I am … a dark-skinned person. … I’ve grown up in communities that at one time … were 100% African American. … I feel like I’ve had … some of the traditional … urban city African American experience.
—An African American state legislator of Hispanic descent, quoted in Lemi (Reference Lemi2018, 735)
Political scientists have studied the relationship between being a racial minority elected to office—a descriptive representative—and substantively advancing the interests of one’s racial group through policy—substantive representation—for decades (Dovi Reference Dovi2002; Gamble Reference Gamble2007; Reference Gamble2011; Mansbridge Reference Mansbridge1999; Minta Reference Minta2011; Tate Reference Tate1994; Reference Tate2003). Yet representatives of color do not experience racialization in the same way. Colorism, a system of oppression rooted in white supremacy that favors lighter skin (Hunter Reference Hunter2007), structures the life experiences of people of color globally. As the vignette in the epigraph above shows, darker-skinned representatives, in this case Black representatives, may believe that they have a unique set of life experiences that lighter-skinned representatives may lack. In fact, skin tone defines who is perceived as a descriptive representative (Monk Reference Monk2022), who enters the pool of candidates for office (Brown and Lemi Reference Brown and Lemi2021), who wins (Brown and Lemi Reference Brown and Lemi2021; Burge, Wamble, and Cuomo Reference Burge, Wamble and Cuomo2020; Weaver Reference Weaver2012), and who may use their interpersonal relationships in the legislature to pass their bills (Brown Reference Brown2014; Lemi Reference Lemi2018). While skin tone has been associated with changes in political opinion and attitudes toward elected officials, little is known about the relationship between skin tone and substantive representation.
We argue that colorism structures the personal experiences and electoral incentives, and therefore the motivations, of Black members of Congress to explicitly advocate for Black Americans (Grose Reference Grose2011). In this study, we focus on two forms of outreach as outlined by Eulau and Karps (Reference Eulau and Karps1977). The first is policy responsiveness, which is measured through the primary sponsorship of legislation. The second is rhetorical representation, a form of symbolic responsiveness, which we measure through the analysis of Black-centered statements in press releases (Eulau and Karps Reference Eulau and Karps1977). Eulau and Karps (Reference Eulau and Karps1977, 241) define symbolic representation as “public gestures of a sort that create a sense of trust and support in the relationship between representative and represented.” Recent research has demonstrated that these symbolic actions have the power to build trust with constituents and lay the groundwork for substantive actions to be taken in the future by framing debates and defining solutions (Dietrich and Hayes Reference Dietrich and Hayes2023; Garcia, Stout, and Tate Reference Garcia, Stout and Tate2026). Together, both forms of representation shape the policy landscape for Black Americans.
We empirically assess the relationship between skin tone and substantive representation among Black US House representatives in the 114th, 115th, and 116th Congresses. While colorism affects multiple communities (Wilkinson and Earle Reference Wilkinson and Earle2013), we focus on African Americans for several reasons. First, Black voters, while not a monolith, tend to vote similarly with many shared interests, which makes it easier to define substantive representation (Tate Reference Tate2003; White and Laird Reference White and Laird2020). Second, the relatively large number of African Americans in the US House provides an opportunity to explore variation in skin tone and Black-centered substantive representation. Third, the nature of colorism has particular historical meaning for African Americans and the broader trajectory of Black political representation (Brown and Lemi Reference Brown and Lemi2021; Burge, Wamble, and Cuomo Reference Burge, Wamble and Cuomo2020; Davis Reference Davis1991; Hochschild and Weaver Reference Hochschild and Weaver2007).
We test our hypotheses with a novel dataset consisting of hand-coded and automated content analyses of 204,806 press releases and 18,025 bill introductions for the 114th, 115th, and 116th congressional sessions and a crowdsourced measure of skin tone. We find that having darker skin is positively associated with speaking out on issues relevant to African Americans in press releases, but not in legislation. We advance the scholarship on descriptive representation theoretically, conceptually, and empirically. First, theoretically, we show that fully estimating the relationship between descriptive representation and various other forms of representation requires taking into account how descriptive representatives look (e.g., Brown and Lemi Reference Brown and Lemi2021; Lemi Reference Lemi2018). Although prior research has considered how gender (e.g., Minta and Brown Reference Minta and Brown2014), district demographics (e.g., Canon Reference Canon1999), and institutional contexts (e.g., Peay Reference Peay2021; Preuhs Reference Preuhs2006; Whitby Reference Whitby2002) might produce different substantive outcomes within groups of descriptive representatives, less research has explicitly theorized how and why the very markers of racial group membership, like skin tone (Monk Reference Monk2021), are associated with different substantive outcomes. This research demonstrates that future research must take seriously how the very characteristics that define descriptive representatives (Mansbridge Reference Mansbridge1999, 628) may correlate with substantive outcomes.
Second, this article demonstrates the need to refine existing concepts of descriptive representatives. Although descriptive representation by race has predominantly focused on nonwhite group membership (Mansbridge Reference Mansbridge1999), there is growing evidence that scholarship must refine the components of what constitutes a descriptive representative (e.g., Lemi Reference Lemi2018; Reference Lemi2021; Reference Lemi2022). Separate from just “controlling” for appearances in theoretical accounts of descriptive representation, this article demonstrates the value of considering a descriptive representative’s degree of racialization via descriptive characteristics (Monk Reference Monk2022). We demonstrate that descriptive representatives differ in their behavior according to how they look.
We proceed in the following way. We first review the literature on Black political representation. We then theorize how historical and contemporary colorism structures the relationship between skin tone and representation for Black members of Congress. We then present our data and findings. We close with ideas for future research and an invitation to political scientists to integrate skin tone into their studies of descriptive and substantive representation.
Black Political Representation
A plethora of scholarship has shown the value of electing Black representatives (Bobo and Gilliam Reference Bobo and Gilliam1990; Broockman Reference Broockman2013; Mansbridge Reference Mansbridge1999; Stout, Tate, and Wilson Reference Stout, Tate and Wilson2021). From empowering Black constituents (Bobo and Gilliam Reference Bobo and Gilliam1990), to shaping deliberations (Gamble Reference Gamble2011), to providing important oversight (Minta Reference Minta2011), to sponsoring legislation (Canon Reference Canon1999; Gay Reference Gay2007), Black legislators enhance the political representation of Black Americans. Black legislators are also more likely to engage in nonlegislative activities like speaking publicly about racial issues in their public speeches (Gillion Reference Gillion2016), highlighting the work of civil rights leaders (Dietrich and Hayes Reference Dietrich and Hayes2023), and using social media to advance Black political movements (Bonilla and Tillery Reference Bonilla and Tillery2020).
