Subject: Census 2000 News Alert Date: Tue, 12 May 1998 18:27:12 -0400 From: Keri Monihan <kmonihan(a)ccmc.org> May 12, 1998 NEWS ALERT House Census Panel Challenges Portrayal of 1990 Census as a Failure Hearing Scheduled to Review Data Collection on Census 'Long Form' At a May 5 hearing to review the results of the 1990 census, the chairman of the congressional census oversight panel suggested that the last decennial count was not the failure portrayed by many stakeholders. Rep. Dan Miller (R-FL), chairman of the Subcommittee on the Census, called 1990 "a pretty good census, the second most accurate in census history" because it counted 98.4 percent of the population. On the other hand, the chairman said, "[s]ampling was the failure in 1990." Wade Henderson, executive director of the Leadership Conference on Civil Rights, disagreed, noting that the 1990 census was the first "in five decades [to be] less accurate than its predecessor" and that the "differential undercount was the highest ever recorded." "To those who are willing to settle for similar results in 2000," Mr. Henderson asked, "how will we explain this to persons who are among the undercounted?" He also reminded the subcommittee that many of their colleagues, both Republicans and Democrats, had called the 1990 census a failure and supported the use of sampling methods to correct the undercount at that time. The subcommittee also heard from a panel of two statisticians and one demographer. Dr. Jerry Coffey, a mathematical statistician retired from OMB; Dr. Philip Stark, Professor of Statistics at UC Berkeley; and Kenneth Darga, a demographer with the State of Michigan, said that based on their review of Census Bureau evaluations of the 1990 census provided by the subcommittee, they believed the methods used to measure how many people were missed were too flawed to give an accurate result. Dr. Stark said that the Bureau's undercount figures reflected more problems with the methods themselves than an actual undercount. "Adjustment puts in far more error than it takes out," Dr. Stark concluded. Dr. Coffey said that the larger sample for the post-census quality check survey would not solve the problems of statistical errors in the adjustment methods. Dr. Stark was quoted in The Washington Times on May 6 as saying that he disagreed with the findings of several National Academy of Sciences panels because "I place more trust on evidence than on letterheads. There is no data and no mathematical theory to support the use of sampling in this way." The subcommittee did not invite witnesses who believe the Census Bureau's methodology is sound or who had participated in the evaluations of the 1990 census. Rep. Thomas Sawyer (D-OH), chairman of the census oversight panel during the 1990 count, and Rep. Thomas Petri (R-WI), senior Republican on the census subcommittee when planning for the 2000 census began, also testified and urged their colleagues to reduce the political rhetoric surrounding the census. Congressional hearing notice: The House Subcommittee on the Census will hold a hearing on Thursday, May 21, to review the proposed questions that will be asked on the so-called 'short' and 'long' forms in the 2000 census. The hearing will begin at 1:00 p.m., in Room 2247 Rayburn House Office Building. The subcommittee has not yet announced a witness list. Race and ethnicity update: The Census Bureau's various advisory committees will hold a joint meeting on June 3 to discuss the tabulation of multiple responses to the race question in the 2000 census and other surveys. A task force set up by the Office of Management and Budget (OMB) has been working to develop standards for tabulating the data when respondents check off more than one race. OMB decided last October, as part of a comprehensive review of its policy on racial and ethnic categories, to allow more than one response when Federal agencies collect data on race. The day-long advisory committee meeting, which is open to the public, will take place at the Holiday Inn Hotel & Suites, 625 First St., Alexandria, VA. Please call Carol McDaniel (301/457-2308) at the Census Bureau for more information. Legal update: Several new parties have weighed in on the side of plaintiffs challenging the constitutionality of sampling in the census. The State of Wisconsin and more than a dozen organizations, including Americans for Tax Reform and the American Conservative Union, filed amicus curiae briefs in the lawsuit filed by the House of Representatives at the direction of Speaker Newt Gingrich. Wisconsin sued to prevent an adjustment of the 1990 census, fearing the loss of a congressional seat to California. Population projections show that the state is likely to lose a congressional district following the 2000 census. The Pennsylvania counties of Bucks and Delaware, and DuPage County, Illinois, have moved to intervene in Glavin v. Clinton, which also asks the courts to rule that the Constitution does not permit sampling. Cobb County, Georgia, is an original plaintiff in Glavin. More than two dozen Georgia counties sued the government during the 1990 census to force a correction of the undercount using sampling methods. The post-census survey showed that Bucks and Delaware Counties were overcounted in 1990. Stakeholder activities: The chair of the 2000 Census Advisory Committee, Mayor Ann Azari (Ft. Collins, CO) sent a letter on April 17 to Commerce Secretary William M. Daley, expressing concerns raised by committee members at their March quarterly meeting. Referring to ongoing debates over key census issues, Mayor Azari said that the "credibility" of both the Census Bureau and the census process were at stake and urged "reinforcement that the Census Bureau is a credible, capable, professional organization [that] is the best in class at what they do." Noting the continued uncertainty over how the next census will be taken, the Advisory Committee urged the Secretary to provide comprehensive information on the census process to Congress so that an "intelligent and informed" decision on census methods could be made. Azari also said that new census methods have been the subject of controversy in the past, noting that Congress questioned the transition to a mail-based census in 1970 because the procedures hadn't been tested on a national scale. The Census Bureau, Azari wrote, "does have a history of innovation." The committee also encouraged the prompt nomination of a permanent Bureau director and attention to the "basic building blocks," such as address lists and promotion, that are needed to take a good census. The next quarterly meeting of the 2000 Census Advisory Committee will be June 11-12, at Census Bureau headquarters in Suitland, MD. The meeting is open to the public; however, if you plan to attend, please contact Ms. Pat Ellis at 301-457-2095 to ensure proper clearance and to obtain an agenda. The Subcommittee on the Census has notified us that Chairman Dan Miller will lead a discussion sponsored by The Heritage Foundation entitled, "Virtual Representation: Is Census Sampling Good Enough?" Matthew Glavin, President of the Southeastern Legal Foundation and lead plaintiff in a lawsuit challenging the constitutionality of sampling in the census, and Lee Price, Chief Economist (and former Acting Under Secretary), Department of Commerce, will respond to the chairman's remarks. The event will take place on May 19, at 12:00 p.m., at the Heritage Foundation offices. Please call Paul Love at 202-675-1752, if you would like to attend or for further information. Questions about the information contained in this News Alert may be directed to TerriAnn Lowenthal at (202) 434-8756 or, by e-mail, at <TerriAnnL(a)aol.com>om>. Please feel free to circulate this information to colleagues and other interested individuals.