Seen as Latino,Assumed Lower Class: Racialized Class and Immigrant Status Perceptions in the United States

Abstract

Although racial and class stereotypes are intertwined, few studies have examined the degree to which racial perceptions are linked to assumptions about social class. Among Latinos, racial self-identities often do not align with racial classifications by others, complicating debates about their place in the U.S. racial order. This study draws on unique survey data in which respondents classify the race, immigrant status, and social class of people in photographs who self-identify as Latino, Black, or White. In contrast to theories positing Latinos as a group in between Black and White Americans, findings show that self-identified White, Black and Latino observers alike tend to perceive Latinos as lower in socioeconomic standing than Black Americans. However, class perceptions of self-identified Latinos vary by their perceived race and immigrant status. This study suggests that ideas about the hierarchical positioning of racial groups at the macro-level, especially for Latinos, may not correspond to how socioeconomic stereotypes are experienced at the individual level, which vary by perceptions of race and immigrant status.

Studies show that people of different social class backgrounds experience differential treatment—by teachers, physicians, and others—and that people of lower social class experience negative stereotyping as lazy, incompetent, or immoral (Cozzarelli, Wilkinson, and Tagler 2001; Fiske and Markus 2012; Lott and Saxon 2002). Yet although racial differences in objective indicators of socioeconomic status are well documented, surprisingly little research systematically examines the degree to which class stereotypes are linked to race (for an exception, see Freeman et al. 2011). Race-based class stereotypes, however, are the implied mechanism in a large literature on “money whitening” (Roth, Solís, and Sue 2022; Schwartzman 2007; Telles 2002; Telles and Lim 1998). Moving beyond objective indicators to examine perceptions of social class is valuable because “humans experience inequality . . . in social interactions” (DiMaggio 2012). Understanding people’s perceptions of social class therefore provides insight into how everyday interactions may be shaped by assumptions about class in addition to assumptions about race.

Although the growth in the Latino ¹ population has transformed U.S. racial demographics, few studies have considered how people perceive Latinos’ location within the racialized socioeconomic order. Latinos have alternately been theorized as a group in between White and Black Americans in status (Bonilla-Silva 2004; Feliciano 2001; O’Brien 2008), a group in the process of assimilating into Whiteness (Yancey 2003), or a racialized minority group similar to, or even positioned below, Black Americans (Feagin and Cobas 2014; Massey 2009). Yet we know little about how Americans perceive Latinos across these possibilities, particularly when it comes to imagined socioeconomic standing.

To investigate these issues, we draw from a unique nationally representative online survey of White, Black, and Latino adults, in which respondents classified the race, immigrant status, and social class of photos of people who self-identify as Latino, Black, or White. We ask, how do racial classifications and perceptions of immigrant status relate to assumptions about social class, especially for the growing population who identifies as Latino? We find that in contrast to theories of Latinos as an in-between group, Latinos are more often perceived as lower in socioeconomic standing than Black Americans. At the same time, class assumptions vary depending on how respondents perceive a person’s race and immigrant status. Individuals seen as Latino only, especially when assumed to be undocumented, are likely to be perceived at the bottom of the class structure. These findings suggest that macro-level ideas about the hierarchical positioning of racial groups, especially for Latinos, may not capture how socioeconomic stereotypes operate in everyday interactions, where judgements vary by perceptions of race and immigrant status.

Cognitive Links between Race and Class

The U.S. racial stratification system has historically been marked by the sharp boundary between Whites at the top of the socioeconomic hierarchy and Blacks at the bottom (Bashi Treitler 2013; Jung 2015). Consequently, this has fostered a long-standing association between Blackness and lower social class in the minds of many Americans (Cox and Devine 2015; Fiske et al. 2002). This association is so entrenched that middle-class Blacks often feel compelled to signal their social class to counter negative stereotyping and discrimination (Lacy 2007).

Just as class perceptions may be shaped by racial categorizations, race perceptions may be shaped by socioeconomic status. Several studies from Latin America indicate that observers classify higher socioeconomic status individuals as White (Roth et al. 2022; Telles 2002; Telles and Lim 1998). Although findings in the United States context are more mixed, some evidence suggests racially ambiguous individuals of lower socioeconomic status are more likely to be perceived as Black than as White (Saperstein and Penner 2012; though see Kramer, DeFina, and Hannon 2016). Consistent with the “money whitening” theory, when subjects observe a counter-stereotypical pairing (e.g., when primed with an “educated” Black person), they may cognitively “lighten” the individual (Ben-Zeev et al. 2014).

That perceptions of race and class are linked is consistent with intersectional scholarship that emphasizes how people simultaneously occupy multiple social categories which affect their life experiences (Collins 2015). As Ridgeway and Kricheli-Katz (2013) argued, intersectional systems of inequality at the macro-level affect how social stereotypes are experienced at the individual level. For example, Heiserman (2023) argued that some “prototypical” categorical intersections, such as Black and lower class, are subject to particularly pronounced negative stereotypes. Browne, Tatum, and Gonzalez (2020) argued that people’s experiences are shaped by the “shared cognitive image of the modal group member and their key attributes based on intersecting stratification systems” (p. 81). Their research, drawing on in-depth interviews with middle-class Latinos, is one of few studies to apply these ideas to understanding Latinos’ experiences (Browne 2020). We extend research on intersectionality by assessing how perceivers draw connections between membership in White, Black, and Latino racial categories and social class.

