Associations Between Race, Representation, and Issue Areas

Scholars have long investigated the impact of racial prejudice in shaping political behavior at the mass and elite levels (for a review, see Cramer Reference Cramer2020; Jardina and Piston, Reference Jardina and Piston2019; Stephens-Dougan Reference Stephens-Dougan2021). Many studies have found consistent evidence of the deleterious impact of racial prejudice on the status and representation of African Americans. Whites’ negative racial views predict their preferences on policies purported to benefit African Americans (Gilens Reference Gilens1999; Peffley and Hurwitz Reference Peffley and Hurwitz2002; Tesler Reference Tesler2012; Wallsten et al. Reference Wallsten2017). In explaining the relationship between racial attitudes and policy preferences, scholars argue that 1) the mass public has established a cognitive link between the policy issue area in question and the Black community, 2) this link primes their racial attitudes, and 3) these “top of the head” considerations are brought to bear when forming their policy preferences and candidate evaluations (Zaller Reference Zaller1992).

These presumed race-policy associations are also marshaled by political candidates, namely Republican candidates, to mobilize electoral support from racially conservative White voters (for review, see Stephens-Dougan Reference Stephens-Dougan2021). Candidates elicit electoral support by either portraying themselves as opponents of policies that are associated with African Americans or by painting their opponent as a champion of these policies (Banks and Bell Reference Banks and Bell2013; Mendelberg Reference Mendelberg2001; Peffley and Hurwitz Reference Peffley and Hurwitz2002; Valentino, Hutchings, and White Reference Valentino, Hutchings and White2002). According to theories of racial priming, due to the change in racial norms, candidates can no longer be explicit about their antipathy toward African Americans and expect to emerge victorious on Election Day. Rather, in seeking to mobilize racially conservative White voters, these candidates instead rely on the use of implicit racial appeals and rest on the widespread perceived associations between particular public policies and the Black community (Haney-Lopez Reference Haney-López2015; Mendelberg Reference Mendelberg2001).

Given the growth in the number of African American representatives, a large body of scholarship also focuses on how racial descriptive representation influences substantive representation. Studies conclude that while the presence of African Americans in legislatures does not directly always result in the passage of legislation on civil rights, criminal justice reform, or social welfare, descriptive representation does yield greater general attention to these issues (Canon Reference Canon1999; Tate Reference Tate2003; Griffin Reference Griffin2014; Griffin and Newman Reference Griffin and Newman2008; Grose Reference Grose2011). At the same time, how politicians claim different issue priorities can influence whether they are perceived to represent racial minorities. For example, research on the presidency shows that Barack Obama, in order to undermine the perception that he would disproportionately cater to the interests of African Americans, failed to substantively represent the interests of the Black community during his tenure in the White House (Harlow Reference Harlow2009; Smith, King, and Klinkner Reference Smith, King and Klinkner2011; Harris Reference Harris2014; Stephens-Dougan Reference Stephens-Dougan2022; but see Nteta, Rhodes, and Tarsi Reference Nteta, Rhodes and Tarsi2016; Haines, Mendelberg, and Butler Reference Haines, Mendelberg and Butler2019).

Not all Americans may have unified perceptions of race, representation, and issue positions. White Americans are more supportive of punitive policies when they believe that African Americans are the target of such policies (Gilens Reference Gilens1999; Peffley and Hurwitz Reference Peffley and Hurwitz2002; Wallsten et al. Reference Wallsten2017). When they view policies as being beneficial to African Americans, they are less supportive (Haney-Lopez Reference Haney-López2015; Kinder and Sanders Reference Kinder and Sanders1996). If White and Black Americans have similar perceptions of the target populations of candidates’ issue priorities, then political strategies based on racial appeals might have less polarizing effects, potentially leading to a more unified public response to policy initiatives.

Schneider and Ingram’s theory (Reference Schneider and Ingram1993) on the social construction of “target populations” argues that policymakers often craft policies that favor groups perceived positively within society, typically those with more political power, while imposing stricter policies on negatively viewed, less powerful groups (Kreitzer and Smith Reference Kreitzer and Smith2018). This perspective is key in understanding how societal stereotypes and the political power of different groups shape policy agendas. It also informs the broader discourse on how political candidates’ issue priorities might align with these societal constructions, influencing their perceived effectiveness in representing diverse racial groups. In this context, it is particularly significant to understand how issue priorities set forth by politicians are perceived in terms of representing African Americans and other racial groups. The way candidates prioritize certain policies can be seen as a reflection of their commitment to addressing the concerns of these groups.

