The Washington Post recently reported that according to the U.S. Census Bureau’s Current Population Survey (CPS), the number of black and Hispanic registered voters fell sharply between 2008 and 2010—declining 7 percent and 5 percent, respectively.
As the article notes, comparing presidential election years with off-year elections is not ideal because, in general, registration numbers decline significantly after a presidential election for all groups.
McDonald cites concerns not about the survey itself, but about how the registration and voting data are reported. For example, under voter registration the CPS has a “no response” category which includes, “those who were not asked if they were registered as well as those who responded 'don't know,' and 'refused.'” The CPS treats this category the same as the category for those who have affirmatively stated they were not registered.
McDonald suggests that it would be more appropriate to treat these “no responses” as missing and not include them in calculating registration or voting rates.
Factoring missing data in and comparing similar election years—2006 and 2010—McDonald finds essentially the same registration rates in both years for Hispanics (63.5 percent in 2006 and 63.7 percent in 2010) and an increase in Black registration rates (75.9 percent in 2006 and 79.3 percent in 2010).
|Hispanic (CPS unadjusted)||57.9%||53.7%||59.4%||51.6%|
|Hispanic (CPS adjusted)||68.5%||63.5%||71.3%||63.7%|
|Non-Hispanic Black (CPS unadjusted)||69.0%||61.3%||70.2%||63.2%|
|Non-Hispanic Black (CPS adjusted)||82.7%||75.9%||85.7%||79.3%|
Note: CPS adjusted rates treat “no responses” as missing data.
Source: Michael McDonald, “Is Minority Registration Really Declining?” Huffington Post, May 7, 2012.