The U.S. racial hierarchy has long been dominated by a primary divide between White Americans at the top of the socioeconomic order and Black Americans at the bottom, but links between race and class perceptions may be changing. As Roth (2012) argued, societal cultural understandings of the meanings of racial categories vary across time and place. Subjective perceptions of social class reflect larger cultural schemas and a sense of social standing in the larger society, and not just objective socioeconomic indicators. For instance, historically, White Americans with similar objective social class indicators (e.g., education, income) have tended to rate themselves higher in subjective social class, whereas Black Americans have tended to rate themselves lower (Jackman and Jackman 1973). Yet, more recently, Cohen et al. (2017) found evidence of a weakening of this historically “caste-like” racial pattern. Specifically, poorer and less educated White Americans no longer view themselves as holding higher subjective class status than would be expected on the basis of their objective class position. Nevertheless, Black Americans continue to perceive their own class standing as lower than that of Whites with comparable objective indicators (Cohen et al. 2017). Although this research focuses only on racial differences in individuals’ perceptions of their own class status, it raises questions about how people perceive the class standing of others, and how these perceptions extend to ethnoracial groups beyond the Black/White binary.

Although Black individuals are often seen as lower in social class than White individuals, some studies also associate lower class status with Latinidad. Browne et al. (2020) found that middle-class Dominican and Mexican immigrants in Atlanta recognize the stereotype of a typical Latino as working class. These immigrants sought to distinguish themselves from that stereotype by emphasizing their middle-class status, like middle-class African Americans in Lacy’s (2007) study. Survey experiments also suggest cognitive associations between Latinidad and lower class status. Garcia and Abascal (2016:427) showed that respondents associated Spanish names with lower class backgrounds. Similarly, Schachter, Flores, and Maghbouleh (2021) found that White Americans tend to categorize profiles of lower occupational status individuals as Latinos (Schachter 2021, Table 2).

Social categories intersect, but as Monk (2022) argued, the boundaries between categories (including both race and class) are not always clearly delineated. Monk argued that inequality hinges on “not the superordinate categories themselves, but the various properties (or cues) associated with categories . . . and the mental representations, abstract ideas . . . and stereotypes triggered by these properties” (p. 14). Monk’s theory of typicality suggests that the extent to which any individual’s characteristics coincide with what is “typical” for their social category shapes variation in life experiences and inequalities. This suggests the value of examining how people place others in racial and class categories simultaneously while recognizing that who is viewed as part of a stereotypically lower status racial group may not align with self-identified membership in that category. For example, a light-skinned person might self-identify as Latino but be seen as White and assumed to have high socioeconomic status. This matters because class stereotypes can affect life experiences, resulting in diverging experiences among those asserting the same ethnoracial identity (e.g., as Latino). This is crucial for Latinos who are not as consistently placed in racial categories on the basis of phenotypic cues as Black and White Americans, leading to more divergence between racial self-identities and observer classifications (Abascal 2020; Feliciano 2016; Laniyonu and Donahue 2023; Roth 2010).

Latinos in the U.S. Racial Structure

Scholars debate the place of Latinos, now the largest “minority” group in the United States (Stepler and Brown 2015), in the U.S. racial hierarchy (Bonilla-Silva 2004; Feliciano, Lee, and Robnett 2011; Frank, Akresh, and Lu 2010). Some argue that Latinos occupy an in-between place in a hierarchy in which Whites remain at the top and Blacks at the bottom (Flores-González 2017; O’Brien 2008), whereas others suggest that they are assimilating into Whiteness (Alba and Islam 2009; Yancey 2003).

Another view, recognizing the growth of the African American middle class and highly educated Black immigrant groups, posits that Latinos are now at the bottom of the socioeconomic hierarchy. For example, Massey (2009) argued that “after occupying a middle socioeconomic position between whites and blacks for generations, the economic fortunes of Mexicans have now fallen to levels at or below those of African Americans” (p. 25). Although Massey focused on Mexican Americans, the largest Latino group, qualitative studies suggest that because of tendencies to place others into “typical” categories, the typical Latino is often imagined as a working-class, undocumented Mexican immigrant (Browne 2024; Jiménez 2010).

Perceptions of Latinos’ social class may shed light on this debate. If Latinos are in “the racial middle,” or “becoming White,” they might be perceived as lower class less often than Blacks, but more often than Whites. Conversely, Latinos in general may be perceived as a lower socioeconomic group than Black Americans if others’ perceptions of Latinos are influenced by ideas about “typical” Latinos: working-class Mexicans.

The phenotypic diversity among self-identified Latinos, driven partly by their ethnic and geographic diversity, complicates their place in the racial hierarchy (Feliciano and Robnett 2014). Some Latinos are racially classified as White, others as Latino, and others as Black (Feliciano 2016; Feliciano and Robnett 2014; Itzigsohn 2009; Rodriguez 2000; Roth 2012). Bonilla-Silva (2004) theorized a triracial hierarchy in which assimilated White Latinos are accepted as White, light-skinned Latinos have “honorary White” status, and dark-skinned Latinos fall into the “collective Black” category. This theory suggests phenotypic variation in both objective and subjective dimensions of socioeconomic status among Latinos. Existing studies show that darker skinned Latinos have lower socioeconomic outcomes (Arce, Murguia, and Frisbie 1987; Bonilla-Silva 2004; Espino and Franz 2002; Frank et al. 2010; Hunter 2005; Roth 2010; Telles and Murguia 1990). This heterogeneity implies that perceptions of social class may depend upon observers’ racial classifications. Put another way, an observer who perceives a person as Latino may simultaneously perceive them as lower class, while another might see that same person as White and middle class.

Observer Characteristics

Observer characteristics can also shape perceptions (Feliciano 2016; Harris 2002). Social-psychological research suggests that Whites invoke positive stereotypes for in-group members and negative ones for out-groups (Gaertner and McLauglin 1983). However, this in-group/out-group pattern may not be universal. Implicit association tests show people of all races attribute positive traits to higher status racial groups and negative traits to lower status groups (Fiske 2005). White observers may be more likely than Black or Latino observers to perceive Black or Latino targets as lower class. On the other hand, Black (and possibly Latino) Americans might classify in-group members as lower class, recognizing the persistent racial hierarchy in the United States and broader societal standing beyond just socioeconomic factors (Cohen et al. 2017).

