Latino? Latinx? Latine? A Call for Inclusive Categories in Epidemiologic Research

Abstract

Use of the word “Latinx” has risen in popularity among both academics and nonacademics to promote a gender-inclusive alternative to otherwise linguistically gendered terms of “Latino/a.” While critics claim the term is inappropriate for populations without gender-diverse individuals, or those of unknown demographic composition, increasing usage and among younger communities signals an important shift in centering the intersectional experiences of transgender and gender-diverse people. Amid these shifts, what are the implications for epidemiologic methods? We provide some brief historical context for the origin of the word “Latinx” along with its alternative “Latine” and discuss the potential consequences of its use for participant recruitment and study validity. Additionally, we provide suggestions for the best use of “Latino” compared with “Latinx/e” in several contextual circumstances. We recommend using “Latinx” or “Latine” in large populations, even without detailed data on gender, since there is likely gender diversity in the population, albeit unmeasured. In participant-facing recruitment or study documents, additional context is needed to determine which identifier is most appropriate.

Editor’s note:  The opinions expressed in this article are those of the authors and do not necessarily reflect the views of the American Journal of Epidemiology.

“Latinx” was introduced into mainstream usage in the early 2000s, with its modern origin in online forums by gender-diverse Latinx people (1). The vernacular use of “Latinx,” however, likely dates back to the 1990s during feminist protests across Latin America, where the vowels in protest signs were crossed out as a statement about the rejection of patriarchal society (2). As the usage of Latinx rose in popularity in the United States, some critics argued that the shift reflected neocolonial imperative upon the Spanish language—since the “x” is seldom used in the language—and was difficult for monolingual Spanish speakers to use fluently. Today, trans Latine activists have elevated the use of “Latine” as an equal alternative to “Latinx” that promotes fluency in Spanish and integration into other terms like pronouns (i.e., “elle”); however, it is also important to recognize that indigenous language speakers have been using “x” for centuries, and both are acceptable for different audiences. This commentary uses Latine to reflect the current activist discourse. While the vicissitudes of the term have been extensively described in the social sciences, greater efforts are needed by epidemiologists to understand and incorporate this term and its alternatives.

To strive for health equity in epidemiologic research, it is imperative that our categorizations of race and ethnicity are well-defined so that subgroup disparities can be exposed and generalizable information can be accurately attributed to a defined population. As research continues to demonstrate the burden of disease among historically marginalized communities whose intersections with power are further associated with disparate health outcomes, tensions around whether to use “Latino/a” or “Latinx” underscores the critical importance of querying who is in the category. For example, as of 2022, the US Census Bureau estimates that Hispanics or Latinos make up 19% of the nation’s population (3). Intracommunity debate about the use of “Latinx” versus “Hispanic” or “Latino” is divided primarily across generational lines, with older generations less likely to identify with “Latinx” than younger generations (4). In the US Census, all people of “Cuban, Mexican, Puerto Rican, South or Central American, or other Spanish culture or origin” are classified under a pan-ethnic category titled “Hispanic or Latino” (5). This is the phrasing most often seen in the epidemiologic literature, reflective of the language used in population data sets. Using any pan-ethnic identifier, irrespective of which form, requires critical consideration, especially in health disparities research. “Hispanic and Latino” encompasses a wide diversity of experiences, cultures, and languages; each country has unique experiences with colonization that continue to have impacts on people’s health. Most individuals who fall under the “Hispanic or Latino” label view their identities in a more granular context, identifying more with a national or regional label (e.g., Puerto Rican, Boricua, Mexican, Chicano/x) than the pan-ethnic “Hispanic or Latino.”

Although not always a feasible unit of analysis due to limits on sample size, these identifiers are the most precise way to measure ethnicity across Latin America and can illuminate intracommunity disparities that would be overlooked in the aggregate “Hispanic or Latino” population. In aggregate, people of Latin American descent in the United States have a lower smoking rate than the national average. However, there is significant heterogeneity across subgroups; for example, Puerto Ricans have much higher smoking rates than both the national average and other subgroups of people of Latin American descent (6). In studies where subgroups create underpowered analyses, however, a pan-ethnic term like “Latino” or “Latinx/e” may be more efficient and allow for some detail retention.

