WPCr 3W\`z:պ əDMZM^ 1bbM 0U 0B 0w 1 0! 0 1j 0b 0Di A X!b)" 01" 0" 0# 0J$ 0$% 0D% 0K%7& 1u&(r' 0'f5(a7( 72K( 72}( 0( 0n) 09* 0* 0+ 0.,2,@J; 0E< 0< 0i= 0= 0>t? 0? 04@ 0 N@ 0&A 0A\B 0e"CUNC 0C 0DU6EU>E 0EMF 1F BG 0n.G D-G D5G 0G AH D5I B!I D5I BJ D5;J BpJ D5J BJ D1J D5K B*EK D5oK B*K CK CK CL CL C6L CPL CjL CL CL CL CL CL 0M BN 0N 0N 0O 0sP 0bQ 72QMRfRRM4Ro8Ra7M7f7 0D7M(9(9,9MB9?F9a0:0:rD:M: 0C:: 0D; 0D< AQK< Ak<M= 1 =P=f==$>M>>>M>M>M> 72?Us@4?Mt? 1x? 0N?MM@ 1Q@ 1@M_A 1cA 1AMqB 1uB 1B 1CM D 1DD 1DS4E 1E 1F 1F4GMJG"NGNGpG"GGMG 1G 1IH 1HH 0PWIIIIII"I"!J!JaCJWJmJmJ"JJJJJJJKK)Kp?KLLL B -LLLnMvNNNNNNNN"NOO2O2OHOU@PQ2Q!STUUUU@WWf'Xa)XU,=X=XfiXakXfXaXXXfSZUZa[["2[lT\^abceiiUNii3jU@Dl@llanananEpEp[p[p[ppppppppppppppppppppppppppXqZ]qPq\\CENSUS\PRED-TR2-2-HP9000-1A.PRED.DECENNIAL.HQ.CENSUSdttddtdt0(hH  Z6Times New Roman RegularX($/ s.:i+003|x7U(!:!*$! !2Ɠ   .3  0  ! !(xir$0  0` (#(#0 ` (#` (#2#(  0  )3  0 (# (#(9 Z 6Times New Roman Regular3#37=CIQYag1.a.i.(1)(a)(i)1)a)  wxyzpo"Gaeimquy}Bullet ListBullets List EEGMSY_ekqAutoList631.1.1.1.1.1.1.1.(EGMSY_ekqAutoList681.1.1.1.1.1.1.1.  wxyzxYIcgkosw{Large BulletLarge Bullet\  `&Times New Roman   (EGMSY_ekqAutoList691.1.1.1.1.1.1.1.*+ (_2623  ..*D+D (_25   ," <DL,23  ..," <DL,   (O;$0  2#  a  .3  0` (#(#  wxyzxCcgkosw{TrianglesTriangle Bullet(b$0  0` (#(#2#   .3  0 ` (#` (#*5+5 (_24  ) <DL)23  ..) <DL)    wxyzh<EGKOSW[_cAutoList14)1)a)   (2ـ$ !    0  ڀ(#$  0  1/0X&OLE 2.0 Box <=8C HKKKK $ـ$  18  ڀ  _Shortformdataitemsarethecensusdataitemsthatweretobecollectedforallpersonsand  households.- -*2+2 (_23 ` &<<DL&23  ..&<<DL& ` */+/ (_22  #DL#23  ..#DL#  *,+, (_21   DL 23  .. DL  *)+) (_20 h DDL23  ..DDL h *&+& (_19  L23  ..L     (3$ !  ( $ Figure  1  Level 1Level 2Level 3Level 4Level 57#137=CIQYag1.a.i.(1)(a)(i)1)a)*#+# (_18   L23  .. L  Table_Ai)2i)*>> (_17  2( 4 <DL223  Ԁ2( 4 <DL2  *DD (_16   ," <DL,23  Ԁ," <DL,   *55 (_15  ) <DL)23  Ԁ) <DL)  *22 (_14 ` &<<DL&23  Ԁ&<<DL& ` *// (_13  #DL#23  Ԁ#DL#  *,, (_12   DL 23  Ԁ DL   BurchamCanvasssing LUCAlistersgeocodingungeocodedgeocoded MAFlisterCavassing BSAsPrecanvass DSF USPS DSFsgeocodeunmerging DMAFgeocodesmapspot DeliverabilityWindham ACF D.Cresidentialscorrrections QIP TEAsEurichHaltermanKillion TXERuhnke GEOVitranoConnectiutmailbackmiscodedReults http(www.geo.census.gov gasbul0total_colblkultea.htmlnonBlackovercountA.C.Ereinterviewsubsampling NPC Jeffersonville&misclassificationnonmoverssubnationalrecomputed$postratificationnonresponseinmovers CAPICantwell HUCSvacantsmisclassified ImelFeldpauschEnumertaionBeaghenNonmatchesWetroganCresce DSSD EESC MES ICMpoststratumurbanicityBiemernonminoritiesunweightedMaRCSCitroredisticting No.BincreasaesBelinMulry M.H. STSD GQs TESsubsampledovercounts poststratifingMcGrathZelenakAcurracyreweightedFeldpouschCensal E.A. PFUcoeffeicients unti CVsStarsinicSisselAsiala piiiFeldpaush pviiunenumerated pivunduplicationmultiunitsunduplicatestreetname ppiv GburSprattFarberIncensalSchaferovercounted HUCF HCUF D.R.multiunit NRCOnoninterviews LCO LCOsnonhouseholdRelistingKresjaKostanichinpercentRelisingrelistedmulitPetroniESCAPRaglin DSEssubsampleKrejsa DSEPostcensalnoninterview"poststratifyingpoststrataKeathleyNonmatchpercentserrneous&HispanicscomparedPostratifyinmoveroutmoverssubselectionhousemates NRFUnonresponding nonrespondentsZuWallackSalganikCromarSekarDemingWolterstratgepostratumnonsamplingundercoverage p.vsocio EFUA.C.E.ppiiinonmovercorrectormailout PHCWhitfordH:\MURAWSKI\LATESTMO\1CPOSTAG.TIF*)) (_11 h DDL23  ԀDDL h *&& (_10  L23  ԀL  (## &_9   L23  Ԁ L  (>> &_8  2( 4 <DL223  2( 4 <DL2  (DD &_7   ," <DL,23  ," <DL,      (55 &_6  ) <DL)23  ) <DL)  (22 &_5 ` &<<DL&23  &<<DL& ` (O$(// &_4  #DL#23  #DL#  (,, &_3   DL 23   DL  1, 2, 3,Level 1Level 2Level 3Level 4Level 54#;2Quick 1.L  .0 L(9 Z6Times New Roman Regular8ow6Plain Text  KQ<6X9`(Courier NewKXXXSR\  `&Times New RomanS  ()) &_2 h DDL23  DDL h <6X9`(Courier New\  `&Times New Roman(&& &_1  L23  L    wxyzh fEGKOSW[_cAutoList79"""""""")1)a) d<:Default ParaXXXXXX)!dxdxVVVV)!dxdx0;.header   !XXX   !1/0X I&mage <=8C HKKKK]]]])!dxdx  d____)!dxdxP Pdaaaa)!dxdx dcccc)!dxdx d__)! dxdxffff)!dxdx&0 d dhhhh)!dxdx&P d dd ddd(d2d>>???@@@AAABBBCCCDDDEEEFFFGGGHHHIIIJJJKKKLLLMMMNNNOOOPPPQQQRRRSSSTTTUUUVVVWWWXXXYYYZZZ[[[\\\]]]^^^___```aaabbbcccdddeeefffggghhhiiijjjkkklllmmmnnnooopppqqqrrrssstttuuuvvvwwwxxxyyyzzz{{{|||}}}~~~o}sϏO/??oL߯O? ?O??O???̅{3OD_/_3߬.7_/o.ߓ3o/_2;)ߜ?A?>2_?_O/+.? ?5_/o+_?1oo//_7) O:+?AO=?0/=?$o-+?._ O??4+0?/"4?)$_8+o@o=??O/;?'?.+?. ?3_,/?/$3?O*&_6+@=?OO. /? ?: O_/?.?*?. 2_ _o/ ,.o Oo? o0&_2O*/&6+@<O- O_?8_  _ .O*?.o1_? ,??- O 0&_1*O'5*? ?<￟? ߿_. ?7/ ??0Ͽ? ߂0o __2/ _-߿/ ?Ͽ. 2 ￿O 1￟? ߿_.Ͽ? OO 6 ς?_ <￟? ߿_ ??0??6/ ?2 o3 /3/O0 0o4? _ 3 ??2  7? ?> o? ?? ?O/__5/ /2 3 3?0 /0o/4? ? ?3 ?O2 / _7? ?> ? ?_ _0O5O _?2 3 20 ?/?4 _ 2 _2 / 6 > > O /4 _?2 3 ?3?/?0 ??/_o4 ? _2 3  _6 = O>  /?/3 ?2 ￏO?3? 2_1 ??/o4  2 3  5 ??=_>  .?3/ 2 Q? ?21 ?./4 _ 2 3 ? 5 Z>  ?.2 ?/2 Q????2?߿1 o.o￯4  2 ?3  _5 Z> ??.oM o$3 Qo??2?M .?Q  O2?4  /5 Z/?? =???.?N [ Q??1?N .?Q /?2??4 ? 5 Y O=????.?M \ QO??1?N .Q ??2??4 O 5 Y_ <????.?L \ Q/ ??1/N . Q ??2??4  4 X__ <????.?L/ o] Q? ??1 M . P ?O2??4  4 X <????.?/K ] Q?? o??1 M . oP ?2??4  4 X?// ;????.O J O^ Q? ??2 M .O 2??4 o 5 W O;????. ?J/ ^ Q? ??2 L .? oO  2??4 ? 5 Wo :????. H ^ Q?/ ??2_ K . N O3??4 /?5 W__ :????/ G ?_ Q? ??2 J . M _3??4  o5 V :????/ Fo _ Q? /??3 I /? L /4??4  5 V? 9????/ _D/ _ Q? ??3 H / K o4??4 _6 U _9????0 C ` Q? ??4/ OF 0O ?I o5??4 _6 U 8????1/ o@ ?` Q?o ???4 E 0 H /6??4  6 UO _ 8????1 ? _` Q? ??5?C 1F //6??4 _7 T  8????2= ` Q? ??6__A 2?D /7??4 O7 TO  7????3?; ` Q?O O??7_? 3oB __8??4 _ 8 So  _7????49O ` 8? ??8o> 4A ߿/:??4 _ _9 :  6????5/8? a!8?  ??9?< 5/?_;??4 / ?:!:? O 6????6_?6? a!8? ? o ??:/: 6?_=_=??4 ? O:!9  /6????85? a!8?  ??;o9 8<????4 ? o<!9_ 5????9_3? a!8?   ??=/O7 9??:o<??4=!9o _5????;/2? a!8? /  ???6 ;9:??4o>!9 4????=?0? a ߃9?  ??@_5 <_7 O/8??4o@ σ9/O 4????>/? a Q?  / ??B?_3 >O/6 o O6??4B P /4????@O.? a Q?  ??D/2 @?/5 _ 5??4?B P_ 3????B/- a Q?  ??F_2 B/4 ? 4??4  A O_o _3????D /, ` Q?o ???G_1 C_4  3??4 o @ O 2????Eo o, ` Q? ??I/ 0 E? 3 / 2??4  ?@ O/? 2????G + ` Q? ??K ?0 G 3 / O2??4  ? N ?2????H/ _+ O` Q?O O??L / H 2  1??4 ? > No 1????I_ +` Q? ??M O/ I/ 2  1??4  ?> N(o1????J/ +O _ Q? ??N . J 1 O 0??4 o = M)1????K O+ _ Q?? o??O . K 1  _0??4  < M?*0????L + O_ Q? ??P? o. L_ ?1 /0??4  O< L+?0????M + ^ Q? ??P ?. L 1  /??4 ? ; Lo,/??OM +_ ^ Q?/ ??Q./MO 1  /O4  : L?? o/_oM + /^ Q? ??Qo .?M 1  /4 _ O: K /o?N , ] Q? /??Q .OOM 1  /?4  9 KO? .??N, ?] Q? ??Q .o?O 1  /?4  8 Jo ?./ ON?, \ Q? ??Q.?O 1  / O4 / O8 J - OO?- \ Q?o ???Q?.OO1  05  7 J/o ./OOO. [ Q??Q_/N?1 / 0OO5 _ 6 I .OO_._?/ Z Po/?1o/Oo.ooo1  0_5  O6 I ? -OO.0/ X 1?O?1?O///??o1  /1O4  5 )_! ?- O/O/_0 /W _?1??1?/0/?2  2/O5 / 4 O?(?" -/O_/_?2 /#1 1??11_?/?2 O 2_5  /4 '# ?-_O?0?_3_ / ?1 1? ?12 o02  _4O?6 _ 3 '?o# -O? 0?o?4/ 1 1 ?1O3_ 0o3o  4 6  2o ?&$O O- ? o1?_5/ o1 o1 ?0_?413? O 6? o7  1? ?&?/% -? O _1?7/ 1 /1/ ?0?_5 oO24?  _ ?7 _7 / O1? %& _- _  ?2 O _9/ ߏ/-Ͽo ? O-_? ?0 ? 7 ?/2 ? ? 1߿ Ͽ_ /8 ?4￟/ o ?(ϯ_ ?/ "/_߿/ o+ ?  ? 3 o_ _;_/oￏ_ -?,?,o ?0 /? 9_ ?ߏ2 O 2(/: ? 5?o_(-"O o+ ?  OO/ 4?? >?/-/,?,_ ?0/_o? ; /_O 3Oo? 3'o= OO/ 6?O(-"? ?, O_/  /5!@O$/-?,? ?0!O=?4!/3%?/7/?'-" /-/ o6oC? 0-?,__?0@o65#?Co7 '-O"/o _9ϏO_G__2-,??1o?C_o9?6/_F_9'-?!o0__/?ϟo?/Lo/?5﮿,ϐ?6O/_HOO?_/O9ߝN_/;'߭!ߒ3_/ϟO?/?_G߿??/?ORo????_Ƈ;o???_OϏ????oHϯo???OϟO?/?_lϟO?/?_AJPO?[?_o/Oo?P/[_O Po[??/   $.O]  TO)F 0ooo/?_No%?c o  O&/-] _ TO+G O/?/L??$c  &/.] OU*OG .?Jo$c /'o._oWO,I0Ooϯ/H$c O(.O/`/_YO-I?0/ _G$Od ?(.a_o o?Z?-J0??O F%d ?(o /Z-??JO0o/ /)??O [oo.oJ/0O_?)? /\?O/o??-J/0_O?o(O] /o-I_0o__O ?/o _?߃?/o//__?/o? ߏ?/ _?o?? /o/?oO?? ?ߏ?O?O o_ o?? ߂o?__O?_ _o//?/O//o?ߏ?__/Oo? O?o _?_/O _?o___/O/_o o /??_ ?  ?_/  ?? ??__    OO_O__?/ _ O  o? ? ?_?/ ? ??? ??? _?O_ O /_o_? O? _ o _ // o  /  O?_/ oO   ?/ ??/  / ? ???? oo/ oo/ o_? __/?/ _ߟ/_? ?_?oO   O_/ o/ o o/o/  o _ OOo?? //_ _?߯_/o _/o __? _ ? _o /  ?oO  o / o_ ?/? ? /o ? ? ?o_ //? __ _//? _oO /o_? /? /oo/??  ? ?O/ _ /  oO ? O/  O o/ _  _/ o  _/? o//_? _ o??oo_? _ / Oo/_? /?O? ? /_O //? ?/ O o /?     _ ? _ ?OO_o   ?// ?_o_ /o_   ?/߿/?/_  / O??/? O  / /o _ o_O/ O//  ?? O_   !_/ / _ Oo/o   _O ?/ ? _O _ o/ ??// O _/ O?/_   o    /_?? O߿? /_ _ o/ ?_? O o" O ? ?/_? /? O/ o O/  /?   //O/  // O? __ o /  ?ooO _  ??#/O _ ? /o  _ /? ?_ O_   ?O_Oo_ ?  / _ ? O ?_O/?  /?O /O  #  ??? /? ?   _o   oo O OoO_? /O? O_O/ _/ o _o ?_ _ # / ?o O?o?O _  ?/OOo_ ?/??  _ _?o?/??__  ??_? ?? ? Oo _O? ?/  // / #o O  ?o   O?O_O//  _  //__ ? OO_o/_/_  ?? /_?_/o _ oo  ?/ O$   ?O? _o ? ?/  /?  /  _// _/ o/ oo  _ ooo/  ?/ _  _  ?$   o O  ?_ OO   _? OO _?ߏ _/ O_ o _ _/o ?? _ _oO  _? /? ? /O o$? O/ ?/ O_O  o_??   o  o OO?o _?_/o  ?o?? / /? O  /o /o   o  O ?$? ? O__??/O  ?/O? _"_  ?O   _  /O   ??_ _ ?// o /o o ???o? ?    _ # o _?oo o O/O%/ ?  _  ?? _ ?_O O_??oO/  /  ?? o?O o /  ? o$  /o?O?_  /+/   /?  _ _ /? _O/o/?/  O  o ?? ? _ _ _/O?$o  _?oo_? ?O ?+_ O    O  _/? oO   _ o o _ //  /__O?o$/  / ?Oo_?//  ?+ O _   _ O  _o/O?? ?/?  _ //__o   _  ?_$? ? O O///_ ? o_  /O    _  /OOoO??//o??_ O_ / /? o/ / /  ??  o ?#_  o /__O ?_ O?/  o/   _ O  ? O  ?/_?/ ? oo_??o  ?  /?O O/  _?? o /  // ?O$ / o?/ ? ?/ __o O O/  o  /? o o O   O  o / _? ?/   O /O _ ??  ? _ ? ?O/O/$? _O/ o O O ?_/  _?? /  _ /  o O?? ?/ OO/?o? oo? o ?o _ O/    _oo/_$ / ? ߿  ?_/? OOOߏ/ O ? _ oO  ? oOO_?ߟ?o ?/?O/oO// _O / ?_ _ ??????O?   /!o O_ߟ/O  oO/oO? _o/??? o/Ϗo / _ oO? /o?o????/o??__oo ? ?OO/      ? ?  _!O / ? ?O ?_  o _  ?  /  o /?_ _??/_   ?  /?? / _   _ o_?/ Oo  ? oo O? _?  o ? /?/ ? O  ? ? //? o  O  ?  o  _   //_ ?O/?߄  O?/!ߊOO/oτ?o ?OO?_  _? o?/_ ߄  OO?o?oO _?O  //o ?? 4߿%_V _@aD\߿m߿`Ͽ*Ͽ)Ͽ@\bi\Ooj_\_Poj?\?Q /jo[oQ_? j[Qoko[oQ??l/[/Sn?Z?S_ /noZX߯oq???o?????/??_O??/?__????????_??o????o?/???/???????_/???o??__???????/OO?o??_????/o??o???/??O_????_????//??_???????__?/??OO??????OO??_o__??_o_????O???o߿/??__????/?????/_??/??O????_????o??/?__???O/????????o????o_????_????/O??o??o_????/??/????o??????߿???__O/??oϿ/O??????/???_O/oO??/????O????/?/_???_????/???_/???_o_????/?o???__??o????oo???/?/???/?????????/O???????/_O?OO?????o????_???O/???o????/O???O??o????O??o??_???????O????o_?/??O????????????o??O_?/??/OO??o_o/???o?_Oo??/??OO?O???????o??o?/O?????O/OO??o????O/o/OO??_?/OO_OO__?o/?oOoooo_O?OO/?O??O/??o/O_?O/O_/_????/?/O/O??/O__?/??_??O_/?_O??__/???o_O?_?oO??o?/?O_ooo?OO?o?_/oo?_??O?o???/?O_?///??_oO?_?_/O?__?O_/ߏ/Ͽo?????????????????O_????/??߿Ͽ_/?￟/o?ϯ_????????????????//_߿/o??o___/oￏ_??o?/?_?ߏO/??o_Oo?OO/??????//?_?/_o?/_OOo?oOO/?O??O_//O???O?///?//oo??__?o?oOo_ϏO___??o?_o?/__?o__/ϟo?/o/?￿Ͽ?O/_OO_/O߿_/￿߿￿߿_/ϟO?/?_߿??/?Oo????_o???_Ϗ????oϯo???OϟO?/?_ϟO?/?_O??????????_o/Oo?????/_Oo??/OOooo/?_o?oO/_OO?/??/OO?o/ooOOoϯ/OO//_O?/_O?_oo????O?o/??Oo//??Oooo/O_??/?O/o??/_O?oO/o_o__O????/o_????????/o//__?/o?ߏ?/_????o?????/o/?oO?????ߏ?O???Oo_o??o????__O???__o//?/O//o?ߏ?__/Oo????O?o_????_/O_???o___/O/_oo/??_??_/????__OO_O__?/_Oo???_?/???????_?O_O/_o_?O?_o_//o/O?_/oO?/??//?????oo/oo/o_?__/?/_ߟ/_??_?oOO_/o/oo/o/o_OOo??//__?߯_/o_/o__?_?_o/?oOo/o_?/??/o???o_//?___//?_oO/o_?/?/oo/????O/_/oO?O/Oo/__/o_/?o//_?_o??oo_?_/Oo/_?/?O??/_O//??/Oo/?_?_?OO_o?//?_o_/o_?/߿/?/_/O??/?O//o_o_O/O//??O__//_Oo/o_O?/?_O_o/??//O_/O?/_o/_??O߿?/__o/?_?OoO??/_?/?O/oO//?//O///O?__o/?ooO_??/O_?/o_/??_O_?O_Oo_?/_?O?_O/?/?O/O???/??_oooOOoO_?/O?O_O/_/o_o?__/?oO?o?O_?/OOo_?/??__?o?/??__??_????Oo_O??////oO?oO?O_O//_//__?OO_o/_/_??/_?_/o_oo?/O?O?_o??//?/_//_/o/oo_ooo/?/__?oO?_OO_?OO_?ߏ_/O_o__/o??__oO_?/??/Oo?O/?/O_Oo_??ooOO?o_?_/o?o??//?O/o/ooO???O__??/O?/O?__?O_/O??__?//o/oo???o??_o_?oooO/O/?_??_?_OO_??oO//??o?Oo/?o/o?O?_///?__/?_O/o/?/Oo???___/O?o_?oo_??O?_OO_/?oO_oo_///__O?o//?Oo_?//?O__O_o/O???/?_//__o_?_??OO///_?o_/O_/OOoO??//o??_O_//?o///??o?_o/__O?_O?/o/_O?O?/_?/?oo_??o?/?OO/_??o///?O/o?/??/__oOO/o/?ooOOo/_??/O/O_???_??O/O/?_O/oOO?_/_??/_/oO???/OO/?o?oo?o?o_O/_oo/_/?߿?_/?OOOߏ/O?_oO?oOO_?ߟ?o?/?O/oO//_O/?__????????O?/oO_ߟ/OoO/oO?_o/???o/Ϗo/_oO???/o???o????/o????__oo??OO/??_O/??O?_o_?/o/?__??/_?/??/__o_?/Oo?ooO?_?o?/?/?O??//?oO?o_//_?O/?O?/OO/o?????o?OO????__?o?/_OO?????o?oO_?O//o??߿__߿߿ϿϿϿo__o??/oo_?ooo?????//??_/o߯oTABLE B(F$0  0` (#(#0 ` (#` (#0 (# (#0h(#(#0h(#h(#0(#(#0p(#(#2#     )3  0p(#p(#    @@DH:\MURAWSKI\LATESTMO\1CPOSTAG.TIFH:\Murawski\Latest Models\1cpostag.tifT4b̤"ͤͤnJ' ']&I{s u&TABLE D h  \R3'Avery 5366 File Folder -\(2C$ !  ݃ Level 1 Level 2 Level 3 Level 4 Level 5($$   1  (3$ !  /XC<< c/]Inline Te&xt" <<C8;GMSY_ekqRomanRomanI.I.I.I.I.I.I.I.http://www.census.gov/dmd/www/ace2.htmlLevel 1Level 2Level 3Level 4Level 5  9       I.(9 Z(Times New Roman (ƓGIOU[agmsAutoList207I.I.I.I.I.I.I.I.>$"Small Circle"0    wxyzh2EGKOSW[_cAutoList26)1)a)  wxyzhBEGKOSW[_cAutoList27)1)a)  wxyz`REGKOSW[_cAutoList28)1)a)  wxyzhBEGKOSW[_cAutoList33)1)a)  wxyzhREGKOSW[_cAutoList34)1)a)  wxyzhrEGKOSW[_cAutoList36)1)a)  wxyzhEGKOSW[_cAutoList37)1)a)  wxyzhEGKOSW[_cAutoList38)1)a)   wxyzpEGKOSW[_cAutoList39"""""""")1)a) I    #"Xd# qCEKU[agmsAutoList31.1.1 1.1.1.1.1.1.(EGQ[eoyAutoList171.1 1.1 1.1 1.1 1.1 1.1 1.1 1.1 (UEGMSY_ekqAutoList181.1.1.1.1.1.1.1.  9                 wxyzh#EGKOSW[_cAutoList40)1)a)  wxyzh3EGKOSW[_cAutoList41)1)a)  wxyzhCEGKOSW[_cAutoList42)1)a)((3/Q$ !   d                      TABLE F                                 f $ـ$  12  ڀ  _ԀDataiscollectedforblocksusuallydefinedbyphysicalfeatures.AftertheCensusis  completed,blocksaresplitbyotherboundaries,suchaspoliticalboundaries,tocreatenew  blocks.Dataistabulatedbasedonthenewblockdefinitions.   )Hairline d    d $ـ$  1  ڀ  _ԀTheCensusDayresidentsareboththenonmoversandtheoutmovers.Nonmoverslivedatthe  sampleaddressonCensusDayandatthetimeoftheA.C.E.interview.Outmoverslivedthereon  CensusDay,butnotonthedayoftheA.C.E.interview.                              $ـ$  2  ڀ  _ԀXaGXXXAconflictinghouseholdreferstothehouseholdsatamatched,non-vacantaddressor  individualhousingunit,wheretheA.C.E.householdandcensushouseholddonotcontainany  matchedorpossiblymatchedpeople.#XXXXaGJ#(9 Z(Times New Roman      $ـ$  3  ڀ  _ԀThecurrentresidentsarethepeoplewholivedinthehousingunitatthetimeofthePES  interviewin1990,whicharethenonmoversandinmovers.Thenonmoverslivedatthesample  addressonCensusDayandatthetimeofthePESinterview.Theinmoversdidnotlivethereon  CensusDaybutmovedtotheaddressbeforethedateofthePESinterview. $ـ$  4  ڀ  _BaseDAestimatesrefertotheDemographicAnalysisestimatesproducedinJanuary2001by  theCensusBureauforCensus2000. $ـ$  5  ڀ  _TheprocessofrevisingtheDemographicAnalysisestimatesmadeuseofCensus2000long  formdatatoreviseestimatesoftheforeignbornpopulation. $ـ$  6  ڀ  _TheMarchCurrentPopulationSurveywasreweightedusingtheCensus2000countsbyage,  race,sex,andHispanicoriginforthiscomparison.  $ـ$  7  ڀ  _Thisfigurediffersfromthe1.18percentusuallyquotedfortheXaGXXXMarch2001#XXXXaG#A.C.E.becausethe  A.C.E.andDAestimatedifferentpopulations.ThebaseoftheDApercentisthetotal  population,whilethebaseoftheA.C.E.percentisthehouseholdpopulation,whichexcludes  groupquarters.(9 Z(Times New Roman TABLE A,cAZArial XTABLE MTABLE E  $ـ$  8  ڀ  _Correlationbiasreferstothetendencyforcensusenumeratedpeopletobemorelikelyincluded  intheA.C.E.thanpeoplemissedinthecensus.TheDAsexratiosandMarch2001A.C.E.data  areusedtoproducecorrelationbiasestimatesformales.Adultfemalesareassumedtohaveno  correlationbias. $ـ$  9  ڀ  _Doublesamplingreferstothesubselectionofthefinalsamplefromapreselectedlargersample.TABLE C     $ـ$  10  ڀ  _Missingmatch,residence,orenumerationstatusarereferredtobelowasunresolvedmatch,  residence,orenumerationstatus.Sometimeswealsorefertothemasunresolvedpersonstatus. b $ـ$  11  ڀ  _Theimputationcellestimatorseparatedpeoplewithresolvedandunresolvedmatchorresident  orcorrectenumerationstatusintogroupscalledimputationcellsbasedonoperationaland  demographiccharacteristics.Withineachimputationcell,theweightedproportionofmatchesor  residentsorcorrectenumerationsamongthecaseswithresolvedstatuswascalculated,andthat t valueimputedforallunresolvedpeopleinthecell.(Cantwellet.al,2001.)  $ـ$  13  ڀ  _SyntheticerrorisnotincorporatedintotheA.C.E.varianceestimates.Theassumptionisthat  thecoveragerateisuniformoverallareaswithinpost-strata.Syntheticerrorisintroducedtothe  extenttheareasdeviatefromthisassumption.Theaccuracyofthismethodologymaydecreasein  areaswherelocalizedeffectsnotreflectedinthepost-stratificationaffectthetruesampling t variance.Thediscrepancybecomeslargerasthepopulationofanareadecreases.Thus,caution ` shouldbeusedincomparisonsbetweenareasofdifferentsizes.(Starsinicet.al.,2001.)  $ـ$  14  ڀ  _UnresolvedpeoplearethoseforwhomtheCensusBureaudidnothaveenoughinformationto  accuratelycodewhethertheyshouldbecountedaserroneousorcorrectenumerations.  $ـ$  15  ڀ  _ConflictingpeoplearetheseforwhomtheCensusBureauobtaineddifferentresidency  informationinPFUandEFUandtheCensusBureaucouldnotdeterminewhichwascorrect.  $ـ$  16  ڀ  _Discrepantenumerationsincludefalsification(theamountisuncertain),butdonotinclude  honestmistakesmadebytheinterviewersorrespondents.Apersonisclassifiedasdiscrepant  duringthematchingoperationifthreeknowledgeablerespondentsindicatenotknowinghimor  herineithertheEFUorproductioninterview.  $ـ$  17  ڀ  _& t XaGXXXFortheDressRehearsalICM,theCensusBureauplannedtoconductQAononlyaportionof  thework,butlogisticalconcernsnecessitateda100percentQA.Forthe2000A.C.E.,QAwas  doneonasamplebasisoncethematcherreachedaspecifiedlevelofproficiency(periodically,  thematchingsoftwarereevaluatedthedecisiontosample).#XXXXaGV#XaGXXXԀThesampleQAinvolveda t dependentrematchon1/6oftheclerks(thelowestlevelofmatchers)and1/10ofthe ` technicians(themiddlelevelofmatchers)work.Inaddition,casesmeetingspecial mustdo L  criteriawerereviewed.(Byrne,2001.)#XXXXaG%#'t:"J      (9 Z 6Times New Roman Regular    ) $ـ$  20  ڀ  _ԀXaGXXXIn1990thesearchareawasthesampleblockclusterandsurroundingringsofblocks.The  surroundingblocksconsistedofoneringforurbanareas,tworingsforsuburbanareas,anda  largerareaforthemostruralareas.In2000thesearchareawasonlyoneringofsurrounding  blocksforallareasandthesearchwastargetedtocertainclusters.#XXXXaGK#(9 Z(Times New Roman http://www.census.gov/pred/www/  $ـ$  21  ڀ  ݀____Not_Ԁahousingunitcanbenonresidentialordidnotexistasahousingunitinthesearcharea.  Ahousingunitidentifiedasnotbeingfoundwithinthesearchareamayhaveexistedasahousing  unitoutsidethesearcharea.  $ـ$  19  ڀ  _Thedatadefinedpopulationincludescensuspersonrecordswithsufficientdatatobeaccepted  forfurtherprocessing.DatadefinedrecordsforCensus2000musthaveatleasttwocompleted  items.Oneitemmaybename,definedasatleastthreecharactersinthenamefield.Records  thatarenotdatadefinedarewholepersonimputations.    $ـ$  22  ڀ  _ԀThesenumbersarefromtheunweightedconflictinghouseholds. http://www.census.gov/dmd/www/EscapRep.htmlhttp://www.census.gov/dmd/www/EscapRep2.htmlhttp://www.census.gov/dmd/www/ace2.html !  _X NXaGXXXTRX3'LetterX3' Letter3'Letteretter3'LetterT 8NaGNXXdd8(  N XNXaG   Census2000Testing,Experimentation, D andEvaluationProgram0 h 0h(#h(#0(#(#0p(#(#0p(#p(#0 (#(##NXaGXN )#N&%XNXaGAugust28,2003#NXaGX%N&M#N XNXaG~ (# (# __________________________________________________________ u    `     h     K   `     h      p #NXaGXN # TopicReportSeries,No.4  % #XXXNXaGM# LXX#XX L##XCoverageMeasurement  fromthePerspectiveof 5 March2001A.C.E.#XX##( & X0  #XX& ( ~# LXXA(#(# #XX L#XaGXXXQualityassuranceprocedureswereappliedthroughoutthecreationofthisreport.#XXXXaG0# LXX  #XX L#Thistopicreportintegratesfindingsandprovidescontextandbackgroundforinterpretationof  resultsfromCensus2000evaluations,tests,andotherresearchundertakenbytheU.S.Census  Bureau.Itispartofabroadprogram,theCensus2000Testing,Experimentation,andEvaluation  program,designedtoassessCensus2000andtoinform2010CensusPlanning.XaGXXX o  XXaG#XaGX  # XXaG# L  #  L  Preparedby @ 0 @` 0@ ` (#` (## L } ##XX L; # LXX  L0@ (# (#0@h(#(#RitaPetroni /&# and# L I #  L#XaGX * # XXaGԀDannyR.Childers# L  #  L#XaGX  # XXaG i'$ AV) xdME(xA (%  0 @  @` #XaGX  # XXaG0@ (#(#0@ (# (#0@h(#(#Planning,Research,and#XaGX v # XXaGԀ#XaGX  # XXaGEvaluationDivision# L o ##XX LN#XXԀ?#XaGX?# XXaG ( &! and#XaGX (##XXXXaG#XXԀ?#XaGX?# XXaGDecennialStatistical#XaGX # XXaG# L #  LԀStudiesDivision# L ##XX L^#XX#XX# )F'"  4 XX   #XX 4#4 XXn[@:*&`z0r+$  `(*E-$ `$ `e- n (#(#(#(# #N-*%(#(# # X8XXdNXXd8   TRX3'Avery 5366 File Folder -X3'Letter3'Letter366 File Folder -Xd3'LetteretterT  CONTENTS #XX 4#   (Ɠ83TableofContents""J(#.H H _i_Ԉ  ListofTables""H(#.  iii  1. Introduction""J(#. 1 t  *,X` XX*0  1.1TheCensusDesign#""J(#.  !(#(##3 L  0  1.2CoverageMeasurement#""J(#.rr$(#(##4 8  6,X, X,X` X60  1.30(#(#Differencesin1990PostEnumerationSurveyand2000AccuracyandCoverage $ t  Evaluation.#""J(#. (#(##8 `  2.DemographicAnalysis:ComparisonwithMarch2001A.C.E.CoverageEstimates""J(#.