“A Single Census Question To Measure To Measure Race, Spanish-Orgin and Ancestry” by Reynolds Farley (January-February 1995 P&R Issue)

By Reynolds Farley (Click here to view the entire P&R issue)

The Censuses of l980 and 1990 asked distinct questions about race, Spanish-origin and ancestry. This permits us to describe the vibrant tessellation resulting from the many cultures, ethnicities and linguistic groups now represented in the United States. We also have the detailed data required by the Voting Rights Act of 1965 to ensure equity in drawing up legislative districts down to the smallest local level. Nevertheless, the three questions seem overlapping and were, perhaps, confusing to many respondents. Furthermore, they produced an array of data many users immediately adjusted in order to separate non-Hispanic whites and non-Hispanic blacks from Latinos.

How will the federal statistics system -especially the Census- measure race and ethnicity in the future?

We do not know, but we may confidently make several assumptions.

First assimilation has effectively integrated whites from European origins and may now be effectively incorporating many Asians into the economic and social mainstream. But data from the 1990 Census reveal that many blacks, Native Americans and both native and foreign-born Latinos are toward the bottom of the economic ladder. Spokespersons for these groups will insist upon obtaining data pertinent to addressing these substantial social and economic gaps.

Second, court rulings may strongly influence the federal statistical system. Counts could mandate an adjustment for undercount in the next Census or could, in forthcoming redistricting decisions, place relatively great or relatively little weight upon racial and Spanish-origin data obtained from the Census. Although there is no precedent for such a ruling, federal courts could conceivably mandate the race question for the Census, including the specific categories to be listed.

Third, immigration from Asia and Latin America remains at very high levels, leading to the rapid growth of an increasingly diverse minority population. In addition, data from thc 1990 Census report sharp rises in racial intermarriage.

Fourth, as a society we are unlikely to develop a consensus about how to classify persons by race or ethnicity, so Congress and federal agencies will be subject to convincing but conflicting argument, including suggestions to delete the race question, to make it open-ended or to include many mixed-race categories. These challenges will undoubtedly increase as the number of people with parents from two racial groups or with a Spanish and a non-Spanish parent increase. Can a person be one-half Chinese and one-half white for purposes of redistricting a state legislature?

Fifth, if there is any consensus emerging, it is that, for purposes of federal policy, persons may be members of only one major group and that almost all persons may reasonably be classified into one of the following five groups:

Whites not Spanish in origin
Blacks or African-Americans not Spanish in origin
Hispanic or Latino
Asian or Pacific Islanders not Spanish in origin
Native Americans not Spanish in origin

In view of this, it would be appropriate to test the question shown below. Each individual would be asked to identify with one, and only one, of the five major groups. Then, individuals would have the opportunity to identify with a particular cultural, ethnic or national origin group. The word race would not appear on the questions,just asitdid not appear on the 1980 Census enumeration schedule.

This proposed question has three advantages. First, rather than asking the current confusing array of three distinct questions about Spanish-origin, race and ancestry, only one question would be used. Second, the question is congruent with the actual practices of most of those who use 1980 and 1990 Census data-including federal and state agencies, those working in litigation and those investigating demographic trends. Third, this question would permit all respondents -not just a sample- to identify with an ethnicity, an ancestry or a national origin.