WPC8 p%/C34d[ #:n0MOe_Oû76cڽX i$M-t'xnh֕VC$=>,2\+eTO<+:~Zۏkrׯ÷D)`g4Xъ G>p^ Zp{HsN0[L'<V< ;(wZa׉J anP3("ClތEqI [2!u,1'7xy=MzR 'ZcdG`[#?&A͸^^/iK^C;,p$ jJf'Hμw !I?Υn]IDqyf 74j,AJ/R퀬WaIgGX`i"(Xm{DžU %;t0v9L6f+z3I<63҆U<#UN %  0(UN7^ w4 mJ 0l 0DNMMU@E 0!c 1uz 0w 0f 0 0MD 0H 0 0 1@bxffoU.!K!f "f"f""f("M*"a." 0DB"a"a"6" 0 D" 0D(# 1l#M# 1# 1~$M%M %M %M%a%)%a& 0D%&@i&M'f'UJ''M)K)f*a*f*a*f.*K0*f{*f}**pDE+fBfBfBaBaBcB BE D 57E BH*lE AYE 0DEa3F(GFaoF 72FF CFFFFFMFFFF BFaFGGMHHHHHH C H CHHHHaHaHfHaHHHHH B I A)IfIaIMI AI 0KJ D+JJfKKKaL AYL 0FM AFM D3MMK"NKmN B*NfNfNaNUNNKHOHO 0DO 0zOaQPaePUFyPU@PP BPU>QQQfZQa\Q\Q\Q\Q\Q\Q\Q\Q i[pQpQpQpQpQpQpQpQpQpQpQpQpQpQpQpQpQpQpQpQpQpQpQpQpQpQpQpQpQpQpQpQpQpQpQpQpQpQpQpQpQpQpQpQpQpQpQpQpQpQpQpQpQpQpQpQpQpQpQpQpQpQpQpQpQpQpQpQpQpQpQpQ@JQQQoRRRR?7\\CENSUS\PRED-TR2-2-HP8000-3.PRED.DECENNIAL.HQ.CENSUS0(hH  Z6Times New Roman RegularX($(hH  Z6Times New Roman Regularij. -Y3|x$U @  $    4    _ɀXXԀCaseswithaSSNthatislessthan9digitsorcontainsallninesorzerosarealsotreatedasmissing.#X#pXɀS#ɀXX#p#X#pXɀ#((3$ !      0  (#$  0   (G Z(Times New Roman (;3$2#  0  .3  0  3#37=CIQYag1.a.i.(1)(a)(i)1)a)(O;$0  2#  a  .3  0` (#(#(b$0  0` (#(#2#   .3  0 ` (#` (#(xir$0  0` (#(#0 ` (#` (#2#(  0  )3  0 (# (#($0  0` (#(#0 ` (#` (#0 (# (#2#(  a  )3  0h(#(#(F$0  0` (#(#0 ` (#` (#0 (# (#0h(#(#2#(   )3  0h(#h(#($0  0` (#(#0 ` (#` (#0 (# (#0h(#(#0h(#h(#2#  0  )3  0(#(#({$0  0` (#(#0 ` (#` (#0 (# (#0h(#(#0h(#h(#0(#(#2#  a  )3  0p(#(#xAGaeimquy}Bullet ListBullets List#O*OLE*  <5WPCNN55!5 X  .1  `& & MathTypep "-\ Times New Roman- 2 d( 2 ) 2  ( 2  ( 2  ) 2 x) [Times New Romany- 2 >(J 2 B/> 2  )JSymbol- 2 q 2 bq 2 (q 2 q |Symbol- 2 a [Times New Roman- 2 i> 2 Cj> 2  m 2 i> 2 j>Times New RomanM- 2 tk 2  VARSymbol- 2 O- 2  2 - [Times New Roman- 2 2p 2 '1p 2 2p & "Systemn-\  `Tms Rmn A 2'  _  AppendixB!0 e lTABLE G(3$ !  TABLE BTABLE C ,_   AppendixA6b$Triangle0 (3$ !  (5hCEKQW]cioAutoList1I.I.I.I.I.I.I.I.(uhCEKQW]cioAutoList5IIIIIIIIIIIIIIII(hCEKQW]cioAutoList6IIIIIIIIIIIIIIIITABLE A  $    2    _ԀɀXXOddsarecalculatedastheexponentialofthecoefficientinthemodel.#X#pXɀU#TABLE H(3$ !   6  $    1    _ԀɀXXAlthoughUAAsareexcludedfromthedenominators,theresponseratesarenotreturnratessincethe  vacancystatusoftheunitsisunknown.#X#pXɀU#kIA`2WP TypographicSymbols  SSNnonresponse SSNsmailoutmailback SQTprenoticeDillman SFC LFC DMAFA.C.E LCA HCA UAA UAAs AQE USPS PMPDSCMOpairwiseHochbergReltnshpA.C.E.undercoverage A >(  _  AppendixETABLE DЬTABLE Z A 2'  _  AppendixD?Level 1Level 2Level 3Level 4Level 5WPWin 6.0/OLE 1.0 Prefix Information MarkerEquation.COEE2ࡱ> Root EntryF*7zOle 7CompObjjBOlePart 'F! WP 7 [3] FMicrosoft Equation 2.0 DS EquationEquation.COEE29q% vN X  .1  `& & MathTypep "-\ OlePres000Equation Native Times New Roman- 2 d( 2 ) 2  ( 2  ( 2  ) 2 x) [Times New Romany- 2 >(J 2 B/> 2  )JSymbol- 2 q 2 bq 2 (q 2 q |Symbol- 2 a [Times New Roman- 2 i> 2 Cj> 2  m 2 i> 2 j>Times New RomanM- 2 tk 2  VARSymbol- 2 O- 2  2 - [Times New Roman- 2 2p 2 '1p 2 2p & "Systemn-jkCI (q i -q j )t (a/2m) (VAR(q i -q j )) )12 METAFILEPICTvvN X  .1  `& & MathTypep "-\ Times New Roman- 2 d( 2 ) 2  ( 2  ( 2  ) 2 x) [Times New Romany- 2 >(J 2 B/> 2  )JSymbol- 2 q 2 bq 2 (q 2 q |Symbol- 2 a [Times New Roman- 2 i> 2 Cj> 2  m 2 i> 2 j>Times New RomanM- 2 tk 2  VARSymbol- 2 O- 2  2 - [Times New Roman- 2 2p 2 '1p 2 2p & "Systemn-'ĜTABLE PTABLE K Y  $    3    _ԀɀXXThesimultaneoussignificanceofthegeneralandspecificnotificationwastestedbysummingthe e parametersandcomparingtheresulttozeroinaFtest(H0:_ specific_ +_ general_ =0).TheresultsshowthatF=4.593atp Q =.033.#X#pXɀU#dB  ԍ(17,pages2and5,R6)C  d   `   dXXXX)!dxdx&0 d dZO  , deUU(3$ !  TABLE FTABLE Li)dd  dTABLE ACLevel 1Level 2Level 3Level 4Level 5Level 1Level 2Level 3Level 4Level 5d dTABLE ETABLE AATABLE AB% %dNZ I&mage <=8C HKKKKJTABLE QNZ&OLE 2.0 Box <=8C HKKKK( $ Figure  1  'dxd쀄Level 1Level 2Level 3Level 4Level 5Level 1Level 2Level 3Level 4Level 5TABLE ADZM << deUU(n$ (  1  ) NZ&OLE 2.0 Box <=8C HKKKKVVVV'dxd A >(  _  AppendixC A >(  _  AppendixB&P d dTABLE J(hH  Z 6Times New Roman Regular A 2'  _  AppendixC($$  0  (;2r$ !  XX  0  .TABLE OTABLE A<4 9Z .Courier New Regular)0rZ (Garamond AntiquaK Kd\  `&Times New RomanTABLE AViiiKCC$ H:\MURAWSKI\LATESTMO\1CPOSTAG.TIFWPC   !!!"""###$$$%%%&&&'''((()))***+++,,,---...///000111222333444555666777888999:::;;;<<<===>>>???@@@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?????//??_/o߯o@DH:\MURAWSKI\LATESTMO\1CPOSTAG.TIFH:\Murawski\Latest Models\1cpostag.tifT4b̤"ͤͤnJ' ']&I{s u& !  _e   XX  0  0` (#(#0 ` (#` (#0 (# (#0h(#(#0h(#h(#0(#(#0p(#(#Census2000Testing,p(#p(# 0  0` (#(#0 ` (#` (#0 (# (#0h(#(#0h(#h(#0(#(#0p(#(#Experimentation,:p(#p(# 0  0` (#(#0 ` (#` (#0 (# (#0h(#(#0h(#h(#0(#(#0p(#(#andEvaluationProgram$tp(#p(#   AV) xdVE]xA ^ 0  0` (#(#0 ` (#` (#0 (# (#0h(#(#0h(#h(#0(#(#0p(#(#November13,2001mp(#p(# #X#pX E#G+XX#p   #X#pXG+!#xQvXX#pC AnalysisoftheSocialSecurityNumber   NotificationComponentoftheSocial   SecurityNumber,PrivacyAttitudes,and   NotificationExperiment#X#pXvxQt# ?   X#pXXX#p XX#pFINALREPORT # #X#pX # XX#p#  #  #X#pX P##X#pXXX#p#?+ ` hp x X?X#pXXX#pThisresearchpaperreportstheresultsofresearchandanalysisundertakenbytheU.S.Census G Bureau.Itispartofabroadprogram,theCensus2000Testing,Experimentation,andEvaluation 3 (_TXE_)Program,designedtoassessCensus2000andtoinform2010Censusplanning.Findings o fromtheCensus2000_TXE_ԀProgramreportsareintegratedintotopicreportsthatprovidecontext  [ andbackgroundforbroaderinterpretationofresults. XX#p G #  7##X#pX  #   `     h      XX#p0  0` (#(#0 ` (#` (#0 (# (#0h(#(#0h(#h(#0(#(#0p(#(##X#pX g # XX#pJenniferA._Guarino_," p(#p(# 0  0` (#(#0 ` (#` (#0 (# (#0h(#(#0h(#h(#0(#(#0p(#(#JoanM.Hill,and#X#pX A #u XX#p#Ep(#p(# u  u#X#pX u # XX#p0  0` (#(#0 ` (#` (#0 (# (#0h(#(#0h(#h(#0(#(#0p(#(#HenryF._Woltman_/% p(#p(# AV) xdMEh&xA i&! #   ##X#pX  #0  0` (#(#0 ` (#` (#0 (# (#0h(#(#0h(#h(#0(#(#0p(#(# XX#pPlanning,Research,and x&! EvaluationDivision#X#pX ;#'#p(#p(#   8#pXXdd8q=-)`~0+<3 `(*E+<3<3E+q )(%  _ TABLEOFCONTENTS     EXECUTIVESUMMARY......................................................................................................iv  1.0  BACKGROUND...........................................................................................................1L (#(# 2.0  METHODOLOGY.......................................................................................................4 `(#(# 3.0  LIMITATIONS...........................................................................................................12$ (#(# 4.0  RESULTS....................................................................................................................13 (#(# 5.0  RECOMMENDATIONS............................................................................................23\(#(# AppendixA..............................................................................................................................26  p AppendixB..............................................................................................................................27 4    ` AppendixC..............................................................................................................................28   AppendixD..............................................................................................................................30  l  AppendixE..............................................................................................................................35  0"      Thispageintentionallyblank     EXECUTIVESUMMARY   h      p     u XX#p #X#pX u#X9 XXX#pThepossibilityofusingadministrativerecordsfromotherfederalgovernmentagenciesto  supplementcensusdatahasbeeninvestigatedforsometimeattheCensusBureau.Theuseof  administrativerecordscouldpotentiallyincreasecompletenessofmeasurementbyreducing  respondentburdenwithshorterquestionnairesandimprovedataqualitybyeliminating t  memory/respondenterrors.Torealizemaximumbenefits,SocialSecurityNumbersforeach `  individualareneededtolinkresponsestoadministrativedata.#X#pX XX9(#Thepurposeofthisanalysisisto L  assesstheeffectsofSocialSecurityNumberrequestsanddifferentnotificationsofadministrative 8  recorduseoncensusresponsebehavior,specificallyformreturnandformcompleteness. $ t  InadditiontotheSocialSecurityNumberrequestsforPerson1andallhouseholdmembers,two L  notificationsaretested.Generalnotificationofadministrativerecorduseinformsthehousehold 8  thatcensusdatamaybelinkedtodatafromotherfederalgovernmentagencies,whilespecific $  notificationgoesfurthertonametheagenciesfromwhichdatamaybesoughtforlinking.   ThisstudyisdesignedtodetermineifSocialSecurityNumberrequestsandadministrativerecord  usenotificationaffectcensusresponse,formcompletenessamongformsreturned,andresponse  totheSocialSecurityNumberitem.Moreover,ifsuchaneffectexists,thisstudyinvestigates p whethertheeffectdiffersdependingonthe_ subpopulation_ Ԁtowhichthetreatmentsare \ administered. H Historically,experimentalanalysesconductedbytheCensusBureauandtheircontractorshave  p involvedaseriesofpairwisecomparisonstomeasurestatisticaldifferencesinresponserates  \ amongtheexperimentaltreatments.TheexperimentaldesignusedinSPANcalledfora H differentandmoresensitivemethod.Throughoutthisreport,logisticregressionanalysisisused 4 todeterminetheeffectofeachtreatment(SSNrequest,notificationofadministrativerecorduse)   onresponse.   LookingattheeffectoftheSocialSecurityNumberrequest,theresultsindicatethat:   (Ax3"3"  3$2A3  0    ThepresenceofaSocialSecurityNumberrequestforoneorallhouseholdmembers l! resultsinasmall,yetsignificant,decreaseinmailresponsetoCensus2000.Logistic X"  regressionresultssuggestthattheSocialSecurityNumberrequestforallhousehold D#! memberswoulddecreaseresponseby2.1percentinhighcensuscoverageareasand2.7 0$" percentinlowcensuscoverageareascomparedtonorequest.3$=$݌%l #(#(# Ќ  "3"  3&2A3  0    TheslightdecreaseinresponseduetotheSocialSecurityNumberrequestisnot &D"% statisticallydifferentbetweenlowcoverageareas(expectedtocontainahighproportion '0#& oftheBlackandHispanicpopulationsandrenters)andhighcoverageareas.3&&݌($'(#(# Ќ  "3"  3(2A3  0    ThedeclineinresponseduetotheSocialSecurityNumberrequestismodestconsidering *%) pastqualitativeresearchassessingtheextenttowhichpeopleopposedthisrequest (Singer +&* etal(1992);AguirreInternational(1995),Singer(forthcoming)).However,thedeclineis |,'+ comparabletothe3.4percentdropinresponseduetotheSocialSecurityNumberrequest  citedinthe1994SimplifiedQuestionnaireTest(Dillmanet.al.(1994)).3((݌(#(# Ќ  "3"  3E+2A3  0    HouseholdsreceivingtheSocialSecurityNumberrequestshavehigherratesof t incompleteformreturncomparedtootherhouseholds.3E+r+݌`(#(# Ќ  Thefindingsfromtheanalysisofnotificationofadministrativerecorduseonresponseandform 8  completenesssuggestthat: $ t "3"  38-2A3  0    Includinggeneralnotificationofadministrativerecorduseinthecoverletterwiththe  L  censusformcausesasmall,butsignificant,decreaseinresponse,whiletheinclusionof  8  specificnotificationdoesnot.38-e-݌$ (#(# Ќ  "3"  3.2A3  0    Whenadministeredsimultaneously,thecombinationoftheSocialSecurityNumber   requestandspecificnotificationofadministrativerecordusedecreasesresponse(70   percent)comparedtotheconditionwherespecificnotificationistheonlytreatment(73.5   percent).3./݌p(#(# Ќ  "3"  302A3  0    Interestingly,itemnonresponsetothePerson1SocialSecurityNumber(15.5percentfor H Person1only,15.8percentforallhouseholdrequest)issignificantlyhigherwhenno 4 notificationisincludedinthecoverlettercomparedtothecasewhenspecific(11.5  p percent)orgeneral(12.6percent)notificationispresent.301݌ \(#(# Ќ  Basedonthesefindingsthefollowingrecommendationsaremade: 4 "3"  3r32A3  0    IfSocialSecurityNumbersarerequestedinfuturecensuses,theCensusBureaushouldbe   preparedforadecreaseinmailresponse.3r33݌(#(# Ќ  "3"  342A3  0    ThedeclineinmailresponseduetotheSocialSecurityNumberrequestcouldbegreater  thantheseresultsconveyifexpandingtherequesttotheentirenationfuelsapublic l debateoverprivacyconcerns.344݌X (#(# Ќ  "3 "  3l62A3  0    IfmorecompletedataonSocialSecurityNumbersaredesiredabovehighermail 0"  response,notification(generalorspecific)ofadministrativerecorduseshouldbe #l! includedinthecoverletterwiththecensusform.3l66݌$X"(#(# Ќ   T,'+ )O1.0  BACKGROUND(#(#  X#pXXX#pTheCensusBureauundertakesaprogramofexperimentationduringdecennialcensusesto  measuretheeffectivenessofnewtechniques,methodologies,andtechnologiesinthespecial t environmentthatadecennialcensusgenerates,suchasmasstemporaryhiring,promotionand ` outreachincoordinationwithlocalgovernments,thenationalpaidadvertisingcampaign,andthe L  nationwidedistributionofCensus2000publicuseforms.Resultsfromexperimentsform 8  recommendationsforsubsequenttestingandultimatelyhelpdesignthenextdecennialcensus $ t (_ Neugebauer_ ,1999).  ` Decennialcensusesbeginningin2010mayrelyonexpandeduseofadministrativerecords  8  informationobtainedfromotherFederalagencies.Becausetheuseofadministrativerecordshas $  wideimplicationsondecennialmethodology,itisimportanttocollectbehaviorandattitudinal   dataonseveraltopics.TheyincludehowthepublicrespondstorequestsforSocialSecurity   numbers(SSNs)oncensusquestionnaires,howthepublicrespondstodifferentlyworded   notificationsabouttheCensusBureaususeofadministrativerecords,andthe#X#pXXX#p8#X#pXXX#ppublicsattitudes#X#pXXX#p]>#X#pXXX#p   onprivacyandconfidentialityissuespertainingtotheuseofadministrativerecordsinadecennial p census(_ Neugebauer_ ,1999)#X#pXXX#p>#X#pXXX#p. \ TheSPANexperimentconsistsofthreemajorcomponentstoachievetheresearchobjectives. 4 Thefirstcomponentusesalistassistedrandomdigitdial(_ RDD_ )telephonesurveytocollectdata  p onthepublicsprivacyconcerns;itisreferredtoastheStudyofPrivacyAttitudesin2000  \ (SPA2000).#X#pXXX#p?#X#pXXX#pThesecondanalyticalcomponentinvolvesthevalidationofSSNscollectedfrom H fourexperimentalpanelsthatrequestit.ThisevaluationexamineswhatpercentageofSSNs 4 obtainedintheexperimentarevalidbypanel.#X#pXXX#pA#X#pXXX#pThethirdcomponentanalyzestheeffectsof   differentnotifications,twostrategiesforobtainingSSNinformation,andnotificationcombined   withtheSSNrequestonresponsebehaviorandiscalledtheSSNNotificationcomponent  (_ Neugebauer_ ,1999).#X#pXXX#pB#Thisreportcontainsafullanalysisofthethirdcomponent.  X#pXXX#pSpecificresearchobjectivesoftheSPA2000componentofSPANincludedeterminingthe l publicsopinionoftheFederalgovernmentandtheCensusBureauingeneral,ofexpandinguses X  ofadministrativerecords,andoftheCensusBureauspossibleinterestincollectingSSNsinthe D! future.SPA2000isconductedattwodifferenttimes;oncebeforethepaidadvertisingand 0"  promotionprogramandtheenumeratorrecruitingprogrambegan,andonceimmediatelyafter #l! CensusDay(_ Neugebauer_ ,1999)#X#pXXX#pD#X#pXXX#p.Comparingresultsbetweenthesetwotimeperiodsindicates $X" whetherthe censusenvironmenthasaneffectonthepublicsprivacyattitudes.Each $D # measurementgroupcomprisesanational_ RDD_ Ԁsampleof2,000households.Thepre %0!$ measurementsurveyoccurredfromJuly14,1999toOctober17,1999.Datacollectionforthe &"% post-measurementsurveybeganonoraboutApril8,2000andconcludedaroundJuly12,2000. '#& #X#pXXX#pG#NotethatSPA2000resultsareprovidedasaseparatesectionofthefullreport. (#' TheSSNvalidationcomponentwilladdressthefollowingquestions:ArereportedSSNs |*%) accurate?WhatarethecharacteristicsofhouseholdsthatprovideanddonotprovidetheSSN? h+&* WhatistheeffectoftheshortformgeneralnotificationonSSNreporting?Whatistheeffectof T,'+ theshortformspecificnotificationonSSNreporting?Thefinalreportinvolvesthevalidationof  SSNscollectedfromthefourpanelsthatrequestit.Thisevaluationexamineswhatpercentageof  SSNsobtainedintheexperimentarevalidbypanel.InadditiontocountsofverifiedSSNs,the  validationprocessalsoprovidescountsofthenumberofimpossible_ SSNs_ --SSNsnotfoundon t theSSAs_ NumIDent_ ԀFile--andthenumberofpartiallymatched,forexample,casesthatmatched ` Nameonly,NameandDateofBirth,orNameandSex.Exactmatchesobviouslytranslateto L  validationoftheSSN,whereasthepartialorequivocatedmatchesarereferredtoasindirect 8  validationX#pXXX#p(_ Neugebauer_ ,1999)#X#pXXX#p"O#.NotethattheSSNvalidationresultsareprovidedinaseparate $ t sectionofthefullreport.  ` Thegoalofthe_ SSN_ Ԅnotificationcomponent,andhencethispaper,istoassesstheeffectsof  8  differentnotificationsandrequestsforSSNinformationonoverallresponsetoCensus2000and $  dataqualityassuggestedbyformcompleteness.Todate,noempiricalresearchhasmeasuredthe   effectsofaSSNrequestorpublicnotificationofadministrativerecorduseonmailresponseor   formcompletenessinadecennialcensusenvironment,althoughresearchhasbeenconducted   duringmidcycletests.The_ SSN_ Ԅnotificationcomponentisapanelexperimentdesignedto   comparemailresponseandformcompletenessratesamongvariouspanelswiththe p aforementionedtreatments. \ 1.1PastResearch 4 Paststudiesintheprivacyandconfidentialityrealmshowthatpeoplewhoaremostconcerned  \ withprivacyparticipatelessinsurveysandcensusesthanthosewhoarenotconcerned(_ Kulka_ , H Holt,Carter,andDowd,1991;Singer,_ Mathiowetz_ ,and_ Couper_ ,1993;GatesandBolton,1998). 4 Tostudythisphenomenon,qualitativeandquantitativeanalyseshavebeenconductedtoassess   publicopinionandresponsebehaviortoSSNrequestsoncensusforms.Whilethequalitative   researchsuchasthe1992focusgroupsindicatedextremenegativereactiontoaSSNrequest,a  _ mailout_ /mailbacktest(the1992SimplifiedQuestionnaireTest(SQT))indicatedasmallerthan  anticipatedactualdecrease(3.4percent)inmailresponserates(Dillmanetal(1994);Singeretal  (1992);AguirreInternational(1995)).Notethatthisdecreaseoccurredinconjunctionwitha l shorter,respondentfriendlyquestionnaire,aprenoticeletter,areminderpostcard,anda X  replacementquestionnaire.Also,amongrespondentslistedontheSSNcensusform,justover1 D! in10failedtoprovideaSSN(Bates,1992).Thesefindingswereunexpectedandseemingly 0"  contradictedtheanticipatedextenttowhichrespondentswouldresistprovidinganidentifierwith #l! datalinkingimplications. $X" Forfurtherinvestigation,aquestionaskingrespondentswillingnesstoprovidetheirSSNson %0!$ censusformswasincludedinaseriesofsurveysaimedatmeasuringprivacyattitudesofU.S. &"% residentsovertime.Singer(forthcoming)reportsthatthepercentageofrespondentswillingto '#& providetheirSSNonacensusformdeclinedfrom68%in1996to55%in1999and56%in (#' 2000.Thedropinwillingnesswassignificantbetween1996and1999,buttherewasnofurther )$( significantchangebetween1999and2000. |*%) Pastresearchonnotificationofadministrativerecorduseisqualitativeinnatureandtherefore T,'+ doesnotindicatetheeffectofnotificationoncensusresponse.Pastfindingsrevealthatfocus  groupparticipantsaregenerallyunsureaboutwhateffectnotificationwillhaveoncensus  response.Somebelievethatnotificationofadministrativerecordusewillhavenoeffecton  response,whileothersbelievethatnotificationwilldecreaseresponse.Withregardtothetypeof t notification,focusgroupsadministratorsnotethatmanyoftheparticipantsdidnotunderstandthe ` taskofratingwhichnotificationwasmostpersuasiveinincreasingparticipation,andinstead L  ratedthenotificationspecimensbywhichuseofrecordstheyfeltwasmostjustifiable(Aguirre 8  International,1995). $ t 1.2Hypotheses  L  Giventhepastresearch,severalapriorihypothesesweredevelopedpriortothedataanalysis. $  Forthetreatmentsforwhichpastresearchislimited,hypothesesweredevelopedbasedon   expectationsfromprivacyresearch.   1.WithregardtotheSSNrequest,itishypothesizedthatmailresponserateswilldropwhena   SSNrequestispresent,withalargerobservableeffectinareasoftypicallylowcensuscoverage, p whereresponseisalreadylow,comparedtohighcoverageareas. \ 2.Furthermore,itisanticipatedthattherequestforSSNwillincreasetheamountofincomplete 4 formsreturnedcomparedtonoSSNrequest.  p Thereislittleguidancefrompastresearchdrivingthehypothesesabouttheeffectofnotification H onresponse. 4 1.Intheabsenceofpastquantitativestudies,wesuspectthatnotificationofadministrative   recordusewillcausesignificantdropsinmailresponseandincreasesintheamountof  incompleteformsreturned,withspecificnotification(includingagencynames)havingastronger  effectthangeneralnotification.  2.Moreover,itisexpectedthatrequestingSSNinadditiontoprovidingeithertypeof X  notificationwilldecreaseresponsecomparedtoprovidingnotificationalone. D! 3.Withrespecttoitemnonresponse,wesuspectthattheSSNitemforPerson1willbemissing #l! atahigherratewhengeneralorspecificnotificationisincludedwiththeSSNrequest. $X" 4.Lastly,webelievethatnotificationofadministrativerecordusewillincreasetheamountof %0!$ incompleteformsreturnedinamorepronouncedwaywhencoupledwiththelongform &"% comparedtotheshortform,duetoprivacyissuesraisedpubliclyregardinglongformquestions '#& duringCensus2000.B  d(1,page1,R1)dC   (#' ЀThe_ SSN_ Ԅnotificationcomponentprovidesabetterunderstandingofthepotentialramificationsof |*%) requestingSSNonbehaviorregardingquestionnairereturn,especiallyinadecennialcensus h+&* environment.3 X,[XX3 T,'+ Ї 2.0  0(#(#METHODOLOGY (#(# 2.10  PanelDesign(#(# Theexperimentaltreatmentsareimplementedwithintenpanelsinthisexperiment.Households ` selectedforthisexperimentwererandomlyassignedtoeachpanel.Thepanelsdivideintoseven L  shortandthreelongformpanels.Generaldescriptionsofthepanelsarelistedbelow: 8  1b  BX,X X,[XB X1bl0 4   Twoshortformandtwolongformpanelshavedifferentlywordednotificationson  ` administrativerecorduseinthecoverletteraccompanyingtheform.1bl m݌ L 4(#4(# Ќ  Xm1b  X1bMn0 4   TwoshortformpanelshaveformsmodifiedwithaSSNrequesteitherforall $  householdmembersorforonlythepersoncompletingtheform(i.e., PersonOne).   Notification,beyondthestatementinformingrespondentsthatprovidingSSNis   voluntary,isnotapartofthesepanels.1bMnvn݌ 4(#4(# Ќ  1b  1bCp0 4   TwoshortformpanelscombinethenotificationaspectandSSNrequestforall p householdmembers.1bCp^p݌\4(#4(# Ќ  1b  1bDq0 4   Therearetwocontrolpanels,oneshortformandonelongform.1bDq_q݌44(#4(# Ќ  Morespecifically:  \  4  ShortFormPanels z        *   LongFormPanels 4  &13%XX#pCONTROLFORM(SFC)0  0(#(#0(#(#0*(#(#0*(#*(#CONTROLFORM(LFC) (#(# (1)AllSSNRequest0 z 0z (#z (#0(#(#0(#(#0*(#(#0*(#*(#(7)SpecificNotification(#(# (2)One(Person1)SSNRequest0  0*(#(#0*(#*(#(8)GeneralNotificationl(#(# (3)AllSSNRequest,GeneralNotification D (4)AllSSNRequest,SpecificNotification l (5)GeneralNotification D  V(6)SpecificNotification V}u#X#pX%&13r#    XhnEachpanelreceivesthefullcomplementofcensusmailoutmaterialsinthesamesequenceand "! timingastheofficialCensus2000schedule.Experimentallettersandformsaretheofficial |#" censusformsreceivedbythesampledhouseholds(seeAppendixAfordetails). h$# Therearetwonotifications,referredtoas generaland specific.Thenotificationiswrittenin @&!% thelettersaccompanyingthequestionnairesanddescribeshowandwhytheCensusBureaumay ,'|"& useadministrativerecordsdatafromotherFederalagencies.Thegeneralnotificationmentions (h#' theCensusBureauspossibleuseofstatisticaldatafromotherFederalagencies,whilethe )T$( specificnotificationgoesfurthertonametheFederalagencies.Thegeneralnotificationis: )@%)  XɀXX#pToimprovethequalityofc_ ensus_ Ԁstatistics,theCensusBureausometimesusesrecordsfromother +'+ governmentagencies.Usingotheragenciesrecordshelpsmakethecensusmorecomplete.Bymaking ,', betteruseofgovernmentrecordsthatalreadyexist,theCensusBureaumaybeabletoaskyoufewer  questionsinthecensus.#X#pXɀy#ɀXX#p t #X#pXɀ{# XyThespecificnotificationwordingis: L  XE" ,Xrm ,hXX,XEɀXX#pToimprovethequalityofcensusstatistics,theCensusBureausometimesusesrecordsfromother $ governmentagencies,suchastheSocialSecurityAdministration,theInternalRevenueService,oragencies   providingpublichousingassistance.Usingotheragenciesrecordshelpsmakethecensusmorecomplete. \  Bymakingbetteruseofgovernmentrecordsthatalreadyexist,theCensusBureaumaybeabletoaskyou  p fewerquestionsinthecensus.