\\X9 8  0  2.1InternationalMigration#`"`"I(#.PP'(#(##10   0  2.2MeasurementofVitalEvents#`"`"I(#.22+(#(##11   0  2.3ResultsofRevisedDA#`"`"I(#.66%(#(##12  0   (#(# 3.March2001A.C.E.:PersonCoverage`"`"I(#.014 p 0  3.1SamplingandEstimation#`"`"I(#.  ((#(##14 H Ѐ0  0(#(#3.1.1Sampling#`"`"I(#. (#(##15 4 0  0(#(#3.1.2WeightTrimming#`"`"I(#.rr$(#(##16  p 0  0(#(#3.1.3MissingData#`"`"I(#. !(#(##16  \ 0  0(#(#3.1.4DualSystemEstimation#`"`"I(#.^^+(#(##20 H 0  0(#(#3.1.5SyntheticEstimation#`"`"I(#.NN)(#(##22 4 0  0(#(#3.1.6VarianceEstimation#`"`"I(#...((#(##22   0  3.2A.C.E.PersonInterviewing#`"`"I(#.*(#(##NXaGXXX@`"`"I(#.24#XXXNXaG#NXaGXXXԈ   0  3.3ErrorSources#`"`"I(#. (#(##26  0  #XXXNXaGV#0(#(#3.3.1ErroneousEnumerations,IncludingDuplicates#`"`"I(#.jjA(#(##27  0  0(#(#3.3.2CensusOmissions#`"`"I(#.%(#(##32   0  0(#(#3.3.3BalancingError#`"`"I(#. $(#(##34 l! 0  0(#(#3.3.4CorrelationBias#`"`"I(#. %(#(##35 X"  0  0(#(#3.3.5Conditioning#`"`"I(#. !(#(##36 D#! 0  0(#(#3.3.6LateAdditions#`"`"I(#.p p #(#(##36 0$" 0  0(#(#3.3.7CensusImputations#`"`"I(#.'(#(##36 %l # 4.A.C.E.:HousingCoverage`"`"I(#.  &37 &D"% ̀ 4.1HousingUnitCoverageStudyNXaGXXX`"`"I(#.<<.38 ($' #XXXNXaG"#Ԁ4.2ConflictingHouseholds`"`"I(#.FF)41NXaGXXXԈ )%( #XXXNXaGa##Ԁ4.3HousingUnitFieldOperationsandInstrumentsNXaGXXX`"`"I(#.@43 *%)  +&* 5.ConclusionandRecommendations`"`"I(#.\\+44  0  5.1DataCollection#`"`"I(#. (#(##45  0  5.2SurveyDesign#`"`"I(#. (#(##45 t 0  5.3Estimation#`"`"I(#.l l (#(##46 ` #XXXNXaG##0  5.4CoverageMeasurementEvaluations#`"`"I(#.TT0(#(##47 L  6.References`"`"I(#.  47 $ t 0  6.1ESCAPReports#`"`"I(#.  (#(##47  L  0  6.2ESCAPIIReports#`"`"I(#. (#(##48  8  0  6.3EvaluationReports#`"`"I(#. "(#(##50 $  0  6.4OtherReferences#`"`"I(#.t t (#(##50   7._Acknowledgements_`"`"I(#. 54   Appendix`"`"I(#.55  p  XaGXXXXXaG  LISTOFTABLES #XaGX(#  (6X` X,X, X6Table1:#XXXXaG(#XaGXXX  PercentNetUndercountforMajorGroups:March2001#XXXXaG)#XaGXXXA.C.E.and1990PES#XXXXaG*#XaGXXX""J(#. X8#XXXXaGq*#XaGXXXԈ  Table2:#XXXXaG*#0  ResidentPopulationTotalsfromCensus2000,DemographicAnalysis,andtheMarch  2001A.C.E.:April1,2000#`"`"I(#.+(#(##12 t  Table3:0  PercentNetUndercount,byRace,Sex,andAge:1990and2000#`"`"I(#.))M(#(##13 L  Table4:0  March2001A.C.E.and1990PESMissingDataRates(weighted)#`"`"I(#.L(#(##18 $ t  Table5:0  DistributionofCVsforPopulationEstimatesbyGeographicalAreaforMarch2001 L  A.C.E.and1990PES#`"`"I(#.#(#(##23 8  Table6:0  Comparisonof1990and2000PercentNetUndercountofHousingUnits#`"`"I(#.S(#(##38   Table7:0  NationalHousingUnitCoverageEstimates#`"`"I(#.mm8(#(##39  Table8:0  OverallPercentE-sampleHousingUnitDuplication#`"`"I(#.99A(#(##41 p   \ _Oa 4 XXZ_R _c TQ OaOa#OadO *OajOadOhOaOadOOa(Oa)Oa+Oa5Oa,Oa-[ hOa _cNOa.1.Introduction#XX 4/#   TheCensus2000Testing,Experimentation,andEvaluationProgramprovidesmeasuresof  effectivenessfortheCensus2000design,operations,systems,andprocessesandprovides  informationonthevalueofnewordifferentmethodologies.Theresultsandrecommendations  fromtheseanalysesprovidevaluableinformationcrucialtoplanningthe2010Census.By t  providingmeasuresofhowwellCensus2000wasconducted,thisprogramfullysupportsthe `  CensusBureausstrategytointegratethe2010planningprocesswithongoingMasterAddress L  File/TIGERenhancementsandtheAmericanCommunitySurvey.Thepurposeofthereportthat 8  followsistosynthesizeresultsfromrelatedCensus2000evaluations,experiments,andother $ t  assessmentstomakerecommendationsforplanningthe2010Census.Census2000Testing, `  Experimentation,andEvaluationreportsareavailableontheCensusBureau'sInternetsiteat: L  4 IO  5  http://www.census.gov/pred/www/66IO6#  7 6.ThereportsanddocumentationoftheExecutiveSteering 8  CommitteefortheA.C.E.Policy(ESCAP)arelocatedat: $  49IO  5  http://www.census.gov/dmd/www/EscapRep.html67IO84  7^18,   4:IO  5  http://www.census.gov/dmd/www/EscapRep2.html68IO84  79,and   4;IO  5  http://www.census.gov/dmd/www/ace2.html69IO94  7:.  TheCensusBureauconductedtheAccuracyandCoverageEvaluation(A.C.E.)expectingit p couldbeusedtoadjusttheCensus2000resultsforallnon-apportionmentpurposesifitimproved \ thecensusdata.TheoriginalMarch2001A.C.E.estimatesbecameavailableintimetocorrect H theCensus2000redistrictingfiles.OnMarch1,2001,theCensusBureaureleasedthe Report 4 oftheExecutiveSteeringCommitteeforAccuracyandCoverageEvaluationPolicywhich  p reportedthat TheExecutiveSteeringCommitteeforA.C.E.Policy(ESCAP)isunableto  \ conclude,basedontheinformationavailableatthistime,thattheadjustedCensus2000dataare H moreaccurateforredistricting.Accordingly,ESCAPrecommendsthattheunadjustedcensus 4 databereleasedastheCensusBureausofficialredistrictingdata.NXaGXXXԀ#XXXNXaGz>#(ESCAP,2001.)NXaGXXXԀ#XXXNXaG>#XaGXXXdB  ԍ(51, H recommendationpage,R1)C  #XXXXaG?#d   TheESCAPnotedthedifferencebetweentheA.C.E.estimate,a3.3millionnetundercount,and  DemographicAnalysis(DA)results,a1.8millionnetovercount.TheCensusBureauconducted  furtherevaluationsoverthenextsixmonthstoexaminethisdifferenceanddetermineif   Census2000data,otherthanredistrictingdata,shouldbecorrected.TwoplannedA.C.E. l! evaluationprograms,theMatchingErrorStudy(MES)(Bean,2001)andtheEvaluation X"  Followup(EFU)(RaglinandKrejsa,Report3,2001),identifiederrorsintheA.C.E.ThePerson D#! DuplicationStudy(Mule,Report20,2001)usedcomputermatchingtoidentifyduplicatesacross 0$" theentirecountryandFeldpausch(2001)examinedtheenumerationstatusassignedtothe %l # Esamplefortheseduplicates.AdamsandKrejsa(2001)re-codedtheenumerationstatusto &X!$ reduceanyoperationalandproceduralerrorsintheoriginalenumerationstatuscoding. &D"% Additionalevaluationsaddressedotherconcerns(ESCAPII,2001)includingA.C.E.balancing, '0#& contamination,andmissingdata.TheDAestimateswereinvestigatedfurtherresultingin ($' revisions(particularlymigrationestimates)andrevisedDAestimates.(Robinson,Report1, )%( 2001.)DuetouncertaintywhetherallerrorsassociatedwiththeA.C.E.(e.g.duplicationerror) *%) werecaptured,resultsfromthetotalerrormodeldesignedtosynthesizeindividualerrorsandthe +&* associatedlossfunctionanalysiswerenotused.(Petroni,2001.)dB  ԍ(59,p2,R1)C  dNXaGXXX  #XXXNXaGF#OnOctober17,2001,theCensusBureaureleased ReportoftheExecutiveSteeringCommittee  forAccuracyandCoverageEvaluationPolicyonAdjustmentforNon-RedistrictingUseswhich t reported, ` ? X4` X%XX` X? 0 4  44 0 4 TheExecutiveSteeringCommitteeforA.C.E.Policy(ESCAP) 8  recommendedonMarch1,2001thatunadjustedcensusdatabeusedfor $ t redistricting.Afterassessingconsiderablenewevidence,ESCAPnow  ` recommendsthatunadjustedCensus2000dataalsobeusedfor  L  non-redistrictingpurposes.Theeffectofthisnewevidenceisthatthe  8  AccuracyandCoverageEvaluation(A.C.E.)overstatedthenetundercount $  byatleast3millionpersons.ThecauseofthiserrorwasthattheA.C.E.   failedtomeasureasignificantnumberofcensuserroneousenumerations,   manyofwhichwereduplicates.ThisleveloferrorintheA.C.E.   measurementofnetcoverageissuchthattheA.C.E.resultscannotbeused   intheircurrentform.Thisfindingofsubstantialerror,inconjunctionwith p remaininguncertainties,necessitatesthatrevisions,basedonadditional \ reviewandanalysis,bemadetotheA.C.E.estimatesbeforeanypotential H usesofthesedatacanbeconsidered.TheCensusBureauwillreleasethe 4 remainingCensus2000dataproducts,post-censalestimates,andsurvey  p controlsusingunadjusteddata.Itis,however,reasonabletoexpectthat  \ furtherresearchandanalysismayleadtorevisedA.C.E.estimatesthatcan H beusedtoimprovefuturepost-censalestimates.(ESCAPII,2001.)XaGXXXdB  ԍ(1,p  p 1,R22)C  #XXXXaGN#d444 CX` X X4` X%X$HC cHCoupledwiththerevisionstotheDAestimates,theinconsistencywithDAwasexplainedbythe   failureoftheA.C.E.tomeasurealargenumberofcensuserroneousenumerations.Theearlier  concernsinA.C.E.withbalancing,contamination,andmissingdatawerealsoresolved.The  levelofothererrorswasbelievedtobesmallbycomparisonandthereforewasnotamajorfactor  inthesecondESCAPdecision.(Hoganetal.,2002;MulryandPetroni,2002;ESCAPII,2001.) l XaGXXXdB  ԍ(59,p2,R2)C  #XXXXaG7R#dInOctober2001,theCensusBureaureleasedapproximateestimatesoftheundercountforthree D! race/Hispanicorigingroups.(Thompsonetal.,2001.)These RevisedEarlyApproximations 0"  correctedestimatesoferroneousenumerationsforcensusduplicatesandforothererroneous #l! enumerationsidentifiedintheA.C.E.evaluationsbutnotinthefullA.C.E.Esample.This $X" illustratedthecorrectioneffectsonnetundercountestimatesandonpossiblecoverage $D # differences.TheCensusBureaulaterusedthesamemethodsanddatatoexpandthecalculations %0!$ tosevenrace/Hispanicorigingroups.(Fay,2002;Mule,2002.)Thesepreliminaryestimates &"% showed,liketherevisedDAresults,asmallnetundercountandthatthedifferentialundercount '#& wasreduced,butnoteliminated.Theseresultsonlyprovideddataatthenationallevelforbroad (#' populationgroups.Furthermore,thesepreliminaryapproximationswerebasedonasmallsubset )$( ofA.C.E.dataandonlypartiallycorrectedforerrorsinmeasuringerroneousenumerationsusing |*%) Fayslowerbound.(Fay,2001,Fay,2002.)Potentialerrorsinmeasuringomissionswerenot h+&* accountedfor.XaGXXXdB  ԍ(59,p2,R3)C  #XXXXaG!X#d  EventhoughtheESCAPrecommendedtwice .XXNOT #XX.X#tocorrectthecensuscounts,theyhad  concernsaboutdifferentialcoverageinCensus2000.Theythoughtfurtherresearchonrevised t coverageestimatescouldbeusedtoimprovethepost-censalestimates.Workonrevised ` estimateswouldprovideabetterunderstandingofCensus2000coverageerrorthatcouldbeused L  toimprovethe2010Censusanddevelopbettermethodsforthe2010coveragemeasurement 8  program.Hence,workbeganonrevisingtheA.C.E.estimatestocorrectfordetectederrors.B  ԍd(59, 8  p3,R4)C  dThe $ t resultscanbefoundinthe TechnicalAssessmentofA.C.E.RevisionII.(U.S.CensusBureau,  ` 2003.)  L  ThefollowingreportisdividedintothreesectionsdiscussingthecomparisonofMarch2001 $  A.C.E.coverageestimateswithDemographicAnalysis,March2001A.C.E.personestimates,   andA.C.E.housingunitestimates.Theconclusionsandrecommendationsarebasedonthe   March2001A.C.E.,workdonepriortotheA.C.E.RevisionII.Therecommendationsforthe   2010CensusassumetheCensusBureaudecidestodoasimilarapproachtoundercount   estimationusingacoveragemeasurementsurvey.TheresearchandresultsfromA.C.E.Revision p IIareoutofscopeforthisreportduetoresourceconstraints.Thecoverageestimatesfrom \ A.C.E.RevisionIIareconsideredbetterthantheMarch2001A.C.E.estimates.Infactthe H earliercoverageestimatesareconsideredflawedandarenotindicativeoftheCensus2000 4 coverageerror.Toavoidmisunderstandings,earlierflawedestimatesarereferredtoasMarch  p 2001A.C.E.intheremainderofthereport.Alistoftheevaluationsandotherreportsusedin  \ thistopicreportareinthereferencessection.Allevaluationsdiscussedinthisreportarebased H ontheMarch2001A.C.E. 4 Thehousingunitcoverageestimatesdidnothavetheexhaustiveevaluationthatwasconducted   forthepersonestimatesandA.C.E.RevisionIIdidnotattempttorevisethehousingunit  coverageestimates.Thedifficultiesinidentifyingresidenceinthepersonestimatesshouldnot  affectourabilitytoidentifytheexistenceofahousingunitonCensusDay.However,tothe  extentnotidentifyingerroneouscensusenumerationsorduplicateswaslargeforthehousehold, l thenthiscouldhaveaffectedthereliabilityofsomeofthehousingestimatesbyhouseholder X  characteristicssuchasoccupancystatus,race,orownerstatus. D!  &  4 XX1.1TheCensusDesign#XX 4d#  #l! Census2000paralleledthedesignofotherrecentU.S.DecennialCensusesinmanyrespects. %l # '#dCensus2000attemptedtoenumerateallpeoplelivingintheUnitedStatesonApril1,2000. &X!$ Mostofthepopulationwasenumeratedbymeansofmailbackquestionnairesdeliveredtotheir &D"% homesinMarch2000.Themailbackquestionnaireasked, Howmanypeoplewerelivingor '0#& stayinginthishouse,apartment,ormobilehomeonApril1,2000?,thenaskedrespondentsto ($' answerquestionsforeachperson.Respondentsweregivenguidanceaboutwhomtoinclude )%( ( fosterchildren,roomers,orhousemates; peoplestayinghereonApril1,2000whohaveno *%) otherpermanentplacetostay; peoplelivingheremostofthetimewhileworking,evenifthey +&* haveanotherplacetolive)andwhomtoexclude( collegestudentslivingawaywhileattending  college; peopleinacorrectionalfacility,nursinghome,ormentalhospitalonApril1,2000;   ArmedForcespersonnellivingsomewhereelse; peoplewholiveorstayatanotherplacemost  ofthetime).Peoplewhodidnotrespondbymailwereenumeratedinpersonbyenumerators t whovisitedtheirhomesduringNonresponseFollowup(NRFU)betweenApril27andJune26. ` Inmostcases,NRFUinterviewersspoketoamemberofthenonrespondinghousehold,butafter L  theyhadattemptedrepeatedlytocontactnonrespondinghouseholds,theywereallowedtoobtain 8  basicdataabouttheresidentsfromproxyrespondents,suchasneighbors,landlords,orother $ t nonhouseholdmembers.(Martin,Fay,andKrejsa,November2002.)dB  ԍ(24,p5,R2)C  d  ` Peopleingroupquarters(suchascollegedormitories,homelessshelters,andnursinghomes)  8  wereenumeratedinseparateoperationsatthefacilities,whereenumeratorslistedthenamesof $  thepeoplelivingorstayingthereandleftIndividualCensusReportsforeachpersontocomplete.   Enumeratorspickeduptheformsand,ifnecessary,conductedinterviewswithnonrespondents.   (Martin,Fay,andKrejsa,November2002.)XaGXXX   Twobasicerrorsaffectedthepopulationtotalfromthecensus:omissionsofpersonswhoshould p havebeencounted,anderroneousenumerationsofpersonswhoshouldnothavebeencounted, \ suchasfictitiouspersonsorpersonscountedmorethanonce.Personswereconsideredomittedif H theywerenotcountedintherightgeographicareaanderroneouslyincludediftheywere 4 incorrectlycountedinadifferentarea.TheA.C.E.definedomissionsanderroneous  p enumerationswithrespecttoarelativelysmallgeographicareacalledthesearcharea,whichwas  \ typicallyablockorgroupofblocksandinsomecasesblocksimmediatelysurroundingthe H sampleblocks.Underthisapproach,peoplecountedinthewrongblockwereclassifiedas 4 omittedfromwheretheyshouldhavebeencountedanderroneouslyenumeratedwheretheywere   counted.#XXXXaGXo#(Martin,Fay,andKrejsa,November2002.)XaGXXXԀdB  ԍ(24,p5,R3)C  d#XXXXaGs#XaGXXX   Inthedecennialcensus,theCensusBureauattemptedtoenumerateeachpersonathisorher   usualresidenceasofApril1st,definedastheplacewhereapersonlivedorsleptmostofthe  time.Thebasicusualresidenceprinciplewasbasedon31residenceruleswhichapplytospecial l circumstances:forexample,peoplewhowerestayinginmosttypesofinstitutionalsettingsor X  othergroupquarters(e.g.,dormitory,shelter,ornursinghome)onApril1stwereenumerated D! there,eveniftheyalsohadanotherresidence.Theinstructionsonthecensusformdescribedthe 0"  mostcommonlivingsituations,butrespondentsoftenfindtherulesself-contradictoryandthe #l! terminologyconfusing.(Martin,Fay,Krejsa,2002.)#XXXXaGSt#XaGXXXdB  ԍ(60,p1,R1)C  d#XXXXaGx#XaGXXXԀAlso,sometypesofnoninstitutionalgroup $X" quartersallowedtherespondenttoindicatetheyshouldhavebeencountedattheirusualhome. $D #  &  XXaG1.2CoverageMeasurement#XaGXuy#  &"% TheA.C.E.attemptedtomeasurenetundercount'&iythroughasamplesurvey.Conceptually,an ($' independentsampleofthepopulation,thePsample,wasusedtoestimatetheomissions.A )%( sampleofcensusenumerations,theEsample,wasselectedfromcensusenumerationsto *%) determineerroneousenumerations.Togethertheywereusedtoestimatethenetcoverageerror. +&* Omissionsanderroneousenumerationsaredefinedwithinthesearcharea.Theyarenotmeantto  beestimatesofgrosserrors.Followingtheprecedentofthe1990coveragestudy,theA.C.E.  geographicallyoverlappedthePsampleandEsamplebyselectingthemfromthesamesampleof  blocks.#XXXXaGx#K+ ` hp x X` XKXaGXXXdB  ԍ(24,p6,R5)C  d#XXXXaG}#Seethe"TechnicalDocumentationforMarch2001Estimates"inKostanich(2003)for t moredetails.XaGXXX ` ThehousingunitstageofA.C.E.wasacombinationofsamplingandoperationalactivitieswhich 8  resultedintheselectionof300,000P-samplehousingunitsfromthe50statesandtheDistrictof $ t Columbia,#XXXXaG{~#XaGXXXexcludingareasofremoteAlaska#XXXXaG#XaGXXX.dB  ԍ(61,p2,R1)C  #XXXXaG*#dTheselectionwasindependentofanycensus  ` operation.AdetaileddescriptionoftheA.C.E.samplingplancanbefoundinZuWallack,  L  Salganik,Cromar,andMule(2000).  8  TheA.C.E.comprisedseveraloperations,butfiveprimaryoperationsarecriticalintheanalysis   thatfollows.FormoredetailsseeChilders(2001).The GlossaryofSpecializedandTechnical   TermsUsedintheESCAPReportandSupportingDocumentsisahelpfulsourcefordefining   termsusedthroughoutthisreport.(ESCAPII,ChapterVIII,2002.)      wxyz(o"p8" *"! !   *2o"3  0    AninitialinterviewofP-samplehouseholdswasconductedbyphone(April24through \ June13)orbypersonalvisit(June18throughSeptember11),usingacomputer-assisted H instrumentonalaptop.Theinterviewestablishedboththecurrentresidentsand,if 4 different,theCensusDayresidentsofthesampledhousingunits.XaGXXXTheinterviewwas  p conductedonlywithahouseholdmemberforthefirstthreeweeksofinterviewing.Ifthe  \ interviewwithahouseholdmemberwasnotsuccessfulafterthreeweeks,aninterview H withaproxyrespondent,whichisanonhouseholdmember,wasattempted.#XXXXaGׅ#ԀThe 4 ComputerAssistedPersonalInterview(CAPI)instrumentwasdesignedtoobtainaroster   ofthecurrentresidentsandtheresidentsonCensusDay,measuretheirdemographic   characteristicssuchassex,age,race,andethnicity,anddeterminewhethereachidentified  CensusDayresidentshouldhavebeenincludedinthecensusinthehousingunitor  somewhereelse. *݌(#(# Ќ! !  " *"! !   *2o"3  0    InOctoberthroughNovember2000,P-sampleCensusDayresidents %#  1      ׀werematchedtothe X  census.IftheA.C.E.interviewestablishedtheCensusDayaddressforaP-sampleperson D! andifthatpersonmatchedapersonenumeratedintheCensus,thentheP-sampleperson 0"  wasconsideredmatchedtothecensusandthecorrespondingE-samplepersonclassified #l! asacorrectenumeration. *݌$X"(#(# Ќ! !  " *"! !   *2o"3  0    SomecategoriesofP-samplepeoplewhodidnotmatchtothecensusweresenttothe %0!$ A.C.E.PersonFollowup(PFU)becausetheCensusBureauwasnotabsolutelycertain &"% abouttheinformationprovidedintheoriginalinterview.Thecategoriesincludednot '#& matchedpeoplefromproxyinterviews,fromconflictinghouseholds %#  2      ,andfrom  householdswheresomepeoplematchedXaGXXX.ThefollowupinterviewidentifiedP-sample  peoplewhowerenotresidentsofthehousingunitonCensusDaywhowerethen  removedfromthePsample.Thefollowupinterviewalsoincludedallnotmatched t E-samplecasesinordertoidentifyerroneousenumerations.#XXXXaGQ#Becausetheywerematched, ` themajorityofcensusenumerationsintheEsamplerequirednoseparatefieldwork. L  FollowupinterviewswereconductedinpersonfromOctober19throughNovember21, 8  2000,usingaquestionnairepreprintedwithnameandaddressinformationaboutthe $ t samplehousehold.Interviewswereacceptedwithnon-householdproxiesif  ` knowledgeablehouseholdrespondentswerenotfoundaftersixcontactattemptson  L  differenttimesondifferentdays. *݌ 8 (#(# Ќ! !  XaGXXX" *"! !   *2o"3  0    #XXXXaG#NXaGXXXDespiteextensiveattemptstointervieweveryhousingunitinthePsample,therewere   householdsforwhichtheCensusBureausimplycouldnotobtaininformation.To   accountforthemissinginformationfromthesehouseholdsnotinterviewed,theCensus   Bureauappliedanoninterviewadjustmenttointerviewedunits.#XXXNXaGe#Furthermore,although   mostP-samplepeoplewereassignedaresidenceandmatchstatus,andE-samplepeople p anenumerationstatus,asmallnumberofpeopleremainedwithoneormoreofthese \ statusesunresolved.Thatis,theCensusBureaumaynothavebeensureifapersonwas H actuallyaresidentofthehousingunitonCensusDay,orifanotherpersonwascorrectly 4 enumeratedinthecensus.Forthesepeoplewithanunresolvedstatus,theCensusBureau  p assignedaprobabilityofhavinglivedintheblockclusteronCensusDay,having  \ matched,orhavingbeencorrectlyenumerated.SeeCantwellet.al.(2001)fora H discussionoftheA.C.E.missingdataprocedures. *݌4(#(# Ќ! !  " *"! !   *ę2o"3  0    XaGXXXTheresultsofmatchingthePsampleandEsamplewereusedtoproducepopulation   estimatesusingDualSystemEstimation(DSE).TheDSEisatechniquethatestimatesthe  truepopulationusingestimatesofthenumberofcensusenumerationscorrectlyincluded  inthecensusfromtheEsampleandtheratioofthenumberofpeoplewhoshouldhave  beenincludedandwerecorrectlyincludedinthecensustothetotalpopulationfromthe l Psample.SeeSekarandDeming(1949),Wolter(1986),andHogan(1993)#XXXXaGX#B  C  XaGXXXformore X  informationaboutDualSystemEstimation.Estimatesofthepopulationaremadewithin D! estimationcells,calledpost-stratadefinedbygeographyanddemographicvariables.The 0"  sumoftheestimateofthepopulationacrossestimationcellsistheDualSystemEstimate #l! ofthepopulation.ThenetundercountisthedifferenceintheDualSystemEstimateof $X" thepopulationandthenumberofpeoplecountedinthecensus.Thepercentnet $D # undercountisthenetundercountdividedbythenumberofpeoplecountedintheDual %0!$ SystemEstimateofthepopulation.#XXXXaG # *ę݌&"%(#(# Ќ! !   '#& Oa     Table1comparespercentnetundercountestimatesfromtheflawedMarch2001A.C.E.andthe  1990PostEnumerationSurvey(PES)formajorgroups.TheA.C.E.estimateswerelaterrevised.  (SeeU.S.BureauoftheCensus,2003.)Comparisonshereshowresultsbeforethoserevisions.  