Voter Evaluations of Biracial-Identified Political Candidates

Today, identity expression and acceptance represent an important area of political advocacy and representation. Yet, how responsive are voters to new racial identity cues promoted by political leaders? Using candidates with interracial backgrounds as a case study, we assess whether voters are responsive to candidates who assert a mixed-race identity or if voters primarily rely on other traits, such as the candidate’s family background, in determining their support of that candidate. Using an experimental design, this study presents participants with various hypothetical candidates who vary both in their racial heritages (i.e., candidates with Asian and White interracial parents or Black and White interracial parents) and identity choices (i.e., as single-race minority, single-race White, or biracial). We then compare how the mixed-race, single-race minority, and White participants evaluate the candidate. We expect that the mixed-race participants will be most supportive of candidates who signal a common in-group identity by identifying specifically as “biracial”. On the other hand, the single-race minority and White participants should be more likely to adhere to the one-drop rule or hypodescent in their evaluations, meaning they will provide more positive evaluations of interracial candidates who identify as a single-race minority. Our study finds that the single-race minority and White participants completely overlook racial identity cues and instead focus on the description of the candidate’s family heritage along with their own assumptions about hypodescent. The mixed-race participants, on the other hand, show strong support for biracial-identified, in-group political candidates This study adds to a burgeoning literature on racial perception and on political


Introduction
In recent years, Americans have increasingly had the opportunity to support elected officials who openly share with the public about their mixed-race heritage or interracial family background. President Barack Obama could be looked to as a model demonstrating how American voters were open towards political candidates who emphasized a family narrative centered around interracial marriage (Obama 2004). Since Obama's election, the number of elected officials who openly share about their mixed-race background has blossomed to include nationally recognizable leaders such as Kamala Harris, Tammy Duckworth, Charles Rangel, and Laura Richardson, along with many others representing local-level office. Yet, while the American public may be open towards narratives celebrating interracial marriages, there continue to be constraints on how elected officials can racially identify (Lemi 2018). Elected officials may share about their interracial family backgrounds, but most tend to identify with only one race. President Obama personally These studies also primarily focus on how voters who identify with only one race (as Asian, Black, Latino, or White) respond to candidates described as having an interracial family background. This current study expands on our existing knowledge in many ways. First, ours is the first such study to examine how voters respond to differing identity cues among candidates who simultaneously divulge their interracial family background. Second, our study includes a unique sample consisting of a sizeable number (n = 234) of participants who report having two or more racial backgrounds, along with samples of participants who have single-race backgrounds (as Asian American, Black, or White). This allows us to explore whether mixed-race participants may respond to racial identity cues differently than single-race participants. Third, given the longstanding racial tension between Whites and Blacks, as well as the tremendous salience of Barack Obama as a representative of mixed-race candidates, scholarship on the mixed-race experience tends to focus on mixedrace populations with Black and White backgrounds. Yet, demographic patterns show that interracial coupling involving an Asian American partner is on the rise (Wang 2012); so, for this study, we focus on mixed-race candidates with Asian and White heritage. By focusing on mixed-race identities involving Asian and White heritage, this study offers new insights into how the longstanding rule of hypodescent applies to non-Black minority populations.
In the following sections, we integrate diverse literatures from the social sciences and review what we know about the role of race in candidate evaluation and how the introduction of mixed-race identities complicates established assumptions. After reviewing the literature, we then present analyses from our original, embedded survey experiment that tests responses to mixed-race heritage candidates who provide varied cues about their racial identity.

Race, Candidate Evaluation, and the Influence of Hypodescent
This study seeks to expand on our understanding of how individuals rely on race in making an assessment and offering support for elected officials. One assumption that political scientists have generally come to expect is that most individuals do not (or cannot) process large amounts of information when they are asked to make political decisions. Rather than taking in new information to make decisions, individuals rely on cognitive shortcuts to help make decisions about new situations or to account for missing information (Feldman and Conover 1983). For this reason, a candidate's race is understood to offer a powerful cue to voters which signals the kind of politics a candidate promotes. However, there exists an important variation in how voters use the race of the candidate as a political cue. White voters, for example, tend to rely on race as a vehicle for political stereotyping, such as assuming that all Black elected officials are politically liberal (Leslie et al. 2019).
For voters of color, however, descriptive representation is a powerful tool in redressing historical inequalities in political representation, and its myriad benefits are well documented. For example, early work theorizes that descriptive representation leads to increased communication between voters and elected officials and fosters more productive policy deliberation between same-race representatives (Mansbridge 1999). Recent research demonstrates that elected officials who share the same race as their constituents are known to enjoy more electoral support (Harris 2012), engage more regularly in substantive representation (Grose 2011;Clark 2019), and improve political efficacy and trust in the government among their constituents (Merolla et al. 2013;Fowler et al. 2014), and embolden citizens to learn about politics (Wolak and Juenke 2021) and to participate in the electoral process (Barreto 2010;Tate 2001;Grumbach and Sahn 2020;Bobo and Gilliam 1990). However, shared racial heritage alone is often not sufficient for reaping the benefits of descriptive representation among people of color. Rather, gender and skin tone, to name a few, are important factors for determining whether or not same-race candidates can activate feelings of a common in-group identity among voters (Lemi and Brown 2020;Burge et al. 2020). Between-group differences may also exist, with some evidence suggesting that Blacks may be more responsive to in-group identity cueing than Latinos (Adida et al. 2016).
Yet, while we have learned that race is an oft-used cue in voting, a limitation of these studies is that race is typically assumed to be unambiguous and straightforward for voters. This process is also complicated by the fact that voters may be unsure how to place a candidate's priorities even if they are known to be mixed-race. One study, for example, shows that mixed-race candidates may engender doubt about their priorities among voters, even if they share a common racial group heritage (Lemi 2021). Indeed, many works show that the ability to racially categorize is far from an effortless process. In reality, there are a number of traits that one can rely on when sorting others into racial categories. Social scientists have pointed to factors including phenotype or visual appearance, cultural markers such as clothing, or the type of first or last name, while legal precedent has in the past relied on science and common knowledge to define a racial category (Rickford 2016;Correll et al. 2002;Greenwald et al. 2003;Haney Lopez 2006). Moreover, most of the literature on candidate evaluation has studied the perceptions of White voters (for example Berinsky et al. 2011;Sigelman et al. 1995), while a smaller but growing scholarship has evaluated viewpoints of Black, Latino, and Asian American voters (for example : Barreto 2010;Junn and Masuoka 2008;Tate 2001). While existing studies have established important findings, as diversity increases we can expect that the use of race in candidate evaluation will become increasingly more complex.
In academic scholarship, mixed-race individuals represent a unique case study for exploring the many different ways individuals perceive and process racial categories. Studies on mixed-race populations have served as a critical intervention for how we understand the ways that individuals use race as an informational cue. In particular, a growing field in psychology explores how external perceivers presented with the faces of mixed race or racially ambiguous-appearing faces categorize those faces as members of different racial groups (e.g., Blascovich et al. 1997;Chen and Hamilton 2012;Chen et al. 2014;Freeman et al. 2010;Gaither et al. 2016). From this literature, a prominent finding is that perceivers who identify with only one race tend to rely on the rule of hypodescent when classifying mixed-race faces into racial categories. For example, studies find that when computer generated faces are presented to have both Black and White heritage or Asian and White heritage, the perceivers generally classify these faces as "Black" or as "Asian", respectively, and not as "White" (Ho et al. 2011;Pauker et al. 2018). These studies further show that the rule of hypodescent is applied more often to mixed-race Black and White faces than to mixed-race Asian and White faces.
Based on these studies, we can begin to develop new questions for the study of racial categorization in candidate evaluation. We learn that even though race might be recognized to be complex, there is a tendency to classify others into a small number of established, mutually exclusive racial categories. This poses a number of constraints for leaders who might want their mixed-race identities to be acknowledged and recognized by constituents. This also suggests that leaders may not be able to strongly influence how voters perceive them as voters may continue to rely on longstanding cultural practices based on one's racial heritage, such as the rule of hypodescent.