#X#pXɀ|#ɀXX#p  4  IX,X ,Xrm ,hX\|I#X#pXɀ~# XI|BecauseprovidingtheSSNisvoluntary,thecoverletterfortheshortformpanelswiththeSSN l   requestcontainsanadditionalstatement: X   XɀXX#pE" ,Xrm ,hXX,XEToimprovethequalityofcensusstatistics,theCensusBureausometimesusesrecordsfromother 0  governmentagencies.Forthatpurpose,weareaskingforyoursocialsecuritynumber;however,providing D  yoursocialsecuritynumberisvoluntary.IX,X ,Xrm ,hXI    XW#X#pXɀj#NotethatthecoverletterforshortformpanelswiththeSSNrequestandstandardnotification, @ whichisusedinthecontrolpanels,issimilartotheofficialCensus2000materials,butcontains ,| theadditionalstatement. h 2.2SampleDesign @         "  N ThesamplewasselectedonSeptember9,1999fromtheJuly1999versionoftheDecennial  MasterAddressFile(DMAF)_ mailout_ /mailbackuniverseof92,575,792addresses,which  excludessamplesfortheAccuracyandCoverageEvaluation(A.C.E.)listing,thecontamination  evaluation,aswellascongressionaladdresses.Becausethesamplinguniversecomprisesonly  _ mailout_ /mailbackareas,list/enumerateandupdate/leaveareasareexcludedfromtheSPAN. x Table1belowcontainstheuniversesizeatthepointthatthissamplewasselected. d Thesampleisequallyallocatedtotwostratathatreflectanticipateddifferencesintheraceand < tenurecompositionofthepopulationand,basedonpreviouscensusexperience,differencesin ( x  theCensus2000mailreturnrates.Ingeneral,censusquestion_ naire_ Ԁdesignresearchprojectsprior !d! to1993denotedthesestrataas lowresponse(_ LRA_ )and highresponse(_ HRA_ )areasbasedon "P" stratificationatthe1990districtofficelevel.ForCensus2000,strataarebasedon1990census "<# tractlevelraceandtenuredataandaredenotedaslowandhighcoverageareas(LCAandHCA #($ respectively).TheLCAstratumisexpectedtocontainaveryhighproportionoftheBlackand $ % Hispanicpopulationsandrenteroccupiedhousingunits.TheHCAstratumcontainsthe %!& remainingaddresses.TheHCAstratumcomprisesapproximately81%ofthetotalDMAF &!' universeasofSeptember9,1999.dB  ԍ(2,page9,R1)C  d '"(  Table1.X#pXXX#pUniverseSize:_ Mailout_ /MailbackAreas*  `)$* *Vddd Xdd Xdd X(#(#,dd ,dd ,dd ,dd +  $L*%+L*% $FormType '*&, 'HCA '*&- 'LCA '*&. 'Total *&/ Short  0,'0 63,378,681(68.5%)  0,'1 15,184,672(16.4%)  0,'2 78,563,353(84.9%)  0,'3 Long  d 11,336,039(12.2%)  d 2,676,400(2.9%)  d 14,012,439(15.1%)  d Total   74,714,720(80.7%)   17,861,072(19.3%) ;1  `A92575792`A;92,575,792>42 `A92575792  `A >ɀXX#p*Excludessamplesselectedfor_ A.C.E.listing_ Ԁsample,A.C.E.contaminationevaluation#ɀɀŋ#ɀ,andcongressionaladdresses.#X9 X#    #X#pX XX9#Allfiguresinthisreportareweightedtoaccountfor_ oversampling_ ԀoftheLCAstratum.The `   inverseofthesamplingintervalforeachstratumwithinapanelistheweightforeachcase L   containedinthatpanelandstratum. 8   2.3IssueswiththeSPANShortFormControl(SFC)Panel  ` AspartofCensus2000datacollection,thedataprocessingteamcompiledcheckinrates 8  showingthenumberofformsreturnedonadailybasisforallexperimentalpanels.Fromthe $  veryfirstdaysofprocessing,theSPANSFCpanelwasseriouslylaggingbehindallother   experimentalpanelsandCensus2000intermsofthenumberofformsreturned.Additionally,   preliminarymailresponserates,computedwiththeinclusionofUAAs,revealedthattheSPAN   SFCresponseratewasabout10%lowerthanotherpanels.dB  ԍ(3,page2,R1)C  dFurtherinvestigationrevealedthat  theUAAratefortheSPANSFCwasaround19%whileitwas8to12%forotherpanels(see p Table1).dB  ԍ(4,page2,R1)C  dInvestigationintotheseissuesrevealedthatthemailingofthisparticularpanelwas \ potentiallyproblematic.Therefore,analternatecomparisongroupswassoughtforthis H experiment. 4 OtherexperimentssuchastheAlternativeQuestionnaireExperiment2000(AQE2000)andthe  \ ResponseModeandIncentivesExperiment(_ RMIE_ )areimbeddedinCensus2000inadditionto H SPAN.Eachexperimentcontainsitsowncontrolpanelinwhichhouseholdsreceivethenon 4 experimentalstandardCensus2000formsthroughmailout.Duetoinherentproblemswiththe   SPANshortformcontrolpanel,theshortformcontrolgroupfortheAQE2000hasbeen   substitutedfortheSPANshortformcontrolpanelinthisanalysis.TheAQE2000shortform   controlpanelhasthesamesamplesizeandallocationtostrataastheSPANcontrolpanel.dB  ԍ(2,page9,R2)C  d !         " 2.4EffectiveSampleSize l # Themailoutsamplesizeforeachpanelwasalittleover5,200addresses,totalingabout52,000 D"% householdsintheUnitedStatesselectedforthe_ SSN_ Ԅnotificationcomponentofthisexperiment 0#& (SeeTable2formoredetail).Thesamplewasrandomlyassignedtoeachexperimentalpanel $l' andisallocatedequallytotheHCAandLCAstrata.Thisallocationisnotnecessarilyoptimum %X ( forestimatingtheoverallresponserates,butisdesirableifreliableestimatesofresponseratesare %D!) examinedbystratum.TheUSPSreturnedabouteighttotenpercentofexperimentalforms(see &0"* AppendixC)thatweremailedineachpanelasundeliverableasaddressed(UAA).Thesecases '#+ neverhadtheopportunitytorespondandarethereforeexcludedfromthedenominatorofthe ($, responserates. )$- TherewerefourcasesacrossallSPANpanels(2intheSPANSFC,and2elsewhere)forwhich |+&/ thehouseholdresponsewassubmittedviatheInternet(SeeTable2).dB  ԍ(4,page1,R2)C  dSinceresponseviathe h,'0 Internetwasnotanexperimentaltreatment,thesehouseholdsdiscoveredtheCensusInternetsite  ontheirown.Thesecasesareexcludedfromtheanalysissincetherespondentshadother  motivationsforrespondingasidefromtheSPANexperimentaltreatment.  DuetoerrorintheUAAflagontheDMAF,fifteenrespondentcasesacrossallofthepanels ` includedinthisexperimentweremistakenlylabeledUAAs(NotethattheUAAratesin L  AppendixCdonotincludethesecases).dB  ԍ(4,page3,R3)C  ԚdFurtherinvestigationofthesecasesrevealedno 8  systematicerrorintheassignmentoftheUAAflag,pointingtotheconclusionthatthesecases $ t weremistakenlymarkedUAAsduringdatacaptureoperations.Assuggestedbythedata  ` processingteam,itispossiblethatacasecouldbelabeledUAAeventhoughCensusreceiveda  L  completedformifthequestionnairesorterinthedatacapturecenterwassetincorrectlytoUAA.  8  Thistypeofsituationoccurredinthe1998DressRehearsal.dB  ԍ(5,pages1and2,R1)C  dThesecasesareincludedas $  respondentsforthisanalysisandwerenotsubtractedfromthedenominatorintheresponserate   calculationsasweretheotherUAAs.dB  ԍ(6,page1,R1)C  d       Moreover,there weretwocasesacrossalloftheSPANpanelsforwhichtherewereduplicate   records.Thisproblemmaybeattributedtothefactthatthequestionnairesweremistakenly t printedwithaleading 1"instreetaddressnumber.Census2000experimentalquestionnaires ` withaleading 1"weresupposedtobedestroyedandreprintedcorrectly.However,areport L fromadataadministratorindicatesthatsomeerroneouslyaddressedquestionnairesmayhave 8 beenmailedtohouseholds,basedonobservationsattheNationalProcessingCenterin $t _ Jeffersonville_ ,Indiana.dB  ԍ(7,page1,R1)C  dForthepurposesofthisanalysis,anycasewithaduplicaterecordis ` excludedsinceitisnotpossibletodeterminewhichrecordcontainsthecorrectinformation.  L  Table2.MailoutSampleSize,UAAs,Duplicates,andElectronicResponsesbyPanel   *!WVTdddd dd dd dd V(#(#,dd ,dd ,9dd ,dd ,dd ,edd ,dd +  /X /Panel KA2PX KInitialMailout P SampleSize*dB  ԍ(8,pages1and2,R1)dC  5X$<" 5UAAs P dB  ԍ(4,page2,R1)C  Ԛd RXA2<" 66X RCaseswith P duplicate < records RXA2( " X RElectronic P  Responses <  dB  ԍ(4,page1,R2)C  Ԛd RXA2( " X RFinal P  Mailout < SampleSize ( dB  ԍ(8,pages1and2,R1)C  d B82" ""X B (X X (HCA =X,!l"X =LCA AX0!l"X A 4*XX 4  X   X  ,s X ,ShortForm: dsO! s @18372@d18372 sqC " @18372 @s @@18369@@18369 spC " @@18369 @@s *@3477*@3477 vsaB " *@3477 *@s # v 0 ss^4 " # s  @2@s2 sm? "  @2 @s @36741@36741 c`NC " @36741 @s cSPANShortForm  ^ Control(NOTUSED) c`N! J` @2626@c2626 `oB ^" @2626 @` |@2622|@2622 `nB ^" |@2622 |@` P@970P@970 `lA ^" P@970 P@`  &@11&@11 `j@ ^"  &@11 &@`  @2@ 2 `l? ^ "  @2 @` @5248@ 5248 WMB ^!" @5248 @` W AQEShortForm z " Control f # (SFC) I?R $ @2624@I 2624 j`3z %" @2624 @ @2625@j 2625 i_3z &" @2625 @  @498 @i 498 [Q2z '"  @498  @ # [ 0 ND%z (" #  # N 0 \R%z )" #  @5249@\ 5249 ?53z *" @5249 @ ? AllSSNRequest " + (Panel1) @6 , @2624@@ 2624 j`3" -" @2624 @ @2624@j 2624 i_3" ." @2624 @ ~@492~@i 492 [Q2" /" ~@492 ~@ # [ 0 ND%" 0" #  # N 0 \R%" 1" #  @5248@\ 5248 ?53" 2" @5248 @ ? OneSSNRequest >3 (Panel2) @6 *z4 @2627@@ 2627 j`3>5" @2627 @ x@2620x@j 2620 i_3>6" x@2620 x@ |@458|@i 458 [Q2>7" |@458 |@ # [ 0 ND%>8" #  # N 0 \R%>9" #  @5247@\ 5247 ?53>:" @5247 @ ? AllSSNRequest, ; GeneralNotification < (Panel3) @6 = @2626@@ 2626 j`3>" @2626 @ @2629@j 2629 i_3?" @2629 @ 0@5180@i 518 [Q2@" 0@518 0@ # [ 0 ND%A" #  # N 0 \R%B" #  @5255@\ 5255 ?53C" @5255 @ ? AllSSNRequest, RD SpecificNotification  >E  (Panel4) @6 *F @2624@@ 2624 j`3RG" @2624 @ @2624@j 2624 i_3RH" @2624 @ @497@i 497 [Q2RI" @497 @ # [ 0 YO%RJ" #   ?1?Y 1 g]0RK"  ?1 ? @5248@g 5248 ?53RL" @5248 @ ? GeneralNotification ZM (Panel5) @6 FN @2624@@ 2624 j`3ZO" @2624 @ @2626@j 2626 i_3ZP" @2626 @ X@523X@i 523 [Q2ZQ" X@523 X@ # [ 0 ND%ZR" #  # N 0 \R%ZS" #  @5250@\ 5250 ?53ZT" @5250 @ ? SpecificNotification U (Panel6) @6  V ~@2623~@@ 2623 j`3W" ~@2623 ~@ z@2621z@j 2621 i_3X" z@2621 z@ ~@491~@i 491 [Q2Y" ~@491 ~@ # [ 0 YO%Z" #   ?1?Y 1 g]0["  ?1 ? |@5244|@g 5244 Js53\" |@5244 |@ JLongForm: ZsE2"]s @7867@Z 7867 soB2"^" @7867 @s @7830@ 7830 soB2"_" @7830 @s ,@1675,@ 1675 slB2"`" ,@1675 ,@s  @2@ 2 h^?2"a"  @2 @s # h 0 ]S%2"b" #  @15697@] 15697 @642"c" @15697 @ @ LongFormControl #d (LFC) @6 $e @2624@@ 2624 j`3#f" @2624 @ h@2612h@j 2612 i_3#g" h@2612 h@ H@585H@i 585 [Q2#h" H@585 H@ # [ 0 ND%#i" #  # N 0 \R%#j" #  t@5236t@\ 5236 ?53#k" t@5236 t@ ? SpecificNotification &n!l (Panel7) @6  'Z"m z@2621z@@ 2621 j`3&n!n" z@2621 z@ `@2608`@j 2608 i_3&n!o" `@2608 `@ P@522P@i 522 f\2&n!p" P@522 P@  ?1?f 1 YO0&n!q"  ?1 ? # Y 0 \R%&n!r" #  m@5229m@\ 5229 ?53&n!s" m@5229 m@ ? GeneralNotification (#t (Panel8) @6 v)$u x@2620x@@ 2620 j`3(#v" x@2620 x@ d@2610d@j 2610 i_3(#w" d@2610 d@ @568@i 568 f\2(#x" @568 @  ?1?f 1 YO0(#y"  ?1 ? # Y 0 \R%(#z" #  n@5230n@\ 5230 C97(#{" n@5230  n@ C ɀXX#p*ThesamplesizesvaryslightlyfromtheoriginalsamplesizelistedintheProgramMasterPlan.Aftertheplanwas *%{ finalized,thereweresomeGQunitsdiscoveredintheoriginalsamplethatwereremovedpriortomailout.#X#pXɀ#  V+&|  ,j'}  2.5Measurements  @*ThemailresponseratereportedinthispaperissimilartotheofficialCensus2000mailresponse  rateinthatitisameasureofrespondentbehaviorwithregardtothereturnoftheexperimental t questionnaire.dB  ԍ(9,allpages,R1)C  dThemailresponserateisdefinedasthenumberofnonblankquestionnaires ` returnedbymailforthetreatmentgroup(orpanel)dividedbythenumberofquestionnaires L  mailedoutlessthosereturnedbytheUSPSasUAA.QuestionnairesreturnedasUAAare 8  flaggedbyIDontheDMAF. $ t Ѐ        "   ` ForCensus2000,blankformsareidentifiedusingastandardcensusalgorithm(Memofrom  L  Hoganto_ Miskura_ Ԁand_ Longini_ ).dB  ԍ(10,allpages,R1)C  dThisalgorithmtreatsquestionnaireswithlessthan2completed  8  itemsamongrespondentreportedhouseholdcount,tenure,andall100%personitems(race, $  Hispanicorigin,gender,age,dateofbirth)asblanks.Duetointenselaborinvolvedin   replicatingthecompleteblankformeditproceduresintheaforementionedmemoaswellasthe   factthattheprocedurecouldnotbecompletelyreplicatedduetoinherentlycomplicated   programmingintheOpticalCharacterRecognition(OCR)andKeyFromImage(_ KFI_ ),amore   simpletechniqueisusedwhichefficientlydetectsthemajority(ifnotall)oftheblankforms.In p thismodifiedtechnique,blankformsaredefinedasreturnedquestionnaireswherethenumberof \ completeditemsforthehousehold,person1andperson2islessthantwo.Thefirststepinthis H algorithminvolvesidentifyingvariableswithmissingdataforallofthe100%itemsforpersons1 4 and2aswellasrespondentreportedpopulationcountandtenure(Notethatrespondentreported  p householdcountisdenotedasmissinginthisalgorithmifitiszero).Ifthenumberoffilled  \ itemsfromthislistislessthan2,thenthesecasesaremarkedasnonrespondentssincetheyare H technicallyblankaccordingtotheblankformsalgorithm.ManualcheckingoftheSFCandpanel 4 8(SPANLFwithgeneralnotification)casesthataremarkedasblanksbythisalgorithmrevealed   thateachmarkedcaseistrulyblankaccordingtothespecificationsinthe#K3memo.Notethat   thisalgorithmvariesslightlyfromthestandardcensusproceduresgiveninthereferenced  memorandum.Thealgorithmusedheretreatsquestionnaireswithatleast2itemscompleted  acrosspersons3through6asblankifthereisnotonecompleteditemforthehousehold,person  1orperson2;however,thelikelihoodoffindingaquestionnairewiththiscompletionpatternis l thoughttobelow.SeeAppendixBforalistingofthenumberofblankformsbypanel. X  Formcompletenessismeasuredinafewways.Itemnonresponserates(SeeAppendixD) 0"  indicatethepercentofdatamissingforaparticularquestionnaireitemoverallformsreturnedby #l! respondinghouseholds.Forpersonlevelcharacteristics,itemnonresponseratesarecalculated $X" forpersonnumberslessthanorequaltothenumberofpersonsinthehouseholdasreportedby $D # therespondent. %0!