[CTPP] BREAKING NEWS ALERT

January 25, 1999 SUPREME COURT DECISION ON CENSUS IS ISSUED TODAY The Supreme Court today ruled 5-4 against the use of statistical sampling for narrow purposes of reapportionment, the distribution of seats in the U.S. House of Representatives among the states. But the Court noted that the Census Act as amended "now requires the Secretary to use statistical sampling in assembling the myriad demographic data that are collected in connection with the decennial census." (p. 21) The full text of the Court's ruling in Department of Commerce v. House of Representatives can be found on the World Wide Web at http://supct.law.cornell.edu/supct/html/98-404.ZS.html Further information and reactions will be released as they become available with a full News Alert later today. Census Bureau Releases Dress Rehearsal Results Plan for Census without Sampling Sent to Congress The Census Bureau released the first set of population counts from its Census Dress Rehearsal conducted last year in Sacramento, CA; Columbia, SC, and surrounding areas; and the Menominee American Indian Reservation, WI. Following the schedule required by law for the decennial census, the Bureau published totals for each of the sites within nine months of the dress rehearsal Census Day in late April. More detailed population numbers and related socio-economic information will be available later this year. The population of Sacramento, where the Bureau executed its original Census 2000 plan that includes sampling and statistical methods, was 403,313. The Bureau also reported that without the use of sampling methods, the count would have been 349,197 (a difference of 54,116). The quality check survey (called a post enumeration survey, or PES) measured a net undercount of 6.3 percent, or 25,572 people, which was corrected with scientific techniques to reach the final count. Other uses of sampling and statistical estimation added 28,544 (7.1 percent of the city’s total population) to the count, as well. The Bureau’s 1998 funding bill included a provision sought by sampling opponents that requires the agency to publish at least two sets of census counts, produced with and without the use of statistical methods. The Bureau also implemented its original census plan in Menominee County, WI, home to an American Indian reservation of the same name. The site-level count was 4,738, 3.0 percent (or 143 people) more than the 4,595 people counted without using statistical methods. The Bureau plans to visit all unresponsive households on Indian reservations, instead of conducting follow-up visits on a sample basis, but it will conduct the quality check survey in these areas in order to adjust for undercounts. According to the PES conducted in 1990, that census missed more than 12 percent of American Indians living on reservations. In the Columbia, SC, area, the Bureau conducted the dress rehearsal using only traditional census methods and reported a population total of 662,140. The Bureau administered a post-census survey to measure accuracy but has not yet reported the results of that quality check. More information about the dress rehearsal is available on the Bureau’s web site, at http://www.census.gov/Press-Release/www/date.html. "Traditional" census plan released: The Census Bureau has sent to Congress a plan for the 2000 census that does not include sampling and statistical methods to finish the count of unresponsive households or to correct undercount and overcounts measured by a post-census survey. The plan is based on four broad strategies, three of which mirror the strategies developed for the original census design that includes some sampling: building partnerships, simplifying the census process, and using state-of-the-art technology to improve efficiency and accuracy. The fourth strategy – using "special techniques to improve coverage" – replaces "us[ing] statistical methods" in the original plan. The Bureau has been preparing to implement two census plans since the fall of 1998, when Congress tacked such a requirement onto the agency’s annual funding bill. In its "traditional" census plan, the Bureau says that improvements to the 1990 census design will "enhance the quality of Census 2000". However, it also concludes that, "consistent with the judgements of most other experts in this area, [a] census using statistical methods would be more accurate generally, succeed in reducing the differential undercount of minorities and children, and be more cost-effective than the alternative." The traditional plan calls for expanded operations associated with the direct counting effort, such as opening more local census offices (520 v. 476 under the original plan) and hiring more enumerators (not quantified in new plan). More enumerators will be needed to follow-up with all households, not just a sample, which do not respond by mail. Enumerators also will visit all housing units identified as vacant or nonexistent by the Postal Service, instead of the sample follow-up originally planned. And instead of a quality check survey to provide the basis for correcting undercounts and overcounts using statistical methods, the Bureau will employ "coverage improvement" operations to try and increase accuracy. After the follow-up phase is completed, census workers will double-check housing units that were listed as vacant by enumerators on a previous visit, but which did not have an undeliverable postal address. They also will visit all households for which a questionnaire returned by mail was not filled out or was subsequently lost. The alternative plan includes a post enumeration survey to measure the accuracy of the counts, but those results will not be used to correct the official figures. The Bureau also has decided to accept census responses via the Internet, regardless of the final census design. This option will be available only for households receiving the short form (respondents must provide an identification number printed on the form they receive at home) and only in English. The Bureau noted that its "Census 2000 Operational Plan Using Traditional Census-Taking Methods" may change as a result of congressional funding decisions or evaluations of last year’s dress rehearsal. The original Census 2000 plan, unveiled in February 1996, has also been modified over time based on recommendations from the Bureau’s advisory committees and outside evaluators such as the Commerce Department’s Inspector General and National Academy of Sciences panels. Congressional field hearings continue: The House Committee on Government Reform will hold a field hearing in Phoenix, AZ, on January 29, to examine "community based approaches for a better enumeration." A primary focus will be efforts to improve the count of American Indians living on reservations. According to the post-enumeration survey conducted as part of the 1990 census, the undercount for this population was over 12 percent. The hearing will be held in the Phoenix City Council Chambers, 200 West Jefferson St., starting at 2:00 p.m. The panel’s census subcommittee held its first field hearing in Miami last December. (The committee has not formally organized its subcommittees for the 106th Congress.) Congressional committee assignments: The House and Senate continue to fill seats on committees and subcommittees that oversee and fund the census. Because of space considerations, we will provide that information in a separate News Alert later this week. Questions about the information contained in this News Alert may be directed to TerriAnn Lowenthal at (202) 484-2270 or, by e-mail at <terriann2k(a)aol.com>om>. Please direct all requests to receive News Alerts, and all changes in address/phone/fax/e-mail, to Census 2000 at <Census2000(a)ccmc.org> or 202/326-8700. Please feel free to circulate this information to colleagues and other interested individuals.