A New Question of Race: How the US Census is (and is not) Revising Its Collection of Racial Data

A sheet from the 1910 U.S. Census for Manhattan is displayed above. Since 1790, the United States has conducted its national census every 10 years. Photo courtesy of the U.S. Census Bureau.

The U.S. Office of Management and Budget (OMB) is currently proposing changes to the U.S. Census that allow for the overdue counting of Latino and Middle Eastern and North African (MENA) populations for which racial data has been historically overlooked. Through these updates, the U.S. Census would become more inclusive and reflective of the present U.S. racial climate, and produce more accurate data that will aid in creating programs and legislating to support these populations. The changes could be applied and added to all federal forms as early as next summer and be used for the 2030 census.

Currently, the Census is required to employ six categories within the race question based on 1997 OMB and congressional standards: White, Black or African American, American Indian or Alaska Native, Asian, Native Hawaiian or Other Pacific Islander, and Some Other Race. In addition, the U.S. Census asks a separate question regarding ethnicity, asking whether a respondent is Hispanic or Latino. However, the OMB is proposing several changes to the required categories and the formatting of the racial questions. 

The U.S. Census doesn’t clearly define their definition of race, instead specifying that their racial categories reflect a “social definition of race” that involves national origin and sociocultural groups and is not biological, anthropological, or genetic. However, in a general sense, race is often defined as the social categorization of people on shared physical traits according to the American Psychological Association. Conversely, ethnicity is reserved for groups with common identity-based characteristics such as ancestry, language, or culture. For the U.S. Census ethnicity is only defined through the binary of being “Hispanic” or “non-Hispanic.” However, scholars heavily disagree on the definitions of race and ethnicity. It is often difficult to draw the lines of distinction between such terms, as well as others including ancestry, nationality, and heritage, and often results in arbitrary applications.

The first proposal includes combining the race and ethnicity questions. The two questions would become a singular race question, with a Hispanic or Latino category added to the existing racial categories. The ethnicity question of the Census originated in the 1970 long-form Census to estimate the size of the Hispanic population in the U.S., the second attempt at measuring the Hispanic population in the United States after the 1930 Census incorporated a “Mexican” category that was heavily protested by Mexican-American populations at the time. The question was placed on all forms of the census in the 1980 census and in the 2000 census, was moved before the race question to increase the response rate. The reasoning behind the separation of ethnicity and race questions was due to the belief that Hispanic people could be of any race. 

However, a recent Pew Research Center survey of multiracial Americans found many Hispanic people do not separate their Hispanic background from their racial background, which echoes past findings. The same survey found that only 44% of Hispanic populations believed the census questions reflected how they see their own race and identity. The lack of a clear distinction between race and ethnicity has contributed to a lack of consistency in responses for many Hispanic respondents with the census. For example, in the 2015 National Contest Test, a survey conducted by the OMB to test and gather data on different race and ethnicity question designs, 43.5% of self-reported Hispanics reported not belonging to any federally recognized race group. The lack of a clear category for Latino and Hispanic populations has caused the “Some Other Race” racial category to grow into the second-largest category with more than 50 million people, or 1/7 of the U.S. population, identifying within the category. With such a large “Some Other Race” category, certain racial groups, like Hispanic and Latino populations, are lost within the large catch-all. Within the 2020 census, the undercounting of Hispanic populations occurred at a rate of 4.99%. A more modest undercount of 3% of Latinos was estimated to cause a $930 loss in funding for Medicaid for states, which disproportionately affected Latinos as they account for one-third of all Medicaid participants. Just within Los Angeles County, a 2% undercount could mean a $20.5 million funding reduction in the Supplemental Nutrition Assistance Program, a program that provides food-purchasing assistance to needy families. These examples are just two of the many programs and legislation that are based on US Census data, reflecting how necessary accurate population data can be. In tests conducted by the U.S. Census, the “Some Other Race” category decreased from 10.2% to 1.0% when the race and ethnicity questions were combined. The OMB concluded having a singular question helped the bureau overcome challenges in measuring Hispanic and Latino populations and creating clearer demographic data for the populations. 

The second proposal includes adding a MENA category to the current race categories. Currently, MENA populations, including but not limited to Lebanese, Iranian, Egyptian, Syrian, Moroccan, and Israeli people, fall under the category of “white.” However, many people within MENA don’t identify as white. Through 67 focus groups with a total of 768 participants, the Census Bureau found that MENA populations felt the racial category was “inaccurate” or “wrong.” Through several online surveys involving non-MENA white and MENA populations, both populations classified MENA-related traits, such as ancestry, names, and religion, as non-white racial markers. In addition, MENA names are distinctly seen as MENA, not white. Given that names are a primary basis for differentiation and discrimination, especially for labor and housing markets, a lack of a MENA category results in such discrimination being overlooked by the government. In the same study, it was found that when members of MENA are not offered a MENA label, 80% choose to identify as white, but when a MENA category is offered, only 10% choose to identify as white. Due to the difficulty MENA populations experienced when responding to the existing OMB race categories, the large population identifying as “Some Other Race” is also attributed to the MENA population, causing significant undercounting and similar issues as for Hispanic and Latino populations. For example, the Census Bureau estimates that 1.9 million Arab Americans are living in the U.S., while the Arab American Institute estimates it to be 3.7 million. Incorporating a MENA category with the question of race for the U.S. Census can allow for a better understanding of the MENA population within the U.S. and their needs as a racial group. 

