The Census: Statistical Sampling is a Hot Potato
Friday, April 17, 2009
Try dropping the words "statistical sampling" while in conversation with a Republican politician and watch the sizzle. Why does such a nerdy blending of adjective and noun inflame GOP lawmakers like House Minority Leader John Boehner and Vista's conservative Congressman Darrell Issa? Issa is the top Republican on the House Oversight and Government Reform Committee.
Before I get to the answer, here's a little background. It starts with the U.S. census which counts every person who lives in the 50 states, the District of Columbia, and U.S. territories such as Puerto Rico and Guam every 10 years. The census is mandated by the United States Constitution and its results are about power and money. Already, the reasons for the political heat are taking shape.
Let's start with money. The more people in a state, county, city or town, the more of the $300 billion of federal money will be distributed each year for road projects, schools, educational programs, social services and hospitals. Then there's power. The data gathered by census takers is used to reapportion the 435 seats in the U.S. House of Representatives, and also draw the boundaries of congressional district lines and state legislative lines. In short, the nation's 435 House seats are divided proportionately among the 50 states based on their percentages of the national population. So, for example, California now has 53 members and is the most populous state. With numbers comes political clout. Unfortunately, some surveys indicate that California's double-digit unemployment has taken its toll with more people moving out of the state than are moving in. So it is possible that California could lose one or two seats in the House once the numbers are in.
And therein lays the root of the controversy. Will the census head count be accurate or will there be large populations that refuse or are unable to be counted? Remember that every person must be counted, including illegal immigrants, urban minorities, hermits, the homeless, students, children, and even those living overseas. For California with its high numbers of Asian and Hispanic immigrants -- both legal and illegal -- an accurate count from returned questionnaires and follow-up home visits is unlikely.
The 1990 census missed an estimated 8 million people - mostly immigrants and urban minorities. In 2000, about 4.5 million people were not counted, mostly blacks and Hispanics. So the concept of using statistical sampling in addition to an every-person count is gaining favor. A sample population count would provide statistics from which would be extrapolated a calculation for the larger population.
Republicans generally oppose statistical sampling, while Democrats favor it. The basic explanation for their disagreement is that statistical sampling is expected to disclose larger numbers of urban immigrants and minorities than does the head count. These groups tend to vote Democrat and will increase the power of those urban areas in terms of congressional districts and spending. Also, Republicans are more likely to advocate sticking with the U.S. Constitution which mandates a count of every person rather than taking a sample.
What is bringing all this to the forefront now is President Obama's choice earlier this month of Robert Groves as the new Director of the U.S. Census Bureau, an ardent advocate of statistical sampling and not popular with Republicans. Additionally, Republicans are concerned that the White House plans to take an active role in the 2010 census, and that the White House is significantly influenced by minority political organizations.
Between now and April 1, 2010, expect the sizzle to become a full-out political conflagration if Professor Groves, the sociologist and survey expert chosen by the Administration to head up the census, does indeed turn to his statistical training to figure out how many people live in this land of ours.
To view PDF documents, Download Acrobat Reader.