Blumer’s (1958) group position theory also suggests that class perceptions vary by observer racial self-identity. According to this theory, collective images of other racial groups reinforce group positions. Accordingly, White Americans, regardless of their own social class, are likely to view Blacks and Latinos as lower class to reinforce their own superior group standing. Extending Blumer’s insights to racial minorities, Black Americans may perceive Latinos as lower class more often than Latinos do, responding to perceived threats to their own social position. Conversely, Latinos might be more inclined to view Blacks as lower class. How Black, Latino, and White Americans differ in associating race with social class for both their own racial group and others remains an open question.

Unlike group position theory, which suggests that class perceptions of racial groups are driven by collective images rather than individual experiences, intergroup contact theory suggests that direct interactions between different social groups can reduce prejudice and negative stereotyping (Allport 1979; Pettigrew and Tropp 2006). According to intergroup contact theory, individuals with close or frequent contact with out-group members may, regardless of their own self-identity, be less likely to invoke class-based stereotypes about the out-group. Accordingly, we consider how individual interactions with members of different racial groups influence class perceptions.

Perceivers’ social class identities may also influence their assessments of others’ social class. Social projection theory suggests that people tend to assume others are similar to themselves because of greater familiarity with in-group members (Robbins and Krueger 2005; Ross, Greene, and House 1977). Thus, lower class observers may tend to classify others as lower class, whereas those of middle or higher status may assume others are, like them, at least middle class. However, social desirability may also play a role, especially among the more educated, who might be less comfortable classifying someone as lower class (Heerwig and McCabe 2009).

Immigrant Status, Race, and Class Perceptions

Previous research suggests that attitudes toward Latinos vary on the basis of perceptions of their immigrant status. For example, Schachter (2021) found that White respondents do not negatively stereotype U.S.-born Latinos but that Whites negatively stereotype Latino immigrants even more than they do African Americans. Schachter concluded that attitudinal questions about how respondents view Latinos may “simultaneously overestimate anti-U.S.-born Latino sentiment and underestimate anti-Latino immigrant sentiment” (p. 520). This divergence in Latino stereotypes may also relate to associations between social class and immigrant status.

Beyond immigrant status alone, studies suggest a strong association between undocumented immigrants and lower socioeconomic status, which may partly stem from demographic realities, as we discuss further below. Qualitative studies reveal how Latinos work to project images of legality or U.S. citizenship to avoid scrutiny as unauthorized immigrants (Browne et al. 2020; García 2014; Jiménez 2010). One strategy to project a “legal” or U.S.-born image is to signal a higher socioeconomic status, revealing recognition that illegality is associated with lower class status. Indeed, recent works show how the “racialization of illegality” is linked with social class (Browne 2024). For example, García (2014) found that undocumented immigrants use strategies to “pass” as U.S.-born Americans that signal middle-class status, such as changing into “white collar” clothing and driving newer model cars (García 2014). In Atlanta, Browne and Odem (2012) argued that “unskilled illegal” has become the default Latino racial category.

Quantitative studies also reveal that White Americans often assume immigrants, especially undocumented immigrants, are lower class (Flores and Azar 2023; Flores and Schachter 2018). Flores and Azar (2023) found that White Americans commonly ascribe these attributes to “most immigrants”: “non-White” (87 percent), “do not speak English” (79 percent), “are poor” (69 percent) and have “low-level jobs” (71 percent). They also found that White Americans’ views diverge depending upon whether they envision undocumented or legal immigrants, describing “illegal” immigrants as Latino and poor, but “legal” immigrants as Asian and well educated. These findings suggest that traits such as Latinidad, poverty, and illegality are clustered in the minds of White Americans, leading to the expectation that people perceived as Latino and undocumented are most likely to be seen as lower class.

Accuracy of Class Perceptions

Are class perceptions of ethnoracial groups accurate? Assumptions about associations between social class and race may reflect, to some degree, accurate understandings of the average socioeconomic circumstances of different ethnoracial groups in the United States. Categorizing someone as lower class is not inherently negative if it aligns with objective class indicators. Stereotypes can result from “ordinary cognitive processes” that reflect “deeply embedded, shared mental representations” (Brubaker, Loveman, and Stamatov 2004:39–40). The association of the Latino category with undocumented immigration and lower socioeconomic status partly reflects the reality that because of U.S. policies, a large number of Latino undocumented immigrants are only able to work in low-status jobs (Massey 2009). Moreover, people may express their social class through cultural signals that may be visible and correctly perceived by others (Becker, Kraus, and Rheinschmidt-Same 2017; Moya and Fiske 2017).

Yet perceptions regarding race, class, and immigration often do not match reality. For example, White Americans overestimate the size of the undocumented population in the United States and wrongly believe most immigrants are poor and uneducated (Flores and Azar 2023). Moreover, although subjective class identities usually relate to objective indicators like occupational status, income, and education, most people identify as middle class (Hout 2008). The correspondence between individuals’ subjective class positions and objective indicators varies by their class position; the working class tend to inflate their class status, whereas the upper middle class do the opposite (Sosnaud, Brady, and Frenk 2014). A key question is whether observers’ class perceptions align with target individuals’ objective class positions. Regardless of accuracy, however, stereotypes by their nature are often applied to some individuals they do not fit. How such dissonance varies by race is an important question for understanding unequal life experiences. For example, we currently lack a clear understanding of whether middle-class Blacks and Latinos are similarly presumed lower class.

Research Questions

This study addresses several questions. First, which racial groups are most often associated with lower class status in the United States? Second, do perceptions of social class vary by observed race, particularly among self-identified Latinos, for whom racial self-identity and observed race frequently diverge? Third, how do perceptions of social class vary by observers’ own racial self-identity and socioeconomic status? Fourth, how are perceptions of immigrant status related to perceptions of lower class status, particularly for Latinos (both self-identified and those categorized by others as Latino)? Fifth, do observers accurately perceive others’ social class, and (how) does this vary by observed race?