TheDSEshowsCensus2000undercountedthenationalhouseholdpopulationandthat t undercountsdifferedbypopulationsubgroups.Relativetothe1990census,Census2000 ` showedanapparentimprovementintheoverallpercentnetundercountandthedifferential L  undercountsofcertainpopulationgroups.Thenationalpercentnetundercountofthehousehold 8  populationforCensus2000is1.18percent(standarderror,0.13percent)comparedtothe1990 $ t Census1.61percent(standarderror,0.20percent).TheCensus2000coverageshowed  ` differentialundercountratesamongtherace/origindomains,tenure,andage/sexgroups.B  d(10,p1,  R1)C  dThe  L  percentnetundercountfortheNon-HispanicBlackandHispanicdomainsislowerfor  8  Census2000comparedtothe1990Census,whichresultsinadifferentialundercountreduction $  relativetotheNon-HispanicWhiteand Someotherracedomain.TheCensus2000percent   undercountforNon-Ownersandforchildrenages0to17islowerthanin1990.(Davis,2001.)B  d 8  (10,p2,R2)dC         Z     & r!  Table1:PercentNetUndercountforMajorGroups:March2001A.C.E.and1990PES XXXXXXdB  ԍ(10,p3,R3)C  Ԛ#O##XXn#d  *ddd Xdd Xdd X(#(#, ,4% ,% , , , , , +  0V# 0ЀMarch2001A.C.E.* BV]3$T"V B1990PESXX NVA)T"  V] N 9V,! p3V 9 <V/! p3V <Net 9V,! p"V 9Standard ;V],! p"V ;Net 9V,! p"V] 9Standard <V/! p "V < 9V,! p "V 9 ND, p "  V N '2 3 ' * 2 3 *Undercount 6,!2" 6Error ,]2" ,Undercount 6,!2"] 6Error * 2" * 6,!2" 6 JV92"   JCharacteristic CV2$V C(Percent) @V/!"V @(Percent) =]V,!"V =(Percent) @V/!"]V @(Percent) DV3!"V DCharacteristic#XX#XX QG/  V Q '~ 3 ' * ~ 3 * 6,!~ 3 6 ,]~ 3 , 6,!~ 3] 6 * ~ 3 * 6,!~  3 6 C9~ !3 C Total OEb " zG?1.18zG?O1.18 yoBb #U zG?1.18 zG? p= ף?0.13p= ף?y0.13 o]`3b $U p= ף?0.13 p= ף? (\?1.61(\?o1.61 yoBb %U (\?1.61 (\?] ?0.20?y0.20 OE3b &U ?0.20 ? OTotal QG/b '  Q 'R (3 ' * R )3 *Ј 6,!R *U 6Ј ,]R +U ,Ј 6,!R ,U] 6Ј * R -U * 6,!R .3 6 C9R /3 CRace/OriginDomain** -#B 0 -Ј 6,!B 1U 6Ј ,]B 2U ,Ј 6,!B 3U] 6Ј .$B 4U .Race/OriginDomain QG/B 5  Q '2 63 'NonHispanicWhite LB2 7 q= ףp?0.67q= ףp?L0.67 yoB2 8U q= ףp?0.67 q= ףp? Q?0.14Q?y0.14 o]`32 9U Q?0.14 Q? (\?0.68(\?o0.68 yoB2 :U (\?0.68 (\?] )\(?0.22)\(?y0.22 KA32 ;U )\(?0.22 )\(? K 6,!2 <3 6NonHispanicWhite&Other ?52 = ? '"r >3 'AIOffReservation LB"r ? = ףp= @3.28= ףp= @L3.28 yoB"r @U = ףp= @3.28 = ףp= @ HzG?1.33HzG?y1.33 M]>3"r AU HzG?1.33 HzG? MЈ 6,!"r BU] 6Ј * "r CU * 6,!"r D3 6 ?5"r E ? 'b F3 'NonHispanicBlack LBb G \(\@2.17\(\@L2.17 WMBb HU \(\@2.17 \(\@ W0.35 N]?b IU HzG@4.57HzG@N4.57 yoBb JU HzG@4.57 HzG@] ?0.55?y0.55 KA3b KU ?0.55 ? K 6,!b L3 6Black ?5b M ? 'R N3 'Hispanic LBR O @2.85@L2.85 yoBR PU @2.85 @ RQ?0.38RQ?y0.38 o]`3R QU RQ?0.38 RQ? (\@4.99(\@o4.99 yoBR RU (\@4.99 (\@] = ףp=?0.82= ףp=?y0.82 KA3R SU = ףp=?0.82 = ףp=? K 6,!R T3 6Hispanic ?5R U ? 'B V3 'NonHispanicAsian LBB W Q?0.96Q?L0.96 yoBB XU Q?0.96 Q? {Gz?0.64{Gz?y0.64 o]`3B YU {Gz?0.64 {Gz? zG@2.36zG@o2.36 yoBB ZU zG@2.36 zG@] = ףp=?1.39= ףp=?y1.39 KA3B [U = ףp=?1.39 = ףp=? K 6,!B \3 6AsianorPacificIslander ?5B ] ? '2 ^3 'HawaiianorPacificIslander LB2 _ ffffff@4.60ffffff@L4.60 yoB2 `U ffffff@4.60 ffffff@ )\(@2.77)\(@y2.77 M]>32 aU )\(@2.77 )\(@ MЈ 6,!2 bU] 6Ј * 2 cU * 6,!2 d3 6 ?52 e ? '"f3 'AIOnReservation LB"g (\@4.74(\@L4.74 yoB"hU (\@4.74 (\@ 333333?1.20333333?y1.20 p]a3"iU 333333?1.20 333333? q= ףp(@12.22q= ףp(@p12.22 zpC"jU q= ףp(@12.22 q= ףp(@] )\(@5.29)\(@z5.29 KA3"kU )\(@5.29 )\(@ K 6,!"l3 6AIOnReservation ?5"m ? 'n3 ' * o *Ј 6,!pU 6Ј ,]qU ,Ј 6,!rU] 6Ј * sU * 6,!t3 6 C9u CTenure -#v -Ј 6,!wU 6Ј ,]xU ,Ј 6,!yU] 6Ј .$zU .Tenure QG/{  Q '|3 'Owner LB} )\(?0.44)\(?L0.44 WMB~U )\(?0.44 )\(? W0.14 N]?U {Gz?0.04{Gz?N0.04 yoBU {Gz?0.04 {Gz?] zG?0.21zG?y0.21 KA3U zG?0.21 zG? K 6,!3 6Owner ?5 ? '3 'NonOwner LB @2.75@L2.75 yoBU @2.75 @ p= ף?0.26p= ף?y0.26 o]`3U p= ף?0.26 p= ף?  ףp= @4.51 ףp= @o4.51 yoBU  ףp= @4.51  ףp= @] Q?0.43Q?y0.43 KA3U Q?0.43 Q? K 6,!3 6NonOwner ?5 ? '3 ' * 3 *Ј 6,!U 6Ј ,]U ,Ј 6,!U] 6Ј * U * 6,!3 6 C93 CAge/Sex -#r -Ј 6,!rU 6Ј ,]rU ,Ј 6,!rU] 6Ј .$rU .Age/Sex QG/r  Q 'b3 '017 * b *1.54 6,!bU 60.19 N]?bU q= ףp @3.18q= ףp @N3.18 yoBbU q= ףp @3.18 q= ףp @] (\?0.29(\?y0.29 KA3bU (\?0.29 (\? K 6,!b3 6017 ?5b ? 'R3 '1829Male LBR )\(@3.77)\(@L3.77 yoBRU )\(@3.77 )\(@ {Gz?0.32{Gz?y0.32 o]`3RU {Gz?0.32 {Gz? ffffff @3.30ffffff @o3.30 yoBRU ffffff @3.30 ffffff @] HzG?0.54HzG?y0.54 KA3RU HzG?0.54 HzG? K 6,!R3 61829Male ?5R ? 'B3 '1829Female LBB ףp= @2.23ףp= @L2.23 yoBBU ףp= @2.23 ףp= @ (\?0.29(\?y0.29 o]`3BU (\?0.29 (\? p= ף@2.83p= ף@o2.83 yoBBU p= ף@2.83 p= ף@] Gz?0.47Gz?y0.47 KA3BU Gz?0.47 Gz? K 6,!B3 61829Female ?5B ? '23 '3049Male LB2 (\?1.86(\?L1.86 yoB2U (\?1.86 (\? RQ?0.19RQ?y0.19 o]`32U RQ?0.19 RQ? = ףp=?1.89= ףp=?o1.89 yoB2U = ףp=?1.89 = ףp=?] {Gz?0.32{Gz?y0.32 KA32U {Gz?0.32 {Gz? K 6,!23 63049Male ?52 ? '"r3 '3049Female LB"r Q?0.96Q?L0.96 yoB"rU Q?0.96 Q? (\?0.17(\?y0.17 o]`3"rU (\?0.17 (\? )\(?0.88)\(?o0.88 yoB"rU )\(?0.88 )\(?] ?0.25?y0.25 KA3"rU ?0.25 ? K 6,!"r3 63049Female ?5"r ? 'b3 '50+Male MCb п-0.25пM-0.25 zpCbU п-0.25 п  ףp= ?0.18 ףp= ?z0.18 p]a3bU  ףp= ?0.18  ףp= ? zG-0.59zGp-0.59 zpCbU zG-0.59 zG] (\?0.34(\?z0.34 KA3bU (\?0.34 (\? K 6,!b3 650+Male FV5b F @V/!R3V @50+Female cVR!RV HzG-0.79HzGc-0.79 VpCRU HzG-0.79 HzGV (\?0.17(\?0.17 ]VpBRU (\?0.17 (\?V ףp= -1.24ףp= -1.24 VpCRU ףp= -1.24 ףp= ]V (\?0.29(\?0.29 aVPBRU (\?0.29 (\?V a @V/!R3V @50+Female#XXذ#4 XX<20R  V <#XX 4H#XX2000netundercountisforhouseholdpopulation. B 1990netundercountisforthePESuniversewhichincludednoninstitutional,nonmilitaryGroupQuartersinadditiontothe   householdpopulation.Asaresult,the1990estimatesmaydifferfromtheCommitteeonAdjustmentofPostcensalEstimates R! (CAPE)results.SeeBryantetal.(1992)andThompson(1992). "R The1990HispanicdomainexcludesBlacks,AsianorPacificIslanders,andAmericanIndiansonReservation. " Anegativenetundercountdenotesanetovercount. b# B  C  *TheseestimatesareconsideredtobeunacceptableandweresubsequentlyrevisedasexplainedinU.S.BureauoftheCensus $b (2003). $  **SeeDavis(2001)fordefinitionsofRace/OriginDomains.'r! r%  #XX# 4 XX$XX$1.30  Differencesin1990PostEnumerationSurveyand2000Accuracyand &"" CoverageEvaluation#XX 4# '6#(#(# The1990PESand2000A.C.E.werebasedonthesamemethodology,buttherewere )6% differences.Themajordifferencesareasfollows: *"& " *"! !  -S` XX- *[2o"3  0    Thesamplewas166,000housingunitsin1990dB  ԍ(66,p1049,R2)C  dB  C  and300,000in2000.B  d(67, 8  p1,R1)dC   [݌,'(#(# h     ь! !  " *"! !   *2o"3  0    In1990theuniversewashousingunitsandnoninstitutionalnonmilitarygroupquarters.  Theuniversefor2000washousingunitsonly.dB  ԍ(67,p1,R2)C  d *݌(#(# Ќ! !  " *"! !   *R2o"3  0    In1990largeblocksubsamplingwasaclericaloperation.In2000housingunitmatching  wasconductedbeforetheinterviewingallowingthelargeblocksubsamplingtobedone t bycomputer.dB  ԍ(67,p1,R3)C  d *R݌`(#(# Ќ! !  " *"! !   *]2o"3  0    Theinterviewwasapaperoperationin1990andin2000theinterviewwascomputer L  assisted.dB  ԍ(67,p1,R4)C  d *]݌8 (#(# Ќ! !  " * "! !   *2o"3  0    In1990thePsamplewasthecurrentresidents %#  3      ׀andtheCensusBureaumatchedthe $ t inmoverstotheirCensusDayaddress.In1990thesearchareawasthesampleblocksand  ` oneringofsurroundingblocksinurbanareas,tworingsofsurroundingblocksinrural  L  areas,andinalargerareaofblocksforthemostruralareasofthecountry.In2000the  8  PsamplewastheCensusDayresidentsandtheywerematchedtothecensus $  enumerationsintheblockclusterandsurroundingblocksforselectedclusters.(Thisis   referredtoastheTargetedExtendedSearch.)dB  ԍ(67,p2,R5)C  d *C݌ (#(# Ќ! !  " * "! !   *2o"3  0    In1990allwholehouseholdP-samplenotmatchedpeopleweresentforafollowup   interview.In2000P-samplenotmatchedpeoplefromhousingunitsinterviewedwith   householdmemberswerenotsentforfollowup.dB  ԍ(18,p52,R4)C  d *e݌p(#(# Ќ! !  4 XX &  2.DemographicAnalysis#XX 4;#4 XX:ComparisonwithMarch2001A.C.E.Coverage H Estimates #XX 4# \ DemographicAnalysis(DA)isawell-developedtoolforevaluatingpopulationcoverage.The \ DAis']ananalyticapproachthathasbeenextensivelyusedattheCensusBureautomeasure H coverageofthenationalpopulationineverycensussince1960(SiegelandZelnik,1966;Fayet. 4 al.,1974,1988;Robinsonet.al.,1993;Robinson,March2001.)  p DemographicAnalysisrepresentsamacro-levelapproachforestimatingthenetundercountby H comparingaggregatesetsofdataorcounts.Thedemographicapproachdiffersfundamentally 4 fromthesurvey-basedA.C.E.ThetraditionalDApopulationbenchmarksaredevelopedforthe   censusdatebyanalyzingvarioustypesofdemographicdataessentiallyindependentofthe   census,suchasadministrativestatisticsonbirths,deaths,authorizedinternationalmigration,and   Medicareenrollments,aswellasestimatesoflegalemigrationandnetunauthorizedimmigration. ! ThedifferencebetweentheDemographicAnalysisbenchmarksandthecensuscountprovidesan "  estimateofthecensusnetundercount.DividingthenetundercountbytheDAbenchmark l#! providesanestimateofthenetundercountrate.(Robinson,March2001.)B  d(22,p5,R1)dC   X$" DemographicAnalysisestimateswereinconsistentwithMarch2001A.C.E.estimates.The 0&!$ CensusBureauexpecteddemographicanalysistopositahigherestimateofthetotalpopulation 'l"% thantheMarch2001A.C.E.becauseofthepresenceofcorrelationbias,andthatthetwo (X#& [    estimateswouldgenerallyagreeonthecoverageofcertainpopulations.Instead,theBaseDA  estimates %#  4      ׀werelowerthanboththeCensus2000populationcountsandtheXaGXXXMarch2001#XXXXaG#A.C.E.  estimates.Inresponse,theCensusBureaudevelopedAlternativeDAestimatesbydoublingthe  unauthorizedimmigrationassumedinthe1990sh %#  5      .Doingsoyieldedanumberofforeignbornfor t 2000consistentwiththeMarch2000CurrentPopulationSurvey. %#  6      ׀Still,theAlternativeDA ` estimatednumbersproducedinFebruary2001,weresignificantlylowerthantheXaGXXXMarch2001 L  #XXXXaGq#A.C.E.TheAlternativeDAindicatedthatCensus2000undercountedthepopulationby0.32 8  percent,whiletheXaGXXXMarch2001#XXXXaGS #A.C.E.producedanetundercountestimateof1.15percent. %#  7       $ t TheCensusBureauconcludedthattheinconsistentestimatesofthetotalnationalpopulation  L  derivedfromoneormoreofthreeexplanations:  8  "0  Allavailable1990censusdata,includingthecensusresults,the1990coverage   measurementsurvey,andthe1990DAestimates,significantlyunderstatedtheNations   population,butCensus2000foundthispreviouslyunenumeratedpopulation. (#(# "0  DAunderestimatedpopulationgrowthbetween1990and2000.p(#(# "0  TheMarch2001A.C.E.overestimatedtheNationspopulation.H(#(# Furtherresearchondemographicanalysisfocusedontwomaintopics:internationalmigration  p andmeasurementofvitaleventslikebirthsanddeaths.(ESCAPII,2001.)B  ԍd(1,p4,R1)dC  d  \ d &  4 XX2.1InternationalMigration#XX 4@#  4 TheCensusBureauregardedtheinternationalmigrationassumptionsasthemostuncertain 4 componentinthe''4demographicanalysisestimatescompletedbyMarch1,2001.Researchafter   March1,2001focusedprimarilyonthoseinternationalmigrationcomponentsthatarelesswell   measured(e.g.,emigration,temporarymigration,andunauthorizedmigration).Italsoincluded  researchintolegalimmigrationandthedemographiccharacteristicsofmigrantsusedinthe  Oa    March2001DAestimates.  Partoftheanalysisinvolveddiscussionswithindependentexpertsondemographicanalysisand  internationalmigration.ParticipantsofaMarch20,2001meetingexplainedhowtheDA t estimatesdifferedfromtheXaGXXXMarch2001#XXXXaGm#A.C.E.estimates,anddiscussedhowtoprioritize ` short-termandlong-termresearchactivities.Attendeesincludedexpertsfromthestatistical L  community,academia,stateagencies,theCensusBureausadvisorycommittees,professional 8  organizations,andinternationalorganizations.Theseexperts,almostunanimously, $ t recommendedfocusingoncomponentsofinternationalmigrationbecauseoftheuncertaintyof  ` associatedassumptionsandestimates.  L  TheCensusBureausoughtexperthelponSeptember24,2001,aftercompletingtheoriginal $  researchactivities(validationofthe1990estimatesandupdated2000estimates)thatproduced   therevisedDAestimates.Althoughtheseexpertsgenerallyagreedwiththemethodsusedto   calculatecomponentsofinternationalmigration,theyhadconcernsabouttheassumptions   regardingtheundercountofinternationalmigrants.Specifically,theybelievedtheundercount   assumptionof15percentforunauthorizedmigrants,whichtheRevisedDAincorporated,was p toohigh,especiallygiventheMarch2001A.C.E.undercountsforotherhard-to-enumerate \ groups.Inaddition,theyurgedrenamingtheresidualmigrantcategoryastheresidual H foreign-born,orseparatingtheresidualforeignbornintoknowncomponents( quasi-legal 4 migrants)andtheimpliedunauthorizedmigrantpopulation.Subsequentsensitivityanalysis  p incorporatedbothofthesesuggestions.  \ Thesensitivityanalysisofassumptionsaboutvariouscomponentsoftheforeign-bornpopulation 4 showedthatthetotalnumberofforeignborndidnotvaryenoughtohavemucheffectonthetotal   populationDAestimate.Forexample,thelowerboundassumptionof3.3percentnet   undercountoftheforeign-bornequatedtoapopulationof281.3million,ormorethanthree  millionpeoplelowerthantheMarch2001A.C.E.totalpopulation.Theupperboundassumption  of6.7percentwasconsistentwitha282.5millionpopulation,whichisstillmorethantwo  millionlowerthantheMarch2001A.C.E.totalpopulation.TheseresultsledtheCensusBureau l toconcludethattheRevisedDAwasanappropriatebenchmarkforassessingCensus2000and X  theXaGXXXMarch2001#XXXXaG!#ԀA.C.E.estimates.(ESCAPII,2001.)dB  ԍ(1,pp4&5,R2)dC   D!  4 XX2.2&  MeasurementofVitalEvents#XX 4#  #l! OtherresearchexaminedtheremainingassumptionsunderlyingtheDAcomponentsofchange, %l # '#includingthebirth,death,andMedicarecomponents.Althoughdeathandtheelderlypopulation &X!$ sizeestimatesdidnotchangemuch,thisresearchchangedtheestimatesofhistoricalbirthsby &D"% revisingtheassumptionsaboutregistrationcompletenessofbirthssince1968.ThepreviousDA '0#& estimatesassumeda99.2percent(the1968level)registrationofallbirthsinyearssince1968 ($' (thelastyearoftestingbirthregistrationcompleteness).FortheRevisedDAestimates, )%( registrationcompletenesswasassumedtograduallyreach100percentby1985(thefirstyear *%) natalitystatisticswerereportedelectronicallyfromalltheStates),andremainedat100percent +&* _R 9   through2000.Thisrevisionloweredtheestimatednumberofbirthsfor1968-2000by715,000,  whichloweredtheRevisedDAestimateofthetotalpopulationin2000bythesamenumber.  (Robinson,October2001.)B  d(1,pp5&6,R3)   ЎdC     &  4 XX2.3ResultsofRevisedDA#XX 4##  ` TheresearchundertakenbetweenMarchandOctoberallayedtwofundamentalconcerns:(1)'t#the `  AlternativeDAdidnotcapturethefullgrowthofthepopulationbetween1990and2000,and(2) L  the1990DAwaslowerthanthetruepopulation.Theresearcheffectonimmigration,births,and 8  deathsledtoRevisedDAestimates,producedinSeptember2001,whichwereslightlydifferent $ t  fromtheAlternativeDA.TheinconsistencybetweentheAlternativeDAandtheMarch2001 `  A.C.E.estimatesdidnotresultfromunexplainedproblemsinDA.TheseresultsledtheCensus L  BureautoconcludethattheMarch2001A.C.E.overestimatedtheNationstotalpopulation. 8  TheRevisedDAloweredtheestimatednetundercountratesfrom1.85to1.65percentin1990,   andfrom0.32to0.12percentin2000,butdidnotaltertheDAfindingthattheestimatednet   undercountratein2000wassubstantiallylowerthanin1990.(Robinson,October2001.)The  RevisedDAcontinuedtoestimatealowernetundercountthantheMarch2001A.C.E.,andwas  veryclosetotheAlternativeDAestimateusedinMarch.TheRevisedDAestimatedanet p undercountof0.3million,or0.12percent,comparedwiththeMarch2001A.C.E.estimateofa \ netundercountof3.3million,or1.15percent.PopulationtotalsfromtheBaseDA,Alternative H DA,andRevisedDA,alongwiththeCensus2000countsandtheMarch2001A.C.E.estimates, 4 areshowninTable2.  p *56Fdd 4% 4%      (#(#,P",2","",P""+  9/HH 1 d9  & 8 Table2:ResidentPopulationTotalsfromCensus2000,DemographicAnalysis,andthe )y March2001A.C.E.:April1,2000 fVVU0e 1 d 1 df,  0  Source ]VVL4;VVd d]  TotalPopulation ZP5;# 1 VVd (Z 9/!X 9BaseDA `V!X )TA 279598121)TA`279,598,121 _UGX" )TA 279598121 )TA _ QG,X 3 (  (Q 9/!l! 9Census2000 `V!l" R(ưA 281421906R(ưA`281,421,906 _UGl#" R(ưA 281421906 R(ưA _ LB,l$3 (  (L 9/! % 9RevisedDA(RevisedRegistrationCompletenessAssumption) `V! & rP˰A 281759858rP˰A`281,759,858 _UG '" rP˰A 281759858 rP˰A _ LB' (3 (  (L 9/!!) 9AlternativeDA(DoubleUnauthorizedImmigration) `V!!* ԰A 282335711԰A`282,335,711 _UG!+" ԰A 282335711 ԰A _ XVG,!,3 (  ( X @V/!"-V @March2001A.C.E.* gVV!".V A 284683782Ag284,683,782 fVUG"/" A 284683782 AV f<20"03 (  V <3,XF` XS` X30 F *Thisestimateisconsideredtobeunacceptableandwassubsequentlyrevisedasexplainedin #0 U.S.BureauoftheCensus(2003).$1F(#F(# '8 ),AsshowninTable3(seeAppendixAforanoteregardinginconsistenciesinraceclassifications a&!3 betweenDA,the2000A.C.E.,andthe1990PES),theRevisedDAimpliedagreaterreduction M'"4 thantheMarch2001A.C.E.inestimatednetundercountinCensus2000comparedwiththe1990 9(#5 census.TherevisedDAreducedtheestimatednetundercountrateby1.53percentagepoints, %)u$6 from1.65percentin1990to0.12percentin2000.Incontrast,theMarch2001A.C.E.estimate *a%7 of1.15percentnetundercountin2000was0.43percentagepointslowerthanthe1.58percent *M&8 estimateinthe1990PES.Additionally,bothDAandtheMarch2001A.C.E.estimateda +9'9 reductioninthenetundercountratesofBlackandNonBlackchildrencomparedwith1990.Both ,%(: _c 9   methodsalsoestimatedareductioninthenetundercountratesofadultBlackmenandwomen.  OaOaNOa.Oa FhOaOaOaOaTherevisedDAandMarch2001A.C.E.estimatescontinuedtodisagree.TheDAfounda  reductionintheestimatednetundercountratesofNonBlackmenandwomeninCensus2000 t comparedwiththeratesofpreviouscensuses.TheMarch2001A.C.E.indicatednochangeora ` slightincreaseinestimatedundercountratesforNonBlackadultsasagroup. L  *dFdP"P2"2""P""P56(#(#,dd,dd,dd,dd,,dd,9dd,dd,dd,1dd,dd,wdd+  8.$ t$ t 0 8 @     & T Table3:PercentNetUndercount,byRace*,Sex,andAge:1990and2000 dB  ԍ(1,p8,R4)(28,p3,R1)C  d _VVN/T  0    0_ @  @ ? GVV6$  #VV G @ RevisedDemographic    Analysis GVV6$x  "VV GPES/A.C.E XN/  " 0  VV 0 X   @ Category bX$ !  @1990@ bB  1990 yE " @1990 @  @@2000@@ B  2000 aWE " @@2000 @@   aPES1990 @6$ "   @March2001   A.C.E.** [VVJ/ " 0    0[3 ffffff@2.05 ffffff@V a bVQ!?3V Q?0.07Q?b0.07 aVPB@3 Q?0.07 Q?V a bVQ!A3V Q@2.64Q@b2.64 aVPBB3 Q@2.64 Q@V a bVQ!C3V {Gz?1.28{Gz?b1.28 aVPBD3 {Gz?1.28 {Gz?V a QG,E3 0  V 0 Q =3! F   =  NonBlack ^T$ G  HzG?1.08HzG? ^V  1.08 ZPB H3 HzG?1.08 HzG?   Z \R! I3  (\ҿ-0.29(\ҿ \-0.29 [QC J3 (\ҿ-0.29 (\ҿ   [ [Q! K3  zG?1.18zG? [1.18 ZPB L3 zG?1.18 zG?   Z [Q! M3  )\(?1.01)\(? [1.01 ZPB N3 )\(?1.01 )\(?   Z QG, O3 0    0 Q 9/!TP   9 \R!TQ  `@37637`@ \0-17 }sCTR `@37637 `@  Q?1.12Q? }1.12 ZPBTS3 Q?1.12 Q?   Z [Q!TT3  HzG?0.54HzG? [0.54 ZPBTU3 HzG?0.54 HzG?   Z [Q!TV3  Gz@2.46Gz@ [2.46 ZPBTW3 Gz@2.46 Gz@   Z [Q!TX3  RQ?1.27RQ? [1.27 ZPBTY3 RQ?1.27 RQ?   Z QG,TZ3 0    0 Q 9/!8[   9 9/!8\   9Male,18+ [Q!8]  ףp= ?1.74ףp= ? [1.74 ZPB8^3 ףp= ?1.74 ףp= ?   Z [Q!8_3  (\?0.17(\? [0.17 ZPB8`3 (\?0.17 (\?   Z [Q!8a3   ףp= ?1.19 ףp= ? [1.19 ZPB8b3  ףp= ?1.19  ףp= ?   Z [Q!8c3  zG?1.43zG? [1.43 ZPB8d3 zG?1.43 zG?   Z XVG,8e3 0    0X @V/!lfV @ @V/!lgV @Female,18+ bVQ!lhV )\(?0.44)\(?b0.44 aVPBli3 )\(?0.44 )\(?V a cVR!lj3V RQ-1.27RQc-1.27 bVQClk3 RQ-1.27 RQV b bVQ!ll3V (\?0.34(\?b0.34 aVPBlm3 (\?0.34 (\?V a bVQ!ln3V )\(?0.44)\(?b0.44 aVPBlo3 )\(?0.44 )\(?V a<20lp3 0  V <9,XF` X,XF` X9 F   Aminussigndenotesanetovercount. pp 0 F 0F(#F(#*SeeappendixAforanoteregardinginconsistenciesinraceclassificationsbetween \q DA,theMarch2001A.C.E.,andthe1990PES.Hr(#(# 0 F 0F(#F(#**Theseestimatesweredeterminedtobeunacceptableandweresubsequentlyrevised 4s asexplainedinU.S.BureauoftheCensus(2003). pt(#(# 0 F 0F(#F(#***Thesefiguresdifferfromthe1.61and1.18percentsquotedinTable1becausethe  \u A.C.E.andDAestimatedifferentpopulations.ThebaseoftheDApercentisthetotal  Hv population,whilethebaseoftheA.C.E.percentisthehouseholdpopulation,which !4w excludesgroupquarters." x(#(# 'TT ?0 F 0F(#F(#B  C  Ԛ (#(# Demographicanalysisprovidedevidencethatcorrelationbias %#  8      ׀wasnotreducedbetween1990 $z and2000.ComparisonsoftheDAandMarch2001A.C.E.sexratios(menper100women) % { showedthatcorrelationbiasinthesurveyestimateswasnotreducedforBlackmenbetween &!| 1990and2000.TheMarch2001A.C.E.sexratiosforBlackadultsweremuchlowerthanthe l'"} expectedsexratiosbasedonDA,implyingthattheMarch2001A.C.E.didnotcapturethehigh X(#~ undercountrateofBlackmenrelativetoBlackwomen.Thesizeofthisbiaswasaboutthesame D)$  TQOa     asinthe1990coveragemeasurementsurvey.