (or leans Republican), a “pure” Independent, or a Democrat (or leans Democratic). The baseline party identification used in the models is Democratic.

Table 3. Effects of inequality on support for race-based policies, by party of the respondent

Note: ${{\rm{\;}}^{{\rm{***}}}}p$ < .001, ${{\rm{\;}}^{{\rm{**}}}}p$ < .01, ${{\rm{\;}}^{\rm{*}}}p$ < .05. Dependent variables are coded so that higher values indicate more support for race-based policies. Results from OLS regressions predicting support for the dependent variable. Includes white respondents who passed at least one attention check.

Here, we find statistically significant and positive coefficients on the interaction between the treatment condition and Republican partisan identification. This is the empirical pattern anticipated by the structural hypothesis, rather than that anticipated by the racialization hypothesis. In the first treatment condition, Republicans became more supportive of three of five race-based policies after hearing about structural inequality (compared to those in the control condition). Democrats did not significantly change their position on these dependent variables in response to the treatment. Compared to Republicans in the control condition, Republicans who read an article that highlights the consequences of structural racism in the pandemic are significantly more likely to support policies that would redress racial disparities in health care, prioritize access to vaccines for Black people, and increase funding for Black schools. Independents, on the other hand, were not moved by this first treatment condition.

The second treatment condition, the equity condition, inspired even greater effects. For Republicans, these effects were slightly larger and were statistically significant for an additional variable—support for government tax breaks to businesses in Black neighborhoods—which makes sense, given Republicans’ general enthusiasm for tax breaks (but note that the positive treatment effect still demonstrates that the treatment boosted this support further among Republicans). Additionally, this second treatment condition inspires support among Independents. Independents who are exposed to the article about structural racism and successful government attempts to confront aspects of this inequality with a race-based policy results in greater support for government efforts to address racial disparities in healthcare (the first dependent variable) and prioritizing vaccine distribution to Black Americans (the second dependent variable).

Statistically significant partisan differences only appear on two of the five dependent variables for the third treatment condition among Republicans—and on none of the dependent variables for pure Independents. Recall that this condition discusses inequality but also reports on a fabricated policy that is likely to reproduce inequality. Perhaps because it reports on a governmental policy that will likely make racial inequality worse, respondents in this treatment condition are turned off from support for race-based policies designed to address racial inequality.

But in all of these models, where there is movement in response to the treatments, it is Republicans and sometimes Independents who are shifting their attitudes—not Democrats. Of course, it may be that white Democrats shift less than others because they already have very high levels of support for race-based policies, even without encountering the treatment. To investigate this possibility of ceiling effects for Democrats, we plot the predicted level of support for government efforts to reduce racial disparities in health outcomes by partisanship and treatment condition in Figure 2. This dependent variable ranges from 1 to 6. As is clear, Democrats in the control condition are quite high on this variable, averaging about 5.25. This persists for all of the conditions—indicating that there are ceiling effects for Democrats.

Figure 2. Support for government efforts to reduce racial inequalities in healthcare, by party

Note: These predicted values correspond to the first model in Table 3 that predicts white respondents’ agreement that the government should implement policies that reduce racial inequalities in health care.

On the other hand, Republicans and Independents in the control condition are more lukewarm toward race-based policies that would reduce healthcare disparities, only averaging around 3.5 (for Republicans) and a 4 (for Independents) on the 6-point scale. Republican and Independent support for this policy increases by about .5 points on this 6-point scale in the first and second treatment conditions—with the largest jump in the second treatment condition. Even the third condition inspires an increase in support for this race-based policy among Republicans, though to a lesser degree. For Republicans and Independents, the information about the structural roots of racial disparities in the pandemic appears to disrupt their baseline understandings about the nature and causes of racial inequalities, prompting greater openness to policies that combat these inequalities.

The analysis thus far provides consistent support for the structural hypothesis. First, across the full sample of white respondents, information about structural racism inspires an increase, not a decrease, in support for race-based policies. Moreover, these positive effects are especially pronounced among Republicans, as the structural hypothesis anticipates. Next, we turn to respondents’ qualitative responses to explore whether the logic behind the structural hypothesis—that situating racial inequality as the result of systemic as opposed to individual failings is received as “new” and “surprising” to some (even if this supposed newness results from willful ignorance (Kraus et al., Reference Kraus2019; Mills, Reference Mills1997)) and that this information about the structural nature of inequality is in conflict with underlying egalitarian ideals.