We answer these questions using a novel survey with a race-stratified sample of Americans asked to evaluate photographs drawn from online dating profiles in terms of race, immigration status, and social class.

Data and Methods

Data

This study draws on an online survey collected from December 2022 to January 2023 by YouGov, a survey research company with an opt-in, online panel of adults living in the United States. The sample consists of 2,850 respondents, 950 each whose “best” self-identified race was White, Black, or Latino (see Appendix Table 1 for descriptive statistics of the sample). With weights, the race subsamples are representative of self-identified Whites, Blacks, and Latinos in the United States in terms of age, gender, U.S. region, and education, and in terms of unobserved traits, under standard assumptions. ² The Latino sample also includes oversamples of Latinos from the Spanish-speaking Caribbean (Cuba, the Dominican Republic, and Puerto Rico). The survey was offered in both English and Spanish.

The survey took about 15 minutes to complete. For the main part of the survey, respondents were asked to view 35 photos and rate their “best guess” of the race, immigrant status, and social class of the person in the photo. After a screen with instructions, ³ each photo was shown at the top of the screen, followed by three questions about respondent’s perceptions of the target photo, presented in the following order: race, immigrant status, and social class. This order may lead respondents to consider the race and immigrant status of the person in making their guesses about the person’s social class, in line with the directionality implied by the framing of our research questions. ⁴ However, as all three questions were visible at the same time, it is likely respondents were making simultaneous guesses; these cognitive processes cannot be disentangled.

The photos were selected from a bank of 576 high-quality ⁵ photos collected for a prior study from a public dating website (see Feliciano 2016). Profiles had been posted on Match.com by daters living in the New York, Los Angeles, Chicago, and Atlanta metropolitan areas and were downloaded in full by a team of research assistants. The profiles had been randomly selected within the key strata of interest in the original dating preferences study: race/ethnicity, gender, sexuality (e.g., men seeking women, men seeking men), and metropolitan area; other profile characteristics were not part of the sample selection process (see Feliciano 2016; Feliciano and Kizer 2021).

Although only the main photo drawn from the profile was included in the survey instrument, we linked the survey data to the self-identified race/ethnicity and educational attainment the person in the photo had reported in their dating profile. Given our interest in the Latino population, we oversampled self-identified Latino photos but also included some self-identified White and Black target photos.

Each respondent was randomly shown a total of 35 photos: 21 from the self-identified Latino-only target profiles, 4 from self-identified Latino/White target profiles, 2 from self-identified Latino/Black target profiles, 4 from self-identified Black-only target profiles, and 4 from self-identified White-only target profiles. The ratios of Latino/White photos to Latino/Black photos to Latino-only photos resemble these ratios in the full set of high-quality photos selected for the survey.

Variables

The key dependent variable for this study is perceived social class on the basis of the question “If you had to guess, would you say this person was most likely . . .?” The available categories, drawn from the National Election Surveys, were upper class, upper middle class, middle class, lower middle class, working class, and lower class. The order of the Likert-type scale was randomized (i.e., from upper to lower or lower to upper). Given our substantive interest in which racial group(s) are perceived to be at the bottom of the socioeconomic hierarchy, we focus our analysis on whether photos were perceived as lower class, which we operationalize by combining the lower class, working class, and lower middle class responses (vs. a reference category of middle class or higher). ⁶

We include several independent variables of substantive interest. The first is the survey respondent’s ethnoracial self-identity (respondent racial self-identity), and includes three categories—White, Black, or Latino—on the basis of the single category respondents felt described them best. Second, some analyses consider photo racial self-identity (the ethnoracial self-identity listed in the dating profile of the person in the photo); the possible categories include White only, Black only, Latino only, Latino/Black, and Latino/White. ⁷ Third, following Roth’s (2016) terminology, we measure photo observed race on the basis of survey respondents’ best guess of the target’s race/ethnicity. Respondents were asked, “What is your best guess of this person’s race/ethnicity?” and instructed, “Please check more than one box only if you think they are mixed or fall into several boxes at once.” The options included White, Hispanic/Latino(a), Black/African American, Middle Eastern or North African, East Asian, South Asian, American Indian or Native American, and some other race/ethnicity. These categories led to 153 possible combinations. The most common responses included White only (29 percent), Latino only (21 percent), and Black only (16 percent) (see Appendix Table 2 for a comprehensive list of observed race responses). For the sake of parsimony, we include in regression analyses the three observed race categories of most theoretical interest: White only, Black only, and Latino only; we collapse other combinations into an “other” category. In some descriptive analyses, we also consider these observed race categories: Latino/White, Latino/Black, White alone or in combination (AIC), Black AIC, and Latino AIC.

We also consider respondents’ perceptions of immigrant status on the basis of a question about whether the target photo was most likely “born in another country and came to the U.S. illegally,” “born in another country and came to the U.S. legally,” or “born in the United States.” These response options were presented in random order.

We also include additional variables pertaining to our research questions about how race and social class perceptions relate to one another. First, we use the respondent’s educational attainment as a proxy for their social class, collapsed into three categories: high school graduate or less, some college or two-year degree, and bachelor’s degree or higher. We also included a variable to control for the degree of interaction respondents have with various racial groups. This is based on the “frequency of having a conversation with someone who is [racial/ethnic group].” For our analyses, we focus primarily on their reported frequency of interactions with Latinos. The variable is measured on a seven-point scale, ranging from “not at all” to “every day.” Finally, for analyses of the accuracy of class perceptions, we use the educational attainment reported in the target’s dating profile to provide an indication of the actual social class of the person in the photo. ⁸ Here, we focus descriptive analyses on two categories: high school educated or less (as a proxy for lower class) and bachelor’s degree or higher (as a proxy for middle class or above). We also conduct supplementary regression analyses using the full range of the educational attainment categories available on the dating profiles of the people in the photos (high school or less, some college, associate’s degree, bachelor’s degree, graduate degree, PhD or postdoctoral, no answer) as independent variables.