(ESCAPII,2001.)dB  ԍ(1,pp6-8,R4)dC  B  C    TheDAestimatesdohaveafewlimitations.First,themajorDAestimatesareavailableonlyat  thenationallevelandonlyfortwobroadracecategories:BlackandNonblack(AllOtherRaces t Combined).B  d(23,p7,R1)dC  ԀAnotherconcernregardingDAestimatesistheuncertaintyofthemeasured ` undercounts.TheaggregateadministrativedataandestimatesusedtoconstructtheDA L  benchmarksarecorrectedforvarioustypesoferrors.Thereareassumptionsinthisestimation 8  process,someofwhichcanbevalidatedandsomeofwhicharebasedonquitelimited $ t information.B  d(23,pp7-9,R2)dC  Third,theracecategoriesintheDAestimateslargelyreflecttheraceassignedin  ` theparticularadministrativerecordatthetimeoftheevent(birth,death,orenrollmentin  L  Medicare).TheDAestimatesofnetundercountarebiasedtotheextentthatpeoplewhoare  8  classifiedasaparticularraceinDA(e.g.,Black)reportedadifferentraceintheA.C.E.B  d(23,pp9&10,R3)dC  Fourth, $  theDAcoversthetotalpopulationwhiletheA.C.E.islimitedtothehouseholdpopulation.The   differenceintheuniverseisthegroupquarters(GQ)population.TheGQpopulationisincluded   intheDAestimates,andcannotbeseparated,buttheGQpopulationisexcludedfromtheA.C.E.   universe.(Robinson,October2001)B  d(23,p10,R4)dC  .   TheCensusBureaushouldcontinuetouseDAasacoverageevaluationtool.Forthe2010 \ Census,theCensusBureaushouldalsoinvestigatewaystomeasureuncertaintyintheDA H estimatesofundercountandtoexpandDAestimatetomorerace/ethnicitygroups. 4  4 XX&  3.March2001A.C.E.:PersonCoverage #XX 4u#  \ '  vWebegintheexaminationofpersoncoveragebysummarizingrecommendationsforsampling  \ andestimationandevaluationstudiesandrecommendationsforpersoninterviewing.Wefollow H thiswithsummarizationsofbasicresultsorevaluationsforspecificerrorsources:erroneous 4 enumerations,censusomissions,balancing,correlation,conditioning,reinstatedlateadditions,   andCensus2000imputations.Foreacherrorsource,wealsoproviderecommendationsfor   futureconsideration.   4 XX &  3.1SamplingandEstimation#XX 4x#     InJanuary1999,theSupremeCourtruledagainsttheuseofsamplingforcongressional "  ' yapportionment.(DepartmentofCommercev.HouseofRepresentatives,525U.S.316,119S. #! Ct.765(1999).)(U.S.CensusBureau,December2002.)ThischangedtheCensusBureaus $" plansforthecoveragemeasurementsurvey.TheA.C.E.,asubsampleoftheIntegratedCoverage l% # Measurement(ICM)survey,replacedtheICM.TheICMwouldhaveproducedestimatesofthe X&!$ populationforeachstatedirectlyfromthestatesamplewithsufficientreliabilityfor D'"% apportionment. TheA.C.E.wasaqualitychecktoevaluatethecensuscoverageandpossibly 0(#& correctfornetcoverage,butcouldnotbeusedforapportionment.(U.S.CensusBureau, )l$' December2002.)B  ԍd(9,pvii,R3)dC  Ԁ *X%( Thetimingofthedecisionagainstsamplingforapportionmentimpactedsamplingandestimation +0'*  OaOa     inthreekeyways.  6r` X(#,XF` X6"0 r TheCensusBureaudidnothavetimetoredesigntheA.C.E.sampletomeettheproduction  schedule,sotheCensusBureaudesignedtheA.C.E.samplebasedontheICMusinga t doublesampling %#  9      ׀approach.B  d(9,pvii,R4)dC  `r(#r(# "0 r Themulti-phasesamplingmeantthattheCensus2000DressRehearsalvarianceestimation 8  procedureshadtobediscardedandnewproceduresresearchedanddeveloped.B  d(9,pvii,  L  R5)dC  $ tr(#r(# "0 r Thestate-basedpost-stratificationplanhadtobediscardedandanationalpost-stratification  L  researchedanddeveloped.(U.S.CensusBureau,December2002.)B  d(9,pvii,R6)dC   8 r(#r(# BelowisasummaryofassessmentsofthemajorA.C.E.sampling(U.S.CensusBureau,October   2002)andestimationstepsinCensus2000(U.S.CensusBureau,December2002)and   suggestionsforthefuture.TheseassessmentswereobtainedbytheDecennialManagement   Divisionthroughdiscussionswithkeyprofessionalstaffinvolvedinsamplingandestimation   planningandimplementation.Exceptforthemissingdatacompensationstep,noformal p evaluationswereconducted. \ Themajorsamplingandestimationstepsforthe2000A.C.E.are:B  d(9,piii,R2)dC   4 "0 r Sampling \r(#r(# "0 r WeighttrimmingHr(#r(# "0 r Missingdatacompensation4r(#r(# "0 r Dualsystemestimation r(#r(# "0 r Syntheticestimation r(#r(# "0 r Varianceestimation.r(#r(#  &  4 XX3.1.1Sampling #XX 4#  ThePsamplecontainedapproximately300,000housingunitsinthe50states(excludingareasof   remoteAlaska)andtheDistrictofColumbia.B  d(11,p3,R4)dC   l! ' &  Forplanningthe2010Census,theCensusBureaushouldconsiderthefollowing D#! recommendations: 0$" "0 r Exploretheprosandconsofadoublesamplingapproachfromtheperspectiveof &X!$ '%Uestimationandfieldconcerns.ConsideriftheCensusBureaushoulddevelopadesign &D"% allowingtheflexibilitytoapplyeitherastateornationaldesigninanticipationofthe '0#& possibleneedtochangedesignsaswasdoneinCensus2000.($'r(#r(#  )%( Oa    "0 r Doadditionalanalysistoobtainbettermeasuresofsizeofcensusblocks,critical  informationforthesampledesign.Improvedmeasuresofsizemakespossiblebetter  controlofworkloads,weights,andvariances.r(#r(# "0 r UsetheCensus2000methodorasimilarlydefinedmethodforsamplingsmallcensus ` blocks,thosewithzerototwohousingunits.Thisresultsinvarianceestimation L  efficienciesandreducestheeffectthesmallblockclustershaveontheestimates.(U.S. 8  CensusBureau,October2002.)B  d(11,pp13&14,R9)dC  $ tr(#r(#  &  4 XX3.1.2WeightTrimming#XX 4#   L  TheCensusBureaudesignedA.C.E.weighttrimmingtoreducesamplingweightsforclusters L  ' thatwouldhaveanextremeinfluenceonthedualsystemestimatesandvariances.Theytrimmed 8  theweightforonecluster.(Mule,AmericanStatisticalAssociation,2001.)B  d(30,p4,R1)dC   $  Forthe2010Census,theCensusBureaushouldconsidertheserecommendations:   "0 r Buildathresholdstandardintotheweighttrimmingprocedure.Ifthetotalweighttobe  trimmedorthechangeinthemeansquareerrorbydoingthetrimmingarebelowagiven p threshold,thentheweighttrimmingwouldnotbeimplemented.\r(#r(# "0 r Controlweightsatthepost-stratumlevelratherthanattheclusterlevel.InCensus2000, 4 onlyoneclusterneededtrimming.However,atthepost-stratumlevel,apost-stratumhad  p severalclusterswithhighweights. \r(#r(# "0 r Scheduleweighttrimmingduringdualsystemestimationinsteadofbeforemissingdata 4 processingtoallowtheCensusBureautotakeintoaccounttheeffectofweighttrimming   ontheDualSystemEstimates.(U.S.CensusBureau,December2002.)B  d(9,pp1&2,R4)dC   r(#r(#  &  4 XX3.1.3MissingData#XX 4X#   MissingdataoccurredintheA.C.E.if,afterallfollowupattempts,thereremainedhouseholds ! 'Lnotinterviewedorhouseholdswithportionsofthepersondatamissing,suchasageorrace. "  Sometimesthemissingitemmighthavebeenthestatusofwhetherapersonmatched,wasa l#! residentonCensusDay,orwascorrectlyenumerated.TheCensusBureauusedstatistical X$" modelstoaccountformissingdata.B  ԍd(1,p13,R13)C  dAsshownbelow,thelevelandpatternofmissingdatainthe D% # March2001A.C.E.wascomparabletothatofthe1990PES.Theeffectofthemissingdataon 0&!$ theoverallMarch2001A.C.E.qualitywassimilartothatexperiencedbythe1990PESand 'l"% documentedinthePstudies.(Macket.al.,1991;Gbur,1991;West,1991.) (X#& B  d(7,p38,R1)C  dAdditionalstatisticalmodelstoaccountformissingdataweredevelopedtoassesstheeffecton (D$' theestimatesofusingalternativemodels.(Keathley,Kearney,andBell,2001.) )0%( Imputeddemographiccharacteristicsusedtoaccountformissingpost-stratificationvariables +'* resultinincreasedclassificationerroraswellassyntheticerrorandpossiblycontributeto  correlationbias.B  d(34,pp1-3,R1)dC  ԀHighlevelsofmissingdata,particularlyformatch,residence,orenumeration  status  %#  10      ,alsoincreasevariance.TheCensusBureaudidnotevaluatehowthistypeofmissing  dataincreasesvariancebecausethemeasureofsamplingvariancelargelypickedupthis t component. ` TwoimportantchangesfortheCensus2000couldhaveaffectedmissingdatarates.First,the 8  levelofmissingdataintheA.C.E.interviewcouldhavebeenhigherbecauseofachangeinhow $ t theCensusBureautreatedmovers.In1990,theCensusBureauonlyneededtointerviewthe  ` currentresidents,whereasinCensus2000,interviewersneededinformationaboutthecurrent  L  (A.C.E.InterviewDay)residentsandtheCensusDayresidents.Ontheotherhand,theA.C.E.  8  eliminatedtheneedtogeographicallycodetheCensusDayaddressofinmovers,thuseliminating $  onepotentialsourceofmissingdata.Second,theComputerAssistedPersonalInterview(CAPI)   instrumentkepttheintervieweronthecorrectsetofquestionsandallowedfortightmanagerial   control.B  d(7,pp39&40,R2)dC     TheMarch2001A.C.E.missingpersondemographiccharacteristicsimputationprograms p operatednearlyidenticallytothoseusedforthe1990CensusPES.(U.S.CensusBureau, \ December2002.)B  d(9,p3,R14)dC  ԀTheMarch2001A.C.E.usedadifferentstatisticalmodeltoaccountfor H missingdataformatchandresidencestatusthanthe1990PES.TheCensusBureaubasedthe 4 1990modelonhierarchicallogisticregression,whilethe2000modelusedthe ImputationCell  p Estimator. h %#  11      ׀Theinputdataandbehavioralassumptionsbetweenthetwomodelsweresimilar  \ butnotidentical. H TheamountofmissingdataintheMarch2001A.C.E.waslow.Thislowlevelminimizesthe   effectofthemissingdataassumptionsonthefinalestimates.TheCensusBureaufound:   "0 r March2001A.C.E.hadhighinterviewrates.Amongoccupiedhousingunits,the  CensusBureauhada97.1percentinterviewrateforCensusDayand98.8percentfor  A.C.E.InterviewDay.Thiscomparesto98.4percent(unweighted)inthe1990PES. l Becauseofthehighresponse,mostofthechangesduetothenoninterviewadjustment X  factorsappliedwereverysmall.Thisresulthelpskeepdownthevarianceofsurvey D! weights.0" r(#r(#  #l!  N TRX3'Avery 5366 File Folder -X3'Avery 5366 File Folder -3'Letter.19  ףp= ?   Z [Q!83'Letter366 File Folder -XdT    "0 r Alowproportion,2.2percent,ofpeoplehadunresolvedresidence.Themissingdata  proceduresassignedanaverageresidentprobabilityof82.6percenttopeoplewith  unresolvedresidentstatus.Asdesigned,thiswaslowerthantheaveragerateamongpeople  withresolvedstatus(98.2percent).tr(#r(# "0 r Only1.2percentofthesamplehadunresolvedmatchstatus,comparedto1.8percentinthe L  1990PES.TheCensusBureauassignedanaveragematchrateof84.3percenttopeople 8  withunresolvedmatchstatus,comparedto91.7percentforthosewithresolvedstatus.The $ t lowrateofunresolvedmatchstatusimpliesonlyasmalleffectontheestimation. `r(#r(# "0 r About2.6percentoftheEsamplehadunresolvedenumerationstatuscomparedto  8  2.3percentinthe1990PES.Theaveragerateofcorrectenumerationforpeoplewith $  unresolvedstatuswas76.2percentcomparedto95.9percentforthosewithresolvedstatus.   (Cantwellet.al.,2001.)B  ԍd(7,pp39&40,R2)dC   r(#r(# "0 r Similartothe1990PES,March2001A.C.E.hadlowratesofmissingdemographicdataas   showninTable4.Therewerefewproblemsgatheringanswerstoallquestionsabout p respondentsinA.C.E.interviewsforthePsampleorfromcensusformsfortheEsample. \ Thissuggeststhatthepost-stratificationresultsaccuratelyreflectedrespondentstrue H characteristics,andshouldhelptoreduceheterogeneity(i.e.thepossibilityfordifferent 4 peoplewithinpost-stratatohavedifferentchancesofbeingcountedinthecensusandinthe  p A.C.E.),sinceimputationdeterminesthepost-strataforonlyasmallnumberofpeople.  \ (Farber,2001.)dB  ԍ(17,p20,R1)C  dHr(#r(#  & $ Table4:March2001A.C.E.and1990PESMissingDataRates(weighted) dB  ԍ(17,p20,R2)C  Ԛd   **Vd ddddddddd,dd,9dd9dddd1dd1ddwddw(#(#*,&dd ,dd , dd ,dd , dd ,dd , dd ,dd ,dd +  +! +XXMissingCharacteristic#XX#XX KA2$t" KMarch2001A.C.E.#XX\#XX 1'$t" 11990PES#XX#4 XX =3$$t" =#XX 4u#XX ( f (Psample @6$z" @Esample @6$z" @Psample @6$z" @Esample#XX#XX <2$z " <#XX>#XXRace WM! ! ffffff?1.4ffffff?W1.4 J@2 "3 ffffff?1.4 ffffff? J ZP! #3  @3.2 @Z3.2 YOA $3  @3.2  @ Y WM! %3 @2.5@W2.5 J@2 &3 @2.5 @ J [Q! '3 '@11.8'@[11.8#XX#XX ZPB (3 '@11.8 '@ Z 9/! )3 9HispanicOrigin ZP!( x* ffffff@2.3ffffff@Z2.3 YOA( x+3 ffffff@2.3 ffffff@ Y ZP!( x,3 333333 @3.4333333 @Z3.4 YOA( x-3 333333 @3.4 333333 @ Y =3!( x.3 =NotAvailable @6$( x/" @NotAvailable#XX_#XX <2$( x0" <Age ZP!!1 333333@2.4333333@Z2.4 YOA!23 333333@2.4 333333@ Y ZP!!33 333333@2.9333333@Z2.9 YOA!43 333333@2.9 333333@ Y ZP!!53 ffffff?0.7ffffff?Z0.7 YOA!63 ffffff?0.7 ffffff? Y ZP!!73 333333@2.4333333@Z2.4#XXF#XX YOA!83 333333@2.4 333333@ Y 9/!!93 9Sex ZP!"(: 333333?1.7333333?Z1.7 YOA"(;3 333333?1.7 333333? Y ZP!"(<3 ?0.2?Z0.2 YOA"(=3 ?0.2 ? Y ZP!"(>3 ?0.5?Z0.5 YOA"(?3 ?0.5 ? Y ZP!"(@3 ?1.0?Z1.0#XXF#XX YOA"(A3 ?1.0 ? Y 9/!"(B3 9Tenure aVP!0$C ffffff?1.9ffffff?a1.9 `VOA0$D3 ffffff?1.9 ffffff?V ` ZP!0$E3V  @3.6 @Z3.6 YOA0$F3  @3.6  @ Y ZP!0$G3 ffffff@2.3ffffff@Z2.3 YOA0$H3 ffffff@2.3 ffffff@ Y ZP!0$I3 @2.5@Z#XX#XX2.5'$ #XX# YOA0$J3 @2.5 @ Y1'%0$K3   1B  C  WhilethemissingdatarateswerelowandtheactualmissingdatatreatmentstheCensusBureau 8&!L usedfortheA.C.E.hadsmallimpactsontheestimates,thetreatmentofmissingdatacanhavea $'t"M largeeffectontheA.C.E.estimatesundercertainassumptions.TheCensusBureauconsidered, (`#N invariouscombinations,sevenbasicmethodsforaddressingthenoninterviewandunresolved (L$O personstatuscomponentsofmissingdataintheMarch2001A.C.E.estimates.TheCensus )8%P BureauusedeachresultingalternativemodeltocomputenewDSEs.Thealternativesconsidered *$&Q showedthechoiceofstatisticalmodeltoaccountformissingdatacanhaveasubstantialeffect +'R ontheresultingestimatesofcoverageerror,causingtheDSEstobeoverorunderstated.The ,'S  jOa 9 9   CensusBureauchosetorepresenttheeffectsofthesealternativemodelsintheformofincreased  uncertaintyintheMarch2001A.C.E.estimates.  TheCensusBureauusedtheDSEsresultingfromthealternativemodelstocalculateameasure t ofvariationsimilartoasamplingerror.Thisevaluationfoundlargenon-samplingvariability ` fromtheuseofalternativemissingdatamodels.Atthenationallevel,theevaluationfoundthe L  overallmagnitudeofthevariationresultingfromallcombinationsofthealternativestatistical 8  modelsusedtoaccountformissingdatatobeabout530,000.Argumentscanbemadethatthis $ t measureunderstatestheactuallevelsofvariationduetomissingdatabecauseitassumeseach  ` alternativewasequallylikely.(Spenceret.al,2002.)B  d(1,pp13&14,R14)(8,p14,R1)dC    L  TheCensus2000unresolvedenumerationstatusrateswereslightlyhigherthanthosein1990, $  butwerenotviewedashighenoughtocausemajorconcern.(Liu,Jones,andFeldpausch,2001.)   Thealternativemodelanalysisindicatedthatmissingdatahadalargereffectthananticipated.   ThiscouldhavebeenduetochangesinthemethodsforincorporatingmoversintotheDSE,orto   amorediversesetofalternativemodelsthanusedinevaluationofthe1990missingdata   procedures.B  ԍ(ESCAPII,2001.)d(1,p14,R15)dC  Ԁ(Macket.al.,1991.) p Forthe2010Census,theCensusBureaushouldconsiderthefollowing: H "0 r IftheCensusBureauexpectslownoninterviewratesin2010,thenusetheA.C.E.2000  p noninterviewadjustmentmethodology.Ifnot,theninvestigateandconsideralternativesfor  \ noninterviewadjustment.B  ԍd(9,p2,R1)dC  ԀIftheCensusBureautakesthesuggestioninSection3.2to H minimizeproxyinterviews,examinetherelativetradeoffbetweendealingwithproxiesand 4 withmissingdata. r(#r(# "0 r WritetheA.C.E.characteristicimputationprogramfromscratchratherthenbaseingitona  previouslywrittenprogramwhichwasnotwellunderstood.r(#r(# "0 r Imputemissingcharacteristicsusingthesamemethodologyimplementedforthecensus, l especiallyfortheEsampletoincreaseconsistencyofdemographiccharacteristics.X r(#r(# "0 r LinktheP-andE-samplefilesforeachblockclusterandusetheavailableinformationfor 0"  matchedunitsforcharacteristicimputation,especiallyiftheP-samplepersonsaremissing #l! thecharacteristics.Whenpeoplematch,considerusingthecensusinformationtoimpute $X" forthosemissingcharacteristicstoreduceinconsistenciesbetweenthetwosamples.$D #r(#r(# "0 r Designaflexibleimputationplansothatpriortoimputing,theCensusBureaucanexamine &"% thedataanddetermineanoptimalimputation.'#&r(#r(# Ѐ"0 r Formissingage,imputeanage(anumber),ratherthananagecategory.Imputinganumber )$( ratherthanacategoryprovidesmoreflexibilityforestimationandlaterevaluations.|*%)r(#r(#  h+&* Oa    "0 r Whileformissingtenure,race,andHispanicorigin,thenearest-neighborhotdeckappears  toworkwell,evaluateifalternativescouldprovideimprovements.r(#r(# "0 r Modelalternativesforseximputationtoseeifimprovementsarepossible.dB  ԍ(9,pp3&4,R6)C  tr(#r(# d"0 r ForthePsample,determinetheprobabilitiesofresidentstatusandmatchstatusjointly. L  Thishelpsaccountforthedependencebetweenthetwo.8 r(#r(# dd"0 r Whengatheringinformationtobeusedtoassignprobabilitiesforunresolvedpersonstatus,  ` concentrateoninformationpertainingtotheinterviewoperationsandfieldprocedures(i.e.  L  whatwenton),andlessondemographicinformation(i.e.race,ethnicity,tenure,etc.).In  8  Census2000,theformerinformationbetterclassifiedcasesforthepurposesofassigning $  probabilities.d r(#r(# d"0 r Evaluatealternativessuchaslogisticregressionandrelatedsoftwaretoassignprobabilities   forunresolvedpersonstatus,butbecarefulnottoover-model.dB  ԍ(9,p5,R2)dC  (U.S.CensusBureau,   December2002.)Thereisthepotentialtoimproveprobabilityestimates,butalsothe p potentialtoincreasevariability.\r(#r(# "0 r Useconsistentcodingandeditingforraceandotherpost-stratificationvariablesinA.C.E. 4 andcensustoincreaseconsistency. pr(#r(#  &  4 XX3.1.4DualSystemEstimation#XX 4 #  H Dualsystemestimationmeasuredthedegreeofpopulationnetcoverageerrorobservedduring H the'censusenumeration.Itaccomplishedthisbycomparingthecensusenumerationresultsto 4 A.C.E.resultstocalculatedualsystemestimatesseparatelybypost-stratabasedongeography   anddemographicvariables.PopulationsnotincludedindualsystemestimationwereGroup   Quarterspersons,Service-BasedEnumerationpersons,andpersonsinRemoteAlaska.B  ԍd(9, D! pp5&6,R7)dC    TheCensusBureaudevelopedpost-stratainreferencetoexperienceinallpreviouscensuses,but   especiallythecensusesof1980and1990.Inthosecensuses,sincethenetundercountwas l! significantlylargerthanzero,theybelievedgrossomissionswasthedominanterror,withgross X"  erroneousinclusionsbeingsmaller.TheCensusBureaualsoconcludedthatthedeterminantsof D#! netundercountwouldprimarilyfollowsocio-economicgroupings. 0$"  r  `  %l # Thefactorsthatcausederroneousinclusionsprobablydrivethedifferentialerrorsby &X!$ demographicgroupandgeographicareaasmuchasthosethatcausedomissions.The &D"% CensusBureauconsideredpoststratifyingseparatelyforomissions(Psample)anderroneous '0#& enumerations(Esample).TheCensusBureaurejectedthisapproachbecauseofthetight ($' schedule.(Hogan,2002.)dB  ԍ(12,pp3&4,R1)C  d )%(     *%) dO    & L  For2010post-stratificationplanning,theCensusBureaushouldconsiderthese  recommendations:  "0 r Post-stratifythePandEsamplesseparatelytobetteraccountforbothomissionsand t erroneousinclusions.(U.S.CensusBureau,December2002.)B  ԍd(9,p7,R8)dC  `r(#r(# "0 r Identifypost-stratathatbetteraccountforvariabilityofsocioeconomicgroupings.8 r(#r(# "0 r UsegeneralizedDSEsinsteadofpostratification.Thisapproachuseslogisticregression'Lto  ` estimateprobabilitiesofinclusioninthecensusandofcorrectenumerations.  L  Post-stratificationmakesuseofestimationcellswhichmustbeofsufficientsize.  8  GeneralizedDSEsarenotlimitedbysizeconstraintsandthusoffergreateropportunityto $  reducebiases. r(#r(# &  For2010DSEimplementationplanning,theCensusBureaushouldconsiderthefollowing   recommendations:   "0 r Considerusing Procedure Btohandlemovers.In2000theCensusBureauused \  ProcedureC'withafewexceptionswhereweused ProcedureA. ProcedureCand L  ProcedureAmatchthenonmoversandoutmoversattheCensusDayaddressandwithin 8 thesearcharea. ProcedureCand ProcedureAratherthan ProcedureBwereusedin $t 2000becauseitiseasiertomatchwithinthesearcharea.Oneproblemwith ProcedureC ` and ProcedureAwasthatinterviewswithwholehouseholdsofoutmoverswereproxy L interviews. ProcedureBwasusedin1990.Forthisprocedure,thenonmoverswere 8 matchedtotheCensusDayaddressandsearcharea.Theinmoverswerematchedtotheir $ CensusDayaddressrequiringcollectingtheCensusDayaddressfortheinmover,obtaining  thecensusgeographyfortheaddress,andmatchingtothataddressandsurroundingblocks.  Thisinmovermatchingwastimeconsumingbecausethecensusquestionnaireswere  printedfrommicrofilmforclericalmatching.Theentiremovermatchingprocesscouldbe  improvedin2010sincenamesarecapturedfortheentirecountry.pr(#r(# "0 r Developimprovedmethodstodetecterroneousenumerationsandtoincorporateduplicates H! intothecoveragemeasurementsurveyestimates.(SeeSection3.3,Erroneous 4"  Enumerations,IncludingDuplicates.) #p!r(#r(# "0 r Useexternaldatasources(e.g.administrativerecords,DemographicAnalysis,American $H # CommunitySurvey)toimprovecoverageestimates.%4!$r(#r(# "0 r Includegroupquartersinfuturecoveragemeasurementsurveystoimproveundercount ' #& estimates.B  ԍd(9,pp7&8,R9)dC  Ԁ(U.S.CensusBureau,December2002.)(#'r(#r(#   )$( _c    &   4 XX3.1.5SyntheticEstimation#XX 4#   ThelastoperationintheMarch2001A.C.E.estimationprocess,syntheticestimation,provided  population'estimatesforsmallgeographicareassuchasblocks,tracts,counties,and  congressionaldistricts.TheCensusBureauformedthesesmallareaestimatesbyapplying  coveragecorrectionfactors(i.e.theratiooftheDSEtothecensuscountforeachpostratum)to t  thecensuscountsatthedifferentgeographiclevels.Forexample,theCensusBureauformeda `  block-levelsyntheticestimatebydistributingapost-stratumsdualsystemestimatetoblocks L  proportionaltothesizeofthepost-stratumspopulationwithintheblock.Foruseinallcensus 8  dataproducts,theCensusBureauconstructedrounded,adjustedsyntheticestimatesatthe $ t  tabulationblocklevel %#  12      .Datausersthenproducepopulationestimatesforanygeographicareaof `  interestbyaggregatingblocks.PopulationsnotincludedinsyntheticestimationwereGroup L  Quarterspersons,Service-BasedEnumerationpersons,andpersonsinRemoteAlaska.B  ԍ   Ўd(9,p9,R10)C  d 8  ForCensus2010,theCensusBureaumaywanttoconsideramodelingapproachtoDSEs.Inthis   case,theCensusBureauwouldnotneedaseparatesyntheticestimationprocedure.B  d(9,p11,R11) \ ЎdC     (U.S.CensusBureau,December2002.)   &  4 XX3.1.6VarianceEstimation#XX 4L# p  TheCensusBureauexpectedthesamplingvariancesandcoefficientsofvariation(CV)tobe p lowerfortheMarch2001A.C.E.comparedtothe1990PESbecause: \ "0 r ThehousingunitsamplesizefortheA.C.E.wasalmostdoublethatofthePES 4 (approximately300,000'@versusapproximately165,000). pr(#r(# "0 r Bettermeasuresofpopulationsizewereavailableduringsampleselectionofclusters.Hr(#r(# "0 r Samplingweightswerelessvariable. r(#r(# Asexpected,theimprovementsledtomuchsmallersamplingvariances.Theactualreduction   waslargerthanthe25percentexpectedreductionduetotheincreaseinsamplesize.B  ԍd(13,p1,R1)dC  ԀThisisseen ! fromTable5.Also: "  "0 r XaGXXXTheCVsdeclinedforforty-sevenstates,withanaveragereductionof36.8percent.B  ԍd(13, %0!$ p1,R2)dC  X$"r(#r(# #XXXXaG#"0 r XaGXXXAttheCongressionalDistrictlevel,themedianCVdecreasedbyabout40percent,from 0&!$ 0.499percentto0.297percent.'l"%r(#r(# #XXXXaG#"0 r ThemedianCVdecreasedbyroughly50percent,from0.629percentto0.314percent,for (D$' placeswithacensuspopulationgreaterthan100,000.)0%(r(#r(#  *&) Oa    "0 r ThemedianCVdecreasedbyabout40percent,from0.510percentto0.310percent,for  countieswithacensuspopulationgreaterthan100,000.(Starsinicet.al,2001.)B  ԍd(13,p2, ` R3)C  d  r(#r(# Table5:DistributionofCVs  %#  13      ׀forPopulationEstimatesbyGeographicalAreaforMarch t 2001A.C.E.and1990PES  ` *~d d&dd &dd  dd dd  dd dd  dd dd dd V(#(#,,X,X,,,,,,X,X,+  AVVV.L L  AXXArea d   UVVD2 d N1VVV USource UVVD2d " ppVV UNumber UVVD2d  " ppVV UMeanSize UVVD2d  " ppVV UMean d   CV [VVJ2 d " VV P[Marginof d   Error* WVJ8 d" PVV WDistributionofCVs aVN/d "   V a @-, |V @ @-, | @ @-, | @ @-, | @ F3, | PF @V/!, |P @Minimum @V/!D "V @Q1 @V/!D "V @Median @V/!D "V @Q3 @V/!D "V @Maximum ^VK,D "   V ^ State** J@2$ tpp8V JA.C.E. YO!$ t"  I@51I@Y51 }s@$ t"  I@51 I@ 4KUA55820354KUA}5,582,035 ]SE$ t" 4KUA5582035 4KUA ]0.310% \R!$ t" @28506@\28,506 [QC$ t" @28506 @ [0.159% 9/!$ t" 90.220% 9/!$ t" 90.240% 9/!$ t" 90.378% 9/!$ t" 90.804% WVVH,$ t"    W 3) <VV 3PES YO! T "  I@51I@Y51 }s@ T!"  I@51 I@ @RA4955153@RA}4,955,153 ]SE T"" @RA4955153 @RA ]0.449% \R! T#" @36623@\36,623 [QC T$" @36623 @ [0.322% 9/! T%" 90.369% 9/! T&" 90.406% 9/! T'" 90.496% 9/! T(" 90.933% ZVG, T)"    Z 9/! 4*V 9 9/! 4+" 9 9/! 4," 9 9/! 4-" 9 ?5! 4." P? ?5' 4/"P ? 9/! 40" 9 9/! 41" 9 9/! 42" 9 9/! 43" 9 ^VK, 44"    ^ Congressional   5 Districts*** J@2t 6:V JA.C.E. ZP!  7" 0{@4350{@Z435 }sA  8" 0{@435 0{@ ^#A653103^#A}653,103 \RD  9" ^#A653103 ^#A \0.330% [Q!  :" @3546@[3,546 ZPB  ;" @3546 @ Z0.156% 9/!  <" 90.250% 9/!  =" 90.297% 9/!  >" 90.375% 9/!  ?" 90.948% WVVH,  @"    W 3) @VV 3PES ZP! A" 0{@4350{@Z435 }sA B" 0{@435 0{@ ޯ!A579567ޯ!A}H+ 4 <DL!r` X(#H579,567 \RD C" ޯ!A579567 ޯ!A \0.557% [Q! D" @5309@[5,309 ZPB E" @5309 @ Z0.299% 9/! F" 90.420% 9/! G" 90.499% 9/! H" 90.628% 9/! I" 92.007% ZVG, J"    Z 9/! KV 9 9/! L" 9 ?5! M" ? E;' N" E E;' O" E E;' P" E E;' Q" E E;' R" E E;' S" E E;' T" E dVQ2 U"    dPlaces> d V 100,000**** J@2d W=V JA.C.E. ZP!d X" n@245n@Z245 }sAd Y" n@245 n@ t:A315037t:A}?+ 4 <DL!X?315,037 \RDd Z" t:A315037 t:A \0.343% [Q!d [" @1776@[1,776 ZPBd \" @1776 @ Z0.213% 9/!d ]" 90.283% 9/!d ^" 90.314% 9/!d _" 90.361% 9/!d `" 91.435% WVVH,d a"    W 3),| aVV 3PES ZP!D b" `h@195`h@Z195 yAD c" `h@195 `h@ T|A335637PT|A335,637 JD d"P T|A335637 T|A qn{?0.00673qn{?0.673% KD e" qn{?0.00673 qn{?  @3718P @3,718 HD f"P  @3718  @ uim?0.00363uim?0.363% KD g" uim?0.00363 uim? N ^u?0.536N ^u?0.536% ID h" N ^u?0.536 N ^u? l`y?0.629l`y?0.629% ID i" l`y?0.629 l`y? 7ܘ~?0.7477ܘ~?0.747% ID j" 7ܘ~?0.747 7ܘ~? vݰm?1.702vݰm?1.702% VoTD k" vݰm?1.702   vݰm?  9/!$t lV 9 9/!$t m" 9 9/!$t n" 9 9/!$t o" 9 ?5!$t p" P? ?5'$t q"P ? 9/!$t r" 9 9/!$t s" 9 9/!$t t" 9 9/!$t u" 9 ^VVK,$t v"    ^Counties> Tw 100,000**** J@2x@VV JA.C.E. ZP!Ty" `@524`@Z524 }sATz" `@524 `@ A409345A}?+ 4 <DL!X?409,345 \RDT{" A409345 A \0.368% [Q!T|" b@2481b@[2,481 ZPBT}" b@2481 b@ Z0.201% 9/!T~" 90.274% 9/!T" 90.310% 9/!T" 90.405% 9/!T" 91.498% YVVVH,T"    Y :V)VVV :PES aVP!4"V |@458|@a458 VyA4" |@458 |@V DsA400593PDsA400,593 VJ4"P DsA400593 DsAV s.Ueu?0.00534s.Ueu?0.534% VK4" s.Ueu?0.00534 s.Ueu?V ~@3519P~@3,519 VH4"P ~@3519 ~@V +eXg?0.00285+eXg?0.285% VK4" +eXg?0.00285 +eXg?V +ٱq?0.432+ٱq?0.432% VI4" +ٱq?0.432 +ٱq?V ZӼt?0.510ZӼt?0.510% VI4" ZӼt?0.510 ZӼt?V 5x?0.5915x?0.591% VI4" 5x?0.591 5x?V ;X0_?1.483;X0_?1.483%dZX4" ;X0_?1.483   ;X0_?V d#XXk #XX*MarginofErroriscalculatedas1.645'standarderrorofthepopulationestimate.  ** Stateincludesall50statesandtheDistrictofColumbia. \ ***103rdCongressionalDistrictsforthePES;106thCongressionalDistrictsfortheA.C.E..DoesnotincludetheDistrictof  \ ColumbiaorPuertoRico.   ****Countiesandplaceswithcensuscountsofmorethan100,000intherespectivecensuses,2000forA.C.E.and1990for l PES. l #XX?<#4 XXB  d(13,p1ofappendix,R4)dC  #XX 4>#  TheCensusBureauusedreplicationmethodstoestimatethevarianceduetoA.C.E.andPES  samplingandestimation.UnlikethePES,theMarch2001A.C.E.replicatevariancesofthe  censusestimateshadthreecomponents:  "0  Varianceduetothemulti-phasesamplingofblockclustersfortheA.C.E.h(#(# "0  VarianceduetosamplingfortheTargetedExtendedSearch.T(#(# "0  VariancefromestimatingthemissingdatainA.C.E.B  ԍd(9,p11,R12)dC  Ԁ@(#(# Thevariancecomputationaccountedforsomeofthecomponentsofvarianceduetomissing !h data,butitisunknownwhetheritlargelyaccountedforthevarianceduetomissingdatasincethe "T computationdidnotincludethevariancecomponentduetotheselectionofstatisticalmodelto "@ accountformissingdata.ThealternativemodelsevaluationconductedbyKeathley,Kearney, #, andBell(2001)indicatedthiscomponentmaybelarge.B  ԍd(32,p1,R1)dC   $   h.Oa  . .  TheCensusBureaushouldconsiderwhethertheuseofcomplexvariancemethodswouldbe  morebeneficialinproductionorinanevaluationoftheproductionvariances.Theyshouldalso  considerdevelopingconfidencemeasuresforA.C.E.thatreflectsyntheticerrorandother t nonsamplingerroraswellassamplingerror. `  4 XX&  3.2A.C.E.PersonInterviewing #XX 40F# 8  Byrne,Imel,Ramos,andStallone(2001)examinedtheA.C.E.personinterviewoperation.The 8  '8 OFinterviewingoperationhadtwophases:telephoneandpersonalvisit.Thepersonalvisitalso $ t  usedaNonResponseConversionOperation(NRCO)totryconvertingthenoninterviewsbyusing `  thebestinterviewers. L  &  Datesoftheoperation: $  *r XX*"0 ' TelephonePhase0 '(#'(#0 (# (#0h(#(#April24,2000-June13,2000 h(#h(# 'IH"0 ' PersonalVisitPhase0'(#'(#0h(#(#June19,2000-September11,2000h(#h(# "0 ' NonresponseConversion0'(#'(#0h(#(#July27,2000-September11,2000dB  ԍ(68,p6,R1)C  dh(#h(# TheA.C.E.plannedallinterviewactivitiestoendonSeptember1,2000.However,onelocal \ censusoffice,Hialeah,Floridarequiredmoretimetocompletethecensusdatacollection H operations.ThisresultedinadelayforthesubsequentA.C.E.personinterviewinginHialeah 4 untilAugust18,2000-September11,2000.Allotherofficesfinishedinterviewingonschedule,  p September1,2000.dB  ԍ(68,p6,R2)C  dXaGXXX  \ The2000A.C.E.didnotuseapaperformasusedinthe1990PES.TheCensusBureauused 4 computerassistedpersonalinterviewing(CAPI)software.   #XXXXaGL#Byrne,Imel,Ramos,andStallone(2001)dB  ԍ(68,p1-2,R3)C  dprovideresultsoftheinterviewingoperation:  "0 ' Almostall(99.9percent)interviewsresultedinasatisfactoryoutcome.TheCensusBureau   classifiedonly0.12percentofallinterviewsaseitherrefusal,languagebarrier,orno l! knowledgeablerespondentnoninterviews.X" '(#'(# "0 ' TheCensusBureaucompleted29percentofthetotalA.C.E.workloadduringthetelephone 0$" phase.Asaresult,theinterviewphaseendedwithmuchlesstimetranspiringbetween %l # CensusDayandthedayoftheinterview,potentiallyreducingrecallbias.The &X!$ CensusBureauclassifiedover99percentofthetelephonecasesascompleteorpartial &D"% interviewsconductedwithahouseholdmember.'0#&'(#'(# "0 ' TheCensusBureauclassified84percentofthepersonalvisitinterviewsaseithercomplete )%(  orpartialinterviewsandfound14percenttobevacantonInterviewDay.Thisaccountsfor *%) Oa    98percentofthepersonalvisitworkload.Oftheremaining2percent,1.9percentwere  nonexistentunitsonInterviewDayand0.2percentwerenoninterviews.'(#'(#  "0 ' Interviewersconverted70.8percentofthecasessentforNRCOfromthetelephoneand t personalvisitphasestocompleteinterviewsand14.1percenttopartialinterviews.Ofthe ` remainingcases,11.4percentconvertedtovacantunitsand1.5percenttononexistentunits. L  Only2.2percentoftheNRCOcasesfinishedasrefusals.8 '(#'(# "0 ' Automatingtheinterviewingenhancedthequalityofdatacapturedintheinterviews,  ` expeditedtheturnaroundtimeforreassigninginterviewsandprovidingfeedbacktothe  L  interviewers,andinstilledtheinterviewerswithasenseofprofessionalismandpurpose. 8 '(#'(#  "0 ' TheQualityAssurance(QA)operationhelpedkeeptherateoferrorlowandindicatedahigh   levelofdataquality. '(#'(# TheQAofpersoninterviewinghelpedensurecorrectresultsfromthetelephoneandpersonal   visitphasesoftheoperation.TheQAsamplewasfromtwosources:afivepercentrandom p sampleofthetotalcaseloadandtargetedcasesselectedbytheQAsupervisorsbecausetheywere \ likelytocontaininaccurateinformationorinsufficientdataquality.Only190casesfailedthe H QA.Forallsuchcases,theCensusBureauobtainedandusedareplacementinterviewinthe 4 survey.TheCensusBureaueffectivelyweededoutseveralinterviewerswhoseworkhad  p egregiouserrors.TheCensusBureauaccomplishedmorebytargetingforproblematiccasesthan  \ throughrandomlysamplingcases.Thelowfailurerateintherandomsamplemeanttheerrorsin H personinterviewingwereundercontrol.dB  ԍ(68,p2,R4)C  d 4   HighlightsoftheQAresults:   "0 ' Theoverallfailurerateforthetargetedcases(0.85percent)wasdramaticallydifferentfrom  therandomlyselectedcases(0.13percent).Thispatternheldforbothtelephoneand  personalvisitinterviews,suggestingtargetingwaseffectiveinidentifyingcaseslikelytofail l thequalityassurance.dB  ԍ(68,p2,R5)C  dX '(#'(# "0 ' Becauseofthedataeditsandautomatedskippatterns,aswellasthequickturnaroundtime 0"  forcasestogetassignedandcompletedinQA,automatingboththeoriginalperson #l! interviewandtheQAreinterviewenhancedtheoverallqualityandefficiencyoftheperson $X" interviewoperation.dB  ԍ(68,p2,R6)C  d$D #'(#'(# Wolfgang,Byrne,andSpratt(2003)examinedthecharacteristicsofpeopleandhouseholdsby &"% respondenttype(i.e.,interviewwithahouseholdornonhouseholdmember).Amongtheoriginal '#& A.C.E.personinterviews,theagegroupbelow18hadthelowestpercentinterviewwitha (#' nonhouseholdmemberorproxy.Theagegroupbetween18and29hadthehighestpercent )$( proxyinthepersoninterview.dB  ԍ(69,p10,R1)C  dTheownershadalowerpercentproxythanthenon-owners.dB  ԍ(69,p11,R2)C  dThe |*%) racecategorycontainingNon-HispanicBlackhadthehighestpercentproxyandallotherrace h+&* N    categorieswerenotsignificantlydifferentfromeachother.dB  ԍ(69, t p12,R3)C  dThepeopleinmulti-unitstructures  hadahigherpercentproxythanpeopleinsingleunitstructures.B  ԍd(69,p16,R4)C  dSinglepersonhouseholdshada  higherpercentproxyinterviewthanlargerhouseholds.dB  ԍ(69,p20,R6)C  d  Thepeoplewhodidnotmatchtothecensushadalargerpercentproxythanpeoplewhomatched. ` dB  ԍ(69,p19,R5)C  dTheCensusBureausentP-samplepeopletofollowupbecausewewerenotabsolutelycertain L  abouttheinformationprovidedintheoriginalinterview,suchasnotmatchedpeoplefromproxy 8  interviews,fromconflictinghouseholds,andfromhouseholdswheresomepeoplematched. $ t AmongthefollowupinterviewsfortheA.C.E.,25.6percentwereproxyinterviewsintheA.C.E.  ` originalinterviewandtheCensusBureaugotahouseholdmemberasarespondentinfollowup  L  foronly35.3percentoftheoriginalproxyinterviews.dB  ԍ(69,   p24,R7)dC  Forfollowupinterviews,thepercent  8  proxyfortheP-samplepeopleremovedbecausetheywerenotresidentsofthehouseholdorwere $  fictitiouswaslower(15.4percent)thanfortheP-samplepeoplewhowerenotremoved(33.5   percent).dB  ԍ(69,p22,R9)C  d   Wolfganget.al.examinedcharacteristicsofthematchedpeoplelookingattherespondenttype.   Whenthe2000A.C.E.responsesforvariouscharacteristicsdidnotagreewithcensusresponses, p theyfoundahigherpercentageofproxyresponsesthanwhentheresponsesagreed.B  ԍd(69,p22,  \ R8)C  d \ Thesedataraiseissuesaboutthedataqualityprovidedbynon-householdmembers.Thepercent 4 ofproxyrespondents(5.5percent)inthe2000A.C.E.interviewraisesquestionsabouttheeffect  p ofproxydataontheundercountestimates.The2010coveragemeasurementprogramshould  \ minimizeproxyinterviewsandonlyacceptinterviewsfromknowledgeablerespondents. H 4 XX &  3.3ErrorSources#XX 4l#  H ThreestudiesproducedsubstantialinformationonerrorcomponentsassociatedwiththePandE H 'lsamples:theMatchingErrorStudy(MES),theEvaluationFollowup(EFU),andthePerson 4 DuplicationStudies.   &  0 ' "MatchingErrorStudy '(#'(# 0 ' TheMatchingErrorStudy(Bean,2001;Bean,2002)providedtheP-samplematchingerror "  ' nrateandtheE-sampleprocessingerrorrate.Expertmatchersclericallyrematchedallofthe l#! peopleinaone-fifthsubsampleoftheA.C.E.sampleclusterstodeterminethematchcode. X$" TheythencomparedthesecodestothematchcodesassignedtoproducetheMarch2001 D% # A.C.E.estimates.B  ԍd(1,p9,R5)dC  0&!$'(#'(# 0 ' &  "EvaluationFollowup(X#&'(#'(# 0 ' TheEFU(KrejsaandRaglin,Report3,2001;Krejsa,2001;RaglinandKrejsa,Report16, )0%( '(p2001;AdamsandKrejsa,2001;Krejsa,2003)consistedofareinterviewofasubsampleof *&) householdsintheone-fifthsubsampleofA.C.E.clustersusedintheMatchingErrorStudy. +'* TheCensusBureauusedtheEFUinterviewresultstomeasuretheE-sampleclassification ,'+        accuracyofcorrectanderroneouscensusenumerations.Theyalsousedtheresultsto  measuretheP-sampledataaccuracyregardingmoverstatusandCensusDayresidence.B  ԍd(1, $ t p9,R6)dC  '(#'(# &  0 ' "PersonDuplicationStudiest'(#'(# 0 ' ThePersonDuplicationStudies(Feldpausch,Report6,2001;Fay,2002;Thompson,Waite, L  't/tandFay,2001;Mule,Report20,2001)tookadvantagethatCensus2000recordedname 8  informationinthedatacapturesystem.Forthefirsttime,thisnewinformationpermitted $ t theCensusBureautoconductnationwidecomputermatchingtomeasurecensusduplication.  ` ThesestudiesalsoexaminedhowwelltheA.C.E.accountedfortheseduplicates.Whilethe  L  A.C.E.matchedrespondentsinthesameblockandsurroundingblocks,thisnewtool  8  permittedtheCensusBureautosearchforduplicatesthroughoutthecountry.Becausethe $  CensusBureaulackedresourcestoconductbothacomputerandclericalmatchtotheentire   country,thePersonDuplicationStudiesinvolvedonlycomputermatching.Thisresultedin   anunderstatementoftheactualduplicationlevel.Thesestudiescomparedtheresultsofthe   EFUwiththePersonDuplicationStudiestodetermineiftheEFUcorrectlymeasured   duplications.B  ԍd(1,pp9&10,R7)dC  p'(#'(# SomeerrorcomponentsproducedfromtheMatchingErrorStudy,EvaluationFollowup,and H PersonDuplicationStudiessuggesttheMarch2001A.C.E.overestimatedthenetundercount 4 whileotherstudiessuggestthenetundercountwasunderestimated.Theresultsfromtheseand  p otherstudiesarediscussedbelowasweexamine:  \ "0 ' Erroneousenumerations4'(#'(# "0 ' Censusomissions '(#'(# "0 ' Balancingerror '(#'(# "0 ' Correlationbias'(#'(# "0 ' Conditioning'(#'(# "0 ' Reinstatedlateadditions'(#'(# "0 ' Census2000imputationsl'(#'(#  &  4 XX3.3.1ErroneousEnumerations,IncludingDuplicates#XX 4!}#  D! EvaluationsindicatedthattheMarch2001A.C.E.didnotmeasureasubstantialportionofthe D#! 'D!}Census2000erroneousenumerations.Themeasurementoferroneousenumerationsiscriticalto 0$" boththenationalnetundercountandtosub-nationalestimates.ThiserrorresultedintheMarch %l # 2001A.C.E.overstatingthenetCensus2000undercountbyatleastthreemillionpeople,witha &X!$ rangeofthreetofourmillion.(ESCAPII,2001.) &D"% TheEFUandPersonDuplicationStudiesdescribedaboveprovidedsubstantialinformation ($' regardingthemeasurementoferroneousenumerations.TheinitialEFUresultsgaveevidenceof )%( asignificantunderstatementintheMarch2001A.C.E.measurementoferroneousenumerations. *%) Becauseofthesizeoftheunderstatement,theEFUwasextensivelyreviewed.TherevisedEFU +&*  #Oa Fh  # F# F  alsoindicatedaproblemwithunderstatingtheerroneousenumerations.TherevisedEFUhada  highlevelofunresolved  %#  14      ׀orconflictingcases h %#  15      .ThePersonDuplicationStudiesfounda  significantnumberofduplicateenumerationsnotmeasuredbytheMarch2001A.C.E.andthat  theEFUdidnotpickupsignificantportionsofthiserror.ThePersonDuplicationStudies t resolvedaportionofthecasesleftunresolvedorconflictingbytheEFUReview. ` ThreeandahalfpercentoftheEFUsamplechangedenumerationstatusasrecordedbythe 8  March2001A.C.E.TheEFUre-codedabout2,800,000estimated(SE223,000) correct $ t enumerationsas erroneousenumerationsandre-codedabout900,000estimated(SE99,000)  `  erroneousenumerations,as correctenumerations.(KrejsaandRaglin,Report3,2001.)The  L  EFUfoundanestimatednetdifferenceof1,900,000.Also,about4,500,000estimated(SE  8  353,000)casesintheEFUcouldnotberesolved.ThisstudyshowedtheMarch2001A.C.E. $  overstatedthenetundercountbyaminimumofabouttwomillionpeople.(ESCAPII,2001.)   Forcomparison,the1990evaluationstudy(West,1991)oferroneousenumerationsfoundthe   1990PESunderstatedthenetundercountbyabout360,000estimatedpersons.Westalsofound   about1,273,000estimatedE-samplepeoplecouldnotbematchedorwereunresolved.B  ԍd(2,p37, H R1)dC  Ԁ(In1990,   theEFUwasareinterviewusingthe1990PersonFollowup(PFU)form.In2000,theEFUwas p designedtodifferfromthe2000PFUform,includingmoreresidencyprobes.) \ BecauseoftheEFUsimplicationsfortheMarch2001A.C.E.estimates,furtherEFUanalysis 4 wasconducted.BettertrainedmatchinganalystsfromtheNationalProcessingCenter(NPC)  p reviewedasubsampleoftheEFUandproductioncases.ThisreviewoftheoriginalEFU  \ confirmedtheerrorsintheMarch2001A.C.E.sidentificationoferroneousenumerations. H About1,800,000estimated(SE189,000)enumerationscodedascorrectinproductionwerethen 4 codederroneousintheevaluation,whileabout361,000estimated(SE46,000)enumerations   codedaserroneousinproductionwerethencodedascorrectinthereview.(AdamsandKrejsa,   2001.)Consequently,thenetdifferenceinthe correctenumerationto erroneousenumeration  and erroneousenumerationto correctenumerationcellswasestimatedtobe1,450,000,  ratherthantheinitial1,900,000.However,thereviewidentifiedover15millionestimatedcases  whichcouldnotberesolvedorhadconflictingA.C.E.andEFUinformation.(Adamsand l Krejsa,2001.)Thecodingoferroneousenumerationswasconservativebecausethepurposeof X  thereviewwastodetermineiftheoriginalcodingdesignatedtoomanyerroneousenumerations. D! Thiscreatedalargenumberofconflictingcases.Dependingonassumptionsregardingthe 0"  enumerationstatusoftheseconflictingcases,theestimatedoverstatementofthenetundercount #l! couldrangefromabout1.45millionto5.9millionpeople.(ESCAPII,2001.) $X" ThePersonDuplicationStudiesfoundasignificantnumberofduplicateenumerationswere %0!$ incorrectlymeasuredintheMarch2001A.C.E.orintheEFU.Furthermore,uponcombiningthe &"% PersonDuplicationStudiesresultswiththereviewoforiginalEFUresults,theCensusBureau '#&  Oa!Oa!     couldexplainsomeoftheunresolvedandconflictingcases.Basedonthiswork,theydeveloped  morerefinedrangesfortheMarch2001A.C.E.overstatementlevel.Directestimatesproduced  fromthePersonDuplicationStudiesindicatedtheMarch2001A.C.E.errornotmeasuredwas  aboutthreemillionpersons.Inaddition,theCensusBureauexpectedfurtherrefinements(Fay, t 2002)tothetreatmentoftheunresolvedandconflictingcaseswouldleadtoaboutanadditional ` 800,000errors.Thus,theyreducedtheestimatednetundercountoverstatementrangetothreeto L  fourmillionpersons.(ESCAPII,2001.)B  ԍd(1,pp10&11,R8)C  d 8  Martin,Fay,andKrejsa(2002a;2002b)conductedapreliminaryevaluationoftheA.C.E.Person  ` FollowupandEFUquestionnairestounderstandtheirsuccessinidentifyingerroneous  L  enumerations.Theyexaminedtheconsistencyofresidencyreportinginthetwosurveys,and  8  usedduplicationratestoassessthevalidityoftheclassificationsproducedbythetwo $  questionnaires.Theyexaminedresponsestoquestionnaireitems,anddidnotincorporate   informationfromclericalcodingofinterviewersnoteswhichformedthebasisofofficial   estimatesofCensus2000coverage.TheA.C.E.PersonFollowupquestionnaireattemptedto   determineCensusDayresidencewithonlyafewglobalquestions;theEFUquestionnaireaskeda   largernumberofmoredetailedquestions.B  ԍd(3,p1,R1)C  dMartinet.al.sanalysisindicatedhighlevelsof p inconsistentreportingofmovesinandoutofhouseholds,secondresidences,andstaysingroup \ quarters.TheyfoundthatB  ԍd(3,p33,R3)dC  Ԁidentificationofenumerationerrorsappearstohavebeengreatly H enhancedbytakingintoaccountinformationfrombothsurveys,becauseeachquestionnaire 4 addedinformationabouterrorsthatwerenotidentifiedbytheother.Neitherquestionnairecould  p besaidtorepresenta goldstandardforreportingaccuracy.