Experiences of Ethnic Discrimination Among US Hispanics: Intersections of Language, Heritage, and Discrimination Setting

Abstract

Background:

Informed by Latino Critical Race Theory, the present study examined how intersections between English use/proficiency, Spanish use/proficiency, and heritage group shape the varying experiences of ethnic discrimination reported by US Hispanic adults.

Methods:

The study utilized data from 7,037 Hispanic adults from the 2012–2013 National Epidemiologic Survey on Alcohol and Related Conditions-III. Multivariable binomial logistic regression modeled language use/proficiency, heritage, and demographic characteristics as predictors of past-year self-reported perceived ethnic discrimination, overall and in six different settings.

Results:

Both English and Spanish use/proficiency were positively associated with increased adjusted odds of reporting ethnic discrimination overall, in public, or with respect to employment/education/ housing/courts/police; however, with respect to being called a racist name or receiving verbal/physical threats/assaults, a positive association was observed for English, yet not Spanish. Results also indicated a significant interaction between English use/proficiency and Spanish use/proficiency when predicting past-year ethnic discrimination overall or for any of the six types/settings examined, although the relationship between language use/proficiency and ethnic discrimination varied by Hispanic heritage group.

Conclusion:

Study findings emphasize that experiencing some form of ethnic discrimination is relatively common among US Hispanic adults, yet the prevalence and types or settings of ethnic discrimination vary widely on the basis of demographics, immigrant generation, heritage, and the interplay between English and Spanish use/proficiency.

Contemporary news reports are replete with examples of anti-Hispanic acts across the country: a woman in Illinois was berated for wearing a shirt with the flag of Puerto Rico—a US territory (Cox & Keilman, 2019); restaurant employees in New York were reviled for speaking Spanish and threatened to be “kicked out” of the country (Hajela, 2018); and, in an act of the utmost violence, 23 individuals in El Paso, Texas were killed by a shooter who reportedly sought “to kill Mexicans” (Aguilera, 2020). In 2019, the US Federal Bureau of Investigation (FBI) reported 693 victims of anti-Hispanic hate crimes (FBI, n.d.). Discriminatory acts based on race/ethnicity represent symptoms of historical, far-reaching, pernicious systemic racism, engendered by economic domination/exploitation and the imposition of racist attitudes, ideologies, and institutions (Feagin, 2006), leading to a variety of detrimental outcomes including compromised health.

A substantial body of studies, including longitudinal analyses (Cave et al., 2020), document the negative effects of racial/ethnic discrimination on health. Findings from systematic reviews and meta-analyses indicate that racial/ethnic discrimination is positively associated with adverse outcomes in general health (Paradies et al., 2015), physical health (Cave et al., 2020; Dolezsar, et al., 2014; Korous et al., 2017; Paradies, 2006; Paradies et al., 2015; Pascoe & Richman, 2009), and mental health (Paradies, 2006; Paradies et al., 2015; Pascoe & Richman, 2009). The association between racial/ethnic discrimination and a host of adverse outcomes has also been documented specifically in US Hispanic populations, not only for mental and behavioral health, but also for educational and occupational measures (Lee & Ahn, 2012).

While an abundance of research focuses on ethnic discrimination as a predictor of various outcomes, the present study situates ethnic discrimination as the outcome of interest. Guided by principles from Critical Race Theory (Delgado & Stefancic, 2017) and LatCrit (Latino-critical) theory (Hernández-Truyol, 1997; Valdes, 2005), the study focuses on the multiple dimensions of identity and the diversity among “Hispanics.” The “Hispanic” category, which groups together people from numerous heritages, represents a social construction, a pan-ethnic label imposed by a dominant society (Delgado & Stefancic, 2017). The present study, therefore, aims to document and quantify variations in experiences of ethnic discrimination across different intersecting identities of the individuals grouped into the “Hispanic” category.

Literature Review

Diversity of Hispanic Populations

Hispanics trace their origins to regions including Cuba, Mexico, Puerto Rico, and South or Central America (US Census Bureau, 2020), yet data from the 2019 American Community Survey (U.S. Census Bureau, n.d.) reveal substantial variation between Hispanic heritage groups in terms of: median age (28 years for Mexican heritage vs. 48 years for Cuban heritage); educational attainment (among ages 25 and older, 12% with less than a high school education for South American heritage vs. 51% for Guatemalan heritage); nativity (29% foreign-born among those of Mexican heritage vs. 62% among South American heritage); English proficiency (46% of Central American heritage reporting speaking English less than very well vs. 17% of Puerto Rican heritage); medical insurance rates (39% of Honduran heritage uninsured vs. 8% of Puerto Rican heritage); rates of poverty (25% in poverty among Honduran heritage vs. 4% among Uruguayan heritage); and home ownership (63% home owners among Bolivian heritage vs. 28% among Dominican heritage).

Political affiliation and racial identification also vary by Hispanic heritage, with 58% of Cuban-heritage voters identified as Republican, compared to only 32% among non-Cuban Hispanics (Krogstad, 2020), and 86% of Cuban-heritage Hispanics self-identifying as “white” on census questionnaires, compared to 50% of Puerto Rican-heritage Hispanics (Pew Research Center, 2006). Hispanic heritage groups also differ in their geographic settlement patterns in the US; for example, Hispanics of Mexican heritage account for 78% of the Hispanics in the Los Angeles-Long Beach area of California, compared to only 3% of the Hispanics in the Miami, Florida area (Brown & Lopez, 2013). The diversity of US Hispanics is further increased by variation within Hispanic heritage groups in terms of nativity, immigrant generation, length of time lived in the United States, and place of residence.

Types and Measures of Ethnic Discrimination

Measuring ethnic discrimination is an ongoing challenge (Williams et al., 2003), and the results of studies on ethnic discrimination in Hispanics may depend in part on the measures utilized. While some research utilizes overall measures or scales of ethnic discrimination (Alamilla et al., 2010; Torres et al., 2012), other studies focus on specific types or settings of ethnic discrimination, such as discrimination in health care (Lauderdale et al., 2006). The most prevalent type of ethnic discrimination reported among Hispanics varies between studies, based on the types of ethnic discrimination assessed, with some research indicating that racial/ethnic slurs are most commonly-reported (Findling et al., 2019), while other studies conclude that “exclusion/rejection” is the most frequently-reported subtype of ethnic discrimination among Hispanics (Ornelas et al., 2016).

Ethnic discrimination may be explicit (e.g., overt acts such as being called a racist name) or implicit (e.g., mistrust of a patient’s presenting complaints due to racial/ethnic profiling) and also includes structural or institutional-level barriers to resources and opportunities (Krieger et al., 2011). Structural/systemic discrimination operates at levels beyond the interpersonal, often in imperceptible ways. Racial discrimination evolves in a system of power and encompasses the actions not only of individuals but also of institutions that disadvantage racial groups with comparably less power (Gee & Ro, 2009). Structural discrimination based on race/ethnicity limits minority groups’ access to resources, opportunities, and power (Churchwell et al., 2020) via social marginalization, economic dependence, and political disenfranchisement (Bruch et al., 2019). Structural discrimination crosses myriad domains (National Research Council, 2004), including labor markets (e.g., hiring and promotion processes), education (e.g., acceptance, retention), housing (e.g., steering, loan pricing), criminal justice (e.g., policing, sentencing), and healthcare (e.g., access to, and quality of, care). Such discrimination stifles economic, social, and political advancement by restricting social capital, networks of opportunity, and intergenerational wealth transmission (Feagin, 2006; Pager & Shepherd, 2008). Exclusion and subordination of Hispanics is also manifest through the criminalization of immigration infractions, even as immigrant labor is exploited for economic gain at the expense of health and human rights (Churchwell et al., 2020). Eliminating racial inequities may be more a function of changes at the structural than interpersonal level (Gee & Ro, 2009). Nevertheless, examining perceptions of discrimination at the interpersonal level is also important, due to the negative psychological impacts of discrimination that is perceived (Gong et al., 2017).