We include several control variables that relate to the survey respondent’s characteristics (see Appendix Table 1). These include whether the respondent identified as more than one race (one race, two races, three or more races), age, gender (female vs. male), region (Midwest, Northeast, South, and West), and nativity. Nativity is measured as place of birth, with options including the United States, U.S. territories, and some other country. We combined the latter two into a “not U.S. born” category. ⁹ Political affiliation is measured by self-identified political party, with the options Democrat, Republican, independent, other, and “not sure.” We combined the latter three categories to form a three-category political affiliation variable.

Analyses

To address our research questions, we use descriptive and regression analyses, all with sampling weights applied. We treat each respondent-photo dyad as a case, and our unit of analysis is photo observations. We begin by assessing how perceptions of the class status of White, Black, and Latino Americans differ and how perceptions of class status vary by the measurement of race—self-identification or observation classification—using descriptive statistics. Then, focusing only on self-identified Latino photos, our main regression analyses predict perception as lower class using linear probability models with fixed effects for photos. Although our findings are robust to alternative modeling approaches, ¹⁰ the main advantage of using linear probability models for dichotomous outcomes is their interpretability (Hellevik 2009). These models account for unobserved variation in the photos themselves, allowing us to assess the impact of observer characteristics and observers’ perceptions of the race and immigrant status of the person in the photo. Because the same respondent assesses multiple photos, we provide robust standard errors that adjust for the clustering of observers. We use these regression models to assess how perceived social class varies by observed race among self-identified Latino target photos, how observer characteristics relate to class perceptions, and how perceptions of immigrant status relate to perceptions of lower class status.

To provide a sense of the accuracy of class perceptions, we conduct additional supplementary linear probability models with random effects for photos and respondents that include the educational attainment reported in the profile of the photo target. We also use weighted descriptive statistics to show how observers’ perceptions of the social class of target photos align with an indicator of social class as reported in the target photo’s profile (educational attainment), and how this varies by observed race among self-identified Latino target photos.

Findings

Which Ethnoracial Groups Are Most Often Perceived as Lower Class? Variation by the Measurement of Race

Table 1 reports the share of cases in which a photo was perceived as lower class on the basis of either the target photo’s racial self-identity (panel A) or the target photo’s observed race (panel B), sorted in order from lowest to highest. Both panels show a racialized ranking, with Whites most often placed at the top.

Table 1.

Perception as Lower Class by Race of Target Photo, Self-Identified and Observed.

	%	n
A. Perceived as lower class, from lowest to highest percentage, by racial self-identity of target photo
White only	34.3	11,399
Latino/White	44.2	11,220
Black only	47.4	11,399
Latino only	49.6	60,023
Latino/Black	51.5	5,699
B. Perceived as lower class, from lowest percentage to highest, by observed race a of target photo
White only	37.8	29,184
Latino/White	47.3	3,199
Black only	49.5	15,784
Latino/Black	55.6	1,255
Latino only	59.6	20,773

Note: Chi-square tests of associations within each panel were significant at p < .000.

a

Observed race categories are not comprehensive. For ease of comparison, we present only the observed race categories that are equivalent to the target photo self-identity categories available in our data.

In Table 1, panel A, which uses racial self-identity, White-only target photos are far less often perceived as lower class than any other group. Specifically, 34 percent of self-identified White-only target photos are perceived as lower class, compared with 44 percent of Latino/White target photos, 47 percent of self-identified Black-only target photos, and 50 percent of self-identified Latino-only target photos. Latino/Black target photos are perceived as lower class slightly more often, at 52 percent. ¹¹

Panel B, which uses an observed measure of race, shows a similar ordering. Photos perceived as White only are least likely to be seen as lower class (38 percent). They are followed by those perceived as Latino/White (47 percent), then by Black only (50 percent). Target photos perceived as Latino and Black are more likely to be perceived as lower class (56 percent) than photos perceived as Black only are. Finally, target photos perceived as Latino only are at the bottom of the hierarchy, with nearly 60 percent perceived as lower class.

Comparing panels A and B in Table 1 shows broadly similar social class rankings: Whites at the top, Blacks in the middle, and Latinos toward the bottom. The difference is sharper, however, when race is based on observer classification: Latinos fall even further down the hierarchy with a greater difference from others. For example, panel A shows only a 3-point difference between self-identified Black-only and Latino-only photos (47 percent vs. 50 percent). In contrast, panel B shows a 10-point gap between those classified as Black only (50 percent) and those classified as Latino only (60 percent). Overall, both panels indicate Latinos (except Latino/Whites) tend to be perceived as lower class more than Black Americans are. However, Latino-only target photos are more often perceived as lower class when observers classify them as Latino (60 percent) than when we consider photos of people who self-identify as Latino (50 percent), a 10-point gap. This discrepancy highlights how much perceptions of class status among Latinos depend on whether race is measured by self-identity or by external classification. In contrast, although other racial groups are also somewhat more likely to be perceived as lower class when race is based on observer perception rather than self-identification, the gap is much smaller. For example, the perceived rate of lower class status for Black individuals differs by only 2.1 percentage points when comparing self-identified race (panel A, 47.4 percent) to observed race (panel B, 49.5 percent)

How Perceived Social Class Varies by Observed Race among People Who Self-Identify as Latino

The variation in perceptions of Latinos across race measures stems from the frequent discrepancy between how Latinos self-identify and how they are perceived by others. At least on the basis of photos, which provide only visual cues, observers often do not racially classify self-identified Latinos as Latino (see Appendix Table 3). In fact, self-identified Latino-only photos were only categorized as Latino AIC in 39 percent of cases. ¹² In contrast, we find upward of 84 percent agreement in observed race of self-identified Black or White target photos (see Appendix Table 3). Given low rates of racial misalignment, class perceptions are similar for White and Black target photos regardless of whether we use the target photo’s self-identified race or observed race. We focus all subsequent analyses on target photos who self-identify as Latino AIC only.