(Martin,2001.)Theyalsofound  \ thattheEFUidentifiedsmallbutsignificantnumbersoferroneousenumerationsamongthe H matchedcases,andthesecaseshadhighratesofduplication.B  ԍd(24,p3, D! R1)dC   4 Martin,Fay,andKrejsa(2002a;2002b)andMartin(2001)identifiedthefollowingasresearch   relatedtoerroneousenumerationsidentificationneededfor2010:  "0 ' Reviewresidencerulesandcriticaldefinitionstocreateasimplerclassificationschemethat  reliesonclearerdefinitionsthatcanbeunderstoodbyinterviewersandrespondentsinthe l field.B  ԍd(3,p33,R4)(4,p8,R1)dC  X '(#'(# "0 ' Devoteresourcestolongtermcoveragemeasurementinstrumentimprovements.Research 0"  andinstrumentdevelopmentandtestingareneededtoaddressconceptual,recall,and #l! comprehensionissuesaffectinggroupquartersresidenceandmultipleresidencereporting.B  ԍ  Ўd(4,p9,R2)dC  Ԁ $X" Afurthergoalforresearchanddevelopmentshouldbetoreducerelianceonexpensive,time $D # consumingandlabor-intensiveclericalcodingoperations.Accurateclassificationsshould %0!$ beproducedbythestandardizedquestionsintheinstrument,reducingtheneedfor &"% interventionandinterpretationbyanalystsandclerks.B  ԍd(3,pp33,R5)dC  '#&'(#'(# "0 ' DevelopacoveragemeasurementdesignthatbetterintegratesinstrumentsfortheA.C.E., )$( followupsurveys,andevaluations.Explorehowtoimprovecoveragemeasurementby |*%) maximizingthepotentialeachdatacollectioninstrumentoffers.Forexample,itwouldbe h+&*  dO"Oa"     desirableforthePFUinstrumenttodomoreprobingaswasdonebyEFU.TheEFUand  PFUinstrumentsmightalsobedesignedtoidentifycensusorA.C.E.omissions.The  Censusinstrumentsmightcollectadditionalinformationtofacilitateresidence  determinations.dB  ԍ(4,p9,R3)dC  t'(#'(# TheEFUprovidedinformationregardingwhethertheMarch2001A.C.E.accuratelyidentified L  Census2000discrepantenumerations. %#  16      ׀Thisstudyshowednoproblemwiththeidentificationof 8  discrepantenumerations.(Krejsa,2001.)B  d(25,pi,R1)dC   $ t Mule(Report20,2001)andJones(Report0.16,2003) foundhigherduplicationratesforcertain  L  subgroupsandareas.BothfoundhigherduplicationratesamongNon-HispanicBlacksand  8  Hispanicscomparedtootherrace/ethnicitysubgroupsandamong18to29yearoldmales $  comparedtootherage/sexsubgroups.B  d(20,ppiv&v,R1)(21,pp10&11,R2)dC  Additionally,Jonesfoundduplicationmoreprevalentin   smallmultiunithousingstructuresandmobilehomescomparedtootherhousingtypesubgroups,   intheNewYorkandBostonregionalofficescomparedtootherregionaloffices,amongrenters   comparedtoowners,andamongpersonsinduplicatehousingunitsorinhousingunitsaddedto   thecensusinventoryafter1990comparedtopersonsincludedinotherways.dB  ԍ(21,p.v,R4)(21,p.21,R5)C  dThesefindings p suggestthatitmaybebeneficialtotargetthesesubgroupsandareasinconductingperson \ duplicatesearches.B  d(21,  p21,R1)dC   H Finally,Feldpausch(2001)examinedtherelationshipbetweentheE-samplepeopleidentifiedas  p duplicatesoutsidethesearchareaandtheircorrespondingMarch2001A.C.E.enumeration  \ status.Inpastcensuses,theCensusBureauhadnowaytoevaluatethecodingofpeople H duplicatedoutsidethepost-enumerationsurveyssearcharea.ForCensus2000,analyses 4 precedingthePersonDuplicationStudiessearchedforduplicatesthroughoutthecountry.Mule   (Report20,2001)conductedcomputermatchingtodeterminetheextentofduplicate   enumerationsnotfoundintheMarch2001A.C.E.ThisallowedhimtoevaluatetheMarch2001  A.C.E.codingofpeopleduplicatedoutsidethesearcharea.  Apersonenumeratedtwicebythecensuswasduplicated.Therecordofapersonenumeratedin l thecorrectplaceshouldhavebeencodedasacorrectenumeration.Therecordofaperson X  enumeratedinanincorrectplaceaccordingtocensusresidencerulesshouldhavebeencodedas D! anerroneousenumeration. 0"  FeldpauschfoundalowerthanexpectedpercentageerroneousenumerationforE-samplepeople $X" duplicatedtopeopleingroupquarterswheretheresidentswerenotallowedtoclaimusualhome $D #  elsewhere(45.5percentforcollegedormsand16.5percentforothergroupquarters).Thisrate %0!$  dO#Oa#     shouldhavebeenclosertoonehundredpercentbecauseforthemajorityofthesepeopletheir  usualresidencewasprobablythegroupquarters.   ForthoseE-samplepeopleduplicatedtopeopleingroupquartersallowedtoclaimusualhome t elsewhere,Feldpauschfoundahigherthanexpectedpercentageerroneousenumeration ` (12.5percent).Theerroneousenumerationrateforthesepeopleshouldhavebeenclosetozero L  becausethehousingunit,notthegroupquarters,wasprobablytheirusualresidence. 8  ThepercentageerroneousenumerationforE-samplepeopleduplicatedtopeopleinhousingunits  ` outsidetheA.C.E.searcharea(14.2percent)waslowerthantheapproximate50percentone  L  mighthaveexpected.Onemightexpect50percentbecausehalfofthetimethewronghousing  8  unitshouldbeinsample,resultingincodingtheresidentsaserroneous. $  &  Somepossibleexplanationsofthesefindingsare:   "0 ' Theinstructionsindicatingwhotoincludeonthecensusquestionnairemaynothavebeen   'Ecompletelyunderstoodorwereignoredbytherespondent.Examplesoflivingsituations p causingproblemsincludecollegestudents,peopleinlocaljails,andpeopleinnursing \ homes.H'(#'(# "0 ' Therespondentmaynothaveknownahouseholdmemberwasenumeratedelsewhere. p'(#'(# "0 ' Somegroupquartersenumerationsmayhaveusedadministrativerecordsnotreflecting H residentsasofApril1,2000.4'(#'(# "0 ' Residenceinsomegroupquartersistemporary,suchaslocaljails.Somepeoplecountedin   thesemaybeusualresidentsofthesamplehousingunit.'(#'(# "0 ' Thecomputermatchingofduplicatesoutsidethesearchareamightbeincorrect.The  CensusBureaudoesnotthinkthiswaslikely,becausetheCensusBureauonlylookedat l thosecasesthathadahighprobabilityofbeinglinkedcorrectly.X '(#'(# "0 ' TheMarch2001A.C.E.failedtocompletelyidentifyerroneousenumerationsduetoother 0"  residence.Thepercentotherresidencewas1.4intheMarch2001A.C.E.and2.3inthe #l! 1990PES(thesepercentagesreflecttheredistributionofpeoplewithunresolvedstatus). $X" TheresultsoftheEvaluationFollowupalsomeasuredthisphenomenon.$D #'(#'(#     EvidencesuggeststheMarch2001A.C.E.didnotcodesomepeopleaserroneousenumerations &"% whoshouldhavebeenbecausetheylivedinotherresidences.B  d(26,ppiii&4,R1)dC  Somepeoplewereidentifiedas '#& erroneousenumerationsbecausetheyshouldhavebeencountedatanotheraddress.Inmanyof (#' thesecasestheywerealsocountedwheretheyshouldhavebeencounted,makingthem )$( duplicates. |*%)  h+&* ThesestudiesshowthattheCensusBureauneedstoconductfurtherresearchtobetteridentify  erroneousenumerations,includingduplicatesinthe2010census.Thisresearchshould  investigatechangesinresidencerules,datacollectionprocedures,codingproceduresand  instrumentsandimprovedestimationapproaches.FurtherworkbyMartin,Fay,andKrejsa t (2002a;2002b)suggeststhattheCensusBureauconsidersendingasubsampleofmatchedcases ` tofollowupduringproductionbecausetheremaybeundetectederroneousenumerationsamong L  them. 8  AdditionalinformationabouterroneousenumerationsforCensus2000maybefoundinAdams  ` andLiu(2001);Bean(2001);Bean(2002);Feldpausch(2002);Liu,Jones,andFeldpausch  L  (2001);Liu,Byrne,andImel(2001);andRaglinandKrejsa(Report16,2001).4 XX  8  #XX 4Y# 4 XX3.3.2 &  ԀCensusOmissions#XX 4#  6  TheCensusBureauusedthePsampletomeasurecensusomissions.Therefore,matchingof :  P'sampletothecensus,theclassificationofP-samplemoverstatusandCensusDayresidence, & andthedeterminationofP-samplediscrepantenumerationswerecriticalaspectsoftheP-sample  processing.TheMESproducedinformationaboutmatchingaccuracy.TheEFUprovided  informationabouttheaccuracyoftheclassificationofmoversandCensusDayresidenceandthe  lackofdiscrepantenumerations.B  ԍd(1,pp11&12,R10)dC    IntheA.C.E.clericalmatching,clerksexaminedcomputerizedP-sampleresponsesandcensus ^ responses.Theyalsohadaccesstothescannedimagesoftheoriginalquestionnaires.Because J ofautomationofthematchingoperation,allmatchingcouldbedoneatonelocation,insteadof 6 sevenasin1990.(NationalResearchCouncil,2001.)B  d(5,p101,R1)dC   "r TheCensusBureauconductedaMatchingErrorStudyfortheCensus2000DressRehearsal J IntegratedCoverageMeasurement(ICM)andforthe1990PES.TheMESfortheCensus2000 6 DressRehearsalwasunabletomeasuresignificantmatchingerrorbecauseofa100percentQA " duringtheCensus2000DressRehearsalICM %#  17      .B  C  ԀThematchcodediscrepancyrates(which  representthesizedifferencebetweentheperson-levelICMandMESmatching)forthePsample   werelessthanonepercentinallsites.B  d(6,p1,R1)dC   !   "  The1990MESfoundthePESgenerallytendedtooverestimatetheP-samplenonmatches,   especiallywhenmatchingcentralcity,minoritypersons.Byevaluationpoststratum(basedon  region,urbanicity,andminoritystatus)thebiasesduetomatchingerrorrangedfrom  approximately0.7percentto1.3percentofthepopulationsizes.Ofparticularconcern,thePES t overestimatednonmatchesforBlacksbyabout4.5percentwhichequatedtoanapproximately ` 0.7percentpositivebiasinthetotalBlackpopulation.B  ԍd(33,p5,R1)dC  Ԁ(DavisandBiemer,#H-2,1991.) L  Reductionsinmatchingerrorfrom1990to2000provideevidencethatchangesmadefrom1990 $ t improvedthequalityoftheA.C.E.matchingprocess.Evenwiththeseimprovements,matching  ` errorfromthePandEsamplescombinedinflatedthenationaldualsystemestimateby483,938  L  withastandarderrorof92,877andthereforeoverstatedtheundercountestimate(holdingall  8  othererrorsconstant).(Bean,2001;Bean2002.)B  d(6,pi,R3)dC  Tofurtherreducematchingerrorinthefuture, $  plannersshouldcontinueeffortstoimprovethematchingprocess.Improvementstothe   matchingprocessmaybemadeby:   "0 ' simplifyingthetargetedextendedsearch(TES)matchingproceduresandimprovingthe   qualitycontrolfortheTESclusters;p'(#'(# "0 ' identifyingwaystofurtherensurethatmatchersupdatecaseswithinsufficientinformation \ formatching;andH'(#'(# "0 ' clearlydefiningrulesforcodingcasesasdiscrepantorunresolved.(Bean,2002.)4'(#'(# TheEFUshowedthatmisclassificationofmoversintheMarch2001A.C.E.mayhaveresulted  \ inanestimatedunderstatementofabout450,000inthenetundercount.(RaglinandKrejsa, H Report16,2001.)Thisfinaleffectresultsfromsignificantchangesinmoverstatus.These 4 changesinvolvedalargenumberofmoversbecomingnonmoversandviceversa.TheEFU   estimatedthatabout4.5millionpeopleclassifiedas moversinproductionbecameEFU    nonmovers,andabout2.4millionpeopleclassifiedas nonmoversinproductionbecameEFU   movers.Atthenationallevelthereisasmallestimatedneteffectofabout65,000onthe  accuracyofthemeasurementofcensusomissions.B  ԍd(1,p12,R11)C  d  TheCensusBureauwasconcernedabouttheEFUmeasurementofmoverswhobecame X  nonmovers,specificallyaboutwhethertheEFUmeasuredtoofewmovers,duetoits D! questionnairedesign.TheEFUrequiredlessdetailedinformationforclassifyingapersona 0"  nonmoverthanforclassifyingapersonamover.Anexaminationofthebiascausedbymover #l! statuschangesindicatedtheeffectofmover-to-nonmoverchangeswasgreaterinabsolutevalue $X" thantheeffectofnonmover-to-moverchanges.Eventhoughtheneteffectsoftheseerrorscancel $D # atthenationallevel,assessmentofthesubnationaleffectsrequiresfurtherresearch.B  ԍd(1, 8  p12,R12)C  d  %0!$  Martin,Fay,andKrejsa(2002a;2002b)examinedtheEFUquestionnaireinregardtomover '#& identification.Resultsindicatedhighlevelsofunreliabilityinmeasurementofmoversinandout (#' ofhouseholds.B  d(3,p31,R2)C  dResearchandinstrumentdevelopmenttestingareneededtoaddressissues )$( affectingthereportingofmoves.B  ԍd(3,p32,R6)dC   |*%)  h+&* TheEFUalsodemonstratedthatA.C.E.didnothavealargeproblemwithdiscrepant  enumerations.TheEFUidentifiedaweightednet326,855P-sampleresidentswhoshouldhave  beenremovedbecausetheywerediscrepant.Upto23,879weightedpeoplewereexcludedas  P-sampleresidents,butidentifiedaspotentiallydiscrepantbytheEFU.(Krejsa,2003.)B  d(27, \ piv,R1)dC   t TheMESandEFUstudiessuggesttheCensusBureaushouldcontinueeffortstoimprovethe L  matchingprocessandconductresearchtoimprovethereportingofmoversfor2010. 8  AdditionalinformationaboutcensusomissionsandmoversforCensus2000maybefoundin  ` Liu,Jones,andFeldpausch(2001);Liu,Byrne,andImel(2001);andWolfgang,Adams,Davis,  L  Liu,andStallone(2001).  8   4 XX3.3.3BalancingError#XX 4f#    Balancingerrormayoccurifthesearchareasformatchestothesurroundingblocksdonotequal   searchareasforcorrectenumerationsinthesurroundingblocks,theCensusBureaucodedthe   datainconsistently,ortheCensusBureauintroducedP-samplegeocodingerror.SincetheA.C.E.  usedarandomsampleindependentfromthecensus,theCensusBureauexpectedcensushousing  unitsshouldbeerroneouslyincludedwithinA.C.E.sampleclustersasoftenascensushousing p unitsshouldbeerroneouslyexcludedfromA.C.E.sampleclusters.Afteradjustingforthe \ P-samplecoverage,iftheproblemsaboveoccurred,theweightednumberofmatchesto H surroundingblocksmaynothaveequaledtheweightednumberofcorrectenumerations,creating 4 balancingerror.(AdamsandLiu,2001.)B  d(29,pi,R1)dC  Ԁ  p TheA.C.E.carriedoutmatchinginadefinedsearchareaconsistingoftheA.C.E.sampleblocks H (clusters)andatargetedareaofblockssurroundingorborderingtheA.C.E.blocks(i.e.Targeted 4 ExtendedSearchArea).TheMarch2001A.C.E.foundthreemillionmorematchesin   surroundingblocksthancorrectenumerations,indicatingbalancingerrorasapotentialproblem.   Thiscouldhaveaffectedtheaccuracyoftheestimates.TheCensusBureauidentifiedvarious  scenariosthatcouldexplainthedifferences.Also,theCensusBureauconductedevaluationsto  investigatethesourceofthisdifference,identifythescaleofanyerror,andassesswhetherits   magnitudecouldsignificantlyaffecttheaccuracyofadjusteddata. l! TheevaluationattributedmostofthethreemilliondifferencetotheA.C.E.listinghousingunits D#! intheblockssurroundingthesampleblocks.Thishadlittle,ifany,effectontheDSE.However, 0$" theevaluationsdetectedabout246,000additionalA.C.E.people(SE82,000)locatedoutofthe %l # surroundingblocksduetoP-samplegeocodingerrors.(AdamsandLiu,2001.)Theevaluations &X!$ alsoestimatedanadditional195,000people(SE56,000)incorrectlyidentifiedashavingbeen &D"% correctlyenumerated,althoughfoundoutsideofthesearcharea.Theseerrorsresultedinan '0#& overstatementofthenetundercountbyabout450,000persons.TheEFUandMESresults ($' includedportionsoftheseerrors.Whileadditionalworkisrequiredtoresolvethepotential )%( effectsofbalancingerror,theCensusBureaubelievesthatmostoftheconcernsregarding *%) balancingerrorhavebeenaddressed.B  ԍd(1,p14,R16)dC   +&* ЇTherelationshipsbetweenvariablesusedindefiningpost-stratafordualsystemestimationand  variablesrelevanttosamplingofTargetedExtendedSearchB  d(16,p9,R2)dC  Ԁcasesrevealednoconcernwith  geocodingerrororinsightforimprovinggeocodingerror.(Wolfgang,Stallone,andAdams,  2002.)B  d(16,piv,R1)dC   t  &  4 XX3.3.4CorrelationBias#XX 4#  L  'L  Correlationbiasreferstothetendencyforcensusenumeratedpeopletomorelikelybeincluded L  intheA.C.E.thanpeoplemissedbythecensus.Itcanresultfromcausaldependence,which 8  occurswhentheactofbeingincludedinthecensusmakessomepeoplemorelikelyorlesslikely $ t  tobeincludedintheA.C.E.Correlationbiascanalsoresultfromheterogeneitybias,whicharises `  whendifferentpeoplewithinpoststratahavedifferentchancesofbeingincludedinthecensus L  andalsodifferentchancesofbeingincludedintheA.C.E.Tocausethistypeofbias,these 8  chancesofinclusionmustbecorrelated,aswhenthoselikelytobemissedbythecensusarealso $  morelikelytobemissedbytheA.C.E.Thistypeofheterogeneitywouldresultinadownward   biasintheDSEs.InMarch2001,theCensusBureauassessedpossiblecorrelationbiasinthe   A.C.E.estimatesbycomparingtheA.C.E.andDAresults  (Bell,B-12*,2001).ThesecorrelationbiasestimatesusedDAestimatesasofFebruary26,  2001.ThecorrelationbiasestimateswererecomputedinOctober2001tousetheRevisedDA p estimates(Bell,Report10,2001).B  d(1,p12,R17)dC   \ Althoughthereisevidenceofandreasontoexpectsomecorrelationbiasindualsystem 4 estimates,theCensusBureaudidnotcorrectforitinDSEsproduceduptoMarch2001.Forthe  p 1990PEStherewasconcernaboutthenewnessofthemethodologyforestimatingcorrelation  \ biasandaboutthetimerequiredtofitacorrelationbiasadjustmentintotheproductionschedule. H Therewasalsoconcernthatalternativemodelsforcorrelationbiascouldbeusedthatprovided 4 thesamefittothedatabutyieldeddifferentsubnationalestimates.Thislatterconcernwasalso   presentforthe2000A.C.E.,leadingtothedecisionmadeintheplanningstagesnottoadjust   thoseestimatesforcorrelationbias.Anotherpartofthethinkingbehindthisdecisionwasthe  that,inthepresenceofcensusundercounts,DSEswithoutadjustmentforcorrelationbiasare  conservativeinthattheymovetheestimatesintherightdirection,thoughperhapsnotfully   correctingfornetundercoverage.Thisthinkingtiedinwithconcernsthatadjustmentfor l! correlationbiascouldovershootthetruth,atleastforsomepopulationgroups,ajustifiable X"  concernifotherbiasesintheDSEsarepositivetendingtomakethemoverestimates(aswasthe D#! casewiththeMarch2001A.C.E.estimates).A.C.E.RevisionIIfacedadifferentsituation, 0$" however,because(i)itcorrecteditsDSEsforotherbiases(suchastheunderestimationof %l # erroneousenumerationsduetoduplication),and(ii)therewasevidenceofsomenetovercounts &X!$ andsomesmallerundercountsforvariousgroupsinthe2000census.InthissituationDSEs &D"% withoutadjustmentforcorrelationbiaswerenotseenasconservative,astheycouldestimate '0#& overcountsforgroupsthatweretrulyundercounted,andthusmoveestimatesfurtherfrom,not ($' closerto,thetruth.Becauseofthis,thetreatmentofcorrelationbiaswasreconsideredand )%( A.C.E.RevisionIIdidcorrectitsestimatesforcorrelationbias.B  d(31,pp1&2,R1)dC  ԀForthe2010census,theCensus *%) Bureaushouldre-evaluatewhethertoadjustforcorrelationbias. +&* Ї &  4 XX3.3.5Conditioning#XX 4#   Conditioning,orcontaminationbias,referstothesituationwheretheA.C.E.influencedthe  'census.Asinthe1990Census,B  C  contaminationbiaswasnotaprobleminCensus2000,as  researchdidnotidentifyanystrongevidenceofitspresence.B  ԍ(Bench,2001).d(19,P1,R1)dd(1,  \ pp14&15,R19)C  dTherefore,procedurestoavoid  contaminationbiasappeartowork.(Bench,2001;Bench2002.) t   4 XX3.3.6&  ԀLateAdditions#XX 4#  L  TheCensusBureaumadesubstantiallymorelateadditionsthaninthe1990census.Late L   'L additionsrefertopersonsincludedinthefinalcensuscountwhowereexcludedfromA.C.E. 8  matchinganddualsystemestimationbecauseoftheirlateinclusion.ForCensus2000,thelate $t  additionsconsistedexclusivelyofhousingunitstemporarilyremovedfromthecensusbecause `  theCensusBureaususpectedtheyduplicatedotherhousingunits,butwhichtheylaterreinstated L  intothefinalcensusafterfurtherresearch.ThehousingunitswerereinstatedaftertheA.C.E. 8  matchingprocessstarted(i.e.thematchingprocessdidnotinfluencethedecisionofwhatto $ reinstate).Ifthereinstatedpeoplewereasmallpercentageofthecensus,oriftheirA.C.E.  coverageratewassimilartotheA.C.E.coveragerateforcensuspeopleincludedintheA.C.E.,  thentherewouldbeminimaleffectontheDSEs.(Hogan,Q-43,2001.)TheCensusBureau  validatedthisassumptionbyclericallymatchingthereinstatedpeoplecollectedinA.C.E.and  censusinevaluationclusters(a1/5thsampleofA.C.E.clusters),attemptingtomimicasbestas p possiblewhatwouldhavehappenedhadtheybeenamongthecensuspeopleintheproduction \ matchingoperations.(Raglin,2001.) H Basedonthisadditionalwork,theCensusBureauconcludedthatexcludingreinstatedcensus  p peoplefromtheA.C.E.hadlittleeffectontheDSE.TheMarch2001A.C.E.coverageratemay  \ havebeenoverestimatedby0.034to0.082percentagepoints.(Raglin,2001.)dB  ԍ(1,pp15,R20)C  d H  &  4 XX3.3.7CensusImputations#XX 4 #    Census2000experiencedahigherrateofwholepersonimputationsthanthe1990census.The   ' CensusBureauexcludedwholepersonimputationsfromA.C.E.matchingactivities,butreflected !  theminthecensuscoverageerrorasmeasuredbytheA.C.E.TheCensusBureauexamined "  whetherCensus2000designfeaturesexplainedthewholepersonimputations(andthusshould #! havenodiscernibleimpactontheA.C.E.),concludingthatthekind,level,andpatternofwhole $" personimputationsinCensus2000raisednoissuesrelativetotheaccuracyoftheMarch2001 l% # A.C.E.adjustment.(ESCAPII,2001.) X&!$   D'"% Approximately5.77millionpersonshadalltheircharacteristics(shortformdataitems' %#  18      )   imputedinCensus2000,comparedto1.97millionpersonsinthe1990census.TheCensus  Bureauaddedapproximately1.2millionofthesepersonstothecensuscountthroughacount  imputationprocess.TheCensusBureaucountedtheremaining4.6millionpersonsdirectly t throughthecensusenumerationprocess,butimputedalltheirpersoncharacteristicsbecause ` informationaboutthemwassubstantiallymissingfromthecensusrecords.(Nash,2001.) L  Researchintothesourcesofthewholepersonimputationsidentifiedchangesinthewaydata 8  werecollectedforlargehouseholdsascontributingtothelevelofhousingunitsrequiring $ t imputation.Furthermore,thecountimputationratewascomparabletotherateexperiencedin  ` the1970and1980censuses.(ESCAPII,2001.)  L  TheCensusBureaualsoexaminedcharacteristicsoftheimputedpersons.Itfoundsimilar $  distributionsbetweentheage,raceandsexcharacteristicsofthepopulationrequiringsomeform   ofimputationandthedata-definedh %#  19      ׀populationwiththeexceptionoftheagecategoryunder18.   