Ethnic Discrimination in Hispanics: A Critical Race Perspective

While Hispanic ethnicity is defined irrespective of race (US Census Bureau, 2020), ethnicity and race are interconnected in societies with racial inequities (Krieger, 2001), such as American society. Individuals of Hispanic ethnicity in the US are subject to racialization—a social, economic, and political process that considers a population/group as different, underserving, or a threat, potentially producing mistreatment in the forms of stigmatization, exclusion, and punishment (Gans, 2017). Insights from Critical Race Theory (CRT) suggest that social groups in the US are racialized based upon the needs, beliefs, and appraisals of the majority group in power (Delgado & Stefancic, 2017). According to the CRT tenet of differential racialization (Delgado & Stefancic, 2017, p. 79), many Hispanics are subjected to nativist and xenophobic narratives in which they are viewed as undeserving immigrants taking American jobs or exploiting the welfare system; others are ridiculed for speaking accented English; others are considered un-American for speaking Spanish; others are disdained because of their Hispanic names, or required to produce proof of lawful presence on the mere presumption of foreignness; and others are told to return to their home countries, even when the only country they can call home is the United States.

Critical Race Theory describes racism as “ordinary,” deeply embedded in the institutions and systems that form American society (Delgado & Stefancic, 2017). The impacts of systemic racism are far-reaching and permeate core societal components, such as values, identifications, and institutions, with consequences for current and subsequent generations (Feagin, 2006). It follows that, as subjects of racialization (Delgado & Stefancic, 2017), all Hispanics would be impacted by structural discrimination, considering the historical and consistently-disfavored position of Hispanics in the US. Nonetheless, this structural discrimination is often so prevailing and profoundly foundational that its far-reaching effects are not clearly perceived or identified (Delgado & Stefancic, 2017).

The “voice-of-color thesis” of Critical Race Theory (Delgado & Stefancic, 2017, p. 11) posits that racialized individuals, such as US Hispanics, possess competence in discussing race and racism-related issues, by virtue of their social position as a minority group in a racially-stratified society. By recounting their experiences with racial discrimination, people of color allow others to better comprehend the multifaceted nature of racism, while also combating the presumptions, prejudices, and myths propagated by the dominant group (Treviño et al., 2008). Critical Race Theory therefore facilitates an examination of experiences and perceptions of ethnic discrimination, many of which may be located at the interpersonal level. It should be noted, however, that assessing ethnic discrimination via subjective self-reports of experiences does not purport to capture all instances of ethnic discrimination or measure an objective “reality,” as usually defined by the dominant group. Instead, examining the experiences of ethnic discrimination—as Hispanics perceive and identify them—helps in the framing of a problem. By placing marginalized groups’ perspectives at the “center” of social issues (Ford & Airhihenbuwa, 2010), the realities and issues affecting these groups can be more fully comprehended and, perhaps, deconstructed in time.

Prevalence of Ethnic Discrimination Among US Hispanics

Studies based on nationally-representative data from 2012–2013 indicate that approximately 40% of US civilian, non-institutionalized Hispanic adults reported experiencing any form of ethnic discrimination in the past year (Cano, 2019; Cobb et al., 2020), and a more recent national survey of Hispanics, conducted in 2018, estimated a similar prevalence of past-year ethnic discrimination, at 38% (Lopez et al., 2018). Prevalence of perceived ethnic discrimination among Hispanics varies by gender, with men reporting higher levels of discrimination than women (Ornelas et al., 2016) for some, yet not all, types of ethnic discrimination (Arellano-Morales et al., 2015; Findling et al., 2019). Rates of self-reported perceived ethnic discrimination are generally higher among younger, rather than older, Hispanic adults (Araujo-Dawson, 2015; Arellano-Morales et al., 2015; Cobb et al., 2020; Ornelas et al., 2016). Findings regarding variation by socioeconomic status are mixed; some studies report lower odds or prevalence of ethnic discrimination among Hispanics with the highest income level, relative to the lowest income level (Cobb et al., 2020; Findling et al., 2019) yet higher odds of ethnic discrimination among Hispanics with higher education levels (Findling et al., 2019); other studies find no significant associations between ethnic discrimination and education or income yet report associations between several types of ethnic discrimination and employment status (Arellano-Morales et al., 2015).

Population-based studies have consistently reported the lowest levels of ethnic discrimination in Hispanics of Cuban heritage, relative to other Hispanic subgroups (Arellano-Morales et al., 2015; Ornelas et al., 2016; Salas-Wright et al., 2019), with higher levels of ethnic discrimination generally reported among Puerto Rican and Mexican-heritage Hispanics (Arellano-Morales et al., 2015; Ornelas et al., 2016). Sociodemographic differences between heritage groups may account for some of the variation in levels of ethnic discrimination, yet the lowest rates of ethnic discrimination have been reported among Cuban-heritage Hispanics even after controlling for sociodemographic and acculturation-related factors (Arellano-Morales et al., 2015). Beyond heritage group differences, variation in ethnic discrimination has also been reported by geographic region (Arellano-Morales et al., 2015; Cobb et al., 2020) and skin tone (Araujo-Dawson, 2015; Gonzalez-Barrera, 2019).

Ethnic Discrimination and Acculturation-Related Measures

Some evidence suggests that self-reported perceived ethnic discrimination among Hispanics is associated with acculturation-related measures, although the direction of the association has varied between studies, and the association also depends on the measures used for discrimination and acculturation (Arellano-Morales et al., 2015; Findling et al., 2019; Lee & Ahn, 2012). Acculturation is regarded as a multi-faceted construct involving changes in practices, identifications, and values in a culture of heritage and culture of reception (Schwartz et al., 2010), yet proxy measures such as nativity, age at time of immigration, length of stay in the US, and language use have been used to approximate acculturation. Higher levels of lifetime ethnic discrimination have been reported among US-born, versus foreign-born, Hispanic adults (Arellano-Morales et al., 2015; Ornelas et al., 2016); however, higher levels of ethnic discrimination among foreign-born than US-born Hispanics have been reported for the specific experiences of feeling “unaccepted” (Garcini et al., 2018) or being paid unequally or overlooked for promotions (Findling et al., 2019). A national survey of Hispanics in 2018 found that the highest prevalence of ethnic discrimination was observed among second-generation Hispanics, followed by first-generation Hispanics and finally third-or-higher-generation Hispanics (Lopez et al., 2018).

Higher levels of ethnic discrimination have also been reported among participants who responded to a survey in English, as opposed to in Spanish (Ornelas et al., 2016), consistent with the notion that assimilation to the dominant US culture is accompanied by greater exposure to racial/ethnic discrimination (Abraído-Lanza et al., 2016). Finally, Hispanic adults identified as bilingual/bicultural reported the highest prevalence of ethnic discrimination in a latent class analysis based on acculturation and ethnic identity (Salas-Wright et al., 2015). On the one hand, foreign-born Hispanics may often encounter discrimination as a result of being not only Hispanic, but also immigrants, particularly in the midst of the contemporary anti-immigrant environment (Mann-Jackson et al., 2018); on the other hand, US-born Hispanics may be consistently exposed to othering and ethnic discrimination across multiple domains and throughout the life course, therefore learning the dynamics of race and power in the US and their position in society (Viruell-Fuentes, 2007).

Intersecting Identities and Language: Insights from Latino Critical Race Theory

As a branch of CRT, Latino Critical Race Theory (LatCrit) underscores the “multidimensional identity” of Hispanics, comprising various “interdependent intersections” on the basis of race, gender, color, ethnicity, culture, and language (Hernández-Truyol, 1997, p. 883), among other identifications. Language, a particularly-defining marker of identity, contributes to the process of othering for Hispanics (Hernández-Truyol, 1997). Speaking Spanish exposes Hispanic ethnicity in xenophobic environments (Padilla, 2001), serving as “signifier of foreign and outsider status, and a target of nativist hostility” (Davis & Moore, 2014, p. 678).

Language use/proficiency may gauge educational opportunity or time lived in the US (Halim et al., 2017), but also acts as a “tangible index of cultural differentiation” (Alba & Nee, 2003, p. 219) between and within ethnic groups. Within Hispanic populations, language represents an indicator of intergenerational change (Alba & Nee, 2003) and cultural identity, with a relatively high prevalence of Spanish among first-generation Hispanics, Spanish and English often spoken among second-generation Hispanics, and English primarily used by third-generation Hispanics. Language use/proficiency has also been associated with health (Browne-Yung et al., 2013; Walsh, 2018), individual well-being (Browne-Yung et al., 2013; Pitt et al., 2015; Little, 2019; Farr et al., 2018; McCarty et al., 2018; Walsh, 2018), and community well-being (McCarty et al., 2018).