Assignment and Identity: Mixed-Race Individuals Hold Different Understandings of Race Compared to Single-Race Individuals
Another important intervention of this study is to investigate the unique ways that single-race and mixed-race individuals may hold different assumptions about race and how these differing views inform political preferences. Along these lines, our study was designed to test how single-race voters might respond differently than mixed-race voters to varying identity frames presented by candidates with mixed-racial heritage. In this section, we first discuss how single-race individuals might respond to mixed-race identity cues. We then discuss how mixed-race individuals present different modalities for understanding race, particularly in their emphasis on the importance of identity choice.
A distinctive tension experienced by mixed-race individuals today is between the belief that one's race is assigned to them by others or that one's race is a reflection of Soc. Sci. 2022, 11, 171 5 of 31 one's personal preferences for racial self-identification (Masuoka 2017;Does et al. 2021). As we discussed in the previous section, historically, mixed-race individuals have been classified into a single-race category, most often a non-White category. Yet, many mixedrace individuals hold attachments to multiple races and often want to assert an identity that communicates this hybridity (Root 1992;Hall 1992). Mixed-race activist groups have long advocated for the authority and autonomy to self-identify how they wish (Dacosta 2007;Williams 2006). Through this emphasis on chosen identity, mixed-race individuals assert that others should honor these identities rather than rely on other traits to define race. In practice, we have seen how visible public figures have pushed the public to recognize distinctive mixed-race identities and labels. Most famously, athlete Tiger Woods asserted his race as "Cablinasian" or an amalgam of Black, indigenous, and Asian heritage (Cole and Andrews 2011).
While we have seen public figures assert their own preferred racial identities in the public sphere, it is uncertain to what degree those in the general public will, in turn, accept and reference those identities in their evaluations. We might expect that those who identify with only one race will strongly reject mixed-race identities as mixed-race identities may be perceived to challenge the standing racial order governed by the rule of hypodescent. Take for example, an individual who has one minority parent and one White parent and who attempts to assert their identity as White. Historically, social policing upholding the rule of hypodescent denies them opportunity to do so. Such constraints on who can identify as White have also been present in our legal system (Gross 2008;Haney Lopez 2006), and discrimination and violence have been employed to enforce this system. Hypodescent in American culture has been used to preserve the White racial category and thus whiteness has been subject to the most stringent social policing (Reuter 1918;Spickard 1991;Curington 2016;Strmic-Pawl 2016). Mixed-race people historically have only been classified into one of the single-race minority groups regardless of what their personal identity preferences may be (Sims and Njaka 2020). In this way, it may be the case that the majority of single-race individuals may strongly reject mixed-race identities given their perceived challenge to the rule of hypodescent.
In contrast, research has found that due to their personal experiences, mixed-race individuals understand race differently from those who identify with only one race. As a result, mixed-race individuals respond to racial cues in distinct ways. Studies in psychology have found that mixed-race individuals are more willing to acknowledge race as a complex characteristic. Mixed-race individuals are much more likely to categorize racially ambiguous faces as "Multiracial" compared to participants who identify with only one race (Iankilevitch et al. 2020). Researchers theorize that this may be due to the fact that mixed-race individuals tend to hold a less essentialized notion of race than do single-race individuals (Bonam and Shih 2009;Shih et al. 2007), meaning that they are less likely to consider race as biologically based, immutable, and meaningful for determining individuals' values and qualities (Haslam et al. 2000;Prentice and Miller 2007). Qualitative evidence also suggests that mixed-race individuals are less likely to assign others to a race that does not match that person's chosen identity (Does et al. 2021). Building from this, we would expect mixed-race individuals to not only be more responsive to the nuances of different racial cues offered by leaders but also more willing to use those offered cues in their own decision making.
The growth of the mixed-race population represents a cultural shift away from racial classification norms based on external perception by adding a new framework in which categorization could be based on an individual's own agency and choice Masuoka 2017;Dacosta 2007). The existing literature therefore offers useful insight into how mixed-race individuals may respond to race cues in politics differently from those who identify with only one race. In the next section, we build from the literature to develop the expectations specific for our study.

Expectations on Responses to Racial Identity Cues
The core intervention of this study is to evaluate how mixed-racial heritage (i.e., interracial family background) and racial-identity cueing influence participants' perceptions of candidates. Specifically, we examine the extent to which a candidate's own assertion of their racial identity is seen as relevant and important information for participants when presented alongside information about the candidate's mixed-race heritage. For example, if a candidate is described to be of both Asian and White heritage but asserts that their preferred identity label is "White", are participants more likely to pay attention to the candidate's preferred identity and evaluate them as they would a typical White candidate, or will they instead focus on the candidate's Asian and White heritage as they construct their evaluations? Furthermore, what if the candidate identifies as "Asian" or "biracial"?
Questions such as these are of critical importance given that individuals with interracial backgrounds are indeed known to vary in the ways that they racially identify, and often systematically, according to a number of factors. For example, gender (Davenport 2018;Sims and Joseph-Salisbury 2019;Strmic-Pawl 2016) and class (Masuoka 2017) are important predictors of identity choice among individuals with interracial backgrounds, (i.e., identification as a woman, affluence, and educational attainment are all positively correlated with identifying with two or more races rather than as single-race), and some individuals are even known to change the way they racially identify quite often (Doyle and Kao 2007;Harris and Sim 2002;Liebler et al. 2017). Those with mixed-racial heritage tend to negotiate their racial identities situationally and often use racial identity cues strategically to signal commonality with others as they strive for acceptance (Renn 2004;Townsend et al. 2012;Saenz et al. 1995;Xie and Goyette 1997;Shih and Sanchez 2009;Shih et al. 2019). However, while identity cues are used interpersonally to establish commonality, it is not presently clear whether the use of identity cues by political candidates will succeed in influencing the perceptions of others of their racial identities or trigger in-group support.
The specific expectations for this study are rooted in the existing literature. Social science research demonstrates that, all else being equal, voters prefer candidates who share a common in-group identity over candidates who do not, especially if those identities are based on race (Bobo and Zubrinsky 1996;Schuman 1997;Allport 1954;Turner et al. 1987;Huddy 2013). As a consequence, our base expectation is that all participants in general will be more supportive of candidates they perceive as sharing a common racial identity than candidates who do not. Key here is that we are going to comport the evaluations of participants from various racial groups as the litmus tests for whether or not individuals from those groups are focusing on information about the candidate's racial heritage or about their asserted racial identity. For example, ample research demonstrates that Whites have long held a preference for White candidates as compared to minority candidates (Williams 1990;Reeves 1997;Lewis-Beck et al. 2010, but see Highton 2004. So, if we observe that White participants are rating the Asian and White heritage candidate who asserts their identity as "White" more favorably than those who identify as "Asian" or "biracial", we will take it as evidence that White participants are in fact taking the candidate's specific identity cue into account and favoring the candidate with the same racial in-group identity, rather than relying on information about their mixed-race heritage or interracial family background. Similarly, a plethora of literature demonstrates that minorities tend to have strong preferences for minority candidates compared to White ones (Dawson 1994;Tate 1994). However, given the pervasiveness of the one-drop rule as a tool for racial classification, as well as the relative newness of the modern mixed-race population, we do not expect that single-race participants will factor identity cues into their evaluations. Indeed, one of the few studies that includes mixed-race candidates argues similarly that single-race voters may be befuddled by mixed-race categorization and thus unsure of which racial groups' interests the candidate will support (Lemi 2021). As noted above, throughout American history race has not been something that an individual can decide for themselves; rather, one's racial classification is ascribed to them by others based on a set of carefully predetermined and curated-often biological-characteristics meant to preserve boundaries protecting Whiteness (Davis 1991;Curington 2016). We also detailed the literature above which demonstrates that single-race individuals are more likely to hold essentialist conceptualizations of race and racial identity and to rely on this conceptualization when categorizing mixed-race faces (Ho et al. 2015). This propensity for single-race individuals to rely on the perception of race as biological rather than socially constructed and chosen leads us to hypothesize that single-race participants will generally dismiss candidates' identity cues altogether, and thus not differ in their evaluations of candidates with known mixed-race heritage or interracial family backgrounds.
On the other hand, given the considerable literature detailing how mixed-race individuals' own experiences lead them to adopt a more nuanced conceptualization of race, we expect that mixed-race participants will be responsive to candidates' identity cues. Mixedrace individuals have grappled personally with having to make choices about how they racially identify and so are found to have much less essentialized notions of race and to rely on the chosen racial identity decisions of others rather than make their own assumptions (Does et al. 2021). For these reasons, we expect that mixed-race participants will differ in their evaluations based on the different identity cues candidates assert. In particular, we expect that mixed-race participants will be the most supportive of candidates who cue their specific racial identity as "biracial". Given the dramatic increase in the biracial-identifying population in both size and salience over the past few decades, we argue that "biracial" might plausibly represent a distinct in-group, such that candidates who make the effort to emphasize their membership as a biracial may elicit additional support among mixed-race participants.
Lastly, though we hypothesize that single-race participants will not take identity cues into account, we acknowledge that one type of identity assertion-identification as Whitemay be incorporated into both single-race and mixed-race participants' evaluations and engender negative sentiments from all. Although research shows that mixed-race individuals are more likely to accept the identity cues provided by others, research also shows that mixed-race individuals are more interested in identity expression rather than completely overturning the existing racial order (Dacosta 2007). Mixed-race identity activism has, to date, never advocated for the right of mixed-race individuals to self-identify as White (see Williams 2006). In fact, negotiations between Multiracial activists and congress in the 1990s resulted in the stipulation that individuals who identify as both White and with a minority group on the census would by default be counted with their minority group for some purposes (U.S. Office of Management and Budget 2000). This agreement is a corollary to deeply entrenched practices of hypodescent that continue to place pressure on mixed-race individuals to remain more loyal to their minority counterparts than to Whites (Masuoka 2017). So, it may be the case that interracial candidates identifying singularly as White are perceived to violate the norms and expectations of hypodescent so severely that they are disfavored among all-Whites may disapprove because this act attempts to overstep the White/non-White boundary they have defended for centuries, and single-race minorities and mixed-race individuals may disapprove because it represents an abandoning of one's responsibility to underserved minority groups.
The hypotheses for our study are as follows: Hypothesis 1 (H1). Mixed-race participants will take identity cues into account when assessing hypothetical candidates.
Hypothesis 2 (H2). Mixed-race participants will be most supportive of the candidate who identifies actively as biracial.
Hypothesis 3 (H3). Single-race participants will take not identity cues into account when assessing hypothetical candidates.
Hypothesis 4 (H4). All participants will reject a candidate with an interracial background who self-identifies as White.

