Disaggregating Race/Ethnicity Data Categories: Criticisms, Dangers, And Opposing Viewpoints

Conducting health equity research relies on complete, accurate information about race and ethnicity. However, data quality issues, including race/ethnicity misclassification and data incompleteness, remain challenges. Disaggregating the standard minimum US federal race (American Indian or Alaska Native, Asian, Black or African American, Native Hawaiian or Other Pacific Islander, White) and ethnicity (Hispanic or Latino) categories into smaller groups may reduce missing data on questionnaires, uncover health inequities, and allow appropriate allocation of resources to meet community needs.

For these reasons, the Robert Wood Johnson Foundation (RWJF) supports collecting and analyzing disaggregated race/ethnicity data, especially in a health context. Their National Commission to Transform Public Health Data Systems, as well as numerous policy analyses and study findings, recommend city-, county-, and state-level policies that improve collection of demographic information. This includes using race, ethnicity, and primary language to identify and disaggregate locally relevant race/ethnicity groups.

A recent success story for data disaggregation proponents is New York State Law S.6639-A/A.6896-A, signed by Governor Kathy Hochul in December 2021. This bill requires state agencies’ race/ethnicity questionnaires to include disaggregated response options for Asian Americans (AA) and Native Hawaiian or Other Pacific Islanders (NH/PI), allowing constituents to identify with a more specific group. Examples of such groups under the AA category include Korean, Tibetan, and Pakistani, while the NH/PI category includes Samoan and Marshallese. The signing coincides with heightened anti-Asian racism, and the disproportionate impact of COVID-19 on culturally and linguistically distinct AA and NH/PI communities.

This measure results from broad support and years of grassroots advocacy. Civil rights groups and AA and NH/PI leaders informed the text of S.6639-A/A.6896-A and effectively pushed its passage into law. Still, a small group of data disaggregation opponents have had their voices amplified in press coverage surrounding the bill, claiming the law will be used to divide Asians by reducing their overall size and subsequent political leverage.

These dissenting opinions are not new; they have hindered similar legislation proposed in other states and have even motivated attempts to preemptively ban disaggregated data collection. Below, we summarize opposing views that leaders and advocates should prepare to address to promote well-rounded, informed discourse as data disaggregation efforts move forward. We also recommend existing resources that aim to address pushback and guide sustainable, systematic data disaggregation.

Privacy And Surveillance Concerns

Widespread collection of disaggregated race/ethnicity data on government questionnaires may spark privacy concerns, especially for members of smaller populations. For example, stratifying health and vital statistics by disaggregated race/ethnicity and locality may inadvertently reveal respondents’ identities. It is important for reporting agencies and data managers to tighten privacy procedures to avoid re-identification. One option is to suppress public data with cell sizes for confidentiality reasons, while making these detailed data available upon request by researchers and organizations seeking to improve emergency response services and other efforts. Data suppression criteria can vary depending on confidentiality of individuals, statistical reliability, or data quality.

Data disaggregation requirements may also provoke opposition from historically surveilled groups, as the census has been used to target and harm specific racial and ethnic groups. As an example, the “Mexican” race option on the 1930 census aimed to monitor and segregate this growing immigrant group at the time. Over the years, civil rights legislation has worked toward establishing protections from such profiling but racial profiling remains a salient issue. Ongoing surveillance of Arab Americans, for instance, can lead to non-response on demographic questionnaires, especially questions with more granular race/ethnicity categories. Despite past misuse, the census is a critical tool for informing budgets, political redistricting, and resource allocation.

Consequently, disaggregation policies should include local processes to transparently explain the intended use of collected data. Policy makers must work with trusted community partners to communicate commitments to confidentiality. Community-based organizations may co-develop public education on how disaggregated data may work to combat systemic discrimination across sectors. Moreover, policies to train and educate data managers, analysts, and collectors are necessary to avoid repeating or evoking past harms for communities that now have higher mistrust of governmental surveillance.