The Present Study

Although prior literature has documented variation in experiences of ethnic discrimination among Hispanics based on demographic characteristics, heritage, region, skin-tone, and acculturation-related measures (Arellano-Morales et al., 2015; Araujo-Dawson, 2015; Cobb et al., 2020; Findling et al., 2019; Gonzalez-Barrera, 2019), the present study is centered around the intersection of multiple identities as they relate to different types and settings of ethnic discrimination. In consideration of the defining role of language in shaping the ways in which individuals “interact with and make sense of the world” (Davis & Moore, 2014, p. 681), the study focuses on English and Spanish use/proficiency as dimensions of Hispanics’ identities which interact with other identifications to shape the ways in which Hispanics experience and perceive ethnic discrimination. The study departs from prior research on language and discrimination in three ways. First, the study examines language use/proficiency on a continuum, in contrast to the frequently-used categorical measures of language use/proficiency (e.g., bilingual; Jones et al., 2019; Salas-Wright et al., 2015) or acculturation proxies with discrete measures (e.g., nativity; Cobb et al., 2020; Findling et al., 2019). The use of continuous measures for language proficiency reflects the numerous possibilities of levels of English and Spanish use/proficiency among Hispanics. Second, the study examines English use/proficiency and Spanish use/proficiency both independently and interdependently, via moderation analysis, capitalizing on a bidimensional scale (Marin & Gamba, 1996). Many of the language and acculturation scales employed in prior studies are unidimensional, zero-sum, in which the degree of acquisition of a second language signifies a corresponding degree of loss of a first language (Marin & Gamba, 1996). The bidimensional conceptualization of language, in contrast, is consistent with LatCrit’s multi-dimensional focus (Valdes, 2005). Finally, also guided by intersectionality and multi-dimensional analysis (Valdes, 2005) the present study examines the extent to which relationships between language identifications and ethnic discrimination experiences vary between Hispanic heritage groups, rather than solely for Hispanics as an overall population.

Methods

Participants

Data were obtained from the 2012–2013 National Epidemiologic Survey on Alcohol and Related Conditions-III (NESARC-III), a cross-sectional, nationally-representative survey of the civilian, non-institutionalized, adult (≥18) population in the United States (excluding US territories). NESARC-III employed multi-stage probability sampling, oversampling areas with relatively high proportions of ethnic minorities. Self-report data were collected by lay interviewers utilizing computer-assisted personal interviewing, and Spanish questionnaires were available. Detailed information about NESARC-III methodology is available elsewhere (Grant et al., n.d.). The limited-access data from NESARC-III was obtained after submitting a data use agreement and notice of Institutional Review Board exemption to the National Institute on Alcohol Abuse and Alcoholism. The study utilized data from the 7,037 participants in NESARC-III who self-identified as of “Hispanic or Latino origin.” With approximately 1.4% of data missing across the variables included in the multivariable analyses, listwise deletion was utilized.

Measures

Outcome variables.

NESARC-III used a modified version of the Experiences of Discrimination (EOD) questionnaire (Krieger et al., 2005) to assess ethnic discrimination. Participants who self-identified as Hispanic or Latino were asked if they had experienced discrimination “because you are Hispanic or Latino” for six different settings or types of ethnic discrimination: 1) discrimination in obtaining healthcare or insurance; 2) discriminatory treatment when receiving care; 3) discrimination in public (on the streets, in stores, or in restaurants); 4) discrimination related to employment, education or training, housing, the courts, the police, or any other setting; 5) being called a racist name; or 6) being verbally or physically threatened or assaulted. Respondents were asked to report the frequency (ranging from never to very often) with which they had experienced discrimination in each of the six settings in the past year. Consistent with prior studies (Cano, 2020; Cano & Takeuchi, 2020; McCabe et al., 2010), responses were dichotomized (never=0; any response besides never=1). These six dichotomous variables indicated whether or not a participant reported experiencing ethnic discrimination (at any frequency) in the past year in each of the six settings. Finally, an overall dichotomous ethnic discrimination variable (Carliner et al., 2016; Cobb et al., 2020) was created to indicate whether a participant reported experiencing ethnic discrimination in the past year at any frequency, in any of the six settings.

Predictors and covariates.

Language use/proficiency was assessed via the Bidimensional Acculturation Scale (BAS; Marin & Gamba, 1996), a 24-item questionnaire with items related to language use, language proficiency, and language of electronic media use, with 12 questions regarding Spanish and the same 12 questions regarding English. The BAS was originally designed as a “measure of acculturation” with two domains, Hispanic and non-Hispanic, guided by the principle that acquiring a second culture does not depend on retaining or losing one’s heritage culture (Marin & Gamba, 1996). Concurrent validity of the scale has been documented in prior research (Marin & Gamba, 1996). In the present study, the BAS was considered a measure of language use/proficiency, rather than acculturation, as advances in the conceptualization of acculturation have identified several other elements of acculturation beyond language (Schwartz et al., 2010).

Results of principal-component factor analyses indicated that a one-factor solution explained 79.8% of the variance in the set of 12 items of the English scale and 72.1% of the variance in the set of 12 items of the Spanish scale. Accordingly, each scale (English or Spanish) was treated as a unidimensional measure. In these data, the reliability coefficient was 0.98 for the English scale and 0.96 for the Spanish scale. Scoring of each item ranged from 1 to 4, with higher scores indicating greater English or Spanish use/proficiency. A mean score was computed for the English scale, hereafter referred to as English use/proficiency; a separate mean score was computed for the Spanish scale, hereafter referred to as Spanish use/proficiency.

Immigrant generation was coded based on participants’ responses to questions regarding their birthplace and the birthplace of their parents and grandparents. First generation indicates that a participant was born outside of the 50 states or District of Columbia ([DC]; individuals born in US territories such as Puerto Rico were included in the “first generation”). Second-generation participants were born in the 50 states/DC to at least one parent born outside of the 50 states/DC. Third-generation participants were born in the 50 states/DC to US-born parents yet had at least one foreign-born grandparent. Finally, higher-than-third-generation indicated that the participant, both parents, and all grandparents were born in the 50 states/DC.

Sociodemographic characteristics included gender (men/women), age category (18–29, 30–44, 45–64, 65+), educational attainment (less than high school, high school/ General Education Diploma, some college, Bachelor’s degree or higher), total combined family income in the past year (0–19,999; 20,000–34,999; 35,000–69,999; 70,000+), and heritage. Heritage was based on the participant’s report of the country of heritage or ancestry with which they identified most (excluding the US). The following heritage categories were utilized: Mexico; Cuba; Dominican Republic; Central America; Puerto Rico; Spanish-speaking South America; and “other,” for countries not included in the aforementioned categories.

Statistical Analysis

Analyses were conducted in Stata/MP 16.1 with the svy suite of commands; variance estimates were produced using Taylor linearization. First, two-way tabulations (with Pearson chi-square statistics corrected for the survey design) calculated prevalence estimates (and standard errors [SE]) of past-year self-reported perceived ethnic discrimination overall, and in each setting, by demographic groups (Table 1). Next, the distribution of participants’ English use/proficiency and Spanish use/proficiency scores was depicted as a density plot, using R (Figure 1). Binomial logistic regression models were subsequently utilized to calculate adjusted odds ratios (with 95% Confidence Intervals [CIs]) and predicted probabilities of self-reported perceived ethnic discrimination by language use/proficiency; predicted probabilities were plotted with Stata’s margins suite of commands. The following models were examined:

Table 1.