Largely attributed to their life experiences as Black persons (Mansbridge Reference Mansbridge1999, 628), scholars have shown that Black legislators bring distinct perspectives and priorities into Congress. Through their lived experiences, Black legislators hold insights into issues that are prioritized and faced by Black communities (Brown Reference Brown2014; Minta Reference Minta2011). Black legislators bring their experiences to bear on fighting for Black communities in policy making (Grose Reference Grose2011; Minta Reference Minta2011; Tate Reference Tate2003). The experiences of Black Americans and Black legislators are not monolithic, and some research suggests that Black legislators are often keenly aware of and sympathetic to the experiences of Black Americans whose experiences differ (Garcia, Stout, and Tate Reference Garcia, Stout and Tate2026; Maeda and Hashimoto Reference Maeda and Hashimoto2020; Tajfel and Turner Reference Tajfel, Turner, Worchel and Austin1986). As a result, Black legislators may be intrinsically motivated to help resolve problems faced by members of the Black community even if they do not experience those problems themselves.
Yet the kind of representation Black legislators provide is not homogeneous. On the one hand, external factors shape their legislative behavior. Electoral circumstances, like the race of primary opponents (Canon Reference Canon1999) and district demographics (Gay Reference Gay2007; Grose Reference Grose2011; Casellas and Wallace Reference Casellas and Wallace2015), as well as their institutional position (Gamble Reference Gamble2011; Hawkesworth Reference Hawkesworth2003; Tyson Reference Tyson2016), have all been shown to influence the representation that Black representatives provide to Black Americans. On the other hand, the combination of certain ascriptive characteristics, such as race and gender, also significantly impacts how Black legislators behave within Congress.Footnote 1 For example, Black women legislators have been shown not only to advocate for distinct issues (Brown Reference Brown2014; Orey et al. Reference Orey, Smooth, Adams and Harris-Clark2006; Smooth Reference Smooth2011) but also to use distinct tactics to advance these priorities (Hawkesworth Reference Hawkesworth2003).
Another characteristic on which Black representatives vary is skin tone (Hochschild and Weaver Reference Hochschild and Weaver2007). In this study, we explore how skin tone, a marker of race and descriptive representation, is associated with the representation of Black interests. We argue that colorism shapes the experiences that members of Congress bring to office as well as the electoral incentives that Black members of Congress face to explicitly advocate for Black Americans.
Theory and Hypotheses
Whereas past research has focused on how personal experiences with racism shape Black representatives’ behavior, we argue that personal experiences with colorism as a Black person also shape how Black representatives advocate for Black communities. We theorize that the impact of colorism varies based on the type of representation a descriptive representative provides due to (1) divergence of lived experiences based on skin tone, and (2) divergence of electoral constraints based on skin tone. We use Monk’s (Reference Monk2022) infracategorical model of inequality and treat skin tone as a meaningful marker of difference among Black US representatives from the perspective of non-Black observers in predominantly non-Black institutions: Congress and the electoral system (see also Chirco and Buchanan Reference Chirco and Buchanan2022). We view skin tone as both a marker of descriptive representation that denotes membership of a racial group as well as a layer of differentiation that shapes the degree of one’s racialization as Black (Hannon et al. Reference Hannon, Keith, DeFina and Campbell2021; Hutchings et al. Reference Hutchings, Jefferson, Lewis and Yadon2016; Mansbridge Reference Mansbridge1999; Monk Reference Monk2022; Ostfeld and Yadon Reference Ostfeld and Yadon2022b; Yadon Reference Yadon2025).
Colorism is a form of discrimination rooted in white supremacy that favors lighter skin (Hunter Reference Hunter2007; Monk Reference Monk2021, 79; strmic-pawl, Gonlin, and Garner Reference strmic-pawl, Gonlin and Garner2021). Its origins in the US can be traced back to slavery. Laws, like Virginia’s 1662 statute, declared all children born to enslaved women to be enslaved themselves, tethering legal status to maternal lineage and over time to visible Blackness (Wright Reference Wright1995). Within this system, lighter-skinned enslaved people, often the children of white enslavers who had raped enslaved women (Dixon and Telles Reference Dixon and Telles2017, 407), received differential treatment. They were more likely to work in the household, have access to better nutrition (Bodenhorn Reference Bodenhorn2002), and have an increased likelihood of literacy and specialized training in a trade (Russell, Wilson, and Hall Reference Russell, Wilson and Hall1993). They were also more likely to be manumitted (Monk Reference Monk2015, 401), and, after the Civil War, were better positioned to convert those advantages into economic and social mobility (Bodenhorn Reference Bodenhorn2006; Davis Reference Davis1991; Russell, Wilson, and Hall Reference Russell, Wilson and Hall1993).
These hierarchies were institutionalized through law and social practice, establishing a durable stratification system in which skin tone became intrinsically connected to opportunity. For instance, in an effort to track the amount of “Black blood” in Black individuals, the US Census Bureau included categories such as “mulatto” to account for fractions of Black ancestry from 1850 to 1920, formally encoding the idea that lighter skin conferred a distinct societal position (Nobles Reference Nobles2000; Pratt, Hixson, and Jones, Reference Pratt, Hixson and Jonesn.d.). In fact, in the post-Reconstruction era, most Black representatives had mixed ancestry (Hochschild and Weaver Reference Hochschild and Weaver2007, 650).