Table 2 presents results of linear probability models ¹³ predicting perceptions of lower class status for those target photos that self-identified as Latino, with control variables for respondent characteristics. Model 2 adds a measure of perceived immigrant status. These models include fixed effects for photos and provide robust standard errors that adjust for clustering of observers. Importantly, the fixed effects allow us to assess patterns when observers view the same photo, such as whether observers who perceive a particular target photo to be Latino also tend to perceive that target photo as lower class, while observers who classify that same target photo as White tend to perceive that target photo as middle class.

Table 2.

Linear Probability Models Predicting Perception as Lower Class among Self-Identified (Alone or in Combination) Latino Target Photos, with Fixed Effects for Photos.

	Model 1		Model 2
	Coefficient	Robust SE	Coefficient	Robust SE
Observed race of photo (reference: White only)
Black only	.045***	.012	.040**	.012
Latino only	.147***	.009	.110***	.009
Other	.041***	.009	.010	.009
Respondent best self-identified race (reference: White)
Black	.033*	.014	.032*	.014
Latino	.072***	.016	.070***	.016
Respondent self-identifies as multiple races (reference: one race)
Two races	.009	.018	.011	.018
Three or more races	.132*	.056	.108	.056
Respondent party affiliation (Democrat)
Republican	−.006	.016	−.011	.016
Other	.004	.013	.011	.013
Respondent education (reference: high school or less)
Some college or AA	−.047***	.013	−.040**	.013
BA degree and above	−.070***	.015	−.064***	.014
Respondent female	−.016	.011	−.013	.011
Respondent age	−.001**	.000	−.001**	.000
Respondent foreign born	−.013	.019	−.026	.019
Respondent region (reference: Midwest)
Northeast	.002	.019	−.001	.019
South	−.035*	.016	−.036	.015
West	−.042*	.018	−.043	.018
Contact with Latinos	−.010**	.003	−.008**	.003
Perceived immigration status (reference: U.S. born)
Foreign born, migrated legally			.056***	.008
Foreign born, migrated illegally			.248***	.012

Note: N = 74,674; robust standard errors were adjusted for clustering on observers.

*

p < .05. **p < .01. ***p < .001.

Regression results in Table 2, model 1, confirm that among self-identified Latino target photos, those that observers perceive as Latino are much more likely than those perceived as White to be perceived as lower class as well, net of photo and observer characteristics. Specifically, compared with self-identified Latino target photos that are perceived as White, self-identified Latino target photos that are perceived as Latino have a probability of being perceived as lower class that is almost 15 percentage points higher. A target photo perceived as Black is also more likely to be perceived as lower class than if perceived as White (4.5 percent), but the association between Latino classification and lower class perception is much stronger. Moreover, additional statistical tests reveal that target photos perceived as Latino are significantly more likely to be perceived as lower class than target photos perceived as Black. These results, consistent with the descriptive results, suggest that self-identified Latinos categorized by others as Latino will be most likely to be perceived as lower class, followed by those classified as Black, with those classified as White least likely to be perceived as lower class.

(How) Do Observer Characteristics Relate to Class Perceptions?

The results in Table 2 suggest that respondents who identify as White, Black, or Latino are not equally likely to classify Latino-identified target photos as lower class. As model 1 shows, Latino respondents’ probability of perceiving a photo as lower class overall is about 7 percentage points higher than that of White respondents, whereas Black respondents’ probability of perceiving a photo as lower class is only about 3 percentage points higher than that of White respondents. Further analyses show that self-identified Black and Latino respondents do not significantly differ from one another, and interaction tests show no significant differences in the effects of respondents’ race by the observed race of the photo, challenging group position theory and theories that suggest people stereotype out-group members more negatively than in-group members. Differences in class perceptions between White respondents and Latino or Black respondents are not driven by ethnoracial differences in their social class backgrounds, self-identification with multiple racial categories, party affiliation, education, gender, age, nativity, region, or contact with Latinos, as these are also included as controls in the model.

Respondents’ social class relates significantly to their social class perceptions of others. Using educational attainment as a proxy for social class, those of higher social class are less likely to classify others as lower class. For example, the probability that college-educated respondents classify a photo as lower class is about 7 percentage points lower than that of respondents with high school education or less. On the basis of an objective indicator of social class (which likely correlates to subjective social class), these results suggest people tend to include others as members of their social class in-group, in line with social projection theory. These findings may also reflect social desirability bias among those who are more educated.

Table 2, model 1, also shows a small net difference in perceiving a target photo as lower class depending on the degree of contact respondents have with Latinos. Respondents who report more frequent conversations with Latinos are less likely to perceive self-identified Latino target photos as lower class. The net effect size is modest, such that the difference between the least contact (none) to the most (daily) is only 6 percentage points. We caution readers not to conclude that contact has a causal effect on perceptions. The measure of contact is self-reported, and the kinds of people who interact with more Latinos may differ from those who do not. Nevertheless, these findings suggest that differences by respondents’ racial self-identity and education are not driven by contact with Latinos. Moreover, interaction tests between contact with Latinos and target photo observed race show that those with greater contact with Latinos are not any less likely to assume observed-Latino photos are of lower class people, in contrast to intergroup contact theory.

How Do Perceptions of Immigrant Status Relate to Perceptions of Lower Class Status?

Table 2, model 2, adds perceptions of immigrant status to the analysis and shows that perceptions of immigrant status relate strongly to perceptions of lower class status. Self-identified Latino target photos that observers perceive as foreign-born are more often perceived as lower class than those assumed to be U.S. born. Yet the association between immigrant status and lower class status is much stronger for photos perceived as unauthorized migrants, ¹⁴ which is the strongest of any association in our models: photos perceived as unauthorized have a 25 percentage point higher probability of being perceived as lower class than photos perceived as U.S. born.