Taken together, the through line connecting these bodies of scholarship is the importance of how public policies and issue areas are perceived to be connected to the representation of African Americans. These perceptions are thought to be indirectly responsible for low levels of mass support for policies that might benefit African Americans and, thus, racial inequalities in substantive representation. Yet, missing from each account is systematic empirical evidence about racialized perceptions of politicians’ issue priorities. It is not clear who the public considers to be the “relevant” target populations (Schneider and Ingram, Reference Schneider and Ingram1993). A candidate prioritizing criminal justice reform might be perceived as advocating for African Americans, viewed as a disadvantaged group in this policy domain. Conversely, emphasizing policies like climate change or education might not evoke the same association. Existing research relies on top-down assessments of such cognitive connections, leading to questions concerning the reliability and validity of these assumptions. The implications of understanding these perceptions are profound. If these perceptions drive support for policies that might benefit African Americans, then reducing racial inequalities in representation is dependent on the racialized “pictures in our heads” about issue areas (Lippman Reference Lippmann1922).

Race-Policy Associations by Race and Partisanship

While our results suggest that the mass public is more likely to identify certain issues, most notably criminal justice and poverty, with African Americans, given the ubiquity and strength of both partisan and racial divides in public opinion, it remains possible that these connections differ among partisan and racial groups. Differences in these associations could potentially complicate assumptions about the racialized nature of issue areas and, consequently, the strategic behavior of politicians who focus on certain policies to appeal to particular groups. For instance, if White individuals believe that a candidate emphasizing criminal justice is best suited to represent African Americans, but Black respondents do not, it could imply that using this issue to appeal to African Americans would be costly for politicians without any real gains.

To better explore differences by race, we oversampled African American respondents to uncover if they identify a distinct set of issues relative to White respondents. Second, we are also interested in heterogeneity by partisanship. The parties have sorted along racial lines (see Mason and Wronski Reference Mason and Wronski2018), and research on party issue ownership has demonstrated that certain policies are perceived as being closely aligned with one party or the other (Goggin and Theodoridis Reference Goggin and Theodoridis2017). Given differences in partisan affiliation and issue ownership, it is important to know whether Democrats and Republicans have different views on which issues influence Black individuals.

Figure 2 shows the predicted proportion that selected African Americans as being represented best for each issue priority compared to any other racial group. Predictions were estimated while holding constant the respondent’s race, education level, party affiliation, ideology, and their perceptions of the candidate’s race, party affiliation, and ideology (at the modal levels). People were more likely to select their own racial group as being impacted by each issue area. That is, African American respondents were significantly more likely to select African Americans as being represented best compared to White respondents (and the same pattern was found for White respondents; see Appendix E), but overall, the order of issues perceived to impact African Americans is consistent regardless of the race of the respondent.

Figure 2. Predicted proportion selecting African Americans by partisanship and race

Note: Conditional adjusted predictions from cluster-robust models regressing whether the respondent selected African Americans as being represented best (1) or another group (0) for each issue priority. Models include controls for race, education, party affiliation, ideology, and perceptions of the candidate’s race, party affiliation, and ideology. Left: results by party ID, including leaners. True independents and people who answered Not Sure for party affiliation are excluded. Right: results by race of respondent. Respondents who answered Asian, Middle Eastern, Native American, or Two or more races are excluded. 95% confidence intervals.

Republicans were slightly more likely to select African Americans for all issue areas than Democrats, though these differences are not statistically significant. Again, a similar pattern as the full sample is found across subgroups. Democrats, Republicans, White respondents, and Black respondents alike were likely to say that African Americans would be best represented by candidates who focused on criminal justice, poverty, and crime, and least represented by candidates who focused on climate change, immigration, and the military. The results for the representation of White Americans are shown in the Appendix, and again, few differences are found between White and Black respondents and Democrats and Republicans for the ranked importance of issues. The unanimity in responses across both partisan and racial groups demonstrates how robust and widespread these evaluations are. While all of these models control for the respondent’s level of education, we also replicate these analyses split by the most and least educated (see Appendix D) to more clearly account for the possibility that these perceptions are only shared by those who may have relatively sophisticated understandings of politics. The results do not markedly differ by level of education, with the same issues being associated with African Americans and White Americans.