Method
To examine our hypotheses, we conducted an original, embedded survey experiment which includes a highly diverse sample of both single-race and mixed-race participants. Each participant was given a short vignette to read about a hypothetical political candidate with interracial heritage who is running for the U.S. Senate. Within these vignettes, we vary both the racial heritage of the hypothetical candidate as well as the specific way they choose to racially identify. After being presented with each vignette, the participants were surveyed about their general attitudes toward the hypothetical candidate as well as the likelihood with which they would vote for said candidate.

Participants
The participants were recruited either through a university (private university in Massachusetts) participant subject pool, in-person recruitment in a university student center, and online advertisements in exchange for a chance to win a USD 25 gift card through a raffle. Our dataset includes 273 single-race White participants (56% female; M age = 28.14, SD = 10.43); 139 single-race minority participants (67% female; M age = 28.08, SD = 10.83); and 234 mixed-race participants (75% female; M age = 27.36, SD = 9.11). To measure race, we rely on two items which ask participants to indicate the race of both their mother and father. Mixed-race participants are identified on the basis that they indicated two different races among their mother and father (see Appendix A for the exact wording of each item used in this study). Single-race White and minority participants are similarly identified on the basis that they indicated only White or same-race minority heritage among their mother and father.
It is important to note a few implications involved with our choice to identify mixedrace participants based on information about their parents (mixed-race by heritage) rather than their own identity choices. First, theories of descriptive representation suggest that individuals who strongly identify with their racial group will be most likely to support same-race candidates (Stout et al. 2021). So, it may be the case that mixed-race participants who actively identify as "biracial", "Multiracial", or "mixed-race" would be more supportive of biracial-identified candidates than those identifying as single-race. However, research also shows that identity choice among those with mixed-race heritage is extremely labile, such that they can be more likely to change the way they racially identify at different times or in different contexts than to identify the same way (Doyle and Kao 2007;Liebler et al. 2017). On the other hand, individuals' symbolic or affective relationships with the racial groups in their heritage are known to be quite stable over time (Sears and Funk 1999); so, we suspect there will be minimal differences, if any, between participants who are mixed-race by heritage and those who are mixed-race by self-identification in their evaluations of candidates. 2 Moreover, enumerating mixed-race participants by heritage was essential for enabling us to procure a sufficient sample size for statistical analysis. Still, in our analysis we also include various disaggregations that account for mixed-race individuals who self-identify specifically as "Multiracial" (see the "Race [Self]" item in Appendix A) and with two distinct racial groups (i.e., those who self-identified as both Asian and White-Asian-White-identified participants-and those who self-identified as both Black and White-Black-White-identified participants).
Moreover, it is also important to note that sampling mixed-race participants in social science surveys is often a very difficult task. While a random sample of mixed-race adults would be ideal, the prohibitive cost structure involved in randomly sampling this subpopulation made it unfeasible. To overcome these constraints, we leveraged a diverse student population of young college-going individuals who were known to be more likely to have mixed-racial heritage than the general American population, and we relied on snowball sampling through the mixed-race participants' own networks, using online social media. It is therefore important to acknowledge the possible limitations associated with snowball sampling and college-student samples, especially in terms of generalizability (Dosek 2021;Peterson 2001;Hanel and Vione 2016). However, given that our study is principally concerned with whether or not certain treatment effects and mechanisms indeed exist-rather than on making generalizable claims about a specific group's characteristicswe feel this approach is valid. We also point out that our sample is actually fairly large in comparison to many recent studies and therefore represents a unique contribution given the dearth of political science research on this important population.