$ Inasecondapproachtomeasuringformcompleteness,theofficialCensus2000DataDefined '#& PersonAlgorithmcreatesastandardcensusmeasureindicatingwhetherahouseholdrequires (#' countdiscrepancyfollowupand/orwholepersonimputationtobeconsideredcompletedueto )$( missingitemsonatleastonehouseholdmember.ApersonrecordisdatadefinedinCensus |*%) 2000ifatleasttwoofthe100%dataitemsarecompleted(SeeMemofromHoganto_ Longini_  h+&* andMarx$XXXX#p)#X#pXX$X#.dB  ԍ(11,page2,R1)C  d Forthisanalysis,thenumberofpersonsrecordedbytherespondentaslivinginthe T,'+ _ householdiscomparedtothenumberofdatadefinedindividualsinthathouseholdtodetermine  ifthenumbersmatch.Ifthenumberofdatadefinedpeopleislessthantherespondentreported  householdcount,theformfailsthecountdiscrepancycoverageeditandrequires  followup/imputationinordertoobtaincompletedata(SeeMemofromHogantoLongini t attentionStoudt).dB  ԍ(12,allpages,R1)C  dInadditiontobeingameasureofdataquality,thismeasureisdirectlyrelated ` toincreasedcostsassociatedwithfollowinguphouseholdsthatfailthisspecificcoverageedit. L  Lastly,aformcompletenessindicatoriscreatedtodeterminewhichhouseholdshaveatleast $ t somemissingdataontheircensusforms.Caseswithatleastsomemissingdataareidentifiedas  ` thosehouseholdswithcompleterespondentreportedhouseholdcountsthataremissinganyofthe  L  followingforthehouseholdorhouseholdmembers:householdcount,tenure,sex,dateof  8  birth/age,race,andethnicity. $  2.6AnalyticProcedures   Theanalysisoftheexperimentalquestionnairedataisconductedbymeasuringthepairwise   differencesinSSNitemnonresponseratesamongthepanelsandbymodelingtheresponserates, p formcompletenessrates,andcountdiscrepancyfollowupratesusinglogisticregressionmodels. \ Theanalysisisdesignedsothatstatementsaboutthesignificanceoftreatmenteffects(i.e. H differencesinresponserates)canbemadeaboutalltestssimultaneouslywhilemaintaininga90 4 percentconfidencelevel(theCensusBureauStandard).  p Thepairwisecomparisonscomponentofthisstudyprovidesestimatesoftheactualdifferences H amongpanelswithregardtothevariablesofinterest.Usingamultiplecomparisonprocedure 4 (MCP),thecomparisonswerecarriedoutsothatstatementsabouttheentirefamilyofpairwise   comparisonsaremadewhilemaintainingthe90percentsimultaneousconfidencelevel,wherea   familyisdefinedasacollectionofinferencesforwhichitismeaningfultotakeintoaccount  somecombinedmeasureoferrors.TheuseofaMCPisappropriateif,inorderforafinal  decisiontobecorrect,itisnecessarythatallinferencesbesimultaneouslycorrect.Specifically,  MCPsrequirethatlargerdifferencesexistbetweenindividualtreatmentstobeconsidered l significant,whenmanytreatmentsarebeingcompared.Thesignificanceoralphalevelforeach X  responseratecomparisonwithinafamilyinthisanalysisisadjustedusingtheBonferroniMCP D! wheretheformulafortheconfidenceintervalaroundthedifferenceis 0"  kY7'#`^  `.E#  # k߀ #l! whereiandjarethetreatmentmeans,t(/2m)istheupperpointofthetdistributionform $I " pairwisecomparisons,andVar(i󀄀j)isthevarianceofthedifference. %5!# Thelogisticregressionapproachprovidesaquickandeffectivemeansforevaluatingwhether ' #% differencesinresponse,formcompletenessandcountdiscrepancyfollowupratesforeachofthe (#& mainexperimentaltreatmentsareinfluencedbythepresenceofothertreatments.Interpretation )$' ofthelogisticregressionresultsisbasedonparameterestimatesandoddsratiosofsignificant *%( experimentaltreatmenteffects.Theseestimatesassessthemagnitudeoftheimpactofthe m+&) treatmentonahouseholdsoddsofmailingbackthecensusquestionnaireandhaving_ missing Y,'* dataonthereturnedform.Inaddition,theyassesstheCensusoddsofhavingto  followup/imputeduetocountdiscrepancies.  Parameterestimates,whichareestimatedcoefficientsforthepredictorvariables,aredirectly t relatedtooddsratios.Theparameterestimateforanyvariableisthenaturallogarithmofthe ` oddsratio.Hence,_ exponentiating_ Ԁtheparameterestimateyieldstheoddsratioforthatvariable. L  Theoddsratioofagivenexperimentaltreatmentdescribestheoddsofhavingthedependent 8  variableequaltooneascomparedtothebaselinetreatment.Thefollowingexample $ t demonstratestheutilityofthisrelationship.  ` Consideramodelwithoneindependentvariable,sayanindicatorofwhetherahousehold  8  receivedanexperimentalformwiththeSSNrequestforallpersons.Thedependentvariablein $  themodelisresponsetothecensus.Ifthehouseholdmailedbacktheirquestionnaire,the   responsevalueis1.Ifahouseholddidnotreturntheirquestionnaire,theresponsevalueis0.   Suppose100personsarecrossclassifiedbytheSSNrequestindicatorvariableandresponseto   thecensusasfollows(NOTE:dataarefictional):   *WVdddd dd 9dd 9dd dd edd edd !WVT(#(#,dd , dd , dd ,dd +  \\    DidnotreceiveSSNrequest   ReceivedSSNrequest   Total   Respondedtothe @ Census >4 ,|  5@215@>21 f\1@"  5@21 5@  6@226@f22 f\1@"  6@22 6@  E@43E@f43 =31@"  E@43 E@ =DidnotRespondto  theCensus =3   @6@=6 e[0"  @6 @  I@51I@e51 f\1"  I@51 I@  L@57L@f57 =31"  L@57 L@ =Total >4 h   ;@27;@>27 f\1h!"  ;@27 ;@  @R@73@R@f73 g]1h""  @R@73 @R@ Y@100Y@g100B86h#" Y@100  Y@ BAlogisticregressionmodelisfittothesedatatopredicttheprobabilityofrespondingtothe  p$ census.TheoddsofrespondingtothecensusforahouseholdthatdidnotreceivetheSSN  \% requestistheratiooftheprobabilityofthenonSSNrequesthouseholdresponding(21/27)tothe H& probabilityofthishouseholdnotresponding(6/27).Theresultingoddsforhouseholdsnot 4' receivingtheSSNrequestforallpersonsis21/6.Theoddsofrespondingtothecensusfor  ( householdsreceivingtheSSNrequestforallpersonsaresimilarlycalculatedas(22/73)/(51/73)= ! ) 22/51. "* TheoddsratiofortheSSNrequestindicatorvariableisdefinedastheratiooftheoddsfor $, householdsnotreceivingthisexperimentaltreatmenttothatforhouseholdsreceivingthis l% - treatment.Inthisexample,theoddsratiois(21/6)*(22/51)=8.11.Thismeansthattheoddsof X&!. respondingtothecensusare8.11timeshigherforhouseholdsnotreceivingtheSSNrequestthan D'"/ forhouseholdsreceivingthistreatment.Notethattheclosertheoddsareto1,thelessimportant 0(#0 theindependentvariableisinpredictingresponsetothecensus.Iftheoddsoftheexperimental )l$1 indicatorabovewere1,thenhouseholdsnotreceivingtheSSNrequestforallpersonsand *X%2 householdsreceivingthistreatmentwouldbeequallylikelytorespondtothecensus. *D&3  +0'4 2.7VarianceEstimation  Inordertotakeintoaccountthestratifiedsampledesigninthedataanalysis,_ WesVarPC_ ԀVersion  2.12isusedtocomputestandarderrorsforallestimatesandmodelsusingareplication t methodology._ WesVarPC_ ԀVersion2.12requiresatwoPrimarySamplingUnit(_ PSU_ )perstratum ` designinordertouseastratifiedjackknifevarianceestimationmethodology.Sincethereisno L  clusteringinoursampleandonlytwostratatowhichthesampleisallocated,asimplejackknife 8  approachassuggestedbyBobFay,seniormathematicalstatisticianattheU.S.CensusBureau,is $ t substitutedforthestratifiedjackknifewhencomputingtheresponseratesanditemnonresponse  ` rates.Thismethodologyrequiresforming256replicatesbynumberingobservations  L  consecutivelywithinstrata.Notethatreplicatesamplesarecombinedacrossthetwostrata,  8  whichisgenerallyavoidedinanalysesthatinvolvestratifiedjackknifereplication.Theproposed $  designbalancesthereplicatesbyselectingthemfrombothstrata.     Twentyreplicatesareusedincomputingthestandarderrorsforanalysesinvolvingitem   nonresponsesincethesamplesizeismuchsmallerwhenonlyrespondentsareconsidered.   X9 XXX#pQualityassuranceprocedureswereappliedtothedesign,implementation,analysis,and \ preparationofthisreport.Theproceduresencompassedmethodology,specificationofproject H proceduresandsoftware,computersystemdesignandreview,developmentofclericaland 4 computerprocedures,anddataanalysisandreportwriting.Adescriptionoftheproceduresused  p isprovidedinthe#X#pX XX9U#BXXXX#p>#X#pXXBXh#X9 XXX#pCensus2000EvaluationProgramQualityAssuranceProcess.   ! #X#pX XX9#  \ T14 <DL!T$X,XT u XX#pX#pX u3.0  LIMITATIONS#u XX#p#4(#(# X#pX u 3.1PopulationCoverage #u XX#pI#  4 X#pX uTherearecertainlimitationsofthisexperimentwhichareinherentinthedesign.First,the 2 samplingframedoesnotentirelyrepresenttheCensus2000universe.Addressesselectedforthe  AccuracyandCoverageEvaluation(A.C.E.)initiallistingsamplesareexcludedfromthe   universe,asareaddressesselectedforA.C.E.contaminationevaluation.Alsoexcludedarethose   householdsnotinthe_ mailout_ /mailbackuniverse,namelythoseinthelist/enumerateand ! update/leaveareaswhichtendtobemoreruralthanaddressesinthe_ mailout_ /mailbackuniverse. ~"  Sincethegoalofthisexperimentistoevaluateresponsebehaviorwithregardtomail j#! alternatives,ourpopulationofinferenceisbydefinitionthe_ mailout_ /mailbackuniverse. V$" However,addressesaddedthroughcoverageimprovementprogramsbetweentheprintingof B% # addresslabelsinSeptember1999andthequestionnairemailoutinearlyMarch2000arenot .&~!$ includedinthesamplingframewhichmayresultinaslightundercoverageofthetarget 'j"% population.AlsonotethatCoverageEdit_ FollowUp_ Ԁ(_ CEFU_ )andlargehouseholdfollowupwere (V#& notperformedforexperimentalcases. (B$' Furthermore,nonEnglishspeakinghouseholdsareexcludedfromthisexperimentsincethe *&) SPANquestionnairesandformshaveonlybeenprintedinEnglish.Thislanguagerestrictionis +'* incontrasttotherestofCensus2000,whererespondentscanrequestquestionnairesinavariety  oflanguages.Thecostassociatedwithtranslatingexperimentalformsisprohibitive,considering  thesmallnumberofexperimentalhouseholdsexpectedtobenefitfromthetranslation.Although  thislanguagerestrictionlimits_ generalizability_ ,nodifferentialeffectsareexpectedacrossanyof t thepanelsduetotherandomassignmentofhouseholdstothetreatmentgroupsinthis ` experiment. L    3.2UndeliverableForms $ t  r_#Someexperimentalquestionnairepackagessenttothoseselectedinthesamplemaybereturned  L  duetothefactthattheyareUAAs,whichisconsideredbytheUSPStobeundeliverableas  8  addressed(seesection2.3formoredetail).Thesampleselectedinthisexperimentwilllikely $  haveahigherUAAratethanCensus2000sincethesampleisselectedpriortofinalverification   ofaddressesbytheUSPS.However,duetotherandomassignmentofhouseholdstotreatment   groups,theassumptionismadethattherearenodifferentialUAAratesacrosspanelsforthis   experiment.&  CensustestswithinthelastdecadewhichreportedUAAstatisticsrepeatedlyfound   noevidenceofdifferentialpanelUAArates.PairwisecomparisonsofweightedUAArates p amongtheshortformandlongformpanelsinthisexperimentindicatenosignificantdifferences, \ replicatingresultsexhibitedinpreCensus2000responseratetests.SeeAppendixCfor H statisticalcomparisonsofUAArates. 4 'i&  3.3CausalAssumptionabouttheEffectoftheTreatment  \  r(Incomparingpanelresponseacrossexperimentaltreatments,acriticalcausalassumptionismade 4 regardingrespondentbehavior.Theassumptionisthateachrespondentwhoprovidesdatafora   particularexperimentaltreatmenthasbeenbothexposedtothewrittennotificationand/orSSN   request,andknowinglydecidestoengageinthebehaviorsuchasreturningtheformwithfull  awarenessoftheexperimentaltreatmentsembeddedinthemailoutpackage.Somerespondents  maynothaver_ ead_ Ԁthequestionnairecoverletterandthereforemaynothavenoticedthe  notificationofadministrativerecorduse.Thereisnowaytoascertainadirectcausalrelationship l betweentreatmentandrespondentbehavior.However,duetotherandomassignmentofsample X  casestotheexperimentaltreatments,weassumethattherearenodifferentialconfounding D! motivationaleffectsacrosstreatments. 0"  #u XX#p# X#pX u4.0  RESULTS#u XX#pz-# X#pX u$X"(#(# 4.1MailResponseRates? #  1        %0!