INTRODUCTION
A central question in the study of group attitudes is how people include, or exclude, others. Of particular interest is the process of categorization and who is considered a member of the in-group. While many studies examine the attitudes of the most prototypical group in shaping the boundaries of inclusion, less work compares them to less prototypical groups. While peripheral group members still have a claim to that particular group identity, their membership is itself precarious. Scholarship shows the importance of dominant groups in maintaining group-based hierarchies (Sidanius et al. Reference Sidanius, Feshbach, Levin and Pratto1997), yet less is known about whether less prototypical group members help to sustain group-based hierarchies.
When considering how peripheral group members compare to those who are prototypical in their attitudes toward group boundaries, there are a few possibilities. One is that due to their marginal position within that group, peripheral group members may draw boundaries in a similar fashion as prototypical members to enhance their sense of membership within that group (Ellemers and Jetten Reference Ellemers and Jetten2013). Indeed, scholars have shown that when their position within that group is threatened, peripheral group members express more restrictive attitudes (Pérez and Kuo Reference Pérez and Kuo2021; Pérez, Robertson, and Vicuña Reference Pérez, Robertson and Vicuña2023). Another possibility is that peripheral group members may form a bond with others who are on the fringe of the group over a shared sense of discrimination and therefore develop more positive attitudes of the group in question compared to prototypical group members (Pérez, Vicuña, and Ramos Reference Pérez, Vicuña and Ramos2023).
Considering the racial hierarchy in the United States as an example of one such socially constructed group-based hierarchy, Black Americans are viewed as relatively inferior yet still insiders, while Latinos are both inferior and foreigners (Kim Reference Kim1999; Zou and Cheryan Reference Zou and Cheryan2017). Whites are viewed as the most prototypically American group (Devos and Banaji Reference Devos and Banaji2005), whereas the relationship that Black Americans have toward their American identity is more complicated (Carter Reference Carter2019; Carter and Pérez Reference Carter and Pérez2016). Given their long-standing treatment as second-class citizens (Collins Reference Collins2001), Black Americans are aware of their position as “outsiders within” (Carter Reference Carter2019), yet still value their American identity (Masuoka and Junn Reference Masuoka and Junn2013). Historically, their Americanness has served as a premise for claiming equal rights, illustrating the significance of the identity for the group (Carter Reference Carter2019). To answer the question of how prototypical group member attitudes compare to those who are less so, on the question of how the boundaries of the superordinate group are defined (Dovidio, Gaertner, and Saguy Reference Dovidio, Gaertner and Saguy2009; Hornsey and Hogg Reference Hornsey and Hogg2000), I compare white and Black American perceptions of whether Latinos are American.
Beyond exploring the relationship between prototypical and peripheral members, studying perceptions of who is an American has important implications for understanding who is considered part of the national imagined community. Perceiving commonality is an important component of improving overall attitudes (Danbold and Huo Reference Danbold and Huo2022; Hornsey and Hogg Reference Hornsey and Hogg2000). The contours of national identity are specifically tied to the allocation of social citizenship (Bloemraad et al. Reference Bloemraad, Kymlicka, Lamont and Son Hing2019), as national identity symbolizes a group bound together by shared beliefs, characteristics, and obligations (Miller Reference Miller1999; Theiss-Morse Reference Theiss-Morse2009). As a group who has historically been excluded from definitions of who is an American, whether Latinos are viewed as American signifies their full political inclusion.
The heterogeneity of the Latino population complicates this question, yet it has been underexplored systematically across various markers of difference. As a group with diverse physical appearance, immigration, and socioeconomic statuses, among other differences, it is likely that attitudes toward Latinos will be contingent on these varying characteristics. Different types of Latinos will likely be perceived differently by white versus Black Americans.
I draw on a survey experiment of white and Black Americans to examine their perceptions of whether Latinos are considered American. The conjoint survey experimental design allows me to manipulate multiple characteristics—including photos to illustrate physical appearance, immigrant background and legal status, education, occupation, language ability, ethnic/national identification, religion, and political affiliation. Specifically, the conjoint experimental design allows me to disaggregate the category of “Latino” to understand the complexities of attitudes toward the group. By using photos to reflect the diversity of Latinos, I expand how survey experiments traditionally evoke racial categorization using words (Abrajano, Elmendorf, and Quinn Reference Abrajano, Elmendorf and Quinn2018) to more realistically capture how a person’s appearance can impact attitudes (Dowling Reference Dowling2014; Roth Reference Roth2010).
I find that perceptions of Latinos are shaped by a series of ascriptive characteristics and that the negative effect of legal status transcends intergenerationally. In addition, this paper also finds that white and Black Americans draw different boundaries of the national imagined community. Black respondents have more positive perceptions of undocumented people, while whites valorize whiteness, greater adaptation, and higher socioeconomic status. On the whole, the most important characteristic in assessing perceptions of who is American for Black respondents is political affiliation, whereas for whites, it is language and legal status. Whites’ perceptions of the boundaries of the national imagined community are strongly shaped by their attachment to their national identity—where strong attachments are associated with drawing more restrictive boundaries—whereas this is not the case for Black respondents. On the contrary, when examining the relationship between attachment to racial identity and perceptions of who is an American among Black respondents, strong identifiers draw less restrictive boundaries than weak identifiers.
These findings support existing theories that suggest that whites’ perceptions are driven by cultural threat, whereas Black respondents’ perceptions are shaped by economic concerns. The results suggest that higher socioeconomic status individuals are a greater concern for Black respondents. Overall, the findings suggest that Black Americans view those with experiences of marginalization as more American. In addition, the findings demonstrate that Black respondents’ perceptions are driven by a sense of political threat, above other considerations, which has been underexplored. Black perceptions also challenge long-standing results about attitudes toward immigrants—notably the preference for higher socioeconomic status (Hainmueller and Hiscox Reference Hainmueller and Hiscox2010; Hainmueller and Hopkins Reference Hainmueller and Hopkins2015) and the categorical rejection of undocumented people (España-Nájera and Vera Reference España-Nájera and Vera2020; Wright, Levy, and Citrin Reference Wright, Levy and Citrin2016).