Over time, the government’s adoption of the “one-drop” rule collapsed skin tone gradations into a single legal boundary between Black and white (Weaver Reference Weaver2012, 164). This consolidation strengthened racial solidarity as a basis for political organizing but did nothing to erase the skin tone hierarchy (Hochschild and Weaver Reference Hochschild and Weaver2007; Weaver Reference Weaver2012). The disparities in life conditions that colorism produced did not dissolve with legal reform. Rather, they compounded across generations, embedding skin tone advantage and disadvantage into the structures of wealth, education, and opportunity.Footnote 2
Colorism today is not a relic of the past, but a foundation for present-day inequality. Skin tone remains consequential across nearly every domain of Black American life. Darker-skinned Black Americans face worse health outcomes, lower educational attainment, lower wages, and a higher exposure to the punitive law system than their lighter-skinned counterparts (e.g., Blake et al. Reference Blake, Keith, Luo, Le and Salter2017; Crutchfield, Fisher, and Webb Reference Crutchfield, Fisher and Webb2017; Eberhardt et al. Reference Eberhardt, Davies, Purdie-Vaughns and Johnson2006; Hamler et al. Reference Hamler, Nguyen, Keith, Qin and Wang2022; Hargrove Reference Hargrove2019; Keith et al. Reference Keith, Nguyen, Taylor, Mouzon and Chatters2017; Keyes, Small, and Nikolova Reference Keyes, Small and Nikolova2020; King and Johnson Reference King and Johnson2016; Ryabov Reference Ryabov2013; Stockstill and Carson Reference Stockstill and Carson2022; Uzogara and Jackson Reference Uzogara and Jackson2016; White Reference White2015; Yadon Reference Yadon2022; but see Branigan et al. Reference Branigan, Freese, Patir, McDade, Liu and Kiefe2013; Reference Branigan, Wildeman, Freese and Kiefe2017). These are not merely statistical disparities; they represent divergent lived experiences shaped by how dark or light skin is perceived by institutions, gatekeepers, and everyday citizens. A darker-skinned Black American navigating the labor market, the criminal justice system, or the healthcare system encounters a meaningfully different set of obstacles than lighter-skinned Black Americans, even when both share the same racial identity. This divergence of lived experiences rests at the core of our argument. Therefore, we hypothesize that darker skin is associated with stronger levels of representation.
We theorize two pathways through which colorism influences representational outcomes. First, direct experiences with colorism may inform a Black representative’s desire to substantively represent Black people (Gamble Reference Gamble2011, 298–99; Mansbridge Reference Mansbridge1999, 628). Because skin tone shapes the material conditions, daily encounters, and institutional obstacles that Black Americans navigate, darker-skinned Black legislators are uniquely positioned to represent the interests of Black constituents broadly.
Having faced the sharpest and most frequent experiences of racialization, darker-skinned legislators bring firsthand understanding of the burdens that colorism compounds. Direct experiences with colorism may inform a Black representative’s desire to substantively represent Black people (Gamble Reference Gamble2011, 298–99; Mansbridge Reference Mansbridge1999, 628). Through interviews, Black legislators highlight how life experiences related to colorism impact how they approach their jobs as legislators. In an interview with Brown (Reference Brown2014, 300–1), one legislator described experiencing colorism as a child and drew on the experience to declare, “When I’m in the legislature people can say what they want about me. I know who I am. I’m here for a purpose,” that purpose being “to advocate for the voiceless.” As such, the kind of representation Black legislators provide to African Americans may vary based on these experiences. Legislators’ personal accounts of their skin tone suggest that colorism may influence the representation Black legislators provide to Black communities by shaping their personal experiences and motivations to advocate for Black Americans. Experiencing colorism in everyday life or within the legislature may make racism a particularly salient issue and lead darker-skinned Black representatives to be more likely to advocate for Black Americans (e.g., Brown Reference Brown2014; Lemi Reference Lemi2018). This may be particularly true given that Black legislators who face more discrimination may be acutely aware of the costs of racial advocacy for Black Americans (Gause Reference Gause2022).
By contrast, lighter-skinned Black representatives, including lighter-skinned representatives with non-Black parentage, may have fundamentally different experiences with colorism and anti-Black racism than darker-skinned Black representatives, including darker-skinned representatives with non-Black parentage (see Davenport Reference Davenport2018; Lemi Reference Lemi2018; Masuoka Reference Masuoka2017). As a result of differences in experiences with colorism—not just socialization from parentage—Black legislators may vary by skin tone in the degree to which they explicitly advocate for Black Americans.Footnote 3
Second, colorism may structure a Black representative’s behavior before and after their election. Generally, it is harder for darker-skinned Black people to get elected. Historically, most Black legislators have been lighter skinned (Hochschild and Weaver Reference Hochschild and Weaver2007). Among Black women, the pool of individuals who compete for elected office has also been found to skew lighter skinned (Brown and Lemi Reference Brown and Lemi2021). Scholarship on candidate evaluation has found a relationship between skin tone and candidate evaluation (e.g., Burge, Wamble, and Cuomo Reference Burge, Wamble and Cuomo2020; Chirco and Buchanan Reference Chirco and Buchanan2022; Lemi and Brown Reference Lemi and Brown2019; Orey and Zhang Reference Orey and Zhang2019; Terkildsen Reference Terkildsen1993; Weaver Reference Weaver2012). White voters in particular tend to favor lighter-skinned Black candidates over darker-skinned Black candidates (Chirco and Buchannan Reference Chirco and Buchanan2022; Terkildsen Reference Terkildsen1993; Weaver Reference Weaver2012).Footnote 4 Colorism influences campaigns, ultimately shaping who gets elected to Congress.
As a result, lighter-skinned Black representatives may see more career opportunities in politics given their ability to attract some white support (Strickland and Whicker Reference Strickland and Whicker1992). Given that Black people do not make up a majority in any state, those hoping to win statewide positions may focus on running a race-neutral campaign or engage in racial distancing or emphasizing their closeness to white people via “whitewashing” (Hutchings et al. Reference Hutchings, Nichols, Gause and Piston2021; Stephens-Dougan Reference Stephens-Dougan2020; Reference Stephens-Dougan2021).Footnote 5 Research demonstrates that these race-neutral campaign appeals are an effective way for candidates of color to attract white support (Perry Reference Perry1991; Price Reference Price2016; Stephens-Dougan Reference Stephens-Dougan2020).
Consequently, lighter-skinned representatives may deracialize their campaigns in the hope of attracting a broader base of support when running for elected office (Strickland and Whicker Reference Strickland and Whicker1992). While lighter-skinned legislators may still craft bills and issue press releases that benefit Black Americans, they may do so in a way that is more conservative or does not explicitly mention African Americans to avoid losing support from white voters. By contrast, darker-skinned candidates may cut their losses with white voters and invest their resources in explicitly advocating for Black interests, a group they might rely on (Burge, Wamble, and Cuomo Reference Burge, Wamble and Cuomo2020; Grose Reference Grose2011). Taken together, skin tone may be associated with the extent to which Black US House Representatives advocate for Black Americans. Therefore, our first hypothesis is as follows:
Hypothesis 1: on average, darker-skinned Black legislators issue more Black-centered press releases and sponsor more Black-centered legislation than their lighter-skinned counterparts.