Procedure
All the participants were given the same biography of Tom Smith, a hypothetical candidate for Senate, to review. This biography outlined the candidate's career path in politics, his educational background, and his marital status, and it highlighted a non-racial policy specialization (i.e., the environment; see Appendix A for the exact wording). We did not include a photograph of the candidate because previous research has shown that racial phenotype has its own unique impact on participant evaluations (e.g., Pauker et al. 2009;Young et al. 2013), and we were interested primarily in the role that an asserted racial identity cue plays on voter evaluations in contrast to the described family background of the candidate.
Participants were randomly assigned to one of six different treatment conditions, which varied on the bases of the racial heritage of the hypothetical candidate as well as the specific racial identity they cued (see Appendix A for the exact wording of these treatments). The specific treatment conditions are the following: (1) a candidate who has one parent of Asian American descent, one parent of European American descent, and who identifies as Biracial (Asian and White heritage: Biracial-identified); (2) a candidate who has one parent of Asian American descent, one parent of European American descent, and who identifies as Asian (Asian and White heritage: Asian-identified); (3) a candidate who has one parent of Asian American descent, one parent of European American descent, and who identifies as European American (Asian and White heritage: White-identified); (4) a candidate who has one parent of African American descent, one parent of European American descent, and who identifies as Biracial (Black and White heritage: Biracial-identified); (5) a candidate who has one parent of African American descent, one parent of European American descent, and who identifies as Black (Black and White heritage: Black-identified); and (6) a candidate who has one parent of African American descent, one parent of European American descent, and who identifies as White (Black and White heritage: White-identified).
An important note about our data collection effort is that our main goal was to focus on exploring multiracialism outside of the Black and White paradigm. We did this in order to limit concerns that some of our findings represent a spillover of affective predispositions toward Barack Obama, who is perhaps so salient as a mixed-race political candidate that attitudes toward him may be primed by any treatments including candidates with Black and White heritage. As a consequence, we began collecting data first 3 using treatments involving candidates with Asian and White heritage. Once we were confident that we could collect a sufficient sample of mixed-race participants within the conditions involving candidates with Asian and White heritage, we began including three additional treatments, which included candidates with Black and White heritage who identify either as Black, biracial, or White. All told, the sample sizes for each treatment condition are: Asian and White heritage: Asian-identified = 144; Asian and White heritage: Biracial-identified = 145; Asian and White heritage: White-identified = 130; Black and White heritage: Biracial-identified = 41; Black and White heritage: Black-identified = 44; Black and White heritage: White-identified = 141. See Appendix C for more information about the sample, sizes, means, and standard deviations for each participant race × treatment group cell.
After reading the candidate's biography, the participants were first asked to offer evaluations of the candidate; these were measured in two ways: a standard feeling thermometer measure in which participants rated how warmly they felt toward the candidate on a scale of 1 (very cool) to 10 (very warm) and the participant's likelihood of voting for the candidate on a scale of 1 (very unlikely) to 4 (very likely). For easier comparison, we then standardized all variables to a scale of 0 to 1.

Mixed-Race Participants
The first part of this analysis explores whether or not the mixed-race participants will take the specific identity cued by each hypothetical candidate with mixed-race heritage into account (H1). If so, they should be especially favorable toward candidates who actively identify as biracial (H2). To test this, Figure 1 (see also Table 1) presents the mean values of the feeling thermometer and vote intention scores among the mixedrace participants for each treatment condition that includes candidates with mixed-race heritage. The x-axis category, labeled "White", represents the mean ratings of the two conditions in which candidates with Asian and White heritage and Black and White heritage identify as "European American". "Biracial" denotes the two conditions under which these candidates with Asian and White heritage and Black and White heritage identified as "Biracial". "Minority" denotes the candidates with Asian and White heritage and Black and White heritage who identified with their minority racial group (i.e., as Asian or Black; see Appendix C for more information about the mean, standard deviation, and samples sizes for each participant race × treatment condition group). Additionally, Table 1 presents the results of t-tests corresponding to Figure 1. Note that in this analysis, each figure will be followed by a table of the same number, which indicates the magnitude of the difference of means between all the possible treatment group comparisons and presents p-values to highlight where the differences are statistically significant. group by choosing to identify as White. Moreover, six out of six tests confirm that mixedrace participants differentiate between candidates based on the identities they assert (support for H1).

Figure 1.
Mean feeling thermometer and vote intention ratings of hypothetical candidates. All six treatments are represented here and grouped by each identity cue (e.g., "White" denotes candidates from both the Asian and White heritage and Black and White heritage treatment conditions who identified as "European American". Sample includes only mixed-race participants who indicated two or more racial heritages among their parents. X-axis denotes different treatment groups based on the identity cued by our hypothetical candidates. Note: error bars denote 95% confidence intervals. While the initial results indicate that the mixed-race participants have preferences for biracial-identified candidates when the treatment conditions are grouped, we acknowledge that much of way voters feel about mixed-race candidates may be informed Figure 1. Mean feeling thermometer and vote intention ratings of hypothetical candidates. All six treatments are represented here and grouped by each identity cue (e.g., "White" denotes candidates from both the Asian and White heritage and Black and White heritage treatment conditions who identified as "European American". Sample includes only mixed-race participants who indicated two or more racial heritages among their parents. X-axis denotes different treatment groups based on the identity cued by our hypothetical candidates. Note: error bars denote 95% confidence intervals. Table 1. Results from difference-of-means t-tests for all possible treatment group comparisons in Figure 1.

