How a Citizenship Question in Census 2020 Would Lead to Undercount of Latino Immigrants
What is it about?
Some socially marginalized populations are more likely than others to be left out of the census count. This gives rise to disparities in federal funding and political representation for these groups and the communities they live in. A survey of Latino Immigrants in California about their willingness to self-respond to the census, respond to enumerator visits, respond to requests for information about neighbors, together with analysis of availability of administrative records, shows that adding a citizenship question to Census 2020 will result in serious undercount of this large population. The undercount will include, in addition to non-citizens, many citizens in their households and social networks.
Why is it important?
The analysis is important because it shows that efforts to compensate for pronounced and uneven levels of response to the decennial census occasioned by adding a sensitive question about citizenship cannot be "cured" by the standard procedures proposed for enumerating households that do not self-respond: reminder postcards, reference to administrative records, proxy interviews with neighbors, and imputation of household size and characteristics using the information from households that do respond. The analysis explains how inclusion of this highly sensitive question initiates a cascade of errors which gives rise not only to differential census undercount but, also undermines the reliability of data on demographic characteristics such as age and, ultimately, makes the sought-for information on citizenship unreliable as well.