$ AsofApril18,2000,theCensusBureaureportsthat66.6%and54.1%ofshortandlongCensus '#& 2000formsrespectivelywerereturned(SeeGAOreport).dB  ԍ(13,page2,R1)C  dNotethatthisfigureisnotdirectly (#' comparabletotheresponseratesinthisexperimentsinceUAAsareincludedinthedenominator  ofthisfigurewhereastheyareexcludedfromthedenominatorsofthefiguresbelow.  Additionally,thiscensusfigureincludesthereturnofformsfromcasesoutsideofthe  _ mailout_ /mailbackuniverse.Likewise,responsefilesfortheSPANexperimentwerecreatedon t June14,2000andthusincludemailreturnssentinbeyondtheApril18thcutoff. ` Historically,experimentalanalysesconductedbytheCensusBureauandtheircontractorshave 8  involvedaseriesofpairwisecomparisonstomeasurestatisticaldifferencesinresponserates $ t acrosspanelswithvariousexperimentaltreatments(SeeTreat(1996);Dillman,WestandClark  ` (1994);Dillman,ClarkandTreat(1994)).TheexperimentaldesignusedinSPANcalledfora  L  differentandmoresensitivemethod.Throughoutthisreport,logisticregressionanalysisisused  8  todeterminetheeffectofeachtreatment(SSNrequest,notificationofadministrativerecorduse) $  onresponse.Theadvantagesofusingthistechniqueintheassessmentofsignificantfactorsare   many.First,theeffectofthetreatmentonresponsecanbedeterminedinthepresenceofother   treatmentsorcontrolvariables,addingcredibilitytothefindings.Controlvariablesusedinunit   oritemnonresponsemodelsincludestrata,lengthofthecensusform,anddiscreteorcontinuous   measuresofthedemographicsofPerson1onthecensusform. p Secondly,logisticregressionmaximizespowerinstatisticaltestssincepanelssharingacommon H treatmentareallincludedintheassessmentofthattreatmentseffectonresponse,holdingany 4 otherexperimentaltreatmentsinthosepanelsconstant.Forinstance,thetreatmentinvolvingthe  p SSNrequestforallhouseholdmembersisinstitutedinthreeseparatepanels(panels1,3and4)  \ asshowninTable2.Notethatpanels3and4containnotificationtreatmentsinadditiontothe H SSNrequest.Whereasapairwisecomparisonapproachwouldassesstheeff_ ect_ ԀoftheSSN 4 requestusingonlythecomparisonofpanel1tothecontrolpanel,thelogisticapproach   simultaneouslycontrolsthenotificationtreatmentinpanels3and4.Inthisway,panels3and4   canbeincludedintheassessmentofthefulltreatmenteffectoftheSSNrequestforallhousehold  members.Specifically,testsofpanel3(allSSNrequest,generalnotification)versuspanel5  (generalnotification)willcontributetothestudyoftheeffectoftheSSNrequestonresponseina  logisticregressionmodelwhereasitwillnothaveanyinfluencewhenusingasimplepairwise l comparisonapproach.dB  ԍ(14,page1,R1)C  d X  Finally,thelogisticregressiontechniqueprovidesresultsthataremoreinsightfulwithrespectto 0"  designingexperimentsfor2010planning.Logisticregressionresultsprovideoveralltreatment #l! effects,givingexperimentdesignersmoreinformationandflexibilitywhendesigningafollowup $X" experimentwiththespecifictreatments.Pairwisecomparisonresultsalonearemoredifficultto $D # usefordesigningsubsequentexperimentswhenexperimentalpanelsinvolvemorethanone %0!$ treatmentsandlessthanthefullsetoftreatmentcombinationsaretested. &"% MailresponseratesforthevariouspanelsinthisexperimentareincludedinAppendixE.Note (#' thattheresultsofsimplepairwisecomparisonsoftheresponseratesacrosspanelsmaynotfully )$( agreewiththeresultsillustratedbythelogisticmodelbelowduetothedifferencesinthe |*%) approachesasmentionedabove.Withregardtotheresponserates,itisinterestingtonotethat h+&* responseratesareconsistentlylowerintheLCAthantheHCAacrossallexperimentalpanelsin T,'+ accordancewithstratificationexpectationsasshowninTablesE.1andE.2intheAppendix.  Additionally,longformmailresponseratesarelowerthanshortformratesthroughoutthis  report,consistentwithCensus2000.    4.1.1WhatistheEffectoftheSSNRequest? ` Basedonpastresearchfindings,ithasbeensuggestedthattherequestfortheSSNofPerson1or 8  allhouseholdmemberswilldecreaseresponse,withamorepronouncedeffectinlowcensus $ t coverageareascomparedtohighcoverageareas.   `  InordertoassesstheeffectoftheSSNrequestinthepresenceofothertreatments,logistic  8  regressionanalysisisusedtomodelahouseholdsoddsofrespondingtothecensus.TheSimple $  ModelshowninTable3investigatestheeffectoftheSSNrequestforoneorallpersonsand   notificationonresponse,whilecontrollingforstrataandformlength.Thismodelassumesthat   theeffectofatreatmentonresponseisconstantwithintheotherexperimentaltreatment.The   purposeofthemodelswiththeinteractiontermsistorelaxthisassumptionassociatedwiththe   SimpleModel.Thatis,theinteractiontermshelpdetermineifatreatmenteffectdiffersamong p _ subpopulations_ Ԁandothertreatments.TheSSNStrataInteractionModeldetermineswhetherthe \ effectoftheSSNrequestonresponsediffersbasedonthe_ subpopulation_ Ԁfromwhichitis H requested(i.e.HCAversusLCA).Lastly,theTreatmentInteractionModel,whichisdiscussedin 4 thenextsection,testswhethernotificationofadministrativerecordusedecreasesresponseinthe  p presenceofaSSNrequestcomparedtonotificationalone.  \ *'V/dddd  dd  dd  dd WV(#(#,dd ,Fdd ,dd ,dd +  $   $Table3.LogisticRegressionCoefficientsPredictingtheLogOddsof 4 RespondingtotheCensus C9)L"  i CVariable ;V,!H"i ;Simple H _ Modela_  4 dB  ԍ(17,page5,R1)C  d ;V,! "Vi ;SSNStrata H Interaction 4  Model  ! dB  ԍ(17,page3,R2)C  Ԛd ;V,! ""Vi ;Treatment H# Interaction 4$ Model  % dB  ԍ(17,page4,R3)C  Ԛd ;1& &"  Vi ;SSNRequest: ,V<"' , ;V,!<"("V ; ;V,!<")"V ; -#!<"*"V -ЀForPerson1=1 ,V# + , .099* ;V,!# ,"V ; ^VO!# -"V V--.071V-^Є.071 OEC# ." V--.071 V-V OЀForHousehold=1 ,V<% / , .113* ;V,!<% 0"V ; ^VO!<% 1"V A`"-.053A`"^Є.053 OEC<% 2" A`"-.053 A`"V OЀForPerson1orHousehold=1 ,V& "3 , ;V,!& "4"V ; .105* ;V,!& "5"V ; 6,!& "6"V 6Notification: ,V<(#7 , ;V,!<(#8"V ; ;V,!<(#9"V ; -#!<(#:"V -ЀGeneral=1 ,V) %; , .090* ;V,!) %<"V ; .094* ^VO!) %="V T㥛 -.063T㥛 ^Є.063 OEC) %>" T㥛 -.063 T㥛 V OЀSpecific=1 OV@<+&? l-.037lOЄ.037 VqC<+&@" l-.037 lV ˡE-.041ˡEЄ.041 VpC<+&A" ˡE-.041 ˡEV ~jt?.019~jt?.019 WMB<+&B" ~jt?.019 ~jt?V WFormType: ,V d , ;V,! d"V ; ;V,! d"V ; -#! d"V -ЀLongForm=1 ,V  , .454* ;V,! "V ; .454* ;V,! "V ; .454* 6,! "V 6Strata: ,Vd  , ;V,!d "V ; ;V,!d "V ; -#!d "V -ЀHighCoverageAreas=1 ,V  , .757* ;V,!"V ; .761* ;V,!"V ; .757* -#!"V - Interactions: ,Vd , ;V,!d"V ; ;V,!d"V ; -#!d"V -ЀSSNRequestforeither*Strata ,V , ^VO!"V ~jtx-.006~jtx^Є.006 ]VNC" ~jtx-.006 ~jtxV ] -#!"V -ЀGeneralNotification*SSNforHousehold ,Vd  , ;V,!d "V ; ^VO!d "V Q-.060Q^Є.060 OECd " Q-.060 QV OЀSpecificNotification*SSNforHousehold ,V  , ;V,! "V ; ;V,! "V ; .120* -#! "V -Intercept ]VN!d ! Vi ~jt?.429~jt?].429 ~VoBd ""  ~jt?.429 ~jt?VVi Q?.430Q?~.430 ~VoBd #"  Q?.430 Q?VVi |?5^?.402|?5^?~.402RHFd $"  |?5^?.402  |?5^?VVi R*ɀXX#pIndicatesstatisticalsignificanceat=.1.#X#pXɀ`# 8 $ aɀXX#p_ A_ ԀtestofthecombinedeffectoftheSSNrequestforallhouseholdmembersandPerson1revealsthatanyrequest $l% forSSNdecreasesresponse.Therefore,theSSNstratamodelcombinesthesetreatments.#X#pXɀ7a#Ԁ X& TheSimpleModellogisticresultsinTable3suggestthattherequestfortheSSNofPerson1as 0( wellastheSSNrequestforallhouseholdmemberssignificantlydecreasesresponsetothe ) census,whilecontrollingfornotificationandtypeofarea(oddsofrespondingtothecensus * decreaseby9.5%forthePerson1SSNrequest,11%fortheSSNrequestforallhouseholds + members< #  2      ).Whilethisdecreaseinresponseissignificant,itisalsofairlysmall.Thedropin  , oddssuggestsabouta2.1%B  d(47,page1,R1)C  ddecreaseresponseinhighcoverageareasanda2.7%dB  ԍ(47,page1,R2)C  ddropinlow !- coverageareaswhentheSSNrequestforallhouseholdmembersispresent.Thisdropin p". responsesupportsourhypothesesformedinlightofpastprivacyresearch.However,incontrast \#/ toourinitialhypotheses,the_ SSN_ ԄStrataInteractionmodelrevealsnodifferentialeffectsofthe H$0 SSNrequestonresponse(_ SSN_ *Strata=.006)betweenareaswhichdifferwithregardtotheir 4%|1 demographicsandpropensitytorespondtothecensus,whilecontrollingfornotification.Since  &h2 SSNisauniqueidentifierinadditiontoanexcellenttoolforlinkingtoadministrativedata,one  'T3 wouldconjecturethattherequestforSSNwouldfurtherdissuadethosewhotendtoavoidthe '@4 censusfromreplying.However,thedatadonotsupportthishypothesis,suggestingthatthe (,5 slightdropinresponseduetotheSSNrequestisthesameinlowandhighcoverageareas. )6 )O )e )  4.1.2Whatistheeffectofnotificationofadministrativerecorduseonresponse? Q Recallthathypothesesregardingtheeffectofthenotificationofadministrativerecorduseon ) responsearebasedonintuition,ratherthanonempiricaldatagiventhelackofpastresearchin  thisarea.Intheabsenceofpastdata,itissuspectedthatnotificationofadministrativerecorduse   willcausesignificantdropsinmailresponse,withspecificnotification(includingagencynames)   havingastrongereffectthangeneralnotification.Moreover,basedonpastqualitativestudy,itis  t expectedthataddingeithertypeofnotificationtotheSSNrequestwilldecreaseresponse  ` comparedtonotificationalone.3 X,[XX3  L   LogisticregressionresultsgiveninTable3allowdeterminationofthenotificationeffectson $  response.   Testingofthesimultaneoussignificanceofgeneralandspecificnotificationinthe u  SimpleModelrevealsthat,takentogether,notificationofadministrativerecordusedecreases a  mailresponse_ U #  3      _ .Lookingattheeffectofeachnotificationtypeseparately,thelogisticresults M  showthatgeneralnotificationcausesasmall,yetsignificant,decreaseinresponse,whilespecific 9  notificationdoesnot.Sincethisfindingdisagreeswithourinitialhypotheses,wefurther % comparedtheparametersassociatedwithgeneralandspecificnotification.Themagnitudeofthe  parametersandthereforetheeffectonresponsebetweengeneralandspecificnotificationisnot  statisticallydifferent(H0:_ specific_ 󀄀_ general_ =0,p=.12).dB  ԍ(17,pages2and5,R4)C  d   F WithrespecttotheeffectofnotificationonresponseinthepresenceoftheSSNrequest,the \ TreatmentInteractionModelinTable3suggeststhatspecificnotificationhasamoreharmful H effectonresponsewhenSSNisrequestedthanthecasewherespecificnotificationistheonly 4 experimentaltreatment.ThecombinationofgeneralnotificationandtheSSNrequestdoesnot   causefurtherdropsinresponsecomparedtogeneralnotificationalone. q  SincethehypothesisregardingtheeffectofnotificationinthepresenceofaSSNrequestdidnot I distinguishbetweengeneralandspecificnotification,anotherlogisticmodelwasruntoassess 5 theeffectsofeithertypeofadministrativerecordnotificationwiththeSSNrequestcomparedto ! notificationalone(notshown).dB  ԍ(17,page4,R5)C  dTheSSNnotificationinteractiontermisinsignificant(p=.1056),   indicatingthatingeneral,thereisnotsufficientevidencetoconcludethatnotificationof   administrativerecordusefurtherdiscouragesresponseinthepresenceofaSSNrequest !  comparedtonotificationalone.Thisfindingcontraststheinitialhypothesisofamore "l! pronounceddropinresponseduetothepairingofnotificationwiththeSSNrequest. #X" 4.2ItemNonresponseRates %0!$ Totheextentthatdataforparticularitemsaremissingfromthereturnedquestionnaires,the m'#& amountoffollowupbycensusinordertoresolvecountdiscrepanciesormeasurementerrorcould Y(#'  F  e    increase.Inthissection,westudytheeffectsoftheSPANexperimentaltreatmentsondata  qualityasmeasuredbymissingdata.  4.2.1WhatistheEffectofSSNrequestonItemNonresponse?  GiventhelevelofresistanceshowninthepastforSSNrequest,wehypothesizethattherequest  forSSNwillincreaseitemmissingdata.Logisticregressionmodelswillbeusedtoassessthe   validityofthishypothesis.ItemnonresponseratesinAppendixDindicatethepercentofdata  t missingforaparticularquestionnaireitemoverallformsreturnedbyrespondinghouseholdsby z ` experimentalpanel.Forpersonlevelcharacteristics,itemnonresponseratesarecalculatedfor f L  personnumberslessthanorequaltothenumberofpersonsinthehouseholdasreportedbythe R 8  respondent.Onestrikingfeatureofthesedataistherelativelylowratesofmissingdataacrossall >$  treatments. *     TheofficialcensusDataDefinedAlgorithm(discussedaboveinsection2.5)isusedtocreatea   dichotomousvariableindicatingwhetherahouseholdrequiresfollowupbasedonacount   discrepancy.Usingthismeasureasthedependentvariable,theoddsofhavingtofollowupwitha  householdduetoacoveragefailurearemodeledusingtheexperimentaltreatmentsandtheir  interactionsaspredictorvariables.Casesforwhichtherespondentreportedhousehold  populationcountismissingorzeroareexcludedfromthesemodelssinceitisnotpossibleto  determinehowmanyindividualsareexpectedtohavecompletedata. Thedependentvariablein p themodelishighlyskewedinthesensethatthereareveryfewcasesthatrequirefollowupor v\ imputationasdeterminedbytheDataDefinedPersonalgorithm.Morespecifically,outof bH 28,525householdswithcompletedrespondentreportedhouseholdcountsacrossallpanels,only N4 216requirecountdiscrepancyfollowup.dB  ԍ(19,page1,R1)C  dThisapproachwasreplacedbyanalternativemeasure :  ofdataquality.