Past-year Prevalence of Self-Reported Perceived Ethnic Discrimination, Overall and in Six Settings, Among US Hispanic Adults, 2012–2013, by Selected Demographic Characteristics

		Type or Setting of Ethnic Discrimination, %(SE)
	Ethnic discrimination, any type/setting	Obtaining health care/ insurance	Receiving care	In public (streets, stores, or restaurants)	Employment, education, housing, courts, police, other	Called a racist name	Verbally/physically threatened/assaulted
Overall	40.0(0.9)	18.1(0.7)	16.4(0.6)	27.7(0.9)	21.4(0.7)	16.5(0.6)	7.9(0.4)
Gender
Men	41.5(1.2)	18.4(0.9)	16.4(0.8)	27.6(1.2)	23.0(1.0)	19.7(0.8)	9.1(0.6)
Women	38.4(1.0)	17.8(0.7)	16.5(0.7)	27.7(0.9)	19.8(0.8)	13.3(0.6)	6.8(0.5)
p value	<0.05	0.55	0.96	0.91	<0.01	<0.001	<0.01
Age, years
18–29	42.1(1.7)	16.0(1.1)	14.2(1.0)	28.9(1.3)	21.8(1.1)	22.2(1.3)	8.7(0.7)
30–44	45.2(1.4)	21.3(1.1)	19.7(1.0)	32.9(1.3)	25.3(1.2)	17.1(0.9)	8.7(0.7)
45–64	36.9(1.5)	18.6(1.1)	16.7(1.0)	24.3(1.3)	19.5(1.1)	12.9(0.8)	7.5(0.7)
65 or older	20.9(1.9)	11.0(1.4)	9.9(1.3)	13.0(2.0)	9.9(1.5)	5.4(1.2)	3.3(0.8)
p value	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001
Educational attainment
Less than high school	40.5(1.6)	23.7(1.2)	20.9(1.2)	27.8(1.4)	22.6(1.4)	15.2(1.0)	8.6(0.7)
High school/GED	40.5(1.5)	17.7(1.0)	15.7(0.9)	28.2(1.4)	22.1(1.2)	17.4(1.1)	8.3(0.8)
Some college	41.0(1.6)	15.9(0.8)	15.3(0.8)	28.6(1.4)	21.9(1.2)	18.4(1.1)	8.0(0.7)
Bachelor’s or higher	35.5(2.1)	10.6(1.3)	9.7(1.1)	24.1(2.0)	15.9(1.5)	13.3(1.4)	5.4(0.8)
p value	0.19	<0.001	<0.001	0.27	<0.05	<0.05	0.05
Family income, $
0–19,999	39.1(1.5)	21.0(1.1)	18.8(1.2)	26.9(1.3)	21.8(1.2)	17.3(1.1)	8.6(0.8)
20,000–34,999	40.8(1.6)	21.0(1.3)	17.8(1.1)	29.5(1.5)	22.5(1.4)	16.5(1.0)	8.4(0.8)
35,000–69,999	43.9(1.5)	17.5(1.0)	17.0(1.0)	30.2(1.5)	22.6(1.2)	17.1(1.0)	8.3(0.8)
70,000 or higher	34.2(1.8)	10.5(1.1)	10.2(1.0)	22.5(1.6)	17.4(1.4)	14.4(1.3)	5.7(0.8)
p value	<0.001	<0.001	<0.001	<0.001	<0.05	0.33	0.08
Heritage†
Mexico	44.1(1.5)	21.0(0.9)	18.2(0.9)	31.6(1.3)	24.4(1.1)	18.8(0.9)	9.4(0.7)
Cuba	19.0(2.9)	9.2(2.0)	5.5(1.9)	9.7(2.1)	8.4(2.5)	4.9(1.5)	2.2(1.0)
Dominican Republic	29.4(3.7)	12.8(2.3)	15.2(2.9)	19.8(2.6)	15.2(2.2)	8.1(1.9)	5.3(1.4)
Central America	45.3(2.3)	23.3(1.8)	21.7(1.8)	30.7(2.2)	25.9(1.9)	18.6(1.9)	10.6(1.3)
Puerto Rico	40.9(2.6)	14.9(2.1)	15.5(1.8)	27.7(2.7)	20.7(2.4)	17.4(2.4)	6.5(1.1)
South America	34.8(2.2)	19.5(1.7)	17.9(1.7)	19.7(1.9)	17.2(2.1)	10.0(1.4)	5.0(1.1)
Other	32.9(1.8)	10.8(1.0)	10.7(1.0)	23.0(1.9)	16.2(1.7)	14.7(1.2)	6.0(0.9)
p value	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001
Immigrant generation§
1st	40.8(1.3)	23.6(1.0)	19.9(0.9)	27.4(1.0)	22.8(1.0)	14.8(0.8)	7.9(0.6)
2nd	41.0(1.6)	13.3(1.0)	13.5(1.0)	29.5(1.6)	21.2(1.1)	19.4(1.3)	8.4(0.8)
3rd	38.2(2.1)	9.8(1.1)	11.9(1.2)	27.1(2.2)	19.1(1.7)	17.9(1.6)	7.8(1.1)
>3rd	35.3(2.2)	11.2(1.2)	11.3(1.1)	25.0(2.2)	17.5(1.7)	16.2(2.1)	6.9(1.4)
p value	0.14	<0.001	<0.001	0.35	<0.05	<0.05	0.79

Notes: Weighted results from the National Epidemiologic Survey on Alcohol and Related Conditions-III (NESARC-III).

^†Heritage country is irrespective of birthplace;

^§1st generation, foreign-born; 2nd generation, US-born with at least one foreign-born parent; 3rd generation, US-born with at least one foreign-born grandparent; >3rd generation, US-born with no foreign-born parents or grandparents. P values are for the corrected chi square statistic (i.e., design-based F), corresponding to the bivariate analyses of each demographic category and a given type/setting of ethnic discrimination. Exact p values are reported for values larger than or equal to 0.05. Abbreviations: US, United States; SE, standard error; GED, general education diploma

Figure 1. Distribution of participants’ combinations of English use/proficiency scores and Spanish use/proficiency scores.

Note. Unweighted results from the National Epidemiologic Survey on Alcohol and Related Conditions-III (NESARC-III), 2012–2013. The color reflects the proportion of the sample with any given English/Spanish score.

1. Ethnic discrimination, overall and in each of six settings, predicted from English use/proficiency and Spanish use/proficiency, adjusting for gender, age, educational attainment, family income, and heritage. Main effects of English use/proficiency and Spanish use/proficiency were modeled first, followed by interaction effects of the continuous measure of English use/proficiency by the continuous measure of Spanish use/proficiency (Table 2 and Figure 2).

2. Ethnic discrimination, overall, predicted from: a) the interaction of heritage and English use/proficiency, adjusting for gender, age, educational attainment, family income, and Spanish use/proficiency (Table 3 and Figure 3, left panel); b) the interaction of heritage and Spanish use/proficiency, adjusting for gender, age, educational attainment, family income, and English use/proficiency (Table 3 and Figure 3, right panel).

3. Ethnic discrimination, overall, disaggregated by Hispanic heritage, predicted from English use/proficiency and Spanish use/proficiency, adjusting for gender, age, educational attainment, and family income. Main effects of English use/proficiency and Spanish use/proficiency were modeled first, followed by interaction effects of the continuous measure of English use/proficiency by the continuous measure of Spanish use/proficiency (Table 4 and Figure 4).

Table 2.

Association between past-year self-reported perceived ethnic discrimination (overall and by type/setting) and language use/proficiency, for US Hispanic adults, 2012–2013

		Type or Setting of Ethnic Discrimination
	Ethnic discrimination, any type/setting	A. Obtaining health care/ insurance	B. Receiving care	C. In public (streets, stores, or restaurants)	D. Employment, education, housing, courts, police, other	E. Called a racist name	F. Verbally/physically threatened/assaulted
	AOR (95%CI)	AOR (95%CI)	AOR (95%CI)	AOR (95%CI)	AOR (95%CI)	AOR (95%CI)	AOR (95%CI)
Main effectsa
English use/proficiency	1.26*** (1.16–1.38)	1.02 (0.92–1.13)	1.09 (0.98–1.22)	1.30*** (1.17–1.43)	1.16** (1.04–1.30)	1.42*** (1.25–1.61)	1.25** (1.06–1.48)
Spanish use/proficiency	1.22*** (1.12–1.33)	1.41*** (1.27–1.57)	1.29*** (1.14–1.45)	1.21*** (1.09–1.34)	1.18** (1.07–1.31)	1.11 (0.99–1.25)	1.12 (0.95–1.30)
Interaction effectsb
English use/proficiency × Spanish use/proficiency	1.65*** (1.44–1.89)	1.66*** (1.45–1.89)	1.58*** (1.37–1.83)	1.62*** (1.42–1.85)	1.68*** (1.48–1.92)	1.54*** (1.35–1.76)	1.71*** (1.43–2.06)

Notes: Weighted results from the National Epidemiologic Survey on Alcohol and Related Conditions-III (NESARC-III). All models are adjusted for sociodemographic characteristics (i.e., gender, age, educational attainment, family income, and heritage). Although analyses represent associations only, the term “effects” is utilized for convention in the context of moderation analysis. Abbreviations. AOR, adjusted odds ratios; CI, confidence interval.