Thehighproportionofyoungerpeopleinthewithinhouseholdcategorycausedtherelatively   higherpercentofthepopulationunder18intheimputedpopulation.Thisreflectedthefactthat   largehouseholds(greaterthansix)likelyhavechildrennotabletobeaccommodatedbythe p six-personmail-returnform,whichwouldrequireimputationiftheircharacteristicswerenot \ collectedinthecoverageeditfollowup.(WetroganandCresce,2001.)dB  ԍ(1,pp15&16,R21)C  d H Forthe2010census,theCensusBureaushouldidentifywaystoreducewholeperson  p imputations.XaGXXX  \ #XXXXaG*# &  XaGXXX XXaG& A 4.A.C.E.:HousingCoverage 4  #XaGX##XXXXaG#XaGXXXTheCensusBureauconductedstudiestoexaminehousingunitcoverage.Thesestudieslooked 4 ' atthenetundercountaswellastheP-samplenonmatchesandE-sampleerroneousenumerations   forhousingunits.Theyexaminedtypesoferroneousenumerations,includingthehousingunit   duplicationforCensus2000.Anotherstudyexaminedconflictinghouseholds,whichare  matchedhousingunitswithdifferentpeopleinthePsampleandtheEsample.  Thecensusflaggedapproximately2.4millionhousingunitsaspotentialduplicates,which l! A.C.E.processingexcluded.(Nash,Memorandum78,2000.)#XXXXaG\#ԀdB  ԍ(70,p6,R1)C  dTheCensusBureaureinstated X"  onemillionofthesehousingunitsintothecensus.(Nash,Memorandum82,2000.)B  d(71,p1,R1)dC  Noneofthe D#! studiesinthissectionincludedthesereinstatedhousingunits. 0$" XaGXXXThestudiesexaminedthefieldoperationsandinstrumentsusedduringthehousingunitphaseof  theA.C.E.andmaderecommendationsabouttheiruseincoveragemeasurementfor2010.   #XXXXaG#XaGXXXXXaG&  4.1HousingUnitCoverageStudy #XaGX##XXXXaG#XaGXXX ` #XXXXaG#XaGXXX* KXr X*Barrett,Beaghen,Smith,andBurcham(2003)examinedtheresultsoftheHousingUnit `  '`9CoverageStudy(HUCS).TheHUCSmeasuredtheCensus2000housingunitcoverageusing L  datafromtheA.C.E.UsingDSE,itestimatedthenetcoverageofhousingunitsenumeratedin 8  Census2000.TheHUCSestimatednonmatchesfromthePsampleofhousingunitsand $ t  erroneousenumerationsfromtheEsampleofhousingunits.Togethertheyestimatethenet `  undercountofhousingunits. L  &  CoverageofhousingunitsenumeratedinCensus2000wascomparabletothehousingunit $  coveragein1990. Table6shows:dB  ԍ(72,p9,R1)C  d   #XXXXaG #XaGXXX"0  #XXXXaG##XaGXXXAnetundercountofhousingunitsof0.61percentinCensus2000and0.96percentin  '"1990.Thenetundercountswerebothsignificantlydifferentfromzero,butnot  significantlydifferentfromeachother.p(#(# 0   (#(# #XXXXaG $#XaGXXX"0  Foroccupiedhousingunits,therewasnoobservablesignificantdifferencebetweenthe H 2000and1990coverage,withanetundercountof0.33percentin2000and0.53percent 4 in1990. p(#(# "0  Anetundercountforvacantsat3.37percentin2000wasnotsignificantlydifferentfrom H the4.71percentnetundercountin1990.4(#(# **Sd dXXXXXXXX~(#(#*,dd ,'dd ,'dd ,dd ,dd ,,dd +  )V   )&  Table6:Comparisonof1990and2000PercentNetUndercountof  p HousingUnits(Standarderror) ;* \V ; @ Status D3$ " D 2000HUCS D3$ " D 1990HUCS <2$  " < 9/! "\! 9 National 9/! "\" 90.61(0.16) 9/! "\#3 9 9/! "\$3 90.96(0.24) 9/! "\%3 9 9/! "\&3 9 9/!#' 9Occupied 9/!#( 90.33(0.13) 9/!#)3 9 9/!#*3 90.53(0.21) 9/!#+3 9 =V,!#,3 = =V,! %\ -V =Vacant =V,! %\ .V =3.37(0.98) =V,! %\ /3V = =V,! %\ 03V =4.71(1.26) =V,! %\ 13V =1'% %\ 23  V 1' )Barrett,Beaghen,Smith,andBurcham(2003)#XXXXaG%#XaGXXXalsofoundthecoverageforoccupiedhousingunits 'd"3 consistentwith1990forvariousresearchcategoriessuchastenureandtypeofenumerationarea. (P#4 "0  #XXXXaG0#XaGXXXIn2000,theundercountforvacanthousingunits(3.37percent)wassignificantlygreater )(%6 thanforoccupiedunits(0.33percent).In1990,theundercountedforvacanthousing *&7 units(4.71percent)wassignificantlygreaterthanforoccupiedunits(0.53percent).#XXXXaG*1#XaGXXXdB  ԍ(72,p9,R2)C  Ԛ#XXXXaG2#dXaGXXX+'8(#(# #XXXXaG3#XaGXXX#XXXXaGp3#"0  Asin1990,the2000coverageforhousingunitsnotoccupiedbyownerswasnot  significantlydifferentthanforthoseoccupiedbyowners.Thenetundercountforowner  occupiedhousingunitswas0.12percentin2000and0.37percentin1990.Thenet t undercountforhousingunitsnotoccupiedbyownerswas0.57percentin2000and ` 0.80percentin1990.XaGXXXdB  ԍ(72,p15,R3)C  Ԛ#XXXXaGv5#dL (#(# "0  Thenetcoverageofhousingunitsinsmallmulti-unitstructures(2to9housingunits)was $ t significantlybetterin2000(-0.17percentnetundercount)thanin1990(2.25percentnet  ` undercount).XaGXXXdB  ԍ(72,p19,R4)C  Ԛ#XXXXaG7#dThenetovercountof0.17percentwasnotsignificantlydifferentfromzero.  L  XaGXXXdB  ԍ(72,p19,R5)C  Ԛ#XXXXaG7#d (#(# "0  TheCensusBureauovercountedoccupiedhousingunitsinsmallmulti-unitstructures   (-1.30percent)in2000,butundercounted(2.11percent)themin1990.XaGXXXdB  ԍ(72,p20,R6)C  Ԛ#XXXXaG&9#d (#(# "0  Thesizeofthemetropolitanstatisticalareadidnotimpactcoverageofhousingunitsin   mailout/mailbackareas.Foroccupiedhousingunits,theCensusBureaufoundno p significantdifferencesbetweenthenetundercountsformailout/mailbackareasinsmall \ (0.53percent),medium(0.30percent),orlarge(0.11percent)metropolitanareasversus H allothertypesofenumerationareas(0.22percent).XaGXXXdB  ԍ(72,p24,R7)#XXXXaG;#dXaGXXXd(72,p26,R8)C  Ԛ#XXXXaG<#d4(#(# XaGXXXAsTable7dB  ԍ(72,p10,R9)C  #XXXXaG<#dXaGXXXshows,the2000percentofE-sampleerroneouslyenumeratedhousingunitswas  \ slightlybetterthanthe1990percent.Thedifferenceof0.53percentagepointwasstatistically H significant.B  ԚC  ԀThepercentnotmatchedwasnotsignificantlydifferent(3.62percentin2000vs. 4 3.57percentin1990).Comparisonsweremadebuttheyshouldbeusedwithcaution.The   searchareaswerenotthesameforthe1990and2000HousingUnitCoverageStudies. %#  20      ׀   ЋXXaGB  C  *dSddd 'dd ''dd 'dd dd ,dd ,(#(#, dd ,'dd ,dd ,sdd , dd +  )V )&  #XaGXЋ?# Table7:NationalHousingUnitCoverageEstimates(standarderror) 7V& \V 7 AV0!"V A 2000   (inpercent) DV3$x "V D 1990    (inpercent) <2$x !"V <NetUndercount 9/!!H" 90.61(0.16) 9/!!H#3 9 9/!!H$3 90.96(0.24) 9/!!H%3 9 9/!!H&3 9P-sampleNonmatches 9/!x#' 93.62(0.15) 9/!x#(3 9 9/!x#)3 93.57(0.20) 9/!x#*3 9 =V,!x#+3 =E-sampleErroneousenumerations =V,!$H ,V =2.31(0.11) =V,!$H -3V = =V,!$H .3V =2.84(0.20) =V,!$H /3V =1'%$H 03  V 1'L#AЋXXaGB  C   &d!0  #XaGXЋHG##XXXXaG!=#XaGXXXBarrett,Beaghen,Smith,andBurcham(2003)alsoclassifiedtheerroneousenumerationsbytype  oferroneousenumeration.Thetypesoferroneousenumerationsaregeocodingerrors,duplicates,  andnotahousingunit. %#  21      ׀Morethanhalf(57.0percent)oftheerroneousenumerationsin2000  werebecausetheydidnotexistashousingunitsinthesearchareaoncensusday.In1990,37.3 t percentoftheerroneousenumerationswereclassifiedasnotahousingunitand33.4percentof ` theerroneousenumerationswereduplicates.Thetypeoferroneousenumerationsinoccupied L  andvacanthousingunitswereexamined.Ofthevacanthousingunits,66.0percentofthe 8  erroneousenumerationswereclassifiedasnotahousingunit.Ofoccupiedhousingunits,the $ t largestpercentageoferroneousenumerationswasfortheduplicatedhousingunits(40.7percent).  ` B  d(72,p27,R10)#XXXXaGG#dXaGXXXC  ԚCorrectlyenumeratingvacantunitscontinuestobeachallenge.Estimatesofnetundercoverage,  8  P-samplenonmatches,andE-sampleerroneousenumerationsforvacantunitsweresignificantly $  greaterthanforoccupiedhousingunits.Ourestimatesattributedalmost75percentofthevacant   erroneousenumerationstothoseclassifiedasnothousingunits.TheCensusBureauhada   difficulttimedecidingwhetheranaddressidentifiedahousingunitwhennoonelivedthere.A   proxyprovidesinformationaboutvacantunitsortheCensusBureaubasesthedeterminationof   vacancystatusonobservationfromthefieldstaff.Theproxyrespondent(ortheobservationof p thefieldstaff)maynotbesufficientlyknowledgeable,especiallyaboutvacantboardedupunits \ andunitsunfitforhabitation.Confusionastowhethertoincludeortodeletethesetypesof H vacantunitsfromthecensusinventorystillexists. 4 Smallmulti-units(twotoninehousingunitsatthebasicstreetaddress)remainproblematic.  \ Althoughnetcoverageofhousingunitsinsmallmulti-unitsimprovedover1990,small H multi-unitshadthehighestpercentofP-samplehousingunitnonmatchesandE-samplehousing 4 uniterroneousenumerationsamongtheothersizesofstructures.   InfuturecensusestheCensusBureaucouldcustomizeaddresslistbuildingoperationsand/or  censuscoverageimprovementoperationstotargetsmallmulti-unitstructureswithtwotonine  housingunitsandvacantunitstoimprovecoverage.Clearinstructionsandtrainingonwhat  unitsmeetthehousingunitdefinitionmayminimizeconfusionofwhatunitstoincludeordelete l fromthecensusaddresslist. X  #XXXXaGL#XaGXXXJones(Report0.10,2003)examinedcensushousingunitduplicationasmeasuredbytheA.C.E. 0"  Duplicationforthesecoveragemeasurementprocesseswastheamountofduplicationwithinthe #l! searcharea.Duplicationwithintheentirecountrywasnotmeasuredbythesecoverage $X" measurementprocesses.Table8dB  ԍ(73,p4,R1)C  #XXXXaGT#dXaGXXXshowsthat#XXXXaGV#XaGXXXԀabout25percentoferroneousenumerationswere $D # duplicates,adecreasecomparedtothe1990Census.Thesearchareasweredifferentinthe1990 %0!$ and2000HousingUnitCoverageStudiesasdescribedinthefootnoteonthepreviouspage. &"% Comparisonsshouldbeusedwithcaution. '#&  Oa-Oa-     *Vdd dd 'dd 'dd sdd s dd (#(#,#dd",dd",dd",dd",dd"+  +! +&    Table8:OverallPercentE-SampleHousingUnitDuplication#XXXXaG"W#XaGXXX CV2$P CYear @V/!"V @ 9!XXaGPercentofErroneous  Enumerationsthat 0 wereduplicates @V/!0"V @PercentofE-sample  housingunitsthat 0 wereErroneous 0  Enumerations @V/!  "V @Percentof   E-sample 0  housingunitsthat 0  wereduplicates @V/! "V @Estimated  numberofcensus 0 duplicates#XaGX! 9[# [Q!0"V @1990@[1990 |rB $" @1990 @ 33333@@33.433333@@|33.4 {qB $" 33333@@33.4 33333@@ ffffff@2.8ffffff@{2.8 {qA $" ffffff@2.8 ffffff@ ffffff?0.95ffffff?{0.95 ~tB $" ffffff?0.95 ffffff? -A971505-A~971,505 ~tD $" -A971505 -A @@2000@@~2000 |rBT " @@2000 @@ 8@24.88@|24.8 {qBT " 8@24.8 8@ ffffff@2.3ffffff@{2.3 {qAT " ffffff@2.3 ffffff@ = ףp=?0.57= ףp=?{0.57 ~tBT " = ףp=?0.57 = ףp=? `)$A660656`)$A~660,656 cVVDT " `)$A660656 `)$A  c'Z4*( $ "   V 4Jonesalsofound:dB  ԍ(73,p15-16,R2)C  #XXXXaGZ#dXaGXXX @  "0  Morehousingunitduplicationinsmallcitiesandruralareas. (#(# "0  Morehousingunitduplicationamongunitsinsmallmulti-unitstructuresthanamong   singleunitstructures.!(#(# "0  Morehousingunitduplicationamongvacantunitsthanamongoccupiedunits.Single d# unitsaremorefrequentlyduplicatedwhentheyarevacant.P$(#(# "0  ArelativelyhigherhousingunitduplicationpercentageonAmericanIndianreservations.(x&(#(# "0  Duplicateaddressesreferringtothesamehousingunitwereseldomidentical.P((#(# &  Jonesevaluationsuggeststhefollowingforthe2010census: (* "0  #XXXXaG?d#XaGXXXDuplicatesearchandunduplicationeffortsshouldtargetsmallcitiesandruralareas, , 'kgmulti-unitstructuresinsmallcitiesandnon-mailout/mailbackareas,andsmallmulti-unit - structuresinthelargeandmediumsizedcities..(#(# #XXXXaGh#XaGXXX"0  #XXXXaGi#XaGXXXAttemptsshouldbemadetoimprovetherecordingofalladdressinformationfortherural `!0 areasthatarenotmailoutandmailback.L"1(#(#  #XXXXaGi#XaGXXXXXaG&  4.2ConflictingHouseholds #4 j# 4#XaGXj# $$t3 Liu,Feldpausch,andSmith(2002)examinedconflictinghouseholds'$$ kidentifiedaftercompleting $&t!5 allpersonmatchingandhousingunitmatching.Aconflictinghouseholdreferstothehouseholds '`"6 atamatched,non-vacantaddressorindividualhousingunit,wheretheA.C.E.householdand 'L#7 censushouseholddonotcontainanymatchedorpossiblymatchedpeople.TheA.C.E.sample (8$8 found4,369unweightedconflictinghouseholdaddresses.Personsintheseconflicting )$%9 householdsaccountedfor1.2percentoftheP#XXXXaGXk#XaGXXXsampleand1.3percentoftheEsample.dB  ԍ(74,piv,R1)C  #XXXXaGDn#dXaGXXX *&:  +&; Oa.    #XXXXaGn#XaGXXXThecensushouseholdwasmorelikelytocontainerrorsthantheA.C.E.householdwhen  householdswereconflicting.Thisconclusionisbasedontheerrorsmeasuredbywhetherthe  peopleshouldhavebeenincludedineachhousehold.AnE-samplepersondeterminedtobe  erroneousandaP-samplepersondeterminedtobenotaresidentofthehouseholdonCensusDay t arebotherrors.Inconflictinghouseholds,theEsamplecodedahigherpercentofpeopleas ` confirmederroneousenumerationsthanthePsamplecodedasconfirmednonresidents L  (26.9percentvs.5.1percent).dB  ԍ(74,piv,R2)C  #XXXXaGo#dXaGXXX 8  OaOaLookingatthepeoplewithunresolvedresidenceorenumerationstatus,therewasahighdegree  ` ofuncertaintyinconflictinghouseholds.  L  "0  ThePsamplehadahigherpercentofpeoplewithunresolvedstatusthantheEsample $  had(30.4percentvs.26.3percent);dB  ԍ(74,piv,R3)C  #XXXXaGr#dXaGXXX (#(# "0  Peoplefromconflictinghouseholdshadasignificantlyhigherunresolvedrate   (30.4percentforthePsampleand26.3percentfortheEsample)thanpeopleinmatched   (1.8and1.5percentrespectively)andnotmatched(3.7and10.8percent)housingunits.dB  ԍ(74,piv,R4)C  #XXXXaGt#dXaGXXXp(#(# "0  Includingimputationforunresolvedresidenceandenumerationstatus,theCensusBureau H estimatedalargernumberofE-sampleerroneousenumerationsthannonresidentsamong 4 thePsample(1,355,026vs.436,900).dB  ԍ(74,piv,R5)C  #XXXXaGv#dXaGXXX p(#(# Amongtheconflictinghouseholdaddresses, %#  22      ׀dB  ԍ(74,piv,R6)C  #XXXXaG/x#dXaGXXX H "0  #XXXXaGEy#XaGXXXTheCensusBureaufoundmoreE-samplewholehouseholderroneousenumerations   (1,057)than#XXXXaGy#XaGXXXP-sample#XXXXaG]z#XaGXXXԀwholehouseholdnonresidents(646); (#(# #XXXXaGz#XaGXXX"0  #XXXXaG-{#XaGXXXTheCensusBureaufoundsimilarnumbersofP-sample#XXXXaG{#XaGXXXaddressesofwholehousehold  unresolved(1,132)andE-sampleaddressesofwholehouseholdunresolved(1,070).(#(# "0  #XXXXaG{#XaGXXXThefollowupinterviewsindicatedthatfor1,302addressesthe#XXXXaG|#XaGXXXP-sample#XXXXaGj}#XaGXXXԀhouseholdrather X  thantheE-samplehouseholdlivedatthesampleaddressonCensusDay.Therewere D! 688addresseswheretheE-samplehouseholdandnottheP-samplehouseholdlivedatthe 0"  sampleaddressonCensusDay.#l!(#(# Renters,Hispanics,Blacks,peopleofage18-29,malesofage30-49,andhouseholdsin $D # multi-unitstructureshadhigherratesofconflictinghouseholdsthantheircounterparts.dB  ԍ(74,p  iv,R7)C  #XXXXaG}#dXaGXXX %0!$   &"%  /Oa/     &   Thisevaluationsuggeststhatfor2010,theCensusBureaumaywanttodirecteffortstoreduce  conflictinghouseholdsby:  "0  Probingformultiplehouseholdslivingatanaddress.'t(#(# "0  Developingmethodstoensuredeliveryofcensusformsinmulti-unitapartmentstothe L  intendedoccupantsoftheapartment.8 (#(# "0  Improvingtrainingforcensusandcoveragemeasurementinterviewerstoidentifythe  ` correctaddress. L (#(#  XXaG[&  4.3HousingUnitFieldOperationsandInstruments #XaGX# $  #XXXXaG:#XaGXXXGreen,Watson,Smith,Barrett,Byrne,andSpratt(2003)#XXXXaG\#XaGXXXdB  ԍ(75,piv,R1)C  #XXXXaGք#dXaGXXXexaminedtheA.C.E.housingunitphase $  #XXXXaGL#XaGXXXfieldoperationsand'instrument.Todeterminehowthefieldoperationsperformedandidentify   improvements,theyfocusedonresultsfromthefollowinghousingunitoperations:   #XXXXaGɅ#XaGXXX"0  #XXXXaG#XaGXXX&  AddressListing:August1999-December1999(#(# 0  TheAddressListingrecordedinformationforallhousingunitswithinthesampleof p 'iblockclustersinIndependentListingBooks.\(#(# &  #XXXXaGJ#XaGXXX"0  InitialHousingUnitFollowup:February2000-April20004(#(# 0  TheInitialHousingUnitFollowupoccurredtogetmoreinformationonhousingunits  p thatcouldnotbematchedduringtheInitialHousingUnitMatchingoperation. \(#(# '4"0  #XXXXaGÈ#XaGXXX&  Relisting:April2000-May20004(#(# 0  TheRelistingrevisitedhousingunitsandconductedanewlistingoperationinclusters   'thattheoriginallisterhadlistedinthewrongblock. (#(# 0   (#(# #XXXXaGs#XaGXXX"0  #XXXXaG#XaGXXX&  TargetedExtendedSearch2:January2001-April2001(#(# 0  TheCensusBureauperformedthesecondTargetedExtendedSearchtoascertainifsome   ofthe'vhousingunitsdeterminedtonotexistashousingunitsonCensusDayactually l! existednearbyashousingunitsoutsidethecluster.X" (#(# 0   (#(# #XXXXaGW#XaGXXX"0  #XXXXaGa#XaGXXX&  FinalHousingUnitFollowup:March2001-May20010$"(#(# 0  Housingunitsthatwereaddedtoordeletedfromtheinventoryofhousingunitssince %l # January2000wereprocessedandtheresultswerethenusedforhousingunitestimation.&X!$(#(# '0$Ҏ#XXXXaG#AddressListingwasmoresuccessfulinmailout/mailbackareasbecausethoseareashadmore '0#& citystyleaddresses.Ruralareasaremoredifficulttolistaccurately.Thehousingunitfollowup ($' oftheA.C.E.housingunitsidentifiedhousingunitsthatshouldnothavebeenlistedinthe )%( sampleareasbecauseofgeocodingerrorortheywerenothousingunitsonCensusDay.One *%) reasonthefollowupwasnecessarywasbecausetheCensusBureaulistedhousingunitsunder +&*  Oa0Oa0     constructionandfutureconstructionwhentheCensusBureaulistedA.C.E.housingunits  betweenAugustandDecember1999.Relistingoperationswereundertakenforablockcluster  with80percentormoreofthehousingunitsgeocodedincorrectly.Thelistingofhousingunits  forthePsampleneededtobeascompleteaspossible.Lessthanonepercentoftheclusterswere t relisted. ` ThesecondTargetedExtendedSearchoperationprovidedevidencethatthereweresomehousing 8  unitsclassifiedaserroneousenumerationsduringthehousingunitfollowupthatwereactually $ t geocodingerrors.  ` ResultsfromcomparisonsofInitialandFinalHousingUnitFollowupinterviewerresponse  8  patternsverifiedsuspicionsthatsomequestionswerenotbeingunderstoodbytheinterviewers $  duringtheinitialphase.Greenet.al.recommendthatforfutureapplicationstheCensusBureau   giveasmuchtestingattentiontotheinterviewersinstrumentsastheCensusBureaudoestothe   instrumentsusedbyrespondents.Inparticular,theCensusBureaushouldconductcognitive   testingonfuturefollowupinstruments.    4 XX&  5.ConclusionandRecommendations#XX 4h#  \ The2000A.C.E.waswellthoughtoutandwelldesigned,butitencounteredsomeunexpected \ problems.TheseproblemsresultedinthefinalestimatesproducedfromtheMarch2001A.C.E. H beingdeclaredunacceptableforapportionmentpurposes.TheA.C.E.requiredapreciseand 4 accuratemeasurementofresidenceandenumerationstatus.AssumingtheCensusBureau  p continuestouseacoveragemeasurementsurveytomeasureundercountinthefuture,wehave  \ somerecommendations.Moreresearchisneededtodesigninterviewinstrumentstoaccurately H identifypeoplewhoshouldbecountedinthehousehold.Livingsituationshavebecomemore 4 complexwithmultipleresidencesandmobility.Moreresearchisneededtounderstandsources   oferror,suchasrecallerrorandlackofknowledgeonthepartofproxyrespondents,andto   devisequestionsandcategoriesthatcanbereliablyandaccuratelyreported.Intensive  questionnairedesignresearchandtestingareneededtoimprovethequalityofcoverage  measurements.(Martin,Fay,andKrejsa,2002aand2002b.)Designsrobusttothiserrorshould   alsobeinvestigated. l! Thecomplicated,overlapping,andcounterintuitivecensusresidencerulesmakeitdifficultto D#! measurecoverageaccurately.Therulesneedtobesimplifiedandempiricallyevaluated. 0$" TheCensusBureauwillcontinuetoresearchissuesdiscoveredwiththeA.C.E.,particularlythe &X!$ '\duplicatesandtheirestimationordetection.Thisresearchmayleadtodevelopmentofmethods &D"% toimprovefuturepopulationestimatesthatcombineinformationfromthecensus,A.C.E.,and '0#& theA.C.E.evaluations,includingthePersonDuplicationStudies. ($' NXaGXXXBothcensustakingandcoveragemeasurementevolveandimprovewitheachcensus.The *%) Census2000willhelprefinebothcensusandcoveragemeasurementprocessesforfuture +&* Oa1    censuses.Wecombinedrecommendationsfor2010intofourcategoriesdatacollection,survey  design,estimation,andcoveragemeasurementevaluations.   &  NXNXaG5.1DataCollection#NXaGXN #  t Reviewtheresidencerulesandcriticaldefinitions.TheCensusBureaushouldcreatesimpler t  'trulesthatcanbeunderstoodbyeveryoneastheyfillouttheircensusformsandastheyare `  appliedinthefield.Therulesshouldbeconsistentwiththewayspeoplethinkabouttheir L  residence.#XXXNXaG#NXaGXXXԀThedifficultyidentifyingerroneousenumerationsinthe2000A.C.E.mayhavebeen 8  duetothecensusresidencerules. $ t  Improvemethodstoidentifyduplicatesandremovethemfromthecensus.Peoplewithmultiple L  residencesorotherplaceswheretheycanbecountedcancauseduplication.Weneedtoidentify 8  whichduplicatesshouldberemoved. $  ЀContinuallyimproveinterviewinstrumentsusedincoveragemeasurementandtheevaluation.   TheCensusBureauneedsinstrumentdevelopmentandtestingonconceptual,recall,and  comprehensionissuesforgroupquartersresidence,multipleresidencereporting,andmover  reporting. p Reduceexpensive,timeconsumingandlabor-intensiveclericalcodingoperations.Accurate H classificationsshouldbeproducedbystandardizedquestionsintheinterviewinstrument,witha 4 reducedneedforinterpretationbythematchers.Thecodingofresponsesshouldbeconsistent  p andhavedataediting.  \ Redesignthehousingunitfollowupinstrument.TheCensusBureaudiscoveredP-sample 4 housingunitgeocodingerrorsafterpersonmatching.Therefore,thefollowupofP-sample   housingunitnonmatchesdidnotidentifyhousingunitslistedasbeingintheblockclusterin   error.  #XXXNXaG # &  4 XX5.2SurveyDesign#XX 4u#    Thinkmoreabouthowmoversaretreated.In2000theCensusBureauused ProcedureCwith "  afew' iexceptionswhereweused ProcedureA. ProcedureCmatchesthenonmoversand l#! outmoversattheCensusDayaddressandwithinthesearcharea. ProcedureCratherthan X$"  ProcedureBwasusedin2000becauseitiseasiertomatchwithinthesearcharea.One D% # problemwith ProcedureCwasinterviewswithwholehouseholdsofoutmoverswereproxy 0&!$ interviews. ProcedureBwasusedin1990wherethenonmoverswerematchedtotheCensus 'l"% Dayaddressandthesearcharea.TheinmoverswerematchedtotheirCensusDayaddress (X#& requiringcollectingtheCensusDayaddressfortheinmover,obtainingthecensusgeographyfor (D$' theaddress,andmatchingtothataddressanditssurroundingblocks.Thisinmovermatching )0%( wastimeconsumingbecausethecensusquestionnaireswereprintedfrommicrofilmforclerical *&) matching.Theentiremovermatchingprocesscouldbeimprovedin2010sincenamesare +'* Oa2    capturedfortheentirecountry.Astreamlinedprocessthatisfastandeasywouldneedtobe  developedforgeocodingandmatchingfortheinmoversin2010.TheCensusBureaumaywant  toconsiderusing ProcedureBinthefuture.  Considermakingthesearchareatheentirecountry.Then,censuscorrectlyenumeratessomeone ` bycountingthemonceinthecountry.Anerroneousenumerationoccurswhenacensuspersonis L  duplicated.Thisdesignrequiresacomputermatchingalgorithmthataccuratelymatchesthe 8  P-samplepeopletoallcensusenumerationsinthecountryandidentifiescensusduplicates.A $ t followupinterviewwouldbeneededtoidentifyE-samplenotmatchedpeoplewhowere  ` erroneouslyenumeratedbecausetheydiedbeforeCensusDay,werebornafterCensusDay,or  L  didnotlivewithintheUnitedStatesonCensusDay.Anadvantageofthesearchareabeingthe  8  entirecountryismakingtheproblemwithresidencerulesnotafactorsinceapersonisnot $  erroneouslyenumeratedwhencountedatthewronglocationaccordingtocensusresidencerules.   Alternatively,thesearchareacouldbethestateorothersmallerarea.Computermatchingwithin   asmallerareawouldbeeasierthanmatchingintheentirecountry.Thefollowupinterview   wouldalsoneedtoidentifyE-samplenotmatchedpeoplewhowereerroneouslyenumerated p withinthesearchareabecausetheydidnotlivewithinthesearchareaonCensusDay. \ Anotherdesignthatexploitsanexpandedsearchareaisanyaddressmatching.Inanyaddress 4 matchingthepersoninterviewobtainsalladdresseswheretheP-samplepeoplecouldbe  p enumerated.Theresearchshoulddetermineifanonhouseholdmembercanprovidethese  \ addresses.Analogously,thisdesignidentifiesduplicatesinthecensusbyaskingpeoplenot H matchedintheEsampleforallplaceswheretheycouldbeenumerated,whichrequires 4 geocodingandsearchingtheseaddresses.Thisdifficultprocessofgeocodingandsearchingthe   addresseswasconductedformatchingmoversXaGXXXtotheirCensusDayaddressinthe1990PES.For   2010theprocesswouldneedtobemademoreefficient.Improvedcomputertechnologyshould  makemovermatchingmoreefficient.TheCensusBureaucouldalsoconsidercollecting  alternateaddressesonthecensusquestionnaire.  Thedesignfor2010couldcombinetheautomatedandanyaddressmatching. X   &  XXaG5.3Estimation#XaGXǼ#  0"  #XXXXaGӺ#Considerasamplingplanflexibleenoughtoimplementeitherastateornationaldesign.XaGXXXBuilda 0$" '0"planintotheweighttrimmingprocedurethatincludesathresholdcriterion.#XXXXaG#XaGXXXԀImputemissing %l # characteristicsusingthesameproceduresinthecensus,especiallyfortheEsample.#XXXXaGl#Consider &X!$ linkingtheP-andE-samplefilestouseinresolvingmissingdata.Whengatheringinformation &D"% usedtoassignprobabilitiesforunresolvedpersonstatus,useinformationpertainingtothe '0#& interviewandlessdemographicinformation. ($' UsegeneralizedDSEs,amodelingapproachthatcomputestheprobabilityofcapturebasedon *%) demographiccharacteristics.UseboththePandEsamplewhendeveloping_the +&* Oa3    post-stratificationplans.Re-evaluatewhethertocorrectforcorrelationbiasintheDSEs.  Considerwhethertheuseofcomplexvariancemethodswouldbemorebeneficialinproduction  orinanevaluationoftheproductionvariances.  TheCensusBureaushouldcontinuetouseDAasacoverageevaluationtool.Forthe2010 ` Census,theCensusBureaushouldalsoinvestigatewaystomeasureuncertaintyintheDA L  estimatesofundercountandtoexpandDAestimatetomorerace/ethnicitygroups. 8   &  4 XX5.4CoverageMeasurementEvaluations#XX 4 #   ` TheCensusBureaushouldcontinuetodevelopprogramstoevaluatethecoverageperson `  estimationandconsiderhowbesttosynthesizeindividualmeasurederrors.TheCensusBureau L  ' shouldconsiderevaluatingthehousingunitcoverageestimates. 8  4 XX dOOaOaOaOa&  6.References #XX 4#   4 XX 6.1ESCAPReports #XX 4#  Bell,WilliamR., AccuracyandCoverageEvaluation:CorrelationBias,ReportB-12*,  'DFebruary28,2001.  Byrne,R.,Imel,L.,Ramos,M.,andStallone,P.,NXaGXXX AccuracyandCoverageEvaluation:Person  InterviewingResults,ReportB-5*,February28,2001.#XXXNXaG# p Cantwell,P.,McGrath,D.,Nguyen,N.,andZelenak,M.F., AccuracyandCoverage H Evaluation:MissingDataResults,ReportB-7*,February28,2001. 4 Childers,D.,Byrne,R.,Adams,T.,Feldpausch,R.,NXaGXXX AccuracyandCoverageEvaluation:  \ PersonMatchingandFollow-upResults,ReportB-6*,February28,2001.#XXXNXaG# H Davis,P., AccuracyandCoverageEvaluation:DualSystemEstimationResults,ReportB-9*,   February28,2001. !  ESCAP, ReportoftheExecutiveSteeringCommitteeforAccuracyandCoverageEvaluation #! Policy,March1,2001. $" Hogan,H., AccuracyandCoverageEvaluation:DataandAnalysistoInformtheESCAP X&!$ Report,ReportB-1*,March1,2001. D'"% Navarro,A.andOlson,D., AccuracyandCoverageEvaluation:EffectofTargetedExtended )l$' Search,ReportB-18*,February28,2001. *X%(   *D&)  [4dO4     _Robinson,J.G., AccuracyandCoverageEvaluation:DemographicAnalysisResults,Report   B-4*,March2,2001.  Starsinic,M.,Sissel,C.,andAsiala,M., AccuracyandCoverageEvaluation:Variance t EstimatesbySizeofGeographicArea,ReportB-11*,February28,2001. ` 4 XX 6.2&  ESCAPIIReports #XX 4z# 8  Adams,T.andKrejsa,E., ESCAPII:ResultsofthePersonFollowupandEvaluationFollowup 8  FormsReview,Report24,October12,2001. $ t  '8 Adams,T.andLiu,X., ESCAPII:EvaluationofLackofBalanceandGeographicErrors L  AffectingPersonEstimates,Report2,October11,2001. 8  Barrett,D.,Beaghen,M.,Smith,D.,andBurcham,J., ESCAPII:Census2000HousingUnit   CoverageStudy,Report17,October17,2001.   Beaghen,M.,Feldpausch,R.,andByrne,R., ESCAPII:AnalysisofNonmatchesand  ErroneousEnumerationsUsingLogisticRegression,Report19,September27,2001. p Bean,S., ESCAPII:AccuracyandCoverageEvaluationMatchingError,Report7, H October12,2001. 4 Bell,W., ESCAPII:EstimationofCorrelationBiasin2000A.C.E.EstimatesUsingRevised  \ DemographicAnalysisResults,Report10,October16,2001. H Bench,K., ConditioningofCensus2000DataCollectedinAccuracyandCoverageEvaluation   BlockClusters,Report14,September19,2001.   ESCAPII,"ReportoftheExecutiveSteeringCommitteeforAccuracyandCoverageEvaluation  PolicyonAdjustmentforNon-RedistrictingUses,October17,2001.   ESCAPII, GlossaryofSpeializedandTechnicalTermsUsedintheESCAPReportand X"  SupportingDocuments,2002. D#! Fay,R.,"EvidenceofAdditionalErroneousEnumerationsfromthePersonDuplicationStudy, %l # Report9,PreliminaryVersion,October26,2001. &X!$ Fay,R., EvidenceofAdditionalErroneousEnumerationsfromthePersonDuplicationStudy, '0#& Report9Revised,March27,2002. ($' Feldpausch,R., CensusPersonDuplicationandtheCorrespondingA.C.E.Enumertaion *%) Status,Report6,October13,2001. +&* Oa5    Feldpausch,R., ESCAPII:E-sampleErroneousEnumerations,Report5,March13,2002.  Keathley,D.,Kearney,A.,andBell,W., ESCAPII:AnalysisofMissingDataAlternativesfor t theAccuracyandCoverageEvaluation,Report12,October11,2001. ` Krejsa,E., ESCAPII:A.C.E.EnumerationsErrors:AnalysisofCensusDiscrepantPersons, 8  Report4,September21,2001. $ t Krejsa,E.andRaglin,D., ESCAPII:EvaluationResultsforChangesinA.C.E.Enumeration  L  Status,Report3,October15,2001.  8  Liu,X.,Byrne,R.,andImel,L., ESCAPII:AnalysisofMovers,Report15,October2,2001.   Liu,X.,Jones,J.,andFeldpausch,R., ESCAPII:AnalysisofUnresolvedCodesinPerson   Matching,Report11,October10,2001.   Mule,T., ESCAPII:PersonDuplicationinCensus2000,Report20,October11,2001. \ Nash,F., ESCAPII:AnalysisofCensusImputations,Report21,September24,2001. 4 Petroni,R., ESCAPII:Accuracyofthe2000CensusandA.C.E.EstimatesBasedonUpdated  \ ErrorComponentsTotalErrorModel,Report8,October12,2001. H Raglin,D., ESCAPII:EffectofExcludingReinstatedCensusPeoplefromtheA.C.E.Person   Process,Report13,October9,2001.   Raglin,D.andKrejsa,E., ESCAPII:EvaluationResultsforChangesinMoverandResidents  StatusintheA.C.E.,Report16,October15,2001.  Robinson,J., ESCAPII:RevisedDemographicAnalysisResults,Report1,October13,2001. X  Thompson,J.,Waite,P.,andFay,R., Basisof RevisedEarlyApproximationofUndercounts 0"  ReleasedOctober17,2001,Report9a,October26,2001. #l! Wetrogan,S.andCresce,A., ESCAPII:CharacteristicsofCensusImputations,Report22, $D # October12,2001. %0!$ Wolfgang,G.,Adams,T.,Davis,P.,Liu,X.,andStallone,P., ESCAPII:P-sampleNonmatch '#& Analysis,Report18,October11,2001. (#'   )$( Oa6    4 XX 6.3&     EvaluationReports #XX 4d#  Barrett,D.,Beaghen,M.,Smith,D.,andBurcham,J., Census2000HousingUnitCoverage  'Study,ReportO.3,February21,2003.  Bean,S., EvaluationofMatchingError,ReportN.14,June20,2002. t  Bench,K., ContaminationofCensus2000DataCollectedinAccuracyandCoverage L  EvaluationBlockClusters,ReportN.1,August22,2002. 8  Green,A.,Watson,R.,SmithD.,Barrett,D.,Byrne,R.,andSpratt,S., EvaluationofHousing `  UnitFieldOperationsandInstrumentsfortheAccuracyCoverageEvaluation,ReportN.19, L  April3,2003. 8  Jones,J., HousingUnitDuplicationintheCensus2000,ReportO.10,April8,2003.   Jones,J., PersonDuplicationintheSearchAreaMeasuredbythe2000Accuracyand  CoverageEvaluation,ReportO.16,January21,2003.  Krejsa,E., DiscrepantResultsintheAccuracyandCoverageEvaluation,ReportN.10,March \ 25,2003. H Liu,X.J.,Feldpausch,R.,andSmith,D., AnalysisofConflictingHouseholds,Report0.4,  p October15,2002.  \ Smith,D.,Barrett,D.,andBeaghen,M., AnalysisofDeletedandAddedHousingUnitsin 4 Census2000MeasuredbytheAccuracyandCoverageEvaluation,ReportO.19,   February25,2003.   Wolfgang,G.,Byrne,R.,andSpratt,S., AnalysisofProxyDataintheAccuracyandCoverage  Evaluation,ReportO.5,March19,2003.    Wolfgang,G.,Stallone,P.,andAdams,T., TargetedExtendedSearchAnalysis,ReportN.17, X"  September5,2002. D#! 4 XX 6.4OtherReferences %l #  #XX 4#Bryant,B.E.etal., AssessmentofAccuracyofAdjustedVersusUnadjusted1990CensusBase 'l"% forUseinIncensalEstimates:Recommendation,ReportoftheCommitteeonAdjustmentof (X#& PostcensalEstimates,U.S.CensusBureau,Washington,D.C.,1992. (D$'   )0%( Oa7    Byrne,R., MaRCSSpecificationsforQualityAssurance(QA)forPersonMatching,DSSD  Census2000ProceduresandOperationsMemorandumSeriesChapterS-QA-13,February7,2001.   Cantwell,P., AccuracyandCoverageEvaluationSurvey:SpecificationsfortheMissingData t Procedures,DSSDCensus2000ProceduresandOperationsMemorandumSeriesNo.Q-48, ` July9,2001. L  Childers,D., The1990HousingUnitCoverageStudy,ProceedingsoftheSectiononSurvey $ t ResearchMethods,  AmericanStatisticalAssociation,506-11,1992.  ` XaGXXXChilders,D., CoverageofHousingin1990DecennialCensus,1990DecennialCensus  <  PreliminaryResearchandEvaluationMemorandumNo.253,October7,1993, (  U.S.CensusBureau,1993.   #XXXXaG#Childers,D.R., TheDesignoftheCensus2000AccuracyandCoverageEvaluation(A.C.E.),   DSSDCensus2000ProceduresandOperationsMemorandumSeries,ChapterS-DT-1,Revised,   January26,2001. t Davis,M.C.andBiemer,P., EstimatesofP-sampleClericalMatchingErrorfroma L RematchingEvaluation,1990CoverageStudiesandEvaluationMemorandumSeries,#H-2, 8 July11,1991. $t Davis,M.C.andBiemer,P., MeasurementoftheCensusErroneousEnumerations-Clerical L ErrorMadeintheAssignmentofEnumerationStatus,1990CoverageStudiesandEvaluation 8 MemorandumSeries,#L-2,July11,1991. $ Fay,R.,Passel,J.,andRobinsonJ.G., EstimatesofCoverageofPopulationbySex,Race,and  Age:DemographicAnalysis,CensusofPopulationandHousing:1970Evaluationand  ResearchProgram,No.PHC(E)-4,Washington,D.C.:U.S.GovernmentPrintingOffice,1974.  Fay,R.,Passel,J.,andRobinsonJ.G., TheCoverageofPopulationinthe1980Census, \  EvaluationandResearchReports,PHC80-E4,Washington,D.C.:U.S.GovernmentPrinting H! Office,1988. 4"  Gbur,P.M., PESEvaluationProjectP3:EvaluationofImputationMethodologyfor  $\" UnresolvedMatchStatusCases,1990CoverageStudiesandEvaluationMemorandumSeries, $H # #C-2,July10,1991. %4!$ Hogan,H.,"The1990Post-EnumerationSurvey:OperationsandResults,Journalofthe ' #& AmericanStatisticalAssociation,88,1047-1060,1993. (#'   )$( Oa8    Hogan,H., AccuracyandCoverageEvaluationSurvey:EffectofExcludingLateCensus  Adds,DSSDCensus2000ProceduresandOperationsMemorandumSeriesNo.Q-43,  March22,2001.   Hogan,H., FiveChallengesinPreparingImprovedPostCensalPopulationEstimates,DSSD ` A.C.E.RevisionIIMemorandumSeriesPP-1,January25,2002. L  Hogan,H.,Kostanich,D.,Whitford,D.,andSingh,R., ResearchFindingsoftheAccuracyand $ t CoverageEvaluationandCensus2000Accuracy,AmericanStatisticalAssociationJoint  ` StatisticalMeetings,ProceedingsoftheSectiononSurveyResearchMethods,2002.  L  Kostanich,D., AccuracyandCoverageEvaluationSurvey:ComputerSpecificationsforPerson $  DualSystemEstimation,DSSDCensus2000ProceduresandOperationsMemorandumSeries   No.Q-48,March22,2001.NXaGXXX   Kostanich,D., #XXXNXaG#TechnicalDocumentationforMarch2001Estimates,"DSSDCensus2000   ProceduresandOperationsMemorandumSeriesNo.Q-85,April18,2003.NXaGXXX p Mack,S.,Schindler,E.andSchafer,J., PESEvaluationProjectP1:AnalysisofReasonable H Alternatives,1990CoverageStudiesandEvaluationMemorandumSeries,#A-9,July9,1991. 4 Martin,E.A., InstrumentDifferencesandtheirPossibleEffects:ComparisonoftheEvaluation  \ Followup(EFU)andthePersonFollowup(PFU)Instruments,NoteforDistributionList, H October12,2001. 4 #XXXNXaG#Martin,E.A.,Fay,R.E.,andKrejsa,E.A., AnalysisofQuestionnaireErrorsinSurvey   MeasurementsofCensusCoverage,paperpreparedforpresentationatthe2002American  StatisticalAssociation2002a.  Martin,E.A.,Fay,R.E.,andKrejsa,E.A., AnalysisofQuestionnaireErrorsinSurvey l MeasurementsofCensusCoverage,ProceedingsoftheSectiononSurveyResearchMethods, X  AmericanStatisticalAssociation,2002b. D! Mule,T. WeightTrimmingintheAccuracyandCoverageEvaluation,Proceedingsofthe #l! SurveyResearchMethodsSection,AmericanStatisticalAssociation,2001. $X" Mule,T., RevisedPreliminaryEstimatesofNetUndercountsforSevenRace/Ethnicity %0!$ Groupings,DSSDA.C.E.RevisionIIMemorandumSeriesPP-2,U.S.BureauoftheCensus, &"% April4,2002. '#& Mulry,M.andPetroni,R., ErrorProfileforPES-CasImplementedinthe2000A.C.E., )$( AmericanStatisticalAssociation,ProceedingsoftheSectiononSurveyResearchMethods,2002. |*%)  h+&* Nash,F.,"OverviewoftheDuplicateHousingUnitOperations,Census2000Informational  MemorandumNumber78,DecennialManagementDivision,U.S.CensusBureau,November7,  2000.  Nash,F.,"ResultsofReinstatementRulesfortheHousingUnitDuplicationOperations,Census ` 2000InformationalMemorandumNumber82,DecennialManagementDivision,U.S.Census L  Bureau,November21,2000. 8  NationalResearchCouncil, The2000Census:InterimAssessment,PaneltoReviewthe  ` 2000Census,ConstanceF.Citro,DanielL.Cork,andJanetL.Norwood,editors,Committeeon  L  NationalStatistics,Washington,DC:NationalAcademyPress,2001.  8  OBrien,J.L., OvercountEvaluation:HousingUnitEnumerationDuplicationStudy-Results,   PreliminaryEvaluationResultsMemorandumNo.44,February28,1983.   Olson,D., AccuracyandCoverageEvaluation:TechnicalDocumentationforMarch2001   Estimates,DSSDCensus2000ProceduresandOperationsMemorandumSeries#Q85,April p 18,2003. \ Robinson,J.G.,Ahmed,B.,DasGupta,P.,andWoodrow,K.A., EstimationofPopulation 4 Coverageinthe1990UnitedStatesCensusBasedonDemographicAnalysis,Journalofthe  p AmericanStatisticalAssociation,Vol.88,No.423,p.1061-1071,1993.  \ Sekar,C.C.andDeming,W.E.,"OnaMethodofEstimatingBirthandDeathRatesandthe 4 ExtentofRegistration,"JournaloftheAmericanStatisticalAssociation,44,100-115,1949.   Siegel,J.S.andZelnik,M. AnEvaluationofCoverageinthe1960CensusofPopulationby  TechniquesofDemographicAnalysisandbyCompositeMethods,ProceedingsoftheSocial  StatisticsSectionoftheAmericanStatisticalAssociation,Washington,D.C.:American  StatisticalAssociation,pp.71-85,1966. l Spencer,B.D.,Kearney,A.T.,Keathley,D.,Petroni,R.,Belin,T.,andMulry,M.H., D!  QuantifyingBiasfromMissingDataProceduresinthe2000A.C.E.,September26,2002. 0"  Thompson,J.H., CAPEProcessingResults,U.S.CensusBureauMemorandum,Washington, $X" D.C.,1992. $D # U.S.CensusBureau,DecennialManagementDivision, Census2000AccuracyandCoverage &"% Evaluation(A.C.E.)Estimation,ComprehensiveOperationalAssessment,December10,2002. '#& U.S.CensusBureau,DecennialManagementDivision, Census2000AcurracyandCoverage )$( Evaluation(A.C.E.)Sampling,ComprehensiveOperationalAssessment,October28,2002. |*%)  h+&* Oa:    U.S.CensusBureau, TechnicalAssessmentofA.C.E.RevisionII,  4tIO  5  http://www.census.gov/dmd/www/ace2.html6YIOmESCA  7 ,March12,2003.  West,K., PESEvaluationProject9A:AccurateMeasurementofCensusErroneous t Enumerations-EvaluationFollowup,1990CoverageStudiesandEvaluationMemorandum ` Series,#K-2,July11,1991. L  Wolter,K.M.,"SomeCoverageErrorModelsforCensusData,JournaloftheAmerican $ t StatisticalAssociation,81,338-346,1986.  ` Zuwallack,R.,Salganik,M.,Cromar,R.,andMule,V.,JR.,"FinalSampleDesignforthe  8  Census2000AccuracyandCoverageEvaluation,ProceedingsoftheSurveyResearch $  MethodsSection,AmericanStatisticalAssociation,441-446,2000.    4 XX7.Acknowledgments#XX 4A#    Wewouldliketothankthemanypeoplewhomadethisdocumentpossible.Wedrewliberally  fromthemanyreferenceddocuments.Withoutthesedocumentsourreportwouldnotexist.We  valuethecommentsreceivedfromthemanypeoplewhoreviewedourreport.Inparticular,we p wanttoacknowledgecommentsfromFlorenceAbramson,TammyAdams,ArjunAdlakha, \ MiriamBalutis,MichaelBeaghen,SusanneBean,BillBell,DickBlass,CynthiaClark,Jim H Dinwiddie,DebbieFenstermaker,HowardHogan,AnneKearney,RuthAnnKillion,Donna 4 Kostanich,ElizabethKrejsa,KenMacDonald,BetsyMartin,TomMule,MaryMulry,FayNash,  p MagdaRamos,GreggRobinson,DavidWhitford,GlennWolfgang,andMaryFrancesZelenak.  \ WealsowanttoacknowledgeJimTreat,DavidWhitford,andDarleneMoulforconductingfact H checkingandMaryUntchfortheinvaluableclericalsupportsheprovided. 4        Appendix  Inconsistenciesinraceclassifications   TheracecategoriesintheDAestimateslargelyreflecttheraceassignedtheparticular ` administrativerecordsatthetimeoftheevent(birth,death,orenrollmentinMedicare).TheDA L  estimatesofthenetundercountarebiasedtotheextentthatpeoplewhoareclassifiedasa 8  particularraceinDA(e.g.,Black)reportedadifferentraceinthecensus. $ t Theeffectofthenew markoneormoreinstructionfortheCensus2000questiononrace  L  complicatesthetraditionalcomparisonofDAestimatesbyracewithcensusracetabulations.In  8  fact,theCensus2000tabulationsdonotincludeacategory Blackthatiscomparableto1990or $  earliercensustabulations.TabulationsfortheBlackpopulationfor2000containtabulationsof   thenumberofpeoplewhoreportedBlackonlyandtabulationsofthenumberwhoreportedBlack   whetherornottheyreportedotherracesaswell.   Todealwiththereportingmorethanonerace,wepresentalternativeDAestimatesofcensus p undercountusingtwomodels:(1)Model1comparesthe2000DAestimatesforBlackswith \ Census2000tabulationsforpeoplewhoreportedBlackonly,and(2)Model2comparesthe2000 H DAestimatesforBlackswithCensus2000tabulationsforpeoplewhoreportedBlackwhetheror 4 nottheyreportedanyotherrace.Attheyoungestages,thedifferencesbetweenthetwomodels  p arethegreatest.Thetablesandfiguresshowtheaveragesofthetwomodelestimatesfor  \ comparisonwiththehistoricalDAestimatesand2000A.C.E.results.Theseaveragesarenot H necessarilythebestpointestimates;researchonthedetailedCensus2000raceandethnicitydata 4 toconductedlaterthisyearmayprovideabasisfordeterminingatwhichpointalongtheModel1   toModel2rangeofcensusracetabulationstheDAestimatemightbestbecompared.   AfinalinconsistencyaffectsracecomparisonsoftheDAandA.C.E.estimates.In1990,the9.8  millionpeople(mainlyHispanics)whoreportedtheirraceas OtherRace-NotSpecifiedinthe  censuswereredistributed(forDAestimation)tothecategoriesWhite;Black;AmericanIndian, l Eskimo,orAleut;andAsianorPacificIslandersothatthecensuscountswereconsistentwith X  theracecategoriesofthehistoricaldemographicestimates.Asimilarmodificationtomakethe D! censusracecategoriesmorecomparablewiththehistoricaldemographicdatawasagainusedin 0"  2000fortheDAestimation. #l! Theinconsistenciesintheracedataplaceevenmoreimportanceontheuseofsexratiosfor $D # makinginferencesaboutcoveragebyracialcategoriesinCensus2000.Specially,totheextent %0!$ thattheinconsistenciesinreportingandthenumbersmarkingmorethanoneraceareaboutthe &"% sameformenandwomen,theinconsistencieswilltendtocanceloutinthecalculationofsex '#& ratios.Wefoundthisassumptionheldtrue:inCensus2002,thesexratiosforpeoplewho (#' reportedBlackonlyarenearlyidenticaltothesexratiosforpeoplewhoreportedBlackwhether )$( ornottheyreportedotherraces.(Robinson,2001.)B  d(22,pp7-8,R2)dC