This is a followup to my April 17th post regarding the Citizenship Question on the Census. In that post I cited the several federal court lawsuits, and the accompanying correspondence from the US Department of Justice and US Department of Commerce.
My key takeaway from all of this is that adding the citizenship question to the 2020 Census would yield highly precise (city block level), yet inaccurate (biased) data.
The “highly precise” data would be made available for over 11 million “blocks” — the finest level that the US Census Bureau releases “short form” census data.
The “inaccurate (biased)” data is due to what we call “non-sampling error” — the results are biased towards an under-count of non-citizens due to respondent’s intentional or unintentional inaccurate (wrong) responses to the citizenship question. It is clear that there is a bias in the citizenship data from the Census, as expressed in the 3/26/18 memo from the US Secretary of Commerce. The Census Bureau estimates that the non-citizen population is under-estimated by 30 percent at the national level. This is important.