The findings also have broader implications for politics. While studies investigating racial attitudes have advanced our understanding of intergroup relations, we know less about the complexities of acceptance for a group that is as heterogeneous as Latinos (Beltrán Reference Beltrán2010). Indeed, that whites view those who do not threaten their group’s prototypicality more positively, while Black respondents tend to view those in a marginalized position more favorably, suggests a substantial divergence in which Latinos are deemed as American by each group. As studies have shown that prototypicality threat among whites leads to support for nativist policies (Danbold, Serrano-Careaga, and Huo Reference Danbold, Serrano-Careaga and Huo2023), and inducing a shared sense of marginalization leads to greater support for pro-Latino policies among Black Americans (Pérez, Vicuña, and Ramos Reference Pérez, Vicuña and Ramos2023), this, in turn, helps advance our understanding of under which conditions policy attitudes are applicable, and whether the underlying sentiments are targeted toward subsets of the Latino population.
Where Do Latinos Fit Within American Identity?
A cohesive sense of nationality emerges from individuals perceiving shared characteristics, including a sense of national culture, with others in the group (Miller Reference Miller1999). One of the conundrums of American identity is that it symbolizes a diverse set of people that make a cohesive “we.” Studies highlight the perception that immigrants are important to the nation (Schildkraut Reference Schildkraut2011) and the endorsement of civic rather than ethnocultural norms (Citrin, Reingold, and Green Reference Citrin, Reingold and Green1990; Citrin, Wong, and Duff Reference Citrin, Wong, Duff, Ashmore, Lee and David2001; Schildkraut Reference Schildkraut2007; Wright, Citrin, and Wand Reference Wright, Citrin and Wand2012). At the same time, studies document the valorization of ethnic proximity to whiteness (Devos and Banaji Reference Devos and Banaji2005; Huynh, Devos, and Altman Reference Huynh, Devos and Altman2015; Rydell, Hamilton, and Devos Reference Rydell, Hamilton and Devos2010), suggesting conflicting notions of who is American.
Considering American identity is crucial because of how it marks an identity from which Latinos are often excluded (Bloemraad Reference Bloemraad2022; Flores-González Reference Flores-González2017; Roth Reference Roth2012). Latino racialization is tied to foreignness and illegality; they are often denied cultural membership into the American imagined community (De Genova Reference De Genova2005; Rocco Reference Rocco2014). Where Latinos fit within the racial hierarchy, in relation to both white and Black Americans, is complicated because of the vast differences between Latinos. The persistence of the label “Hispanic/Latino” represents a racialized ethnic category for the group (Flores-González Reference Flores-González2017; Golash-Boza Reference Golash-Boza2006; Roth Reference Roth2012), suggesting potential proximity to Black Americans as ethnoracial minorities, yet there is evidence of assimilation into whiteness by some Latinos (Alba and Islam Reference Alba and Islam2009). Even so, their categorization is complicated by how others view them (Dowling Reference Dowling2014; Vargas Reference Vargas2015). Yet scholars seldom examine how the group’s heterogeneity (Jones-Correa et al. Reference Jones-Correa, Marrow, Okamoto and Tropp2018) impacts attitudes toward the group. Specifically, the role of physical appearance in shaping how Latinos are perceived by others has not received sufficient attention (Abrajano, Elmendorf, and Quinn Reference Abrajano, Elmendorf and Quinn2018), despite the importance of skin color along other dimensions (Golash-Boza and Darity Reference Golash-Boza and Darity2008; Ostfeld and Yadon Reference Ostfeld and Yadon2022). Typically, skin color has been assessed using the Massey-Martin scale, which rates skin color from zero to ten using a picture of a hand (Massey and Martin Reference Massey and Martin2003). While recent improvements have been made to this scale (Ostfeld and Yadon Reference Ostfeld and Yadon2022), using a hand to depict skin color is a less realistic characterization of how people form stereotypes (Blair, Judd, and Fallman Reference Blair, Judd and Fallman2005).
Consequently, I begin by exploring how different characteristics influence perceptions of which Latinos are American among all respondents. Table 1 highlights key expectations and their corresponding findings. First, research on the role of ascriptive characteristics is mixed. While some research finds that Americanness is associated with whiteness (Devos and Banaji Reference Devos and Banaji2005; Huynh, Devos, and Altman Reference Huynh, Devos and Altman2015; Rydell, Hamilton, and Devos Reference Rydell, Hamilton and Devos2010), other scholars show that ethnocultural components, including being white, are deemed as less important (Schildkraut Reference Schildkraut2007). Nonetheless, scholars have seldom examined physical appearance through photos, which can help to strengthen the external validity of survey experiments.
H1: Footnote 1 Phenotypically non-white profiles will be seen as less American.
H2: American-identifying profiles will be seen as more American.
Table 1. Expectations and Findings
Beyond ascriptive characteristics, other forms of heterogeneity among Latinos include immigration status. Unauthorized status is heavily penalized in perceptions of immigrants (Flores and Schachter Reference Flores and Schachter2018; Hartman, Newman, and Bell Reference Hartman, Newman and Bell2014; Schachter Reference Schachter2016). However, we know much less about how the descendants of immigrants are viewed, despite only 33 percent of Latinos being immigrants as of 2019 (Funk and Lopez Reference Funk and Lopez2022). By definition, being born in America is the primary way that the U.S. confers citizenship (Schildkraut Reference Schildkraut2007). Yet, discussions of whether birthright citizenship should be revoked for the children of undocumented immigrants have been popularized in general media (Chavez Reference Chavez2020). If exclusion is extended beyond the legal criteria of being American (naturalization or birth in the U.S.), then this would suggest how social membership operates beyond this definition.
H3: The undocumented will be seen as less American compared to the grandchildren of legal immigrant parents. The children of the undocumented immigrants will be seen as less American as well. However, the other categories of immigrants or descendants of immigrants should not experience any penalty.
How Do White and Black Attitudes Differ?