Despite the clear benefits of increased representation within demographic data, there remain costs that marginalized populations have to bear. These costs are associated with the increased visibility of being identified and counted within racial data. In the early years of the U.S. Census, “race records were largely used to control and track different populations; categories were formed in relationship to whiteness, and white officials determined the race boxes,” creating a centrality of whiteness within the survey. The centrality pervaded the treatment and collection of data in minority groups through events such as the counting of enslaved workers as three-fifths of people and the conditional counting of indigenous populations only if they renounced tribal affiliation until 1890. Following the bombing of Pearl Harbor, the U.S. Census also shared racial demographic data with the military to aid in the internment of Japanese Americans. For this reason, Mexican Americans heavily fought the creation of the category of “Mexican” to remain under the category of “White” and retain the ability to be a citizen (as you had to be white to be a citizen at the time) and better assimilate into American society. For MENA populations, this fear remains acutely present. MENA Americans have experienced heightened discrimination in the post-9/11 “war on terror” environment and Trump-era immigration policies, causing them to be especially sensitive to the collection of their racial data since it could be used similarly. 

Many countries, due to the fear of racialization and abuses of data, steer away from explicit racial categories, as the U.S. is one of the only countries still using racial categories. Instead, they use alternate categories that implicitly track race, such as language spoken at home, citizenship, country of birth, and more. For example, France uses a color-blind approach to their legislature under the guise of universalism, having banned the collection of racial data in 1978. 

Recently, however, the lack of racial data collection has left French citizens without numbers for the rates of discriminatory events for people of color, such as police stops, workplace and housing discrimination, and COVID-19 deaths. The case of France demonstrates how, without concrete racial data, it is difficult for populations to track and create change in their community, making combating discrimination and disparities difficult. Discrimination is going to occur whether or not racial data is collected. For the populations not accurately counted within the census, that discrimination is effectively erased as it cannot be quantified. In the current system, MENA populations are rendered invisible. Although MENA immigrants tend to have lower incomes, resources and programs meant to target health disparities are not allocated to them given their ‘white’ status. In addition, since they are not legally classified as a minority, cases of discrimination and violations of civil rights become more difficult to prove. Therefore, due to the high cost that MENA citizens are currently paying due to their census exclusion, they have been found to still largely want to get out of the category of “white” despite the increased security risk. Though the demographical data is potentially running the risk of increasing discrimination, it is unquestionably hiding discrimination faced every day by millions of Americans.

The collection of data that reflects and encompasses the complex racial makeup of the country is a necessary aspect of the development and correction of equity issues within society. Since the Civil Rights movement, the U.S. census has tried to move away from racial demographics as a tool of control and more towards such ideals of advocacy and equity for racial groups. In the early 1970s, the U.S. Commission on Civil Rights stated that “racial and ethnic classification can be justified only if the data produced have a legitimate use in combating discrimination, planning programs, or conducting program evaluation.” Today, the census racial data is used to fuel local, state, tribal, and federal programs and is a critical foundation for U.S. research and policies. The race data ensures equal opportunities for racial groups through monitoring compliance with the Voting Rights Act, allocates funds for school districts for bilingual services under the Bilingual Education Act, and identifies segments of the population that may not be receiving needed medical services under the Public Health Service Act. In addition, the data is essential for researchers, advocacy groups, and policymakers in areas such as education, employment, engineering, housing, healthcare, and more. Programs for specific populations, such as the Indian Housing Block Grant Program, depend on the racial data gathered by the census to better understand and allocate resources within the programs. 

Therefore, gathering specific and accurate data about Latino, Hispanic, and MENA populations can ensure that the disparities and discrimination these populations endure will be identified and researched. The data will allow for increases in funding towards necessary programs meant to support vulnerable populations and work in conjunction with discrimination-focused legislation already in place. It will allow policymakers to create policies and programs to help solve these issues and ensure they are listening and adhering to the voices of these populations advocating for their interpretation of their race. Even if policymakers do not act on the data, the numbers create a foundation for communities and organizations to advocate for and better understand marginalized populations, in addition to providing them with a tool to pressure representatives with. However, in addition to enacting these reforms, the census must also continue to update its collection of races to match the ever-changing racial landscape of the US. 

For groups that fall in between existing racial categories, such as populations from Afghanistan, Brazil, Turkey, and more, the ability to self-identify through inputting specific racial groups within the existing categories is key. By writing in individual racial identities within larger categories, such as “Asian” or “Some Other Race,” such populations would still have the opportunity to be counted despite their lack of a dedicated category. There is also a need to update our categorization of Indigenous populations, as the current system undercounts Indigenous communities, limiting the effect of the programs and policies meant to support them. As the U.S. Census continues to play a role in policy-making and research, it must persist in modernizing its collection of racial data to ensure it represents, supports, and advocates for all races.

Ines Legrand (BC ’27) is a staff writer for CPR. She is studying political science and economics. Outside of CPR, she enjoys drinking tea, overspending at bookstores, and watching movies. She can be reached at ial2128@barnard.edu.