multilevel/0000755000176200001440000000000012750607301012435 5ustar liggesusersmultilevel/NAMESPACE0000644000176200001440000000051712750500663013662 0ustar liggesusersimport("stats") import("MASS") import("nlme") importFrom("graphics","lines","plot") exportPattern("^[^\\.]") S3method(summary, rgr.agree) S3method(summary, rgr.waba) S3method(quantile, rgr.waba) S3method(summary, agree.sim) S3method(quantile, agree.sim) S3method(summary, disagree.sim) S3method(quantile, disagree.sim) multilevel/data/0000755000176200001440000000000012750500645013351 5ustar liggesusersmultilevel/data/klein2000.rda0000644000176200001440000013747412750500644015464 0ustar liggesusers7V DJP\PB4H4*%%2ev<woEp'hLpʮ-b ݨ˾vO\녚EJG8-jQh ΋߭? UW >IO SmF`_9H䏋fdpHPo -D U{ɂz!yiIV_q8\2ؓ4jJ.u+%iN*ԽvsuSg6evL ?yY,:r8wYXrt\'GZ)A&~ o~j|y`93ThK)ߒ 3b{s.6uoi$]p8]$Whvedz56FyWa''B:ܱNUL.f S+3ʠD]WΞf|*n`d _QV: niʼ+fHl]JkeswLcGү$~l\:Q.8yl&.7] - dKrޢճĩ'g Nt}2L|a5]s,-ɦO4KSI2"4C'qL?\?T󢏮0`Kn &~gGEԕw>z{FQV8ō{\utGD[\ݢ cic3?\1N`.D+È@jrTo HƯc,Y,Z0&]g,rK%Iг\/A^f/,Zx͙̌/CŃ߷ 3v,zg =e*8ӜMpȮGi*ڻU)DfYסv;{B儶NrR طi⽡ 8ӏH0|Ar(hǾslvC\wzAdpy3X#SorS֯ h-BC38te6Y ߋ~:-_30o:zUՖ2S?*ʧ5nez=hY.&3u6w{vz!{# ?!@UXlwMUR?{A2:0'`5c`7NxOpfΘULҰlbL pفfmvsRV6[M$B8ΜX9da{S>-쏒m_SY&A;c,H&Q<:u\ Cٽ)}ϋ?wi`*OE2$ڿĵȆM8ߺ:+JPe1ި1#bNϧl4.3EK}],AsgXxdsoq4$\ؓ,n!*Հu: [L;ƱDwk)=ج5#iD0x!ϜԸW='(Qwbw[m( ύN[cV-kjl$)-zsݱgqG0Igݨ`3|؞E>Y7q|H <6y`RŎ. ڢ%.ȼd-LR2! .w&ꃩP{݆ܠu~M"svVnhr6A"``n3ɝ3d9O'-6։^?q<%ɫ7( .JI{ . Z/Zq!ݾoHz 8/${ÜoUy1֟ D-5 JNfd,g9 s[SVt1hXz0[|IR/'Z$-^#٨e/4aRAf7y#iCP1n3W.=%邺շeTS>&.Ԓ*-V>X@R{/xġS'dÿrq@_s{Q*Cr~|j4ԒLp*bSW-5/3cfOQq!(~TW NH$wj¶|ݢ$ iϰo0 ^WĆND[NS9AԋV$|by:ET?%nGzJt>Ux@j# U2gt V<K#A8<{̆a C kV2C[ziGp9Ր6__).w%WH0$|On$8D$M'oǎAUU=Z6BʷZ [tbԫntFvZm2ICiVKRT`|Kҹ 7"h}%pl|gM-jb,EPOKZ~LbriľOQCW[h^|9xL ti*iCc7tbfM HSE ڼ$㽨#46gGauTN(1 j;{[u^_GھӮZضCdۗ #"`L5BUHO;jOf#$Ε#;"c6]=-m;4w=8`g  *V.Gq_sCsm 5)/Uʓ?j1?=}Umw×H^O05mn\- ꎙt^41$hR_CJ^2;8 $`#t58Nof?BRڌ3$WV @9y7./1 \a_V|tN>z$F;G @q`36ܾ$hw;JԱ{[aL'u[,kf) vRÖ y:6~92o̺#CeNp6,$vZ/kŋd|gIAWfzzU/ - k~aO<)7pukIm.U.:&͜NuY^$&5UM{tTԆ5W$8 Odr sKwÜQ3[-?Os\&ǤS\I\{W.5]p^9y0=AswDg+s}W.i<ྑGU>W[5Z$OV9gID\I6#%y4q[- */)3iq!FN߶`T<4;K&@KW0Q>9=IMO4=ov+ۑ4 = ֙{#,DV4Z!Y^ʷC}1PGZ3r34+~U9{qgF{}_.KмȰv_CJݓm$cZɬA!]WIc!O|'h[_2vGCA`VbտY,϶SzRwk6-evxc7?K)cokڷUH;Bī?9jyWa**A-`}1%TsmÄ^ e'n5yDbY 抾tMp\v{6Iȯ~gW z RKg?f6ԌzeaKX'ťHBq1W 3V gGk).6/ n=OD4l F3 c0 lP:ǪNe:K˒r/),& | f$i<6Kۯ)KP:Ht+`C#Xm8'$sAhrF}\?,~`Vj^2!$w)#yW@sZg-{->f̖V[I>ZpLm^3l\RB(z$I]Y"y}~2spX@Ij `0dU ]yaVEOQ Θ-g(ΦcSAП\<0YkGtߔ4"*l\҄;%^i}0 8wt9CK'BFB]EŴ?fԸ^{Iwp8<;rv5{1 d́ʉS9 WeJ2Y8ZFfY]TΒ[^H KOB.R%bSWqJIJ.UTU%-Hwa _a+Y$N<Ǥ.$Vj<<=&_R 9 : \7?z^Tmkn'Kr"-:v1)-sǏ XHURcM\; fIOuIh*E ,dϫGBWX~ui:R/({|59;)8!o {BB+k@Ф\@Inla&wTȘ$IXd$KCWnꝖtM=P FSմIq,ݯOT2/űf[u!aXH.X>TO^Fkz:ǜMPpRqK^ACz;޵ KF ߟr4eaebCjC/6 Sumd|F}y{E`vcۗE Q(c*ա$'!9Xvத.LxLu(t%)Jߊ6'nƕ67/thXa&K@߯yw;"l z~}n^ Ҹ{$|"IgF>sEmy+ 駊3M.PX OK0Nc)u^83S%a\^`;Dpz*.qI;$*#IZLk0d.Ahv(ڴ5FUO@cϺ$}X Q|Sx`g Ӕ $Sl v1>X"DYj$,WN-b$[\詠?)Wl ɾ۠|Iܧqu7ty5&DgsFf@5Rhřh r{N@s4.kK}Z^k:{ߎP4CjWz:=aWr)\''qװSW+Ѹya:n?Rvݞv$~ۗ knloGl7_ae9C Yg$V!҆5Iw25|SM)!f^*޿TIE2K]|rYlնzPO߻9 nqf_`eOz7ީ®%H&`Jc:6*J -7o&YRFo{ܠ[Tvڹ1;wBgFS(񣱱(Nd&3bJp0kK螬ْ.$¢3}mzmθr q^-0g7fFZǷ@rKυʽG}vgGX/wD V-fj>X(mbƑs`n>L5aH- MG$ ~B7_Bi!X9x=t\O0ӤaHzc?BAMh#nKs;n_ T׷zVowgsl๗hTUo$wjbkk8{W!_ŀh0@N[YgJ&m_NPwfOj{>n:7VxptKmB6NobkQ[=E?H =k~\; SwBA&=:S,Hگ+~hcqe痒~ֳ*xA^'"ڨ^fS#F{2#}7(S*2xuVLw{:6ŨSGH [xޘB2=2u`NڒUlN+l:Cd"h8νZ/C]0{9ɕ2sRdHPsTJ )pR<;&:ȃ:H$Yurd?6`[D/̔>f*)=I:*yI38S |Y~{P/ٷϩ&} IS &Xh 1e2F/[A߳T-#G}taZK+ G=s_{_(duAxMݜ}^z/K=- Tk96lӛdE>U%e z=b;RnM%dҷ 9$ ?z&0r-gi,̿o"h6 zt|`ibhL:EQ= K}U^i)ͰbRo]8A_R!dJ0{(j+<מk "HEq@ RKmyo:==+EWI&:#!$WAY\w;P v uACڢbu1Hx ӌo_Ԯl>Ep}Q>&aBXTEnB=3/8MU_XkRn>0ۺsK(=Nv6aD5Ku?ݤꎽ(k/S%G wr\nL~xd;;jn"r*z]zK\eNnsLIj\aGMl .7f_~,GC<}m]5\WU&髭R ŘgÀdy{vY>az?D}M?;-4u+{IZGL0pcT}YɪL OtZWϦ`{G3PߴZ?lO"i^4s%ziMحГ!pR`Lf3F0UO^+ =May&/9a.mnu̡tD#xZ.E"w`|Sb߅l wY"w cÇʼ-w5w*efZ4]>qf7W]0NT?%8o piIJ5֝Y $ז_pFTX1< ]n/\֮<%J/NmKϻ*Z uZ3ǷsQVn"ۈZIoH~!e]lƫkFݺ+ko:`V(7Rx#[cG#fwPZut˛9qijZd+AWap#yb;@X^KW\&Ky2PnM')^kWΞp9Mw#}c0_v.-]S-<<O.9@/矓P2re,(Ef冩E`,攷yY\#ĩ<#6Dq`II/9-{Hfzm*L$m MӰk;hmLP+W%Xhz%ן;>3 ze)ȓiN[ZG*߉R6gzu$X 4?lOs!0l~5Sa=|rjyc̵|JU{.BP[ɇ<zh%3ƙ8cۜv/I _s}2H^@i ՞e?aO.=)8ˋjtXCfaG iD4V>WC`b[]U) 7SSѮ8u//݊K]4JqQ́l T;ut09t΅n^$aluB{R8^Ld{bdS{z ԅC$~lŰe,2'f>GCmA.|u7vނק1 9x_֐ߞoĂjw=Mpo <^B{ynzCpb S$mnVJ5-Y)d΋l_ ?OoŻ٪h]k&\}-8]&8]hĩ;+b݇g߁j&UV`Ng8K2eI/uzܕpqq[?_{;aM$A^|^{Ư`LYwX:Wx,IwCh?k7ă};I޼$#bjtdx"-'? Rqۮ.د#^9tȏKpͧ?&h1/36dO]D>T/g nd!>؇r=OX%YE%9?PۗevWBI^,^&Wݗ4qQB'.]"yzj)3ͪmFeN+J9K\Uw$IuZ[!lT1@m'?pvfaɎ:=3{5_1̒$+ftEKLÂlG`du7RM=ҢUۙ'?cs8E><drKɓ,"M໶b3~)F~9- X\d#_Iǻ-D۞MpL;:Sw|w2\`Uo4T_w2?6o{d<[~É+* |I:Nu% ʢxMuE7|_dSmh8deECḮ{{6]xnf<]a%U*!%U| M+UKI~`w G~坲L[y%m +L96E)7cw>% *2'2̋Lzp/ɎP;84&KEбO5*!IgBuL9AKA]P*Lej58 ׆O`O+/C3<7 s +EĜJSnpn:m,vd_^>/H!L=G-qO1FpN=_nGJ_^">D[e!Wa U^V> Ac!V#RBgs~֌E4I+@$}DP)-jɒt'- [Vq1 3x|<@Rv=ٹFBaǽ9`eTRBoa4cYMg$G.v?NPȚxB/^%8$^`` d+*/C3$m`,>8e$| [ ágzgzU?N0l9Ce^<7IUCk&8\}I.tlss/ c\6=ޝOr>L!ƪ}'!8W^aZag ,)nƼ 1+hc۴K ^{񶏈gq9l-%9G4IN,\oǎ˘t;__K5݄uۣ(V+ړ#]eZo;IvCOiw޽C,1m҆aPƴ\>#{BVC _4%Au3Szo-c`xe"si{y;#d\#j[tʮeI;1[`ἔ^3Q3iS@#~0*0!$3P )̫>O-07`$EKYα]mnylJ`/AeT 9"r"Fe`?9N ^]*D8d>ou0n1E'TDn8;8t%Ii3yN -񷠴6A&˪Y*qmJAGv6֐ζQXx6黧@PS]68s8*c"g1wٷIXCㆭܧ*$;*E^MVcGv䶉 _f3{1׷wׂ 'y.>jzi WN zvLf5aaj]3`- [,ߤFT|W33PvZni]Xϥ6駭8dWӳ|NHs*%%,Y:6H[=.L|r O{g^ ?52V28ηqnXZOk`G2džw` Sژo#vvo [W7"?9YeU{jmɷ8D',Ku|5jLai֏}"Hap+2ϴMy1JR璌qŽXKַ:wjiW  㰛04i{Tx>5gtYCEGS3?L+'װGXAj7wow<0QfK Y;S?GLH_,"iCԛcaSS3T=*oW;6^܁ٴ|mHCL#&b6E59˞![{CMMHdӼr" ǰJ @p" Xۜw?vmLn-jCWiM=:k{R=YߨLwo̟'g[qPD}(=TdrKg z%} zs`V[% ,|;v5O KiT$'N Db[u-'% u$d o.H|3) t$8P`[5C-'_A'z e/h?Ǖafнy~:ƥ[(:bpeba)"pNʝ4n')w߉^m-fTINW'EGt^0K5B^y}N՛QK12c$m=(0/qRla0׿;JZa*-x(?p|D51v<0B͊j5 ^z7ܡӭ!'ӰlĺY7AgT6܇eMYƏSC>uu>6ɳ~5͵~m^~n7ӑ`lCjFa3g}-B8Loy$z 4T9{ մ.\05}b &:VFhj1C0go v y>*j8/{$ 'EV>sfcj^L V9ݶ$C]O a~U?Y`_+oف}z8aЎgˁPpztz/ZaO|`0Ac̪+w/kwY%X,d[6\/=>~'Gg)7$-xa~Iv-j`v{i7@ɠza Fh:FM"w1/?4{A?/a"ξ3X?o->ҭX'A>dR̴{Ưw'k7_橄?aϊ↯w+ 6;bp$ѯ7P+[tK: sEq+յq ^}g }1:rq.H bžc7,48LO 0;0ƓaՏU,˔CWq;ۘ.O7u6e[e\8~GY- QtBaaRM9f0ph?lkϠ=;M `YՓؘ.[N/ Kv>WKVSo[a'/Cl$o)Ot9eoa8t$pH>Rx7h/NRmۏFr2b} +"Ka`Kg4ca Y[qhcc;p^פݰ\CxiOhŒE9~ Q}$5`ʜqp"ORa(ݰ £oU\C}㰢KquP(r8卯, <:Evlu/~Ђ xKӡcǹalbMfhx_Y{1xi8(yvuy_{qQg# wU] m$`-}uW!Z>ۘ6HYb`JyךVì+\Yq{ f/:ljc+}/ Cf?Gzh ܥk.\KPG6/1c}>CN0:c b: ^-PiQ +M``.\Z4@cB7b}v]=~=O [Z%pٰB̆/|IJ[Fmcgwa\=Q͍hYj7rpp j ѻvr"[nN=Zz7HM"aH۱o^^im)0~ ёwfVHbbhpbor}oЬ))boSHmrx}^pn)нP恝o+ƱClߚ3tx=j'%Z31Vi{]+ *!8w>j̚6t/*CP? 7vEu S?q.J\u,DtiBw>"CR+~գ^Znib)fy+y5}Y uOߗ{ eY>Ot1 3{硏u.wU0d{_%qlQ ܯ~jrmw7awF;24/ɚܸMRڍ!|ؿ3,WV'`Kz%toĤ :Piuy+T ^ϡROP=>SO y E 'L/}n i%:O"σYjFp g}>rxd@#1,e\aziۭ9"sq^ '{ ~噕fk?nLܿWqNx3}֠hƮӷGI~m85!ÃWk\d%Sn5wF/=03FNkS+qg;"aNf'ߝvŜ{8N)(41kU8,&-ͿvL[qXyȹJXvd;6vw$v͉Іc0QX[g,t \l.u..КȽk̹ xEs. fNW=:n U&U{- , Măc.Pchͣ2_wCI=ZXg1:,e*tn$u=*0z!ԫhǥzNg27O0&Jq@S`"-'sfA8jwkla'>o՗5;pRw?Im¼ޟ}yŷ<(VU++N5چmnZwDkc=hoqO4gl0հNL JN- a~Se+,R]{SJ?'KqU lιL9.lDfu`}oEsJ޽ [6ç> <TN|j7oIc׵)!TX@UG);ri U6b2UvͅWB u;vfv~l្C)M)y~MYVe++^MXe"}C7nzRwa~D!~|aa'L^4yh}t0"ۨLe{y2܎o6;owLemf ]+l"`7SicS>]F;̨Ђ7uQ`x i8ܫ 8]ńv?S>*%)ô;_)~ƂCe2|iuR,#!NY{cCMgPhz(4aX1Ăm{JpNGtU45q?ϘԜAQ+]<:??{̚Ux+9aه_DaҔ跸>{?8m0O|KI)rUIYX|?{n``G@# C{?>,YXhKJ@}ù32`];Y\Slzv覙xߏUM]uQ8'm8JcjB*KL[%y:xa-#qv}w{?Dw$N6|Ėcoض0 )sx;$%뫷nVٹ<]ἎE mc3={ZqZغ:7#{{faλg4vPsGy/7.Ln;v9m:|vA/ X"Uʂ|7ՖuGkJ6gb^C~?tvt]=qNuCgr?o 8\{m%CldBg4,N<ٵ*7\9k5ȭˢ6l[;5xWr=-ꭸn>..f:ZP A^/D\1\>я0Z-Rk'TMϲknڷ&`rBV 5Smgv x^$v>̽nmqн%C8/}G{ 1a<zbO ~1Lj]^1ߪl"q!!xIg&<%M_D~aV0{?Ȥ:_l'T K? 폢֓,TW̲>$9l^6C8_Go}M|<}< Ű:]m$w?`t;}e?beUK$8Ql[GU&_YGM!;AW!cN>mıxP{ n%#]Z-!BQCDz^wC8x:Om75KF Xes 4g)-R`d%q}/ֳм?