Wilbur Ross has added a question to the 2020 census asking if the respondent is a US citizen. The question is being added without going through normal testing procedures.
Update 6-13-19: The House Oversight Committee is in the process of investigating the addition of the citizenship question to the 2020 census, recently voting to hold Attorney General William Barr and Commerce Secretary Wilbur Ross in contempt of Congress for not complying with the committee’s subpoena request for more information. Additionally, President Trump invoked executive privilege to shield documents from the committee that were related to Ross’ decision to add the citizenship question; the issue is currently in front of the Supreme Court.
What happened: Commerce Secretary Wilbur Ross, with some help from the Department of Justice (DOJ), has added a question to the 2020 census asking if the respondent is a US citizen, a question that has not been on the census since 1950. The question is being added without going through the normal testing procedures and over the objection of scientists both within and outside the Census Bureau.
Why it matters: There are a number of problematic issues at stake – the voices of scientists at the Census Bureau are being ignored by the leadership, the citizenship question was put through few of the testing procedures that ensure accurate and high-quality data, and the question has a high likelihood of decreasing minority and immigrant residents participation in the census and thereby leading to an inaccurate population count. Countless scientists depend on accurate census data for studies spanning numerous scientific fields, meaning that the addition of the citizenship question has the possibility of hampering scientific research for years to come.
From February to December 2017, Commerce Secretary Wilbur Ross engaged in a behind-the-scenes campaign to add a citizenship question to the 2020 census, pressuring the Department of Justice (DOJ) to formally request the question from the Census Bureau. Justification by the DOJ for the addition of the citizenship question, as outlined by Secretary Ross in his March 2018 memo, was that the citizenship data would be essential in enforcing a portion of the 1965 Voting Rights Act aimed at preventing racial discrimination. This argument has been more or less refuted by the acting head of DOJ’s Civil Rights Division, who acknowledged that the enforcement of the Voting Rights Act has never depended on citizenship data directly.
The normal procedure for adding a question to the census is a long process, involving extensive testing, review, and evaluation over a 5-year period in order to ensure that the change is necessary and that it will produce useful, quality data. During the process, the Office of Management and Budget (OMB) is consulted, subject matter experts add input on the phrasing and wording of the question, the Census Bureau conducts interviews and field testing, and public comments are solicited and incorporated into the final version.
Federal scientists who have worked at the Census Bureau and outside experts who have advised the agency have objected over the addition of the question and it appears that their voices are being ignored. In the March 2018 memo, Secretary Ross spoke of holding numerous conversations with outside experts, two of whom later said to Science magazine that they told Secretary Ross that they believed it was risky to ask untested questions on the census and that the high burden of proof needed to demonstrate the question’s value had not been met. A member of the Census Scientific Advisory Committee, D. Sunshine Hillygus, has said, “I want to say in no uncertain terms I think this is an absolutely awful decision. I am still dumbfounded that this decision is coming in at such a late date. My view is that this is going to have severe negative implications for data quality and costs.” New documents reveal that the Census Bureau’s lead scientist, John Abowd, sent a memo to the acting head of the Census Bureau in January 2018, stressing that the citizenship question would lead to higher costs and lowered accuracy.
There is some evidence to suggest that asking a citizenship question on the census will lead to fear in the immigrant community, decreased participation in the census taking, and increased costs due to the need to redo sampling in areas of decreased respondents. Before the citizenship question was proposed on the census, the Census Bureau’s National Advisory Committee on Racial, Ethnic, and Other Populations conducted small, non-representative population samplings and focus groups across the country and presented their results in November 2017. Concerns raised by the respondents included: legal residency status, fear of deportation, concern about how the data are used, and which agencies can see the data (especially Department of Homeland Security (DHS) or US Immigration and Customs Enforcement (ICE)). The interviewers administering the census commented that they were being asked by the respondents time and time again about whether being a citizen made a difference in the process.
Many scientists and science organizations have actively opposed the addition of the citizenship question on the 2020 census. For instance, the American Statistical Association, the Consortium of Social Science Associations, and a number of other scientific organizations sent separate letters to Secretary Ross, objecting to the circumvention of researched and proven testing methods during the process of adding this last-minute question to the census. The US census is an incredibly important resource for scientists by providing benchmark information for biomedical science, epidemiological data for cancer registries, population trends for issues of morbidity and mortality, assessing the harmful impacts of toxic waste or scarcity of water supply on populations, and is a needed tool for practically all studies in the social sciences. Without a 2020 census that is fair, accurate, and evidence-based, we risk not only sidelining federal scientists, but sidelining one of the most important tools in science.