Census Bureau: The counting of America
Amid partisan bickering, the Census Bureau is gearing up for its once-a-decade survey. How much does the census matter?
It's time - once again - for Americans to be counted.
It's time - once again - for Americans to be counted.

hy do we have a census?
It’s mandated by the U.S. Constitution. Article I, Section 2 states that an “actual enumeration shall be made within three years after the first meeting of the Congress of the United States, and within every subsequent term of 10 years.” The first census, in 1790, was a modest affair, costing only $45,000 and employing 650 marshals and their assistants. The 2010 census, by contrast, will cost upward of $15 billion. Questionnaires will go to 145 million households, and those that do not respond can expect a visit from one of the 140,000 census workers who will try to ensure that everyone living in the U.S.—an estimated 305 million people—is counted. The massive undertaking, says UC Berkeley statistician Philip Stark, is the nation’s “largest mobilization in peacetime.”

Why is it such a big deal?
The federal government relies on census data to determine the number of congressional seats each state is entitled to. Following the 2000 census, for instance, Arizona, Texas, Florida, and Georgia each gained two House seats, while several states, including Illinois, Ohio, and New York, lost seats. Experts predict that in the wake of the 2010 census, the South and West will each gain five House seats, while the Midwest will lose six and the Northeast four. Washington also uses the census to allocate billions of dollars in funding for everything from health-care programs to infrastructure projects. Every person counted, it is estimated, brings in $1,000 in federal money. Finally, the raw information gathered allows demographers and statisticians to chart the changing face of our nation. The Census Bureau, says agency spokeswoman Shelly Lowe, is “the Fort Knox of data.”

But is it accurate?
Actually, the census has always been notoriously imprecise. George Washington estimated that the original 1790 head count of 3,929,326 was probably off by as much as 3 million—a figure he attributed to tax resisters, those with something to hide, and lazy marshals. Then–Secretary of State Thomas Jefferson, who oversaw the operation, was himself counted twice. Over the decades, counting methods improved, but problems have persisted. The Census Bureau acknowledged that in 1990 it missed 8.4 million people and counted 4.4 million others twice. In 2000, the undercount was estimated at 3 million.

How else could it be done?
Many believe that the Census Bureau would get better results by using statistical sampling methods. As with polling, the process involves obtaining information about a small segment of the population; from that sample, experts make educated estimates about the uncounted portion. Any use of statistical sampling would likely supplement—not necessarily replace—the old-fashioned census methods. But statistical sampling is extremely divisive.

Why is sampling controversial?
Because supporters and opponents tend to break down along partisan lines. Democrats favor sampling because the people who are traditionally hardest to count are the urban poor, minorities, and immigrants, all of whom tend to live in Democratic strongholds and vote Democratic. These groups are often undercounted because they move so frequently and do not trust government employees asking questions. Republicans, by contrast, stress that the Constitution specifies an “actual enumeration” of the population, not an estimate. They also argue that statistical sampling is inferior to counting. “Anyone familiar with public opinion polling can tell you that statistical sampling carries a margin of error,” Republican Reps. Darrell Issa and Patrick McHenry recently wrote. “And error is the enemy of a full and accurate census.”

Is a change in methodology likely?
Not anytime soon. In 1999, the Supreme Court ruled by a 5–4 vote that statistical sampling, which the Democratic Clinton administration had hoped to employ, could not be used to reapportion House seats. The decision did leave open the possibility that sampling could be used to decide how congressional districts are drawn and to determine the flow of money within them. But while the Obama administration may be interested in going that route, as both a political and practical matter, it’s not possible at this point. The 2010 census is “a rocket on the launch pad, and they’re about to ignite it,” says former census director Kenneth Prewitt. “We can’t redesign rocket fuel at this stage.” Nevertheless, the dispute over sampling hangs like a shadow over the 2010 census.

Why is it still an issue?
The new census director, veteran University of Michigan survey researcher Robert Groves, has argued for years that the government has persistently undercounted millions of minorities who typically vote for Democrats, and that statistical sampling will provide a more accurate count. Republicans blocked his confirmation until this week, even though Groves promised that he would not push for sampling in 2010. Still, Republicans remain suspicious, and it’s likely that the partisan argument over sampling will resume at a later date. “What’s not political about the census?” says Harvard government professor D. Sunshine Hillygus. “It’s the basis of the two most important things in politics: money and representation.”

Questions, questions
America’s changing character and sensibilities can be gleaned by the questions officials include in the decennial survey. At various times in the 19th century, citizens were asked how many guns and dogs they owned, how many hogs they had slaughtered, and if any family member had died from an abscess. In 1830, the government recorded the number of deaf, dumb, and blind people in a household; a category for “insane and idiotic” was added in 1840 and remained for 50 years. In 1930, people were asked if they owned a “radio set”; in 1950, they were queried about their TVs. Race has long been a contentious census issue: The first census, in 1790, counted slaves as only three-fifths of a person, and Indians weren’t counted at all. By the 1890 census, many of the questions were racially based, with such specific categories as “mulatto” (one-half black), “quadroon” (one-quarter), and “octoroon” (one-eighth) being included. Marital status was not included on census forms until 1880. In 2010, for the first time, same-sex married couples will be able to officially claim that status.



Subscribe to the Week