Written by: Director Robert Groves
There’s been some controversy over the race question wording in the 2010 Census form. The race question has often changed over the decades, as the country has changed and racial terms have evolved.
The 2010 Census form offers 15 categories for race:
- Black, African Am., or Negro
- American Indian or Alaska Native
- Asian Indian
- Native Hawaiian
- Guamanian or Chamorro
- Other Asian (print race in box below)
- Other Pacific Islander (print race in box below)
- Some other race (print race in box below)
The category “Black, African Am., or Negro” was used in Census 2000, based on research in the late 1990’s that showed there was an older cohort of African-Americans who self-identified as “Negro.” Surprisingly, about 56,000 persons took the time to write in under the “some other race” category the word “Negro.” Above half of them were less than 45 years of age in 2000.
The Census Bureau didn’t do any research on the respondent reaction to the word “Negro” in the 2000’s, but did do tests that showed answers to the ethnicity and race questions tended to change depending on the order of the questions. I think some research on the sensitivity of answers to the presence of “Negro” should have been done last decade, but I am unaware of what limitations there were on the research program then.
Some of the commentary on the question comes from people offended by the term. I apologize to them. I am confident that the intent of my colleagues in using the same wording as Census 2000 was to make sure as many people as possible saw words that matched their self-identities. Full inclusiveness was the goal.
Nonetheless, my review of our other demographic surveys showed that we don’t use the term in them. A few calls around the country showed that the term is not used in most other household surveys. We are currently using the term in the American Community Survey questionnaire (to make it consistent with the decennial census), and a 2010 Census test of the effects of removing the term on answers will inform the next version of the question. My expectation of the test result is that omitting the term has little effect on the response distribution, but we must do the test to make sure about this.
African-Americans need to be fully counted in the 2010 Census, and I hope this controversy doesn’t reduce their participation.