^aThe main effects models regress ethnic discrimination (overall or by type/setting) on English use/proficiency, Spanish use/proficiency, and sociodemographic characteristics.

^bThe interaction effects models regress ethnic discrimination (overall or by type/setting) on the interaction of English use/proficiency and Spanish use/proficiency, as well as English use/proficiency, Spanish use/proficiency, and sociodemographic characteristics. The interaction effects models only show the regression coefficients for the interaction term.

^*p < 0.05;

^**p < 0.01;

^***p < 0.001

Figure 2. Predicted probabilities of six different types/settings of self-reported perceived past-year ethnic discrimination, based on English and Spanish use/proficiency, for US Hispanic Adults, 2012–2013.

Notes: Weighted results from the National Epidemiologic Survey on Alcohol and Related Conditions-III, 2012–2013. The range of scores for language use/proficiency is 1–4, with higher scores indicating greater use/proficiency of English or Spanish. Results for each of six regression models including the continuous-by-continuous interaction of English use/proficiency × Spanish use/proficiency, English use/proficiency, Spanish use/proficiency, sex, age, educational attainment, family income, and Hispanic heritage.

Table 3.

Results of binomial logistic regression of self-reported perceived past-year ethnic discrimination on the interaction of language use/proficiency and Hispanic heritage, for US Hispanic adults

Interaction term	AOR (95% CI)	Interaction term	AOR (95% CI)
English use/proficiency × heritage (ref., Mexican heritage)		Spanish use/proficiency × heritage (ref., Mexican heritage)
× Cuban	0.75 (0.93–1.58)	× Cuban	0.78 (0.51–1.19)
× Dominican Republic	0.92 (0.67–1.27)	× Dominican Republic	1.25 (0.66–2.34)
× Central American	0.96 (0.78–1.19)	× Central American	0.83 (0.65–1.07)
× Puerto Rican	0.87 (0.67–1.13)	× Puerto Rican	1.21* (1.00–1.46)
× South American	0.61*** (0.48–0.78)	× South American	1.41* (1.06–1.88)
× Other Hispanic heritage	0.75 (0.58–0.97)	× Other Hispanic heritage	1.29** (1.09–1.53)

Notes: Weighted results from the National Epidemiologic Survey on Alcohol and Related Conditions-III (NESARC-III). The regression model including the interaction of English use/proficiency and Hispanic heritage adjusts for Spanish use/proficiency, gender, age, educational attainment, and family income. The regression model including the interaction of Spanish use/proficiency and Hispanic heritage adjusts for English use/proficiency, gender, age, educational attainment, and family income. Abbreviations. AOR, adjusted odds ratio; CI, confidence interval; ref., reference category.

^*p < 0.05;

^**p < 0.01;

^***p < 0.001

Figure 3. Predicted probability of self-reported perceived past-year ethnic discrimination based on English use/proficiency (left-hand side panel) or Spanish use/proficiency (right-hand side panel) by Hispanic heritage, for US Hispanic adults.

Notes. Weighted results from the National Epidemiologic Survey on Alcohol and Related Conditions-III, 2012–2013. Results from the regression of past-year ethnic discrimination on the interaction of Hispanic heritage and language use/proficiency, adjusting for gender, age, educational attainment, family income and Spanish use proficiency (in the model examining English use/proficiency) or English use/proficiency (in the model examining Spanish use/proficiency). Abbreviations: Mex, Mexican; Dom Rep, Dominican Republic; Cen-Ame, Central-American; Pto Rico, Puerto Rican; Sou-Ame, South American.

Table 4.

Association between self-reported perceived past-year ethnic discrimination and language use/proficiency by Hispanic heritage group, for US Hispanic Adults, 2012–2013.

	Mexican heritage	Cuban heritage	Dominican heritage	Central American heritage	Puerto Rican heritage	South American heritage
	AOR (95%CI)	AOR (95%CI)	AOR (95%CI)	AOR (95%CI)	AOR (95%CI)	AOR (95%CI)
Main effects modela
English use/proficiency	1.40*** (1.23–1.58)	1.27 (0.73–2.22)	1.43* (1.00–2.04)	1.12 (0.85–1.47)	1.11 (0.74–1.67)	0.75 (0.53–1.07)
Spanish use/proficiency	1.20** (1.05–1.37)	0.94 (0.50–1.78)	1.69 (0.73–3.88)	0.85 (0.62–1.17)	1.39* (1.08–1.81)	1.17 (0.79–1.72)
Interaction modelb
English use/proficiency × Spanish use/proficiency	1.63*** (1.35–1.97)	1.05 (0.48–2.30)	0.99 (0.31–3.16)	2.00*** (1.46–2.73)	1.64* (1.03–2.62)	2.24** (1.24–4.05)

Notes: Weighted results from the National Epidemiologic Survey on Alcohol and Related Conditions-III (NESARC-III). All models are adjusted for sociodemographic characteristics (i.e., gender, age, educational attainment, and family income). Although analyses represent associations only, the term “effects” is utilized for convention in the context of moderation analysis. Abbreviations. AOR, adjusted odds ratios; CI, confidence interval.

^aThe main effects models regress ethnic discrimination (for each Hispanic heritage group) on English use/proficiency, Spanish use/proficiency, and sociodemographic characteristics.

^bThe interaction effects models regress ethnic discrimination (for each Hispanic heritage group) on the interaction of English use/proficiency and Spanish use/proficiency, as well as English use/proficiency, Spanish use/proficiency, and sociodemographic characteristics. The interaction effects models only show the regression coefficients for the interaction term.

^*p < 0.05;

^**p < 0.01;

^***p < 0.001

Figure 4. Predicted probability of past-year ethnic discrimination by Hispanic heritage group, based on English and Spanish use/proficiency, for US Hispanic adults, 2012–2013.

Notes. Weighted results from the National Epidemiologic Survey on Alcohol and Related Conditions-III, 2012–2013. The range of scores for language use/proficiency is 1–4, with higher scores indicating greater use/proficiency of English or Spanish. Results for each of six regression models including the continuous-by-continuous interaction of English use/proficiency × Spanish use/proficiency, English use/proficiency, Spanish use/proficiency, sex, age, educational attainment, and family income.

Results

Prevalence of Ethnic Discrimination Overall and in Specific Settings

Table 1 presents prevalence estimates for the overall measure of ethnic discrimination, encompassing any type of self-reported perceived ethnic discrimination in the past year, as well as prevalence estimates for each of the six different types or settings of ethnic discrimination. Overall, 40.0% (SE, 0.9) of civilian, non-institutionalized Hispanic adults in the United States reported experiencing ethnic discrimination within the prior year (in any setting, at any frequency). The most frequently-reported setting of ethnic discrimination was “in public (on the streets, in stores, or in restaurants),” with 27.7% (SE, 0.9) of US Hispanic adults reporting such discrimination in the past year, while the least frequently-reported type of ethnic discrimination was verbal/physical threat/assault, with 7.9% (SE, 0.4) of Hispanic adults reporting this experience of discrimination in the past year.

Demographics

Relative to women, men reported a slightly higher prevalence of past-year ethnic discrimination overall (38.4% vs. 41.5%, p<0.05) and in three different settings/types of discrimination: employment, education, housing, the courts, or with police (p<0.01); being called “a racist name” (p<0.001); or being verbally/physically threatened/assaulted (p<0.01). The prevalence of overall ethnic discrimination, as well as ethnic discrimination in each of the six different settings, was lowest among Hispanics ages 65 and older. Prevalence estimates were highest in ages 30–44 for all settings of discrimination, with the exception of “being called a racist name,” which was most prevalent among those aged 18–29, and verbal/physical threats/assaults, which were equally prevalent in the age groups 18–29 and 30–44.

Prevalence of the overall self-reported perceived ethnic discrimination measure did not vary significantly by educational attainment; nonetheless, the prevalence of ethnic discrimination in obtaining health care/insurance or when receiving care was more than twice as high among Hispanic adults with less than a high school education, compared to those with a Bachelor’s degree or higher. Hispanics with family incomes at or above $70,000 yearly evidenced the lowest prevalence of discrimination overall and in obtaining health care/insurance, when receiving care, in public, or in employment, education, housing, courts, or with the police; however, no significant differences based on income were observed for being “called a racist name” or verbal/physical threats/assaults.

