NEWS ALERT
September 22, 2000

Census Bureau Reports Final Mail-Back Rates; Court Hears Preliminary Arguments in Virginia Redistricting Case
Plus: Legislators Propose Fixed Term for Census Director; Congressional and Monitoring Board news; and more.

The final mail response rate for Census 2000 was 67 percent, two percent higher than in 1990, Commerce Department and Census Bureau officials reported at a Washington, DC press briefing on September 19. The revised figure reflects an additional three million households that mailed back a form after the April 18 cut-off date - more late forms than in any previous census. The Census Bureau reported a preliminary response rate of 65 percent after the mail-out/mail-back phase of the count ended.

Five states - California, Massachusetts, Nevada, Rhode Island, and Wyoming - increased their response rates by five percent over 1990, meeting the CensusBureau's "'90 Plus 5" challenge to state and local governments. Nearly 9,300 other governmental units also met or exceeded their challenge goal.

Thirteen of the nation's 15 largest cities and 14 of the 15 largest countiesequaled or beat their 1990 response rates.

The difference between the mail-back rates for the short and long censusforms was eleven percent. Census Bureau Director Kenneth Prewitt said census takers closed most of that gap during follow-ups visits to unresponsive households.

Secretary of Commerce Norman Y. Mineta praised career Census Bureau employees for their "tremendous service" to the country, noting that they work largely out of the public spotlight for most of each decade. He also highlighted the importance of census data as "fundamental to the operation of the economy." Commerce Under Secretary Robert J. Shapiro said the 140,000 Census 2000 'partner' organizations from every sector of society made this count "a model of civic engagement by the government." The Census Bureau has begun its campaign to thank organizations and local governments that participated in the partnership program.

Dr. Prewitt called the increase over the 1990 mail-back rate an "historical achievement" because "a decade long trend line in social behavior is hard to change... and reverse." Census 2000 is the first since the all-mail census started in 1970 to improve upon the response of the previous count. The mail response rate dropped from a high of 78 percent in 1970 to 75 percent in 1980 and 65 percent in 1990. The Census Bureau projected a 61 percent response rate for 2000.

The director also said the high response rates are not predictive of the count's accuracy. "You can have a quite good census and still not solve the undercount problem," Dr. Prewitt said, noting that the undercount affects two to three percent of the population and is disproportionately high for some population subgroups. He said the Accuracy and Coverage Evaluation (A.C.E. survey) would give the best measure of accuracy.

The mail response rate represents the percentage of known addresses in the universe of "mail-back" operations from which the Bureau collected a questionnaire or Be Counted form by mail, Internet, or over the telephone. Enumerators later determine that some of those addresses are vacant or nonexistent housing units. The Census Bureau also counts some housing units in rural or remote areas, including many American Indian reservations and Alaska Native villages, by sending enumerators to verify or determine the address and location, and to conduct an interview in person. Housing units counted through these "update/enumerate" or "list/enumerate" procedures are not in the mail-back universe. Next year, the Bureau will release the mail return rate, which represents the percentage of occupied housing units that responded by mail, Internet, or telephone. The return rate is a more precise indication of public participation in the census.

Response rates for all governmental units are available on the Census Bureau's web site at <http://rates.census.gov/>

Legislation to set a fixed term for Census Bureau director: The chairman and ranking Democrat on the House Subcommittee on the Census have made separate proposals to set a fixed term of office for the Census Bureau director.

A bill (H.R. 5257) sponsored by Chairman Dan Miller (R-FL) would establish a ten-year fixed term for the director, currently the only political appointee in the agency requiring Senate confirmation. Rep. Miller called the legislation "the first step in removing partisan politics from the census" and said a fixed term appointment would give the director more independence and more protection from partisan influence. The chairman, who has been at odds with the Census Bureau over the use of statistical sampling to correct for under- and overcounts, said in a press release that the Clinton Administration "has politicized the census and damaged the integrity of our national count." A ten-year term, similar to that of the FBI director, would allow a director "to supervise an entire census from planning to implementation," Rep. Miller said.

Rep. Carolyn Maloney (D-NY) is seeking cosponsors for legislation she plans to introduce that would establish a five-year term of office for the director. Rep. Maloney said her bill would "ensure that the Census Bureau continues to provide the most accurate, non-partisan data possible." By starting five-year terms in 2002, the congresswoman said, her proposal would "guarantee continuity during each decennial census." President George Bush's census director, Barbara Everitt Bryant, did not take office until December 1989, only months before the start of a census she did not help plan.

Traditionally, Census Bureau directors leave office when the President who appointed them does, if not before. The Bureau's director is the only head of a major statistical agency in the United States who serves at the pleasure of the president, instead of a fixed term that usually carries over from one administration to another.

Congressional hearing postponed: The House Subcommittee on the Census has postponed its hearing, originally scheduled for September 26, to review a proposed Commerce Department rule giving the Census Bureau director final say over the decision to release statistically-corrected census numbers next spring. The panel has not announced a new date for the hearing.

State legislative activities update: The U.S. Department of Justice told a federal court yesterday it should dismiss or delay consideration of a lawsuit filed last April by the Commonwealth of Virginia, seeking approval for a new state law that prohibits the use of adjusted census numbers for redistricting.

The court scheduled the hour-long hearing to consider the Justice Department's argument that it cannot determine if the Virginia law adversely affects the voting rights of racial minorities until detailed census data are released next March. Virginia is one of 16 states covered by section 5 of the 1965 Voting Rights Act, which requires Justice Department approval for any changes to election law because of past discrimination in election practices. The state's Attorney General decided to by-pass the "pre-clearance" procedure and sought direct approval for the law in court.

Virginia's Deputy Attorney General, Frank Ferguson, urged the judges to allow the case to proceed, saying a delay might prevent the state from redrawing its legislative district boundaries in time for the 2001 election cycle. Virginia's lawsuit contends that using statistically corrected census data for redistricting would violate the Census Act, which the U.S. Supreme Court said (in a January 1999 ruling) prohibits the use of sampling to derive state population totals used to apportion congressional districts among the 50 states. The state also contends the U.S. Constitution bars the use of statistical sampling in tabulating population counts used for apportionment and redistricting.

A three-judge panel of the U.S. District Court for the District of Columbia is considering the case, Commonwealth of Virginia v. Reno, et al (Civil Action No. 1:00CV00751). Either party to the lawsuit can appeal the lower court's ruling directly to the U.S. Supreme Court. A group of 15 cities and counties, led by the City of Los Angeles, recently joined the lawsuit in opposition to the Virginia law.

Congressional Monitoring Board report due: The eight-member Census Monitoring Board is scheduled to release its next report the first week of October. A primary focus of the report will likely be the panel's visits to dozens of Local Census Offices over the past few months, to observe and evaluate census field operations.

The Board, created in late 1997, has four members appointed by the President and four appointed by Republican congressional leaders. It will operate through September 2001. For information on the Board's activities and copies of reports, visit the web site for the Presidential members at www.cmbp.gov and the Congressional members at www.cmbc.gov.