Answer the Census. The Alternative Is Worse.
April 6, 2000

By Stephan Thernstrom and Abigail Thernstrom.

Mr. Thernstrom teaches history at Harvard. Ms. Thernstrom is a senior fellow at the Manhattan Institute. They are the co-authors of "America in Black and White: One Nation, Indivisible" (Touchstone, 1999).

The census has never before been such a hot political issue. Sure, there have been scattered objections to particular questions over the decades, but the nation's political leaders never threw fuel on the fires. Today, though, the long form has prompted even Senate Majority Leader Trent Lott to tell citizens they can choose which questions to answer.

One in six households gets the long form, which asks where you lived five years ago, what time you leave for work, the age of your house, the size of your electric bill and the details of your bathroom plumbing. "It's just none of their damn business," a Chicago community activist told one reporter.

Census critics argue that the extensive questioning goes far beyond the constitutional mandate for a "Census or Enumeration" of the population at 10-year intervals, for the purpose of apportioning congressional districts. And it's true that the first enumerations did little more than count the total population and distinguish people by age, sex, race and civil status (free or enslaved).

Today's census, however, is far from a recent perversion of the framers' intent. The great change occurred in 1850, when the enumerators for the first time recorded the names of all the people of the nation and asked them such questions as where they were born, what job they held, the value of the real estate they owned and whether their children were enrolled in school.

Subsequent censuses added further questions in response to changing perceptions of society's needs. In 1870, for example, a question on parental nativity made it possible to distinguish second-generation Americans from those whose families had been here longer. At the end of the 19th century, growing concern about the poor prompted new inquiries about home ownership, mortgages, the age of occupied structures and the availability of such amenities as running water, bathtubs, and toilets. The Depression, not surprisingly, prompted the gathering of more informatio about unemployment.

It is thus awfully late in the day to object to a question about how many bathrooms you have. It isn't a secret; your local tax assessor surely knows. Why worry about giving an answer that will make it possible to construct a table showing that 19.4% of American households have four or more bathrooms?

One set of questions that's drawn a lot of criticism asks how people get to work, whether they carpool and what time they leave each morning. The answers will be of obvious use to local officials assessing the need for new highways. A critic from a conservative Washington think tank darkly warns that President Al Gore could use this information in his crusade against the  internal combustion engine. Perhaps, but nothing prevents the Department of Transportation from conducting its own study of the matter, which would cost more and yield less reliable results.

Refusing to answer census questions doesn't kill a dubious government program whose operation requires information about the population. The data themselves are value-neutral, and may in fact be invaluable to critics, who can use them to show that programs are not achieving the intended effects or have other unforeseen negative consequences. And unreliable data can make bad programs worse. Erroneous assessments of labor-force availability, for instance, may encourage employment-discrimination suits that would otherwise be recognized as meritless.

The government is not alone in relying on census results. It's with good reason that the Bureau of the Census is lodged in the Department of Commerce. Many businesses use official economic and demographic data, and would be hard pressed to operate without them. Privacy isn't an issue here, since the Census Bureau releases only aggregate data. The individual forms from this year's census won't be made public until 2072.

Some questions in the 2000 census are objectionable. The use of racial categories rooted in 19th-century social science is deplorable. Having a separate question that asks whether the respondent is Hispanic is absurd—explicable only as an attempt to inflate group numbers. (If Hispanic identity were simply one "ancestry" choice among many, the Hispanic count would be about 10% lower.)

Still, most of the questions are reasonable. And the forms for the 2000 census came before Congress for approval two years ago. Where were Mr. Lott and his fellow critics then?

Today's protest could have a nasty boomerang effect. The Supreme Court has ruled that congressional reapportionment—but not the redistricting of state legislatures—must rely on numbers generated by an actual enumeration. But for purposes of allocating funds, governments may well use adjusted figures based on sampling, which the Clinton administration will also provide. Methodological problems are likely to make these numbers unreliable, but the case for using them becomes stronger if many Americans refuse to fill out their forms.

If Americans don't trust their government today, wait till they see what happens if they leave census forms blank. The result won't be smaller government. It will be government by guesswork. Faulty data will drive funding decisions about housing, education and other programs, and will be used as evidence in discrimination suits involving bank loans, employment and contracting. Is that what Mr. Lott wants?

©2000 The Wall Street Journal

About Abigail Thernstrom: articles, bio, and photo
About Stephan Thernstrom: articles, bio, and photo