Difference of Means p-Value Difference of Means p-Value
Biracial-Minority +0.07 0.048 * +0.07 0.049 * Biracial-White +0.18 0.000 *** +0.15 0.000 *** Minority-White +0.11 0.000 *** +0.08 0.034 * Under these grouped scenarios, we observe positive and significant treatment effects for candidates who cue a "biracial" identity as compared to those who identify singularly either as White or with their minority racial group. The left panel of Figure 1 shows that candidates who identify as biracial are evaluated more favorably by the mixed-race participants on the feeling thermometer scale than minority-identified candidates (+0.07 out of 1; p < 0.05 [two-sample t-testing confirming difference of means is not equal to zero]), and much more favorably than White-identified candidates (+0.18; p < 0.001). In the right panel, we observe that the mixed-race participants also report a higher likelihood of voting for biracial-identified candidates than minority-identified candidates (+0.07; p < 0.05) or those identifying as White (+0.15; p < 0.001).
Next, we are interested in whether the mixed-race participants might be especially unfavorable toward mixed-race heritage candidates who defect from the norms of hypodescent by identifying as White (H4). Evidence from Figure 1 shows that candidates who identify as White are indeed evaluated the least favorably-the mixed-race participants rate the White-identified candidates less positively than the minority-identified candidates on both the feeling thermometer (−0.11; p < 0.001) and the vote intention scales (−0.08; p = 0.03; support for H4).
So far, four out of four tests confirm that mixed-race voters do indeed have a special preference for candidates who make an effort to signal their common in-group identity as "biracial" (support for H2). Moreover, it is apparent that mixed-race heritage candidates may face risks among mixed-race voters if they do not choose to signal membership as a biracial. This is especially the case if they defy expected norms of loyalty to their minority group by choosing to identify as White. Moreover, six out of six tests confirm that mixed-race participants differentiate between candidates based on the identities they assert (support for H1).
While the initial results indicate that the mixed-race participants have preferences for biracial-identified candidates when the treatment conditions are grouped, we acknowledge that much of way voters feel about mixed-race candidates may be informed by the salience of Barack Obama. To more neatly parse out whether this phenomenon is true for mixed-race heritage candidates in general, the following analysis includes only the three treatment conditions which included a candidate of Asian and White heritage. Moreover, as noted above, our main goal is to explore multiracialism outside of the Black and White paradigm; so, the majority of our total sample was treated with the conditions including these hypothetical candidates with Asian and White heritage. However, in Appendix D we include the results for the treatments including the candidates with Black and White heritage, which should be treated as exploratory at best given that many of the treatment × participant cells are quite small (again, see Appendix C for sample sizes). Figure 2 presents the mean feeling thermometer and vote intention ratings for the three treatment conditions including candidates with Asian and White heritage. Notably, we observe that the candidate who identifies as biracial is still the most favored candidate among the mixed-race participants. The effect of cueing biracial identity on the feeling thermometer rating is +0.09 (p = 0.02) as compared to cueing Asian identity and +0.20 (p < 0.001) as compared to those cueing a White identity. Regarding vote intention, the effect of cueing a biracial identity as compared to identifying as Asian is +0.08 (p = 0.066) and +0.18 (p < 0.001) compared to those identifying as White. Four out of four 4 tests continue to support the common in-group support theory among the mixed-race participants (support for H2).
Another interesting observation in Figure 2 (see also Table 2) is that the mixed-race participants continue to take the chosen identities of each candidate into account when making their evaluations. Again, six out of six possible comparisons confirm this to be the case (though two tests provide limited support [p < 0.1]). This evidence supports our theory that mixed-race voters are likely to perceive race and make categorization decisions based on the chosen and preferred identities of others rather than based on information about racial heritage or family background alone (support for H1). Mean feeling thermometer and vote intention ratings of hypothetical candidates from the Asian and White heritage treatment conditions. Sample includes only mixed-race participants who indicated two or more racial heritages among their parents. X-axis denotes different treatment groups based on the identity cued by our hypothetical candidates. Note: error bars denote 95% confidence intervals. Our next analysis examines whether mixed-race participants will continue to give preference to biracial-identifying candidates even if they do not share the same specific racial heritage. To test this, we again analyze the evaluations of the Asian and White heritage candidate treatments, but we now disaggregate the mixed-race participants by the specific biracial subgroup with which they identify. Specifically, we now subset to only two groups of participants that we call Asian-White-identified participants and Black-White-identified participants, who are the individuals who selected first "Multiracial" on the self-identification question (see Appendix A) and then indicated that they self-identified as both Asian and White or Black and White in the open-ended space to specify their identities. Figure 3 presents the mean scores on the feeling thermometer and vote intention for the Asian-White-identified participants and Black-White-identified participants. We find that the favorability boost given to the biracial-identified candidate from the Asian and White heritage treatment condition found in Figure 2 is largely driven by the Black-White-identified participants. Figure 2. Mean feeling thermometer and vote intention ratings of hypothetical candidates from the Asian and White heritage treatment conditions. Sample includes only mixed-race participants who indicated two or more racial heritages among their parents. X-axis denotes different treatment groups based on the identity cued by our hypothetical candidates. Note: error bars denote 95% confidence intervals.
Another interesting observation in Figure 2 (see also Table 2) is that the mixed-race participants continue to take the chosen identities of each candidate into account when making their evaluations. Again, six out of six possible comparisons confirm this to be the case (though two tests provide limited support [p < 0.1]). This evidence supports our theory that mixed-race voters are likely to perceive race and make categorization decisions based on the chosen and preferred identities of others rather than based on information about racial heritage or family background alone (support for H1). Table 2. Results from difference-of-means t-tests for all possible treatment group comparisons in Figure 2.
Our next analysis examines whether mixed-race participants will continue to give preference to biracial-identifying candidates even if they do not share the same specific racial heritage. To test this, we again analyze the evaluations of the Asian and White heritage candidate treatments, but we now disaggregate the mixed-race participants by the specific biracial subgroup with which they identify. Specifically, we now subset to only two groups of participants that we call Asian-White-identified participants and Black-Whiteidentified participants, who are the individuals who selected first "Multiracial" on the self-identification question (see Appendix A) and then indicated that they self-identified as both Asian and White or Black and White in the open-ended space to specify their identities. Figure 3 presents the mean scores on the feeling thermometer and vote intention for the Asian-White-identified participants and Black-White-identified participants. We find that the favorability boost given to the biracial-identified candidate from the Asian and White heritage treatment condition found in Figure 2 is largely driven by the Black-White-identified participants.
comparisons lend evidence confirming that the mixed-race participants do in fact take candidates' asserted racial identities into account when perceiving race, and those identity cues are the basis for mixed-race individuals' evaluations (support for H1). Lastly, 13 out of 16 tests (comparisons including a candidate who cues their biracial identity) confirm our theory that candidates who cue their biracial identity activate common in-group identity support among their mixed-race participant peers (support for H2). Mean feeling thermometer and vote intention ratings of hypothetical candidates from the Asian and White heritage treatment conditions. Sample includes Asian-White-identified participants (i.e., those who self-identified as Multiracial and used the words "Asian" and "White" to describe their racial identity) and Black-White-identified participants. X-axis denotes different treatment groups based on the identity cued by our hypothetical candidates. Note: error bars denote 95% confidence intervals. Figure 3. Mean feeling thermometer and vote intention ratings of hypothetical candidates from the Asian and White heritage treatment conditions. Sample includes Asian-White-identified participants (i.e., those who self-identified as Multiracial and used the words "Asian" and "White" to describe their racial identity) and Black-White-identified participants. X-axis denotes different treatment groups based on the identity cued by our hypothetical candidates. Note: error bars denote 95% confidence intervals.
In Figure 3, the Asian-White-identified participants are more supportive of the candidates with Asian and White heritage who identified as biracial or as Asian than they are of the White-identified candidate. However, they do not differentiate much in their evaluations of the non-White-identifying candidates. On the other hand, the Black-Whiteidentified participants are most supportive of the candidate with Asian and White heritage who cues their common in-group biracial identity compared to the other candidates. Specifically, the Black-White-identified participants indicate that they are more likely to vote for the candidate with Asian and White heritage who identifies as biracial than their Asianidentified counterpart (+0.19; p < 0.01) and rate them +0.09 more favorably on the feeling thermometer scale, though this difference is not statistically significant.
Eight out of 12 tests in Figure 3 (see also Table 3) continue to confirm that the mixedrace participants take the mixed-race heritage candidates' identity cues into account when making their evaluations (support for H1). We acknowledge that five of these tests only provide limited support (p < 0.1), though we also note that these models are fairly underpowered given the disaggregation of the mixed-race participants. We also note that all difference-of-means values are oriented in the theorized direction. All told, 20 out of 24 comparisons lend evidence confirming that the mixed-race participants do in fact take candidates' asserted racial identities into account when perceiving race, and those identity cues are the basis for mixed-race individuals' evaluations (support for H1). Lastly, 13 out of 16 tests (comparisons including a candidate who cues their biracial identity) confirm our theory that candidates who cue their biracial identity activate common in-group identity support among their mixed-race participant peers (support for H2). Table 3. Results from difference-of-means t-tests for all possible treatment group comparisons in Figure 3.

Single-Race Participants
While it is apparent that the mixed-race participants internalize the preferred racial identities of the mixed-race heritage candidates, we are also interested in measuring whether this tends to be the case for single-race minorities and Whites. Above, we theorize that single-race individuals will be less likely to perceive race based on racial identities chosen by others than mixed-race individuals (H3). On the other hand, it may be the case that single-race minorities and Whites alike do differentiate in their evaluations of mixed-race candidates based on their identity choices if candidates defy the expectations of hypodescent by identifying as White (H4).
To explore these propositions, Figure 4 (see also Table 4) presents the mean ratings of the grouped mixed-race heritage candidate treatment conditions (similar to that of Figure 1 and Table 1) among both the single-race minority and the White participants. At first glance, it is apparent that both single-race minorities and Whites generally do not differentiate in their evaluations of mixed-race heritage candidates based on the different identity choices they assert (supports H3). Moreover, it is perhaps surprising in that we observe the single-race participants to be so unresponsive to the chosen racial identities of others that they show no tendency to reject candidates who defect from practices of hypodescent by identifying as White (evidence rejecting H4). We had theorized that singlerace minorities might perceive White-identified mixed-race heritage candidates as unlikely to support minority-group interests and that Whites may perceive identifying as White as a threat to their hegemonic position and privileges. Indeed, in 11 out of 12 tests, the single-race participants are unresponsive to the identity cues of others (support for H3).
The single test here which provides limited statistical significance suggests that White participants may actually be somewhat more favorable toward mixed-race heritage candidates who identify with their minority racial group. In the left panel of Figure 4, we observe that the White participants are slightly more likely to prefer mixed-race heritage candidates who identify as minorities than those who identify as biracial (+0.06; p < 0.5) on the feeling thermometer scale. This is perhaps evidence that White participants may reward mixed-race heritage candidates who comply with the expectations of hypodescent by identifying singularly as a minority, thus protecting the boundary between Whites and all others. 022, 11, x FOR PEER REVIEW 15 of 31 Figure 4. Mean feeling thermometer and vote intention ratings of hypothetical candidates. All six treatments are represented here and grouped by each identity cue (e.g., "White" denotes candidates from both the Asian and White heritage and the Black and White heritage treatment conditions who identified as "European American". Sample includes only single-race participants who indicated having two parents who share the same race (i.e., either two White parents or two single-race-identified minority parents). X-axis denotes different treatment groups based on the identity cued by our hypothetical candidates. Note: error bars denote 95% confidence intervals.   Finally, Figure 5 (see also Table 5) presents mean feeling thermometer and vote intention scores for the three treatments including candidates with Asian and White heritage among the single-race minority and White participants. Specifically, the White participants continue to rate the Asian-identified candidate +0.06 (p < 0.1) points more favorably on the feeling thermometer scale than the biracial-identified candidate. Again, a possible explanation for this is that Whites feel more affectively positive toward mixed-race heritage candidates who align with their personal expectations about race by adhering to the rules of hypodescent. We point out that Whites' preference for the minority-identified candidates is limited to the feeling thermometer scale, which is more similar to likability and not reflected in Whites' stated desire for voting for said candidates.