Further, people who are both gender-diverse and of Latin American descent have been historically excluded from both research of gender-diverse people and research focusing on disparities among people of Latin American descent. This gap in intersectional analysis means that “Latinx/Latine,” “Latino,” and “Hispanic” are being used interchangeably and without regard for the cultural complexities and methodological consequences behind these words. Prior commentaries have elaborated on the discourse surrounding the use of “Latinx” to describe populations that may not contain gender-diverse individuals. Dr. Del Río-González’s 2021 commentary (7) suggests the use of “Latinx” solely when describing populations of gender-diverse individuals; she goes on to say that, in groups of unknown gender composition, the triad of “Latina, Latinx, and Latino” should be used as an alternative. While we agree that Latinx is an intentional acknowledgment of gender diversity, we differ in the belief that researchers will always know whether there are gender-diverse people in the population—especially given the often limited collection of gender identity data. Additionally, while we acknowledge that “Latina, Latinx, and Latino” may seem to be the most precise option, it still deviates from the Spanish linguistic standards that opponents of “Latinx” so frequently cite and does not add anything to the understanding of the study population.

“Latino”—the original all-gender descriptor in Spanish, which is used to describe mixed-gender study populations—linguistically encompasses both “Latino”- and “Latina”-identified individuals in a single term. “Latinx” functions much in the same way, while extending beyond the gender binary, and encompassing those identities within a single term. These terms are inherently not meant to be precise, nor should they be regarded as such; any pan-ethnic identifier deviates from the most precise and authentic self-identity labels (i.e., national identity labels, regional identity labels). As with any pan-ethnic identifier, “Latinx” encompasses a wide variety of identities and cultures that are impossible to fully capture under a single label. Our responsibility as researchers is to critically reflect on the rationale for the choice and to choose the most appropriate and inclusive option that both reduces harm and uplifts community members whose identities have historically been deprioritized.

To improve the quality of epidemiologic research and better serve minoritized populations, researchers should first consider their study design and intended audience when using these terms. In mixed-methods studies where participants have the opportunity of open-ended self-identification, the best practice is to use the terms that people use for themselves. Additional linguistic nuances can be addressed in the resulting papers, thereby providing readers with a descriptive understanding of the study population. In most epidemiologic studies—such as cohort studies—with larger sample sizes, “Latinx” or “Latine” can be used as a mixed-gender descriptor, waiving the assumption that there are zero gender-diverse people in the population. The trans and gender-diverse population is growing (8), and the visibility of gender-diverse populations in the United States and across Latin America is at a historic high, even when suppressed by colonial actors (9, 10). It is too strong an assumption, in most cases, that the study inclusion criteria or demographic data collection is sufficient to deduce a complete lack of gender diversity in the population.

When collecting primary data, prior knowledge of the target study demographics will influence language use in participant-facing documents such as recruitment materials and participant surveys. Recent studies reveal a generational, educational, and linguistic divide in the knowledge and use of these terms (11). In an older cohort, providing multiple vowel endings (i.e., “Latino/a/e”) may increase word recognition compared with the sole use of “Latine” or “Latinx.” In a monolingual Spanish-speaking cohort, “Latine” may be the more accessible option compared with “Latinx” due to its fluency in the Spanish language. While open-ended self-identification of ethnicity would provide the greatest detail, this is often unfeasible and creates subsequent complications with statistical power.

Language in participant-facing documents is especially important when recruiting gender-diverse participants within the Latine community. An understanding of intersectionality is critical to engaging with the lived experiences of this population and to promote effective study recruitment. Intersectionality is best summarized by this quote from Drs. Patricia Hill Collins and Sirma Bilge’s book, Intersectionality: “The events and conditions of social and political life and the self can seldom be understood as shaped by one factor. They are generally shaped by many factors in diverse and mutually influencing ways. When it comes to social inequality, people’s lives and the organization of power in a given society are better understood as being shaped not by a single axis of social division, be it race or gender or class, but by many axes that work together and influence each other.” (12, p. 2).

