According to the Census Bureau, the 1990 Census missed 8.4 million people and double-counted 4.4 million others. The net undercount rate of 1.6 percent (4 million people) in 1990 was 50 percent greater than in 1980. In addition, the difference between the undercount of whites and the undercount of ethnic minority groups (known as the `differential undercount' was the highest ever recorded since the Census Bureau began conducting post-Census evaluations in 1940, missing 4.4 percent of African Americans; 5 percent of Americans of Hispanic origin; 2.3 percent of Asians and Pacific Islanders; and, more than 12 percent of Native Americans living on reservations. While children under the age of 18 represented 26 percent of the total national population that year, they accounted for an incredible 52 percent of the undercount.
In 1990, Robert Mosbacher, Secretary of the U.S. Department of Commerce, decided against adjusting the Census to correct for the undercounts and overcounts despite the recommendation of Census Bureau Director Barbara Bryant to do so and a U.S. Department of Justice memo saying that the use of sampling is both constitutional and legal. As a result, the 1990 Census was the first one in five decades to be less accurate than its predecessor.
Upset with the results and cost of the 1990 Census, Congress unanimously passed and President Bush signed into law the Decennial Census Improvement Act of 1991 (P.L. 102-135) that directed the National Academy of Sciences (the Academy) to study "the means by which the Government could achieve the most accurate population count possible."
In February, the U.S. Department of Commerce and the Bureau of the Census unveiled "The Plan For Census 2000" in response to the 1991 legislation and the subsequent guidelines and recommendations of the National Academy of Sciences. Since 1990, the Census Bureau had worked hard to research, test, and evaluate Census methods to achieve the objectives set by Congress in the 1991 Decennial Census Improvement Act. The Bureau has been guided by recommendations from independent experts, including three panels of the National Academy of Sciences, the General Accounting Office, and the Commerce Department's Office of Inspector General.
In one report to the Bureau, the National Academy of Sciences concluded, "It is fruitless to continue trying to count every last person with traditional Census methods of physical enumeration...[P]hysical enumeration or pure `counting' has been pushed well beyond the point at which it adds to the overall accuracy of the Census...Techniques of statistical estimation can be used, in combination with the mail questionnaire and reduced scale of follow-up of non-respondents, to produce a better Census at reduced costs."
The resulting plan for 2000 combines a more aggressive direct enumeration effort, including several mailings to every household and multiple response options, with modern scientific sampling techniques to complete the count of the final non-responding households and to eliminate the pervasive differential undercount of people of color and the urban and rural poor.
In March, the U.S. Supreme Court ruled that Congress has delegated "virtually unlimited discretion" in how to conduct the Census to the Secretary of Commerce, and given him broad latitude in how the Census is conducted. The Court ruled that the Secretary was within his authority to reject the recommendations to correct the 1990 Census. The ruling came in a lawsuit filed in 1991 by the City of New York and several other state and local governments and outside stakeholders who called on the Secretary of Commerce, Robert Mosbacher, to adjust the 1990 Census to correct the undercount.
In September, the House Committee on Government Reform and Oversight voted 22-12 along mostly partisan lines in favor of a committee report Sampling and Statistical Adjustment in the Decennial Census: Fundamental Flaws that recommended against the Census Bureau's proposed use of sampling to complete the initial count and reduce the disproportionate undercount of children, people of color, and the rural and urban poor. The committee held only one hearing on the Census in 1995, before issuing the report.
In a September letter to the U.S. Senate, the U.S. Conference of Mayors (USCM) said that it is "absolutely unforgivable for the federal government to knowingly undercount millions of citizens of the United States of America when there is an acceptable means of including these persons in the official count of the Census."
Also in September, the American Statistical Association's Blue Ribbon Panel on the Census issued a report endorsing in principle the Census Bureau's plans to incorporate scientific sampling techniques into the 2000 Census "because sampling potentially increases the accuracy of the count while reducing costs."
In January, the American Sociological Association (ASA) Council unanimously adopted a resolution citing sampling "as an important and valid scientific method for containing costs and improving the accuracy of the Decennial Census." The resolution urges the Secretary of Commerce and Congress to "support unequivocally" the use of scientific sampling for follow-up of non-responding households and for reducing the differential undercount in the 2000 Census.
Also in January, the National Academy of Sciences renewed its recommendation that the Census Bureau use scientific sampling methods to supplement the traditional head count. The Academy's National Research Council states in its report Preparing For the 21st Century
Challenges Facing a Changing Society that the "application of modern statistical methods provides an opportunity to obtain more accurate Census results at lower costs than in recent Censuses.
