The Bumpy Journey of Collecting Race and Ethnicity Data of Latinos

By Arturo Vargas

It is 2024, and Latinos in the United States are approaching 20 percent of the nation’s population at some 63 million. The Census Bureau reports this number from those answering the decennial census questionnaire and indicating that they are of Hispanic origin — something that in the history of the nation only began happening fairly recently, about 44 years ago. And the journey to identifying Latinos in the census — and getting the numbers right — has been a bumpy one.

The first time the Census Bureau attempted to count all Latinos as Latinos in a decennial census was in 1970, and it was a last-minute decision. Only after extensive advocacy by the courageous Latino leaders of the era, including National Council of La Raza President Raul Yzaguirre and MALDEF President and General Counsel Vilma Martinez, did the Census Bureau agree to add a Hispanic origin question — and only to the “long form.” The long form, replaced in 2005 by the American Community Survey, was an extensive questionnaire that was sent randomly to one of every six households asking about a variety of housing and demographic characteristics.

The 11th-hour addition of a question that was not thoroughly tested yielded unsurprisingly poor results. The Hispanic origin question was then added to 100 percent of the questionnaires in 1980, a major milestone and civil rights achievement. Yet it was fundamentally flawed.

Ever since the addition of a Hispanic origin question to complement a race question, the proposition to Latino respondents has been vexing. To be clear, Latinos are not confused about their race or ethnicity, but the way the Census Bureau was asking Latinos to self-identify was itself problematic.

When asked to identify a race, many — indeed most — Latinos found the question confusing because the options included only the racial categories approved by the Office of Management and Budget (OMB), which sets standards for federal data on race and ethnicity. These categories in the standards only included White, Black, Asian, American Indian or Alaska Native, and Pacific Islander. Many Latinos did not find themselves in these racial categories, and they either left the question blank or indicated they were Some Other Race — a category that as far as OMB was concerned did not exist. This has placed the Census Bureau in the unwanted and unenviable position of assigning an OMB-recognized racial category to those who left the race question blank or indicated they were of Some Other Race.

Additionally, when asked to identify if they were of Hispanic or Latino origin, Latinos eagerly answered and self-identified. Yet because of the OMB guidelines, Latinos could only indicate their background as being of one single Latino national origin. You have a Mexican father and a Salvadoran mother? According to the Census Bureau at the time, you could be only one. So pick who you love more — mom or dad — because the bureau could not tabulate mixed Latino origins.

This dilemma persisted through the 2000 and 2010 Censuses, even with some slight modifications to the Hispanic origin question over time. Outreach campaigns geared toward Latinos to answer the census were often undermined by the complicated framework of distinguishing between what is your race and are you Latino. Ultimately, many Latinos indicated that they identified racially as Latino, even though Latino is not a race.

We also learned in the 2000 Census that the way the question is framed has dramatic impacts. In that decennial, the Hispanic origin question did not include “Dominican” among the examples in the question of who is Latino. As a consequence, thousands of Dominicans marked that they were “other Latino” — inflating that category and dramatically reducing the number of self-identified Dominican Latinos, even though their numbers had significantly increased over the past decade.

Throughout the decennial censuses, asking about Hispanic origin has been a Census Bureau Achilles heel.

Leading up to the 2010 Census, the bureau engaged in a comprehensive research project to determine how to better capture race and Hispanic origin data. The Census Bureau’s most extensive experiment ever was embedded in the 2010 Census, where multiple approaches to asking about race and Hispanic origin were asked — including a proposal to ask about Hispanic origin and race in one combined question.

NALEO Educational Fund had a very healthy skepticism about the new approaches to asking about race and Hispanic origin. The Census Bureau’s proposal for a combined Hispanic origin and race question was a huge issue. Securing a Hispanic origin question on the decennial census was one of the most impactful civil rights achievements of the time. To consider eliminating that Hispanic origin question was unimaginable. We would have to be thoroughly convinced that a different approach was better.

In 2014, NALEO Educational Fund convened Latino stakeholders and Census Bureau researchers to delve into these issues. The Latino stakeholders — which included advocates, demographers, researchers, and civil rights litigators — provided advice and recommendations to the bureau. The bureau’s 2010 research, which was followed up by extensive focus groups and a second national test (the National Content Test in 2015), demonstrated that asking Latinos about their identity in a combined race and Hispanic origin question yielded more complete results.  

The new approach allowed Latinos to more easily identify with as many backgrounds as they chose — or just one. Latinos could identify ethnically as Hispanic and choose one or more racial categories as well, such as White, Black, Asian, or American Indian. Also important to understanding the rich cultural diversity of Latinos, the combined question would allow Latinos to choose one or more Latino national origin groups, such as Mexican and Salvadoran. But to adopt a combined question, the OMB would have to revise its standards on the collection of race and ethnicity data. 

The OMB was presumably on track to update the standards and adopt the new combined question format for the 2020 Census. This effort, however, was halted by the previous administration in 2017 and was only made public when the Census Bureau published the questionnaire that would be used in the final field test before 2020 — the 2018 “End to End” test (what should have been the 2018 “Dress Rehearsal” but was downscaled due to lack of funding). In early 2018, the Census Bureau released the test questionnaire that once again included the separate race and Hispanic origin questions, although with some modifications.

The results of this two-question approach were exactly what the Census Bureau had predicted: The non-existent racial category of “Some Other Race” catapulted to being the second largest racial group in the nation, outnumbering those who identified as Black or African American.  The Some Other Race category number was driven by Latinos, as more than 90 percent of those who picked this racial category also indicated they were Latino. The bureau had to impute a racial category for millions of people in America once again.

When a new administration took office in 2021, one of the top priorities for NALEO Educational Fund and a growing coalition of census advocates was for OMB to complete the unfinished business of updating the race and ethnicity standards. Fortunately, that process began in Spring 2023 and was completed by March 2024. This included extensive consultations with the public by OMB, including holding town halls and receiving thousands of public comment submissions.

NALEO Educational Fund organized participation by Latino stakeholders in the public comment process as well as engaged with leadership of the Afro-Latino community, many of whom opposed a combined race and Hispanic origin question. The organization held a number of listening sessions and discussions to understand the complexity of the issues and concerns, and these sessions informed the comments we submitted to the OMB about the proposed revisions to the standards.   

With the OMB decision to move forward with a combined question, we are committed to ensuring the Census Bureau implements the new question in a way that fully promotes Latinos’ understanding of the question and the opportunity to indicate as many backgrounds, races, ethnicities, and national origins as fully captures that person’s racial and ethnic identity. This should include additional research on the optimal wording of the question itself and a comprehensive public education campaign.

As 2030 approaches, the Census Bureau may finally have the opportunity to significantly improve the collection of race and ethnicity data of the Latino population. This must be a top priority for the agency. As Latinos become a larger share of the U.S. population, a successful census — and with it a greater understanding of the diversity of the United States — will be impossible without a fair and accurate count and data collection on Latinos.

Arturo Vargas is the chief executive officer of NALEO Educational Fund.