We anticipate that skin tone will matter more for rhetorical outreach through press releases than legislative representation.Footnote 6 We expect to observe differences in rhetorical representation in press releases because the representative and their staff have control over the content of their press release, its timing, and any public relations follow-up required by the representative’s office.Footnote 7 Black representatives are less constrained in the types of press releases that they issue, but they face institutional and partisan barriers in the bills they introduce (Cox and McCubbins Reference Cox and McCubbins2005). In contrast to rhetorical representation, when members of Congress introduce bills, they may make a different set of considerations that have to do with their institutional position and accountability (Cox and McCubbins Reference Cox and McCubbins2005; Grose Reference Grose2011, 25; Rocca and Sanchez Reference Rocca and Sanchez2008; Woon Reference Woon2008). As a numerical minority, Black representatives may not appear to behave differently according to skin tone when introducing bills because racial group interests may be most salient (e.g., Hochschild and Weaver Reference Hochschild and Weaver2007; Kanter Reference Kanter1977). They may experience racism within the institution (e.g., Hedge, Button, and Spear Reference Hedge, Button and Spear1996; Tyson Reference Tyson2016) and as a result behave more cohesively as a group within Congress.
They may also make considerations tied to reelection (Grose Reference Grose2011, 24). If representatives want the bill to be successful, they may consider how the bill fits into the broader package of their party (e.g., Griffin and Newman Reference Griffin and Newman2019; Grose Reference Grose2011), their caucuses (e.g., Peay and Rackey Reference Peay and Rackey2021b; Pinney and Serra Reference Pinney and Serra2002; Tyson Reference Tyson2016), their goals for their stature in Congress (e.g., Mixon and Pagels Reference Mixon and Pagels2007), and their constituents’ reactions (e.g., Grose Reference Grose2011, chap. 2). Introducing bills that are likely to pass may require more internal coordination, more loss of control, and more negotiation with a wider set of stakeholders than simply releasing a statement from one’s own office (Grose Reference Grose2011, chap. 2).
These fundamental differences were exemplified in the summer of 2020 following the death of George Floyd on May 25. Members of Congress who were moved by Floyd’s murder and who wanted to alert the public about their concerns could put out as many press releases as they wanted about the topic, highlighting any aspect they deemed in need of attention, without coordinating with other members of Congress. Along these lines, the vast majority of Black members of Congress issued a statement about police brutality in the months following Floyd’s death. Many of these members issued multiple statements.
In contrast, Black legislators are constrained in their ability to influence the legislative process. Returning to the Floyd example, legislators did not all introduce separate bills addressing racial justice in policing, but instead coordinated their efforts on just two bills: the George Floyd Justice in Policing Act in the House and the JUSTICE Act in the Senate. Given that most of the efforts were centered on just two bills, the nuances in legislative priorities among members are not as well captured in the legislative process, which gives less discretion to individual members. This process, which incentivizes compromise and moderation, may lessen the importance of skin tone in racial representation.
Under forces of polarization and insecure majorities (Lee Reference Lee2016), advancing policy priorities onto the legislative agenda and through the lawmaking process is tough to do. Party leaders have tremendous agenda-setting control (Cox and McCubbins Reference Cox and McCubbins2005; [1993] Reference Cox and McCubbins2007). Aimed at preventing division with the majority party, both parties have come to rely on the Hastert rule, named after former Speaker of the House Dennis Hastert (R-IL), who refused to advance legislation that did not have the support of the majority of the majority party. As such, the support of the party median is needed to advance legislation on the agenda (Cox and McCubbins Reference Cox and McCubbins2005).
Though there has been a growing concordance between the Congressional Black Caucus and the Democratic Party (Tate [Reference Tate2014] 2020), the Democratic Party has been hesitant to include racialized issues on its agenda (Frymer Reference Frymer2011). Even when Black members of Congress promote nonracialized issues, such policies are often viewed as “ideologically radical” (Peay Reference Peay2021) and risk being racialized (Tesler Reference Tesler2012), impeding their advancement. Viewing policy issues advanced by Black legislators as divisive, party leaders prevent their advancement for fear of dividing the party. Darker-skinned Black legislators, as numerical minorities within a numerical minority (Kanter Reference Kanter1977), may face even tighter challenges to getting their bills passed because they may navigate racism and colorism within their party and Congress.
Moreover, committees serve an important role in advancing priorities through the legislative process. Peay (Reference Peay2021) shows that, on average, only 10% of sponsored legislation made it out of committee during the 103rd to the 112th Congresses. Committee chairs serve as gatekeepers, determining which bills will be given attention and which bills will die in committee. However, studies show that Black legislators have been put on less prestigious committees than white legislators (Griffin and Keane Reference Griffin and Keane2011; Rocca, Sanchez, and Morin Reference Rocca, Sanchez and Morin2011; Tate Reference Tate2003), and on committees whose jurisdictions do not overlap with their personal interests or the interests of their constituents.Footnote 8 Still, Black legislators are a growing presence in committees, and during the 117th Congress held seven full committee chairs (Office of the Historian 2023) and 27 subcommittee chairs (Office of the Historian 2025). Despite this, Peay (Reference Peay2021) shows that Black-sponsored bills are disproportionately winnowed in congressional committees, preventing them from advancing to the House.Footnote 9
Considering the obstacles Black legislators face and the constraints on highlighting individual legislative priorities, our second hypothesis makes the following argument:
Hypothesis 2: we expect a Black legislator’s skin tone to be a stronger factor in racial rhetorical outreach than racial legislative representation.
Data and Methods
We focus on two forms of substantive representation that differ in the amount of the legislator’s control (e.g., Grose Reference Grose2011). We measure rhetorical representation by analyzing all press releases from a member in the 114th (2015–17), 115th (2017–19), and 116th (2019–21) Congresses. In each Congress, we obtained press releases for over four hundred of the 435 US House representatives. Our dataset has 204,806 press releases for our period of interest. The only members we do not have data for are those who did not post during our period of interest or whose press releases could not be collected through webscraping (Wickham Reference Wickham2026). In the online appendix, we show that our sample is no different from the complete US House with regards to the race, partisanship, and gender of the representative, or the characteristics of the districts that they represent (see appendix table SA.1).
We measure legislative representation by exploring the legislative summaries of all bills, resolutions, concurrent resolutions, and joint resolutions introduced in these Congresses. We obtained the bill summaries with the names of the primary sponsors from Congress’s official website, congress.gov. We have 18,025 pieces of legislation introduced in the US House of Representatives between January 3, 2015, and January 2, 2021. Unlike press releases, we obtained all available information on legislation because it came from a single source, congress.gov. This combination of press releases and legislation allows us to measure multiple areas of substantive representation for African Americans.