%4"&w`[:RnAA=z8sI{o݁G†`i)_(YZsc8)fx$ tR 5,ļow`Vˊ=&J*nǩ},S>{N$ "C)>$em/%+h(t_fR'i?,̅y~x Autn9c0;?/h?"UCs6ˆm[=Єe\ġW㈛čh[%_\&p};"m r7SC; *ۛ?z-wLW؝g)CGP?I~,l%6,J}OBq@`L?,kay.cmQbJ\ꋱ;bǡkY>|uTT;肣$VJ}wplʈKҞӆxإKbsQ'plɐ )X>O\Y>K2pOoK3V' [ª0a K61 ¼&C08U$\hMp0r$d : Sۏ-Tj5[C؝0w?//q_ v|xw=KKy- Bn %vU3B|7r. viBՄi*uͰ .R\M@au9 *XeL,o3uۆr{Sl ?oUaywƐ''65ڌ$!AS*&:Ak"~!fǝ:*V@V"Ȍ haÏ\a!>V+t ҘCkrcYsMs蕎ͽ*dv mUpwcNQ9ïG \+6wTwNX|~90^A;+tv#^=z.Ѽ8) -y>_.yڣ]Ea)e)bG=Xv8`jԱ!f&q)8Eoo2N_|"2DOaK} }21%&Vgwy4Q_c`Xt7au>?>L<幙ΎqX1ir鱱oY>m-$ Dnی3'`_cѧ0>{} N. [¡+]0knd+b~zk ܿz8˚' ʌu=S\< {-]\~5YC8,$q$[1}v:W*tT [@KGӢW%`ەe9y36\qw4Ni[oS+&(4|mC*[.`7KTk$_rN%g쿬Wi Cd`sgDμ41 ˬBaZy4O}fRN|u }T0q.{ ZZRx)Clio4׈v|HLf/`5k^R8,OP^Rv.L;_5DJYn2'#f.3 R .}/ZZZ#UY1w :提4մJPs%9' c?2 4?&M-K9Vz8Lt#$I0DZ+(Kؗi 9mcgxn==V9^jcۭiI)!C'@:W0Q-| <2HF*ΰffͩuCu5Y].4鼳5-()r ;E*AI> \#n/`Iv-b} tj(gEB 7ǵV+$&|ڗJrb!稭]ǡavh(׭0EPu\G@"wmڦyk杖kC?o!b1c:ΖTaKi=BJݼ4t&lf5AkM0v[ܽY% ;C8=[Bv' 5XV?>bǿ:kEjL`c-W|Zc1)Mry vښXx;{O8ah" L]ijQi=eURZPo-Ƹh9w:]`>ǎ/rK38࠲4yqfׇt0/ !w8X %x RM?ZiAz/vK!IjDw3թ仪6TB `TtPrBf#;v5^1?i%¸zJ(A]ZTSfsw)gr{,3N4=40`˹{t6Wkv=hҚP\JaP~.jLc˺Jޱ`a߫zo h_ ^61}m/y:(  % p"ҳ`W'`*?Qѯ+GИv+}8\ʯ+`FdiEJ`Vzwx, R޷_I{4824& NaT3K*>􃊓w`>J4lokN`/zXLx$^\ަgu:zߟU ccJzݭwNŶom5 _\|{d| G׏%8,$]n>X:Xa}Jz\03;!nG׼w9| {i%L˿J4(=#g}9J+HEl` mlKQ4[ G$>CZRH=bc&Hp\n S0n# SX`7 g$ӔV0nG*~B:cul!-aXű} nry7u>'`Hǃ1ZX_M6li8G%]ȍ;>.й 1}8rAaXA²\- Խszp]ues%hdŀrk'!W(;TtREƽ78/׆v"Taq,b%^.q(g3WBvL~ju'd=: XdglvlN5dсڝ/G7|_}s}sgRZ%v݇@ηZ=Xlb;5sĝ^)5 |P 2Qv`T6õ[Ύ`_@~"61ݜORi:04}%4Z0X0ե{p;N>:6`.@ͦh.]I'oqwglԖ1amZjN>I?c ?u)_Ł5Vn`>}:;9ҤIxum&V){ mZ7SBUOScrI]kDCocK%;f2O:Neu$!7ڊ?۾`-6z0|]Lz{I%ѺId¨3p '?0\;˜ Xڝw c} U|o3nyޢy ,w`LzF=x/*Szf }͉eL^:h?Q,]8@rv2RXWJ3/^%6^(9>\>J`!e \wS* go\)5όcCU:Xf>л{} 92 CGG5h+9D2تp|` /8g f`o}x-9砟ag.qT#Y˦?26q(3aRX6}&RqDoSK6=p!i~Rg:Pە0t Tmz8p<2NgUpTzp(u0PEX6.s2qYўsC8c{URa{Oq]afyG$݌M)֧ 0S->h!8V*p>3v<8kZ1VYu-0h/ waR{g_|*ö(.v); Gd>`땂Nƿ+7<,rk̀uSͫVT{&?<&Z\|ϙ;$곌VY@wuOc%\SO^ֵk_zHچΒ4Z~_Ga]d?]sed[騩ATE+糰;]6٨̯a_$ck 9 cXcXLp &VaUK^Yc෺Uql*7Yp寉tA>\ 6w}Պo/ca4qʼ;?yf؂wy'-ߒCN8ز"s,jهwp>Xj/ <9}̔I0hlbM9 Qϧ.@xd&% |÷qJ͕l 9P<S- {֡=f3_n q޿&Kę;@UݨFUʻ+|򝹲 eax* ̯2<_GRo1zKM{aNj0֡wU&:|pNOHYNǡ6eH L1ض([_CtHܦ0~鳧o6S r]^Rw=`2&,ɫ%_oA&7^#]&޼+Ua}Skf'984QVZAvi9n~ 4 7'ezWI#ۜ-c3:0R_f% DK`=Or!ּbWq|gr@1ޟ?E -FJhg"oFъ3~c,~qstj؆-OQR촪ۿzw`)A ۧ>Tru@a#6kQm/3cky](|*XfGe܉ilv}!q%r;͂^9sH'5ћ*Cp1!S5Mت(b4)ױof.[Un_S0)̂;Nؖ>ɂes kWiC_0o63ɒ O#[@r4v˞:Z& g*IG} yA1J̓fe/ 0pI>^J9B 9gO\9mb>41]Qϋ6|ވF8#Jjwh;hУE%po *=yrýgVg{+wNA ߫p=|S ;y ~锏lщs_GNd^_W I:1Z2i>; rX9$G#6 4YOݟbϥǰE>r"`x9StR0|w-1$=YQ٪ Y [kUb7N`'ɃW6cϘ˦Sxa-fFXNM\!Oٻ]98KEFOCR-re^e+$P"Oh}^P\)355Nv1앥)X q?~s g"o^+ku`vw[Vq}#VG^Y3?T 0.nNO((>Z9߹amK5ֶEHٚOO`>]뱆qۓ g=kw TzcL|$=c˃a/7гpzV!4n۱Jri2֝uЌH Ʀ׏N=.E&78nqQJ"+&|zlV+a؀(81Y\g%dijAO2NǻkBy͌Sذ%b4*vK Dy@.}IC|j}"9KR}Y;a&8y{'kO^r_a{aVeK?Ǥ͌yggT:p}+N}lbF}ljZ57gZ؉OeXEc xX c]MHV\4RL#LE%h^ۋE6|zLO+7 U˰cV7wS>Ҭ}*x(':m&3(q17[Lemr t/Qc!iBäMz ~?$eu'd[fUj#CNеzx vx\ 觚Q6[Qi<+ZZpjEiV4iqơʱ17qhdԟF`{W}s wxAWYs4414RZas`v!:VkR@Q̇Ah%9߰vˆ!Ú?B0|xѥ2V l]2R]܄pxJl=ӒۄD66e&݊ᥰzSԻ9Ehڄ9 e|{^X3bi jGK!iJ`ԫ0Dˊ JAv9;{*.!=r|Hr,h`wUfyP5VH*hWʴ^זv&xWm4%4ac[K(,}$o~g+7ɛ_6S6c3sHA{j>K5ۋyg4&8po-һgUqUaoL훱l:>&}dz]r=yþJWנ髖3k4`WA.q>ySWs[N#ǰ,%qc̍sЙt˘Y LJ\?܁=8X. 柉nu9 CY6GP+oVU Qγ{wZB $f8_4`׸-Cs˷bl~(Y/zY8NшnNeźQ/c#ڭJh6~ zeBno*iR NɚNq_.g@90hxV³/L ݨ~uY5BčZTL0c`>Pa{o<'&ψ_+j5ׅ6s|h< ap|8eZv#gN/1Aixl:6foʭYCC]~̄y!,g&5~fs5!QRF1 q!P F7",p싎,\nqX6gg'C%[-+/&F˚]E@Kw#0cKC谐v8{{\jLGoT.ˋPM_%6™Ry5uU% 2?I*^._>FO2;ZtqyPDv=cq{> a4I(W5b aƻzP?zA`(,%lWJsQ|'v(;eǦV7^aC[VK8f>-teaz{PF%2Tz rn/I됈GcFQufIF^z{u$aF~O}OƛOa-u^<ޱr z# f53pFIMs*ޔtTj;K[=ՄD&bvtM yRq_[Mp.rv܆- Et‚rvRQ{'JCzC ) e3])ɧDҖ.]ij\F9ƉHQ%O8$0||_Sbg"؂" sm/ю01qO.l9qxU2y< /<x "Z)t{Noތuw~wj6mamAeKmJ"mk]+`ќv,y) #>bKݥF9 ?10a ۝XؽNcK{υ 0S-缞g?Ǜ&ƠCl6z-v ZP* K*E;O(w>Uߍ y|7K`a"6r7\Ljaa|a:i}_}by.i$ к~Q 0وu\4a]X4|nP;gT}/uE9oM<2܈/88Ӿ@8r-8B%OX*s D4+sj7z(鉶}GSL1,#jZ^w4zQJkfa1'c sDuUvuG3iĖKnbDgk`m lxո';I)%aeeȜV= ڎCg'] ,W)r&w-zǂZʉW1A<2 /9V E =wOxP*kSlk;|k{Obk7ª,>{ϔf \8vuVP$Q| s:L6w_J8`'vJ׵n˚YgҏS;ݓЩ; T\]6\*=\/)C{AP(Lʉz͋v~]w&tI#VXbRj5z[dBeZrp,,?PaϠ]7qxCTm7S=ZB4cZ^X5&O<:%߇}E89(ȳtR4]%Z]a q]C ]Ⱦj \X;ue\5҈S2˘s>_CTD5~”Nۅ>雽aqEs0ڝ180\Rv :5^GV>ǔ3qBة:QݗξwB|6LrGC5%ϼuթlyzG+ܮԐoXokit}T/fgp-~iMqtˤx/xַ/sl<tpA>x /}Nx,̥ugTWsR]9zo~nxˆ{={gQcr2В}>~9t1e2{/biP]&1 (~3,0!$ZRSnPA&;/l^'pp7uDDr'{WS%*{KâMfL-!RϚG _v~zHk:3)rXSֻQ7a~ͧy0h-Xze[)[KO?iykqg$^RA{.֚xuoaZzEv,챯'L^Fλ$~{ϵFxhJٛ8 pp g~|7C̹i+jX7 DPtpؿ6㣋sl~ttmCȗb %*S1@{P*_(B0c)zO։5-Dۺ駰.wD2~ʷԟ?uz,6hW`i(7do7kz7V|[ ԗ'f`eKcλw_@/-W"r .],Μ? SVzDHJ vd&ڦB|xڊ3/@{+fi;cfPE|VkhwDg A2(o ul9zڮ)SmrCbQ3d-X &my髍%"|b#O pER3,Ӏ~jZyW^ki9,Oow ޗe \3XV-ECcԣa~{u088_9:6Whc۴آv[*XiI*-z@L+E {foLڨ~ 6N\Mgxpu<{0n~7KW`$}Ϋyh j;&+ʑZ ; >yGKrRPw%|&uwrcoY┉qě'6oRh.θ&zZ.H?w&*߽u?㷦[t; %stO-i`'`vɁH*VgE כ\0,AI7A1hs53-sWh°КzLCo: *M e=m3wo Y'  s0vzj?{~2cy{3];C,K؆ziaa-Vhr` h| \' {KQ| Zoı>=6bHMa#c_6{ßSzHWkmO.ttnxI˫/@LKD^Rlu|& JSs[s.6y}of{9jjB |2l8Q\b$}`~RyTQA3́qB6ا,Uv*ΊhalO&Ig@"o t>/qd${-o- g~sF1]aIx4R@}~*zh2Ǹ/oKC c CGCK@|ĕ-Б',5tXW.kn+O~*ObTLM=55}:ŗll6_Jl3~fR*9ߋlX| Odj?źu]$d-bU[hՆ/~,bM,s $򱁭o05wL|Juq\F" ܒ36@݅k 4BE,08?r*JXܴqEݐ6yh[n,ZX<y=raJ0,z<JZ"nq1*-NShP^Pzf*7NpQ$xj6'a,:/K[a=WSbe"%_ɩuJZ0ق\5h{g1'YOWb>K9֥l2>?q=EˬeYrкJ^,/-OΜ1!7_' ټ?@5ga1oU|3g>ܝX#=j}Ȁ_(.̚nfsOZC7ʺT"Cfbb ^ૐ툖(d\eu T2 ̢_at`bV|,8'[{µF)_ ?@ǤI,L`WfiH ȉRXW]3ϟP4կ"ܺq4bWPG9vj^0~w>7aT RzK`ɝ؟ӗbj;d*x_|z/v) A뿽gjYcks}# $gcenlҵe_#/(oNx&*+{3 ; :7k Qϥ @dpy`9Ia>ajg]]ޅ\?KHfSֳXB3}Kȭ-YWtx|Q!/ l_h̻!Jyf~'^xC(WA,ܸ nt{}Cf20p_`o]xx*R3K`͙}aoQ`~{0@o}+-]4sd3g:^^Q[1SU?Xvy"z[roVQaL7|кs5uw}Q?e[S#܂?$y '6 XQy"<t|o\jZ*c[c~ػN:t* q*h˾P}]NdH~/.?KP\̓+DcWCWTHUhirt6kRᑫ0f$U^-|kbܷ;dChA_/ϱ_|U<+gJ`/@>rggcOϕ˖YU(a'=(=?:KM^e%OBԇ'!ߊveX&f>Kk]];{= 2BUkdsYxjq3$/]wקyfe\, fo:m =Z|<p Q.HeĔ;ա"^x I>&c8]"N*G*G^OAow]0pЧW0;,|VCX(I|_$ۉe[̃[rԄ|B䮏Wf^C[ Q(۾E_S~_ƗK=\=?/M \TȻ]8(o5 8?` :_WUoK<[7M_P%,4pJEC :ERڪ]YgPt~PӤ@!]1vC6WL%okX1ۉn<jg,SiP6536-(t9'۵l8I{%?SފͫWBχڣi ȭ ٧}nL@ PfAY)DnW`Hg쒯Kc{7l9y,Qk<‰G3Eu* A٣w=}=w_lB[^[8.1t2j&YTQEl(=Kտ:CyOCf)];n+p?neAc=E_ڔvi nj?!Ҩ^6,TmPz(߆-,QFN]aAZܢ9Ɲ_K^NC)6o?gC}XY+\)_v =胿ZvƲ¿}e^'p_Zb߉BEV*ZJ,ҷa63m K p 8O.^G{X]ja :Ưښ6B\.(:7 .~Nfk7RxzwObV&GraU/|48Δ&a:X%>u@˒3v]QQsy"YR߿9S$)c 1-@3{qJG6׊PAe%e~ɵD@ɌUϲ1SCPakY7bem ENVKݵ+lX?{lxkV:R隌˷PgoNJd ,ZvI%iz`upfHbP' ߽ $sZ{C6-Xh`߉}G|N_}|WaJlOhRr+9JU8צb_t?[|)Ȩ_8]4׊8 S= ~tqiZ3f᤯Vo+4'ZEb<;4yO\]4>%ۗ ޭR YkLvbOVL:hܺL ߃t,%pNDŽ-<'&` }=~h:u<9N/ӾJ27T*[Xׯ1MRdOMe.(vwǒwjFۘ۹8H5SJnjv(gvZj]~~;ː>eq2j—fauXNέ u-y/oc_R]6r"Kjpwis /)yVBz4I?eG-{wُO"Z 1hM>7_M 6ƺTA'h| )l$gm;U{sYo#8z_8 $gQ]J? ʯ6Z"+Ҡ)HX׷9)%1.=!5S1ux@Ã'r6ACuQN,ygy90y:H-[7x b 4˼v-ԝ8b }YW\Z [ǏUܞ^xm-Gzalܧ j0La1͚joȫwXRЁsо-B&IT~nW?0$}T;p+0 m?Sa=ί-?ܵA ?a}LykCbK<#ﯫviXb9~ 3. 3OI~1Egd*r+^ziaN]{}PgjB%UQż ̀!heE͓ωf[T}r[6dxa?vz[?05|ܰ.l/uEC~EHzSlS [t-go(E^߇_´/D„ϩ;;6ơM}pmz p_l~O &N]cTbޠ0s b…Z͐*)vy9UEba3)}!l\s 5ƛן!hbblUShKӖ_M5#ǡ,tUcȒ`J]xReoyK8SЧhqfmlnMyb[R2~)_y)EYc֩gcw!`rUs_V0%7ef* vNcWC" 2ֺu0mK:mBb|}-8qc|hz-Ȝdښo$7}L"֐k`s11}&~)b6v 9}k~~3mߴ`5w N# X=1 n":uFKYBaJ3˟Z;Ut_ wJ^~vJkStD _KAj[ک^=8 '_kB)傗"mcx5%ucC=)"})*'Ŀ{Rfb[reoVȳ-`l:;h Y1VhѴbM$~?WdY k|f8D!fXm?Od=mڿ#XF7֩=W~ݩ*PNtfIDT\8Ar3aO3K=Hڽ>1Ȩ'ZrW/ el8edtL|<_aPySؐ+^J^5|hL)mz}p]SjgWZȡ$;wOG }Ú;?!;~' Vvzz[YT,z?<쨖[ڕW|CM=כ&#wwv-b Wcqk=VnW=xMj9φEqcEs,1O7:mDU1w.v1Tٝض TK??sl<dNO8e%Ze<лrt[ e Ҽ & ,y>U;C&m\Ba\9-(J09ZvM ԣ󇱩/Q<윇ת5a97|pEa/euL,pZbmpxE8l}J V8ݖ-.kl+S!)k]iwΔ,R@Nt3LƖAF3`гk.Qge1~aZ̀kUأunt(LXރM0"~},7ӣV*o`wxz>]MM~,}+sjI{ >VXzM M|ˆ"6U0_I "ږmQ%Mf)y' ˭ޏ۾fvSػwV6׎Vv-[ koˣlWx%-E]QÎS8EAl )ਮj`mewpO[7 zG+Pqrh]Y4lCaT̻U| i; s?qh'e9[b%\s_e\).tTYoM5NC굦'u۵Ñ-N_]"bUV(:D*%Ԕw¸KVEC-ԋEzlG:͡;%w%uU|WFwҊccgd%pͲtNr-؎}[+6&+ܡ*k7dmq[}itL}SOya[g%Visz Y|:ͦF""Ca+N?g+9'!'wa2{H_W9W'L\P >+}Kc^Zl # CwV.qŗԷQ,6=Gi><7'KcӌIvЙt^p UaLyۊ|9].