@* &  Asaproxyfordataquality,theeffectofthetreatmentsonitemnonresponseisassessedby  lookingattheeffectofeachtreatmentonthelikelihoodofahouseholdhavinganymissingdata  among100%personitemsinadditiontohouseholdtenure.Outofthe28,525householdswith  completehouseholdcounts,3,385outof21,068householdsrespondingtotheshortformhadat  leastsomemissingdatawhile1,268outof7,457householdsrespondingtothelongformhadat   leastsomemissing100%items.dB  ԍ(20,page1,R1)C  dThelogisticresultsinTable3belowareusedtoaddressthe !  validityofitem_ nonresponse_ Ԅrelatedhypotheses. "l! *"V=dddd Fdd Fdd dd 'V/(#(#,N dd ,rdd ,dd +  $r#X"r#X $Table4.LogisticRegressionCoefficientsPredictingtheLog ## OddsofReturninganIncompleteCensusForm C9)$ $"  i CVariable 6,!&"%"i 6SimpleModel &"& dB  ԍ(21,page2,R1)C  d 6,!'l#'"i 6NotificationForm &"( LengthInteraction 'l#) Model r(X$* dB  ԍ(21,page3,R2)C  d ;1&^)D%+"  i ;SSNRequest: '*&, ' '*&-" ' *&." ЀForPerson1=1 I? d |?5^?.103|?5^?I .103 j`3 d" |?5^?.103 |?5^? 1Zd?.1071Zd?j.107 ?53 d" 1Zd?.107 1Zd? ?ЀForHousehold=1 '  ' .201* ' " ' .201* ' " 'Notification: 'd ' 'd" ' d " ЀGeneral=1 J@  ~jt-.019~jtJЄ.019 lb4 " ~jt-.019 ~jt Q-.015QlЄ.015 @64 " Q-.015 Q @ЀSpecific=1 I?d  Mb?.008Mb?I.008 j`3d" Mb?.008 Mb? Q?.015Q?j.015 H>3d" Q?.015 Q? HFormType: ' ' '" ' " ЀLongForm=1 'd  ' .189* 'd " ' .243* 'd " 'Strata: '  ' ' " '  " ЀHighCoverageArea=1 'd  ' .820* 'd " ' .820* d "  Interactions: '  ' ' " '  " ЀGeneralNotification*LongForm 'd ' J@d " x&-.067x&JЄ.067 @64d!" x&-.067 x& @ЀSpecificNotification*LongForm '" ' J@#" EԸ-.097EԸJЄ.097 @64$" EԸ-.097 EԸ @Intercept ZP!d% Vi S-1.333SZЄ1.333 }sDd&"  S-1.333 SVi 1Zd-1.3371Zd}Є1.337TJHd'"  1Zd-1.337  1ZdVi TɀXX#p*#X#pXɀa#ԀɀXX#pIndicatesstatisticalsignificanceat=.1.#X#pXɀ# ' Inaccordancewithhypotheses,resultsfromtheSimpleModelaboverevealthattheSSNrequest z X) forallhouseholdmembersisassociatedwithhavingmissingdataonthereturnedcensusform f!D* (oddsofhavingmissingdataincreasebyafactorof1.25).Itisinterestingtonotethattherequest R"0+ fortheSSNofPerson1isindependentlynotassociatedwithhavingmissingdata.However, >#, collectively,anyrequestforSSNseemstoincreasetheoddsofhavingatleastsomemissingdata *$- ontheform(H0:_ SSN_ ԀforPerson1+_ SSN_ Ԁforhousehold=0,p=.026).dB  ԍ(21,pages1and2,R2)C  d %. Notethatweattemptedtomodeltheproportionofmissingdatagiventhesetreatmentsusinga &0 linearregressionapproach.Thedependentvariableishighlyskewedbeyondthehelpof '1 statisticaltransformationssincetherearemanycaseswithnomissingdata.Therefore,the (2 logisticresultsinTable4abovemustsufficeforthisanalysis. )3  v,T!6  4.2.2WhatistheEffectofNotificationonItemNonresponse?  Wealsosuspectthatnotificationofadministrativerecordusewillcausesignificantincreasesin  itemnonresponse,withspecificnotification(includingagencynames)havingastrongereffect  thangeneralnotification.Moreover,webelievethatnotificationofadministrativerecorduse  willincreaseitemnonresponseinamorepronouncedwaywhencoupledwiththelongform  comparedtotheshortform,duetoprivacyissuesraisedpubliclyregardinglongformquestions   duringCensus2000.dB  ԍ(1,page1,R1)C  ԚdLastly,basedonpastqualitativeresearch,weanticipatethattheSSNitem  t forPerson1willbemissingatahigherratewhenanynotificationisincludedwiththeSSN z ` request. f L  FromtheSimpleModelinTable4,itisclearthatnotificationofadministrativerecorduseasa >$  wholedoesnotappeartoadverselyaffectformcompleteness.Individually,neithertypeof *  notificationhasaneffect,nordoesonetypehaveastrongereffectthantheother.Thisfinding   contradictspriorhypothesesofasuspectedcorrelationbetweennotificationandmoreitem   missingdata.   TheNotificationFormLengthModelallowsatestofthehypothesisthatnotificationof  administrativerecordusehasamoreharmfuleffectonformcompletenessonthelongcensus  formthanthecensusshortform.Theinteractionparametersinthatmodelsuggestthatthereare  nodifferentialeffectsofnotificationonformcompletenessbetweenthelongandshortcensus p forms,regardlessofthetypeofnotification. v\ InordertoinvestigatetheeffectofnotificationonthecompletenessoftheSSNitem,item N4 nonresponseratesforSSNarecomputedforthepanelsinwhichthisinformationisrequested. :  Thesefigureswerecomputedacrossallhouseholdsforwhichtherespondentreportedhousehold &  countisgreaterthanorequaltothepersonnumber.   Table5.ItemNonresponseRates(StandardErrors)forSSN   *(V.ddN dd N rdd rdd "V=(#(#,Sdd ,dd ,dd ,dd ,dd ,dd ,dd +  .XXX .Panel @XX/!>$XXX @Person >$ 1dB  ԍ(22,page2,R1)C  d @XX/!*  "XX @Person >$! 2dB  ԍ(23,page2,R1)C  d @XX/!* ""XX @Person >$# 3dB  ԍ(24,page2,R1)C  Ԛd @XX/!* $"XX @Person >$% Ѐ4dB  ԍ(25,page2,R1)C  Ԛd @XX/!* &"XX @Person >$' Ѐ5dB  ԍ(26,page2,R1)C  Ԛd BXXX/!* ("XX BPerson >$) Ѐ6dB  ԍ(27,page2,R1)C  d >X/!* *"XXX > Panel1 (allSSNs) 9/!!+X 915.8% !, (.66) 9/!"|-" 921.6% !. (.92) 9/!"|/" 928.6% !0 (1.21) 9/!"|1" 928.1% !2 (1.63) 9/!"|3" 930.9% !4 (4.21) BX/!"|5" B29.0% !6 (5.00)1'%"|7"  X 1*HVIddSdd Sdd dd dd dd dd dd (V.(#(#,Sdd ,dd ,dd ,dd ,dd ,dd ,dd +  ,X#7# , Panel2 (OneSSN) 9/!$8X 915.5% $9 (.77) 9/!% :" 9n/a 9/!$;" 9n/a 9/!$<" 9n/a 9/!$=" 9n/a BX/!$>" Bn/a1'%$?" X 1*JVyddSdd Sdd dd dd dd dd dd HVI(#(#,Sdd ,dd ,dd ,dd ,dd ,dd ,dd +  ,X&"?&" , Panel3 (allSSNs, &h"@ generalnotification) 9/!r'X#AX 912.6% &h"B (.78) 9/!n'T#C" 917.3% &h"D (.87) 9/!n'T#E" 928.8% &h"F (1.55) 9/!n'T#G" 931.1% &h"H (1.84) 9/!n'T#I" 934.7% &h"J (4.28) BX/!n'T#K" B38.0% &h"L (8.09)1'%n'T#M" X 1*zV{ddSdd Sdd dd dd dd dd dd JVy(#(#,Sdd ,dd ,dd ,dd ,dd ,dd ,dd +  +XX(t$M(t$ + Panel4 (allSSNs, ($N specificnotification) =X,!)%OXX =11.5% ($P (.67) =X,!)%Q"X =15.8% ($R (.76) =X,!)%S"X =22.9% ($T (1.61) =X,!)%U"X =24.5% ($V (2.12) =X,!)%W"X =30.6% ($X (3.56) ?XX,!)%Y"X ?47.3% ($Z (6.19)1'%)%[" XX 1  *&[ InordertoexamineoursuccessatobtainingSSNinformation  #  4      ׀forPerson1,pairwise   comparisonsofPerson1SSNitemnonresponseratesareconsideredamongthepanelsforwhich  thisinformationisrequested(Panels1through4). Eachofthesefourpanelsreceivedsome  degreeofnotificationofthepossibilityofadministrativerecorduseduetothestatementinthe  coverletterexplainingtherequestforSSN.   Table6.MultipleComparisonsofSSNMissingRatesforPerson1byPaneldB  ԍ(28,page1,R1)C  d    **WV,ddSdd Sdd dd dd dd dd dd zV{(#(#,^dd ,edd"+  +X t t +PairwiseComparison =X,! X =Difference 0&!  "X 0Panel1(allSSNs)Panel2(oneSSN) 'r X  '.3% r X  (1.22) '^ D  'Panel1(allSSNs)Panel3(allSSNs,generalnotification) '  ' 3.2%*    (1.07) '  'Panel1(allSSNs)Panel4(allSSNs,specificnotification) 'J0  ' 4.3%*  J0  (.86) '6 'Panel2(oneSSN)Panel3(allSSNs,generalnotification) ' ' 2.9%*   (1.17) 0& 0Panel2(oneSSN)Panel4(allSSNs,specificnotification) 9/!" 9 4.0%*  " (1.02) 0&!" 0Panel3(allSSNs,generalnotification)Panel4(allSSNs,specificnotification) 't '1.1% t (.88)(z`   (ɀXX#p*Statisticallysignificantwhen_ familywise_ Ԁ=.1.#X#pXɀ#     D    | ThecomparisonsshowthatthereisnodifferenceinitemnonresponsetoPerson1SSNwhenthe F, panelrequestingonlyoneSSNiscomparedtothepanelrequestingallSSNs.Fromthe 2 perspectiveofPerson1,theseformsdonotdifferintheirrequestforSSN,andthereforeno  differenceinresponsetothisitemisexpected.Interestingly,theresultsabovesuggestthatlower    responsetothePerson1SSNitemisobtainedwhentheSSNforallhouseholdmembersor ! Person1isrequestedwithoutspecificorgeneralnotificationofadministrativerecorduse.The  " distinctionbetweengeneralandspecificnotificationhasnomeasurableinfluenceonresponseto !# theSSNitemforPerson1. "$ ToallowforcomparisonsofSSNitemcompletionwithindemographicgroups,logistic $x & regressionmodelswereformedcontrollingforthedemographicsofPerson1onthecensusform. ~%d!' Admittedly,pastresearchindicatesthatPerson1isnottherespondentroughly30%ofthetime j&P"( (_ DeMaio_ ԀandBates,1990).However,nodifferentialeffectsregardingthisassumptionare V'<#) expectedacrosstreatmentgroups.  B(($* *~LV]dd^dd ^edd"e*WV,(#(#~,dd ,dd +  $ $ Table7.LogisticRegressionCoefficientsPredictingtheLog ~d OddsofPerson1SSNMissingbyExperimentalTreatmentsdB  ԍ(29,page1,R1)C  d C9)|"  i CVariables 6,!B("i 6SimpleModelwith B( Controls ;1&."  i ;SSNRequest: '  '  " ЀForHousehold=1 I?.  ;On?.009;On?I.009 H>3.  " ;On?.009 ;On? HNotification: '   '   " ЀGeneral=1 '.  ' .275* . " ЀSpecific=1 '  ' .357* ' " 'Strata: '.  ' . " ЀHighCoverageAreas=1 I? Zd;?.061Zd;?I.061 ?53" Zd;?.061 Zd;? ? Controls: '. ' ." ЀPerson1Black=1 J@ }?5^I̿-.221}?5^I̿JЄ.221 C94" }?5^I̿-.221 }?5^I̿ CЀPerson1Hispanic=1 \R!. ;On-.072;On\Є.072 RHC." ;On-.072 ;On RЀAgeofPerson1 \R! MbP-.001MbP\Є.001 RHC" MbP-.001 MbP RЀNumberofPersonsinHousehold 9/!. 9 .053* 0&!." 0ЀRenteroccupiedHousehold=1 \R! Zd;-.122Zd;\Є.122 ]SC" Zd;-.122 Zd; ]  Intercept ZP!.  Vi  rh-1.838 rhZЄ1.838YOM.!"   rh-1.838    rhVi Y  ɀXX#p*Indicatesstatisticalsignificanceat=.1.#X#pXɀ# J0! Inagreementwiththemultiplecomparisonresults,eithertypeofnotificationofadministrative # recorduseyieldshighercompletionforthePerson1SSNitemcomparedtononotificationeven  $ whencontrollingfordemographicfactors.Thisfindingagreeswiththepairwisecomparisons !% above,butcontradictsourinitialhypothesis.Inretrospect,webelievethatrespondentsmayview "& thisnotificationasjustificationfortheSSNrequests.Moreover,thestatementaboutthe #' potentialforshorterfuturecensusformsduetothedatalinkagemaybeapersuasivefactorin $| ( convincingrespondentstoprovidethisidentifier. %h!)  +'0  5.0  RECOMMENDATIONS (#(# Ingeneral,theeffectsofthetreatments(i.e.notification,SSNrequest)ontheresponseratesare  notassubstantialasoriginallyanticipated.However,someeffectsarenoticeable.Forinstance,  itisclearthatrequestingSSN,forallhouseholdmembersorPerson1,causessmall,but  significant,dropsinresponse.Specifically,theoddsofrespondingtothecensusdecreaseby  about9%whenPerson1SSNisrequestedandbyapproximately11%whenSSNisrequestedfor   allhouseholdmembers.Moreover,therequestforSSNcauseshigheramountsofincomplete  t formsreturned. z ` Notificationofadministrativedatauseiscollectivelyassociatedwithlowerresponseaswell, R 8  withgeneralnotificationshowinganindividualeffectwhenseparatedfromspecificnotification. >$  Onceagain,thedropissignificantbutslight.Specificnotificationhasamoreharmfuleffecton *  responsewhenSSNisrequestedcomparedtothecasewhenspecificnotificationistheonly   treatment.Generalnotificationdoesnotcausefurtherdropsinresponsewhencoupledwiththe   SSNrequestcomparedtogeneralnotificationalone.   Notificationdoesnotappeartohaveanynegativeeffectsonformcompleteness.Infact,  notificationofadministrativerecorduseactuallyincreasesresponsetothePerson1SSNitem  comparedtothecasewhennonotificationisgiven.  Giventhesefindings,theCensusBureaushouldbeawarethatanyrequestforSSNinfuture v\ censusesmaydecreasemailreturns.Yet,ifbetterSSNdataaredesiredabovehighermail bH response,theresultsofthisanalysissuggestthatnotificationshouldbeincludedwiththerequest. N4 Obviously,policyandprivacyissueswillbeatthecenterofanydebateoverwhethertoinclude &  SSNrequestsonfuturecensusforms.Curiously,only9ofthe210callsmadetoOperator  Assistance(_ OA_ )regardingtheSPANexperimentincludednegativefeedbackconcerningtheuse  ofSSN.dB  ԍ(30,page1and2,R1)C  dHowever,theCensusBureaushouldbepreparedforpublicairingofprivacyconcernsif  SSNrequestsareextendedtotheentirenation.Thisfindingsuggeststhatfocusgroupsshouldbe  convenedtoupdatefindingsfromfocusgroupsinthe1992SimplifiedQuestionnaireTest  regardingreactiontotherequestforSSN.     +'+ Acknowledgments   TheauthorswishtothankRandall_ Neugebauer_ ԀandDebbieBoltonfortheirinitiation,design,and  implementationofthisparticularCensus2000experiment.Withouttheirongoingsupport,this  analysiswouldnothavebeenpossible.SpecialthankstoRandallfordetailinganddocumenting  theissuesarisingwiththeshortformcontrolpanelinthisexperiment.Additionally,wewould  liketothankJimCopeandCharlesKahnfortheirassistanceinobtainingtheresponsedata.    References  z ` AguirreInternational. PublicConcernsabouttheUseofAdministrativeRecords,contractors R 8  report,July1995. >$  Bates,N. RevisedItemNonresponseResultsforSocialSecurityNumberfromtheSimplified   QuestionnaireTest(SQT),MemorandumtoRobertD._ Tortura_ ԀandSusanM._ Miskura_ ,dated   June18,1992.   Bates,N. FocusGroupReportonDataSharingandNotificationofDataSharing,Census  Bureaureport,CenterforSurveyMethodsResearch,July13,1995.  _ DSSD_ Ԁand_ PRED_ Ԁ(2000) Census2000EvaluationProgramQualityAssuranceProcess,July p 31,2000. v\ Cohn,D. CensusComplaintsHitHome,WashingtonPost,May4,2000. N4 _ DeMaio_ ,_ T.J._ ,andBates,N.A.(1990) WhoFillsOuttheCensusForm?,Proceedingsofthe &  ЀSurveyResearchMethodsSectionoftheAmericanStatisticalAssociation.  Dillman,D.A.,Treat,_ J.B._ ,andClark,_ J.R._ Ԁ(1994) Influenceof13DesignFactorson  CompletionRatestoDecennialCensusQuestionnaires,1994AnnualResearchConferenceand  _ CASIC_ ԀtechnologiesInterchangeProceedings,_ Rosslyn_ ,Virginia,March,1994.    Dillman,D.A.,West,_ K.K._ ,andClark,_ J.R._ Ԁ(1994) InfluenceofanInvitationtoAnswerby !  TelephoneonResponsetoCensusQuestionnaires,PublicOpinionQuarterly,58:557568. "l! Gates,_ G.W._ ,andBolton,D.(1998) PrivacyResearchInvolvingExpandedStatisticalUsesof ^$D # AdministrativeRecord,ProceedingsoftheSectiononGovernmentStatisticsandSectionon J%0!$ SocialSciences. 6&"% Hochberg,Y.,and_ Tamhane_ ,A.(1987).MultipleComparisonProcedures.JohnWiley&Sons. (#' _ Kostanich_ ,D.andHaines,D. AccuracyandCoverage_ Evalution_ ԀSurvey:LogisticRegression )%) Modelingfor_ Poststratification_ ԀVariableSelection,Memorandumfrom_ Kostanich_ ԀtoHogan *&* datedSeptember22,1999. +'+ Ї_ Kulka_ ,R.A.,Holt,N.A.,Carter,W.,andDowd,_ K.L._ Ԁ(1991) SelfReportsofTimePressures,  ConcernsforPrivacy,andParticipationinthe1990MailCensus,Proceedingsofthe1991  AnnualResearchConference.  LettertoHonorableDanMiller,Chairman,SubcommitteeontheCensusfromJ.Christopher  _ Mihm_ ,AssociateDirector,FederalManagementandWorkforceIssues,Subject 2000Census:  InformationonShortandLongFormResponseRates,UnitedStatesGeneralAccounting   Office,June7,2000.  t MemorandumfromHoganto_ Longini_ Ԁattention_ Stoudt_ , SpecificationsfortheIdentificationof f L  CoverageEditFailuresinCensus2000",_ DSSD_ ԀCensus2000ProceduresandOperationsMemo R 8  Series#DD3,May18,1999. >$  MemorandumfromHoganand_ Miskura_ Ԁto_ Longini_ , DefinitionofaBlankFormbyDCS2000",   _ DSSD_ ԀCensus2000ProceduresandOperationsMemoSeries#K3.   MemorandumfromHoganto_ Longini_ ԀandMarx, SpecificationsforlinkingCensusFormsand  SettingtheExpectedReturnPopulationCountontheDecennialResponseFile",_ DSSD_ ԀCensus  2000ProceduresandOperationsMemoSeries#C4,July13,2000.  _ Neugebauer_ ,R. ProgramMasterPlanfortheSocialSecurityNumber,PrivacyAttitudes,and p Notification(SPAN)Experiment,November5,1999. v\ Singer,E.(Forthcoming) PublicPerceptionsofConfidentialityandAttitudestowardData N4 SharingbyFederalAgencies,ChapterinupcomingbookentitledConfidentiality,Disclosure :  andDataAccess:TheoryandPracticalApplicationsforStatisticalAgencies. &  Singer,E.,andMiller,E. ReactionstotheUseofAdministrativeRecords:ResultsofFocus  GroupsDiscussions,CensusBureaureport,CenterforSurveyMethodsResearch,August24,  1992.  Singer,E.,_ Mathiowetz_ ,N.A.,and_ Couper_ ,M.(1993) TheImpactofPrivacyandConfidentiality   ConcernsonSurveyParticipation:TheCaseofthe1990U.S.Census,PublicOpinion !  Quarterly,57,465482. N  z  [        D  "l! Treat,JamesB(1996) TheEffectofQuestionnaireLengthonResponse,Proceedingsofthe ^$D # ЀSurveyResearchMethodsSectionoftheAmericanStatisticalAssociation. J%0!$  +'+  0?,XX4 X X,[X? TableA.1.  PanelsoftheSSNandNotificationcomponent,October29,1999.    *xVdddd dd LV](#(#x,d ,dd ,4dd ,edd , dd ,dd +  0iiV 0XX#pExperimental  PANELS F x  @iV/! "iiV @#X#pX#,T,XX#pPLANNED  < MAIL  OUT (  SAMPLE l   SIZE#X#pX,,T#XX#p     @i/!X  "iV @COVERLETTER @i/!"i @OUTGOING  ENVELOPE @i/!F x"i @QUESTIONNAIRE @i/!"i @RETURN  ENVELOPE#X#pXm# 9iV&!F x"i 9XX#p SHORTFORM    CONTROL WVH( ZiV @5249@W5,249 QGB " @5249 @V QOfficialCensus2000:   S701A.1(L) '( Z 'OfficialCensus   2000:S702A.1 '( Z 'OfficialCensus2000:#X#pXb#XX#pԀS700A.1 '  'Colored:#X#pXK#XX#pԀS   703A.1#X#pX#XX#p 0iV( Z  0ONE Z ! ALLSSNs WVH *"iV @5248@W5,248 QGBZ #" @5248 @V QStatementthatprovidingSSNis Z $ voluntary:S701A.2(L) ' *% 'OfficialCensus Z & 2000:S702A.2 ' *' 'SSNrequestforallhouseholdmembers Z ( andunique_ OA_ #X#pX%#XX#pԀand_ TDD_ Ԁnumbers#X#pX#XX#p:  *) S700A.2 '* 'Colored:#X#pX #XX#pԀ#X#pX#XX#pS Z + 703A.2#X#pX#XX#p 0iV *, 0TWO  - ONESSN WVHf .iV @5248@W5,248 QGB /" @5248 @V QStatementthatprovidingSSNis  0 voluntary:S701A.3(L) 'f 1 'OfficialCensus  2 2000:S702A.3 'f 3 'SSNrequestforonlyonepersoninthe  4 householdandunique_ OA_ #X#pX? #XX#pԀand_ TDD_  f 5 numbers#X#pX #XX#p:#X#pXA #XX#pS700A.3 '6 6 'Colored:#X#pX #XX#pԀ#X#pX #XX#pS  7 703A.3#X#pXA #XX#p 0iVf 8 0THREE 6h 9 ALLSSNswith  : General r ; Notification WVHB<iV @5255@W5,255 QGB6h =" @5255 @V QGenerallywordednotification 6h > languageandstatementthat  ? providingSSNisvoluntary: r @ S701A.4(L) 'BA 'OfficialCensus 6h B 2000:S702A.4 ' C 'SSNrequestforallhouseholdmembers 6h D andunique_ OA_ #X#pX #XX#pԀand_ TDD_ Ԁnumbers#X#pX#XX#p:  E S700A.4 'r F 'Colored:#X#pX#XX#pԀ#X#pX#XX#pS 6h G 703A.4#X#pX#XX#p 0iV H 0FOUR BtI ALLSSNswith J Specific ~K Notification WVHNLiV @5249@W5,249 QGBBtM" @5249 @V QSpecificallywordednotification BtN languageandstatementthat O providingSSNisvoluntary: ~P S701A.5(L) 'NQ 'OfficialCensus BtR 2000:S702A.5 'S 'SSNrequestforallhouseholdmembers BtT andunique_ OA_ #X#pXS#XX#pԀand_ TDD_ Ԁnumbers#X#pX`#XX#p: U #X#pX#XX#pS700A.5 '~V 'Colored:#X#pX/#XX#pԀ#X#pX#XX#pS BtW 703A.5#X#pX#XX#p 0iVX 0FIVE NY General Z Notification WVH[iV @5250@W5,250 QGBN\" @5250 @V QGenerallywordednotification: N] S701A.6(L) '^ 'OfficialCensus N_ 2000:S702A.6 '` 'OfficialCensus2000formwithunique Na _ OA_ #X#pXJ#XX#pԀand_ TDD_ Ԁnumbers:S700A.6 'b 'Colored:#X#pX#XX#pԀ#X#pXp#XX#pS Nc 703A.6#X#pX#XX#p 0iVVd 0SIX e Specific Zf notification YVVH*giVV }@5245}@Y5,245 XVGBh" }@5245 }@VV XSpecificallywordednotification i language:S701A.7(L) .VZjV .OfficialCensus k 2000:S702A.7 Zl  .V*mV .OfficialCensus2000formwithunique n _ OA_ #X#pX#XX#pԀand_ TDD_ Ԁnumbers:S700A.7 .VZoV .Colored:#X#pX#XX#pԀ#X#pXv#XX#pS p 703A.7#X#pX#XX#p 0ViVVZqV 0 LONGFORM  *\r CONTROL YVVVHsViVV t@5236t@Y5,236 XVVGB*\t" t@5236 t@VVV XOfficialCensus2000:S *\u 701B.1(L) .VVvVV .OfficialCensus *\w 2000:S702B.1 .VVxVV .OfficialCensus2000:S700B.1 .VV*\yVV .Colored:#X#pX#XX#pԀ#X#pX!#XX#pS *\z 703B.1#X#pX!#XX#p 0iV{VV 0SEVEN ,| Specific } Notification WVH6 h~iV o@5231o@W5,231 QGB," o@5231 o@V QSpecificallywordednotification , language:S701B.2(L) ' 'OfficialCensus , 2000:S702B.2 ' 'OfficialCensus2000formwithunique , _ OA_ Ԁand#X#pX("#XX#pԀ_ TDD_ Ԁnumbers:S700B.2 ' 'Colored:#X#pX$#XX#pԀ#X#pX\%#XX#pS , 703B.2#X#pX%#XX#p 0iV 0EIGHT h! General "8 Notification WVH"iV o@5231o@W5,231 QGBh!" o@5231 o@V QGenerallywordednotification h! language:S701B.3(L) '"8 'OfficialCensus h! 2000:S702B.3 '"8 'OfficialCensus2000formwithunique h! _ OA_ Ԁand#X#pX%#XX#pԀ_ TDD_ Ԁnumbers:S700B.3 '"8 'Colored:#X#pX(#XX#pԀ#X#pX-)#XX#pS h! 703B.3#X#pXp)#XXXX#p("8  (? ,Xrw ,hX,XX4 X?#X#pXXX)#XX#pz! #Xs%X!z*##X#pX%XsҨ*#Xs%XX#p#X#pX%Xs+#  r  D%v 0%  r TableB.1.NumberofBlankFormsbyPanel dB  ԍ(31,pages2to11,R1)C  d  *PWVjSd dd dd 4dd 4edd e dd  dd V(#(#,m dd ,dd +  +X +Panel =X,!PX =Numberof P BlankForms 6,! <"X 6 ShortFormControl   (SFC) G=v   ,@14,@G 14 =31 "  ,@14 ,@ = AllSSNRequest  ( (Panel1) >4     0@160@> 16 =31 ( "  0@16 0@ = OneSSNRequest b  (Panel2) >4 N   5@215@> 21 =31b "  5@21 5@ = AllSSNRequest,General   Notification   (Panel3) >4    ,@14,@> 14 =31 "  ,@14 ,@ = AllSSNRequest,Specific &X Notification  D  (Panel4) >4 0  9@259@> 25 =31&X"  9@25 9@ = GeneralNotification ~ (Panel5) >4 j  :@26:@> 26 =31~"  :@26 :@ = SpecificNotification  (Panel6) >4   0@160@> 16 =31"  0@16 0@ = LongFormControl(LFC) >4 V  >@30>@> 30 =31V"  >@30 >@ = SpecificNotification  (Panel7) >4   <@28<@> 28 =31 "  <@28 <@ = GeneralNotification B t! (Panel8) >4 .!`"  7@237@> 23 A75B t#"  7@23 7@ A n,&. 5%TableC.1.WeightedUAARatesbyPaneldB  ԍ(32,page2,R1)C  d  *xWVd Sdm dd m dd PWVj(#(#x,udd ,dd +  +X +Panel =X,!PX =UAArates P Overall 6,! <"X 6 AQEorReplacementShortFormControl(SFC) '  ' 8.6% ' " ' AllSSNRequest(Panel1) ' < ' 9.0% ' <" ' OneSSNRequest(Panel2) '   ' 8.0% '  " ' AllSSNRequest,GeneralNotification(Panel3) ' <  ' 9.4% ' < " ' AllSSNRequest,SpecificNotification(Panel4) '  ' 8.5% ' " ' GeneralNotification(Panel5) ' <  ' 9.1% ' < " ' SpecificNotification(Panel6) '  ' 8.3% ' " ' LongFormControl(LFC) ' < ' 10.5% ' <" ' SpecificNotification(Panel7) ' ' 9.2% '" ' GeneralNotification(Panel8) ' < ' 10.2% " <" "TableC.2.PairwiseComparisonsofUAArates(ShortForm)dB  ԍ(32,page2,R2)(33,page1,R1)C  d D *xAWVd dudd udd WV(#(#x,2dd ,edd"+  +X00 +PairwiseComparison =X,!bX =Overall b Difference At,!N"X A SFCPanel1(allSSNs) At,!t A .4% At,!"t A SFCPanel2(oneSSN) At,!N t A .6% 6,!N!"t 6 SFCPanel3(allSSNswithgeneralnotification) '" ' .8% #"  SFCPanel4(allSSNswithspecificnotification)  N!$  .1% t N!% SFCPanel5(generalnotification) /t"&t / .5% /t"'t / SFCPanel6(specificnotification) /tN$(t / .3% $N$)t $ Panel1(allSSNs)Panel2(oneSSN)  % *  1.0% t % + Panel1(allSSNs)Panel3(allSSNswithgeneralnotification) /tN'!,t / .4% /tN'!-t / Panel1(allSSNs)Panel4(allSSNswithspecificnotification) /t(#.t / .5% $(#/t $ Panel1(allSSNs)Panel5(generalnotification)  N*$0  .1%  N*$1  Panel1(allSSNs)Panel6(specificnotification)  +&2  .7%  +&3  Panel2(oneSSN)Panel3(allSSNswithgeneralnotification)  2d  1.4%  2d  Panel2(oneSSN)Panel4(allSSNswithspecificnotification)    .5%    Panel2(oneSSN)Panel5(generalnotification)  2 d  1.1%  2 d  Panel2(oneSSN)Panel6(specificnotification)     .4%     Panel3(allSSNswithgeneralnotification)Panel4(allSSNswithspecific 2 d  notification)   P   .9% t 2 d  Panel3(allSSNswithgeneralnotification)Panel5(generalnotification) /t t / .3% $ t $ Panel3(allSSNswithgeneralnotification)Panel6(specificnotification)  P   1.1%  P   Panel4(AllSSNswithspecificnotification)Panel5(generalnotification)     .6% t   Panel4(allSSNswithspecificnotification)Panel6(specificnotification) /tP t / .1% /tP t / Panel5(generalnotification)Panel6(specificnotification) /tt / .7% (  t ( ɀXX#p*significantwhen_ familywise_ Ԁ=.1.#X#pXɀJ#   TableC.2.PairwiseComparisonsofUAARates(LongForm)dB  ԍ(33,page2,R2)dC   j *xRWVSdd2dd 2edd"eAWV(#(#x,Kdd ,-dd"+  +XVV +PairwiseComparison =X,!X =Overall  Difference 8t#!"X 8 LFCPanel7(specificnotification) /t&Xt / 1.3% /t&Xt / LFCPanel8(generalnotification) /tt / .3% /tt / Panel7(specificnotification)Panel8(generalnotification) /t&X t / 1.0% (&X!  t ( ɀXX#p*significantwhen_ familywise_ Ԁ=.1. #X#pXɀ;O# Bt!  NotethatnoneoftheUAAratesdiffersignificantlyasexpectedfrompastresearchresultsand  $# thisexperimentaldesign.  !$   $' K TableD.1containsitemnonresponseratesforshortformquestionsontheexperimentaland  controlquestionnaires.Personspecificratesarecomputedforhouseholdsforwhichthe  respondentreportedhouseholdcountisgreaterthanorequaltothepersonnumber.Forexample,  nonresponseratesfortheageofPerson4arecomputedforhouseholdswith4ormore  individualsasreportedbytherespondent. ~   TableD.1.ItemNonresponseRatesbyPersonNumberandPanelforShortFormPanels.  V  *x&WV;ddKdd K-dd"-RWVS(#(#x,dd ,dd ,dd ,9dd ,dd ,dd ,9dd ,dd +  +XB tB t +Panel =X,! "X =Item =X,!  "X =Person    1    dB  ԍ(35,pages2and3,R1)C  d =X,!~ "X =Person    2   dB  ԍ(36,pages2and3,R1)C  Ԛd =X,!~"X =Person   3   dB  ԍ(37,pages2and3,R1)C  Ԛd =X,!~"X =Person   4   dB  ԍ(38,pages2and3,R1)C  Ԛd =X,!~"X =Person   5   dB  ԍ(39,pages2and3,R1)C  Ԛd =X,!~"X =Person   6   dB  ԍ(40,pages2and3,R1)C  d :0!~"X :SFC 0  (panel=0) 8.# YY 8Pop 0  countdB  ԍ(34,page2,R1)C  d '  '1.82% 0  (.19) ' !" ' '0 "" ' '0 #" ' '0 $" ' '0 %" ' "0 &" " ! 8 & !Age 'j ' '1.03% j ( (.19) 'V )" '1.60% j * (.31) 'V +" '2.38% j , (.45) 'V -" '3.32% j . (.83) 'V /" '4.52% j 0 (1.16) 'V 1" '3.00% j 2 (1.24) "V 3" " ! r3 !DOB '4 '1.06% 5 (.27) '6" '1.96% 7 (.33) '8" '2.08% 9 (.55) ':" '2.98% ; (.90) '<" '5.65% = (1.73) '>" '2.40% ? (1.18) "@" " ! @ !Sex 'BtA '0.33% BtB (.12) '.`C" '0.35% BtD (.14) '.`E" '0.96% BtF (.33) '.`G" '1.63% BtH (.53) '.`I" '3.74% BtJ (1.13) '.`K" '8.60% BtL (3.05) ".`M" " ! J|M !Race 'N '1.50% O (.19) 'P" '2.32% Q (.29) 'R" '3.41% S (.47) 'T" '4.22% U (.75) 'V" '5.99% W (1.34) 'X" '3.60% Y (1.38) "Z" " ! Z !Hispanic L[ Origin '8\ '3.23% L] (.36) '8^" '2.95% L_ (.42) '8`" '4.78% La (.64) '8b" '3.93% Lc (.81) '8d" '5.60% Le (1.20) '8f" '6.10% Lg (2.51) "8h" " ! "Th !Reltnshp i toP1 'rj ' 'k" '0.92% l (.20) 'rm" '0.87% n (.24) 'ro" '1.07% p (.40) 'rq" '2.40% r (.93) 'rs" '0.60% t (.61) +!ru" +Panel1  $v (allSSNs) 8.#!wNN 8SSN ' $x '15.78%  $y (.66) '!z" '21.55%  ${ (.92) '!|" '28.57%  $} (1.21) '!~" '28.13%  $ (1.63) '!" '30.93%  $ (4.21) '!" '29.03%  $ (5.00) +!!" + ! ", !Pop ^# countdB  ԍ(34,page2,R2)C  d 'J$| '2.79% ^# (.36) 'J$|" ' '^#" ' '^#" ' '^#" ' '^#" ' "^#" " ! f% !Age '% '1.56% % (.25) '& " '1.96% % (.31) '& " '2.14% % (.46) '& " '2.85% % (.72) '& " '5.45% % (1.52) '& " '1.76% % (.77) "& " " ! '" !DOB '6(h" '1.75% 6(h" (.23) '")T#" '2.17% 6(h" (.26) '")T#" '2.43% 6(h" (.39) '")T#" '2.96% 6(h" (.55) '")T#" '3.92% 6(h" (1.36) '")T#" '1.32% 6(h" (.73) "")T#" " ! >*p$ !Sex '*$ '1.18% *$ (.17) '+%" '0.72% *$ (.17) '+%" '1.31% *$ (.32) '+%" '2.80% *$ (.72) '+%" '2.70% *$ (1.29) '+%" '2.20% *$ (.96) "+%" " !  !Race '2d '2.12% 2d (.19) 'P" '3.33% 2d (.31) 'P" '4.99% 2d (.61) 'P" '5.65% 2d (.99) 'P " '6.50% 2d  (2.65) 'P " '5.42% 2d  (2.43) "P " " ! :l  !Hispanic  Origin '  '3.40%  (.33) ' " '3.90%  (.42) ' " '4.83%  (.55) ' " '5.54%  (.93) ' " '6.18%  (2.25) ' " '5.35%  (2.10) " " " !   !Reltnshp  < toP1 ' ( ' ' <" '0.67%  < (.19) ' ( " '0.83%  <! (.29) ' ("" '0.92%  <# (.38) ' ($" '0.75%  <% (.62) ' (&" '1.76%  <' (1.04) +! ((" +Panel2 v ) (oneSSN) 8.#b*NN- 8SSN 'v + '15.45% v , (.77) 'b-" ' 'v ." ' 'v /" ' 'v 0" ' 'v 1" ' +!v 2" + ! ~ 2 !Pop  3 countdB  ԍ(34,page2,R3)C  d ' 4 '1.77%  5 (.28) ' 6" ' ' 7" ' ' 8" ' ' 9" ' ' :" ' " ;" " !  ; !Age 'N < '1.16% N = (.25) ':l >" '1.70% N ? (.36) ':l @" '2.25% N A (.53) ':l B" '1.93% N C (.54) ':l D" '2.63% N E (1.03) ':l F" '8.86% N G (4.06) ":l H" " ! VH !DOB 'I '1.51% J (.29) 'K" '2.52% L (.44) 'M" '3.00% N (.59) 'O" '2.65% P (.79) 'Q" '1.58% R (.84) 'S" '5.93% T (2.85) "U" " ! U !Sex '&XV '0.65% &XW (.21) 'DX" '0.80% &XY (.29) 'DZ" '1.28% &X[ (.36) 'D\" '1.47% &X] (.52) 'D^" '1.97% &X_ (.91) 'D`" '4.64% &Xa (2.42) "Db" " ! .`b !Race 'c '1.77% d (.21) '~e" '2.80% f (.28) '~g" '4.27% h (.75) '~i" '5.09% j (.73) '~k" '6.95% l (1.32) '~m" '9.61% n (3.63) "~o" " ! o !Hispanic 0p Origin 'q '3.18% 0r (.26) 's" '2.67% 0t (.42) 'u" '2.79% 0v (.49) 'w" '2.38% 0x (.72) 'y" '1.56% 0z (.58) '{" '2.93% 0| (1.75) "}" " ! 8} !Reltnshp j~ toP1 'V ' 'j" '0.58% j (.18) 'V" '0.77% j (.37) 'V" '0.97% j (.39) 'V" '0.26% j (.18) 'V" '2.52% j (1.76) +!V" +Panel3   (allSSNswith ! general " notification) 8.##m 8SSN '  '12.55%   (.78) '!" '17.31%   (.87) '!" '28.76%   (1.55) '!" '31.12%   (1.84) '!" '34.70%   (4.28) '!" '37.99%   (8.09) +!!" + ! " !Pop B#t countdB  ԍ(34,page2,R4)C  d '.$` '1.96% B#t (.30) '.$`" ' 'B#t" ' 'B#t" ' 'B#t" ' 'B#t" ' "B#t" " ! J%| !Age '% '1.59% % (.28) '& " '2.50% % (.53) '& " '3.20% % (.78) '& " '5.79% % (.84) '& " '8.24% % (2.67) '& " '8.89% % (4.87) "& " " ! '! !DOB '(L" '1.65% (L" (.29) ')8#" '2.47% (L" (.49) ')8#" '3.57% (L" (.82) ')8#" '5.18% (L" (1.07) ')8#" '8.24% (L" (2.35) ')8#" '6.02% (L" (4.11) ")8#" " ! "*T$ !Sex '*$ '0.79% *$ (.19) 'r+%" '0.88% *$ (.24) 'r+%" '1.35% *$ (.40) 'r+%" '2.78% *$ (.76) 'r+%" '3.69% *$ (1.52) 'r+%" '3.01% *$ (3.03) "r+%" " !  !Race '2d '1.78% 2d (.23) 'P" '3.98% 2d (.44) 'P" '6.05% 2d (.75) 'P" '8.79% 2d (1.18) 'P " '12.90% 2d  (2.60) 'P " '14.19% 2d  (5.19) "P " " ! :l  !Hispanic  Origin '  '3.33%  (.31) ' " '3.97%  (.37) ' " '6.30%  (.89) ' " '5.61%  (1.20) ' " '7.34%  (2.35) ' " '8.17%  (4.47) " " " !   !Reltnshp  < toP1 ' ( ' ' <" '0.93%  < (.22) ' ( " '1.39%  <! (.37) ' ("" '2.68%  <# (.79) ' ($" '2.25%  <% (1.46) ' (&" '3.01%  <' (3.03) +! ((" +Panel4 v ) (allSSNswith b* specific N + notification) 8.#:l , 8SSN 'v - '11.44% v . (.46) 'b/" '15.74% v 0 (.0077) 'b1" '22.88% v 2 (1.66) 'b3" '24.50% v 4 (2.37) 'b5" '30.59% v 6 (4.02) 'b7" '47.34% v 8 (5.46) +!b9" + ! ~ 9 !Pop  : countdB  ԍ(34,page2,R5)C  d ' ; '3.48%  < (.40) ' =" ' ' >" ' ' ?" ' ' @" ' ' A" ' " B" " !  B !Age 'N C '1.39% N D (.20) ':l E" '2.28% N F (.40) ':l G" '3.26% N H (.36) ':l I" '3.31% N J (.68) ':l K" '2.93% N L (1.36) ':l M" '10.33% N N (4.66) ":l O" " ! VO !DOB 'P '1.25% Q (.23) 'R" '2.42% S (.51) 'T" '2.35% U (.58) 'V" '1.94% W (.51) 'X" '4.25% Y (1.39) 'Z" '9.50% [ (4.61) "\" " ! \ !Sex '&X] '0.75% &X^ (.17) 'D_" '0.82% &X` (.19) 'Da" '1.20% &Xb (.39) 'Dc" '2.06% &Xd (.56) 'De" '2.93% &Xf (1.17) 'Dg" '3.35% &Xh (1.85) "Di" " ! .`i !Race 'j '1.94% k (.33) '~l" '3.28% m (.40) '~n" '3.74% o (.53) '~p" '4.94% q (.79) '~r" '8.00% s (1.94) '~t" '12.01% u (4.95) "~v" " ! v !Hispanic 0w Origin 'x '3.24% 0y (.38) 'z" '4.40% 0{ (.56) '|" '5.33% 0} (.83) '~" '5.01% 0 (1.12) '" '5.89% 0 (2.40) '" '1.67% 0 (1.08) "" " ! 8 !Reltnshp j toP1 'V ' 'j" '0.92% j (.21) 'V" '1.38% j (.38) 'V" '0.94% j (.40) 'V" '1.79% j (1.02) 'V" '2.51% j (1.75) +!V" +Panel5 ! (general " notification) 8.## 8Pop   countdB  ԍ(34,page2,R6)C  d '! '2.09%   (.29) '!" ' ' " ' ' " ' ' " ' ' " ' " " " ! " !Age 'B#t '0.90% B#t (.15) '.$`" '1.65% B#t (.29) '.$`" '1.88% B#t (.48) '.$`" '2.02% B#t (.47) '.$`" '3.95% B#t (1.60) '.$`" '5.58% B#t (2.48) ".$`" " ! J%| !DOB '% '0.87% % (.18) '& " '1.64% % (.25) '& " '1.34% % (.31) '& " '1.51% % (.37) '& " '2.24% % (1.03) '& " '5.58% % (2.37) "& " " ! '! !Sex '(L" '0.41% (L" (.12) ')8#" '0.48% (L" (.15) ')8#" '1.16% (L" (.29) ')8#" '1.67% (L" (.43) ')8#" '4.45% (L" (1.70) ')8#" '7.25% (L" (2.55) ")8#" " ! "*T$ !Race '*$ '1.58% *$ (.21) 'r+%" '2.71% *$ (.30) 'r+%" '4.72% *$ (.75) 'r+%" '4.48% *$ (.76) 'r+%" '5.95% *$ (1.11) 'r+%" '7.60% *$ (2.31) "r+%" " !  !Hispanic 2d Origin 'P '2.22% 2d (.29) 'P" '2.26% 2d (.43) 'P" '2.91% 2d (.56) 'P" '2.92% 2d  (.74) 'P " '3.75% 2d  (1.01) 'P " '2.93% 2d  (1.00) "P" " ! :l !Reltnshp  toP1 '  ' '" '0.54%  (.22) ' " '0.86%  (.40) ' " '0.70%  (.28) ' " '2.11%  (.99) ' " '3.00%  (2.17) +! " +Panel6  ( (specific   notification) 8.# - 8Pop  <  countdB  ԍ(34,page2,R7)C  d ' (! '2.48%  <" (.26) ' (#" ' ' <$" ' ' <%" ' ' <&" ' ' <'" ' " <(" " !  D( !Age 'v ) '1.20% v * (.17) 'b+" '1.34% v , (.26) 'b-" '2.36% v . (.48) 'b/" '2.33% v 0 (.58) 'b1" '5.67% v 2 (1.51) 'b3" '12.60% v 4 (4.79) "b5" " ! ~ 5 !DOB ' 6 '0.77%  7 (.18) ' 8" '1.50%  9 (.23) ' :" '1.95%  ; (.43) ' <" '1.69%  = (.44) ' >" '1.31%  ? (.90) ' @" '5.50%  A (2.89) " B" " !  B !Sex 'N C '0.68% N D (.13) ':l E" '0.49% N F (.15) ':l G" '1.16% N H (.34) ':l I" '2.38% N J (.66) ':l K" '3.61% N L (1.04) ':l M" '6.57% N N (3.24) ":l O" " ! VO !Race 'P '1.52% Q (.20) 'R" '2.91% S (.37) 'T" '3.74% U (.60) 'V" '5.18% W (1.00) 'X" '7.80% Y (1.81) 'Z" '8.16% [ (2.97) "\" " ! \ !Hispanic &X] Origin 'D^ '2.29% &X_ (.26) 'D`" '2.45% &Xa (.38) 'Db" '4.47% &Xc (.68) 'Dd" '3.89% &Xe (.71) 'Df" '4.89% &Xg (1.12) 'Dh" '5.94% &Xi (2.56) "Dj" " ! .`j !Reltnshp k toP1 '~l ' 'm" '0.87% n (.21) '~o" '0.93% p (.30) '~q" '0.99% r (.36) '~s" '1.01% t (.61) '~u" '3.28% v (2.32 )"~w" 5 " TableD.2containstheitemnonresponseratesforthetotalincomequestionforthelongform x panels.RatesinTableD.2arecomputedacrosshouseholdsforwhichtherespondentreported vy householdcountisgreaterthanthepersonnumberoftheitemandthepersonofinterest. bz Additionally,theseratesarecomputedoverallhouseholdswherethepersonofinterestis15 N{ yearsoldorolder(inaccordancewithcensusformskippatterns,personslessthan15yearsof : l| ageskipthisquestion). &!X}   "D~ TableD.2.ItemNonresponseRatesbyPersonNumberandPanelforLongFormPanels   *QWVxd6ddd dd dd 9dd 9dd dd 9dd 9dd &WV;(#(#,dd ,dd ,dd ,9dd ,dd ,dd ,9dd ,dd +  +X +Panel =X,!P"X =Item =X,!P"X =Person P 1  < dB  ԍ(41,page2,R1)C  d =X,!("X =Person P 2  < dB  ԍ(42,page2,R1)C  Ԛd =X,!( "X =Person P  3  <  dB  ԍ(43,page2,R1)C  Ԛd =X,!( "X =Person P  4  < dB  ԍ(44,page2,R1)C  Ԛd =X,!("X =Person P 5  < dB  ԍ(45,page2,R1)C  Ԛd =X,!("X =Person P 6  < dB  ԍ(46,page2,R1)C  Ԛd 9/!("X 9LFC v   0&!b  0Total v  Income 0&b  026.60% v  (1.36) 9/!b " 936.19% v  (1.50) 9/!b " 945.22% v  (2.23) 9/!b " 953.09% v   (4.18) 9/!b !" 935.79% v " (5.54) 9/!b #" 958.38% v $ (12.60) 6,!b %" 6Panel7  & (Specific  ' Notification) ( Total  ) Income   * 24.76%  + (1.05) ' ," '37.05%  - (1.24) ' ." '48.15%  / (2.19) ' 0" '54.94%  1 (4.37) ' 2" '48.07%  3 (7.84) ' 4" '45.60%  5 (17.98) ' 6" 'Panel8 :l 7 (General &X 8 Notification) D 9 Total :l : Income  &X ; 26.60% :l < (1.16) '&X =" '38.68% :l > (1.52) '&X ?" '46.51% :l @ (3.35) '&X A" '51.11% :l B (5.80) '&X C" '49.30% :l D (9.00) '&X E" '50.97% :l F (20.85)"&X G"  " TableD.3.StatisticalComparisonsofItemNonResponseRatestotheTotalIncome LH QuestionforPerson1ontheLongForm B  ԍd(42,page2,R2)C  Ԛd 8I *WVdddd dd dd 9dd 9dd dd 9dd 9dd QWVx(#(#,Kdd ,-dd"+  +X$J$ +PairwiseComparison =X,!VKX =Overall VL Difference 8t#!BtM"X 8LFCPanel7(specificnotification) /tNt /1.84% O (1.89) /tPt /LFCPanel8(generalnotification) /t.`Qt /0.00% .`R (1.82) /tLSt /Panel7(specificnotification)Panel8(generalnotification) /tTt /Є1.84% U (1.55)(V  t ( r    `       h      p      *%\\ E؝K *~FVGddKdd K-dd"-WV(#(#~,dd ,dd ,rdd ,rdd +  $ $TableE.1.MailResponseRates(StandardError)forSSN P RequestandControlPanelsbyStrata EV4)6h"  i EExperimentalPanel =VV,! "V =Full   PopulationdB  ԍ(15,page1,R1)C  d =VV,! "VV =HighCoverage   AreasdB  ԍ(15,page2,R2)C  d =VV,! "VV =LowCoverage    AreasdB  ԍ(15,page2,R2)C  Ԛd -#!  "VV -ShortFormControl ,VN   ,73.1% N   (.70) ;V,!: l "V ;75.9% N  (.83) ;V,!: l"V ;60.8% N  (1.03) -#!: l"V -AllSSNs(Panel1) ,V ,72.0%  (.77) ;V,! "V ;75.3%  (.92) ;V,! "V ;57.9%  (1.01) 0&! "V 0OneSSN(Panel2) >V/!&X i >71.7% &X  (.77) >V/!D "Vi >75% &X  (.93) >V/!D "Vi >57.4% &X  (1.06)1'%D "  Vi 1 r *MV[dddd dd rdd rrdd rFVG(#(#, dd ,Fdd ,dd ,dd +  $L L $TableE.2.MailResponseRates(StandardErrors)forNotificationand ~! ControlPanelsbyStrata @ EV4)""  i EExperimentalPanel =V,!Bt#"V =Full Bt$ Population =V,!.`%"V =HighCoverage Bt& Areas =V,!.`'"V =LowCoverage Bt( Areas -#!.`)"V - ShortFormPanels:d B  ԍ(15,pages1and  - 2,R3andR4)C  d '* ' '+" ' '," ' -" ЀShortFormControl '.`. '73.1% .`/ (.70) 'L0" '75.9% .`1 (.83) 'L2" '60.8% .`3 (1.03) L4" ЀGeneralNotification(Panel5) '5 '71.8% 6 (.66) '7" '74.8% 8 (.81) '9" '59% : (1.05) !;" !ЀSpecificNotification(Panel6) 9/! 8< 973.5%  8= (.72) 9/! $>" 976.8%  8? (.86) 9/! $@" 959%  8A (1.00) -#! $B" -ЀAllSSNswithGeneralNotification r"C Ѐ(Panel3) '^#D '69.5% r"E (.72) '^#F" '72.6% r"G (.85) '^#H" '56.6% r"I (1.09) ^#J" ЀAllSSNswithSpecificNotification $K Ѐ(Panel4) '%L '70% $M (.76) '%N" '73.0% $O (.91) '%P" '56.9% $Q (.98) %R"  LongFormPanels: dB  ԍ(18,pages1and )#W 3,R1andR2)C  d 'J'|!S ' 'J'|!T" ' 'J'|!U" ' J'|!V" ЀLongFormControl '("W '63.5% ("X (.81) ')#Y" '66.8% ("Z (.98) ')#[" '48.7% ("\ (1.02) )#]" ЀGeneralNotification(Panel8) '6+h%^ '62.2% 6+h%_ (.80) '",T&`" '66% 6+h%a (.95) '",T&b" '46% 6+h%c (1.04) !",T&d" !ЀSpecificNotification(Panel7) 9/!di 963.9% d (.81) 9/! P"i 967.4% d (.98) 9/! P"i 948.5% d (1.07)1'% P" i 1