Difference of Means P-Value Difference of Means P-Value
Still, we continue to find no evidence that single-race minorities or Whites are especially unfavorable toward mixed-race candidates who identify as White. Sixteen out of 16 tests (eight in Figure 4 and eight in Figure 5) return statistically insignificant difference in Figure 4. Mean feeling thermometer and vote intention ratings of hypothetical candidates. All six treatments are represented here and grouped by each identity cue (e.g., "White" denotes candidates from both the Asian and White heritage and the Black and White heritage treatment conditions who identified as "European American". Sample includes only single-race participants who indicated having two parents who share the same race (i.e., either two White parents or two single-raceidentified minority parents). X-axis denotes different treatment groups based on the identity cued by our hypothetical candidates. Note: error bars denote 95% confidence intervals. Table 4. Results from difference-of-means t-tests for all possible treatment group comparisons in Figure 4. Finally, Figure 5 (see also Table 5) presents mean feeling thermometer and vote intention scores for the three treatments including candidates with Asian and White heritage among the single-race minority and White participants. Specifically, the White participants continue to rate the Asian-identified candidate +0.06 (p < 0.1) points more favorably on the feeling thermometer scale than the biracial-identified candidate. Again, a possible explanation for this is that Whites feel more affectively positive toward mixed-race heritage candidates who align with their personal expectations about race by adhering to the rules of hypodescent. We point out that Whites' preference for the minority-identified candidates is limited to the feeling thermometer scale, which is more similar to likability and not reflected in Whites' stated desire for voting for said candidates.

Difference of Means p-Value Difference of Means p-Value
dates with mixed-race heritage, then we should have seen at least some boost in preference for mixed-race heritage candidates who identify as minorities. The results presented here provide strong evidence that mixed-race individuals do in fact perceive race differently than do single-race individuals. Mixed-race individuals are more likely to rely on the chosen identities of others than are single-race individuals. (Summary: evidence supports H1, H2, and H3 and conditionally rejects H4).

Figure 5.
Mean feeling thermometer and vote intention ratings of hypothetical candidates from the Asian and White heritage treatment conditions. Sample includes only single-race participants who indicated having two parents who share the same race (i.e., either two White parents or two singlerace-identified minority parents). X-axis denotes different treatment groups based on the identity cued by our hypothetical candidates. Note: error bars denote 95% confidence intervals. Table 5. Results from difference-of-means t-tests for all possible treatment group comparisons in Figure 5.  Figure 5. Mean feeling thermometer and vote intention ratings of hypothetical candidates from the Asian and White heritage treatment conditions. Sample includes only single-race participants who indicated having two parents who share the same race (i.e., either two White parents or two single-race-identified minority parents). X-axis denotes different treatment groups based on the identity cued by our hypothetical candidates. Note: error bars denote 95% confidence intervals. Table 5. Results from difference-of-means t-tests for all possible treatment group comparisons in Figure 5. Still, we continue to find no evidence that single-race minorities or Whites are especially unfavorable toward mixed-race candidates who identify as White. Sixteen out of 16 tests (eight in Figure 4 and eight in Figure 5) return statistically insignificant difference in means results between the evaluations of White-identified candidates and others. Moreover, given that the grouped conditions in Figure 4 contain ample amounts of participants, we are confident this is not a sample size issue. These results conditionally reject H4, with the caveat that the mixed-race participants do evaluate White-identified candidates the least favorably.

Difference of Means p-Value Difference of Means p-Value
Lastly, in only two of 24 comparisons do we observe even limited evidence that the single-race participants are responsive to the specific identity cues provided by mixedrace heritage candidates (support for H3). Given that a wealth of literature demonstrates that voters are most supportive of candidates who share their same race, if the singlerace participants were indeed respecting and internalizing the chosen identities of our candidates with mixed-race heritage, then we should have seen at least some boost in preference for mixed-race heritage candidates who identify as minorities. The results presented here provide strong evidence that mixed-race individuals do in fact perceive race differently than do single-race individuals. Mixed-race individuals are more likely to rely on the chosen identities of others than are single-race individuals. (Summary: evidence supports H1, H2, and H3 and conditionally rejects H4).