While there are multiple ways we can measure communicative representation, such as website biographies (e.g., Brown and Gershon Reference Brown and Gershon2016), we focus on press releases for several reasons. First, previous research demonstrates that press releases are a common form of political communication used by elected officials to disseminate key messages (Gershon Reference Gershon2012; Grimmer Reference Grimmer2013; McNair [Reference McNair1995] 2017). Grimmer (Reference Grimmer2013), in interviews with communications directors, found that press releases were often the main form of communication that legislative offices used to connect with constituents. Bennett (Reference Bennett1990) finds that one of the reasons press releases are so effective is that they are often heavily quoted by local and national newspapers and have the potential to shape what the media covers. As a result, their reach often extends far beyond those who directly receive the press release.
Moreover, recent research shows that elected officials often disseminate the same messages across platforms (Garcia, Stout, and Tate Reference Garcia, Stout and Tate2026). As a result, if an elected official speaks about race consistently in press releases, they are likely to do the same on other communication channels like Twitter and Facebook. In contrast to press releases, website biographies are specifically meant to orient the politician to the public, not to respond to issues in the news or the legislature as they evolve. As a result, press releases provide a more dynamic measure of outreach than biographies, which tend to be more static over the course of a congressional session. See Lemi and colleagues (Reference Lemi, Garcia and Stout2026) for the complete dataset.
Outcome Variable: Measuring Black Rhetorical Representation
We define a Black-centered appeal as one that highlights an issue, policy, or public figure that is explicitly tied to African Americans. Measuring Black substantive representation is not an easy task given the multitude of ways that elected officials work to appeal to Black people, which vary from focusing on voting rights, to standing with Black Lives Matter, to highlighting symbolic bills recognizing Black Leaders and key figures from the civil rights movement (Arora Reference Arora2019; Crowder Reference Crowder2020; Gillespie Reference Gillespie2010; Grose Reference Grose2011; McIlwain and Caliendo Reference McIlwain and Caliendo2011; Metz and Tate Reference Metz, Tate and Peterson1995; Minta Reference Minta2011; Reeves Reference Reeves1997; Stephens-Dougan Reference Stephens-Dougan2020; Reference Stephens-Dougan2021; Stout Reference Stout2015; Wamble Reference Wamble2018). In our definition of Black-centered issues, we do not include issues like the minimum wage or welfare that may be implicitly associated with Black voters. Rather, we focus on explicit racial statements because they demonstrate the most direct outreach to African Americans. Given racial spillover into issues such as healthcare (Tesler Reference Tesler2012) and gun control (Filindra and Kaplan Reference Filindra and Kaplan2016), this definition constitutes a more confined measure of racial outreach.
We measure Black-centered substantive representation in two ways. First, we use a combination of hand coding and computer-assisted coding via RTextTools (Jurka et al. Reference Jurka, Collingwood, Boydstun, Grossman and van Atteveldt2013). Second, we use a dictionary-based approach. For the former, we used an iterative approach in which we hand coded nine thousand press releases according to whether they included racial outreach (scored as one) or not (scored as zero), and asked RTextTools to use this information to predict whether Black outreach was present in the 195,806 uncoded press releases. We used a similar iterative approach to code 5,901 legislative summaries and asked RTextTools to use this information to code the remaining 12,124. For the legislative summaries, we supplemented this method with the hand coding of all symbolic legislation because the intercoder reliability scores between the hand coder and the computer-assisted coding were poor in this area. More information about our coding and intercoder reliability statistics can be found in the online appendix.
The combination of hand coding and computer-assisted coding has the advantage of providing reliable results over a large corpus of data. It does particularly well in identifying and handling the valence of outreach. While RTextTools is close to perfect in classifying press releases and legislative summaries where racial outreach is absent, it undercounts the number of press releases and legislative summaries that include racial appeals. While this shortcoming does not significantly lower intercoder reliability scores between computer-assisted coding and hand coders dramatically, it remains a limitation of this approach.
As a result, we supplement this classification strategy with a dictionary-based analysis. We used a dictionary of a little over five hundred words that are tied to racial outreach and searched each press release and legislative summary for the presence of these words (see the online appendix for this list). This dictionary approach allowed us to identify a broader set of racial outreach that the computer-assisted coding may have missed. However, unlike the hand-coding/computer-assisted approach, it does not always do well measuring valence. By using both approaches, we can account for the strengths and weaknesses of measuring racial outreach through textual analysis.
Following this process, we created four outcome variables: (1) the percentage of Black-centered press releases a member issued in each congressional session as measured by a combination of hand and computer-assisted coding, (2) the percentage of Black-centered press releases a member issued in each congressional session as measured by dictionary-based coding, (3) the percentage of Black-centered bill introductions a member issued in each congressional session as measured by hand and computer-assisted coding, and (4) the percentage of Black-centered bill introductions a member issued in each congressional session as measured by dictionary-based coding.Footnote 10 To generate an outcome for each member, we divided the number of press releases and legislative summaries that were coded as including racial outreach by the total number of press releases and legislative summaries put out by the member of the House in each Congress. This provides a percentage of Black-centered press releases and bills for each member of Congress. See the online appendix for descriptive statistics of our measures across different legislators. This analysis largely shows that Black representatives, Democratic representatives, and representatives of districts with a larger percentage of Black people are the most likely to engage in black outreach across our measures. This provides some face validity for our dependent variable.
Explanatory Variable: Measuring Skin Tone
Our question focuses on the relationship between experiences with colorism via skin tone and the representation of African Americans. We measure the extent to which a legislator is likely to experience systemic colorism by measuring skin tone with the commonly used Massey-Martin scale (Massey and Martin Reference Massey and Martin2003). The Massey-Martin scale ranges from a score of one (the lightest skin tone) to 10 (the darkest skin tone).Footnote 11 To be clear, skin tone itself is not a marker of colorism, but individuals with lighter skin experience the social and political system differently to those with darker skin. Measuring skin tone is highly subjective and may be contingent on other information about an individual (e.g., Campbell et al. Reference Campbell, Keith, Gonlin and Carter-Sowell2020; Saperstein and Penner Reference Saperstein and Penner2012).
Since the Massey-Martin scale has weak intercoder reliability (Hannon and Defina Reference Hannon and DeFina2016), some scholars have developed alternative measures, such as the Yadon-Ostfeld skin color scale (Ostfeld and Yadon Reference Ostfeld and Yadon2022a). This scale, like the Massey-Martin scale, has 10 skin tone intervals, but with a finer gradation from lighter skin tones (lower scores) to darker skin tones (higher scores). The Yadon-Ostfeld scale also corresponds to measurements of research participant skin tones taken by a spectrophotometer, a medical device that measures the amount of light reflected off a person’s skin (Ostfeld and Yadon Reference Ostfeld and Yadon2022a, 1814).