Immigrant Generation and Heritage

Prevalence of self-reported perceived ethnic discrimination overall did not vary significantly by immigrant generation, yet significant differences were observed by generation for several specific types or settings of ethnic discrimination. Ethnic discrimination related to obtaining health care/insurance or receiving care was highest in first-generation (i.e., foreign-born) Hispanic adults, while reports of being “called a racist name” were highest among the second-generation (i.e., US-born with at least one foreign-born parent). The prevalence of ethnic discrimination in public or verbal/physical threats/assaults did not significantly vary between generations. When considering specific Hispanic heritage groups, the overall prevalence of past-year ethnic discrimination ranged from 19.0% (SE, 2.9) among adults of Cuban heritage to 45.3% (SE, 2.3) among adults of Central American heritage. Cuban-heritage adults consistently evidenced the lowest prevalence of ethnic discrimination across all six settings/types, while the highest prevalence of various types or settings of ethnic discrimination was evidenced by those of Central American or Mexican heritage.

Ethnic Discrimination by English and Spanish Use/Proficiency

Figure 1 depicts the unweighted distribution of participants’ English use/proficiency and Spanish use/proficiency scores. Each participant received both an English score and a Spanish score, with each score independent from the other; scores ranged from one to four, with higher values indicating higher use/proficiency of the given language. As depicted in Figure 1, the highest proportion of scores (red-shaded area) was characterized by high English use/proficiency with low Spanish use/proficiency, followed by high English use/proficiency with moderate Spanish use/proficiency (orange-shaded area), and high Spanish use/proficiency with low to moderate English use/proficiency (yellow-shaded area); relatively fewer participants evidenced a low English score with a low Spanish score (blue-shaded area).

Language Use/Proficiency and Different Settings of Ethnic Discrimination

Table 2 presents adjusted odds ratios from the regressions of the overall self-reported perceived ethnic discrimination measure, and the six types/settings of ethnic discrimination, on English and Spanish use/proficiency, adjusted for sociodemographic characteristics and heritage. Table 2 includes both main effects models and interaction effects models. In the main effects models (with each type/setting of ethnic discrimination independently regressed on English use/proficiency, Spanish use/proficiency, and sociodemographic characteristics), higher Spanish—yet not English—use/proficiency was significantly associated with higher odds of self-reporting past-year perceived ethnic discrimination in obtaining health care/insurance (AOR, 1.41; 95% CI, 1.27–1.57), or receiving care (AOR, 1.29; 95% CI, 1.14–1.45). Higher Spanish use/proficiency and English use/proficiency were each significantly associated with higher odds of ethnic discrimination in public or in employment, education, housing, the courts, or with the police, as well as the overall ethnic discrimination measure. Finally, higher English—yet not Spanish—use/proficiency was significantly associated with higher odds of reporting being called a racist name (AOR, 1.42; 95% CI, 1.25–1.61) or being verbally/physically threatened/assaulted (AOR, 1.25; 95% CI, 1.06–1.48) based on Hispanic ethnicity.

In the interaction effects model (Table 2, bottom panel), the interaction term of English and Spanish use/proficiency was statistically significant for each of the six types/settings of ethnic discrimination and the overall ethnic discrimination measure. To ease interpretation of these results, Figure 2 depicts predicted probabilities of reporting each of the six types/settings of ethnic discrimination, based on the continuous-by-continuous interaction of English and Spanish use/proficiency scores (adjusting for sociodemographic characteristics and heritage). Results indicate that Hispanic adults with the highest English use/proficiency coupled with the lowest Spanish use/proficiency evidenced the lowest predicted probabilities of self-reported perceived ethnic discrimination in obtaining health care/insurance or receiving care. Those with the lowest English use/proficiency coupled with the highest Spanish use/proficiency evidenced the lowest predicted probabilities of perceived ethnic discrimination in public or being called a racist name. Compared to adults with high scores on only one of the languages (either English or Spanish) and low scores for the other language, adults with high English coupled with high Spanish evidenced higher predicted probabilities of reporting perceived ethnic discrimination in employment, education, housing, or in public.

Ethnic Discrimination and Language Use/Proficiency for Different Hispanic Heritage Groups

Table 3 presents adjusted odds ratios from the regression of self-reported perceived ethnic discrimination on: a) the interaction of heritage group and English use/proficiency (adjusted for gender, age, educational attainment, family income, and Spanish use/proficiency); b) the interaction of heritage group and Spanish use/proficiency (adjusted for gender, age, educational attainment, family income, and English use/proficiency). Predicted probabilities from these regression models are displayed in Figure 3. Predicted probabilities of ethnic discrimination were nearly five times as high for South Americans, relative to Cubans (difference, p <0.001), at the lowest level of English use/proficiency, yet probabilities of ethnic discrimination converged as English use/proficiency increased, such that probabilities did not significantly differ (p=0.61) at the highest level of English use/proficiency. Similarly, predicted probabilities of ethnic discrimination were more than twice as high (difference, p <0.001) for Mexicans, compared to South Americans, with the lowest Spanish use/proficiency scores, yet predicted probabilities of ethnic discrimination did not significantly differ (p=0.46) between Mexicans and South Americans with the highest Spanish use/proficiency scores.

Table 4 presents results from the regression of English use/proficiency and Spanish use/proficiency on the overall ethnic discrimination measure, stratified by heritage group, in both a main effects model and interaction model. In the main effects model, both English use/proficiency and Spanish use/proficiency are included in the model, along with sociodemographic covariates. For those of Mexican heritage, higher scores on either English (AOR, 1.40; 95% CI, 1.23–1.58) or Spanish use/proficiency (AOR, 1.20; 95% CI, 1.05–1.37) were significantly associated with higher odds of reporting perceived ethnic discrimination. Higher English use/proficiency was associated with higher odds of ethnic discrimination among those of Dominican heritage (AOR, 1.43; 95% CI, 1.00–2.04), while higher Spanish use/proficiency was significantly associated with higher odds of ethnic discrimination among those of Puerto Rican heritage (AOR, 1.39; 95% CI, 1.08–1.81).

In Figure 4, when predicting past-year perceived ethnic discrimination based on the interaction of English use/proficiency with Spanish use/proficiency (adjusting for sociodemographic characteristics), patterns differed by Hispanic heritage group, with the Cuban and Dominican heritage groups most notably diverging from the pattern observed in other Hispanic heritage groups. As presented in Table 4 (interaction model), the interaction of English and Spanish language use/proficiency was statistically significant for all Hispanic heritage groups except Cuban and Dominican heritage.

Limitations

Although the cross-sectional data utilized in the study were collected in 2012–2013, prior to notable events in the US sociopolitical environment, these data provide a baseline of the prevalence of ethnic discrimination among US Hispanics prior to the rhetoric and policies surrounding, and consequential to, the 2016 US presidential election. These data also represent the most recent nationally-representative data with detailed measures related to Hispanic heritage and English and Spanish use/proficiency. Nonetheless, the data do not assess critical types of ethnic discrimination (e.g., wealth accumulation, quality of legal representation, steering; National Research Council, 2004), and the data do not include relevant information on skin color, race, immigrant documentation status, or specific area of residence, all of which may play a role in precipitating and shaping discrimination experiences for Hispanic Americans (Arellano-Morales et al., 2015; Gonzalez-Barrera, 2019). While the study examined several layers of the social identities of Hispanic individuals in a relatively large sample, examinations of subgroups disaggregated by other additional layers of identity (e.g., gender plus Hispanic heritage plus immigrant generation plus education level) were precluded due to small numbers in these subgroups. Finally, data on indigenous ancestries and languages were lacking in the data source. Discrimination due to indigenous ancestry likely complicates experiences of discrimination due to Hispanic ethnicity, as discrimination may not only be sourced in outgroup, but also other ingroup Hispanic members (Sanchez, 2018).

Discussion

This study examined self-reported past-year perceived ethnic discrimination among US Hispanic adults, focusing on variation in perceived exposure to discrimination, as well as the types/settings of ethnic discrimination reported. Overall, study findings emphasize that experiencing some form of ethnic discrimination is relatively common among US Hispanic adults, yet the prevalence and types or settings of ethnic discrimination vary widely on the basis of demographics, immigrant generation, language use/proficiency, and heritage. Study findings also highlight the interplay between English use/proficiency and Spanish use/proficiency in shaping perceived ethnic discrimination among diverse US Hispanic subgroups.

