If good decision-making depends on good data, then a lot depends on exactly what data you gather and look at. Data generated by the Census Bureau affects congressional apportionment and where hundreds of billions of federal dollars go every year.
A Census questionnaire runs about two pages. Add a page for every member of the household. The research report on which Census will base its 2020 decennial count form — that takes 380 pages.
If the Census only consisted of a simple nose count, the Census Bureau could skip exercises like the 2015 National Content Test Research Study (NCT) and its hundreds of pages of tables and analysis. But as a matter of national policy, we want to know much more than the total of warm bodies residing in the U.S. We also want to know ages, genders, racial and ethnic makeup, and home ownership. Actually, we want to know way more than that, but Census gets vast quantities of economic information from other surveys. It’s all too much to cram into the decennial count. As it is, response rates have been steadily falling, driving up the costs of conducting the count.
The Census Bureau released its latest National Content Test research — and its proposals based thereon — on the eve of President Donald Trump’s first speech to Congress — the news equivalent of the Friday afternoon before Christmas. But some of the recommendations for the 2020 form are sure to generate controversy. (The form itself requires congressional approval).
Collection of racial information didn’t start with the civil rights movement or even the Civil War. It goes back to 1790. The early national government wanted to know the numbers of free white males and females, other free persons, and how many slaves there were. Recall that in the unamended Constitution, slaves — “other persons” — were counted for apportionment purposes only to three-fifths of their actual numbers. In those days, assistant U.S. Marshals conducted the count.
Black, white. Simple proposition. Today, not so simple. The nation is vastly more diverse in terms of race and ethnicity. People think of themselves differently now. National origin and the nuances it may represent in physical characteristics make racial counting a knotty problem when you’re trying to get it onto a one-page form. Or navigating among competing interests.
For instance, if you’re from Syria, do you consider yourself white? Maybe not quite. But not any of the other choices on the form quite fit either. Census has been discovering and re-discovering for decades that people want the forms and the tallies to represent what they consider themselves to be.
On the specific question of 19 countries, Census proposes adding something long sought by Arab-Americans, namely a MENA self-designation. That stands for Middle East-North Africa. Until now, they’ve had to designate themselves as white.
Another new designation Census has proposed: “Hispanic” as a catch-all category. You might think, hasn’t that been around for a long time? Yes and no. It turns out, the NCT research shows the earlier forms confused people who otherwise consider themselves Hispanic, causing rising numbers of respondents to check off “some other race.” So Census will collapse the old sub-designations into the single Hispanic.
Check out this summary presentation to get a faster sense of what went into the thick NCT report and of how extensive the bureau’s statistical analyses can get.
These are tough waters to navigate. In some sense, if everyone is American, what difference does, or should, race make? In the real world, for reasons too complex to enumerate (pardon the pun) here, racial, ethnic, tribal and cultural distinctions have saturated public policy decision-making. That rolls down to the Census Bureau no less than to any other agency.