A crucial factor to consider regarding how superordinate identity is defined is how this process occurs for those who are less prototypical members of the group, where we can expect the subjective meaning of the identity to vary (Huddy Reference Huddy2001). Prototypical group members want to maintain their advantaged status (Danbold and Huo Reference Danbold and Huo2022; Danbold, Serrano-Careaga, and Huo Reference Danbold, Serrano-Careaga and Huo2023; Hornsey and Hogg Reference Hornsey and Hogg2000; Ostfeld and Yadon Reference Ostfeld and Yadon2022), while peripheral group members do not always seek to improve their relative position within the group. Peripheral members react to their marginality differently; while some want greater inclusion, others are satisfied with their peripheral status or seek alternative group memberships (Dovidio, Gaertner, and Saguy Reference Dovidio, Gaertner and Saguy2009; Ellemers and Jetten Reference Ellemers and Jetten2013). The ingroup projection model also posits that subgroups generally perceive themselves as relatively more prototypical of the superordinate category, even when their subgroup is viewed as less prototypical by others (Wenzel, Mummendey, and Waldzus Reference Wenzel, Mummendey and Waldzus2007), and they often “project” their subgroup’s characteristics onto how they perceive the superordinate group. Thus, we can expect that white and Black Americans’ different positions within the group will result in diverging attitudes.
In the context of American identity specifically, Black Americans’ status has been precarious since the country’s founding. For centuries, Black Americans were denied citizenship rights on the basis of not being white (Haney-López Reference Haney-López2006). Even after they were granted equal protection under the law, Black Americans continued to be disenfranchised and excluded from the benefits of being American (Crooms-Robinson Reference Crooms-Robinson, Bobo, Crooms-Robinson, Darling-Hammond, Dawson, Gates, Jaynes and Steele2012). Du Bois illustrated this complicated relationship as double consciousness—where Black Americans are caught between the “twoness” of their “unreconciled” identities (Du Bois Reference Du Bois1909). Specifically, double consciousness illustrates the attempt at reconciling their national identity with their blackness, despite consistently being excluded from being American.
Despite Black Americans’ struggle for inclusion and the juxtaposition between their identities, racial discrimination does not dampen identification as American for the group (Greene et al. Reference Greene, Gray, Carter and Block2020), with other research suggesting similarities in the importance of American identity with whites. Black Americans perceive themselves to be similarly prototypically American as whites (Masuoka and Junn Reference Masuoka and Junn2013) and are surprisingly similar to whites in their endorsement of national identity norms (Tafoya, Corral, and Leal Reference Tafoya, Corral and Leal2022). Black Americans have some of the most rigid criteria for the important characteristics of an American, suggesting the potential for the group to express restrictive attitudes (Masuoka and Junn Reference Masuoka and Junn2013). One key difference between the two groups is that whites are the most likely among racial groups to identify as American, followed by Black Americans (Greene et al. Reference Greene, Gray, Carter and Block2020).
When considering how these two groups develop attitudes toward Latinos, research indicates that whites perceive cultural threat from those who are dissimilar to them (Danbold and Huo Reference Danbold and Huo2015; Danbold and Huo Reference Danbold and Huo2022; Danbold, Serrano-Careaga, and Huo Reference Danbold, Serrano-Careaga and Huo2023; Lacayo Reference Lacayo2017; Zou and Cheryan Reference Zou and Cheryan2022), and they restrict the boundaries of whiteness under threat (Abascal Reference Abascal2020). The literature on attitudes toward immigration policy, which has largely focused on whites, explores both economic and cultural motivations, finding that cultural rather than economic considerations drive white attitudes (Hainmueller and Hopkins Reference Hainmueller and Hopkins2015).
Existing literature outlines two possibilities for Black attitudes. They may express more positive attitudes than whites, given that they are also considered to be on the periphery of this category and may perceive a shared sense of discrimination (Brown, Jones, and Becker Reference Brown, Jones and Becker2018; Craig and Richeson Reference Craig and Richeson2012; Frasure-Yokley and Greene Reference Frasure-Yokley, Greene, Kun and Pulido2014; Pérez, Vicuña, and Ramos Reference Pérez, Vicuña and Ramos2023). On the other hand, they may express similarly restrictive attitudes as whites in an attempt to stake a greater claim to being an American compared to Latinos (Pérez and Kuo Reference Pérez and Kuo2021), as some work suggests that Black Americans do not perceive Latinos to be a minority group like them and do not view much commonality with the group (Wilkinson Reference Wilkinson2015). Black Americans can be characterized as “conflicted nativists,” where despite their usage of national identity as leverage to claim equal rights, they generally do not express the same policy-restrictive attitudes toward immigration (Carter Reference Carter2019). Specifically, Carter argues that Black opinion on immigration is best characterized as ambivalent: while Black Americans are hesitant to embrace positions on immigration rooted in the same racism that impacts their own group, they do not want the inclusion of immigrants to come at their group’s expense (Carter Reference Carter2019). In addition, Black Americans perceive immigration as an act of self-determination, which they are sensitive to because of their own group’s pursuit for equality, yet they do not want this to come at the expense of their advancement “socially, politically, or economically” (Carter Reference Carter2019, 16). Because of these conflicting findings, the question remains whether Black Americans will see Latinos as a group with whom they have common experiences or whether they will view them as competition.
Given the evidence suggesting the role of cultural threat in impacting white attitudes, which is specifically associated with the loss of their prototypical status (Danbold and Huo Reference Danbold and Huo2022), we can expect that the characteristics that will most threaten whites will include ascriptive characteristics, which are the most rigid boundaries that characterize the exclusivity of the group. On the other hand, Black ambivalence toward immigrants does not arise from cultural threat, suggesting an important divergence between white and Black attitudes (Carter Reference Carter2019).
H4a: White respondents will rate profiles that are phenotypically white as more American than Black respondents.
H5a: White respondents will rate profiles that identify as American or as white as more American than Black respondents.
H4b: Physical appearance will be a more salient characteristics for white respondents than for Black respondents.
H5b: Ethnic or national identification of the profile will be a more salient characteristic for white respondents than for Black respondents.
The racial attitudes’ literature that examines Black-Latino relations largely focuses on economic explanations—particularly, Latino labor as a source of threat for Black jobs, as one of the driving factors causing friction between the two groups (Browne, Deckard, and Rodriguez Reference Browne, Deckard and Rodriguez2016; Gay Reference Gay2006; McClain et al. Reference McClain, Lyle, Carter, DeFrancesco Soto, Lackey, Cotton and Nunnally2007; Wilkinson Reference Wilkinson2015). While some scholars argue that direct economic competition is not a factor in white attitudes (Hainmueller and Hiscox Reference Hainmueller and Hiscox2010), studies that directly compare white and Black attitudes toward immigration find that zero-sum competition applies similarly to both groups (Hutchings and Wong Reference Hutchings and Wong2014). Therefore, we can expect a possibility of economic perceptions playing a similar role among both groups.