In February, the U.S. General Accounting Office (GAO), released its "High Risk Series" report for the 105th Congress. The 2000 Census was one of five new program areas added to the list of Federal activities at risk of failure. According to the report, the inability of Congress and the Census Bureau to agree on Census methods could lead to an "unsatisfactory" Census at a "substantially higher cost."
In April, the National League of Cities, U.S. Conference of Mayors, National Association of Counties wrote in a letter to the Honorable Robert C. Byrd, Ranking Member, Senate Appropriations Committee "...we understand the Census Bureau is planning to count at least 90 percent of the individuals residing in all local areas and then use their statistical sampling technique for the remaining non-responding households. We believe the sampling technique is vital to ensure the most accurate possible Census count, especially in "hard-to-enumerate" areas, rural and urban, where access to households is difficult. We believe the integrity of the Census is important for trust and respect of the government and for corporate and governmental services to all of our constituents."
In early May, Congress revisited the issue when the Senate Appropriations Committee passed the disaster relief spending bill with a provision prohibiting the Census Bureau from preparing to use sampling in the 2000 Census. An amendment offered by Senator Ernest Hollings (D-SC) to strike that language was defeated by a 13-15 vote.
Also in early May, the Consortium of Social Science Associations (COSSA), wrote a letter to the U.S. Senate stating, "Prohibiting the use of sampling will also prevent the Census Bureau from correcting millions of errors in the count. In 1990 ten million people were missed and six million people were counted twice. The Census undercount is not just an urban issue. One-third of those missed in 1990 lived in rural areas, most of them poor and white (emphasis in original)."
Later in May, the Senate agreed to compromise language in an emergency spending bill that allowed the Census Bureau to proceed with its plans to use sampling in the 2000 Census. The amendment offered by Senator Ernest Hollings (D-SC), and accepted by Appropriations Committee Chair Ted Stevens (R-AK), as well as Subcommittee Chairman Judd Gregg (R-NH), prohibited the Census Bureau from making any "irreversible" decisions regarding the use of sampling that would affect the final population numbers used for apportionment.
In June, House and Senate conferees on the disaster relief bill approved a provision that banned all sampling and statistical procedures that affect the Census population counts used to apportion the House of Representatives. In addition to prohibiting the expenditure of funds on these techniques in any fiscal year, the new language amended the Census Bureau's authorizing law to ban current and future use of such methods. The provision essentially barred the use of any sampling or statistical procedure to take the Census except for the administration of the Census long form to a sample of households.
In June, a report from the National Academy of Sciences finds, "Changing, updating, and adapting the Census methods is a proven and desirable course of action. Change is not the enemy of an accurate and useful Census; rather, not changing methods as the United States changes would inevitably result in a seriously degraded Census."
President Clinton vetoed the disaster relief bill because of the prohibition on scientific sampling methods in the Census. After several weeks, Congress removed that provision and substituted the language approved originally by the Senate. It also required a report by the Census Bureau detailing its use of scientific methods in the Census. The President signed the bill into law in mid-June.
In July, Congress revisited the issue when the House Appropriations Committee passed the FY98 Commerce Department spending bill with a provision banning the use of sampling for counting the population in the Census. An amendment was offered by Rep. Mollohan (D-WV) to remove the sampling restriction, however, the amendment was defeated by a vote of 33-25, with every Republican voting against the amendment and every Democrat voting in favor.
In July, the Senate gave final approval to its version of the FY98 Commerce Department spending bill. The Senate bill included $354.8 million for 2000 Census activities, the amount requested by the President, and prohibited the Bureau from making "irreversible" plans to use sampling in the Census. The language essentially permitted the Bureau to continue evaluating the current Census design, which includes a limited use of scientific sampling to supplement traditional counting methods in the 1998 Census Dress Rehearsal.
In the House, the FY98 Commerce Department spending bill reported out of the appropriations committee prevented any funds from being used to implement statistical sampling. An attempt was made on the House floor to strip the anti-sampling language from the bill, however, that attempt failed (197-228) on a mostly party-line vote on September 30, 1997 (Roll Call Vote #475).
In November, Congress passed the FY'98 appropriations bill (H.R. 2267) allocating $390 million to the Census Bureau for 2000 Census activities in FY98. The President agreed to sign H.R. 2267 only after a compromise regarding the Census was worked out. The agreement called for expedited review of the constitutionality of Census sampling and created a Census monitoring board. Also, the compromise required the Census Bureau to prepare for two kinds of Censuses: one that includes scientific methods and one that does not. The Census Bureau agreed to test both designs in the 1998 Census Dress Rehearsal.
