Proceedings, First Workshop 0 Panel to Review the 2000 Census Committeeon NationalStatistics Divisionof Behavioral and Social Sciences and Education NationalResearchCouncil NATIONAL ACADEMY PRESS Washington, D.C. NOTICE: The project thatis the subject of thisreport was approved by the Governing Board of the NationalResearchCouncil, whose membersare drawn from thecouncils of the National Academy of Sciences, theNational Academy of Engineering,and the Instituteof Medicine. The membersof the committeeresponsiblefor thereport were chosen for theirspecial competences and with regard for appropriate balance. The project thatis the subject of thisreport was supported by contract no. 50-YABC-8-66010 betweentheNationalAcademy of Sciences and theU.S. Census Bureau. Any opinions, findings, conclusions, or recommendationsexpressedinthispublication are thoseof theauthor(s) and do not necessarilyreflect theviews of theorganizations or agenciesthat provided support for the project. Additional copies of this report are availablefrom Committeeon National Statistics,2101 ConstitutionAvenue, NW, Washington,D.C. 20418; (202)334-3096;Internet, http://www.nap.edu SuggestedCitation: National ResearchCouncil (2001). Proceedings, First Workshop. Panelto Review the 2000Census. Committeeon National Statistics. Washington,DC: National Academy Press. Printedin the UnitedStates of America (cid:1)c2001 by the NationalAcademy of Sciences. All rightsreserved. ii TheNationalAcademyofSciencesisaprivate,nonprofit,self-perpetuatingsocietyofdistinguished scholarsengagedinscientificandengineeringresearch,dedicatedtothefurtheranceofscienceand technologyandtotheiruseforthegeneralwelfare. Upontheauthorityofthechartergrantedtoitbythe Congressin1863,theAcademyhasamandatethatrequiresittoadvisethefederalgovernmenton scientificandtechnicalmatters. Dr. BruceM.AlbertsispresidentoftheNationalAcademyofSciences. TheNationalAcademyofEngineeringwasestablishedin1964,underthecharteroftheNational AcademyofSciences,asaparallelorganizationofoutstandingengineers. Itisautonomousinits administrationandintheselectionofitsmembers,sharingwiththeNationalAcademyofSciencesthe responsibilityforadvisingthefederalgovernment. TheNationalAcademyofEngineeringalsosponsors engineeringprogramsaimedatmeetingnationalneeds,encourageseducationandresearch,and recognizesthesuperiorachievementsofengineers. Dr. Wm. A.Wulfispresident oftheNational AcademyofEngineering. TheInstituteofMedicinewasestablishedin1970bytheNationalAcademyofSciencestosecurethe servicesofeminent membersofappropriate professionsintheexaminationofpolicymatterspertaining tothehealthofthepublic. TheInstituteactsundertheresponsibilitygiventotheNationalAcademyof Sciencesbyitscongressionalchartertobeanadvisertothefederalgovernmentand,uponitsown initiative,toidentifyissuesofmedicalcare,research,andeducation. Dr. KennethI.Shineispresidentof theInstituteofMedicine. TheNationalResearchCouncilwasorganizedbytheNationalAcademyofSciencesin1916toassociate thebroadcommunityofscienceandtechnologywiththeAcademy’spurposesoffurtheringknowledge andadvisingthefederalgovernment. Functioninginaccordancewithgeneralpoliciesdeterminedbythe Academy,theCouncilhasbecometheprincipaloperatingagencyofboththeNationalAcademyof SciencesandtheNationalAcademyofEngineeringinproviding servicestothegovernment, thepublic, andthescientificandengineeringcommunities. TheCouncilisadministeredjointlybybothAcademies andtheInstituteofMedicine. Dr. BruceM.AlbertsandDr. Wm. A.Wulfarechairmanandvice chairman,respectively,oftheNationalResearchCouncil. iii PANEL TO REVIEW THE 2000 CENSUS JANET L. NORWOOD (Chair), Chevy Chase, Maryland ROBERT M. BELL, AT&T Labs–Research,Florham Park, NewJersey NORMAN M. BRADBURN, NationalScience Foundation, Arlington, Virginia LAWRENCE D. BROWN, Departmentof Statistics,The Wharton School, Universityof Pennsylvania WILLIAM F. EDDY, Departmentof Statistics,Carnegie Mellon University ROBERT M. HAUSER, Departmentof Sociology, Universityof Wisconsin ∗ RODERICK J.A. LITTLE, School of Public Health,Universityof Michigan INGRAM OLKIN, Departmentof Statisticsand School of Education, Stanford University D. BRUCE PETRIE, Canadian Institutefor HealthInformation, Ottawa, Ontario CONSTANCE F. CITRO, Study Director MICHAEL L. COHEN, Senior Program Officer DANIEL L. CORK, Program Officer AGNES GASKIN, Senior Project Assistant MARISA GERSTEIN, Research Assistant MICHELE VER PLOEG, Program Officer ∗∗ ANDREW A. WHITE, Senior Program Officer MEYER ZITTER, Consultant ∗ Served until March 2000 ∗∗ Served as study director until March 2000 v COMMITTEE ON NATIONAL STATISTICS 2001–2002 JOHN E. ROLPH (Chair), MarshallSchool