Political Strength In Numbers

Proponents of aggregation argue it is a source of political strength and solidarity among smaller racial/ethnic groups. As the census evolved into an important mechanism for tracking national diversity, advocacy groups in the 1960s mobilized support from Spanish-speaking communities to advocate for a Hispanic ethnic category. Despite initial resistance to the perceived homogenization of all Spanish speakers, the Hispanic classification became an essential source of solidarity and political power. The fear of segregation from breaking down categories should be addressed by policy makers through clear and nuanced resource allocation and communication.

The research institute, PolicyLink, has discussed ideological underpinnings of the claim that disaggregation pits groups against each other, stating that this point is “more part of the broader ‘culture wars’ and general political divisions in the country than making a critique specific to public health or demographic research.” This was a salient issue in the pushback against a proposed bill, CA AB-1726, to disaggregate AA and NH/PI data in California. Some from the state’s Chinese American community suspected that formalized recognition of historically overlooked communities, such as Cambodian or Hmong Americans, would ultimately deprioritize those of Chinese origin for accessing certain resources and opportunities.

Researchers from the RAND Corporation urge a unified Asian American front in the wake of a three-fold increase in anti-Asian hate crimes as a way to seek support for disaggregation. Their suggested advocacy platform illustrates that within-group diversity does not obviate political coalition building, while also demonstrating the importance of disaggregation to resist homogenization and stereotyping. Furthermore, using data to measure diversity and combat stereotyping can highlight unique health needs that are concealed when data are aggregated.

Misinterpretation Of Results

Racial and ethnic categorizations lack a biological basis and are socially constructed. However, there are arguments that presenting health data by more detailed race/ethnicity may perpetuate theories of biological inferiority and discriminatory behavior. Stratifying data into more specific identities may lead some to pathologize ethnicities and cultures, rather than contextualize ethnic disparities as products of structural inequities. However, it is important to note that such misinterpretations are possible even when analyzing racial/ethnic differences under the current Office of Management and Budget (OMB) standard.  

Hawai’i State Department of Health officials reflected on this challenge during COVID-19 data collection efforts on AA and NH/PI constituents: “Efforts were designed to achieve a balance between highlighting the concerns of specific populations and inadvertently contributing to the stigmatization of groups who have been marginalized and who experience racism.” More research is needed to identify valid and reliable measures of structural racism that can be made widely available. Such metrics can help reorient the public perception of race and ethnicity as health risk factors, shifting the focus to racism as a root cause of disparities.

Evolving Perceptions Of Race And Ethnicity

The OMB, which sets national race/ethnicity reporting standards, cites a lack of consensus in terminology as justification for the delays behind additional racial/ethnic categories in the minimum federal reporting requirements. In 2018, OMB rejected the addition of a Middle Eastern and North African (MENA) Americans from the White race category, declaring that disagreement surrounding the MENA label would lead to reporting inconsistencies.

Significant undercounting and inconsistent response rates currently take place under the current OMB standards. Constrained race/ethnicity options do not match how individuals self-identify and are racialized, leading to high rates of missing data and increasing selection of the “Some Other Race” option. “Some Other Race” responses are often left out of analyses and reporting, even though they may be meaningfully patterned. Additionally, terms used to commonly describe certain races and ethnicities are constantly in flux. For instance, there was not common usage of the term “Hispanic” before the OMB formally included it in federal reporting requirements in 1977. Even today, preference between the terms “Hispanic” and “Latino/a/x/e” varies greatly among members of this diaspora.

The United Nations (UN) Statistical Commission’s guidebook on data disaggregation affirms: “Ethnicity is multidimensional and is more of a process than a static concept, and so ethnic classification should be treated with moveable boundaries.” While complete agreement may never be achieved, local leaders, including states, can establish procedures that allow for a regular updating of labels based on changes in race/ethnicity definitions, community preferences, and shifts in the population demographics.

