ORLANDO, Fla. (AP) — Although a test showed that adding a citizenship question to the 2020 Census wouldn't have had an impact on overall response rates, it did make a difference in neighborhoods that were bilingual and had substantial numbers of non-citizens, Hispanics and Asians, the U.S. Census Bureau said Monday.
In its final report on a test conducted last summer, the bureau said there were lower self-response rates when a citizenship question was added to a test questionnaire in neighborhoods where 5 percent or more of residents weren't citizens; where almost half the population was Hispanic; where 5% to 20% of residents were Asian; and neighborhoods that received bilingual materials.
Self-response rates for the test questionnaire also were lower with a citizenship question when respondents mailed in their answers and within New York and Los Angeles. When a citizenship question was on the test form, the proportion of respondents who identified as Hispanic also was lower, the bureau said.
The test was conducted this summer as part of an effort to fine-tune planning for the 2020 head count next spring. Test questionnaires were mailed to 480,000 households across the U.S. Half of the questionnaires had a citizenship question and the other half didn't.
The bureau announced last October that preliminary results showed the test questionnaire with the citizenship question had a self-response rate of 51.5 percent, and the questionnaire without the citizenship question had a self-response rate of 52.0 percent.
At the start of the test, the Census Bureau didn't know if the question would be allowed since it was being litigated between the Trump administration, which was pushing for the question, and civil rights groups and several Democratic state attorneys who opposed it. Critics said it would reduce participation by Hispanics and immigrant groups.
The U.S. Supreme Court ruled in late June that the question couldn't be on the 2020 questionnaire.