#Yy3tۃc /6ܸ[0fc$_9=B B=/a{LF2>|hY}3HW]::N:q-Xdl"}]Aߦ\R!Vbpud h^7c{оDyO'-ˀLQ.KmtKe .8Y&4 8[7/>۳ (S_oUBOTa/f#iGIлhNX$%>}^8lcu==VG +7HXX<9ې=R9eI=O0m20/v? fПJ1|4$PǬγB_ΟA 2U}dX uyOwȇ_aC8d߈=P~3r64|xc}<3Q $ yV,>܅cǻ7/Ɓs pT-T͉-P3K<F6˰ 9W*};[ss8Gga[&"KҰc ~/.B0j]),Jݓ,AY߶9Y7^x3T>e 4N(ۂ<k7 -cy0K_ FU<{Fg=s[$anXarVPf56ZhMGTFnW/e+y/l!(=v2\RלM8&Lh XO>r .G-8mlB^?;IN{f.r|c ,H?/8= 5;f#RvY]nh9u-0Fh)3v8#E|1P)DF8St q^'Fp]%SM0T2VZ|j5>f3 ok,?S7^kiѵY]e~qSY{~ਬVmO-7@'ew/~_'}1Ju'pLDnWG-iJ{vY>W#J~|0v{'vR[hEgbDYj S{nҎ8rs[7pYzfMp{g#BEdw.f>$VۂL;́_sB.{Ͱĩ-_ͱ[YmأЛ}{P~>$q2 I5yzpq.5TKa,}}/TDհ$~b^ Sn5װ%L;uͯ~Ճ'`٦ڤ"'7?c{CϮ^՚xHxエ%lqC﫱"捻ܺ@6=G_@ QyJ" G ?e.|:/m|o͇]эXtyq,L}k@[ĒLhqwИwG13-y{L?q{,ix펝-%Օ%D2a1y][|Vgvfƭ rJxcRy,ح%f|o$,Z]8&:_NH81}ͮ!8(Q|Mz׬4p.I:vȔ+<9G?a]n)71][NP~>Ẳ(vKj ] =TM?_/h|Z F ,!; C_淚'r*Y|mf "VGMOۉ֭t& oXhSU* Nck1a'G_/{!p9ᡊml:3Wt] jBg@i\}~J_ ^FJzEIEj&a4D;n CMbfb cϜ}ksCzb1?+0P3MWA*Ks;:_k*c5[ f"+L~nőkww`]bZly//`Oǹcpt᧬clɧ6laGjv ((I'@]: _/ٗ-I䖰6_%0-yFgeGvρ8V5[ga]&2S2%su e6o]ep~٦@UyLapXУ-e+ Շc_oɔS1f6:%#aL辐!ocK|kk 1T6ʳqQ:mزT䴋 j }fCvy*Xv<첉)96 γRv;Axz}RV}h+`}fM<4~˦v]O_11D•<^z䉑eC-U'G,.1:}(XK PN=|=7W BǷesL;#r8JTۦC.2H7y:zλr?V@sꙂ0\u3ܹ{-iݟ|gl.P`+ib|~3{;sBقf6'LΰE"fx%*,]~آri런t ̳`KEWcO(ǒH߹F#%֊,?C7-Ys~+--fNl%^cwS=-}KdwB)/,5 &X?l=%j}-N)LlT3``-KpWt]|qSD &D/֚gV7yrOl$ PuQKyhu0u {t{aM *d6C=v{#4ɺ}2˾`sc[}qӳϡT 뻷hF1X#)iz K{b٫* N=(f(8ow[̀& ~19,k&_35 &MW,OqǭN_({1` J%ǚ:R&5wx?^;unӝurg$# '$''>N]fMmF[}ZJ`24ا Lwݣm?B.ZvMo։/wןpy7+( oPoA@M7& o&o&ob&ukPx A5(נ^kPx A5נT^kPy *A5נ4^kx A5hנt^ky :A5נt^k0x 5` ^k0x 5`L^k0y &5:k꼆:k꼆:khkhkh5TUɨFF dNFdT'#QHBjR(F!5 QHBjTRԨF%5*QIJjTRhF#5HFj4RhF'5:INjtRF'5:1HAj Rc 51HAjLRcԘ$5&1IIjLRc:::::iiiiK( B,K( B,K( B,K( B,K( B,K( B,K( B,K( B,K( B,K( B,K( B,K( B,K( B,K( B,K( B,K( B,K( B,K( B,K( B,K( B,K( B,K( B,K( B,K( B,K( B,K*J,K*J,K*J,K*J,K*J,K*J,K*J,K*J,K*J,K*J,K*J,K*J,K*J,K*J,K*J,K*J,K*J,K*J,K*J,K*J,K*J,K*J,K*J,K*J,K*J,KhF,KhF,KhF,KhF,KhF,KhF,KhF,KhF,KhF,KhF,KhF,KhF,KhF,KhF,KhF,KhF,KhF,KhF,KhF,KhF,KhF,KhF,KhF,KhF,KhF,K:N,K:N,K:N,K:N,K:N,K:N,K:N,K:N,K:N,K:N,K:N,K:N,K:N,K:N,K:N,K:N,K:N,K:N,K:N,K:N,K:N,K:N,K:N,K:N,K:N,KA,aKA,aKA,aKA,aKA,aKA,aKA,aKA,aKA,aKA,aKA,aKA,aKA,aKA,aKA,aKA,aKA,aKA,aKA,aKA,aKA,aKA,aKA,aKA,aKA,aK&I,aK&I,aK&I,aK&I,aK&I,aK&I,aK&I,aK&I,aK&I,aK&I,aK&I,aK&I,aK&9m |u)multilevel/data/univbct.rda0000644000176200001440000002554012750500645015521 0ustar liggesusers7zXZi"6!X*&+$])TW"nRʟXgqjn]LW,n] +Uac3& Jc!o+U,Gf=zWE0m,Ht@a\v(}N6k<輹j56Eb=et'jz^$2ϑogĶu `&L%s `#vȾ}5%&&28}F:µ@ArE>,)=`F0-X=;vpP^N*/~P~ȠQjPN^^h{z79󴣁;e?Eë[본g76Z2jb>1cs:: Ǻ&ިgbObԖ^ z{2/=ɋmxẻzb3^Y"WjrQ0 }rDa(2;7kwnz] |qw^l fB-.#8"z +aTGxg-P*K B*{p˻;Z}HuhYa%AQğ%Ab3Y9 uW)H#'+o _:;9SFo&X~߆B0Z`&9Zb4<}M= 9~5Q渦a""dh\Qm큋DT.!aǭ['3zuȁ6 Ү&SEIMqŶLo1g` h%cuD_| *dL'X0_x2?CD!C'6Úh?2pnWޙ~e [vUc ^gЈQA8-)K}wTSLZih-̦@&B(Ê2M r?TvHd_h+^O GksS\-glbS/?n`V,- ߖw­yyln[F,ZS+oךpr84GX|R@.tF?i ı᯾G?VFWi;csx[-X4T&&.U.&Y`چ/eƒ9["(dJ=ZIh;ޢRi_PWx 3rt F-"(.A5tY{ (Ax*$Lb|d1^>,= ;~q5դXwߠI_ǬDʒx#ѭb6dąO ۆ˅l3K1RlaWc*!YWJ>RՀB]s{*%XRd&%HalIGFQ-vؗO[2+iE6/7ƺĎ=dX\CXPjjf+r+%@A Y@$~܌ \`'=L^&HWgfɎH:/.gDpMK7m,Tz8$#QWW:q,(X4ynhN:o%0hnٰSQnHyLJJiN,~LV+šFM1h1kYR]$gT!D 艟>cQ g$wrS/> S2Zz7 ?WyjVY2*`e7} 8 ö[\b-ئ.LNkf|i# F0}6{?Ru R̔{`lYӵx`-3@ D8suS ax-%eL*V:7)yʣW\w_†-J+ZsWTeaǢ3[f0DֲiVdgHCF`%g43?A`iˆLHdu*g>+v(DA _> zVo*7>2K9ܴGBË ;>ŗR_p sj;pV:a1bUr$j^>/vj@y ?_ H4ю`<06`q3 ^Y^{exJfA9٣wqi<`z9*slЃwe_ۭugVC  `β2o!7w$ Pa) |\ P Vm+nY79U~-0smRR-8_k'Bzџ'%qq%Ia &+vwb.cGɂQc\ݨЯCj:d9U`}:#M6Q 2%.[3,xwg01sVfwұu$ZVHPܼ:>{eȧޡ~dxuqv663e$>PUsҕo O_~,c8˻c"{[qrÇ̀J۟Ye\ _g j Im U 8lU88c_m!;=LzrQ%}TKr"RB2pg[ik=}>t߼O!=k!t`x:f{Tw51\KSzu{ Lzˡ=}z “;%XɉuSVqX,!v5 gb܌s@k7@qJ|}Ds(p&qf8;M:$1nHEb Ș>U#~U(1Q˧z S8KV&6I92j(N9f=f(3r8lW&3ܞ``K\g#Rxm&T#É2?H|9& a[.ʏ# tL@ [ՑaE`[+$LV81XivNRӪ"A6r_Gyח6[8ж'Ss[_N3߳OvjrQ:v[E|[icVۙ#IT<<_LSk$2>?ǚP jw~ͪO,X8׾淺ڊI]3gÊ *T䰸"!<^ tB_EwfBp^u7A@@NtM(8 :}vp@>@$ 2:&p}QB9$+אhn)QԆX}ǂ;+BEu^4PjDT٠p5H\78."կBZ>gʼnQfml}F{#heZZrH4c:t@4J (kS$Z!GEu&*$^ק3 ?vQ4rݴH]lY Mk`w.OWK% Q@(SjDu8EoY9♅$T;';:5*p~g-q4c>Pw:g!'{>Ħ! D0v|hsTx /peeT_XRsoMZQӠ ēw&-O| #r~#Kp39?255)^xd˦%-H v6^UTfIx2H2[#ZZu%e7VƄ%.XPrۑ21}?k]ZvEʅa5@UV"\=os^m_;2v@ȴlZݓq)QrTha՛WҏbY#F1s{E)gD*חrM#DKUfs '@$?ZJtbX(pNmV 1_OP}l'r^!],؇sfFwgL>J`\|$+M.04 [AO }O2SH"3 _awmhܜ6*LtX؉!4μ@}'"0Y$Plov7Kg_'=L'Y0칩Jh4?fK?YNէY()JPsYn5w9ͬŗgKܒ5 m.H pZ$i nPŞf[ yDðD)FOHuw*}y}nnlP}I%PFR>kK\psp/!վ"GQ(q{?{]$y93N$~ⅷbGkEGN[5a6,~ ݻ1SV7f/6G&QnLk"xla7/vpz5EмeP*SA},T5t+y0WS'o*젚J'L<} IqSPuAt 8D/Q:6T]6?Pic^VwЀbk R ?Ub@ݟps#w';DKSW1=ߔߋz6|(]ypZQ,CntR*e~joșdiPύ&U:GLKXzrVVbj,iTk˕uhQO䆳wOfFPy7l;iQ剤ٮ ]]?Ψb'Fuzt(V6Ғ%mo%no 5 ¼b$_NU ~*W;yvyʈEHNWUg@u<@GD( Ϥc(,F+DŽwz[N5߫t:$ oL2+YDeT1(l;r<&fq01GO-nJ1f[le0ϑ~!"C\hsѐ]]#,4T:$J4-^>29yGlι;ہxMCO<'\+C2~;>pbcoi,Hzcܫ@NR>.G l{\˸> s:-R˧z^ n þB[ ]}FuA'σ2!ħY9]Dd8=+죵LEc? Oi=#{=-KejLzC vgF!7rT!(8tIߣnѭmV15P,M"պ`¼hgt:ãX8>2L>Р.sҭ=#5UgxuޑTƒ?;,ϯP5Чn*aӬI@Q E`Wɪ+azg}12&WH#k"q~]-ȶV8q>9B-Df2 AhW)HO<]I1{,ր*lV'Bl X=Hf@s}qAŰPNd؀v墵6QL>qb1ZP0O%Uozv@W}%g]UTrvIs6D^kK?l&Q:0ּ{~X O]_$!7|^L ƫ,ED J H6H_ݠd/Z*cwm >@W踚*erV_õsɇ7h.;5z-t,<e7db~Ocb,ƹO~)Hk4:XG qJT~NAHk SՖ+!ΐ!YhT:a̧ˑϬWr&3C/Vu~nQAK6ҳ#tO,b)MsOA5P,FVݚkoAjlbJo6l%H"quIpdzZkLHxui0|gpٍ pYgUu@/)$1>"p̴볅~oص1Mo75"fL4;hۿDP7GՋO}#Es;&["ݬmLSΈ6A\z0|7VQ# 𑤪:RK+_j|Du ]/f_*S9Aqʱ<˪dMrX R}[}8VΤV] 2m5ou@ʗG) |%͎~$eKDI B t|P}qo ߯bPeE%ThzBu*Oa'.tY$qМ{^y{_OO'؋ugG:u37 kTJziClT̡G T"j(MHWGBR^)My(xX//M"X] xtE2N:`0yREL9.A("&k܈__L{mDyO$NQVfԃм5[|=bo>0ojLd6?],{Oy5}L._="R䤱ޱ3\pR:SgX-"&\&͔֓6qae պ<=.Wض9(4AR)i\@JXaKx6?^ɺb s"B2ZpvY0A6p4olTOJN6ፘgH(U "c\΃+hFv MQjBh*!pVӑwXz@~ȕ +VpH#NǴToF_ J}~!r>Gw&,KFQ)clht՗b W0dC6Ã3z]#%S(1e n,*k9[6wc^ L.K"7w'C7N`a5!1)r |Y%6=eKkb# ^}0&ީS9#(w؊B!DvRmp-?P| ]*ol])&'v DRr0ֵ+evC>g__^0%KM>:I.ˡ>hg#4#8r5ѭ;0IRWdmN "uot| ||9OzHZ5~QE0Qj.b;ozso<]d~n$ 2_ߑZ%$W:^[ѡ; >G~s N|r|-mu y`e=3:rA$+#jeqZDńeQ*3H]V):{^it>pIOb'M ;M9@dI6an))Ȋ"Tmo^~$vi~4vko0 )K'СGŤA<^ګ2FB_?yC\L(Ut3*ÕQdM,`GbFW3o7xKx-0ibyqE0)M?!QN|ڪڄ.wz iāeyWHd?So#{w>4 xDQAa^AHkB.C$8|B>Wv4oJŤM1g.U➚RG&޸T-^j(0&3Q9/ZʋpIju&K = 5=X0zK<7sy3D3 nF}ejAbymVG0Szi(@!PSP(X$_Ch.?p`iyplfvO`2@2"ݰ/:RdI#2^}8{@w9FP[xDH.01\0t( !<͡O$3+X4Po_a݁CM=ŁŞf|Qqrq*VFfo@/ r7 ɳ?4=r.OT!NY#d[I޲^4D_#?'l9)=@V#V >0 YZmultilevel/data/bhr2000.rda0000644000176200001440000006647412750500643015135 0ustar liggesusers7zXZi"6!X5 l])TW"nRʟXgqjn]LW$dĦb}Q=n2Z9AeR]5;ԗ~a\횴O׵ynPxoPz)CAFOD'ľ^mN0Dq!}f_Q@&5(I6Ph'eE.G~3^:J|>-lFhyxS(F0Mr$r6tzIRdG"QEhږ2C~0[„RX`:Mdf@ڃǙk(g> i9e>ĽFX;P(5LB80sH? ڼWq-^Xƌ1+38眰Ҭ>.r#S$Y./EyOw:S\7f3;0U20ml9+n2cps4桉&{KbFp`uA4mVRRog{JݡHѡ8:xxnDU %E*@rFbP uɞpl؊ ՝zH1U$dۀ׮;Db4 qi/lȪAd&"n}QyaTVl !lR}J{u_FTѕyrYs,mxƞ& @FLm˹/{?PPLt- BwJ#ӳn@wnDԍ*i7"<:&tEMImspdZr^!~Z |Zs/S;wfwI({6vn"NT(㪼2ScwͮFjM0}* lQ.C&+n(@* .d$HV-Ɏy; 4h(c&]q_'PKk^[冊? ݽ@Uj%dY3 Q1N&;iKfN=816%7yC͙N>Kט*\۶~\axȃ% 嬎Y@bjAo(/}iN?(=d<MB%zaKSp"H9xy&XDZ89z{GK6]"V߿ن~ԜO^y[sI#٨VpՔI)DΨ}_NEHkөvxB3+N`֜(%GS1|yI M{~1m8wb]3YxN,*aK]oΎHi3eXOӔ (? lV΃:廮 j<י02E;%[?/ᄑ@;`F^ݒ!Cz[c[Sk$7WTO٦=[a4ߍūp{mO+"q"x"fJ]3!jtBԌXL|U 0_f [b׊={4{opJXت+@4#"pp=>&7S-tA{57SiP1E3sHg<2:ȶu.1!#_uqhg'c =f&xv)C- HrCp32bqg!DT,kA]I*W6$7Ab@ٯZSК!f{5P#'}&>^& s5 #1Tk$[v9mc zGsW^6C@IQH/ ` \?u5{( Td\1͚U =/k?ՔM٫uէR1O j k/MUOw\ptaJ=!CÑ4jjՖXk])t%K>SSww8k]*J'F?x-6+h d..PѣˌJ+Si$1D9džq)&PQZmߐu"3q\l30Jd+%Ph-Qk0۽O]EGz4Sxsz3AҳPܾŜːg}IQdզP;Kq9zmPL%i,1Uxḣe F5>b)ɏ,a֡Wo|Nê*[UASkZ W1tG3"ziIZ81&"!kC"1H1#I] -]yzR/M)gGF.wr(m28(yFzfF,$ oHYbT|fG\"NKRۡ@9 'y Lry/O>skbdW3 Tz!vQ:_ֻڹS*8tو`Vʼ>jۍfk)L;PF˹h0jy J5_r 90f荱~8 hM#ܷhbo~mmއZϽ(Q%hSaNZء_ :|hCM&PcV-U[ƎY2]c p69;C%zA޳umϝha]{FԋsͩBL]M6BԬie2 s]P!^7 Xk)wD/pp/?