Minimizing Links Between Racism And Inequity

Over-reliance on self-reported ethnicity data may diminish the role of perceived race, or socially assigned race, when studying the effects of racism on health. Perceptions of race based on physical appearance rather than language, culture, or other aspects of ethnoracial identity are crucial for understanding this association. To illustrate, a recent study of MENA individuals’ self-reported race/ethnicity found that few of those with African ancestry, including some Egyptians and Moroccans, “identify as Black regardless of labels offered, based on how their phenotypes are perceived in the United States.” However, most participants preferred the MENA ethnic category and were less likely to select an additional race when given the MENA option. Some researchers are concerned that establishing race/ethnicity response options that align more with self-identification than perceived race may inadvertently discourage reporting of “Black” or “White” race, labels that may feel reductive for those with intersecting identities. This may limit researchers’ ability to evaluate outcomes related to anti-Black racism in MENA and other multiracial populations.

To address racism against those from the African diaspora, a UN Human Rights Council report calls for countries that do not currently collect broad race data to implement the practice. The report also asserts the importance of disaggregating ethnic origin to “unpack and understand the differentiated dynamics of systemic racism.” Leaders must engage with communities to discuss racial/ethnic identity and how to apply these terms when completing patient intake forms and other questionnaires. Scripts and explainers during the data collection process can clarify the importance of reporting self-identity, as well as socially assigned race, for understanding racism’s impact on health. This also raises the importance of allowing multiple responses and in-depth analysis of multiracial individuals.

Resources

Many existing resources outline data disaggregation policy considerations and navigate inevitable challenges for state and local leaders. PolicyLink, the RWJF, and the Office of the UN High Commissioner for Human Rights have released reports on recommended practices for disaggregating data, including considerations for preventing inequitable and nontransparent implementation of these policies and practices. In using these tools, it is important to recognize that these policies and practices can be inequitably implemented and sustained.

These guides can be starting points for locally tailored implementation plans. Similarly, overviews of different collection, management, and analysis methods for disaggregating data are available from institutions such the Urban Institute, University of California–Los Angeles, and National Forum on Education Statistics. Such resources are vital for jurisdictions seeking to adapt individual data governance systems to accommodate new categories and perform meaningful data analyses.

Conclusion

Because existing OMB-mandated race and ethnicity categories span racially and culturally diverse ethnic groups, data aggregation inevitably obscures within-group inequities. As the standard categories are only minimum reporting requirements, the OMB recommends that localities identify priority groups to include in questionnaires. Despite this guidance, there has been limited political will to invest the time, resources, and effort needed to implement more progressive data collection practices, as well as opposition to attempts to codify data disaggregation practices.

State and local decision makers can no longer ignore empirical evidence accumulated over years of public health research that race/ethnicity disaggregation will have positive long-term and immediate benefits. Leaders have a responsibility to follow best practices for overcoming data disaggregation opposition. Ultimately, these efforts will strengthen their community ties, help identify more accurate risk factors of disease, and support targeted, culturally appropriate interventions in the pursuit of health equity.  

Authors’ Note

This manuscript is a part of the Innovations in Data Equity for All Laboratory (IDEAL) initiative led by the NYU Center for the Study of Asian American Health and Coalition for Asian American Children and Families with support by colleagues from the New York Academy of Medicine, NYU Langone Health, and the New York State Department of Health. Work was supported in part by the NYU Global Center for Implementation Science Pilot Awards, National Institutes of Health National Institute on Minority Health and Health Disparities Award No. U54MD000538, National Heart, Lung, Blood Institute Community Engagement Alliance Non-Federal 1OT2HL156812–01, Westat Sub-OTA No. 6793–02-S013 and R01HL141427, Department of Health and Human Services, Centers for Disease Control and Prevention Award Nos. NU38OT2020001477, CFDA No. 93.421 and 1NH23IP922639–01–00, CFDA No. 93.185 and New York State.