In December, Rep. Dan Miller (R-FL) was named Chair of a new subcommittee of the House Committee on Government Reform and Oversight that will oversee the decennial Census.
In January, Census Bureau Director Martha "Marty" Farnsworth Riche announced her resignation (effective January 30, 1998), citing personal reasons for her decision not to serve through the remainder of the second Clinton Administration.
In late January, the Atlanta-based Census Bureau Regional Director James Holmes was named Acting Director of the Census Bureau.
In February, the Southeastern Legal Foundation (SLF), a conservative nonprofit organization, and Representative Bob Barr (R-GA) filed a lawsuit in the Federal District Court for the Eastern District of Virginia. (Barr lawsuit) The suit (Glavin v. Clinton, No. 98-207-A) challenges the use of statistical methods when taking the Census.
Also in February, Speaker of the House of Representatives Newt Gingrich (R-GA) filed a lawsuit in the Federal District Court for the District of Columbia, challenging the use of statistical methods when taking the Census U.S. House of Representatives v. U.S. Department of Commerce, No. 98-CV-456 (Gingrich lawsuit).
In April, a number of interested parties moved to intervene (ask the court if they could be parties to the lawsuit) in both the Gingrich and Barr lawsuits. These parties, which include major cities, localities and civil rights organizations, believe that they have an interest in the outcome and that their position was not fully represented by the original parties to the two lawsuits.
In June, a three-judge panel in the District of Columbia District Court heard all pending motions in the Gingrich lawsuit. The judges are Royce Lamberth and Ricardo Urbana of the U.S. District Court in Washington, D.C. and Douglas Ginsburg of the D.C. Circuit Court of Appeals.
In late June, President Clinton nominated Kenneth Prewitt to serve as director of the U.S. Census Bureau.
In July, Congress revisited the issue when the House and Senate crafted FY'99 spending bills for the Census Bureau (included in the FY99 Commerce Department spending bill). Under the House version, only half of the $952 million allocated for the 2000 Census would be available for the Census Bureau to spend through March 31, 1999. The remaining $476 million cannot be spent until the President, by March 15, 1999, formally requests the funds and gives a cost estimate for completion of the Census. Rep. Alan Mollohan (D-WV) offered an amendment to remove the restrictions on the allocation, however, that attempt failed (201-227) on a mostly party-line vote on August 5, 1998 (Roll Call Vote #388). The Senate bill provides $848 million for the 2000 Census, the amount requested by the President.
The Clinton Administration has said it will veto the appropriations bill if it includes the restrictive language in the House bill. It is likely that the FY99 Commerce Department spending bill will be rolled into a continuing resolution at the end of the 105th Congress which is scheduled to adjourn in early October.
In October, the Clinton Administration said it would veto the appropriations bill if it included the restrictive language in the House bill. As a result, the FY99 Commerce Department spending bill was rolled into the omnibus spending bill that Congress passed and the President signed into law at the end of the 105th Congress.
The omnibus bill allocated $1.027 billion for 2000 census activities for fiscal year 1999 (FY'99), $179 million more than the Bureau's request of $848 million. While appropriation bills usually are funded through the end of the Federal government's calendar year (which ends September 31st), the FY'99 CSJ bill is funded only through June 15, 1999. If Congress and the President have not reached an agreement on census methods and funding levels by that date, funding would stop not only for census activities but for all the activities funded under the Commerce, State, Justice appropriations bill.
In late November, the U.S. Supreme Court heard oral arguments in United States Department of Commerce, et. al. v. United States House of Representatives 98-404 which was consolidated with William J. Clinton v. Michael J. Glavin, et. al. 98-564. At issue was the legality and constitutionality of the Census Bureau's plan to use scientific sampling procedures to supplement the initial enumeration for purposes of apportioning representatives among the states as part of the upcoming census.
In late January, the Supreme Court ruled 5-4 that the federal Census Act of 1976 prevents the Census Bureau from using statistical sampling to determine the population for purposes of congressional apportionment. The decision does not prevent, and suggested that the Census Bureau is required, if feasible, to use statistical sampling to adjust the undercount for non apportionment purposes, such as legislative redistricting and the distribution of billions of dollars in federal monies.
In response to the decision, the Census Bureau indicated it would likely employ the old, out-dated methods for apportionment purposes and combine the old methods with new, modern scientific sampling methods to produce a more accurate count for non-apportionment purposes. Opponents of modern sampling methods in Congress indicated they would seek to prevent the Census Bureau from using these methods in the 2000 census.