Linking Issue Priorities to Candidate Partisanship, Ideology, and Race

What kinds of inferences do voters draw about a candidate based on their stated issue priority? Each issue may elicit different assumptions about the candidate, which may shape respondents’ views on how well a candidate would represent various racial groups. For this reason, the analyses above control for respondents’ answers to the other outcome questions about the candidates, which include perceptions of the candidate’s race, partisanship, and ideology. This helps isolate the direct effects of the issue priority on perceptions of African American representation. By accounting for these other perceptions, we can more accurately assess how specific issue priorities shape voters’ views on racial representation without confounding influences from their other beliefs about the candidate.

We note that these four key outcomes—whether respondents select African American as being best represented, the candidate’s perceived ideology (−1 conservative, 0 moderate, 1 liberal), the candidate’s perceived partisanship (−1 Republican, 0 Independent, 1 Democrat), and the candidate’s perceived race (1 African American, 0 Other race)—are all positively and significantly correlated (p < 0.001) as shown in Table 2. The strongest relationship is between perceptions that the candidate is Black and the belief that the candidate best represents African Americans (r = 0.69). Perceptions that the candidate is a Democrat (r = 0.29) or is liberal (r = 0.23) are also positively associated with perceptions of Black representation, though these associations are notably smaller. Perceptions of candidate partisanship and ideology are unsurprisingly strongly correlated (r = 0.64), indicating that these judgments often travel together. However, both are only moderately correlated with the belief that the candidate is Black, further reinforcing the distinctiveness of racial perception in shaping judgments about representation.

Table 2. Correlations between outcome variables

Note: Entries are pairwise Pearson correlation coefficients. *p < 0.001

To better understand how issue priorities influence these other perceptions, we also estimate separate regressions where the dependent variable is each of the three candidate traits: perceived ideology, perceived party, and perceived race. These models assess how respondents’ inferences about the candidate’s identity are shaped by the issue the candidate prioritizes. The results, presented in Appendix F, show that many issue areas increase the likelihood that a candidate is perceived as liberal and as a Democrat, particularly climate change, poverty, health care, education, and criminal justice. Several issue areas increase the likelihood that the candidate is perceived as Black, including drug addiction, poverty, education, and health care. However, criminal justice produces the largest shift in this perception, with a coefficient more than twice the size of most other issue effects. While the perception that a candidate is Black is not exclusive to criminal justice, this issue appears to serve as a particularly strong racial cue relative to others. This helps explain why criminal justice stands out earlier as most strongly associated with perceived African American representation. Respondents may be drawing conclusions about the candidate’s background based on the issue alone. This likely contributes to the observed effects on perceived Black representation.

Since many of the issues tested in our survey may imply certain ideological leanings (e.g., “addressing the criminal justice system”), it is especially important to understand the potential mediating role of ideology in particular. We conducted a mediation analysis to investigate whether candidate ideology mediates the effect of the selected issue area on perceptions of African American representation. Specifically, does the perceived liberalism of candidates alter the relationship between prioritizing criminal justice and respondents’ perceptions of group representation? The rationale here is that voters may interpret candidates’ stances on critical issues through an ideological lens. Given that criminal justice is the issue most closely associated with Black representation, it serves as an ideal case to use as the treatment, ensuring that our findings above are not solely driven by ideological perceptions. When an issue like criminal justice is listed as the candidate’s priority, voters may perceive the candidate as more liberal. This perception may lead voters to associate those candidates with progressive policies to better serve the interests of Black communities. Conversely, candidates who do not prioritize this issue may be seen as more conservative and, therefore, less favorable in the context of representing Black interests.