H6a: White and Black respondents will rate education levels as equally American.
H7a: White and Black respondents will rate occupation levels as equally American.
H6b: There will not be a difference in the salience of education between white and Black respondents.
H7b: There will not be a difference in the salience of occupation between white and Black respondents.
The economic explanations for Black-Latino conflict focus on the impact of the lower end of the labor spectrum—immigrants, particularly the undocumented (Wilkinson Reference Wilkinson2015). Therefore, unauthorized status can be considered a form of economic threat for Black respondents (Brown, Jones, and Becker Reference Brown, Jones and Becker2018). On the other hand, existing literature suggests that discussions about undocumented immigrants among whites focus on the criminality of their unauthorized status (Brown, Jones, and Becker Reference Brown, Jones and Becker2018). These findings suggest that despite having different motivations (economic versus cultural threat), white and Black respondents both have reasons for being threatened by undocumented immigrants, therefore suggesting more similarities than differences regarding the effect of generation and legal status.
H8a: White and Black respondents will rate generation/legal status characteristics as equally American.
H8b: Generation and legal status will be similarly salient in shaping the evaluations of who is considered an American among white and Black respondents.
Within-Group Differences
The degree of in-group identification is also an important component for capturing the subjective meaning of identities (Huddy Reference Huddy2001), suggesting a point of differentiation within each of these groups. High identifiers are most likely to adhere to in-group norms (Ellemers, Spears, and Doosje Reference Ellemers, Spears and Doosje2002). People who strongly identify as American are more likely to draw strict boundaries (Theiss-Morse Reference Theiss-Morse2009). Scholars find that strong national identity attachment among whites is associated with xenophobia and greater likelihood of drawing restrictive boundaries around the category (Carter and Pérez Reference Carter and Pérez2016).
H9a: Whites who strongly identify with their American identity will express more restrictive attitudes regarding who is American, compared to whites who do not identify strongly with their American identity.
H9b: Whites who strongly identify with their white identity will also express more restrictive attitudes compared to whites who do not identify strongly with their racial identity.
One of the key differences between white and Black Americans is the relationship between national and racial identity. For whites, threats from ethnoracial minorities are at the center of protecting their status as prototypical Americans and whites (Carter and Pérez Reference Carter and Pérez2016; Danbold and Huo Reference Danbold and Huo2015; Jardina Reference Jardina2019). In addition, racial and American identity are highly correlated for whites but less so for Black Americans (Carter and Pérez Reference Carter and Pérez2016; Masuoka and Junn Reference Masuoka and Junn2013). Whites generally think their American identity is more important to them, whereas Black Americans rate both identities as equally important (Masuoka and Junn Reference Masuoka and Junn2013).
Although Latinos may pose a threat to Black struggles for equality, an additional threat in protecting their racial group status is discrimination from whites. As Carter argues, Black Americans have historically been supportive of restrictions toward immigration because they did not want immigrants to jeopardize their group’s progress. Historically, immigrants made faster economic gains, were preferred as laborers, and were dissuaded from forming political alliances with Black Americans (Roediger Reference Roediger2007). At the same time, Black Americans have been reluctant of supporting immigration restrictions that were based on racism because of their own group’s position (Carter Reference Carter2019). Yet, Carter notes that Black nativism is not totally misaligned with white nativist sentiment, as they still endorse some of the same notions of exclusivity around religion, language, and citizenship as whites. Given the group’s historical struggle for full inclusion as Americans, it is likely that those who are strongly attached to this identity will be threatened by the further expansion of this category.
H10a: Black respondents who strongly identify with their American identity will express more restrictive attitudes regarding who is American, compared to Black respondents who do not identify strongly with their American identity.
The existing literature suggests potential conflicting outcomes regarding the relationship between Black racial identity attachment and their attitudes toward Latinos. While some work finds that Black respondents who strongly identify with their racial identity are less likely to support reduced immigration levels (Masuoka and Junn Reference Masuoka and Junn2013), other work finds no association between racial identity attachment and anti-Latino sentiment among Black Americans (Carter and Pérez Reference Carter and Pérez2016). Indeed, Masuoka and Junn argue that perceiving high linked fate symbolizes an awareness of the group’s minoritized status. More recently, Carter and King-Meadows argue that Trump appealed to Black Americans with his anti-immigration messages. They find a range of conflicting findings regarding Black linked fate and immigration attitudes. Black respondents in their study who have high linked fate are more likely to agree that immigrants take jobs, housing, and healthcare from the native-born, yet are less likely to support border security funding (Carter and King-Meadows Reference Carter and King-Meadows2019). Given their more recent empirical findings and the possibility of Trump activating a sense of racial group threat among Black respondents, I hypothesize that those who strongly identify with their racial identity will express more restrictive attitudes.
H10b: Black respondents who strongly identify with their Black identity will express more restrictive attitudes than those who do not identify strongly with their racial identity.
METHODOLOGY
Sample
To examine how native born white and Black Americans view Latinos, I use a conjoint survey experimentFootnote 2 (Hainmueller, Hopkins, and Yamamoto Reference Hainmueller, Hopkins and Yamamoto2014) of 750 White and 750 Black Americans, conducted by YouGov in September of 2020,Footnote 3 which results in a total analytic sample of 15,000 (7,500 for each group); see Ocampo-Roland (Reference Ocampo-Roland2025). The conjoint design presents each survey respondent with a total of ten profiles in a table format that are divided over five tasks. In each task, survey respondents are presented with two profiles side by sideFootnote 4 (Hainmueller, Hangartner, and Yamamoto Reference Hainmueller, Hangartner and Yamamoto2015).