of Business,Universityof Southern California JOSEPH G. ALTONJI, Department of Economics, NorthwesternUniversity ROBERT M. BELL, AT&T Labs–Research,Florham Park, NewJersey LAWRENCE D. BROWN, Departmentof Statistics,The Wharton School, Universityof Pennsylvania ROBERT M. GROVES, Joint Program inSurvey Methodology, Universityof Michigan HERMANN HABERMANN, StatisticsDivision, UnitedNations, NewYork, NewYork JOEL L. HOROWITZ, Departmentof Economics, Universityof Iowa WILLIAM KALSBEEK, Survey ResearchUnit,Department of Biostatistics,Universityof North Carolina ARLEEN LEIBOWITZ, School of Public Policyand Social Research,Universityof California at Los Angeles RODERICK J.A. LITTLE, School of Public Health,Universityof Michigan THOMAS A. LOUIS, RAND, Arlington, Virginia DARYL PREGIBON, AT&T Labs–Research,Florham Park, NewJersey NORA CATE SCHAEFFER, Departmentof Sociology, Universityof Wisconsin-Madison MATTHEW D. SHAPIRO, Departmentof Economics, Universityof Michigan ANDREW A. WHITE, Director vi Preface The Panel to Review the 2000 Census was established in 1998 by the Committee on Na- tional Statisticsof the National ResearchCouncil at therequest of the U.S. Census Bureau. The panel hasa broad charge toprovide anindependentassessmentof thequalityof the2000 census operations and results. It is charged in particular to review the statistical methods of the 2000 census, including the use of the Accuracy and Coverage Evaluation (A.C.E.) Program and dual- systems estimation(DSE), along with other census procedures that may affect the completeness and quality of thedata. As part of its work, the panel held three open workshops on topics related to the A.C.E. and possible adjustment of the census counts for population coverage errors. The panel has prepared proceedings of each workshop to be part of the public documentation of planning for the 2000 census. The first workshop was held October 6, 1999. It considered issues of the A.C.E. design that had not yet been completely worked out by the Census Bureau staff. Topics discussed included methodsandissuesfordeterminingpost-strataforestimation,obtainingthefinalsampleofblock clustersfromalargerinitialsample,andimputingvaluesformissingresponsesoncharacteristics needed todefine post-strata. The second workshop was held February 2–3, 2000. It covered the dual-systems estimation process from beginning to end. The third workshop was held October 2, 2000. It laid out the process the Census Bureauplanned to follow inorder toreach a decisionby March 1 on whether to adjust thecensus counts for purposes of congressional redistricting. This volume provides the edited transcript of the proceedings of the first workshop. (See NationalResearchCouncil,2001band2001cforproceedingsofthesecondandthirdworkshops.) The substance of the participants’ remarks has not been altered. Some text has been added [in brackets] to spell out terms, clarify who is speaking, and make sentences complete. The only material deleted consists of introductions of speakers. A list of invited participants and their affiliations(at the timeof the workshop) is provided at the end of the document. Papers prepared by Census Bureau staff were distributed to participants before the meeting. They are often mentioned in the discussion, and appropriate citations have been added in the text to refer the reader to the relevant bibliographicentry in the references. All of the documents are availableon the Census Bureau’s web site: http://www.census.gov. For explanations of terms and discussion of the design of the A.C.E. as it was implemented in 2000, please consult the interim report of the panel, The 2000 Census: Interim Assessment (NationalResearch Council, 2001a). IwanttothanktheCensusBureaustaffwhopreparedmaterialsandparticipatedinthework- shop. Particular thanks are due to Howard Hogan, who did a superb job in presenting a complex set of materialsto the panel membersand invitedparticipants. vii viii 0 PREFACE I also want to thank the panel members and invited participants for their thoughtful com- mentsand probing questions during the workshop. Finally, I want to thank the panel staff who prepared this proceedings. They include Con- stance Citro, Maria Alejandro, Agnes Gaskin, and Daniel Cork. Janet L. Norwood, Chair Panel to Reviewthe 2000Census Contents Proceedings 1 Introduction 1 Update on Census 2000 1 Overview of A.C.E. 4 Definitionof Post-Stratification 10 Continued Discussionof Post-Stratification 32 RemainingIssues for A.C.E. Sample Design 48 A.C.E. EstimationIssues—MissingData 54 Comments from Panel Members and Guests 56 References 74 List of Invited Participants 76 ix