The results of the mediation analysis show that while the direct effect of criminal justice on the outcome (perception that the candidate best represents African Americans) is statistically significant, the indirect effect through the mediator (perception that the candidate is liberal) is not. The Average Causal Mediation Effect (ACME) estimates the indirect effect at 0.0036 (p = 0.37). In contrast, the Average Direct Effect (ADE), which captures the direct impact of criminal justice on African American representation, is 0.31 (p < 0.001). The proportion of the total effect mediated through perceived ideology is very small (1.15%) and not statistically significant. This indicates that the direct effect of criminal justice significantly increases the likelihood of selecting African Americans as the best-represented group, independent of the perceived liberalism of the candidate.

Race-Policy Associations in Congressional Communication

These results so far reveal how people associate certain policy areas with specific racial groups. The question of whether and how such associations are reflected in elite political messaging remains open. To explore this, we turn to a different form of political communication: congressional newsletters. These messages offer a complementary perspective by examining how members of Congress themselves choose to frame policy issues in routine, constituent-facing communications. This next analysis is exploratory, but it helps assess the extent to which legislators explicitly racialize policy in ways that mirror or diverge from public perceptions.

This analysis draws on 196,422 congressional newsletters collected from the DCInbox project (Cormack Reference Cormack2017), which archives official communications from members of Congress to their constituents. The newsletters span the 111th to the 119th Congress (approximately 2009 through March 2025) and include messages from 74,662 Democrats, 121,746 Republicans, and 14 Independents. Newsletters provide a uniquely valuable window into how members of Congress communicate with constituents outside of campaigns or public floor debates (see Cormack Reference Cormack2016; Blum, Cormack, and Shoub Reference Blum, Cormack and Shoub2023). Unlike media appearances or roll call votes, newsletters reflect strategic decisions by legislators about what issues to highlight and how to frame them for their home audiences. Because newsletters are often written with the intention of demonstrating responsiveness, they are especially well-suited for studying patterns in representation, including how racial identity is invoked in the context of public policy (Cormack and Gulati Reference Cormack and Gulati2024).

The goal of this analysis is to examine how frequently legislators reference African Americans when discussing policy issues in their constituent communications. Each newsletter was broken into overlapping 5-sentence windows. Within each window, the analysis flags a co-occurrence when both a policy term and a racial group term appear. This method captures cases where race is used to frame a policy discussion—for instance, when a legislator discusses the criminal justice system in the context of Black communities. Policy terms were identified using a dictionary of keywords related to nine broad issue areas that mimic the issue priorities outlined above: the economy, health care, education, poverty, criminal justice, the budget deficit, immigration, climate change, and the military. Racial group references were identified using a dictionary of common phrases referring to Black Americans, including both general (e.g., “Black people,” “Black Americans”) and more specific descriptors (e.g., “Black residents,” “Black constituents,” “Black voters,” “Black families”). Since the survey results focus on both Black and White associations, we repeated this analysis to identify policy discussions in reference to both groups.

For each pairing, we calculate the percentage of policy mentions that also reference Blacks or Whites as shown in the left-hand plot in Figure 3. Overall, such co-occurrences are rare: the share of policy mentions that include any reference to Black or White Americans within a five-sentence window is typically under 2 percent. The overwhelming majority of racial references are to Black Americans; references to White Americans are exceedingly rare across all issue areas. This aligns with expectations, given the normative status of whiteness as the unmarked “default” and implicit racial messaging found in campaigns given norms of racial equality (Jardina Reference Jardina2019; Mendelberg Reference Mendelberg2001).

Figure 3. Percent of policy mentions referencing race in Congressional Newsletters

Note: Percentage of newsletters that mention a given policy area and also reference Black or White Americans within a five-sentence window (2009–2025, n = 196,422). The left-hand plot shows results for Black and White references across parties, while the right-hand plot disaggregates results for Black references by party.

Criminal justice stands out as the policy area most often discussed in connection with Black Americans. Across all newsletters, 1.7% of criminal justice mentions reference Black Americans (compared to just 0.08% that reference White Americans). Many of these mentions occurred during the 116^th Congress after the killing of George Floyd and in reference to the Black Lives Matter movement. Poverty follows, with 1.2% of messages referencing Black Americans. Education is also among the top policy areas linked to discussions of Black Americans. While the overall rate of such references is lower than for criminal justice or poverty, it remains notably higher than for many other issues. A closer examination of the education-related messages that mention Black Americans reveals that these references often appear in the context of symbolic recognition, such as Black History Month, or policy debates surrounding affirmative action in college admissions. Discussions of the military and the budget deficit are at the bottom of the list, each only barely ever mentioned in connection with Black Americans.