Adding perceptions of immigrant status to the analysis does not change the other patterns of results described for model 1. Most important, target photos perceived as Latino are still much more likely to be perceived as lower class than those perceived as White. Controlling for perception of immigrant status decreases the percentage difference between photos observed as Latino compared with those observed as White from about 15 percentage points more likely to about 11 percentage points more likely to be perceived as lower class.

Does the effect of Latino observed race on class perceptions depend on target photos’ perceived immigrant status? To examine this, we reestimate model 2 (Table 2) including an interaction term between target photo observed race and perceived immigrant status. Full model results are reported in Appendix Table 5 and the key finding is shown in Figure 1. Two findings are worth noting. First, whether target photos are perceived as a U.S.-born person, a foreign-born authorized migrant, or a foreign-born unauthorized migrant, target photos seen as Latino are perceived as lower class more than those seen as White or even Black. ¹⁵ Specifically, among target photos perceived as U.S. born, the model predicts a 54 percent probability that photos perceived as Latino will be perceived as lower class but only a 47 percent chance a photo perceived as Black will be perceived as lower class. Second, Latino observed race is most strongly associated with lower class perceptions for target photos perceived as unauthorized immigrants. According to the model, there is an almost 80 percent chance that target photos perceived as Latinos who migrated illegally will also be perceived as lower class. This is far greater than the equivalent probabilities for photos perceived as Black or White unauthorized migrants (about 64 percent and 55 percent, respectively). In sum, foreign-born status and, especially, unauthorized immigrant status, is associated with lower class perceptions for all target photos, but especially for those seen as Latino.

Figure 1.

Predicted probabilities of lower class perception with interactions of observed race and perceived immigrant status for self-identified Latino (alone or in combination) target photos.

How Accurate Are Observers’ Perceptions about Social Class?

One explanation for our findings is that the patterns reflect accurate differences in social class across ethnoracial groups. As our own respondent data show (see Appendix Table 1), Black and Latino individuals tend to have completed less education than Whites in the United States. ¹⁶ Our fixed-effects analyses show that different observers looking at the same photo make assumptions about social class that vary on the basis of their simultaneous categorizations of people by race and immigrant status. This implies that some guesses about the social class of the people in the photos accurately reflect their social class, whereas others do not. Although we do not have definitive data on the social class of the person in the target photo, we use their reported educational attainment as a reasonable proxy to determine which social class perceptions match the reported social class of the target.

We ran a supplementary regression analysis including the educational attainment listed in the target photo profile using random effects for photos and respondents (see Appendix Table 7). We find that observers are less likely to perceive target photos with bachelor’s or graduate degrees as lower class than target photos with a high school education or less. This confirms that observers picked up on visual cues of social class status from the photos; their assessments reflected the target photos’ reported class status more often than would be predicted by chance. For example, net of perceived race and immigrant status and observer characteristics, the likelihood that target photos with graduate degrees were perceived as lower class is about 13 percentage points lower than target photos with no postsecondary schooling.

However, controlling for the target photo’s educational attainment does not change the effects of perceived race or immigrant status in any discernable way. The statistical effects of Latino observed race, Black observed race, and immigrant status remain comparable in both magnitude and significance. This refutes the claim that target photos perceived to be Latino, Black, or foreign born (authorized or unauthorized) are perceived as lower class because they are lower class and convey that information via subtle visual cues. Regardless of how target photos are signaling their objective social class, target photos that are seen as not White or not U.S. born are perceived as lower class.

To further illustrate these findings, we examine discrepancies between reported social class and observers’ guesses in two directions. First, we consider it an overestimate if respondents categorize photos of targets who report they have a high school diploma or less education as being middle class or higher. Second, we consider it an underestimate if respondents categorize photos of targets who report having a bachelor’s degree or higher as being lower class.

Figure 2 shows how the accuracy of social class perceptions varies by the observed race of the person in the photo, for photos of self-identified Latinos (AIC). As the figure shows, the direction of discordance between perceived and reported social class varies depending upon the racial group in which the respondent categorizes the photo. For example, the social class of target photos classified as White only is much more often overestimated (54 percent) and less often underestimated (37 percent). For self-identified Latino photos perceived to be Black only, there is no bias toward overestimation or underestimation; the rate is about the same (46 percent). In contrast, photos perceived as Latino only are much more often assumed to be of lower social class than their educational attainment indicates (55 percent); only 39 percent are perceived to be of higher social class than the target photo’s education suggests. These findings show that although assumptions about social class partly reflect accurate understandings of ethnoracial differences in social class at the aggregate level, at the level of individual classification, inaccurate perceptions of social class depend on how self-identified Latinos are perceived racially.

Figure 2.

How does the accuracy of perceived social class vary by observed race among self-identified Latino (alone or in combination) target photo profiles?

Discussion and Conclusion

Latinos have been theorized to be an in-between group, not at the top or the bottom of the U.S. racial hierarchy. However, our findings suggest that, at least in terms of social class, Black Americans are no longer viewed as being at the very bottom of the socioeconomic hierarchy. This may relate to recognition of a growing Black middle class (Lacy 2007). Although many Latinos are also middle class (Agius Vallejo and Vasquez-Tokos 2024), it appears that Latinos, or at least a significant segment of them, are racialized at the bottom of the U.S. socioeconomic hierarchy, as Massey (2009) argued for Mexicans.

Moreover, the cognitive association between Latinidad and lower class status appears to be widely accepted among our representative sample of White, Black, and Latino respondents from across the United States, suggesting a widely shared cultural schema about Latino racialization (Roth 2012). We find only modest differences in how observers’ own characteristics, including their interactions with Latinos, relate to their associations between race and lower class status. For instance, in line with social projection theory, individuals tend to assume others share their social class status. Yet although Black and Latino respondents are more likely than Whites to perceive people as lower class, they are not any more or less likely than White respondents to assume Black or Latino photos are lower class, in contrast to theories of in-group/out-group stereotyping and group position. Black and Latino respondents may perceive more photos as lower class because they have more contact with lower class individuals, a factor we did not measure but which future research might address. We also did not examine variations in associations by metropolitan area. Given the heterogeneity in Latino ethnic groups and locations, future research should explore whether cultural schemas, including ideas about race and social class, vary by geographic area, such as in areas with high proportions of Afro-Latinos or Mexicans.