%4muY@?UGwR*P=nT*`9OG7RشMrcYۃ=GΩ<,J_X*͜/tOԔ2t) \Seڶ֥"8i2(nbZs=!$lIU!jN7-??#b6X_oܱrG3瘿/?V*BAG@`+V>4# D;H|P|;k͝zF{Zho){ :4Ǣuoos@_B7Va 0T?"s2Ōg5}s/d-)GU,3u G_MZ+qw*ֲ>g´Yjh*awgzHKø"w R.a#\ *T~wrl:T=~Q2FuRh-WmV,yʳBB]kFV0I0,=E'k,%UȂb.O&kkj-M{`z@AN96ӽ.2ij[Ӈ c~JBrQ+ҭUs#%hej}ǕfGx5did+۱]2.|UK+7IӺb+ k! LSࠎKɖ\gZuC:B8YGͤmTBg~D3_J!=iD{s !?PiW&mɾ oҒ+~WFJLHE5c>L3ޟVI]*waԏB>Bfb?b60ٔ\a-Ԉ8ijő5ZDRxȭusFa ;(Ewf?VU|g ܪ>VaIo.^<+)7xWGbkχp[eG"R YLU~ W X gźN"Eiֈ%j&]C4nz[}Y%?-Os4|e#a&ڣ fz jXGgq9N;/IÊeIΩB}w&4SO/\kG"k*{8'Ċ.j󾏩wMhqy0A4xT?˲;FϏYkѮa%cmL$`Ҷ{O.E_@Rp?NK2@7'-=Cѿ>ƶ~7Ρ25\R,f67_C$BSyW`M<|*JwrSqZLc֏9~`k QLGer_k$JTXhM3-F{1l,s CiRk }BǗkae|Xq!G |ACr|(3,$XKI}[r뵉WWc®)gnzbؼٰbК+t;1p .-|/Sjjs"C<(#ʻZz)/ǸQRf Ҝ\LK]*\`߃@lD0oъF4vtn- +?% wTkՠ33N{wN@VK-*8n]HM""`YN7ݺ}hR@]5$EEů10ZX{ӎG t6AʲA|4|] ‡m2]ehxxb,q? iMp  sݘƌN[$Go+6@T ܳ-OMw>W\yc,/9+$!'l H )0A #wT;ݡy3,GQKEQo9(vfZnc֟v޺%1A8t52;¬u:m 9kb)[O*bwvx %$D3uw@~=6cs:4*עIar>AwXG6΍:G(Op: nlcj?̪ M m^cX0j0rRHv%K;lWZb/6B+`wT[>,7rnԌ3Un|{Y! /jw4O ?pؘs]GvHH} :e vDy~:#[1N4=QntA. !FJ4ҲnOV_9;A+z/Yyuy(?^kW7:@χW&V0em\pقwHy\CΔp~!y$]ŌQ'[/L&EA'ͯHfSl1M&ѾXE\!}+J&FMƁF}-@4?1JאF/boY;*O: es QqQ]׳k_k< COɒX/q# b9UYE~_iv~LH@obh.Y\/!J}ptorJL158ȮDŽ5 芘?[/ZOsduuaŨ{I(<5zd8#5=V&[<<#3 l\r3z+^z궻Ӫ*nϪ2i8/\1zV(Qy|(fSNx*뻠Xc1 2ZĴK};8L`%B_|~bI LOr)ݢ4U(ہM[ 0|} "٢%O)cl0+{ђAïR52֘);.<| V{THPBE`x%X)go~.>7KzTĬd *)[t2s) ?Ϛ \P_ͭV ' NQ_vAJ+Gw5%^\]lY`4b& Ihk N變u%04quÚ [+ '\VAZT:a@X,ZeaJ[G'M.r~` tF/+yWzmHjF˷_DC=0[`¶:<:`Od#&W<^}GX\ %jչ>T,))`L.}gMy6| ҏhN\+k318ߋXx}En?S{8W.JZ@!Mw]B}_& V|-H{'(1`H닍"@. uK`nИ <780VC~NNLFle9!NJdؐ=E@e O"helٹ坙*EWW_'7|;k8Dkl{&o~Hx ?}.\@jU^ڨ{`i՚cB&p?VNX>G%iҡJ+DsFLCՍOBcg?U6yZ\h|W.902+BDhyж95a/S_h>dR b{ z{u|}H 8edk wio* |UDpXq3`UӇstBd UDY2H &ߚ-2EdMh>+g6eI?Y7_Z`_KƝr%݅D";PeS"n:F~Cf=0zco =]X*̑;qhݤb `{jQڒ = Z<*hb.TfL<ҹQF҃8/n j^}tcXk7(0vtH iϣWNy9n"NhLYɎ8^k GAWc wbbӮ]AP }]@r G 0_ڜ$p*v;-ev/ݯxpJ'S聆ʬvA}Y3_ \ͫӞTkz< 8tSEH$3"9 <Y#Uu P'H{rv.dVi[dKDxO f@.#eJO^Ij*#ĨI.G7ϩ5ׂk`,$#`zɵ&֜X. Zˉ9Bd[RCD<[\N|L9aNizg'R o[UeR>^-$m^;5MsWS˪옴;U#y׻ A`kGR>Hȍi1.%)27w GSۏ#Š>y }ZΫLla9w gQIQHv#\[/T̮L} Sۨ8jO0^iܟ! sEx ɓxLqwwzZyjLH5WZn"YfC5*/0\%^<[Q:w)(bF59ƒ6 4dZlHJIhh!T+\2v*0LoLctY][wlo4 hhfU3W/A2 Ky+Ȃi  >YWvKtf iG/8_Bѵ~S_fͩۿ୉}WHՒ5nQ63{ +G Bӗ3~i2|Фw oF#4Y 7r&ȸ+>L{?YOϱ41$Ĩpsgo4u5;փÆ < (Ԛx>^7I Vj1=7=0qIRꄟ@D;gY;%LFs5'rW OYQށ?D k `/UZ*x(iӾga i] vKC)ky6hbwmr l@]f5ڔ}Lm;7i;+M6g1ɲ˙ɸzڋE4( B[TBj]֤ݮÁ׈ىPk&fRmQ2ހ}ff& 5B(R& OMޘ 7XHG0Z.8TB<=T=TzZy< cӐJ&5 V (!*4?3cZH;h~6To!kL rw8@j_)Rx$yTCB}vy~'H0V1KDIû'ޯ[f3<MC_ݥ| oB4=km​PQ0Z?w ?_$ב΀9?a2%ΩOS6ޱi=nfa_fPO]aZ +E8=:037Q_!Qec2[L `b83Q&n7OfL_H_-Ybucc0[a6(D-;յN _'*_'Tb|ՙєTkq# 5[!rT`ofxY P Eڸ?C TOsvD:a ZBd2E pz6 xs+A+fb7sTf<3,9EJxkFL!F}YR۲@Cgd7W20Ut9jroEZ'C~!"zfMdS9M>yAS/y.tQK|CT"iL}3ׂ/?ʷxBf9|Ėb〣/F bP .Y ':tw⢪4ؿ>wFʜ"ɁQڼHU‰>hȺF뒔kt-(z(Wt }tiA 6Uj/QƼ&J婈81ҙ<H){E5 uKxt4qwIuVNfPufԽ9Iܕ.2i@p?kb&~W,.jꤛ)\ >D_/ *u%,Q/<Nk9eg2%>3/M1|k +^rN?2!­"KΑV+KX[- J 2n@->"x z߱ňt%qb@(Q&5۳8U% 8_!וK?*t:L}EO Vp[+<%鐎X9(L8H*|)ȣC0vɅ~ qjUԓ,I .: kkZ7^@S?b j@WGuC cqe\>h׻D6Z[7D^ӐĿ9׹)/K&;TҺa%E(??{ 9(U2%kr&"ib;+f$s+?D{cz@~BRX{ swHϬ+^ZjH9Ŧe7k-"mf_[y+M]3_G`\(WC_l cA,ދh:%ZwO`C**D3kHaɳG*GT _+4(3vPHXԗEJ*n`a]-9ׅ1,\CRۆ ܖt>g@bzAϬgϝ5躇Nbrp$&-IXZcJ9:^6ʰ:ioL㛕iEw*;kiuTk Zf=ݼ| G3J.0#M|1nGtrԫGni\ eIZ 0^;P2^gfz2Lwk:{,;%4yzt)C+\DTߤcA')TUvQǣPY6L>Ime=xDʦ:1kB>ΒezS5% wLZ .-MÎx`k` 5`.yA8:ʕdF @|GHP!vPvM2D]'hIRڭx`j[ Y&e鹝/4 #ze,.P`dHDl}&J 4bH!;5/,ܕ*K"ײc_ynl䣾4ОВ`JVퟓ4$nU[3yٗݩ \kaOqCcܵDp@ >Wzi`*,Ԁo~@Ɵ 4MoD͇&9M\sPўɹmz~zE dij {WTS/ :qENBw9MThCg 8/~3jӛإB5E|t (!jtg&jY<,sI CY-Z<IloNz^b&PbL Vǿ=;T^ ' v3^ryG-Q䧘#C{\R }?s)1N&wٔK#!"#~]U*c90~ǪF<9wS~[4|N澸n!w6ʔ?{Q̱ lPBiˆ d=*k 6Ma%$M N=%8C@*bxڥr|nMBj;G4sk24aڹ`fr?ncgyۼă((qz{3 H(C]t{R"xYjha>Ljqy-Dڛc6C<שsPiu;/a'a \{W?]#-'٤toS=ϿaL :.?{Yj \o7TA%w`e4sDVʗ+~ɑ='q^]|͐-yFBk8G~{)Xc yOJ}ɗ,cld; 8&T r8)IuuTuw#gJ<u>#v'[svf]ۃDx \qPAG%0r|"k0C͏a-1 <3;k 1ڍksp1{ $˵F}}''ӀGj+[*[xEȝ$xRc:Wxh5k1>bZP/QQ93D2q|K㡠 $Bܚ_Yg*M^LBmd\N%!]P!ugv= kxY3RO V[yNֱ*=#94v&u'Ր&'OwLAq-"Sl}"oD|O3_jra&0OiI$9ۖ51a!jNThM_N .IHnA鶈[65,޺RO 2Arab98@;::S ]}p!l,ZˑOjA@a7Y~fqe ?C ?ñⓍ~#ɴsJf8V)mg N Y;'TY[bEd4#P]eZIĂ{!>5]iLKQ`AA琁j--aА">;y$61l-`Pk :S7=M`.1Ʊ)pg`"oo-,#ʺGR˸OPԖ|1N*҃L)ؤH6\dѐeG/i{Dʻ-P^ Qaff2_:9R2Ɵگb<)V' IJ{k]lߥHG8eϠC}0s]dL,<l.vilf_f?c lvB?}=PS&x$$ҋLm<:V ~ȽDû!X(+݉af OTkUe 阨@+2^C4k2r 6sC4x"3aY09XVt.zdu\qG|ƀ~H?3ל+5X=hBgC=4wл7hMWN?8r;_"U<(n2͙u zC#nDz[XagIUU}Z޶lQ^#v=.Gzj!Ó34f`pʈww%KDqoϊS~aKĔ(PtYz^د+ke?J}T" Wt1c4븞5xl>ףt&m{a5t Pq!a9Rٞ"oWV ynBX]9`6MN=+v;%41JDvrT+*Xcmy> Ovu(SIV] J >VjB\qMPvT#Zgp)# iُ s$+-9#!s(CIpde_~^2Ngk`RH衫)G:N|ytwX&>Bv>3"iIA4rZ[3RA3w{!:1ԩj^3t_wz\:Ae 먉Kjɉ萵KLJEiY  HaƍGC')2\3cO 5+$7TqZ+QQQa%n+ͷ! Z[RZĿ @PuNK.vXU}> B+:5 AFtD>/ʮ1iYj4^'E;d #ZoC">a8a!ma`Ql"*8Qײ!!qyfmpgIp:h9>U5tG߯ш:9쾠\\ )F?9:~U !Fst͗$E֥:ޅ8z%ɞ ĨDЭ##ۀV}pR[` S\TPkU-#CR2<%+!*E{7ր?&UG8sDYY#"o-Y32] iЕF_urRv7U&AYPЗʣQM"Y]lo4n\r d<%sBncDVx[z] (F(Zm5U7zdh$Nk'؁ә@*c$<gKD!O.`+Trj`>zug ,'$Cw3Oc مM}YiEO3=3 WKB&=)?g$ (.UKvCСV%l|N> $eRJOy:Jv˧1Oiz_7"G( *7m#o.Q6 JsP,$4oxw`Q5"ɫv9RS3TVJe*sFC͍? }A73s"l)d EVd\gQ_UJBE/=;q.61 _YޏmZAwYAdUR,I,MBnO8/31"fΏ "j SCSy&D1 luuي\gJRԐn2!ȖBc )WKLZB 5)t*Q^ F+~a`ZA>%)"ϝȂ3h eX~3h)0f#z؍ 8=Eh.R•9NbGV 7(&H Rc%QkgPlBV# z6 g#/E`dʋxjs=!Ҹ݋qI@CY[jD"yr ^Ϊӕ߯sll!\! 7֤ѫe̵A֊yX7`M)t ۷gOU%*ُ6H!B#jnnСUy(7Am:4([Ex jy2:e#u@ q_K]q@RpC$(dW4"¹â${K^F# }zz+Oipj^Rlj PV_eh@bj 75?/zCu""՜/m& XKs>ZWߞ/803r}SY:L8qwZIf FHcz&*[k7p ȫUm#y2O瓰Í/BTGX% 3X^$KC]aCR'oOTF1}jkTB(PG _. L Jop"᝸;*-=FN듾0(8\r1lmi$c~apz%>)1v"'vI  Q{0!YnS{`wZp7ALQNj7MJv ʼn9͌hia#/*%WIKivs|T9Zrį9^Wv6l镤 yz'V:N$"o̒/ .m:(dqi9LsJqṲo)^aͬ,'v' J2(%DXB,GWmTO)z-~7%ox.{yua_Fnvְ׃Yb> #KǧO I~FsR&8 {/uҲ!7V<=aUQBkM^[ƟWX=G0V S!3;+#wnuF*?Oa+ '?س{H0T@@DE37MstMjuh<Dbzl[p놿фGQ~e5 sv뗕߼kоVK'0Y]fٱ_?HPT"w߀q H 9)"U9g1z m[t"2l#So\|! CIxD skPyʸ^lɯZI|.~%LI>:*d@qsl:D( ?dlGVz4Ots=wY=(#ŏ =Yȴ[FV@܆?]?e\W(Nl\4Lʯ^ŧoU&/;qYpRxMAf p)FQ[JDmSS L?1||iVDd{2~'^ig 55#ChRV&Z7S˭oJȊ!IJּAS+ 9AIۢg[,B%.7+U^1sQ%jk_-.ET4'(Sǹߌ`J.OYז+=oYI|Ql=YRx#M4͉%y跊`%{2nzu/e(cj*Vh Mu)w,o˔z_bc/[r  䍌=J{,Z8#&6 GcO%JXcc]|PIx(0|!^վ".+M%E2/9MM6`L9ha,}0:@*ˆg 2ʋ , 7 /tLSh篈ՓH>`Prb}n+5dM|>Hؘ8/@,^U\4r +K :{*39BKr w>/ L%S&~uQ~M쬅vQj03ٹL'Ӊ"2<TĒ5qocLnPijb|IL^q.Ȼ C[?3~xg0 R{'!6'S@5J L"hdY5J}6,$ Iv4a<Q^5gLڎۘ;2D \fe0up,J؂i_Gܩd(36yZrţ+%c̘IE=] ÷k+A38xH^I-hC5&ȝM+{tq8+e| /Z! QO7-t?6t&qCQ#VXw:;/bQaAiNMs?zm2Mȼ{@_J}ߝ)4D cDOꙇ?`^4ԙë&!r*0uIkVPcv9 h"^IR;fG]4p/ȷ7D 9DJ&n2PkסuIPmmXJUϱcmA 9c4Q}!;l]:J/sy pX~Xt߭fMNOUrWn)}lHn:vE3zN*2:0;V i[m݅|#ff AĨ uN&El#M_NPHq5. VnT9d/X_=(hb rY5+#xPjw|*|ATI]] s0}Ij˹i#_ c WHEr?r J9lTLƚO@[.=ߚo{R9K[kdqwKltMYa{~4, -*'tju~ }w+!r Zj\%.!:fF )ö77u*0Jš >GrD@Yɻ,_)D3 jrä\'p/>.y\qg~~6I6%: w1rh@D2~A;Ul/,v2rVǡ?l"$E GJq I(PAiوh$:˨L}W`(YK$p'!9z/KTۚ Չ ;4}5٧jr3 ) ՙ81S z%% <~)MST6c@LYF yr;*n~;{`Yjpb.'x(΁_`[0eWD.c\fa B4U\ý}zSek,| 乒WoP"fN-Uy:VؔȒ3(bC0v1HY_luZ^n<|VmTH NjԺCLm%eseg{wHYi5skBv߫캩߽4A2*IpJP^j1L'ؚ2 .JCG( DoMX>B[A#"6P-xl4RkqG+T=+7i~x4C) T&nFVMy7'I,H^xYR6 Dk~ʚ?#lh^Ͻ'uW5AUGf&7wƲ8̢*~S5m![wq鸅2-lX6>O7q Ņmx ۤZTO8 .f=(%ܘC X{l9~zވ m!K9A2T8)&<5%[CigBHذ|HUv70uj |ڋJҽ\TAAm$ScQUYoto,L"9Ȩ8'F(^oA MqZ_H+ ҊU{T^{?|/hl0Z>Lr " PUo%1t}]GܫּHv%ev~/.AĩLpN=?I NC)tRQ^-(;rYԬͧxlqiɒ`4C v jIripM*Uca&"GnM8_%АCYfˀzk2&םW”5IjJljwӘ)HkkcS2oj/FgyQ"훦9R8B%pd P@t_K)Wىo)wH\i黌F*oYs)PQ#2P9F0JWמ^ ޞtRj;vw7YmrHCֿ,rZ.SJCĵSbPg%&ĶQzEg'w/y+僾XI@(1.=Pq"-0 %>hP= hq 6  遨tֱ9 mn'h£fc|e Jm yc`鹠q+e 8(i& V$6_H]9|0xWi(Ph9{\ I?bevRktha{D28!wqYtm'9b$KL Ră|Mx<,@6$pChAchOBSE6h5@r tJ*4YW=:DVy&` T>)5xu7&MNٺ̜ΈT^DNL͗3SRV$$=úXTmnp*>ݱ1y}b~5":3Y+A|"hČsib#0 xaZ! ٰf>WwP^gXS,eCYm]0g}hL۲pLIUwLDzy fGixF@%{"s}AcVZwps1IgmΉnPs f.K0?0O36`Zy+g99!6_$uKY{7 atH!HYd,D<4`U?vA]I:_zIHm4p`kHyDs`QRZu_btO S@Ų=3Ne=s1d;@.FB*QMNSu6^S2χ ;RRNb dAE@G' KLnĭ3n+XNCΐ_륁þӎd]}*xjfxjeM$IM-\r[Kۂ)>˝K)dP]y,~U%Cg-x/g/v(aH_'_˥0hO$qlC1\i lĈ3,C2LvLݛŌI;~7旀3 >Ca/) n;hnVԄ̓X ׻O \:BUd?YA9c=,d; "glo@ Z9 t蟊|)hgm$S]0Hp%Fj*$(3 .s(bn5%i Onm\ȲUz *@KH8aN.