Breaking the results down by party reveals stark differences in how these topics are communicated. Democrats are far more likely to reference Black Americans across every policy area. For Democrats, 4.4% of all mentions of criminal justice also reference Black Americans. That figure drops to just 0.6% among Republicans. These partisan gaps remain for the remaining issues as well, with more Democrats referencing Black Americans when they discuss policy terms than Republicans.

In some ways, these results closely mirror our earlier findings regarding the bottom-up associations between issue areas and the representation of African Americans. In that analysis, we found that the public most often linked criminal justice and poverty to African Americans and least often linked the budget deficit and the military. Congressional communication reflects that same structure: criminal justice and poverty are the most likely to be discussed in connection with Black Americans, while the budget and military are almost never discussed this way. Unlike the survey results, poverty is actually the top issue for references to White Americans, but the overall number remains extremely small at just a handful of mentions out of tens of thousands of policy messages. These patterns underscore the value of studying public perceptions. Even though race is often not explicitly mentioned, people draw conclusions about who issues are for.

Limitations

Of course, our study is not without its limitations. First, our experimental design, while modeled off of a major nationally representative survey about issue priorities in the United States, introduces implied policy positions alongside some of the issue areas (e.g., “strengthening the military”) that are not perfectly equivalent. While this resembles campaigns that highlight both generic policy areas and the direction of candidates’ positions, future research should aim to clearly differentiate between issue areas, specific policy stances, and partisan affiliation.

To be sure, we find that the same issues that are most strongly associated with African American representation—such as criminal justice, poverty, and education—are also associated with perceiving the candidate as a Democrat, as liberal, and, in some cases, as Black. This overlap reflects the broader social meanings these issues carry and the interrelated ways that voters use issue signals to infer a candidate’s identity and agenda. Our goal in this project is not to disentangle these associations entirely, but rather to understand how people interpret issue priorities in the context of race and representation. The mediation analysis helps address one key component of this overlap: the possibility that the survey items themselves imply an ideological direction. Further work is needed to disentangle these concepts more explicitly, and we cannot say how our results would be different if we cued partisanship, for example, in the experimental vignette. Our minimalist design may amplify the observed relationships since the absence of additional candidate information (such as partisanship) likely increases the weight respondents place on issue priority. Moreover, as we are interested in the implications for representation, the design focuses on the top policy priority of a candidate, which differs from examining broad race-policy associations in general.

A third limitation is that the primary dependent variable, which asks respondents to select a single racial group that is represented best, may depend on respondents’ understandings of group policy priorities, which could vary widely. We were interested in forcing a selection of one group among the set to better answer our intended question: whether a specific issue focus is uniquely associated with the representation of a particular racial group. In this way, respondents have to rely on their pre-existing, stereotypical race-policy associations. By asking whether a candidate would or would not represent each group separately or by allowing respondents to select multiple groups, we would not force these trade-offs to uncover whether there is a singular and unique connection to a specific racial group. This would also encourage respondent anchoring and socially desirable responding. That said, the forced choice between racial groups might lead to the default to the most numerous group (White Americans), potentially obscuring associations made in a different context. And the measure may not capture more inclusive or multidimensional views of representation. Future studies would benefit from a measurement strategy that allows a clearer association between issues and racial groups without the constraint of a forced choice.

The newsletter analysis is limited in its scope as well. We show the textual co-occurrence of race and policy terms within a five-sentence window, which may miss more subtle or implied forms of racial framing that do not include explicit group references. We examine explicit mentions of Black or White Americans and therefore cannot capture coded language, tone, or imagery that may still carry racial meaning. The limited dictionary to find key issue terms may also fail to capture evolving or context-specific references to both race and policy areas. We also cannot say how legislators or candidates make race-policy references in other mediums or why they may do so. Rather, this analysis is one way to contextualize our survey findings by illustrating what kinds of race-policy associations politicians actually make.