English and Spanish Use/Proficiency

English use/proficiency and Spanish use/proficiency were examined as two independent, yet also intertwined, continuous measures, as markers of Hispanics’ multi-dimensional social identities (Hernández-Truyol, 1997). Both English and Spanish use/proficiency were positively associated with increased adjusted odds of reporting ethnic discrimination overall, in public, or with respect to employment/education/housing/courts/police; however, with respect to being called a racist name or receiving verbal/physical threats/assaults, a positive association was observed for English, yet not Spanish. Hispanics who live and work in predominantly English-only settings may consequently report both greater English use/proficiency and more exposure to racist language and threats, compared to peers who spend most of their time surrounded by Spanish speakers in an ethnic enclave (Viruell-Fuentes, 2007). At the same time, those who understand and/or speak English may more easily identify racist language in English, regardless of level of Spanish use/proficiency. In a study of Hispanic adults in four large metropolitan areas, higher scores in English than Spanish were associated with ethnicity-based threats or aggression (Arellano-Morales et al., 2015). It is possible that although many Hispanics are US citizens or speak English fluently, they are nonetheless exposed to ethnic discrimination in various settings, as the spillover effects of nativist public discourse and xenophobic policies are not confined to the foreign-born, but also extend to US-born, English speaking Hispanics.

Study findings also indicated a significant interaction between English use/proficiency and Spanish use/proficiency when predicting past-year perceived ethnic discrimination overall or in any of the six types/settings examined. Through the lens of LatCrit (Delgado & Stefancic, 2017; Hernández-Truyol, 1997; Villalpando, 2003), Hispanics’ varying degrees of English and Spanish use/proficiency represent independent and interdependent dimensions of identity intersecting to shape lived experiences. Findings revealed that the likelihood of self-reporting perceived ethnic discrimination in public or with respect to employment/education/housing/courts/police was higher among Hispanics with the highest English use/proficiency coupled with the highest Spanish use/proficiency (“highly bilingual Hispanics”), relative to Hispanics with high use/proficiency in one language only and low use/proficiency in the other language (English or Spanish “monolingual Hispanics”). In a study predicting ethnic discrimination among Hispanic adults based on latent classes of acculturation-related variables, the strongest association was observed with the latent class characterized as bilingual/bicultural (Salas-Wright et al., 2015). By virtue of speaking Spanish, highly bilingual/bicultural Hispanics may be more likely to be viewed as foreign or “other,” relative to English monolingual peers; yet, as English speakers, they may more easily identify racist speech in English and spend more time in settings in which they are an ethnic minority, relative to Spanish monolingual peers. Close ties to highly-stigmatized Hispanics (such as undocumented Hispanic immigrants), may also place English-dominant Hispanics in a position of increased sensitivity toward noticing, labeling, and self-reporting experiences of ethnic/racial discrimination (Jones et al., 2019). Qualitative research highlights an acute awareness of minority status among second-generation Mexican-heritage women, likely bilingual, informed by discriminatory experiences in a variety of settings beginning early in life and a history of witnessing discrimination against parents or other Hispanics (Viruell-Fuentes, 2007).

Differences Based on Hispanic Heritage

Results of this study indicated significant variation by Hispanic heritage group in the relationship between language use/proficiency and ethnic discrimination. For example, for Hispanics with the highest Spanish use/proficiency level (irrespective of English use/proficiency level), the probability of perceived ethnic discrimination did not differ between those of Mexican and South American heritage; however, the probability of perceived ethnic discrimination was double among Mexicans, compared to South Americans, with the lowest Spanish use/proficiency (irrespective of English use/proficiency level). It is possible that many Hispanics of Mexican heritage may stand out as “foreign” or “minority” due to phenotype, even if they do not speak Spanish, while many of those of certain South American heritages may blend in more seamlessly with the white mainstream US, as long as they do not speak Spanish. According to data from the 2019 American Community Survey (Ruggles et al., 2020), approximately 89% of Americans who identify with Argentinean heritage also identify as of white race, compared with only 68% of those of Mexican heritage. This observation, however, is tempered by variation between and within South American countries, as well as the distinction between self-identifying as white versus being seen by others—and treated by others—as white. These findings, however, suggest that, even if use of the Spanish language in xenophobic/nativist settings can be a marker of otherness or foreignness (Negrón, 2018), the (differential) impact of language may be a function of Hispanic heritage and, likely, race as well.

At the highest level of English use/proficiency (irrespective of Spanish use/proficiency level), probabilities of perceived ethnic discrimination did not differ between South Americans and Cubans, yet probabilities of perceived ethnic discrimination were nearly five times as high for South Americans, relative to Cubans, with the lowest level of English use/proficiency (also irrespective of Spanish use/proficiency level). Some Cuban-Americans with low English use/proficiency may be shielded from “othering” by virtue of living in areas surrounded by individuals with a similar ethnic background and language, more so than many South Americans with low English use/proficiency, as South American communities are relatively more scattered across the US, compared to Cuban communities; approximately 67% of the entire US Cuban population resides in Florida; in contrast, the state with the largest population of South Americans only comprises 27% of their US population (U.S. Census Bureau, n.d.).

The relatively low prevalence of ethnic discrimination among Hispanics of Cuban heritage, relative to other Hispanic heritage groups, is consistent with findings from prior studies (Arellano-Morales et al., 2015; Perez et al., 2008; Salas-Wright et al., 2019). Beyond residence in ethnic enclaves, other possible explanations for the lower prevalence of reported ethnic discrimination among Cuban-Americans include social capital (Portes & Sensenbrenner, 1993), socioeconomic status (Arellano-Morales et al., 2015), racial composition, and historic immigration protections (Pew Research Center, 2016).

Dominican-heritage Hispanics, in turn, also evidenced a pattern of discrimination that differed slightly from other Hispanic heritage groups; that is, a significant interaction between English use/proficiency and Spanish use/proficiency was observed for Hispanics of Mexican, Central American, Puerto Rican, or South American heritage, yet not for Hispanics of Cuban or Dominican heritage. Dominican-heritage Hispanics also evidenced the second-lowest prevalence of ethnic discrimination among Hispanic groups in the study. One possible explanation for this finding is that the measure of ethnic discrimination in the study inquired about discrimination “because you are Hispanic or Latino,” that is, discrimination based on ethnicity, rather than race. Experiences of racial (as opposed to ethnic) discrimination may particularly salient for some Hispanics of Dominican heritage, as socially ascribed race may have stronger impacts on experiencing discrimination for some Dominicans than English or Spanish use/proficiency alone. Hispanics of Dominican heritage constitute the Hispanic heritage group with the lowest proportion self-identifying as “white,” with 15% identifying as Black and 42% identifying their race as “other” in Census questionnaires (Ruggles et al., 2020). Dominicans are more likely to be socially ascribed as Black because of the reflection of African ancestry in their phenotype (e.g., natural/textured hair, darker skin color; Borrell, 2005; Torres-Saillant, 2010). Hispanics who identify as Black may experience higher levels of discrimination (Cuevas et al., 2016), and socially assigned race has been shown to impact the likelihood of discrimination among Hispanics (Vargas et al., 2016). A mismatch between self-perceived race and socially assigned race has also been documented as part of Dominicans’ lived experiences in the US (Figuereo, 2020; Roth, 2010). Although many Dominican-heritage Hispanics reside in ethnic enclaves (Martins et al., 2014), with half of the Dominican-heritage population in the US residing in New York/New Jersey alone (US Census Bureau, n.d.), Black identity is often rejected within Dominican communities, and being socially ascribed as Black may result in intragroup marginalization and experiences of discrimination based on phenotype (e.g., “pelo malo” and colorism) from lighter-skinned members of families and communities (Haywood, 2017; Hordge-Freeman & Veras, 2020).