Before seeing the profiles, respondents are told that they will be shown a series of profiles of individuals living in the United States, and they are instructed with answering each question according to their own personal beliefs. Respondents are told that the two profiles shown reflect people who are immigrants or come from a family with a history of immigration. Next, respondents are presented with a series of questions. The first asks them to choose which respondent they personally believe is more American. Next, respondents are asked how they would rate each person on a scale of 1 to 7, with 1 being not at all American and 7 being very much American.Footnote 5
First, I include a variety of characteristics that represent key ascriptive components. I use a series of ten photos that represent Latinos across the skin color spectrum. The series of images presented to respondents are from the Chicago Face Database (CFD), which compiles standardized images of individuals from different ethnoracial groups (Ma, Correll, and Wittenbrink Reference Ma, Correll and Wittenbrink2015). Images were chosen out of their database of pictures to represent a wide variety of skin tones. Notably, the CFD also provides rating data that rates how the photos are perceived racially, which is important when considering the diversity within the Latino population. I selected the photos using these data and chose a total of ten photos representing people of both genders, keeping skin tone as consistent as possible across gender. The photos represent individuals perceived to be 20 to 30 years of age. Although the first fourFootnote 6 categories of photos are people who self-identified as Latino, they are all perceived differently racially. More details on how the images are perceived racially using the norming data provided by CFD can be found in the appendix under Figure A2. These images are key to developing an understanding of how Latinos may be viewed differently based on their skin color, given the heterogeneity of the group. The full set of images can be found on appendix Figure A2.
In addition, I also examine ethnic and national identification, which considers the different terminology that may be used by Latinos to self-identify (including American, Latino, Latino American, Mexican, and Mexican American). I also incorporate religion (including Presbyterian, Catholic, Evangelical, and not religious). Similarly, I include partisan identification, which captures the importance of partisanship as a social identity (Iyengar and Westwood Reference Iyengar and Westwood2015) and political incorporation for immigrants and their descendants (Hajnal and Lee Reference Hajnal and Lee2011; Sears, Danbold, and Zavala Reference Sears, Danbold and Zavala2016). I also vary people’s spouses, including whether or not the profile is married, to further our understanding of whether intermarriage contributes to greater incorporation (Lichter, Carmalt, and Qian Reference Lichter, Carmalt and Qian2011). In addition, I include an interaction between generation and legal status captured in a characteristic representing a profile’s background, capturing different immigrants (undocumented, legal, or naturalized citizen) as well as U.S.-born children and grandchildren of legal/undocumented immigrants. The study also includes a series of characteristics that assess English language ability (Newman, Hartman, and Taber Reference Newman, Hartman and Taber2012), education, and occupation (Hainmueller and Hopkins Reference Hainmueller and Hopkins2015). A full list of characteristics can be found in the appendix Table A1.
Because the profiles are presented in pairs to respondents, I restrict profile pairs from having the same picture. To make the profiles as realistic as possible, I restrict certain attributes from appearing together in the same profile. These restrictions ensure that education, profession, English language ability, and background create reasonable profile combinations; details can be found in Appendix A. In addition, I ask respondents how important their American and racial identity is to them.
Analytic Strategy
First, I analyze the average marginal component effect (AMCE) of American ratings for all respondents together, to examine aggregate-level causal effects for the group.Footnote 7 Next, I compare white and Black respondents’ American ratings by examining marginal mean differences (Leeper, Hobolt, and Tilley Reference Leeper, Hobolt and Tilley2020). To interpret the magnitude of these differences, I use Cohen’s d, where I divide the mean difference by the standard deviation. This can be interpreted as the standardized difference between two means. I also examine the salience of each set of characteristics (Clayton, Ferwerda, and Horiuchi Reference Clayton, Ferwerda and Horiuchi2021) to compare differences between white and Black respondents. I use the profile choice variable to examine the attribute salience.
After comparing white and Black respondent ratings, I examine intragroup differences on the strength of respondents’ attachment to their American identity. Specifically, I examine the difference in perceptions (using marginal means) between respondents who stronglyFootnote 8 identify with their American identityFootnote 9 compared to those who do not, separately for white and Black respondents. Lastly, I examine how white and Black respondents who have strong attachments to their racial identitiesFootnote 10 differ from those who do not. I divide respondents into two groups: those for whom this identity matters a lot versus those with lower levels of identity importance. Adaptive shrinkage robustness tests to correct for multiple hypothesis testing can be found in appendix Tables A16–19 (Liu and Shiraito Reference Liu and Shiraito2023); most of the substantive results remain the same.
RESULTS
Ratings of Americanness among All Respondents
Figure 1 presents the AMCE results for the rating-based results for all respondents, where respondents are asked to rate how American they view a particular profile (rescaled from 0 to 1). Respondents rate the brown man photo as less American than the white man photo, on average, supporting the first hypothesis. However, there are no significant differences between respondents’ ratings of the other photos, in comparison to the white man photo. In addition, respondents rate Mexican, Hispanic, and Latino profiles as less American on average, in comparison to American profiles, which supports the second hypothesis.
Figure 1. Average Marginal Component Effect of Perceived Americanness among All Respondents
Note: Full results can be found in Table A5.
The most substantial effects emerge when considering the third hypothesis, on background and legal status. Immigrants of all categories are seen as less American relative to U.S.-born grandchildren of legal immigrants. This is particularly the case for undocumented immigrants, who are less likely to be viewed as American (0.21, SE = 0.02). The penalty is smaller for legal immigrants, where the difference in perceived Americanness is 0.04 points (SE = 0.01). Interestingly, the penalty for U.S.-born children of undocumented immigrants is larger in magnitude than those for legal immigrants, with profiles awarded a 0.05 lower rating, on average (SE = 0.01). There is also a penalty for grandchildren of undocumented immigrants, who are 0.02 points (SE = 0.01) less likely to be seen as American, relative to the grandchildren of legal immigrants. These findings support the third hypothesis and suggest the weight of unauthorized status persists across generations, impacting not only views toward immigrants but also native-born descendants of the undocumented. While other effects may be influenced by multiple hypothesis testing (see appendix Table A16), the effect of unauthorized status across generations remains robust to adaptive shrinkage (Liu and Shiraito Reference Liu and Shiraito2023), suggesting its importance.