&:UM>Cj'Sw+T&S3 dW.Z0;[*&*omv4(MC>0 YZmultilevel/data/lq2002.rda0000644000176200001440000004713012750500644014765 0ustar liggesusers7zXZi"6!X{HN])TW"nRʟXgqjn]LV?Ȯk_\+$>ҥ.+q̲źZ:b$2"PPM[DG2ČlF\cc=le4FsAM&Kb8<}wW%TZ#TOFGASS1 +J``{zY4v mh3d'#g$B|jT)먺ʢJ0 v\1I,:ş1 9xNƏ;{&_jA. L#ٳ1xXpRuׅ/A-;f،p^N]%fN %dR8ÿGydLƙTnGP̖;Tr̿+ޢDgqA>Y`P ]"Ô)lJk:^-jj&WGSS"{M%"\g&9"Q^"n/GRsh%zC׼gsc52zI|~6tT+zIP]9%HR)3t#<2lɏyhTʊQ*8N?y'> I5 9K9Q*_Zfe[68Zs+|  ăNŞu.V[f%9p5Y̜d)˙hs-!Ԅɶ֎[Y8u9ڟ px)x o:|#Zɪ$N#8lzM%cѯNqU0eʇEAq] uFFm@jt%k E#jk3mࠗhNЕ7LUVFk\Yᳱ·ʁ%57tSKIC4XtkEkn t !YVƏ]a-=wg?JmG"OM 鞚w{c=TGTj$J zH7`B9 +!ޜotyuxTc%0r7'wL1:bu|' 1J=]-$5M9#+gM}YmF x1^h{y(;bO4"E@5@?$=뾐}bi*rktR>ƣ)WCϨWo, ڠcϾ_36sR1n5bCmK2Xmm 4h1@h{}zƵ mכ2 3ʻ:abv2R!dBѽ{P<ąTS-1 @xJ ɯrƂK/`w5;P&ɡJg^FuK(٭L><.aUC8 ւPlBAؔU|i5L'0[;`H@QMica4V^fg!|.h3?K!o+:6OQV 4$g&!h6Ŏ璯zۃ j.lj9zliYYp~ִ|!J Wi6iڽ+bpI#GUyNX/KC^F BKkvu/_fso\{PS=MpG 73/ml@<^rm!;mq Z*nє\ lT_h;[Y^⵲>f"1;/H,I9=ѽbI =lc atY7/$~?3OK&7%:lju9֬,KvYC_@kaki@m @Օ!"g  o]xİ&м,}žl&9 Uz,8)7at8^#Ѝ5Q*D䇯 ,Qdp;y1N1Ƽ UW9Fy0&P0I=^}j Ad[cg*XSSNXN.DL ,I'N`G\Ɵp o}W0o *㗘\xn 'hY: bP7tctpT{?fgŅh;h-h S=U6M D7;A-rwKB̭-l "GJ8հRq+rUѯ2%%7TPR5n, 9i>&]̓?YT$ qJ:CU$[i%{ )wat zT58OQ:qPWi&RқБ-K婯¾^ǶrQn4FO[%l¯AL-(t+ dIw#i4xۑ?u4)p'{4eC>JYC(`Vzzqoɓ#ERp$~Iٽ)Mqu 虅{_bu`406HCM)7,I14|Dn4*C)lus$o-MϰP>6q˜, G݀u>Ftϭȯ+Uf v1Fv*ɣ'\M{|JHy5_NJPҧ)`AGڹ.)MUtz@Y¹GTo]v2vnNf4Fk*wA]Q" u܂&4?QJxדla0JIw|5TWD|>TY}gQ`4ƨd=ѩVNX ƭG5/ْEatM =A0PVC4uKGlqN%@]G 9u.ggѿ*UeBaX=OD2c3+}*3G'h%'#_qeV|2 O #U{ 3|Sؼ G!Hf}ˆA8?Qꏜ7zKf01)CNΫY hE  lT}U5M*qXW~DNl*M Pɍ==pgb+U|C_5־CKdφ<Or{n=TNԍj~tusK+_b#^VO}Cbл"B,1[C1=O"xgL@Aӹ;%K@ gH>4-T4.IK.qr!P+|(OS~-`XvmQViK_;Zc.F#] :#q3EtjuSW5AX B]⿣me5rďۋ0khS!JAVD! SAupD˔M'>2fMh״`'`fwJǫkBt|K>?#-QoA~ Pf#|* V<>P{(E 1CcChY Dn7»*pNu=rzfوАTzuQ(h"޿0nݓ64@: &l%n$bV%-߼,Y`.x,g05α9)gn(ZS7Udxv= P>^jeq=BNsFZ?/IO`G/Kt5Ӧ;RJ BrjGsN($cy_: IZ?DRM"=p"9%y\CNu!U|G1WJv#,ted~̙:  |.5Ȧ&缋(R/@cA/_b37<\12ZIUHZ ?t|IXkG<4 w/ R6:qĝC/SVq|Tu2.,fKtZ7D:Y7`!|6,+Qy&E1 ci"<8jAu hw1u@X 7vS_剫8d3DGg٣݈}^{8tU) K'8o/-*YBɴ'\ځC3Jj` Mc+ IM\ mZ#Vij?$Yzpz%O+>l gEâg',i䴨nY`]3bf#e Cɷ\M;=${kA\148x˾HUr宯ȡZ狀sp!R9sX`fI?=cQ灨# +0|*%f68ao9PR[p@ݮ%`GaX^z5 / [d4`RM=B?Ag3z^y:@( [WV!SjYGPMN^,ť˘ipD Au{D_ŅkN0W#QLB)|E3 <50ya+]U%:6 mS)yzsUq;wQS|loċ+sh@+}qPkRjdn;YE_XbdʬqE/H3`hma|eFc"5IXwH룄 iY)h@M\P۩gi Q?~DXsfh<όn>̅iDOBҐR` Ѳ{nax$Sv Q̚]Pom~8@Jkj!5WdO$E Ye$; sY7~ή/haϕ\&ŋA#K gnY8#BHNŤ9cg3-mhCZsggLӫShg#&]og_YόaEC9ЯDQF [@=ǻ8tJ+赾2#1B8Vsb.!0GV.9W̖3h΍OGz߮X\ޅ2νAy8 ZN(lΠOڞ Qމ#euÙ =J75#ͮe)lۚ_3V耰F6qdpv|P rfƁ9ic.vu.|s]VxsScŻ5+o騨[CZl_!v1[0ƽTC`Bi 3#tPOtGU$[ݔ)e~DF䜗6kx嘻=-^pǏM@@$ wu#m#D, ߆ >9fͽ 3)gXXjt-:Hs灡y E÷Mw͈׶1qR./GTA{F[Wj PYf戔C m֬h]66MS8@\&,>qn}W1ruFLLWN@*rQ&n9Ct?Mb_kiUa+[#9rdV΀Zq)Ip&RE}#',ōї^W#WA6ǁS;4۷jTdV|H;W}Ba I9]#M> i0##^[t jיm-2D sf*v3\b rp [ᰩ($ WO6f tKfP\בg?ŋCIɯNAZɭXw=q,Y4>mmV*+tC V=:ϻx%KfN4<4(Mkq`9*irL+D.zu#XGH{[3Ap31X+QR|yxmX:mݦ 9,=ܱ̩1 +sS'\Pݛ@/tgh~'PGok =1VoZ{2K ˟ѓ %XsZ¹; &x=ELsBy"J8 ep/6%zp#f-Ol@w;̓Mfxv-ƿRAUe\ j6| XU0F ݂ >4m,,1-ҘM3jHG /N ۃ|(]Ёm~*2w7U\4_fdn % 7l0j[zl%6 G}teإs2!.rmb)WFQ7Gw/WqVVWZ1Drlh1&|Ku$^+նgS Tp~^,څo;΃E%Pi2o/ԲKJnKHgYKS%֭=(K6wc |g%acK&wad7qF7xR.?J2JKQuh{KSNngjO?L< ({復}cQdRů"jjxRƯyf0@D-2ޖ L;k9k!%-R{[&iWdW&!fB=>f@f jj"rhT~(\o@5F|iHeAF\Iс#+boEcmh oWs}֥4\l^QC<ݷI{N?R{&ec ^ԾQ+f$N|F "#3`tf0sY( ˜Tx5wkS,/xGkSG ^|WHJh _xX+:X } j[CLᰭ|A3{1:$KOyZ?7a6[4M͏O0Huysc1AO2jVqV%0@$M޶˟uI;- G',n^]{OT請m±|`ߎlu?ճ4ſuv#s\vr(YOUe{,si+Pfk^UCkyxWnr ߟi}X)b)ަqIKsC UfʂZG 6Hji(_*"K\IrЉԩ3U|[S/U>79oKRfKz+dWf$%uћ6uy\]C`@\YµeM(*`CO &*Ÿv,.7r~RqwU5{ @¬؉9@jCm7PZD5|f +ӃބoQzKiMSzn$ک}I2(Pm_KuK76%nЍQF!%"љW6 |_ʗ&\~@#wv)X= a<\B$ :B;^kC }x<n[мXEhia]ܕ,TnH _8bT=yieݓnMcM?Zt}ħ'CwRQᰴ ]:&?B>_ ?oh[&o4J- <d,3Wt_5H4OqނXGL{@䧕kyw Ê;D#H9y`[*k<~^E pÊ՞:D@ 2RK \%EPDh-K>¢C0\ sR"k8$ {IPա-1 m j (&9b=0rdY9|4 6&7M.$g+JĖKvqx젬8@epmdqjIvLC/?Cv=tŌdqEI km0il̰Bp'ʤq,dӦ9ہhU' !dp@pR᧋v {W?ֈkN0F=5~6Slܥ)C`<4a6x׍<'jTWk2I"#7\6>hUhX>P}-.SgltS](kZQWsg.~h㾘芸DnݭLW0 ãj VuJƺ"X,C|y;yF95P=JmBdhZ)Nc%:S|s!LQR5p S',ޫe٠H[W"ݜ+ˊd; 淺Bw@h\  P?Sc+X)"l) 3$`]j8^U!Y7x%-4pLr~,8~j6kU'q4v 00Lu*yZ97Saxo<}ngoYS1,eo) !Wka~)c>syN vtIA'G I >a:Q*o1;YPeD d[a"'HysKG/ ,͉ e1} l%67Pr@Q3,V Ҿ]T?BS(/{[XiT?Ypm>0@'X>f)pI8BkxnU)[#JLV}8Ox}y+)EZu)Oj8Wl`P9L>OYo!n׹H#=LV{/ R}!'L(r#aQk7f;G<eC6M{G'f:<,n׮AOfľ'!_{0y"FֵJ[G^n8>aP=;Aۖ.T؎jY/-Ć_ҽoKkMl-ExQ^kTa#y_wKLu|e>A\TAey*~ÿP$Q/1\fU7J0zqP8l&"mutRZqh$ /4 Kael_ p'r(—{ź$}0GegK-wYZ$o(9NrY$ %=nRhk񋊙_wە gn)YtGH&inC9Ȋ&`C'/;׃+څ!8#{u5ErK*գ3|AYskqUޢӦ,NHvMLQ 0o0w KKջO1mlo;OtwW k) 1R 0hN/hX2J4ƈem2%E}Hfr'/wru2¤S+3]Ks[ﰲC?g>͏=DM,M '{R3Gxh9IAYloS<iZV(c< <s6#j}&NYo%!*{M/E 8 Mܵ4o.i;Q(XNe3GL]߈ Aa a>d*x\y"C,V4)*Bxg쭹e+d~.1#??f:(YS[ߩ- ޣtQ6TVk`{*? @Ǖ/ߠޓ7:ݕ{xjs giʉ>I+ɑ!MeWgHbXk=fRi χ>|U@5AiL(mXH{c v5Ei6Շe 0OVV+ÍA oAbk 9ʇ!ŐXq'foEVy;]dPGzK_^.-z C}t9f]L痱˖&54qgȺQ#$n2!e:x!T.HRΞ=J% MX;\ZFwaG;p)1%16Ics#+8.Pp4MoWfq&_ڗz5*{$~iC YnwG͈U᝱5хI6Dw80vV>MJwAv]? _a 613`cn.7Yً] (ܿUPOnE8^!T .nUԆK26ޡ Zɿ}T*ۃ}0Lwض8]+T@$ rEkD/ c;$MfT&{Y$vF "oQT@cT;Q-^l-мj@XAP3m+"#Q;ܨYAOl]Ek?W<$`3wN*-K0P- R@B ,aC"+j(a܊ζ_}_T?Qa rV`%8H2EL(lxq;5 B]s^2+t0Gw, ]W\푤޷F)%, 9sE fWYP} >L 0;Mg/8{f} /ly>&̈́fB!]eZ ƌ`!?Zv߲MAh]Z?VJY==~_ 7C $vQR m%}CBfD(Њp!M+Oѯ;Ġ3&c|;ᱥ\zS[gsK/SJ\S)yUB|Ѭqo GG"S\9.a5D;vC_~[ℿC|Q({e"DeW ҟ([O:x.\*bQ֩^SanIy 4xs-ޕmg4"&C򒽊mh@ K/ 9Qr;>zb;td+t=oIJ9I:Q GNٕ `~)s]WNpbe␳oS7t̪tIB[5b`[%\)nJ1)ȟ>5=-(ʔ(tTջePqWĮ_~վeUɵU~'ҀyZ")U8r5FلZR a4^mxmsIiIS\{xV6O^ܒ.-?Vtxa.TNIjbVJuZ l%۞91#Jlse-褚uDDVb5U.v0M^7ػ'eܪ9g`lj}*sX5fկ: _ Bvm|CS11ŴyžԗGRAp[TBu5ы(5Ą1뿚=uMj 9|EV0okҸdɔha?7,fR&D> #BZXPx@I<n0rnnG`CJiE啢q6NX䧏34I⍵/`+ABUiF|ʹdb S_7W/vA j֣]RA6ňGeZg< C邻4}*iV6aSfl^lph::b u4zZS1 0hU\~%lkj~ h`|O@:҅oqC]ᠳYסltl> DzĚw˴avd;?wgԼoQ.zz&Dh4Uc\+Zy =xidR_|R" j@$칭J@J#DSh/eۉ]h9.9:[i.;{C>[Rcp$mlr!X5sT5ECnΣUuYW1 <=K4jvjUĪ ɕ{&vƱ=b q>8 Ҁl}4.'QA&R-gv.reTIM 3'>gɥXC,}Ɛ1]+2rH?P R$ KB\e3N:tƶk"T`8S{`YZ RU_es2Ј4i|ƍs5Q#) 2n*i~CU~TH0څϫ&$?45r;Pf"> W/Y"rh ܽ|/L^+([/pQUmʮV\A,yߟ(HmkuN3?_{ږaSDnA( rVްN-=Bo=p}Mg@#Pilf&hI]Xi`y• `A\[pZx񣣠]= wwy2~iہ"ھB5ڍqվg5f4&˻ Bt"|Ih(6<ǝt)"/t)M0ǀ !|7!6Qޡ؉6'ɱ{`ML-Sf'E+fi$,X览:$#)  #u[NzXs3bĎ (BRV>tSJ@'͛G0tS12N0}kb/CTA԰U~( x%dT~~Zag6lsK (,VZll4Rl e:=$C'3~~e-j!>ѽ9[|lVI׻bLʢ˸wgF=##58.o}r70?{o= 鞿+GG^}E)dydE[< Zj[w.i`.4݂/3MtEdĸC`B9uWtgZGnƖ:]dݾFH^Bz愰+\Xh!䙚˔L"U罍 Y=ca&>opjꉃ{ X>6@.|1}ЊV%ڭactfgOkAE{S#{]r=yԃr;q`r6+e-LXNTc[e^6͈ԇEUw3Ҵ3<;fX~\ uX ʳh '^2Bs<:h<ڀҀ\{P14[i~ A3BYRz**pISe8ҙRkiӠ}vpNU%O8!!&wbwIm$gnj ̈́RԶd"(6 }EfWA 1\,ٍ,عx;5᜶ X@!N83Չ3·8-5"/`;Ӑ=rpɉMYD_K;GB4{q IĹdKWc/.5( Kݏa-r`Z_+ h5J8K3J)imuG(" dԭ8HNW,,f>:x vQ'sM *gA,?dS*cQK-XhrۊԲc+:̃/Jm~%=GCYytۃfu5Qђ) }S}βFw9~KK;e~(PڹY(֑?)̋Ԧ#1_{=H&ߧW۟V"TD"HVa_]Kt@jh"? p_*S1" =B(Ct&8FrGo,F#/GkJ2#2qD(ĈVB@V=Eo$f= `R N̝`[ەI6)A7(5~~ F3cN߻9mHpOb§K*k1fR5p\ .{hc4?T8uԉmS|A`upմYM[>@,n;iZz@˃"&<͚!J9槱I?ɤ|BDcغ9okۜ0C<Rv9 YV.R;Ol5œA_8rMCҭ)T ,>G{0rO@jq̙ }RJ$CXmAl|B+n e7{u\zgTW6Lpl5Y?-3<MKԗe<.Ze>{K`XAbBW{c*5"_M~  +b &O7J?]* F{#Givp"C 7u0$r}TV Ѻ$a9*9CuJh'xCL[,f wR~Ϯ)G=a YŽ[K CT,jL@XD3?#;B =tZB=yGq/y[%=6JSn\YAپ{%],|2Eaǯj l,<0Bw;(K)W̼i|dAb½R`,rh{L8ooBґC#2{Q#!%4,"#;!$[$-5oZ@-IuKԋ^0uҊ/[5/ dnl$y}RtiE>0 YZmultilevel/data/chen2005.rda0000644000176200001440000000220212750500644015260 0ustar liggesusersu;Eݛݙ3יqük6}rprHH&!qR;vd;H8jf%y{Q]UWUnYv$Iɠ9c0X;p#IZ?5N&:!Ȱ5cC'LQRjCc]20|haaAѐ}~mK5n_ly Ӳ^ݥP;ۭ{^|hm0B!=Y?}(Ηjm}ikvauhOaQrqc{?cy>eߡ-wkoǷоؓ^+:xW1c|} GshO爟tr]fg1o\.Lz5E^?+_\ у3^ïZQxG?Oҋ73}\9=Ӳtrףĭ>f)~+GZtqr=j_K-TNJ\=8~N훫q#WhyBqo(oۮWHĿ MR#/"2Wp~An$?'/?"綨}{zxQ/3ɣgF}#.