The U.S. Census Bureau has persistently undercounted Hispanics. Now, the same bureau announced that they may change the way they ask about Hispanic ethnicity in 2020 -- a move that has critics worried that the Latino population could be undercounted even further.
Currently, the Census measures Hispanic ethnicity and race in different questions, but the proposed change would collapse the two categories together into a "race and origin" question.
Although the switch would allow Latinos who identify for example, as black or white, to check both ethnicity and race boxes, Carlos Cortés, a professor emeritus at the University of California Riverside worries that those who take the survey will feel they need to choose one or the other if the topic is presented in one question, resulting in an undercounted Latino population.
"There were 50 million [Hispanics] in the 2010 Census, all it takes is 1 in 10 saying 'Well, I'm white or I'm black so I won't check Hispanic' and suddenly there are only 45 million of us," said Cortés who has written about race and multiculturalism for decades.
Why is the Census considering the change at all ? In 2010, when the bureau offered the race and Hispanic ethnicity question separately, a large portion of Latinos marked their race as "some other race," which in turn rendered "Some other race" the third largest racial category in the U.S., to the frustration of Census takers.
With the aim of "improving the accuracy and the reliability of census results by expanding our understanding of how people identify their race and Hispanic origin," the Census bureau has launched series of massive trials of new polling techniques, according to Census Bureau Director Robert Groves.
The study measured nearly 500,000 households and held 67 focus groups with nearly 800 people. The bureau reportedly found that the population who identified as Hispanic was "not significantly different" with the new question from how people responded to their 2010 Census questionnaires, which combined race and origin. This indicated that the "total proportion of Hispanics was not reduced in a combined question approach," the bureau concluded.
But Cortés is just one of a handful of scholars in the field who remain skeptical.
"When I look in the mirror, I'm clearly white, but I'm also of Hispanic heritage," he said. "There's no conflict when they're separate questions, but when you put them together those who don't look upon Hispanic as an ethnicity will be more likely to check one or the other and not both."
So why measure racial and ethnic categories at all? Well, according to Kenneth Prewitt, the director of Census Bureau during the 2000 Census, the system is needed so that policy makers can understand where disparities lie. Knowing this informs the decisions that they make on issues like healthcare and education.
But Prewitt argued in a 2010 Op-Ed for The Washington Post that the race question is "broken" and that "we can do better." He says that Americans are much more complex than the categorization systems we have developed to measure them, beginning with the 1790 Census which only allowed for three race categories: "White, Black, Red."
Instead Prewitt suggests that a more accurate measurement in the next census would result from asking the following two questions: "What national origin, ethnicity, tribe, language group or ancestry do you consider yourself to be? (List all those important to you)" and "Where were you born, and where were your parents born?"
Prewitt argues that such open-ended questions would allow survey-takers to define themselves as they deem necessary. Their responses would also provide enough information for statisticians to breakdown the answers in the same categories we've been using for hundreds of years, if necessary.
"A statistical portrait of how different groups are faring remains necessary both to erase the inequities of historical racism and to prevent discrimination as the recently arrived strive to participate fully in their new country," Prewitt wrote in the Op-Ed.
"But only if we draw the portrait more carefully than that produced by the 2010 Census."