Differences between White and Black Respondents
Descriptively, whites give higher ratings than Black respondents across the board. When exploring the marginal mean differences shown in Figure 2, the results indicate notable differences between the two groups. When using Cohen’s dFootnote 11 to examine the magnitude of the effect, these effects are modest in magnitude. Upon examining hypothesis 4a, the results indicate that whites rate lighter skinned photos as more American than Black respondents. Similarly, white respondents rate profiles with an identification of “white” more highly than Black respondents (0.08 difference, SE = 0.02, d = 0.28), which supports hypothesis 5a. White respondents also rate other identification characteristics as more American than Black respondents, but the difference in the evaluation of profiles that identify as “white” is the largest effect from this set of attributes.Footnote 12 In addition, contrary to the expectation of hypotheses 6a and 7a, white respondents rate education and occupation characteristics at most levels as more American than Black respondents. However, the largest differences that emerge in these categories between the two groups are at the highest end of the socioeconomic spectrum—in their evaluations of doctors (0.08 difference, SE = 0.02, d = 0.28) and those with advanced degrees (0.05 difference, SE = 0.01, d = 0.17).
Figure 2. Marginal Mean Differences in Perceived Americanness between White and Black Respondents
Note: Positive values indicate more favorable ratings among whites. Full results can be found in Table A6.
Substantial differences also emerge when considering hypothesis 8a, specifically on respondents’ evaluation of undocumented immigrants, who are seen as more American by Black respondents (0.07 difference, SE = 0.02, d = 0.24). When looking at the other legal status categories, the results show that whites rate profiles who are legal immigrants (0.07 difference, SE = 0.02, d = 0.24) or the descendants of legal immigrants more favorably than Black respondents, suggesting that white respondents give more favorable ratings toward profiles with a history of “legal” migration.
Beyond looking at differences in how Black and white respondents rate profiles along every characteristic, I also assess whether certain categories weigh more heavily on each group’s decisions by capturing each category’s salience (Clayton, Ferwerda, and Horiuchi Reference Clayton, Ferwerda and Horiuchi2021). The salience of each group of characteristics captures the extent to which a given set of characteristics is important for whether a profile is selected as the most American profile.Footnote 13 When an attribute’s salience is low, the probability of selecting one profile over another is close to being random, for the majority of the levels of that specific attribute (which would be indicated by a probability of selection of around 50 percent). On the other hand, attributes that are highly salient significantly deviate from 50 percent. A high attribute salience suggests that levels of a specific attribute are main drivers of whether a profile is selected or not.
The left hand panel shown in Figure 3 shows the average deviation (from 50%) of the probability of choosing the person as the most American profile with a specific attribute level for Black versus white respondents, with the deviation from the 45 degree line indicating the salience of each attribute. The difference in attribute salience is shown on the right hand panel of Figure 3, with the differences that are statistically significant at the 0.05 level highlighted in black.
Figure 3. Salience of Each Set of Attributes for White versus Black Respondents
Note: Characteristics that are statistically significant at the 0.05 level are highlighted in Black. Characteristics that are more salient for white respondents are below each line. Full results can be found in Tables A7 and A8.
The attributes that do not significantly deviate from the 45 degree line in Figure 3 (faces, ethnic/national identification, profession, education, religion, and spouse) are not more meaningful for either white or Black respondents, which do not provide support for hypotheses 4b, 5b, 6b, or 7b. However, a profile’s political orientation is more salient for Black respondents than for white respondents. On the contrary, language and background are more salient for white respondents, corresponding to hypothesis 8b. These results suggest the importance of language adaptation for white respondents in classifying who is an American. At the same time, legal status and immigrant generation are also important markers for whites. However, for Black respondents, the political affiliation of a profile is more important in shaping who they categorize as American.
How Do Strong Identifiers Draw Group Boundaries?
Among whites, those who strongly identify with their American identity draw boundaries in line with hypothesis 9a, as shown in Figure 4. Considering the marginal mean differences between whites who identify strongly with their American identity compared to those who do not, there are several differences. The largest difference between strong and weak identifying whites is on their evaluation of undocumented immigrants (0.21 difference, SE = 0.03, d = 0.72), which is a large effect. Strong identifiers rate legal immigrants and descendants of undocumented immigrants negatively compared to weak identifiers, but there is no difference in their evaluation of naturalized citizens and descendants of legal immigrants.
Figure 4. Marginal Mean Differences in Perceived Americanness between White Respondents Who Strongly Identify with Their American Identity versus Those Who Do Not
Note: Positive values indicate more negative ratings among strong identifiers. Full results can be found in Table A9.
White respondents who strongly identify with their American identity are also more likely to give lower ratings to ethnic/national identification categories of Latino, Hispanic, Mexican, and Mexican American in comparison to whites who do not strongly identify with their American identity.Footnote 14 Similarly, strong identifiers rate most of the photos as less American than weak identifiers, with the exception of the Black woman photo. These lower ratings among strong American identifiers are present across a range of skin tones, including notable penalties for the light man photo (0.11 difference, SE = 0.03, d = 0.38) and the brown man photo (0.10 difference, SE = 0.03, d = 0.34). Overall, this suggests a hesitancy to accept a broad range of Latinos as American among strong identifiers.
Considering socioeconomic status and adaptation characteristics, the effects again show that strong identifiers express more restrictions across the board. However, these do not consistently favor those with higher or lower socioeconomic status, suggesting that strong identifiers express more restrictive attitudes across the board, without preferring higher socioeconomic status individuals. Analyses comparing white respondents on the basis of their racial identity strength largely mirror the pattern of results for strong American identifiers, supporting hypothesis 9b. These are shown in Figure A4 in the appendix.
On the other hand, when examining Black respondents, there are no differences between those who strongly identify with their American identity compared to those who do not, ultimately not supporting hypothesis 10a (Figure A3 in the appendix). The results for hypothesis 10b, which explores how Black racial identity attachment shapes views toward Latinos, are shown in Figure 5. In stark contrast to the pattern of effects for white respondents, Black respondents who strongly identify with their racial identity generally have more open boundaries when considering who is an American relative to weak identifiers. The largest difference between high identifiers relative to weak identifiers is their preference for the Black man photo (0.09, SE = 0.03, d = 0.31), which is a modest effect. However, strong identifiers also rate some of the whiter photos, as well as profiles with an ethnic/national identification as white, as more American than weak identifiers, which suggests that their perceptions of who is considered an American are not simply restricted to those they perceive as most similar. Black strong identifiers rate U.S.-born profiles more positively than weak identifiers, regardless of whether their parents or grandparents were undocumented. These results indicate that strong identifiers draw more open boundaries—opening up possibilities for redefining who is conventionally viewed as American.