However, assessments of what is “light” and what is “dark” are socially constructed according to context. For instance, someone who has a light olive skin tone that rarely burns in the sun may be considered “dark” compared with someone who is very fair and easily burns. That same individual with light olive skin is considered “very light” compared with someone with darker skin that never burns. The amount of light reflected off someone’s skin does not tell us about their position in a social and political hierarchy or how others see them. Neither scale “objectively” measures skin tone in a way that is perfectly suited for capturing colorism. Future research might measure the relationship between skin tone and power by asking raters to code photos of people who vary in skin tone and assess how much advantage or disadvantage they have in society. Ultimately, we selected the Massey-Martin scale to connect our findings to other social science research. Creating a new measure that improves upon previous limitations is beyond the scope of this paper.
We also circumvent the challenges presented by using single coders by asking 750 individuals on Lucid Academia to code a random sample of 15 members of the 114th, 115th, and 116th Congresses using their official congress.gov photos without any identifiers (number of ratings = 11,250) (see the online appendix for the demographic characteristics of the sample).Footnote 12
In total, we had about 20 ratings for each member of Congress over our period of interest. This coding represents how the general public in the US might rate the skin tones of these representatives. Each member of Congress was then assigned a summary variable based on their average Massey-Martin score (see Stout et al. Reference Stout, Danielle Casarez Lemi, Bosworth, Chase and Kiefer2024). This score is our explanatory variable. To further check the reliability of this measure, we also hand coded skin tones ourselves using the same photos and the Massey-Martin scale. The hand coding and crowdsourced coding are highly correlated (Pearson’s R = 0.8345).
Overall, on average, Black elected officials (average Massey-Martin score = 5.47) are rated darker than Latinx elected officials (average Massey-Martin score = 3.41) and white elected officials (average Massey-Martin score = 2.52). Similarly, Latinx elected officials are, on average, rated as being darker than white elected officials. These differences suggest that our measure of skin tone is sufficiently valid for our analysis. Notably, this skin tone measure reflects congressional photos, which, of course, may vary in quality, lighting, and angle.Footnote 13 However, these official photos are often what people see as representative of their members of Congress, and we believe that relying on photos is sufficient for establishing a baseline relationship between skin tone and representation.
Estimating the Relationship between Skin Tone and Substantive Representation of Black Interests
To assess the relationship between skin tone and Black-centered outreach, we used ordinary least squares (OLS) regression with random effects to estimate the relationship between Massey-Martin scores and the percentage of Black-centered press releases or bills each member produced in the 114th, 115th, and 116th Congresses. The random effects account for differences across each member of Congress. In total, we ran four models estimating the relationship between skin tone and representation for each type of representation and each of our measures: (1) press releases that were hand coded or coded with computer assistance, (2) dictionary-coded press releases, (3) bill introductions that were hand coded or coded with computer assistance, and (4) dictionary-coded bill introductions.
To isolate the relationship between skin tone and racial representation, we control for several variables. First, we control for the representative’s party. While there are only two Black Republicans in our dataset, Democratic representatives’ greater effort to appeal to Black people may influence our results. Second, we control for whether the elected official is multiracial. Research shows that multiracial elected officials use their memberships of multiple racial groups strategically to gain support for their bills (Lemi Reference Lemi2025; see also Leslie et al. Reference Leslie, Masuoka, Gaither, Remedios and Vinluan2022). Moreover, many multiracial elected officials often have a lighter skin tone than their non-multiracial elected Black counterparts.
We also control for the gender and the ideology of the elected official using DW-NOMINATE scores. We anticipate that women and more liberal elected officials will be more likely to work on behalf of Black people rhetorically and legislatively. We measure district partisanship by their Cook partisan voting index score and use census data to capture each district’s Black population percentage. The electoral incentives in more liberal and racially diverse districts may shape how elected officials address race in office. It may also influence the skin tone of elected officials who succeed in each district. Research shows that Black and non-Black people differ in their preference for candidates based on skin tone (Burge, Wamble, and Cuomo Reference Burge, Wamble and Cuomo2020). Finally, we control for the number of press releases or bills the member introduced in each congressional session. We estimate these models only for Black members of Congress.Footnote 14
Findings
Descriptive Analysis
First, we begin with a descriptive analysis to assess general relationships between skin tone and our measures of representation. Figure 1 presents four scatterplots with best-fitting lines displaying the relationship between Black members’ skin tones and the percentage of Black-centered appeals they issued in press releases and bill introductions.
Descriptive Analysis of the Skin Tones of Black Members of Congress and Substantive Representation
Figure 1 Long description
This multi-panel figure contains four scatter plots arranged in a two-by-two grid. All panels share a common x-axis labeled Massey-Martin skin tone, ranging from 3 to 8.
Top-left panel: Titled Average Percentage of Black Appeals in Press Releases (Hand/Computer Coded). The y-axis ranges from 0 to 0.4. Data points are scattered between skin tones 3 and 8, with a slight positive linear increase in the trend line starting near 0.1 and ending near 0.2.
Top-right panel: Titled Average Percentage of Black Appeals in Press Releases (Dictionary Coded). The y-axis ranges from 0 to 0.5. The data distribution is similar to the first panel, showing a slight positive linear increase in the trend line from approximately 0.1 to 0.2.
Bottom-left panel: Titled Average Percentage of Black Appeals in Legislation (Hand/Computer Coded). The y-axis ranges from 0 to 0.8. The data points are more concentrated near the bottom of the graph, and the trend line is nearly flat, showing almost no change across skin tones, remaining near the 0.15 mark.
Bottom-right panel: Titled Average Percentage of Black Appeals in Legislation (Dictionary Coded). The y-axis ranges from 0 to 0.6. Similar to the bottom-left, the data points are clustered at lower percentages, and the trend line shows a very slight, almost negligible positive slope starting just above 0.1.
