News from
Office of Communications
Stanhope Hall, Princeton, New Jersey 08544-5264
Telephone 609-258-3601; Fax 609-258-1301

Contact: Steven Schultz (609) 258-5729

Monday, May 15, 2000

Multiple Race Option in Census May Be More Popular Than Expected

Study Suggests Multiracial Identifications Will Pose Challenges for Civil Rights Policies

PRINCETON, N.J. -- A study by Princeton researchers suggests that many more people are likely to identify with more than one race in the 2000 census than previously thought. The results of the count, the first to allow respondents to mark more than one race, may pose new challenges for making civil rights policies and tracking social and economic inequality.

Based on a new analysis of data from earlier surveys, Joshua Goldstein and Ann Morning of the Princeton Office of Population Research estimate the number of Americans likely to mark multiple races at some 8 to 18 million, several times greater than previous estimates by the government. The vast majority of likely multiple race respondents chose "White" when faced with the traditional single-response race question, the study found.

The actual results from the census, which was carried out in April, have not been tallied; initial data regarding race are expected by April 2001.

In a report in the May 16 online issue of the Proceedings of the National Academy of Sciences, Goldstein and Morning suggest that such a high degree of multiracial self-identification could further complicate the difficult issue of developing and implementing race-based policies, such as fairness-in-lending and affirmative action programs.

"The advantage of the new system is that it more accurately reflects how people see themselves. The disadvantage is it shifts the burden of assigning single-race labels from individuals to the government," says Goldstein. "Now the government must decide how to treat multiple responses."

The Office of Management and Budget and Department of Justice, for example, recently announced that they will allocate mixed-race individuals back to the minority race, or if there are two or more minority races, back to the race that makes sense in the particular enforcement context.

Under the old system, the government was criticized for pigeonholing people into single-race boxes. Now, says Goldstein, it may be accused of reinstitutionalizing the 'one-drop rule,' the segregation-era notion that a person with any amount of minority ancestry must be a member of that minority.

"Taken together, the issues raised by multiple-race reporting may fuel criticism not only of race-based policies but also of the rationale for the collection of racial statistics in the first place," the paper concludes.

The Princeton researchers estimated that somewhere between 3.1 and 6.6 percent of the US population, some 8 to 18 million people, are likely to have marked multiple races. That compares to the government's estimate of one to 1.5 percent, based on pilot tests of a separate "multiracial" category, which was not ultimately included in the 2000 census. Goldstein and Morning explain that the "mark one or more" option is likely to prove more popular the "multiracial" option because it does not require people to give up the single-race identities they already have.

To make their estimates, Goldstein and Morning reanalyzed data from the 1990 census and from an experimental module in the 1995 Current Population Survey. The 1990 census included a traditional single-race question and an additional question on ancestry that allowed multiple responses. Goldstein and Morning used a method developed at the Bureau of Labor Statistics to translate descriptions of ancestry into classifications of race. The Current Population Survey asked people who had chosen one race whether they would have marked additional races if they had had the choice.

"Although neither of our sources had exactly the same format as the actual 2000 census question, we were reassured by the fact that both techniques produced similar estimates," says Goldstein.

The paper emphasizes that the number of people identified as multiracial depends not only on how many people actually have mixed race ancestry, but also on how popular it is to identify oneself as multiracial. Much of this is unpredictable. "As it happens, Tiger Woods did not win this year's Master's tournament, which took place right before the census. If he had, more people would probably have chosen to mark multiple races," said Goldstein.