Figure 5. Marginal Mean Differences in Perceived Americanness between Black Respondents Who Strongly Identify with Their Racial Identity versus Those Who Do Not
Note: Negative values indicate more positive ratings among strong identifiers. Full results can be found in Table A10.
CONCLUSIONS AND IMPLICATIONS
First, this study examines whether different characteristics of Latinos matter for how they are perceived as American and finds that ascriptive characteristics are important. In addition, even Latinos who are born in the United States, who are by definition American, are seen as less American when their parents or grandparents are undocumented. This is a novel finding, highlighting how dramatically unauthorized status impacts perceptions, even toward the U.S.-born (España-Nájera and Vera Reference España-Nájera and Vera2020; Flores and Schachter Reference Flores and Schachter2018; Schachter Reference Schachter2016; Wright, Levy, and Citrin Reference Wright, Levy and Citrin2016). As a whole, attitudes toward Latinos, with a focus on the U.S.-born, warrant closer investigation.
In comparing how prototypical (white) versus peripheral (Black) group members draw the boundaries of group membership, this paper finds differences in their assessments, as well as how national and racial identity attachment operate for each of the two groups. Despite Black Americans being a less prototypical group, they generally rated a range of characteristics as less American than whites. This is consistent with previous work that finds they have strict definitions of what it means to be American (Carter Reference Carter2019; Masuoka and Junn Reference Masuoka and Junn2013; Tafoya, Corral, and Leal Reference Tafoya, Corral and Leal2022). It is possible that because Black Americans have had to fight for inclusion within the category (Collins Reference Collins2001; Du Bois Reference Du Bois1909; Haney-López Reference Haney-López2006), they may have a higher bar for who should be included.
Examining the differences between white and Black respondents suggests that cultural notions drive whites’ definition of the exclusivity of who is an American. These cultural concerns are primarily rooted in the notion of assimilability (Danbold and Huo Reference Danbold and Huo2022), with whites hesitant to extend the boundaries of who is an American to those who do not fit certain criteria for assimilation, which is consistent with previous work (Abascal Reference Abascal2020; Lacayo Reference Lacayo2017). Whites who are strongly attached to their American and racial identity are invested in maintaining the group’s exclusivity and protecting their prototypicality by limiting who is included (Theiss-Morse Reference Theiss-Morse2009).
Some of the notable differences between the two groups also have implications for politics. Even though the undocumented have been framed as an economic threat to Black Americans, they do not appear to perceive them as such, as they view them much more positively than whites. This suggests that there is potential for coalition-building between the two groups (Jones Reference Jones2019), which may emerge on the basis of a shared sense of discrimination (Jones Reference Jones2022). At the same time, as there is a more pronounced difference in evaluations of higher socioeconomic status individuals, this suggests a few possibilities. High socioeconomic status could be more valued by whites, suggesting its role in the cultural underpinning of how whites draw boundaries of American identity. It could also reflect that Black respondents sense greater economic threat from those with higher socioeconomic statuses. Another explanation relates to the ingroup projection model (Wenzel, Mummendey, and Waldzus Reference Wenzel, Mummendey and Waldzus2007), which suggests that subgroups project their own group’s characteristics onto the superordinate group. It is possible that Black respondents are less likely to perceive higher socioeconomic characteristics as representing their subgroup compared to whites. While it is difficult to definitively say which mechanism is at work, these explanations suggest cultural factors driving white attitudes and economic factors driving Black attitudes. For the second explanation, it is notable that greater economic threat for Black respondents emerges from higher socioeconomic status individuals (Gay Reference Gay2006), rather than lower socioeconomic status individuals (Wilkinson Reference Wilkinson2015), which is the focus of much of the literature. This finding also challenges long-standing findings about the preference for higher socioeconomic status immigrants (Hainmueller and Hiscox Reference Hainmueller and Hiscox2010; Hainmueller and Hopkins Reference Hainmueller and Hopkins2015), indicating that this is more reflective of white rather than Black attitudes. Notably, it appears that Black Americans are more likely to see those who are marginalized as more American, highlighting the greatest potential for coalition-building and political alliances to be specifically between Black Americans and more marginalized Latinos.
The prevalence of cultural concerns among whites and economic concerns among Black respondents is consistent with what the literature suggests, yet prior work has not fully documented the political nature of how boundaries are drawn, particularly for Black Americans. Scholars have documented the importance of party cues within the group (Enders and Thornton Reference Enders and Thornton2022; White and Laird Reference White and Laird2020), but this result underscores the importance of political affiliation for Black Americans for superordinate group relations, which is a novel finding. Other work has found that perceptions of a group’s Americanness impacts perceptions of the likelihood of interracial coalitions (Craig et al. Reference Craig, Zou, Bai and Lee2022), but this shows that the opposite is also the case—that political considerations impact whether people are seen as fellow group members. Specifically, this finding suggests that Black Americans draw boundaries corresponding to the characteristics that will protect their group status—in this case, it is someone’s political affiliation. Partisanship could signal whether they perceive Latinos as prospective allies to their plight against racial discrimination or as potential foes. This finding may also be reflective of the importance of partisan identity to Black Americans’ definition of American identity, as reflected by the ingroup projection model (Wenzel, Mummendey, and Waldzus Reference Wenzel, Mummendey and Waldzus2007).
Among Black Americans, respondents who highly identify with their racial group are more open with who they categorize as American, suggesting that strong racial group attachment indicates awareness of the group’s marginalized social position in society (Masuoka and Junn Reference Masuoka and Junn2013) and willingness to welcome others into the group. As they acknowledge their marginal position as Americans, expanding how group membership is defined can help their overall status (Dovidio, Gaertner, and Saguy Reference Dovidio, Gaertner and Saguy2009). This may also relate to the sensitivity that Black Americans have to the plight of immigrants, which may be heightened among those with greater racial identification (Carter Reference Carter2019). Although previous work argues that Trump successfully used similar appeals for Black and white Americans (Carter and King-Meadows Reference Carter and King-Meadows2019), the results do not support that claim, as perceptions of who is American do not differ between Black respondents with different levels of American identity attachment. Despite Black respondents as a whole having somewhat strict definitions of who is considered an American, these views encompass members who likely make sense of their peripheral status in different ways. As previous work has shown, peripheral group members react to their marginality differently and may deviate in their inclusion goals and behavior toward the group (Ellemers and Jetten Reference Ellemers and Jetten2013).
