Why An Agency Blinked
The Census Bureau blinked. Under pressure to be more customer-friendly--and to spend less public money than it did in its 1990 head count--the Census Bureau plans to scale back the number of subjects covered by its census 2000 questionnaires.
Big users of census statistics have mixed reactions to the plans, which were unveiled on March 31.
"I'm unhappy because the counties really need the data and I want everything for my counties," said Jacqueline Byers, research director for the Washington-based National Association of Counties. "But I understand why the Census Bureau is doing it."
The bureau is cutting back its questions because Congress--and many Americans who live beyond the Capital Beltway--want it to. The more questions asked in the decennial census, the angrier and less cooperative people are about voluntarily filling out the forms and returning them in the mail.
That means the Census Bureau has to spend more money to get what is most likely to be less accurate data. The combination of a resentful public and a bloated bureau budget make Congress really mad. Although a decennial census is mandated in the Constitution, Congress decides what questions get asked.
The decennial survey comes in two sizes. The short form goes to every household in the country. In the year 2000, if Congress concurs with the bureau's plan, the census will ask questions on only seven subjects--down from 12 in 1990, and the fewest number in 180 years. The long form, which goes to about one household in six, would cover 27 additional subjects. The Census Bureau would accomplish the streamlining by transferring five subjects from the short form to the long form, adding one new subject to the long form and dropping five 1990 long-form subjects altogether.
Most of the topics shifted or shafted deal with housing. Subjects moving from the short to the long form, for example, include home value, monthly rent, and number of rooms and units in buildings. Dropped altogether are questions dealing with water, sewage and condominiums.
Not surprisingly, housing specialists are the most exercised about the change.
"If we start losing this data now, it's going to be harder and harder to understand the housing dynamics in this country," said Barry Zigas, executive director for housing impact at Fannie Mae, the government-sponsored home-financing company. "It's ironic. The Administration is committed to a strong effort to increase homeownership, and at the same time, it's dropping this data collection" needed to meet that commitment.
Planners in other fields use these data--such as house value or number of rooms--as a proxy for other kinds of economic indicators. "Indirectly, it affects our industry," said Joan G. Naymark, director of research and planning for Dayton Hudson Corp. in Minneapolis, one of the country's largest retail chains.
The difference between the information obtained from the short and the long forms is in reliability for small geographic areas. Short-form statistics are valid down to the block level; long-form statistics only to the "census tract" or neighborhood level. Planners, whether for county governments or shopping mall developers, like to know the economic characteristics of places block by block.
Nonetheless, "we can live with this if the census is done correctly," Naymark said. "If the long form happens and happens on Census Day [April 1, 2000] and is accurate."
Simply scaling back may not be enough for some Members of Congress. Rep. J. Dennis Hastert, R-Ill., chairman of the House Government Reform and Oversight subcommittee that oversees the census, for example, has floated the notion of separating the short and long forms--collecting only the short form information on Census Day, while obtaining the long-form information at a different time.
Most statistical experts oppose that scheme, maintaining that the result would be less accurate data obtained at a greater cost. Hastert's approach would require two collection procedures with double the overhead, critics say. And even if more people responded to the short form, far fewer than usual would be likely to complete the long form.
The Census Bureau cut back the number of questions by applying a tough standard to each proposed topic. Only statistics that Congress has directed the bureau to collect or that can't be obtained otherwise made the cut.
"That's the wrong ruler," said David A. Crowe, staff vice president for housing policy of the National Association of Home Builders in Washington. "That's too narrow a focus" to apply to the government's statistical gathering, he added.
In any event, simply slashing the number of questions might not result in appreciably broader participation in the next survey, Census Bureau officials concede. So they also intend to use clearer language, explain the reasons for the questions and in general make their forms easy to comprehend and complete.
Even if participation soars, some specialists won't be satisfied because, they say, data quickly become obsolete these days. "The world is changing so fast, said Marie V. Bousfield, a demographer in the Chicago's municipal planning department. "What we really need is continuous measurement, not a snapshot."