On May 4 the Washington Post published what Brian Beutler at Talking Points Memo called an “alarming—and darkly ironic” story stating voter registrations have dropped for African-Americans and Latino Americans. WaPo reporter Krissah Thompson wrote in her lead:
The number of black and Hispanic registered voters has fallen sharply since 2008, posing a serious challenge to the Obama campaign in an election that could turn on the participation of minority voters.
By some accounts, it was true. Census numbers, which measured between the 2008 presidential and the 2010 midterm elections, showed that the number of African-American registered voters had fallen from about 17.3 million to about 16.1 million nationally. But the Obama campaign and a number of academics disputed the Post’s conclusion that Obama might be in trouble, instead saying the Census’s methodology is flawed and that voter registration among African-Americans and Latino Americans is actually up.
It’s hard to tell which way is up when there have been so many laws passed recently mandating voters show ID cards, closing early voting periods, placing booby-trapped restrictions on voter registration groups and changing residency requirements for voter eligibility—not to mention legacy felony disenfranchisement. Conservative advocates like Hans von Spakovsky have suggested that voter ID laws actually help black and brown voter participation, though using flawed methodology.
So while there may have been a few hanging chads in the Post’s report, does that change the reality that registering and turning out voters of color is more difficult now than it was in elections past? If the critics of the Post story are correct, then are people like von Spakovsky right?
The most glaring issue with the Post’s story was that the reporter Thompson used Census Bureau voting data that stopped at November 2010. University of California Irvine professor Marty Wattenberg challenged the Census figures by comparing 2008 and 2012 elections figures in Florida and found that Latino-American registration increased by 9 percent. For Asian-Americans, the increase was 14 percent. Further, Wattenberg told Hasen (and confirmed with me by e-mail) that as recently as 2006, Republicans outnumbered Democrats among Hispanics, but by 2012 the reverse was true.
George Mason University professor Michael P. McDonald attacked the very DNA of the Census figures, saying that the Current Population Survey (CPS), upon which Census voter data is compiled, is flawed for relying on self-reporting from a single person in a household. Further, wrote McDonald, the Census Bureau classifies persons who gave no response—because they refused, didn’t know the answer or just failed to provide one—as a “No” response on voting. Meaning, if a CPS rep called your house and your dad answered, if your father didn’t know whether you voted or not then the CPS interviewer recorded his answer as a “no,” i.e., you did not vote—whether you did or not.