In sum, these results have several important implications for politics, as classifying someone as a fellow American indicates perceived commonality and co-nationals are given benefits not extended to others (Miller Reference Miller1999; Theiss-Morse Reference Theiss-Morse2009). Respondents’ overall position as more or less prototypical within the superordinate group shapes their outlook on what the identity means (Carter and Pérez Reference Carter and Pérez2016; Huddy Reference Huddy2001). As previous work has found that perceptions of who is a fellow group member shape public opinion and prosocial behavior (Theiss-Morse Reference Theiss-Morse2009), these findings suggest a divergence in who is awarded the benefits of social citizenship (Bloemraad et al. Reference Bloemraad, Kymlicka, Lamont and Son Hing2019). Black respondents’ political, and to some extent economic, concerns are central to their evaluations of who should be considered American. The evidence indicates that Black Americans perceive commonality with more marginalized Latinos, therefore suggesting greater potential for coalition-building with certain subsets of the group (Jones Reference Jones2022). This helps to refine previous theories that focused on economic competition between the two groups, which this study suggests is more pronounced toward those in higher socioeconomic status positions. In addition, as studies have shown that a shared sense of marginalization can lead Black Americans to support proimmigration policies (Pérez, Vicuña, and Ramos Reference Pérez, Vicuña and Ramos2023), these findings help to better illustrate which Latinos Black Americans view as sharing this status and under what conditions they might support immigration. On the other hand, whites favor those on the opposite end of the spectrum, who meet their criteria for assimilation, upholding a culturally homogenous view (Danbold and Huo Reference Danbold and Huo2022; Danbold, Serrano-Careaga, and Huo Reference Danbold, Serrano-Careaga and Huo2023). While political alliances between whites and Latinos are not discussed as extensively, these results show that more advantaged Latinos and those who approximate whiteness are more likely to be accepted by whites. As prototypicality threat is linked to support for nativist policies among whites (Danbold, Serrano-Careaga, and Huo Reference Danbold, Serrano-Careaga and Huo2023), it is notable that this may not be targeted toward all Latinos but only some Latinos. These findings suggest that Latinos are not politically incorporating as one group; how others perceive Latinos varies based on the group’s diverse characteristics (Bonilla-Silva Reference Bonilla-Silva2004).
Overall, this work highlights the need for further exploration of whether Latinos are included or excluded and specifically to explore attitudes toward Latinos beyond just focusing on attitudes toward immigration policy. In future iterations, studies should examine attitudes toward Latinos from various origin countries, to fully consider whether this impacts perceptions toward the group (Garcia-Rios, Pedraza, and Wilcox-Archuleta Reference Garcia-Rios, Pedraza and Wilcox-Archuleta2019). In addition, the effect of partisan identification should also be further explored as it relates to Black attitudes beyond co-ethnics. Other future paths for exploration include eliciting physical appearance cues through more realistic means, including photographs and beyond, and continue to strengthen external validity in survey experiments (Abrajano, Elmendorf, and Quinn Reference Abrajano, Elmendorf and Quinn2018). Although this study suggests a potential gendered angle, as it was the brown man photo who was seen as less American, future studies should examine the intersection of gender with skin color, and whether there are gendered dimensions to whether a group is included.
Furthermore, paths for future exploration include exploring how this translates to other groups, contexts, and scenarios. Given that this study was conducted in September 2020, the 2020 police brutality protests, and the coronavirus pandemic’s unique circumstances, warrant future examination of these questions. It is possible that the impact of racial group identity may have changed, particularly for Black Americans. Another critical question is the relationship between these findings and policy—these results suggest support for repealing birthright citizenship, but studies should explore this more directly. Lastly, as studies have examined that one’s status as American can be threatened (Pérez and Kuo Reference Pérez and Kuo2021; Pérez, Robertson, and Vicuña Reference Pérez, Robertson and Vicuña2023), studies should continue to examine how stable these boundaries are and under what conditions they can be redefined.
SUPPLEMENTARY MATERIAL
To view supplementary material for this article, please visit http://doi.org/10.1017/S000305542400131X.
DATA AVAILABILITY STATEMENT
Research documentation and data that support the findings of this study are openly available at the American Political Science Review Dataverse: https://doi.org/10.7910/DVN/CXT4YN.
ACKNOWLEDGMENTS
I would like to thank Chenoa Flippen, Daniel Hopkins, Michael Jones-Correa, Camille Charles, Valerie Martinez-Ebers, Angela X. Ocampo, Sergio Garcia-Rios, and Roberto Carlos for their valuable feedback on earlier drafts of this project. I am also grateful to the anonymous reviewers whose feedback greatly improved the manuscript. Thank you to Muyao Hang for excellent research assistance. I would also like to thank Bryan Wilcox-Archuleta and Anton Strezhnev for their help in setting up the study. Participants in the Politics of Race, Ethnicity, and Immigration Consortium meetings, as well as the University of Pennsylvania’s Sociology of Race, Ethnicity and Immigration Workshop, provided helpful comments and suggestions on earlier versions of this project.
FUNDING
This research was supported by several grants at the University of Pennsylvania: the Graduate and Professional Student Assembly-Provost Award for Interdisciplinary Innovation, the Center for the Study of Ethnicity, Race and Immigration’s Turner-Schulman Graduate Fellowship, the School of Arts and Sciences Dissertation Research Fund, and the Department of Sociology’s Pollak Summer Research Fellowship.
CONFLICT OF INTEREST
The author declares no ethical issues or conflicts of interest in this research.
ETHICAL STANDARDS
The author declares that the human subjects research in this article was reviewed and approved by University of Pennsylvania’s Institutional Review Board and certificate numbers are provided in the text. The author affirms that this article adheres to the principles concerning research with human participants laid out in APSA’s Principles and Guidance on Human Subject Research (2020).
Open access
Comments
No Comments have been published for this article.