Americans live in a permanent state of siege. We are bombarded by telemarketers, direct mail, commercials, faxes and e-mails. It is this constant assault on our time and sensibilities that most threatens what should be a national treasure: the once-a-decade Census.
The Census Bureau plans to mail 98 million forms this week to most homes and apartments (an additional 22 million are being hand-delivered in rural areas). Naturally, the forms arrive with the bills, junk mail and magazines. The Census Bureau predicts that only 61 percent of the forms will be returned. This would be lower than in 1990 (65 percent), 1980 (75 percent) or 1970 (78 percent).
FILL IT OUT.
It's the least you can do. John Kennedy famously said: "Ask not what your country can do for you--ask what you can do for your country." The truth is that our country doesn't ask much. There's no military draft or requirement for national service; we have to pay our taxes and obey the laws. Otherwise, we're left alone. Completing the Census form once a decade is a rare duty (actually, it's also the law).
What we get is a good population count--required by the Constitution for congressional apportionment--and a useful snapshot of social and economic conditions. About five sixths of homes receive the "short" form, which asks basic population questions (age, sex, race, ethnic background). The rest get a longer form that also has questions on citizenship, education, disability, commuting, income, employment, housing, car ownership and utility costs. The answers are inherently interesting and inform political debate.
Everyone knows the Census isn't perfect. How could it be in a country so big and messy? Still, the overall undercount in 1990 was less than 2 percent (1.8 percent by one estimate, 1.6 percent by another). This isn't bad, considering that some people--criminals, illegal immigrants--may avoid government and others don't speak English. The 1990 Census found that there were 26 foreign languages spoken exclusively by people at home. These included Spanish (17.3 million), Chinese (1.2 million), Tagalog, a language of the Philippines (843,000), and Korean (626,000).
Practical problems multiply when people don't respond voluntarily. The Census Bureau hires an army of "enumerators," estimated at 500,000 this year. The idea is to count everyone by visiting all the homes and apartments that didn't respond. The obstacles are enormous, as political scientist Peter Skerry of Claremont McKenna College shows in his forthcoming book, "Counting on the Census?" (Brookings Institution Press. $25.95).
In poorer neighborhoods, homes and apartments "may not have doorbells that work, doorbells at all, identifiable address numbers or mailboxes with names on them," he writes. Some people won't open doors to strangers for fear of crime. Some homes don't have fixed families or households. People simply come and go.
To overcome the problems, the Census Bureau has started an ad campaign and created 98,000 "partnerships" with local groups that (in theory) will help in difficult neighborhoods. It has a Web site (www.2000.census.gov) where people can complete the short form. (Note: every Census form has a 22-digit identification number. To use the Web site, you must supply the identification number on your mailed form.) It has six toll-free phone lines: English (800-471-9424), Spanish (800-471-8642), Chinese (800-471-9401), Korean (800-471-9131), Tagalog (800-470-9897) and Vietnamese (800-471-7913).
Just why public cooperation has eroded is unclear. One cause is immigration, though it can't explain the full decline. Jay Waite, the director of operations for the Census, recounts various theories: people are more mobile; they're more squeezed for time; they're more distrustful of government; their privacy fears have risen; their civic commitment has declined.
What's also clear is that some people don't want to be counted. In poorer areas, writes Skerry, people often distrust outsiders--and neighbors. Some conceal "drug use, parole violations, or welfare fraud." Some won't report income (even from legitimate businesses) that hasn't been declared for taxes. Some worry that the extent of overcrowding, if known to landlords, could "lead to rent increases or evictions."
The fears are exaggerated. The Census Bureau credibly promises not to reveal personal information to anyone, including other government agencies. But the fears are understandable. What's impressive is how well the Census Bureau does despite the problems.
Ironically, you wouldn't know it. The bureau focuses obsessively on the undercount and claims it can be cut sharply through a statistical "adjustment." The Clinton administration pushed this plan for congressional reapportionment (the split of House seats among states) until the Supreme Court ruled it illegal. The administration still wants the adjusted numbers used for redistricting (the drawing of congressional boundaries within states). "This is going to be a matter of tremendous litigation," says Lee Price, a top Commerce Department official. "There will be suits in the 50 states."
Great. As Skerry shows, the effects of the undercount are exaggerated. A General Accounting Office report in 1999 estimated that less than 1 percent of federal grant money to states would have changed if "adjusted" population figures had been used. After the 1990 Census, perhaps one House seat would have shifted among states if "adjusted" figures had been followed, says Price. The impact on redistricting might be slightly larger. However, it's not clear which party would gain or that the "adjusted" figures would be more accurate.
The oversize controversy involving the undercount weakens the credibility of the Census and the public's reasons to cooperate. Why respond if the final count is to be "adjusted" by statisticians? And if the underlying numbers are so crummy, why believe them at all? The Census has always tried to count everyone. This is still the best way. But it can't survive in the face of public indifference or cynicism. So when the Census form arrives, don't think about filling it out. Just do it.
