04/05/2009 03:40 pm ET | Updated May 25, 2011

Groves Update

Some updates and a bit more background on the nomination Robert Groves to run the Census Bureau announced on Thursday.

First, in addition to the Associated Press and article I linked to on Thursday, the Groves nomination was also the subject of stories on Friday by The Washington Post, New York Times and the Detroit Free Press. The Post's Ed O'Keefe also blogged excerpts of the press releases issues by various lawmakers and interest groups in reaction to the Groves nomination.

Second, in a development that should surprise no one, the world of survey and statistical science is quickly rallying behind the Groves' nomination. Seven statistical and social science organizations have signed a letter urging confirmation. These include:

  • Consortium of Social Science Associations

  • Council of Professional Associations on Federal Statistics
  • American Sociological Association
  • American Statistical Association
  • Population Association of America
  • Association of Population Centers
  • AAPOR: the American Association for Public Opinion Research
  • Third, note well the point that the FreeP's Todd Spangler honed in on: The argument over the use of statistical sampling in connection with the upcoming 2010 decennial census is now essentially moot:

    Even before the announcement was made, Groves, 60, a sociology professor and director of the Survey Research Institute at U-M, was being criticized by congressional Republicans. They remembered him advocating statistical sampling in the decennial count of the U.S. population when he was an associate census director.

    It wasn't used because the Commerce secretary at the time rejected it, and by the end of the '90s, the U.S. Supreme Court ruled that sampling could not be used in setting congressional representation.


    There are no plans to use sampling in next year's count. The bureau's mission plan doesn't call for it, and it's too late to start now. Census officials have questioned in the past whether they have an effective way of employing sampling to correct the tally of historically undercounted populations, like young people, transients, African Americans and Hispanics.

    Plus, Commerce Secretary Gary Locke said in confirmation hearings recently that sampling wouldn't be used in next year's census, and, just as in the 1990s, it's his call.

    Finally, some links to further reading: Before the Groves nomination, Andrew Reamer of the Brookings Institution and my colleague Eliza Newlin Carney of National Journal wrote about the issues facing the Census Bureau. And for those who want to get deep into the weeds of the science of the ongoing debate, consider the reports issued by the National Research Council (NRC) of the National Academy of Sciences (NAS), particularly their 1999 report, Measuring a Changing Nation: Modern Methods for the 2000 Census, and last year's Coverage Measurement in the 2010 Census (both have free online versions).

    While the discussion in the NRC/NAS reports is highly technical, a quick skim of the executive summaries reveals the way the 1999 Supreme Court decision barring the use of statistical sampling for counts used for congressional reapportionment has impacted the scientific consensus. Just before the decision, NRC's panel endorsed the use statistical sampling to follow-up with those that initially fail to return their census form. Last year, the NRC report focused more on developing statistical databases and models to improve the way the census works.