Changes in the 2000 census that added a racial category and allowed multirace responses permitted a new flexibility in self-identification, but also diluted the data collection that underpins social justice efforts, according to a former U.S. Census director who spoke at Harvard Monday (Feb. 2).
While the changes enacted for the 2000 census may seem innocuous, they constituted the biggest change between successive censuses in 200 years, according to Kenneth Prewitt, Carnegie Professor of Public Affairs at Columbia University and director of the U.S. Census Bureau during the 2000 census.
In addition, Prewitt said, the changes – particularly the multirace option – create a multiplying factor in race categorization that changes the possible number of racial categories from four in the 1990 census, with the additional possibility of being from a Hispanic ethnic group, to more than 100, when all possible racial and ethnic combinations are considered.
“Racial classification in the U.S. is now in a phase of unprecedented instability and uncertainty,” Prewitt said. “So, policies dependent on [that classification] are also in a period of instability and uncertainty.”
Prewitt spoke at William James Hall Monday afternoon as part of a lecture series on population studies sponsored by the Provost’s Office, the Office of the Dean of Public Health, and the Center for Population and Development Studies.
Required by the U.S. Constitution, the census is conducted every 10 years and is essential in determining how many U.S. representatives each state has. Consequently, the census is as much a political exercise – if not more – as it is a scientific exercise, Prewitt said.
In addition to merely counting heads, the U.S. census gathers all sorts of information on the nation’s citizens. The original intent of this information gathering, Prewitt said, was to inform the nation’s leaders about the characteristics of the population and thereby to inform their policy-making efforts to address the population’s needs.
Race classifications have been part of the census since the beginning and were particularly important to Southern states, where a slave was counted as six-tenths of a person and included in the determination of how many representatives Southern states had in Washington, D.C.
Prewitt traced the evolution of race categories on the census, saying that they have shifted over time but have been consistent in that they existed as discrete categories with everyone fitting into one or another.
Over the past 50 years, the census’ racial data has been used more and more by those interested in social justice. Census statistics have been used to highlight the plight of those downtrodden and forgotten. The statistics have been important in determining societal trends about specific minority groups to determine whether they’re thriving or showing the adverse signs of discrimination and racism.
In recent years, Prewitt said, new questions about the classification system have been asked, including whether there should be one at all. In 1990, he said the census had four racial categories: white, African-American, Native American, and Asian, with Hispanic denoted an ethnic group, not a racial group. Before the 2000 census, almost without debate, he said, Congress voted to add another racial group: Hawaiian/Pacific Islander.
“No one saw a reason to object,” Prewitt said. “The indication to me is that racial classification is now a moving target when pressure is brought to bear – and not all that much pressure was brought to bear [in this case].”
The second change allowed people to check more than one racial group, and about 2.4 percent of the population, 6.8 million people, did so, he said.
These changes turn out to be hot political issues, Prewitt said, with groups who advocate for racial justice opposing any dilution of the classification system. Still other groups, including those who would prefer no racial data be gathered at all, support adding more variability to the system, arguing it is an expression of social identity and that people shouldn’t be forced to say they’re something they’re not just to fit a rigid government category.
With the changes so far, Prewitt predicted even more variability will be inserted into the census. Regional groups such as Arab-Americans, or national groups within broader racial categories, such as Korean-Americans, could make an argument that they shouldn’t be lumped in with the others.
“My sense is that over time, it’ll be difficult to resist expanding more,” Prewitt said.
The initial result of the proliferation of categories may be confusion. In quality assurance interviews after the census was conducted, he said, only 40 percent of those who said they were multiracial in the census also said it in the follow-up surveys.
In addition, he said, just 16 percent of those who identified themselves as white and Native American in the census said the same thing three months later. And 9 percent who said they were of mixed descent from two races gave the same answer in follow-up interviews, but said they were of two different races.
Given the importance of fostering diversity in America today, some may see the proliferation of ways to identify oneself as a positive development. But Prewitt asked whether it is dangerous to make decisions on affirmative action and other social justice programs based on information from such a survey.
“The classification system, if pushed far enough, will implode,” Prewitt said. “If classification becomes useless for [creating] race-based policies, that is the price of individuality.”
For the future, Prewitt said he didn’t think the current system would remain. The 2020 census would almost certainly be very different. While self-identification classes may proliferate, some other sort of classification or grouping may serve the function previously served by rigid racial data and provide a foundation for important social policies.
“I think it’s hard to imagine anything like what we have today will be the same in 2020,” Prewitt said.