Our initial descriptive analysis indicates that skin tone is associated with racial outreach in press releases, but not in the introduction of legislation. Figure 1 suggests a substantial and positive relationship between the skin tones of African American House representatives and the percentage of Black-centered appeals issued in the 114th–116th Congresses. For the hand/computer-coded measure of rhetorical representation in press releases, the Pearson’s correlation coefficient between skin tone and Black outreach is 0.17 and statistically significant (p = 0.0462). The correlation between skin tone and Black representation is slightly stronger for the dictionary-based measure of racial outreach in press releases (r = 0.23, p = 0.008). The results in figure 1 present initial support for our hypothesis that skin tone is positively correlated with an increase in Black-centered press releases.Footnote 15
In contrast, we do not find a strong relationship between the skin tone of African American representatives and the percentage of Black-centered legislation introduced. For both the hand/computer-coded and dictionary-based measures of representation, the best-fitting line between skin tone and racial outreach is essentially flat. Moreover, the Pearson’s correlations for both measures are close to zero (Pearson’s R for hand/computer-coded measure = −0.02, Pearson’s R for dictionary-based measure = −0.01) and statistically nonsignificant. This suggests that skin tone is associated with how Black representatives provide rhetorical representation, but not with how they provide legislative representation. Taken together, these correlations may seem fairly weak, but at the scale of tens of thousands of pieces of legislation and hundreds of thousands of press releases, even a lower correlation is substantively important for the rhetorical representation of Black communities.
Regression Results
Table 1 presents the results of four random-effect OLS regression models predicting the percentage of press releases and primary sponsored legislation that each Black member introduced across three congressional sessions. The models are disaggregated by the form of representation (legislative or rhetorical) and the mode of content coding (hand coding/computer-assisted coding or dictionary based). The models include only Black House members from the 114th, 115th, and 116th Congresses. The models include random effects to control for each individual member of Congress. In addition to skin tone, each model includes all of the aforementioned controls.
Random-Effect OLS Regression Predicting the Percentage of Press Releases and Congressional Bills that Include Black-Centered Issues
Table 1 Long description
A table presenting random-effect O L S regression results. The columns are divided into two main categories: Press release and Congressional bills, each further split into Hand coded and Dictionary methods.
Independent variables and their coefficients (standard errors in parentheses):
* Skin tone: Press release (Hand coded: 0.0163*, Dictionary: 0.0195**); Congressional bills (Hand coded: minus 0.00365, Dictionary: 0.00496).
* Female: Press release (Hand coded: 0.0215, Dictionary: 0.0180); Congressional bills (Hand coded: 0.0377, Dictionary: 0.00604).
* Democrat: Press release (Hand coded: minus 0.0443, Dictionary: 0.0245); Congressional bills (Hand coded: 0.102, Dictionary: 0.0611).
* D W minus N O M I N A T E score: Press release (Hand coded: minus 0.136*, Dictionary: minus 0.0647); Congressional bills (Hand coded: 0.0421, Dictionary: 0.0296).
* Partisan voting index: Press release (Hand coded: minus 0.000356, Dictionary: 3.74e minus 05); Congressional bills (Hand coded: 0.000467, Dictionary: 0.000525).
* Percent Black in district: Press release (Hand coded: 0.00137**, Dictionary: 0.00124**); Congressional bills (Hand coded: 0.00178**, Dictionary: 0.00190***).
* Total press releases: Press release (Hand coded: minus 3.13e minus 05, Dictionary: 0.000130).
* Bills total: Congressional bills (Hand coded: minus 4.67e minus 06, Dictionary: 0.00127).
* Constant: Press release (Hand coded: minus 0.0290, Dictionary: minus 0.0861); Congressional bills (Hand coded: minus 0.0147, Dictionary: minus 0.0601).
Summary statistics:
* Observations: 135 for Hand coded press releases; 141 for all other models.
* Number of candidates: 57 for Hand coded press releases; 59 for all other models.
Significance levels: * p < 0.1; ** p < 0.05; *** p < 0.01.
Note: Standard errors in parentheses. * p < 0.1; ** p < 0.05; *** p < 0.01.
In support of hypothesis 1, skin tone matters for the content of Black legislators’ press releases. Overall, we find that Black members of Congress who are darker skinned, holding all else equal, are marginally statistically significantly more likely to speak out in support of Black political interests in press releases, but not legislation. Consistent with hypothesis 2, skin tone does not predict the content of Black representatives’ legislation in our data. This is true for both legislative summaries that were classified using the hand-coded/computer-assisted approach and the dictionary-based approach.
Diving deeper into hypothesis 1, in the full model for the dictionary-coded measure, Black members of Congress issue about 2% more Black-centered press releases for every one-unit increase in their average Massey-Martin score. Going from the minimum (i.e., lightest) score for African Americans on the skin tone index to the maximum (i.e., darkest) score yields about 7.8% more Black-centered press releases in each congressional session. This relationship is marginally statistically significant (p < 0.10). This translates into about 12 more press releases per congressional session from the lightest to the darkest Black members of Congress.
Similarly, Black representatives issue about 1.7% more Black-centered press releases using our hand-/computer-coded measure of racial outreach for every one-unit increase in a representative’s Massey-Martin score. This relationship is also marginally statistically significant (p < 0.10). These findings are also robust and stronger statistically when we fully interact with skin tone and race (e.g., Block, Golder, and Golder Reference Block, Golder and Golder2023; see table SA.2 in the online appendix).Footnote 16 In short, we find evidence that there are internal differences concerning representation.
Discussion and Conclusion
In this article, we examined the relationship between skin tone and substantive representation of African Americans among Black members of the US House. We theorized that colorism influences personal motivations and electoral incentives to advocate for African Americans. In our analysis of press releases and bill introductions in the 114th, 115th, and 116th Congresses, we find that overall, darker-skinned African American elected officials are more likely to make explicit references to Black people and Black issues in their press releases. However, skin tone is not correlated with the percentage of Black-centered bills members introduce in each congressional session. We attribute our findings to personal experiences, electoral incentives, and the degree of autonomy that representatives have over press releases as compared with bill introductions. In short, we find correlational evidence that colorism plays a role in representational outcomes.
We believe this study demonstrates the value of more nuanced approaches to conceiving of and testing the effectiveness of descriptive representation. While dominant approaches to studying descriptive representation often focus on race alone (but see Brown Reference Brown2014), we incorporate skin tone as an attribute that differentiates descriptive representatives (Monk Reference Monk2022). In doing so, our study links the systemic colorism observed in socioeconomic outcomes to representational outcomes at the elite level. Fully understanding the effects of descriptive representation means considering the diversity among descriptive representatives.
That said, our findings should be taken with a note of caution. Given systematic barriers to electoral participation, Black people made up a smaller proportion of Congress in the 114th through 116th congressional sessions than they did in the wider US population. As a result, we had fewer than 60 members to analyze across these time frames. This small sample size raises concerns about statistical power, and even more so in relation to some of the additional interactions we mention below.
