The mission is daunting: count every person in the United States. The political crosscurrents are strong: the Supreme Court, Congress and the White House vie for influence. The logistics are maddeningly complex: the advertising campaign will run in 20 different languages.
No wonder the last director of the U.S. Census Bureau quit, saying she was "tired of putting out fires." Six months later, in June 1998, President Clinton nominated political scientist Kenneth Prewitt to run the bureau. Prewitt, said Clinton, is "a person of impeccable credentials and proven scientific integrity." But not everybody in Washington rolled out the welcome wagon. Republican Dan Miller, chair of the House census subcommittee, mocked the selection. "The Census Bureau needs a General Schwarzkopf, not a Professor Sherman Klump," said Miller, referring to Eddie Murphy's character in the remake of The Nutty Professor.
Prewitt, PhD '63, knew what he was getting into when he agreed to oversee the once-a-decade census, the massive head-count that officials like to call the nation's largest peacetime mobilization of manpower. Traditionally, it's true, the Census Bureau flies below political radar: it's a nonpartisan agency staffed by science-minded experts, and the information it produces -- while sometimes used by one side or another to score political points -- isn't thought to be politically influenced. This time, though, things were bound to be different. The bureau's proposed statistical methods for doing the 2000 count have come under scrutiny -- and the bureau itself has come under fire. "It looked like the politics were trampling the science," says Prewitt.
Prewitt comes at it from the science side. An academic who has never before worked in government, he spent 17 years as a political science professor at the University of Chicago, a decade as senior vice president of the Rockefeller Foundation and another seven years leading the Social Science Research Council in New York. (His wife, art historian Ann Prewitt, MA '60, still lives in New York; Prewitt commutes to spend weekends with her.) Still, he sees his job as deflecting the politics so that bureau scientists "can do their jobs as professionals."
Those jobs have their origin in the Constitution. The founding fathers called for a census every 10 years to apportion seats in the House of Representatives among the states. The information finds countless other uses -- such as helping the states draw legislative districts and the federal government distribute aid dollars.
The controversy at the bureau began in 1994, when the National Academy of Sciences released a report stating that the 1990 count -- at $2.6 billion, the most costly ever -- was flawed: the head-count missed a lot of heads. According to NAS, more than 10 million people were not tallied. Minority and low-income populations were the most likely to be undercounted, the study suggested. At the same time, some 6 million people were double-counted, in part because many moved mid-census and were accounted for in two different places.
To correct for these mistakes, the academy proposed a method of so-called statistical sampling that the bureau endorsed. In past counts, after most forms had been filled out and returned, census-takers would mount expensive efforts to try to locate people who hadn't replied. Under the proposed sampling plan, the bureau would do aggressive follow-ups, but only in selected regions. From the discrepancies detected in the samples, the bureau would extrapolate the national undercount -- and adjust its totals.
Democrats supported sampling, ostensibly on grounds of scientific principle. But of course they knew that, if sampling worked to find more minorities, it could lead to a favorable redrawing of congressional districts. Republicans opposed sampling, ostensibly because the Constitution's call for an "actual enumeration" requires a traditional head-count. But Republicans also were wary of Democrats gaining an advantage. Last fall, the question -- sampling versus head-count -- reached the Supreme Court. By a 5-to-4 vote, the justices ruled that the sampling method runs counter to the Constitution; census-takers must do a head-count for purposes of House reapportionment.
But the Court did not specifically disallow sampling for other purposes; indeed it seemed to suggest that when it comes to federal funding, for example, statistics based on sampling should be used. So now the bureau intends to release two sets of figures: an actual head-count and an adjusted set based on statistical sampling. Individual agencies can choose which numbers to use.
Even with the big question apparently settled, political meddling remains a fact of life for Prewitt. On a balmy day in June, for example, Prewitt talked with a reporter while awaiting a call from a senator urging that the bureau change the marital status categories. Never mind that millions of census forms were already printed. Prewitt views all the attention with wry equanimity: "We don't feel unnoticed."
With six months to go before the forms are mailed, Prewitt is leading his campaign with zeal. The forms will appear in six languages and guides to filling them out in 50. The ad campaign, in 20 languages, will begin in November. And Prewitt has forged partnerships with some 25,000 organizations to help spread the message that completing a census form is important. "There's never been anything like this," he says of the huge outreach program.
Prewitt believes the results will surprise many people. They will show, he says, a population that moves easily, is disengaged from civic life and speaks many different languages. "The 2000 census will measure America as being more racially and ethnically diverse than people understand," he predicts. "All of a sudden, people are going to say, 'Oh my God, the country looks different.'"
Ever the professor, Prewitt says it will be good for the nation to confront that reality. But he knows he can't dwell too long on the philosophical implications of the census. He needs to focus on logistics. After all, a lot of people are counting on him.
Jesse Oxfeld, '98, is an assistant editor at Brill's Content in New York.