Hispanics of Central American heritage and Mexican heritage evidenced the highest prevalence of ethnic discrimination, followed by Puerto Rican-heritage Hispanics, partially consistent with patterns in prior studies in which Puerto Ricans represented the Hispanic heritage group with the highest levels of ethnic discrimination (excluding multiple-heritage Hispanics; Arellano-Morales et al., 2015; Ornelas et al., 2016). Although Puerto Rican-heritage Hispanics are US citizens by birth and therefore do not face the same immigration barriers as most other Hispanic heritage groups, they are nonetheless frequent targets of anti-immigrant and anti-Hispanic sentiments. Many Puerto Rican-heritage Hispanics may also possess a heightened awareness of the racial/ethnic power dynamics in US society, considering the history of slavery in Puerto Rico, the colonial history of Puerto Rico, and the current status of Puerto Rico as a territory under US sovereignty whose residents are not afforded the same rights (e.g., voting in presidential elections, representation in Congress, eligibility for certain government-funded assistance) as residents of the 50 states/DC (Torruella, 2017).

Settings or Types of Ethnic Discrimination

The study examined both an overall measure of perceived ethnic discrimination and six different types/settings of ethnic discrimination. Findings underscore that demographic patterns observed with the overall measure of ethnic discrimination do not necessarily apply to specific types or settings of ethnic discrimination. For example, the highest income group reported the lowest prevalence of the overall measure of ethnic discrimination, consistent with some prior studies (Cobb et al., 2020; Findling et al., 2019), yet this pattern did not extend to experiences of racist names or verbal/physical threats/assaults, for which no significant differences were observed based on income. Similarly, for the overall ethnic discrimination measure, higher proportions of men than women reported ethnic discrimination, consistent with prior studies (Arellano-Morales et al., 2015; Brondolo et al., 2009; Pérez et al., 2008), yet no differences based on gender were observed with respect to discrimination in obtaining health care/insurance, receiving care, or in public. Conversely, while the prevalence of the overall measure of ethnic discrimination did not vary by educational attainment or immigrant generation, differences based on educational attainment and immigrant generation were observed for specific types/settings of discrimination, such as discrimination in healthcare, employment/education/housing, and being called a racist name. This pattern is reminiscent of findings from a study of Black and Hispanic adults in New York City, in which an overall racism measure did not vary by socioeconomic status, yet socioeconomic variation was observed when examining specific types of racist experiences (Brondolo et al., 2009). Prior nationally-representative studies have also documented variation in demographic patterns in discrimination based on the specific type or setting of ethnic discrimination assessed (Arellano-Morales et al., 2015; Findling et al., 2019; Salas-Wright et al., 2015).

Such results emphasize that global measures encompassing different settings and types of perceived ethnic discrimination may obscure subgroup variations in discrimination experiences among Hispanic adults. Different types/settings of discrimination may have differential impacts on physical and mental health. For example, ethnic discrimination in obtaining healthcare or health insurance may be harmful for physical health through decreasing likelihood of seeking and receiving needed preventive care or treatment (Lee et al., 2009). Prior research has also identified lower quality medical care received by Hispanic and other racial minorities, relative to Whites, highlighting the role of system-level variables, such as sub-optimal interpretation services or fragmented health care, in producing and reproducing differential treatment (Smedley et al., 2003). At the same time, the physical health effects of discrimination in healthcare, most prevalent among first-generation Hispanics, may differ from the mental health impacts of being called a racist name or receiving verbal/physical threats/assaults based on ethnicity, both of which are most prevalent among second-generation Hispanics. In a recent analysis of US Hispanic adults, being called a racist name and being “picked on/threatened” were the two types of ethnic discrimination most strongly-associated with mood disorders, anxiety disorders, or substance use disorders (Cobb et al., 2020).

Although different types and settings of ethnic discrimination may differentially impact mental and physical health, the different types and settings of ethnic discrimination may often overlap. In the present study, for example, an experience such as hearing an ethnic slur while walking down the street could be reported both as “discrimination in public” and “being called a racist name.” The study’s self-reported, perceived ethnic discrimination measure captures the way each individual perceives, remembers, and reports any given experience. Although subjective, this measure nonetheless captures participants’ realities as they perceive them, and any experience perceived as ethnic discrimination (whether the unfair treatment was due to ethnicity, another social identity, or any other reason) potentially impacts mental health (Gong et al., 2017). Some of the types of discrimination measured in the present study, such as “discrimination related to employment, education or training, housing, the courts, the police…” allude to institutional discrimination, yet none of the measures in the present study are conceptualized as capturing structural discrimination. Rather, these measures assess how Hispanics experience discrimination in different settings and how they recall, label, and report these experiences. For example, when a Hispanic employee is passed over for a promotion, the fundamental institutional processes which systematically limit Hispanic employees from such promotion opportunities may not be easily identified, and a supervisor’s bias against the employee is likely unspoken and implicit. Therefore, while the denial of promotion may reflect structural discrimination, it is experienced, perceived, and reported in the present study at the individual level.

Furthermore, the measure utilized in the present study specifically inquired about experiences “because you are Hispanic or Latino,” thereby limiting the discriminatory experiences reported to those which participants attributed to their socially-ascribed “Hispanic or Latino” identity. While the motives for some instances of discrimination can be easily identified and attributed to ethnicity (e.g., ethnic slurs), discrimination also includes unfair treatment for which the motive may be ambiguous, and for which attributions are ascribed by the victim of the mistreatment. Hispanics who experience discrimination may attribute it to any of various identifications besides the Hispanic ethnicity, such as social class, immigration status, language proficiency, skin color, or religion. If making sense of and explaining one’s social identity/position becomes a function of exposure to life in a racialized society (Omi & Winant, 1994), recently arrived immigrants may be less likely to attribute the source of mistreatment to “Hispanic” identification, as opposed to other attributes. Mexico’s 2010 National Survey on Discrimination, for example, reported that financial means was the primary self-reported reason for the violation of personal rights in Mexico, distantly followed by skin color (National Council to Prevent Discrimination, 2011). Moreover, individuals categorized as “Hispanic” vary in their level of identification with the pan-ethnic category of “Hispanic/Latino,” with Cuban Americans reportedly less-likely to self-identify as Hispanic/Latino than other Hispanic heritage groups (Okamoto & Mora, 2014). Lower pan-ethnic identification, therefore, may be partially related to the relatively lower prevalence of self-reported perceived ethnic discrimination in Cuban- and Dominican-heritage groups presented here, if fewer Cubans tend to self-identify as Hispanic/Latino and a substantial portion of Dominicans associate more with a racial than ethnic category.

Conclusion

The diversity of the Hispanic population is reflected in the diversity of experiences of ethnic discrimination that Hispanics perceive and self-report. In the study’s nationally-representative sample of civilian, non-institutionalized Hispanic adults, the relationship between Spanish use/proficiency and self-reported perceived ethnic discrimination varied by level of English use/proficiency. Moreover, the relationship between language use/proficiency and ethnic discrimination also varied by Hispanic heritage group. Patterns of variation in the prevalence of ethnic discrimination overall did not consistently mirror the patterns of variation observed for specific types or settings of ethnic discrimination; as such, relying on an overall measure of ethnic discrimination may overlook the relatively high prevalence of certain specific types of ethnic discrimination among Hispanics with intersecting identities. Insights from Latino Critical Theory (LatCrit) highlight the importance of “centering” the “internal diversities” with the Hispanic/Latino(a) community, as “neglect of these internal diversities serves only to neglect those most in need” (Valdes, 2005, p. 154). At the same time, LatCrit emphasizes “embrac[ing] commonalities,” as identification with the Hispanic or Latino(a) pan-ethnic category is instrumental for solidarity, collective action, and social change (Valdes, 2005, p. 158).

Study findings add to the body of literature about the perceptions and experiences of ethnic discrimination affecting diverse Hispanic subgroups across various settings. A large proportion of the general US population is unaware of racial gaps in health, overlooks the role of social forces in shaping health outcomes, or overestimates the progress attained toward economic equality between racial groups (Williams & Cooper, 2019). Highlighting discrimination and inequities is one part of a broader set of coordinated strategies to dismantle racism, which requires concerted actions to promote and optimize health, enhanced access to the opportunities that produce and reproduce resources, and increased public empathy and political will (Williams & Cooper, 2019). Policies that diminish or terminate social closures, where one group dominates valuable resources by closing off opportunities to other disadvantaged groups, can eliminate persistent inequalities (Tilly, 1998). For example, eliminating barriers to social and health services for non-English speakers and increasing voter registration and voting behaviors have the potential to expand opportunities and reduce the discrimination (Bruch et al., 2019) faced by Hispanics in the United States.