\>U~;ǩq4㞐~Ηx+x\ މ!3x|o^%{ qHqEy<)]o: /ZQ܋ .t[Kn .Zн[FxO߽ އԇ]ҍd)L;uf:'v9}):lî^/vo7anW߮bpo}l?] +}multilevel/data/bh1996.rda0000644000176200001440000034523012750500643014770 0ustar liggesusers7zXZi"6!X])TW"nRʟXgqjn]LV:W~~: G3#q+^V&CɠgEn,R"}U7˪C-%q̨$[GE 7}em?Xzw^$R0pLb۶9xl6&KXO2n"zŒ ֐;2kݥﰞ|3(}Ҧ:/+a[؜c3SI>x5,)۬l3?_W/ɔl{@k(ᒔit' dEbW6@}Hiӓ9s`sGԥ?֣;;j^8;o;"8pi`ulQ7'3ُn7Gq9\.0]R^Bnj; kʙm wrtu|hr 3ޗ̨\k_UG15 G'-NintʭY:y%)gEXηɴM:;jf{-BI+mNn+S*JdL@{y8O`VK9+rSf9<a}D?$j3wmƗ0b-4EDiPseY5S-A N\"PI|%e!K`/ Qކ/v@4V迱L!6ļy&@:(hQVʏ5ŒN&1պr .BEX&qJ,c]?*!QI'`nNUt߄f:|pB_He9qB.>])yxJbfVk gI$aPR]Y?o`hnΩ;Ͻ] BZjSY9ڊ.@ nYu=bGh& KY&swS.ߔ4B`5Ŝ;/4$h,P}TX+@U~z U:^6a5*)SUrd",fq;T^iXXyF%JnZAE?k]۞_,47&h10oQtU)F%["W=a shUa2) VoahJd|QG`i!(]MtGg' b#y6;*| C,*1hXE,cshy+4f;K=YL|2҂"Y @_{Lү'J[U>+w֖&'Q^DML2(p6~Gs몑 5w`6T"MRw EkA"g< `j@:I4c؏AFޗTl[]tyud= 'Sm*)7B:_|ྜྷ\1v>F4 pho RL7X&q>x ll5xbcu#[1-uh">{YeT-83=6^p~еLTr0/Lz侵}7 ^Akuy}H%VAy&/|[ɹcG,/{c"8bn1!6|uD~pųKV:N`'@9|W59V}GoNFe`(D4S(mv*P8e6  5(fc^r?νە-[ȉ]&-"S6pU5VS!0gǥlaQk?'Ⱦi#FO?wy әg.cֻkqhIwV@t Myg|?QNezmW_ĚF\CQ Fןg:UGwZlVGQ9n]gDrJbk;GԬ 7QQ'Yfx&٨QW0V,T` (UvֲUz0~;C2דJI,tu*NU sdw Iq42`x@H#^/~mJ9x0W }Tg(Wd,]z郡⸘ɸD,zɽ-d ԏUfծCAޙb>`.Ɔ)/?dW6<EUq}=QC/gJ! Rߴ!ts,#Lr9M#1a㇓V[:C}Kw f,0I!ږ5+ȄZ] Iqn1_N**e;!M A( -ӕW$|+8UjK6|&u?`{fݗkyrvUunHJdj\Fv [^ )&(W 6@q7F`񋰁K &t<txXwEc%e^ӰଲBcz]`33Ri)o%1ۉO 1_sS2~N̦6RKLAQ>#sC,lj#FZ|31 *I2!mSGg1oT@'^6f ] t䅏t'4͌1"b!;-"@MQ ҽ>Ś':U'[;գܣ#ڊI56f1GpF2!`^)#1,K9LkvKa&=:d]N$s[U D+LQhC2ّ%''߁RG>y%OE|f48B6e,F6dnmF#6RςTta{0ɟ|)`FnrnxDC04^sffР΄꽠BgvE85^ߩQT O0+Ɵ4ohsdQ>^?Bh:zzסgH t&ƥ֕K{5bn+@Q T#_,"eD-tIa_1DvNZҰL4 Uތv-eMtq-e:%I纎5X9χ,$(')iGkː!qʓϘU+3 ksNuΚ#aV(AL%?lqfD÷Z!MBڹVPEPN#UiJ6. ɡtJqaAzxXE+O E5-q~¤(0hߩBbpi(ۜ߉]D|gXbU@8f'9ߘZ7/ЂOn9\˜jx.$m15*6 2[3ܬk"+ʳ##=g-wTgPݒ.U%V7e>a㆝ӡ/ŧo!X ƣY_?dox*ػkJ­Ja-UߊH<au3M}s X[`Ѻ%Ε 0y`m= -: ǭG ,s&2=H/W_S$w(Wγׄk-z%Á {C< Y zE"(O.ލF{!9|j9]azRQF#t߰+&7_}B}}shQ[ Pө\/حm)_ ( d>BIkEhʑ9T?^`K?T|ͧ*[~K?T}-#8QYIů".["{& F&'Hx\$mHBԋL|WF_'bB D7ϧ gq1Wh+mp_&dūC/gvz Va*~n#+? r+@IԅPTV7{8\gh4) <o? *-˙_V\Ɛِ]%])1$&2玜-3_i^;{Y2kC򹒏zNG*txs~sQH' SYi>K0d]Z}0ULg?g̖QAp%GG- lL$zlroInU9NkS9Dw'M, mV- tu8/C1W7T!MBу#C>IE#}p,|],%2'R O.[\9Rrfbjgu_n%Zbyb|r دGu,``Z>0q K3ZgRd8N_lM,NcSBYxo7%<&WO]YL\!±?[j |@՝VJjӁ~7%bz|1M~s!?O~ H"ɾ+‘QmL GeWsDӌNnƸ8¯(\b=w48=gش?iMi&!F105z7Q!]*d<Sro e*2fWo3q`s+fP$" w;eN6c+/>c״C_.,ȓNBs20kQT]o4NF*˛bxVNk w4)w! ʘaI|:57pP$ ij`XLdmKyQ.kNHG2Aͤ4K٦훨{Eq2|ރ̶Y\ f6)u ,Pwצh<'Q1U5J{H~?qxIy}IɄ-ȺJځSWyfUz/A9@]zGǑO4ӥKi Iglۮx>TlG 'd-!x0ѷGS*Dm}+J-Y-_c*`2䭅y k"M +h(ee<7j4SsZ;R]mLec{>^#Q\q-t:qRD `}Om=L7[85}oM+spw?(H#GM\L8U z(NkX7)X5^[e6dl ƍ}M `hEkox.\[g65#RBs 1Kk)BȦ;1Gpgd9 9u!E0(*eIyh)ΓܴYvˎ4yl ~[B 8aFxi? A+|%šcAZ{?rڝL2:XBD%TxU?5>tHF5kTC@EIN|[7s k7hP" Um0_V'ܴ[ r9OyMBQ)}Vn^$8ο)#5{X#Ќ8˳B{rⁿG2;m)mpt;C.8ڽ3dMfoxLm {,r:w@c),>I{7k`䓔)S+l/!Bn.<oŝ|S|uRgi4PDǃm%g;oG悆GȢqWH= gw?;Ow5yPzbQzdzIZf(x $j-$hPƔiݒdKQ@S^>4WszZM]94uwW@n).o<ȭ|ϖ|3g_ߩ %,AJLsb:]'6ra\pj*sdo[UkLw=t(;;m ('>l >Nb?lb;f:2g>=e]T}nP7`C.ɐO!7^*1诶(aѹ@HgRʓ/G!Bql8*+rdSѭ*ӆė"~\^#7rQ@g\'Emy][Z Ҟd ZĶA@C s8 yG,K=:ʛVio1Cm8"hC~am@toQc]Ndw? T 3u; P{ D`I)V Z8m&e΁*CIA~ &cSػjx7[Ʀe(UGTR&M:htdf :/t_PN>Dl.8]uY hK9L7# VKPC_DvLdB*fvJs܄]k a_$y1v4nIc)R ؋ s2U\qi }mS`X1.T"WV2e 3%pmlNSw"2F ^IP1(WOpyz3)]geNۇeydS`5\uTՇ%: ` `Et|mQyqQ8:{%Ii}%nM(Ȼ"+W{{\EdtXX/8`"R&!YQ_aQu՛}0{ BARSxv:Dȳ."&54;uG !1;Y{ zrdl=SA ݞ32 gH r Բ{(f;/ AdBms)]Ǽ1c~™IO=R$2փXRK^hM8GEk9OtCF}l7fyXwF\.ߖJ!"6!$Q _NmRuVGlLu YRnJ3lC MTnOy4b_rJJ:59mlo>7N!Ӑ]6"J.ݕ.mk<QM ]r,fqOwwliCQ QlZ+0̋-VVN*o lAl~kk,ݎL00Bm|F"oqd^.H0"SD_QKf:'Ht6ߵaN 5ezN[Wj ӵFlK7 .3 KT7r{7v)Zb0h)DҢ/zW}H }B 9.7%T )SIΉ'%o-iz|ugFi89eY9ăXժ]OE}w|<sKM:A2>65f_^@ hf7vv3)?Jcj+F` -&UG{S=lU^Cu?T9[gS5M(mDƵ3Y&T! &PpJz9һ|Wr[1bܶݫ(Dd}}ѳ:=ƞw&I!1ӂ{M@X`֘e!$cwO`I7,JՆ}39ˡN/ hnAjL\O-^\Buu! /nnQ0-=OO `UEdM GNL3#hћlJ֨+、zQ[)AT%eTW̓ɘ|'9kzwȑ0{ֹlM$֬}{5"P'JeA ugڮut phx!mku:yb`5o(Jq yQx>F2"A_KlF.=(>uܙ"F܃Q-1J~hVfM@x҂iW RRj郵2Yp3B%(.A'{j{=)W8%Qâq0p PoDWAWeWJf ѽSw/?l۷e(y ,Q^H$5mWb{e!I"43鿴EC%&%zwf4*1sYhC+]>Н+Uf :mNٸ.# nWGi^l}#,E  ׸ݎQ~)fV3}DPKzoRIDp4Հ $=PɹTDavRJw|=)Wۧa+d,w8Cv n6/M^I/B^ p v{T+ČELAbG1ɲm(:{qZ@O-+YBhBPX_6ѤT[ /$؀{H|Uߌ !āHf8^s GĜ/“ѻBaQV:$܊`"x[ZH"#Dk)OOZ??1| qUJ3j /9*.1y hQɩYjSpvQ>leCl(B&sP}%(@r@ H$@W@?+<͎'?dAK19{DUBٮ_cϥ#=59b's7 A8?M'ů֐Af2| t=L'|qj(|/qfإdWȯyn,V2eK& Uhv])H( z ")]7"%CWW``]ۤ%TV eb 'JWȼEW~f-=# IkIgFuDvQuBno[_u*C^cPd~۞CeײpDΏBN<8X@0,Z|x uTv<#?魞O3sd-ϺBIx> r%`bhf*5)rQ%p.{葳? J?d $#.Hv>EX#o)cߨoGZ> 4!|ړFTc4s7QvqA:Itrg>یyLSmLyz3e6lĉQqVHu i@ WgKԷ| u]u#G2<ѨURm leOH^!v? d5+GvcOLڑpW.#n)TIdtdmLHG5W.$E˳wG(4p\Ĉ׏UV}4)b/Mɡ0:p%̿I%[4YyYw.l-ʐeΝQ,&>W)Y9!+CJ+]`9 2q|JbP Ηҍyg,Q!ېتq**\P s/+PK|<3@돶/)1Rs1|Xj'I/6"fM~ z~|ՕziJ}rS<}hRvzTtT`QQyjU20O)[3RthEȩ9*#Dcj}%qyEkGL_ח̓j5Ru@f!{9Hbѓg#Z,aKn1 -G/XIN 2&Wd|07d# PmԎ&W 3(gx7t07xH`$x]orj \,^kkx`l#\ejzg f9 ɷTD~Fnhku_Yg(mS8;'Vqo`NR#O/.5x?;M F佫g䌽ޠ+gmd3il(}Rh`# 1nCw,ݣpO3U&xɾMvk -זA]UP2[6r9кE,rk:FޣM);Ai:@Ioܤl00,0\$'!VrODBbT-~8 b?G$WK\voM cuc%̻lb.ѶZsj=0ҴLRjŢKs;[48MՕww- QJ . #E@3z!}IP["8/~eHٚ{hԠr&Yw:{w!'9+KҼht&n뼱2VlhO`)ݕ뎥$]a+IH]tl32GCAxeL5x oJJ4_Q70!5kDr|lb5 f;UNJ0[8 0x5#۲Q?(Dh'r^ARSi0)C.?v F[Ke,Z'жNv02|;>Tl@CDiCKGX IuŸ́y QY3st ~Z&.) Ubj 9{0nX'QzHO@ߋk*YI3subņ:[;7^RJ̯Y~epWSvQf|8d)\[TCCN&ĖB@Y({c8,M2t h Qa?CP[YkNj&=NrH!7B(_'"#0'J2tWYW:UCvMⳀKm][Aθo/+cKۮNߑCE `?1N,SԌ1SkLD '"RB^2h#3 ~ {Zh+u+} Mv18SQ+=W:kYwWޯ~vQ@"*@ژ\2H=R ;U5;# gY c[ ?h #I  dYh3)7\;;b#&dfߨR>}_=/@mZ2aa+KQ`Qc' lh]T_ SF-lHd \9:m f*ݳT{hz)HΝ}9-^fp4GPry2uɔ.dUCl˹9MQ]Iں;Ujgw`idF7<.NutI<<[AwZb&,Q\^OD]O+sv]뙯}.(oB<} #L427pMIKV ,O:fG'ioqE]ƕkb3ưT8I՚B*=0kLmm _ծq-_alWYSyOJ!&äM- ɯ&V?\Rݞ%g7`{+N"/RD+f1=)l"& *OS5*ʏ`#nS^<)~ YE(OKBX3XN\VBjlI1~%al`BZ7 ̙ԭyT{AL{zX=nL!<>.&^Q Qb;xtGPi!dԱ P)g\ Wv~7<|p#B)6jOLx9a[{2/՝h:Q+dyhh `Mb /O*5׉ce)M. Z~} @hj&D#C7I; i ؑOE6fU5]k0#C( 4o4n%I+ {mЮ?Akm4X|峹vrB l%Dg!ζÁ^)5DLvYkai6zsJ!=QGHR˃"1O;xC"̵]iȀ#_E IY V0#LᾧgLjC`G*C7|w>ι5R!H*1V^lt -.t0_o9Qb6kV㕥Zn%Gw ̲lZ^sE8 u!TÁ{}ΗN2Nk^G΀.*X~&gR,at,13?J<lM&_nnE]*@_aP\zJuVl/= R%";gԧ@G#{oW-RY>;0]wk"/ԅXmrFgԦ5Vi|#NQb~'z:|v*S^w'l 9 h5AF ۤKqV͸s<4 F4O;A)YvQ[́ 9Y`-bϔ3`2fkLqQi9>56~"utITz;w&_`p=YF<(!m^ F!ҸvV]g~ Ưz/26aZJYKz\8ET|íai#/Io+ 5Qɮnc(7O%[I~mN zW;P#|U΄0}HO[lvx t~F[?l :VP[`۴LudBq)[ ϱDGp&wF6q†t ;*ZfeIvw{+~ pmB_F1cFZpRiDѯdFM;Xx 桲Wװ]CO0F&@:ؒnGXa ¿@BNA N8W}rpYe:i/gyJjHˎ2BhP   l\Z9덻p@AR1$Yl[pil/#}A8M \7-AJ╥? OQ~XȆ륭=5U-Aze>ۓm4?f@7~ χap5J_gBk}ʝz">ݲ/`7%n &.*1H[}彝 s@8 .] r2[A%ҙ=8׀^z1}A)qhI5 ԵA [ga ƈM❷rho3loKB#_!8D鼕=^utoϏ@_p&z${_2[I]͓#VJʹu i4?Zu%^BͫJ\ddɔ;+ɗU{@؅QDIDv؟Sjȇ)dl9<'7qSi12omqqS-1SXXOApNG8mt.'ԏ&Ͼ/1狍s\b`aaxxXU6^ N1X 58Qѝk1O"nE!S aY&|Rg@F7P{a띃 y / nl$ȟk(0@8\5)CG'"<(}<(_Ԧ6I!~F)lJ{X lL7nIG5V*7!J8^dTkKku]qe]vm^n6jڅOKX f܈TMS@Z5@nc/_IHvo@H^(qj zш'gRB?lDRɏvP"YEڍtRǏ>kx8aAbe]NHk\>7eΘt!b6]\cl>rpun50iƎMs.ǤAu#"\7R-9էTۯaॴb}+Pܧ-f5ROQgip=2N5؅Tx-HHT~i#UKt:ݾ<Ү!5QojۤZ]5mI}rCaxS}hoa.aM5{}@BvT1<'! L֢ ̳c5NO~ b;aV żK.g*"GPl `")k!]6Y7nEÕ|Aّ\ tamg:gྰ$;FY|T|n[g 蚾Sp`U[x]i/ݣm.E:GF9C̽G"OYW[cOnc]S#jA,09 h_Hoڬ^ K_ȴr{}nKU Rty D[١j#1ݠ%u[Ӌa>A>v8oTB\ٮ꣭ xsoz;f34/ "դ 55Ž=ҾS=(Ϸ' 2X? #/hmpԥC݃SPVgn; DWX^=Ŝ5oxRɪC[?2|ٍg|;u]s٩=ĺܨQwe< Hm:MwV݌PlVg*â*}Vv)K` SlX {y,Ucn5<g{X'0(n~Ѥ4ܰ O .򃽘n,PPiSGy&竺.WtKhVeT}H^Eb;@j%"aKUϗjc]vh~qg7+RaX7=,`wfɋ Yz/dB"~ 3n{hOmrg?Eu=S-SԶ# (F[Y?Xwbh#8Wuz Mi.~oEGaKJ5t(THLKyzVF4i#U.[į-[/[gпX$Hkk֍x| "MT}Z\a >34<ã+\FJDTN]WR_S ~"^0ױmWZ{Z~D%1KP3[-nmG.]5!Bn{TD9eq; t38֖YO91x/.p\MƓhɵ0d\y^wDgIEw0hZR+-JwSX,?4%25ziN^3NXP&=H`}"g焕3u5&)A}QfyRYr&lE׌yEɴV@oFM$2s_Mfn;ډԤ r4 [8gL;ؘέTfohvqle(UMTa;H,$~9Ix[jqLyZ٪ *+vP>MnVIN)S'ogXZxFҍ^75>I.4`.lMr뱣,xыuQfu1sgYPHrHWm D@\j;+qxFo#fpt[*3$Le w6BuU(T"{*v+~G?Ǵ- /cJi2?><ѵ{!^.