Discussion
Following the historic election of Barack Obama in 2008, many in both the media and in academic research were quick to point out how the presence of the first African American candidate for president in the general election motivated turnout for minorities and garnered high approval among racially progressive Whites (Tesler and Sears 2010;Segura and Bowler 2012;ABC News 2008). However, many were also quick to argue that Barack Obama was successful not because he was the first Black candidate on the presidential ballot but because he was adept at drawing on his mixed-race heritage to elicit support from both minorities and Whites. While Barack Obama-like many mixedrace heritage candidates-personally identifies singularly as Black, an oft-asked question surrounding his candidacy has been whether Obama would have been as successful if he had chosen to identify as "biracial". Similarly, it had been an unanswered question whether voters would be responsive to different identity choices made by candidates with interracial family backgrounds, more and more of whom are becoming prominent political figures in United States politics today.
The purpose of the study has been to investigate the role of identity assertion on impressions of mixed-race heritage political candidates. Specifically, we investigate how individuals from three different racial groups (mixed-race heritage individuals, single-race minorities, and single-race Whites) respond to candidates with known mixed-race heritage but who vary in the racial labels they use to describe themselves. We ultimately sought to determine if participants from some racial groups were more likely than others to rely on the mixed-race heritage candidates' stated racial identities or if voters were typically only responsive to information about a candidate's racial heritage.
Overall, the results of this study demonstrate robustly that mixed-race individuals tend to differ sharply from both single-race minorities and single-race Whites in that they are more likely to rely on and internalize the racial identity choices asserted by mixed-race heritage candidates in their evaluations. These findings reconcile well with the previous literature which argues that mixed-race individuals exist on the frontier of a new paradigm of racial classification, one that views race and identity as flexible and as a choice rather than as immutable and ascribed (Masuoka 2017;Does et al. 2021;Sanchez et al. 2009). As a consequence, mixed-race individuals seem more readily available to internalize the chosen racial identities of others and to use those identities cues as the platform upon which they make their categorizations and evaluations.
Next, our results also demonstrate that mixed-race participants were most favorable toward mixed-race heritage candidates who identified specifically as biracial, both in their evaluations on a feeling thermometer scale and in their stated interest in voting for said candidates. These results are consistent with both the common in-group identity theory (Gaertner and Dovidio 2014) and social identity theory (Tajfel and Turner 1986). Additionally, we uncovered that the mixed-race participants often extend their support for biracial-identified candidates even when the voter and the candidate do not share the same exact mixed-racial heritage. For example, Black-White-identified biracial participants were found to be highly supportive of the biracial-identified candidate with Asian and White heritage, though they do not share minority racial group ties.
Another important finding is that mixed-race participants show signs of continued loyalty to minority group interests when faced with the option of supporting mixedrace candidates who identify either singularly as White or with their minority racial group. Specifically, mixed-race individuals were highly unfavorable toward mixed-race heritage candidates who defied the expectations of hypodescent and chose to assert their identification with the majority group: Whites. This evidence suggests that while mixedrace heritage candidates may earn a boost in support among mixed-race voters when they signal their shared identity as a biracial, they may face risks when they identify with a single race, especially when they identify as White.
While mixed-race individuals differed greatly in their evaluations of candidates based on identity cues, we find little to no evidence that single-race minorities or Whites do the same. One exception, however, is that single-race Whites express a slightly higher willingness to vote for mixed-race heritage candidates who identify singularly with their minority group, thus satisfying the expectations of hypodescent. To elaborate on this finding, we note that important scholarship from Critical Mixed-Race Studies traces the roots of modern perceptions of mixed-race individuals back to Whites' intentional separation of Blacks according to the degree to which they were racially mixed with Whites during slavery (Curington 2016(Curington , 2021. Laws against miscegenation also played an important role in preserving White privilege and solidifying the perception that individuals with any minority heritage are de facto minorities, especially if they are part Black. Given this history, it follows logically that Whites may reward mixed-race individuals who "follow the rules" and identify as minorities. On the other hand, our finding that single-race Whites prefer candidates with Asian and White heritage, in particular those who comply with hypodescent, is a bit at odds with recent studies. Recent work argues that modern America's tolerance of multiracialism is partly a consequence of increasing rates of intermarriage between non-Black minorities and Whites, especially those unions between Asians and Whites, which are perceived as less taboo (Curington 2016). Indeed, many scholars have surmised that the children of Asian and White interracial parents show signs of having assimilated with Whites so well that they may be considered "White enough" (Strmic-Pawl 2016; Alba 2020; Leslie and Sears 2022). So, while Americans might herald the increasing prominence of mixed-race elected officials such as Barack Obama and Kamala Harris as symbols of a post-racialism and racial reconciliation (Strmic-Pawl 2014), our evidence suggests that the regressive constraints on mixed-race heritage individuals' identities remain quite intact.
Overall, this study represents a substantial contribution to several social science literatures. While some psychology research has examined how biracial identity labels influence how individuals are perceived (Remedios et al. 2012), this research is one of the few studies to examine how the complexities of mixed-race identity affect how political candidates are evaluated by potential voters (See also Lemi 2021 andClayton et al. 2021). Moreover, the research suggests that hypodescent is often applied in judgments of racial category membership (Ho et al. 2011). Therefore, the present research also makes a contribution by showing that, in addition to relying on visual appearance or phenotype when racially classifying others (e.g., Sanchez et al. 2011;Skinner and Nicolas 2015), social perceivers may consider factors beyond appearance, including identity assertion, to form race-based impressions of others. Additionally, this is one of the few lines of research to also include a sample of mixed-race participants as a comparison group to both single-race White and minority participants. Although a recent Pew report highlights some of the voting behaviors of mixed-race individuals (Parker et al. 2015), we did not know how mixed-race individuals responded to race within elections. This is some of the first evidence to directly compare different racial groups on how they perceive different races of political candidates.
Additionally, this project is an important extension to research in political science which tends to assume that a candidate's race or ethnicity is automatically identifiable. The extant research on descriptive representation has shown that it leads to increased participation in democratic governance among people of color, which is of major importance for advancing the interests of historically marginalized groups (Mansbridge 1999;Stout et al. 2021). However, given that our understandings of race and racial identity are becoming more flexible, we extend these types of work by providing important introspection into how descriptive representation may operate when it concerns candidates with mixed-race heritage. Specifically, we extend these theories in two ways. First, we do so by substantiating that mixed-race heritage candidates may be limited in their ability to leverage their dual-racial group membership to activate feelings of common in-group favoritism among both single-race White and single-race minority group voters. Second, we demonstrate that there may in fact be some propensity for mixed-race candidates to achieve descriptive representative status among mixed-race voters.
This study also poses major implications for non-academic practitioners. Specifically, pundits and campaign workers who speculated that Obama could have been even more successful in his campaigns by identifying as "biracial" or "Multiracial" may be cautioned by our research. While we find that mixed-race heritage candidates may earn additional support among mixed-race voters by emphasizing their mixedness, doing so is unlikely to matter much to single-race voters or may even cost them votes, at least among Whites. These results square well with the findings of others, specifically Lemi (2021), who finds that that single-race voters prefer single-race candidates to mixed-race candidates (when voters share some racial heritage with the candidate). Political strategists and policy makers should therefore look to the specific demographic make-up of their constituency when evaluating how different identity cues may affect electoral or policy support. As the proportion of mixed-race voters rapidly expands, it may become beneficial for candidates to target those groups as a voting bloc by emphasizing their mixed-race identities.
While this study represents an important advancement, more work is needed. First, future work should also explore how the gender of both the participants and the candidates may affect responses to identity cues. From an intersectional perspective, existing research demonstrates that the mixed-race experience is highly gendered. Mixed-race women are known to be more likely than men to be appraised as exotic and more likely to be valued (or devalued) for having Eurocentric features, and as such, they are more likely to experience discrimination and exclusion from their minority group (Sims and Joseph-Salisbury 2019;Buggs 2017aBuggs , 2017bBuggs , 2019Strmic-Pawl 2014Curington et al. 2015Curington et al. , 2020. On the other hand, mixed-race men are more likely to experience traditional forms of racial discrimination which reinforces their minority status and constrains their ability to identify as anything other than single-race minorities (Strmic-Pawl 2016;Davenport 2016). Future studies can more deeply examine whether women's proclivity to identify as mixed-race leads to increased enthusiasm for mixed-race candidates. An exploratory analysis using our data suggests that, yes, women may be more likely than men to extend favorable evaluations of biracial-identifying candidates, but much more data collection is necessary to sufficiently examine this prospect. 5 Additionally, more work can be conducted to investigate the specific mechanisms underlying our observed relationship between identity cues and voters' evaluations. Our study theorized that mixed-race voters will favor biracial-identified candidates because they signal a common in-group identity. Future studies could more neatly parse out whether feelings of common identity mediate or moderate this relationship of support. This could be achieved by including items that measure linked fate, group consciousness, identity centrality, or the degree to which they perceive sameness with the biracial-identified candidate. 6 Moreover, existing research shows that racial resentment is a strong predictor of how voters evaluate candidates with minority heritage (Tesler and Sears 2010) and that those who have essentialist conceptions of race are more likely to categorize individuals according to hypodescent (Bonam and Shih 2009;Shih et al. 2007). Future work should therefore explore whether single-race voters' feelings of racial resentment or tendencies to hold essentialized perceptions of race moderate their response to identity cues.
Additionally, our study could be extended to include both participants and candidates with other mixed-race family backgrounds. Studies should extend this line of work to examine how candidates such as those with Latino and White heritage are perceived by voters and even to dual-minority candidates, such as those with Black and Asian or Asian and Latino heritage. Moreover, we are also conscious that the degree to which individuals' skin color and phenotype match those prototypical of certain racial groups contours the ways that they are racially categorized and perceived by others (Ho et al. 2011;Ho et al. 2015). While our study elides these concerns by not including pictures for each candidate, future studies could take advantage of modern methodological advances (Yadon and Ostfeld 2020) to more deeply explore how variation in skin color and phenotype affects voters' responses to identity cues. It is possible that the abilities of candidates with mixedrace heritage to espouse different racial identities may be more constrained as their visual appearance increasingly matches the expected characteristics of a single racial group.
Finally, we wish to situate this current study within the growing body of scholarship referred to as Critical Mixed-Race Studies (CMRS). Importantly, CMRS scholars have underscored that research on mixed-race individuals and the mixed-race experience has been particularly disjointed in that scholars have predominantly approached these subjects from the silos of their own disciplines (Daniel et al. 2014). Lack of cross-disciplinary communication, as well as a lack of methodological diversity, poses the risk of limiting scholarly achievement on this topic. To this point, we note this study is a collaboration between authors based in political science and psychology. Our aim with theory building has been to incorporate literature predominantly from political science, psychology, and sociology, as well as to acknowledge both the sociohistorical precursors to mixed-raceness as we know it today and the modern accounts. Additionally, we hope that future research extends beyond our methodological focus on quantitative measurement by using qualitative or mixed-methodologies to further engage in this line of inquiry. As the proportion of mixed-race voters continues to grow and more mixed-race heritage individuals run for office, it is imperative that scholars rise to the challenge and continue exploring the ways this unique group will impact the future of American politics. Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.