Furthermore, our findings suggest considerable heterogeneity in how Latinos are perceived in terms of social class on the basis of whether observers even see them as Latino, and whether observers perceive them as an unauthorized immigrant. These insights may clarify debates about the mental representations of Latinos in U.S. society, debates which may hinge on which segments of Latinos are envisioned (Schachter 2021). By some media accounts, Latinos, especially unauthorized immigrants, are racialized as “others” who are “taking over” (Chavez 2008), yet in other media representations, U.S.-born Latinos are accepted as part of the new mainstream (Guzmán and Valdivia 2004). The difference in perceptions may depend upon whether the media portrayals are of Latinos who are immigrants (especially undocumented ones) or U.S. born. Our findings are somewhat consistent with Schachter (2021), who found that White respondents negatively stereotype Latino immigrants even more than they do African Americans, but complicate her findings that White Americans do not negatively stereotype U.S.-born Latinos at all. We find that even among those assumed to be U.S. born, people classified by observers as Latino are more often assumed to be lower class than those classified as White or Black. In contrast, self-identified Latinos who are perceived as White and U.S. born benefit from the assumption of middle-class status. Those self-identified Latinos who are perceived as unauthorized Latino immigrants face the strongest assumptions that they are lower class (Figure 1).

Our findings suggest that boundaries between Latinos and Whites may be blurred for light-skinned Latinos who are perceived as White and U.S. born (Alba 2005). On the other hand, self-identified Latinos who are categorized by others as Latino on the basis of visual cues (<40 percent of all self-identified Latino target photos in our study), and especially those who are also assumed to be unauthorized immigrants, may face exclusion from the mainstream. These findings are in line with scholarship emphasizing variation in the outcomes and experiences of Latinos, especially depending on visual cues (Alba 2005, 2020; Alba, Jiménez, and Marrow 2013; Feliciano and Robnett 2014; Feliciano et al. 2011).

A limitation of this study is that we cannot fully disentangle the multidirectionality of these relationships: observers are likely making simultaneous inferences about racial classification, social class, and immigrant status. Stereotypes and social group membership are often mutually constituted (Cox and Devine 2015). This dynamic is evident in research showing that undocumented Latino immigrants sometimes adopt visual markers of middle-class status, such as particular styles of dress or types of cars, to “pass” as legal U.S. residents (Browne 2024; Garcia 2019). These strategies reveal an awareness not only that undocumented immigrants are often associated with lower socioeconomic status, but also that perceptions of legal status are shaped by perceived class status (Browne 2024; Garcia 2019).

Still, past research suggests that race, along with gender and age, are “primary categories” of classification that people use as inputs into other identities (Kang and Bodenhausen 2015; Petsko, Rosette, and Bodenhausen 2022; Ridgeway and Kricheli-Katz 2013). Moreover, visual cues of class status, such as hairstyles and dress, may be more fleeting or easily changed than racial status cues, such as skin color, which are less alterable. Although this suggests that observed race may more powerfully shape life experiences overall, among racially ambiguous persons, social class cues also shape racial classification (Freeman et al. 2011). Future experimental studies should further examine the cognitive processes and primary cues observers use to place those with nonprototypical phenotypes, such as many Latinos, into social categories.

Taken together, our findings suggest that the life experiences of Latinos likely vary widely on the basis of how they are perceived by others (López et al. 2018). Because people of different social class backgrounds experience differential treatment, this may contribute to differential experiences of discrimination among self-identified Latinos depending upon whether they fit the prototypical Latino appearance (Monk 2022). Those whose physical appearance corresponds to them being classified by observers as Latino may simultaneously be perceived as from lower social class backgrounds and subject to negative class stereotypes, such as that they are incompetent or irresponsible (Cozzarelli et al. 2001; Fiske and Markus 2012; Lott and Saxon 2002). Consistent with theories of intersectional stereotyping and typicality (Collins 2015; Heiserman 2023; Monk 2022), some college-educated Latinos who are perceived as Latino and undocumented on the basis of appearance will be subject to stereotyping as lower class even while that contrasts with their objective class position and legal status. Our findings extend past qualitative research highlighting such experiences among middle-class Latinos by showing how a large sample of perceivers make these assumptions (Browne 2024; Browne et al. 2020).

Our research findings, based on reactions to a photo, are most directly relevant to understanding how the experiences of self-identified Latinos vary in public spaces, where observers rely mainly on visual cues to make initial judgements. In such contexts, observers such as police or store personnel may form initial assumptions about race, immigrant status, and social class on the basis of phenotype, which may lead to differential treatment. However, perceptions and categorizations can change and differ in more interactive settings when additional cues of race, immigrant status, and social class—such as names, accents, and occupations—are available. Future research should explore how perceptions of self-identified Latinos vary in different settings or in response to additional information.

To be sure, racial differences in average socioeconomic statuses are an empirical fact that help account for the associations people make between race and social class. However, assumptions about specific individuals are often incorrect. This study highlights that ideas about the hierarchical positioning of racial groups at the macro-level, especially for Latinos, may not correspond to how assumptions about socioeconomic position are experienced at the individual level, which vary by perceptions of race and immigrant status.

Supplemental Material

sj-xlsx-1-srd-10.1177_23780231251409867 – Supplemental material for Seen as Latino, Assumed Lower Class: Racialized Class and Immigrant Status Perceptions in the United States

Supplemental material, sj-xlsx-1-srd-10.1177_23780231251409867 for Seen as Latino, Assumed Lower Class: Racialized Class and Immigrant Status Perceptions in the United States by Cynthia Feliciano, Zhongze Wei, Maria Abascal and Wendy D. Roth in Socius