Moreover, while our results are consistent across a variety of model specifications, the impossibility of randomizing skin tone to show a causal relationship increases the probability that omitted variable bias could influence our relationship of interest. As the number of Black members in Congress increases over time, we hope that future research will further explore the relationship between skin tone and representation with a larger set of Black elected officials. Additionally, future research should examine how different measures of skin tone correlate with different substantive outcomes by comparing various scales, including the Massey-Martin scale and the Yadon-Ostfeld scale.
This study raises several questions for future research. First, we find a null difference between Black men and women (see table SA.6 in the online appendix) in the exploratory supplementary analysis. On the one hand, previous research indicates that Black women experience colorism and legislate in distinct ways (Brown Reference Brown2014; Brown and Lemi Reference Brown and Lemi2021; Hawkesworth Reference Hawkesworth2003; Lemi and Brown Reference Lemi and Brown2021; Orey et al. Reference Orey, Smooth, Adams and Harris-Clark2006; Smooth Reference Smooth2011). Moreover, recent scholarship demonstrates that Black voters may believe that darker-skinned Black women candidates are more likely to win Black voters’ support than lighter-skinned Black women candidates (Burge, Wamble, and Cuomo Reference Burge, Wamble and Cuomo2020). Yet Black women legislators themselves do not always perceive that skin tone matters in their work (Brown and Lemi Reference Brown and Lemi2021). While colorism may shape general opinions of Black men and women candidates, it does not appear to shape gender differences in legislative and rhetorical representation measured here.
One explanation for the lack of gender differences is that our sample size, while large enough to detect differences by skin tone, is not large enough to detect differences between genders among Black members of Congress. Another explanation is that Black men and women truly do not differ in their behavior in the outcomes measured here. This disconnect between what we observe in evaluations of Black men and women candidates, how Black women experience the legislature, and how Black men and women actually represent their constituents should be explored. They may differ in other forms of representation not measured here, such as staff composition or the location of their district offices (Grose Reference Grose2011). In the exploratory analysis in table SA.6 in the online appendix, we do not find evidence of an interaction between gender and skin tone, but that does not mean one does not exist.
More research needs to be done that explores the content of the topics discussed in Black-centered press releases and legislation. While a large portion of rhetoric by Black legislators, specifically members of the Congressional Black Caucus, on social media is not tied to racial topics (Strawbridge et al. Reference Strawbridge, Clark, Mahoney and Brown2025), when Black legislators do speak about race they may do so in a way that accounts for the diversity of the Black experience in the US (Garcia, Stout, and Tate Reference Garcia, Stout and Tate2026). Black legislators who experience the world differently due to their appearance may highlight different racial topics in their outreach. This could lead to a wider set of issues that affect Black people to gain attention and improve representation for this underrepresented group. We coded legislation and press releases dichotomously in this study and thus are not able to fully assess this hypothesis. However, we implore future research to engage in a more nuanced examination of Black-centered outreach to explore this hypothesis.
We did not find evidence that skin tone influences legislative responsiveness. To maximize the likelihood that legislation completes the process and becomes law requires each member to work with their colleagues to produce the strongest bill possible. Legislative constraints and our dichotomous measure of bill introduction may mask important nuances in the legislative process. For example, while skin tone may not influence bill introduction, it may play a role in less visible parts of the process like responsiveness in committee markups or legislative hearings (Grose Reference Grose2011; Minta Reference Minta2011). We implore future research to use our framework and apply it to different types of representation, particularly those that are less constrained by institutional and party factors.
Additionally, colorism affects many groups racialized as nonwhite. While we focus on African Americans here, future research may consider how colorism structures representational outcomes for Latinx representatives, including Afro-Latinx representatives (see Matos, Morel, and Vásquez Reference Matos, Morel and Vásquez2026). Although our dataset lacked a sufficient sample size to analyze Afro-Latinx representatives specifically such as Congressmembers Ritchie Torres (D-NY) or Adriano Espaillat (D-NY), we would expect similar patterns among Afro-Latinx with darker skin tones. Among Latinx communities, colorism is rooted in anti-Blackness (Hernández Reference Hernández2022), and there is reason to believe that personal experiences with anti-Blackness would motivate Afro-Latinx representatives to produce more Black-centered press releases.Footnote 17 Future research should consider conducting similar analyses at the local level and case studies on Afro-Latinx representatives.
We invite representation scholars who study the intersection of gender, race, ethnicity, and skin tone to explore these questions.
Although our analysis was limited to Congress, future research should replicate our work at the state level (i.e., Clark Reference Clark2019). State-level analyses will increase the overall number and diversity of Black representatives included in the analysis and enable scholars to account for state-level institutional contexts. Other questions to explore include the following: Is there a numerical threshold of Black representation at which skin tone begins to matter, or not matter, for substantive representation? Do Black state legislators perceive colorism within their state legislatures (e.g., Brown Reference Brown2014; Tyson Reference Tyson2016)? How might colorism, sexism, and homophobia structure representational behavior for Black representatives in different legislative contexts (e.g., Brant and Butcher Reference Brant and Butcher2022)? And how do Black state legislators vary by skin tone in their use of other communication platforms more directly under their control, like social media (e.g., Tillery Reference Tillery2021), speeches (Peay and Rackey Reference Peay and Rackey2021a), and hearings (Minta Reference Minta2011)? Moreover, given the correlational design of this study, future research might also look for opportunities to track legislators’ behaviors over longer time horizons, employ matching techniques to compare lighter and darker-skinned representatives, or leverage sudden shifts in skin tone representation in other institutions, such as city councils, to further explore the role of colorism in representational outcomes.
More broadly, we believe this article serves as a call for political scientists to think more expansively about the experiences and impacts of descriptive representatives (see also Xydias Reference Xydias2023). White supremacy is the reason why communities of color—and Black communities in particular—need descriptive representation in American political institutions in the first place (Mansbridge Reference Mansbridge1999). Although the benefits of descriptive representation often focus on the ability of descriptive representatives to be bridges between institutions and their communities (Mansbridge Reference Mansbridge1999), this study highlights how descriptive representatives, as racialized people, navigate the obstacles—and opportunities—presented by white supremacy to represent Black Americans once elected.
Supplementary material
To view supplementary material for this article, please visit http://doi.org/10.1017/S1537592726105660.
Data replication
Data replication sets are available in Harvard Dataverse at: https://doi.org/10.7910/DVN/DRIEKC