+ǔjf}=A1$5qh/aħ{ɷZNYxLx A X|;o&vtm!n W=ÚO:Nq{- ,QzKꈠ3UGaMz}݈<:װ<gIsZ]+rqm<~n0mĎmN>q|\iF_71ixtC{Zܥi!F"c&KYRəyPB<̰I6᝽VucM`(9 [xO( x7c&7V4Kiv [h;B,$8wG1*HVaUAmR46\P[ * MS`AYHe[̺5-ZI>_Ţ-T}"ZoΈ3lK T\{2$K`1C.t5 2U]틡&q&*ϼOO5E7 ڴ|X윌3~sk 3zLl-c^Uɭ)h(I5eM?2L6Z  m?0# ̶Պ [!m٢E?mˀ~/#>5C}65`)hQ<vԱsK9# aic`m. eZo;fZ JO&QX{iu*׿dIt/[9 *bX^vqwn}'FT4vB%̆^6A_ QdYe=#Ȏwr>+b0j-A 9ƳPyyrT->u%RQWV~)VÐmM 8sNd;*I^yS%K^Fs h*BRMx4  ipE]7#6Ă.5"2"PB$|'CiÆ(g.v'dkf.#OCyVZZBGLTpԈ}uzKa}ZZ8M2aNmFic 5#^E9bi.kbj5s*?&xIe Z-XoS$u#F>DXIG!LGblS'Uc}7}?KV d?sᣳ%'{ܿFDdaw3\ 茒(DZUS-xM8&,s(-C0((Vc$`dRdeKk$A4#&?f՗kJ[[7D5>nfaW+q& Op=OYj4.F yIG8hq6r t+HiFa#YiINp}v\竟 ?%]zVsN&lc̿;?H:e~w9ܔ!AH|Q&K.7KW "ěM4Ki)ބ(%yslX a'2\ߊd]_1X$1ڎ3XVTvD%bo+"]D%|Ȱ{ezօ.&1k,0]8SY)."]Ul60%M6[Wp<7\̯4jt'<勾;u(-)ݔ!ݙiY ULve1+dq) /{w$;3:OC$=-+⁌5%CֿG9y1K6td'l|e| gdеW'7$b)x =17Uv1_T)I1L|o}Cr,6נf{_<ڈGkd:{|d-Ύ ¡oe7n *1x#UR I\aTxhaU_H؜ fjΩDffٖ>kzS5'/qU@L?DiP)Y)1̩V%PÑ%yۑxBShYcq1A3u*L ةvTrn)'CV0=P>{!S!Y:+)7ozo ~S'M(G;\wܮ:+#Q%ݺݾg,elOhfW"w5)V9,  u(ۮ&vG?.5LnQʪP%KM;4Ӹ%vr&wǿ~Uiy*r[&ȍۉ=w- T7oN~HgG)C&vi4㼇CT0gjɑE9Y6B=\FP|o{d7'͙nXh[.:aut)_L,{7ǎ %]}ySXR 'd$ _I), #KR@,m8Gn$_[4)˞@^`nrι։2>IG9ߚN,=V0TzW+ P3iS疿Wg_vs(H^0Lgq+[ƴ~^=l/7S]/L?%~C[ڨTj\M::j&{s" fvSw9''W2hUnn?M .)`"컟 `yYAx+͊K0/8#x83EHӶ?Yx`\67yȝnA)x^"Y[ Zh.ޗ!`YWG](/w~yԝ\:n ΃QyVy̥RQz~ /N78E?b1E#g^Zddۯ2L ՋPP"D >V&C PUd[4ecGw;؛m5>ѱ#a_ E'MZр 3Ʋm{~8%ON짍d@h,(zUWz2wrl{ ܕfZ a<aU*Ec](b8+F;hvT=F-xצ)_z$"03};iњ@A/_e23et$ANϞ|l.I7^I`%l>%w&~2D݀G·:oiێTMm[΄9uIfD[}gAHecKX.$ZV_|nWPM˸͉W/&^{&/@IQE {f ӳ[ZRx(͂4*L РtՔuOQvÊpkݠr%9vc$I_Zäju4l2'WKr29"X[uҙpޯ'.3[*f^g2T -! (P3+jR690ex_qؠg~mǦTSۅgjsΖD+@b1#Fĺ}PJwF f;ˏpWR`8ĉJ@2&EEр_~{Ȩ{&i1:\׬\]kzƵ\~"Ҭ,uI=Fnjײ-4?V V]9zoO@˖W:}`խ{4}-$#BZό W?ֿ T$zfQmOY)qq3D$.O{$ ȡ?~*=`# ~gglGwR SuJ&ȃՆp[ \1xcCa$HW',(d2aPJWהxCgܐ6! :G/L.ͣNubLTg *muF`@ E2|,qi0: 6R N CF #qHQE_<6ʪ#Ic;4xD[ojQ6zK;qm񕀒i~K/5˴&ФqmX(b8 rݝY\DP2-&uTioP_$e&L"hRK19Kdc=_$E Bxw>㬁vB8 dT[LFC'@-R4خhEY\4+!zLo1f$2ڳJcIi]ƾEV|V4YQB}-:-(CVk-DO`Q }9JV _?mo璫Miӈ4 ؃B1xպ ȲEMSIjԾ_~ǃM$[l.E;߹n.yGE^вN7$[lUi˜_rue|hSƠHON0ĄpG0DQXm Q )2z?3JeqJ˗scUEj#Qbh$q5[ b{jl{**h]QQw^.-G㠉T)Vq7(΍#Im4\Q) Hjd/ [T₰41Bl,yBk̿?cqT?ht!L@|{HEϭ- OvLĽ"M~ZCz`g CQy^ #؁w&(}b8{ Fע>#Cz(kQٜ+_n>6eG]<+P{/ɋoیT#n94b%exZVmXլc_4R- ܴ O"+_^&u2I l/(7c^?.'luPӔH2up.4T2#oN?5LL ޷CFb6%Huڕ! $?;u{淜Z(4q+6/T8MWc'e-NXݚ<: 2&-dZ(O@onZf*2Q4KyJoѴ4Eqf'◻%u%w­ ivܮ12\'n8>Sng.V*fzs d{cA2Z/)l(A哵TyrH@EƸ>eh˵y `Y=rzdD؈LqVt]͏TYioiBSy$CGEI`B76" 5E {xL~+$=mȟ}Q)Б3^V E@'삎]x gj+*arkmo^@b墯B '4۽ܼe c#da(åyY3xlх\MU4ex9" #AE-Qi_RcVM(p܌l&2"\{np n]isy\8/%'/gm PN=IMw1p[r (uS bQ ˇ VSRp?;a"D!F54<QuXIX~_wVz %( ho"p ٫2:G{$asqڥr+&Ǥ< +$\%?լ\ O15qeGk ^+FQ-WR]Xqu:5SJA*eXC`P aw\6kыF(19 уONp-Z5XQ"nӫ{: )**T2-o|>ʰ?>4URilE>-( ޽./ZLY2<\c? +^m{;1'ݿ *#Q*sMWsjӯf(=Zhv™*wq VسGL}KǏ1AoNj;Okcm.eT?R ܚ:6HW(;fн㔒S_X"N\%=ĭj3 }#'uYɭh^5=TjqZ^S݋[{kC5UR~b\k?"5}E]@M;ߞ ,sUeyym%A8W!@lğhT-A&2,Owʿ;{BZxpIk^.nu[PH. pR$!*{W<=8˞LXJAXm~i$\;|\}* -@ܷkMZVݛMgDe`6T7݌HF睰ih^mR,G2 :QZqsЗ[bUMBҍ7B(s[Cl,kbOKBT83 #d20aȑI3nF[U@4-6˜z¾gh/K*g *$(>Q9oXw"!bV/*N,g%B2LHP&ߜ.IASKҔq@fgZOc݊L ų54 CMKb|ȯ`m%p$wx{Af[T%2 w;TAVjߢZΫe"!Zt+"ǚo:j/N'FD;n`Rd?/J@-{z-;y\ opYc *a :weAG e{Gr:]xcbۧq7CO2v$ y~.Ba]?(O.._qu'. fρ%R)YӋЕEi:sqn7Pwz"ƞWs\dyڂ/l'8Cfmcw VSϨm Vk^elzĄU!?J]ޚmc BU`>{` T0Zj)RRglct"n牪TSUccؔ@eE2URAOFBt0kg&r EҠ%qÅ ur5GP2E3Ϭ2Tht(pU\LۨKr WnQ2 ޙ=-ml-R}&FS#/$"܃lG#)^Kv~;Laݷ$fk'`"bP]y_,*d«qs(j\uұ)ֈail ek2p߱pZ<c6gY1 M?xGHUea˽ VdՓu&׿R_>FV>[D+4įj.\ Gi'0Ā,s]G85ixX@k:Tos (aH Ռ^6F%C]M~4䴕]4CrXR #{̯J$nH1`9.eQ3D{ N % yMro v;DlyLN*^*&\~̐i9 `D7f[9-v YY>y&Xݧ^€cPfmxA:.[\"]O)5JC2աia^8k)j2 .&WtFFB`n\4[g#]0AVCYa حvLi "__.|j"pZ @G(M%(3zl(kg6uPpip^Y ;bܹ`]+æώ|$+],$xjM;,vSM7 GTyk^eet 78"h,Wu.J;5۶XC ~*?T4ϧ& =6[jfTqauukV;J\uhށ97Mc<R?aY2ZZ6FӁ hZc\iH(*TêD58kW$BN߅06 )ޓ牫,6(!1901%]Oz]Ҟl@ 1W :ry^p{IzwJft/o_vڅ#vdsENۺW)l/ζS,!ثҹ[[߬t[n_MVz+/0fm~cP#b(cY׌ܣjܷx]k@;@N<`jJVY:NӍ*q!oeUUgu3Fˬb2 ~vӁ8Ko\q<~,2rkyƉy  O=.i&8($ C;zEø;ġ& P0v^!៙P:vvbN8p$ߙ8JW IvVSꘁgB7%h,7 5UƗ5'jM8 #Gd'-M{xDtI_%7/Z5)* rd%Bx?K&9eF8`FknsOP(,nzGNBp㲿2 -$b<^6J^vLغ({-⊂,nL9d6Ϸgá3G)ϋ$#j1!y_y-&6̱QUm+ ݠG؊Q:wk /!dQI\O<.U<$TT}~x̆w${$ȅD$(g/ 7^L"O-wAS*X7ڢ"އ77*+)Ѣ#O (+ƚژG?ᇓE@-m9s/qRi ^dѿ>%"%3LVn~҈EӬP[2H ?Vw[;K\"!OH3(@4j.Tőas+IxLiK<'^rqzt"G@?"[_Z~":m)͸m\"yvЎ!q p&B$3{49L Gŏ3 ӈU@ }׃OkdR)Ivʟ J v3){>W'x.ۡ~ĴR86 9M URFDtY{zQhWk zfj/3`6gq9|75D¹LKBU7\ܑJQk5wlXx5 g}1J?\|dEHuy1RsylfJx=SKK驹%YxAf=æl|Cٺaf#IJ|-Xzcu_BrY*'K]%V 0&F8j2K܏>'y;8c9XMia+OǰS)$m5ju΃q|2THd1&( eA0W8v> na+^CԊ0?8\6 !E1]}n_Q]63T3͊}@(2]Ʌ8I>BЗ|XgTI\j| T(E7L]K Ғ6log36r|{+&!cؗ7jfUrI'5-(7bA u9i:\FEwbX#=Π'#@{@wy3򧌓Sc ǻQ3,Z/,\lKyIb/}Y#KQ¶zSKziCsUEkO!_嶍I>ȖSq,U(Pop*v=d=fK}iy)a4i4ugeB3i"ZH}Ik[1WB fjT ;KY ]JpBbT^K +쳋aߍGA`;!޷*͙8t13M_[V~(R&d rTKq?A~ΎBeMa =k6X=0WX;{Q mj XGBwGAwWRF”aK ϹE\ z~b$1iS"!K@صSSV`OJHnvRόj$Df _u9j5nzCI]|ɚ#*p lNWЭʷ~W5Ä ,(#ίML@X߼}r cosz ɃւQkjR0 Vz@eN%F/䘫183/7,Jy "jҰ1r-3ċܿ|qS %YsJDla#U# +q37u[܌51F&X5ec8EB^w`T6 q`,;엿CiYM/TH̸RIcb2(1rZ:I}5:{oÌ6 Atu:\4矅4h[\ ֘ž}Uy t1'7=Tj"N0 } i/ܿ7rYt-נʵYn،,@+kЗQ ؊&ՕnU֦aOi?#wݫ>hkW +\n.᫣ɂ>@AG Kv =׉v!UCE \"`9 ,xgZM#43nbZDZLn+$W#$%O"~M 0uyGujThk&b%`ejQ;&R][9L[k7DaVL8P\CeY ,^[6RBfs=s`lc%AGcH07]UH#oc!8M2jIN(fy#dR,ښ°uB^5jAcC\rF!!bFŬ<DHߊ31)HN>'֭OJ QH8Lhu }[FxR䗶N>wmoк9`{b@ctM \Uy-'zb䁨^gy[F$#@3˶(2{~.!ڒ@˜l1m;eFۈJ=CQ$Eª92jQC'1D{9bR &!R`b/'`X?)t5mvYIx O5̿[\:˄uHU4=G+VScTwu{h6s _Ep';xHhTOKʁ#yKҡ^sr[S-/Fɘrѐ8)E hZާFVk EY>Çxcw60XCtM{oxO4(),VHWMAWTo̚Ϣ aރ[Vm*ݩ-%7'b,Xfy0'QgfkmLDVĢR,b5} 0M*n>-RtdlZ$FۍiR~:;< :ͤQ'϶Ds҂C(Tn/;J+8IXa^|\`"ۗoWuEEχjs٠ܦ;e7J.~lJ |@rY'Qhs$'aY~ϰv۪NufnPw;y,צuڔ+(B`ZLLnQnf VsA_E\\ "{L]䢃 9F?{cճʠhGO GJz5PkE?"9f5X0kDis.N𐊞t5^9s{b/@f9,3Vnv]ͽi'+H멳TeG1O[ɲZ7fY!꘢e~|'< -Pe`#JO7? J&@_Á7 @jyb1"U}TertwVKPz8S|\[mP^ HrUf^Hը-YXa4FH(G]b-G~+mL|q)Ʃ+Ǧ(0l䆾 4]+,w߀QM,+z%ҊxL"2CΤtӷ (l42%=_Hgڀ@eY]6Z#+.UEkj{>>ź삱!*FZߩ0cyuJ$j f^ONld]_Hug]Ȏ+ͧ^0@QTFZ;KR+A+MFV;C bKدK2Iܐ孂 LJf$uDsOӽG+lœYlri{? j}m)'"[^;Mճ3c" .0{ ;h`"lMX2Hj+N:-PvQ"5+5Gz \gB0}:=񖙉-h;~1/yovc3_?ZgCY*fnw^اA2Pͺ3:JI"̨ U̥LJ4đ\ɕk7ZM8I(#"r!> b[_>ת6eܡ;$;;QW-6gD(S|= W0>%lC]`CXl#xa/[\EH]X[Õ%?M " 96R~RF*l{/nm p/)cDjELS2W℺~УKA9v,܃JZɯ魱Լy+c1woD@QjP*]i6h" `G <'l@8ڮML~kD ҭ\DTlGi~Q_H9b VN^twj3 g`wR J6(8ͽ&OdaC MMhwCL8}Ev_sݴ;_3eJYC|5̴NWϭGŔݨ|jJvd=G^xd@Q.(YPb;7jv gPhrIQ\;lPgiX0!5{mi@ab PPژ%@:e @5j*v f; ëGFĭ,C![fGG&FS)A@7!"\|7h fA `_- 2pץtzaϪ]>=POkөO;uIPCԚztx%ʰ,cSXLw+tlo9_x]^ iW6:ZspcJط; z^"f9{eY5)c֖l:@j1CME|@ߒzݛmFf+ \E> ٩l˳`u/@tSCu%*[Ez&\1N.qH]x~@N=QtT&iʆn}p)?kJ:ŕ#SNj͊vd"d4l`N4<1DUs^_aH@B^frC[K\S5,Yqo?}jV> [Ky6,dQi%A}䊋AW5$ns0:V]UBi tgGQRC ^dQa `ٷ3[֔<+!^FdeBT ($V; 5;Iy7PgJ~G6mشI&絗-C1C2x]֋CF Å}b(r 8}T=}PEs{mpCSz!yܩP8"y8=>v;P8r<'Ao%G`]yL^>؀pBrqگM0C ?.Ḩ%"Bn,ٶT72~Q1Jrː]㷐(G^݆shz)L)r1B#A,+dND YwnHE!)#oD< \2݅{b銽fԞBɐLFL6h 'j?KeG Yf{}F\N}nw]tP?͸_QI9ϿBZg3j%.ZLnي{b`B! 8FWbeY#,K_&T*j`"[238EGz'l9A?8*g=+~j=mV"Ա"Kr{B4h+%Xh_)0:78&f,\4.YBGYS1.QۣzHL(CO`I/*}|{eްU|9 ^Ͳ:_^\?z9MccC2ƏDMg3+%tcKW(.>,T4T&2Pf])ٻV,a$3iGHŕh81Ww'䋣k,8OoNl:/+sSEeExnifL$5h/P)|w##R^[(![dq)}L}2ˣq F`tapR9̿'. *=[CnOsD"Loŝd<7oHဩدKEbJ LIJww_Cg׋5L^@ܼ)Lg@_&04~(a ;4_mH0M\ifW`ZPZu7bSpVv5$e( C |<7=j:p2P XN$ ;('gz!rR@z!\-U FjYa 38dYn}F){0"mo] Er(+YQㅚZ-9tWaw{*ss`[6ݿ#wo3s &bWM ~˝}r+l'pH.f%fz ;-`Jz]V|qx/ LzmalHSJ/R sŴWoA6@Xb@P^ 3Bi xQ4R[/RqfH--\ @%uE{Nx QP*h7UO mՇ >  J4p|.􀰂k#& Ն=Jcl˫/*8):E5W}azyf>bjaN5I`qx>Vp ,?^'AYa ~RֽG,e#}[>S4/KJm1vڜvSC9 ,Cцi|h=lHY@y5؝mOzi7u_/V>$_1 ]G)}Y򤟹Ey¢ֳܲE">'e%+6(juCDz> \lȾq;9v JR'ih~˺UW$'w k)SiwB{t5 }]w55<{p"G 0}7F,b߾sJ0➺,#!K0*D#$#N|4JfտxF{Tm5PC%9 5&4H]#&B.31=OS "Г v4ȠՔ8 sd`I?d|obTM:lQD%#䞱ve(YmtSXՄAkz`3U{}792‘l++j%-c /^mBF_: # L!l f/zE]7Fnb@CQgjCK2:rr|݄*8l}_aחe m3Bu}%/t 7crnm`u3bLNQ|RQisߦyELźd)tp^;;!ޫR;M <\PR0tL6~^.rś7s^;/` 7Q't@!lt9"1߃}`\'McF|t XxvFpq5%v`JR8r>|#k5d 2Ir9-28Eqi< .⍒ΎNeY+gSy.El}R~C7lnxekdAٛ*{Ž:B/u2jï^%^>$vs!)}KӇ.#AR/<~&d %%nEz~3\'za,&2J}bW)T۔2+r۝Nĥ1${$4+gnխ{xyVA##wh'8Zƿ[xZqQ!sW֞_h.qS(j!ZQ ~cƒTDݴ ח,3E0fN]MĎܷ6dg/cwS]s4ľ0xVȀTe 39>Ufƥ/ZGccDj 4 Tia|EJ){lcvnNGV*hUaiF*wNbOEuYexƋ3fcx/zd =c8 D2U "!