Who is more important in the population estimate of the United States: an impoverished urban family in a subdivided row house, or a wealthy businessman living on 2 acres in the suburbs?
That question is at the heart of a battle in court and in Congress over whether to compensate for the expected undercount of minorities in the 1990 census.
It is a classic contest over raw numbers, the grist of American politics. An agreement to adjust census figures for the undercount of minorities would have an impact on everything from the distribution of federal funds to the allocation of seats in Congress and state legislatures.
An adjustment would mean gains for Democrats and losses for Republicans. It would mean a shift of power from suburbs to cities, from whites to minorities, from some states to others.
”The whole adjustment issue is a political can of worms,” said J. Michael McGehee, a Republican consultant and a member of a panel of experts advising the federal government on the question.
”There are certainly strong political passions involved,” said Kenneth Wachter, another panel member.
The intensity of those passions was evident last week when officials from the nation`s largest cities warned the Census Bureau that the counting efforts in their municipalities were in a crisis.
The cities are afraid the undercount of minorities and of the urban population in general will be even greater this year than it was 10 years ago and cause urban areas to suffer an even greater loss of political power and money.
Fueling those fears were Census Bureau statistics showing that suburbanites were mailing back their questionnaires at a much higher rate than city residents. For example, as of last Saturday, the rate of return in Chicago was just 52 percent, compared with suburban rates that ranged from 69 percent to 74 percent.
That means the Census Bureau will have to send out someone to try to get the form filled out in person. That`s an expensive, time-consuming and not-always-effective operation, particularly in low-income urban areas.
”We`re just very, very concerned about the integrity of the census at this point,” said Jay Michaud, a lobbyist for the City of Chicago. He said there is ”a perception that people are beginning to lose confidence in the census.”
Indeed, the uneven nature of the census process, the way it counts some people better than others, undercuts the basic assumption that everyone should be treated equally.
Federal funds are doled out on the basis of the Census Bureau`s population figures, and those figures are used to determine how many representatives in Congress each state has and how the boundaries of legislative districts within states are drawn.
The discrepancies in the census mean that well-counted places receive more federal funds and have more say in the governing of the nation than their populations would warrant, and poorly counted places have less influence than they should.
Despite its immense political importance, the census is commonly viewed as a purely objective numerical measure of the nation. For that reason, some politicians are leery about taking too obvious a partisan stand on adjusting census figures to compensate for the undercount, whatever its size.
”The census is not a contest. At least, it should not be a contest between Republicans and Democrats to see who can count the most. It shouldn`t be a contest between states over who gets the most congressmen,” said Rep. Tom Sawyer (D-Ohio), chairman of the House Subcommittee on Census and Population.
”Does adjustment improve the relative quality of the count? What gets us closest to the truth? That`s the fundamental reason for having a census.”
Others, such as McGehee, are less circumspect. He argues that the decision on adjustment is too important to ”be conducted in a vacuum in the basement of the Census Bureau by well-meaning, but unrealistic, social science statisticians unconcerned with the political and economic effects of their actions.”
William Kruskal, a University of Chicago statistician and a member of the advisory committee, said, in a letter to the Congressional Research Service:
More Top Picks Best Transport Wheelchairs
”It is assumed sometimes that statisticians magically know the way to adjust. In fact there are lots and lots of ways, and they can lead to quite different results for any given unit. It is one thing to make adjustments in scientific research . . . and quite another to make an inevitable arbitrary administrative adjustment to something as close to the democratic bone as the census.”
Undercounts have been part of the census since the first was carried out in 1790.
”George Washington was disappointed that the inaugural census count of 3,929,214 did not exceed the anticipated figure of 4 million and blamed it on the undercount,” J. Gregory Robinson of the Census Bureau`s population division wrote in a 1988 scientific paper.
Every group is undercounted to some extent, but many more blacks and Hispanics, especially those in cities, are missed than whites. There are also overcounts, but these often involve rural or more affluent areas.
In 1980 the net undercount of blacks (the total undercount minus the overcount) was 1.7 million, or 5.9 percent of the actual black total, according to Census Bureau estimates. In contrast, the undercount for whites and all other groups was 1.5 million, or seven-tenths of a percent of the total.
One Census Bureau estimate indicated that, when Hispanics were treated separately, their undercount was about 4.5 percent.
The Census Bureau has a difficult time counting anyone who lives in a city. A study by Eugene Ericksen, a Temple University sociology professor and co-chairman of the advisory panel, found that the net undercount for 16 major cities, including Chicago, was 4.5 percent-nine times larger than for the rest of the nation. For blacks in those cities, it was 9.1 percent.
If figures are adjusted to compensate for these undercounts, the same amount of federal money would be handed out and the number of congressional and legislative seats wouldn`t change. But more money and political power would go to the cities that are traditionally Democratic strongholds, while less would go to the Republican-dominated suburbs.
”Well-counted suburbs believe they`re being penalized or could be penalized by a correction,” said Kirk Wolter, vice president of statistical design for A.C. Nielsen and a member of the undercount advisory panel, ”when, in fact, from a scientific point of view, it was the other (high-minority)
areas that were penalized by the census, and the correction is simply restoring a proper balance.”
The advisory panel was established last year under a consent decree in connection with a lawsuit filed by Chicago and other cities on the undercount issue. Four members, including McGehee, Wachter and Kruskal, were appointed by the Commerce Department, and four, including Wolter and Ericksen, by the cities.
Wolter, who had been the chief of the Census Bureau`s statistical research division, was one of two top bureau officials who quit over the undercount question.
”The agency in general has become more politicized in the last 8 or 10 years,” he said. ”There is less of a scientific professional quality than there once was.
”In the past several years, there have been more and more intrusions into the bureau`s affairs by the Department of Commerce. It`s an intrusion that`s been influenced by partisan politics, by the ideology of the incumbent administration.”
As part of the consent decree in the lawsuit filed by New York and Chicago, the Commerce Department agreed to consider adjusting for the undercount. But, on April 11, the cities went back to federal court to reinstate the lawsuit, contending that the agency had established guidelines that made an adjustment all but impossible.
Each side in the dispute argues that it is protecting the sanctity of the census, accusing the other side of playing politics.
V. Lance Tarrance, a Republican pollster and a Commerce Department-appointe d co-chairman of the advisory committee, said, ”You can call it either tampering or manipulation, but any time you manipulate the government`s basic statistics, you have to think long and hard. It`s like gene splicing.
”I`m reluctant to gene splice because I think we could create a hybrid monster.”
Ericksen, however, said, ”My view is that this should be decided on statistical grounds and scientific grounds, and let the chips fall where they may. It is a political fight, but it shouldn`t be.”
And that fight, likely to continue for at least another year, is still gaining momentum, even as hundreds of thousands of Census Bureau employees scramble to complete what has increasingly become a troubled 1990 census effort.