To capture existing disparities and explore the possibilities of unrecorded disparities, it is necessary to recruit individuals from this identity intersection (i.e., Latine gender-diverse people). Using gendered language in participant-facing documents systematically excludes trans and gender-diverse individuals from the study; this language may make studies unapproachable or make the participant feel excluded from the intended demographic. This thereby perpetuates the continued underrepresentation of this population in health research. Gendered language may also cause distress among participants, creating a lack of trust that resulting data will be handled effectively, with regard to the participants’ identities, and for the purpose of health promotion (13). With the growing hostile climate towards trans and gender-diverse communities, researchers must be intentional about promoting health equity through language. This may mean questioning the ethics around commonly accepted “best practices,” such as the necessity of congruence between manuscript language and participant-facing documents.

Using validated measures is not a good enough reason to forego critical thinking about our language. As an example from our own research in the Nurses’ Health Studies, participants are asked to classify their sexual orientation identity into one of several categories: completely heterosexual, mostly heterosexual, bisexual, mostly homosexual, completely homosexual, or not sure. These categories may look unfamiliar to readers as we seldom use them in published research in this exact linguistic manner. These measures capture a more expansive view of sexuality than traditional identity-based questions and have been longitudinally employed for decades, but they also have the potential to stigmatize queer community members, depending on the context in which they are subsequently discussed. While these items are collected in this format for consistency, recategorization (e.g., grouping people from “mostly homosexual” and “completely homosexual” into a larger “gay/lesbian” category) has become standard practice in published research. While there are potential methodological challenges when adapting language for publication—especially when transitioning from “Hispanic or Latino” to “Latinx/Latine” due to differences in scope—it is important to consider who will benefit from this shift and who will be harmed by inaction.

A grounding goal in epidemiologic research should be health promotion within a harm-reduction framework. Using gender-inclusive language is a method to promote trust and consequently increase study recruitment (14) while minimizing harm, especially in communities that face pervasive structural violence across intersecting identities. Additional action items we should implement to further this goal include:

• Opt for precision when possible. Consider whether a pan-ethnic identifier is necessary given the size of the data, and what limitations this descriptor may present. For exposures/outcomes with documented disparities by Latine subgroup, avoid the use of “Hispanic/Latino/Latine” and opt for specific ethnicity labels (e.g., Puerto Rican, Cuban, Salvadoran).

• When the use of a pan-ethnic identifier is deemed necessary to the study design or beneficial in the analysis, default to gender-inclusive terms like “Latinx” or “Latine.” These terms would be appropriate for describing large populations with people of varying genders.

• Consult colleagues at the intersection of these identities (15). Intracommunity knowledge is invaluable in research teams and can lead to more culturally competent studies and publications. Mid- and late-career researchers can offer mentorship to early-career members on their team and empower them to provide input on study design, participant-facing documents, and resulting publications.

• To increase the wealth of community knowledge in the field and support epidemiologists in researching their own communities, advocate for institutional policies that support researchers with multiple marginalized identities. For example, create incentive programs for both faculty and trainees to apply for National Institutes of Health diversity supplement awards to support multiply marginalized individuals.

We cannot lose sight of the fact that the gender-diverse Latine people who created and adapted this language are Latine culture. Gender-diverse Latines have always been at the forefront of advocacy and innovation in language and science, despite a lack of external acknowledgement. We, as researchers, must recognize the power of categories as wielded in epidemiologic research and minimize the linguistic and scientific harm of excluding and misrepresenting heterogeneous identities. Epidemiologists must strive to capture the complexity and diversity of our communities; without this approach, we cannot elucidate nor eliminate entrenched health disparities.