Data Availability Statement:
The data presented in this study are available upon request from the first author. On a scale of 1 to 10, please rate how you feel toward the candidate. Ratings between 1 and 4 mean that you feel unfavorable or cool toward the candidate. Ratings from 6 to 10 mean that you feel favorable or warm toward the candidate. A rating of 5 means you feel neither warm nor cool toward the candidate.

Conflicts of
Vote Intention: If you were eligible to vote, how likely is it that you would vote for this candidate?
− Very unlikely (1) − Somewhat unlikely (2) − Somewhat likely (3) − Very likely (4) Biography: Tom Smith, age 45, was born and raised in a small city in the northeast. He is the firstborn son of three children to a mother of European American heritage and a father of African American heritage. He is proud of his heritage and describes himself as a "European American". He is a product of the public school system and went on to earn his bachelor's degree in political science from New York University and a law degree from Georgetown University. Smith has had a lifelong career devoted to public service. He began his career serving on his city's Energy Conservation Commission which oversees oil, electric, and water companies. At the age of 32, he was elected to the city council. Five years later, Smith beat a sitting incumbent and was elected as a representative in the state legislature. Smith and his wife have three children. Table A1. Results from difference-of-means t-tests comparing support for biracial-identified candidates among biracial-identified mixed-race heritage participants and single-race-identified mixed-race heritage participants. Includes only mixed-race participants (those indicating mixed-race heritage by parentage) and only those who evaluated biracial-identified candidates. Results show that both those who we identified as mixed-race by heritage and those who actively identified as "Multiracial" have indistinguishable point estimates in terms of their evaluations of biracial-identified candidates.

Difference of Means p-Value Difference of Means p-Value
Multiracial ID-Single-Race ID 0.00 0.963 +0.01 0.878 Note: + p < 0.1; * p < 0.05; ** p < 0.01; *** p < 0.001.   Appendix D Figure A1. Mean feeling thermometer and vote intention ratings of hypothetical candidates from the Black and White heritage treatment conditions. Sample includes only mixed-race participants who indicated two or more racial heritages among their parents. X-axis denotes different treatment groups based on the identity cued by our hypothetical candidates. Note: error bars denote 95% confidence intervals. Table A4. Results from difference-of-means t-tests for all possible treatment group comparisons in Figure A1. Note: + p < 0.1; * p < 0.05; ** p < 0.01; *** p < 0.001. Figure A1. Mean feeling thermometer and vote intention ratings of hypothetical candidates from the Black and White heritage treatment conditions. Sample includes only mixed-race participants who indicated two or more racial heritages among their parents. X-axis denotes different treatment groups based on the identity cued by our hypothetical candidates. Note: error bars denote 95% confidence intervals. Table A4. Results from difference-of-means t-tests for all possible treatment group comparisons in Figure A1.  Figure A2. Mean feeling thermometer and vote intention ratings of hypothetical candidates from the Asian and White heritage treatment conditions. Sample includes Asian-White-identified participants (i.e., those who self-identified as Multiracial and used the words "Asian" and "White" to describe their racial identity) and Black-White-identified participants. X-axis denotes different treatment groups based on the identity cued by our hypothetical candidates. Note: error bars denote 95% confidence intervals. Table A5. Results from difference-of-means t-tests for all possible treatment group comparisons in Figure A2. Note: + p < 0.1; * p < 0.05; ** p < 0.01; *** p < 0.001. Figure A2. Mean feeling thermometer and vote intention ratings of hypothetical candidates from the Asian and White heritage treatment conditions. Sample includes Asian-White-identified participants (i.e., those who self-identified as Multiracial and used the words "Asian" and "White" to describe their racial identity) and Black-White-identified participants. X-axis denotes different treatment groups based on the identity cued by our hypothetical candidates. Note: error bars denote 95% confidence intervals. Table A5. Results from difference-of-means t-tests for all possible treatment group comparisons in Figure A2. Note: + p < 0.1; * p < 0.05; ** p < 0.01; *** p < 0.001. Table A6. Results from difference-of-means t-tests for all possible treatment group comparisons in Figure A3. Note: + p < 0.1; * p < 0.05; ** p < 0.01; *** p < 0.001. Figure A3. Mean feeling thermometer and vote intention ratings of hypothetical candidates from the Black and White heritage treatment conditions. Sample includes only single-race participants who indicated having two parents who share the same race (i.e., either two White parents or two single-race-identified minority parents). X-axis denotes different treatment groups based on the identity cued by our hypothetical candidates. Note: error bars denote 95% confidence intervals.

Single-Race-Minority Participants
Appendix E Table A7. Results from difference-of-means t-tests comparing support for biracial-identified candidates among women and men. Includes only mixed-race participants and only those who evaluated biracial-identified candidates. Results show that women are slightly more likely (+0.16; p = 0.072) to indicate that they would vote for the biracial-identified candidates than are men.

Appendix F
Though our design did not include items to measure linked fate or group consciousness specifically, we did include items which measure (1) the degree to which participants identify strongly with their race and (2) the degree to which participants feel their racial group is important to them. These items are meant to be interpreted as exploratory at best and find overall that identity strength and racial group importance are positively correlated (but statistically non-significant) with evaluations of biracial-identifying candidates among mixed-race participants.
Item Information: Identity Strength: "On a scale of 1 (not at all) to 7 (very much), how much do you identify with the racial group you selected above?" Coding: Strong Identifier = 7, 6, or 5; Weak Identifier = 4, 3, 2, or 1 Racial Group Importance: "On a scale of 1 (not at all) to 7 (very much), how important is the racial group you selected above to you?" Coding: Weak Identifier = 4, 3, 2, or 1; Low Racial Group Importance = 4, 3, 2, or 1 Table A8. Results from difference-of-means t-tests between high racial group identifiers and weak identifiers. Includes only mixed-race participants. Results indicate that those who strongly identify with their racial group and those who believe that their racial group is important to them are more likely to support the biracial-identified candidates, though these results are not statistically significant. Note: + p < 0.1; * p < 0.05; ** p < 0.01; *** p < 0.001.

1
We use the words "mixed-race", and "biracial" interchangeably to describe the population of individuals who have or acknowledge two or more races in their ancestry. "Biracial" in particular refers to those with who feel attached to two specific heritages (e.g., Black-Whites, Latino-Whites, etc.). We use the term "single-race" to designate those individuals who identify with only one racial group.
2 Indeed, in Appendix B we compare these two groups and find no difference in their opinions. 3 We began collecting data on 1 April 2013 and on 13 May 2013 expanded our study to emphasize the inclusion of treatments with Black and White heritage candidates. 4 However, one test (Biracial-Asian on the vote intention scale) should be interpreted as weak support given the p-value of 0.067.

5
In Appendix E, we conduct an exploratory analysis of this question and find suggestive evidence that women may in fact be more likely to vote for biracial-identified candidates than men.