+7a9=# yUp,iSw /Ahr`)1=lN'n ʂȈ~:d^]`p8<]os) c!+`FuM1kp@9vʈ0o,b 9`a).ԨRQOVcm} jwR՚GA)Ikc 9i äTkUM'ϟ^-”79 yɵoX&O,{hI{jeDe$M|VȨAfj9SG{]"A.5Vϗj]X>P [Rf*4,7Q#~*tG(oiij)zjH `ALםp^ uRB ;"5z[CC[>H[71eB!,g7'*GqĺN[̈́%s돷69zi@ EL[lW05 XfybQ ~.ArY dh;*^)J^1?Jعh~2,Pu>N(&y&5ZX: P hSQ̡ z6.ض/gĥbW'Ud>>).g=&i:/ߊA VWbڞ.kg;icR0y&Ȣ?f_xμuXǡ07o8cۘM"s8I^de Z x㷌(ў?;:B38yt ؛"'!/J;ݍ/x PeP7=η"DXЍ2٭I~ڽ5mrj.]%Ί!nˠJTZqM7v*g&,ϒ/Eq-B|@mNŀ|%ZMRRcϋ@'PKVE@?:V8ׅp[_@F+&@<wsa=aS `~&D'וDA=jl;Te $$=#`Y{ -C8~x={ Ϛ(&pj1L*o/QnaԐ8VX/,?MV! |+41pMO' Q`(=U߉HH9=Z0F$%͒]bcM)Ef@uz9ցāܝx3RQ~g9&#]FWFA+I~@VKcDHoS|B20Ot,<jJP|a 2*c^K6 7s{ 2FXUmʇjEmQC䯙7蹬F~rj.\hפ;ѭס4Ezzpd~ vK͉|gB9Abc2WC=&E[ED zkkLP|rҌaH0LEEI_WgN³fX* gp"]`2 1# y-a:9նU+࢐<akBh´LwNm6eV#muVYN{&plvg_=&Nwh2xScN 9R(H74t°8%2 Kj3_ad @T#.^;/N͛ϼZHq bT3jWfpc@{=ؐb4`u(-;=ր0^j6)slMxd>Jm E88̰J,͵IutF!w"ko+{Ka*h=yP~;cgNqUuUΤ~(V]a%Ax)v/l)3 ;mY7]+U{MbQ|Vo /HINu\liwϹ>~_$e馣+臄.P {15c\V$UR`j` {LX^ 7AU}S푻H yL~d_m=Q?@{e\?M+䐙g٣>g͊yƉ*S ?J/8qrz''y*L<mj9~`g/S VvZDUZc!WNsVT>o`)Uܚ:HDƃ.U\Q;6}#tvy.&A!`k"a8 ,%)`_'%qVu?=ySf=Y?' ߙMޖ\$rpEUV@̬TX#;aNb$C(ދ瑿>!|ѢpR,v$?F׈aQ́(f ݸsʂQoғkkyyv@e\?rmlE7PS `SR:*҇x%R]t>p O4ghi;f_"xh\mWo^/ dT7i{/Uw  ?<)[REP}ƚܠNsXf]qX(DlKJ81uьK%)!J1M/Aן;`eC_S^%HːtMq_0gM 1n1,.z^tz<-}l|k'gi)[ᆉTh&LcKl >xB;t++ʤ՗nWo蝹Q]pֲ2dcLT˓2Rs_1y2/N C3h_ûu$Ǡ}yuIRD`iS췆/+Q/JTC')]`JF u * E boXA5&aI8[F3N`kAz04E9^j~rŚ-끉sN^Մ;av*J"2K:%ʅͱ{gW2$T fK[29a*#6VpX=dPhAEI+W%.C&ҕD2t {!(ʎJm}BH˴쟍f|fh=p9"ikLXmSb;RnXNCX\]o]ӛ:FR= <ݨ\ǕA So2u<#td5i֩G+>2 UP_/Hjy1%y-qa'RmyY=* ʏo Iш 2Rm 2}6NV82ou_~R{N\O]-P"b6=3 KZp[9=׷נ5"ѻH\U3$!:Й, "-?G"{hR ֕{=x)ϐ!MI3䬬r+@?k8 XQ< dVֹ8{Ua(* } ,yȻL2Tώ*2G ,f1E[04u2DsdC=أSq17B{$@$!_WB-Lt̾íy@#I{ j37ב`}KglEx^m'/&مTA֖-Nk#ǂy4"D%/N`4> s7;Pw nÁs9wvW=ie2cxjv>wó&kXO}yWRPEJl@]8u}v6KJe>#CN [&NqT™W7'mD$`W(b麘I J!zQX-߮0d3Έ'FؓVә2Hj(AY|i3zz`` 5WصU4rRMnv5Ǜ>1fNuśU->Q1FˁM!It%q"  u3W =憀}7t-RJ[<r= Pu}_詀Ѝ2Uzh&;*d+΅CK7w='^-m'ٗᇇR;=U-1J{f2n7o6N}m#IC2 olPriLѨcգ|2t*omU0@ŦW"I`!Fq|NúJQ&_Yp046{pm.^\J?̀e'O{aL5H!ǬQSF_趷u?Qbߔ1\8vm  =>n- 6\ܘd UVJFR-[&6w^7UNv/#ÔMw?z/8ZP-sGuDp}[pOnр] wMң_:;OjMmPa0m- U"zi+{wNޠKdiI<7^TETWn?~3h;,67 ˛%ua,`wsՎԕgjc됰%/?jKZy-Q,TZTiI㙖VU1!kzǜ[y(ƼTZNӊK3KUVi-Qu1)*b/ ۡl8ߧi`b(ʢ$pLTMD=ϑ_xwY%~dk,u_o6\@;L"$_yMd8Pa(ߙ2ay#f5Γ St óhhJ_uLAK@ oš~v}$2U";m1\B@ Gbl}|fi֖Ͳ>@2:\d gmĆzL쒱T[]xJS *C3f5NCi(8uDrn'LO -o% 5>M]Em4iȻIA[ie)TecY{:?>Q)q57 8{&kK+6ħҵqK3kiu(͘ cTc&Y8گWta+t3ce*3O R66-m@TNq)f,:7L!X*&ƙ'S 6QXT?NfHHMo_ bMzZ# R*)9m>YiBpΠqqw-jwu\Ưp" #3 (aSl3R8S!т'('FK)?i_(1]uFZ&8 NMSj[80DGM6PyRQI%eruD`|S]x{&`!Ww6o@)Sa?o1Sٖ],pqqZ6묵8{z>kyT@$zWȮl,4nq ֨UGW~(Lk:i")R#C *{i=AN\_K9HZ]kc^N[ G0),\!4gAD m>[pp5T`!RVK߂@n)zz[ԎEg-~˼F=ztH!-Gy {SfDY`5A^H3W<崔A̟eŢU:R>nt}frt85rL@ ӏXYVĥ;С3Eud,>z6bm3 3=h'*! V;E˘X2\g :,q\^+jXz ;RN1T),FX$g4a~| oS%]KylAطn Aށf?,9wڍw5w!avΨWe#A)My<2ezۇ^o -U99l 0*;<+V+ŗdC¿!XX)Z %jՓ3Uy!rO }AF2Sn$ȸfwSYv$FLc5&+lְ< ' )F@"c o10{-@Z-wGfɚv09<IJX0''Od^)LwE\Z ѠYm1%s6,tu0ZmR:2ߞ')|~t5x+U5bPi{'X%͵Q[MR@ /% sA=  /#3_q?Ksc,V7iݜu^-1NAq.!? 5dB{pbkC wrJp6> HJi~影J:\\0 eXJ0UKcgoV\ᆝjDb=wϾY3w=^tnv>))4د. ]|Dq4Շ,Ewt~'I3wHj@‹Pܟډz#xc3p BnDީ- ,fT=*0h[?Kji6o]6Jg@ϳ>2 .oS!}-a:^p<"7f){>zR 8C JHt!f*~V] qZ8\Q&+^MLS٨L?7dS5 2g׉:%JFZW9EǼrtZo&3/4{J>kweK=\&HP]ec-&hF\_[ܪ^W{#'vNK4[<Դ>I'ߍ^aLO6%Thu3 F 4w@fϲP4gBhIb &JKk$*Ma?bp;?,kCb1Tyˋ{sxzO\ij4eaNijKz+ H)آ$2ʗ5Qz~tZg8y8M|!j >5a&%gh(DV#Èn Z(Y8%Q3ʪw(;n(USH"B,L9Qj_V)ij\O`uo\=,3^4|1OX!f 5aZwA3+i,V_Z\ Rܛq33RO3?SWH gUٚQ:9anc|O@szLjlu82`W~ni s^?G_ Ѣs@gWr~l_j#)+ٵ 8$j-F~ۣ/_B'lv@Iv"^'$i}QUȎws>3=Yڪu)7bQ65~J~N%Bb|`bk5Oq+0|ZFW+68R$PE|VօVZVQu:3@_}2]R@2Tc.[nx+x(Zj{9j^0 Sv߮y(}e~ԛwkS~v0 jGyA.L40:ޥ630+-";_[L)d~ccVį{TAh5P\sEP%IzUS8YjGi?YQs>u2GG' \p x[<贳c{\~Z٧*8osGo1Ns1Gژȳ{M %8c#"P _|ugݚZ *i(Ï=4&ҤzSdžh#;QFxA? ?!CR|*4#4X4=DGK;Dz$^,(˻_/[hd"5 .ai1@9}7,*]ܘ$ݪiiyYe#,N*:B7¸I|j6nB:A#,2nu\V}[f}5NBզy7ݢ.:cvVLOT3>dOyc'.cH'J\tF]1ռÚ5])pz!Bq+h#TXtajUz 4j;g̪4+ǢVeVC($_ 7;p r,U{sI)kSL9JP{F?\p&砇]Ȁ1˶.Wf[m28Jbk=@+4={]jW p6lEl3 a9]$U(*{o:C> W=aIf&⬑6CxgCSѠaAbPBnB ;g.P̨bIFr`P;tNT2t:̜V8lw-y5 n14K6@d)ܙ; 68e5igQ'5eldZGhbk :?YrI7Sv$9[ P6(^QKruʄ"*}w S)a2G7KA?@pS&*VtZ:T<1sO5*)yR#}hvր:UIS2XYZD;;?y{7)R_perLHӜłվON3^~S^#jǸ'eg$BYțg&G}. ˃%TYAhd CϛB k p?'>bN w6wx5v= Y[U_!W0+wN D7'K;~a ݑßbBAD7Q=5?zoejj๏m;:;ߠzk ,Fz=KeELt7)kAQ¤,T8&Ϟ\|xx6P)n8+pTBMAjW?GL(f윙{[[PKς |"qu #pdVȝ)>HdfPfL{c1,k3?#.⾠>(hcm@!|WC |%wS#iUYc1|ՓEU`W BB6Oӫd2Z|~ՋUrwVhd*g?1Pd~{cx% JN]ѥ'J)c11`!o6/)AMY C($M6Dt2\]79dƶq#sS Zg-j 59OԚ *Lu>R~xHjk*FB.ItgTf\BK_ϤJھpTТ!K) @tz M  > @/CaRVcҌ9p*3 ؒ%;bͅ"5u#* B=߻WZ)Zqj?6JM9Fnfh<.ѭfVR& ȀZdnq(U1^yE[2L4'dKmeIoV&R=lU>އ*6WZyn#ϭ!i苮="L OUmn)8 :~AOfMj\$qŪhG@b, ױ1͗Fny12+*S6RQX7zbĖ4J%oˮfZ>R^ϟX1aLkT ξ`ьӊ5p:w$m,U;ܣX ңp*ch%JȢ׷yWÊةqz_voY6-u7߻aQ  I( uV޴0ѣؽl{q]&skB7Z'BhIc(Z#wphBmR2\7bImBO]p~Twgvȯ&*XˡkAc.ʚbn]3/$ҁ &UxQ ej;Gv2wkfaõ /\߭^[oP{؈i[O-('q[aD tkIClCUa攀d.Ahf*oyܽXEb KƃGߣ^G~JaC伋K|yD>V*۹x+|b/s5?XX漅`L(_ `zDL3/$aRSYV:*+O䪆 I]FZɮN;-L]oh֌(Sb7g\o8,`v nݢ@V{F<^^ӴJ˃"LDCT`0]!w!+y[ ((vbm61hN=cNxφώ-Db[y o{;͂EՔBV(02s hS/Mp-6>cxi`j_9_#2P~H%G >VQ)\xlz<}wE(6CDtk~cIK7E KJHOi9ҍAH5DLbl T$4ՠBJ5!/YrУ!3}*CwS^L`V.֚h~ &<2ILǯ]NgG`enL܄ tFx-(G\fFŕ6}\y8\rebq&BE3`~+Dy/Ir޳ 0#~NܾYɎb$R{YLѯ ɕSw^܇T|*mXeT`x$; X/pX!#귝Q겲| c%0hyOjͧnM;  ;1ߪo0t.m0LQֿRykRAko_).=ɼ"؛bwd^MVjM%ÃOfh]AD}a-W0/$S0£e{c?dGq<ОxaKD0kDZcIHd%hMއh,0K7DwR.U%hc~OaK}A:Μ(gg?Ӓ_P۰|N `>Q›kc VA;,)lJ&Q*8޺w&Ӏ4_?Up)>>a-~"ֲvv@ۮ #)K+ NW‘Ѐ:Zn5@~حN#H2ŜQS$|_<3;FLPTw0t}է8nDvyug/)|%xZ+ho-6gf2y'CP6֞K%ÝrDk5;'=A"jBX|=]p6f +}>M @ xKj1?HY+5[Њ&6lxMw~O (d??'f{`.1,*Snܚ­׸jMWXX".)1&+ǷV%7=Aa{HDJ30~c ` i~j2Ao/.JReڎ\iN0ZYCByᔵk=&5f/h^(:(4庒qը~ GHgo.=#c`㮮YroPXJ{ J'R^sƀcwC]?!HyWlЍ$~yȶTnT3Ī/o|\yb~H KczFU.R瑢^rd9:|HX?F+n|Wd>+DOwVF݀U>H_1+P|)4ZBAvdwCJmQ9bU^$ w3 w)R7K~\ ; BގII.f K66voPi06W*,77 RFCNv~>қ96`Y2Rx뱐 Ra/cP4 6&7:r:joL+,֩mv{ؽ DMcb;AuZ[0;, O`{03,ז§A\om:F2C҃ e쁆1]OFnR+-Dr [di-܊uɋH҅^iOсcFZpt)i)v >n9QկIA0gF\:N=p8Ss=񶕱Q=h 1u6QT??2{oG$0٫A3|芈:KV7J/t]L&K~R5)ڤٙ1 qP JbfQR3;7`?m dwy>X6c(ɗX'%+z ;*KԟWhȸU`ߟ-}\[Ċ`A]ΝF:$6Wzu:8o;LD 13k,}I!PSs :3=kkg?v򤨹l%vxkYcs LQ9 *n]@SH!z w({h+^'YU`b4%hTWPc ^W %?.6S P'V4?܊T[ܴe˱_GG餂΋SozT3@_uճ'kdPn4jG~&^sD^f7ki +.12p>ӀtnP#|(=+o\\0+I꣛t/k?ʏzƾ2K .eP%$W1;zqhu8KA靛Lɑ3l҃#i4zMִhtaKA=cQ#ʚhJ$$%sI+XT@QLf ee;SO) Yvj֪ Fi6$ jDrl`Di_[f.iZO:] Qp ډc0kh&XT+iXSFFCfdjDkK%"cEv&1оc{@ >0|鯊&Iē'}؀Ub~gi `yU"kw./?Rφwq9H%6јnTgo:)QFR;EpEY%)VBn9wyfRG2CƷW-|el7YX ;Y|'WDj h,Kiֹ Pς%mB/!`+RgXO~362 3/>Y |VJ{`C  _< DkAW~M5U`G.1k[\(aW %S?c䑨V-d5A\HAڝe:}gͭdIq.̩|!4#-D++T5T_k F?ۧ#9v`/k= +v>C&!-,B`?,uc m/ݚ=ŞO3|isb`aVtk&- " epB3 h"B&%!cuV߷4 "S`)CYgrvԗjnw=G㛁( `FhT ކ)r"LuBN;&rT!kܚ2[d'h?vsWEW̿A-sT*Esf}#MaH &E\(ְi0E|y xm r1eRa@;avܚ5eZKVD=vf&l *ś00ق)]QntV> GCe hէE5 }mlTS_^[{A G>@`}$ Ӂ@FJo@W*0CpT@q0;=`]';lSP{by)U#ƵEmp%6/+~b wEch̪(>3?Em|۴b?jWG<ϖIH= QQrAib*q)w=KԊG,%\VlR =y'W`*%8Y#&hĈ?ۂpb Lkչ|'>Y%hTvF(Jϗ!~@e´QtEmTG!t`pronpEU4 yG֛Tc>Iw&<ߡo> UnDko 5f#jpʩ۰>PݦUk*O: !ECV͗ O~7"[bM/4WM>VЖ!^p=u]y3S{2x ֝+IA{eI*ݲPGYx_ io箖r3 Χ`\XC=SD/T޼lok$X?` ]Pef@S0ےy \9×^L8-,jt 7=Ax`&+U<1WΉCR=.*Ŝ.ʑM4LtbAF-:*S'~Ϡen؎bߜUo"@48<ўS7[f|pWպθEnk=m "䰽a1k=)pU2i(% l-LWO.XzL9P?rgWdGsFLۄID[R7mLv q/m/}]C--C,@gJ ܆k1&p)`s0o ٤aQMZ2'!mhTwIk|`K",7|v=\G./mvb$n7(`MXGb| AP8%xAQ55)-y߿!|vık!*'0|mPcxͱ915jf$$8^ (n(A6I·2.ux+ "cj'c! LG1ѳt# X样yShP\dž 3.hA)w緾xW0d$13H<2WG22c\-~ەC I@LINpL>roH *_ 疂72ZLl=A/v~E V~S5;r5{'/G++i͗EXM9UtG@`d.ܫx\!0؋m9Ir/-ւvׅ3g^n(OY\g]uߘ}<(i}`q `MY#rN pkO&:7/!mњ 1^KSe8AbOR8< mO']%01Iygب"jD^aԞq=byaL)1;h8j+yEVhhД1ؗ]Ţ3)`[یC8#d¢;KU"slK+p|3gvX3a梴TqĆzk6_3b߶l$ 4|ZLBT8@L[{4q!Ud-{O ;:ļh!Ӄg4uۮ,-`vSQc2c~pyOr9XeY\LO m 65I+Ah+)CJ8taBO_0B[ZTT!r^"VK/i[w|B]`XܲFY"g5JGs2#b=p\\rTXs@+{.IV<GD5Ǿ . ))ĕplx˹~zSoWN]T;C}tHp!@7j2|\ !|[[$0{d/#7?LToxgvJYje0{lYa5 7\`RICKX$w`UxdY<%[ꞦInh,(Vs+ԋJƄ0#֔Owϧ2(:br1i|a&@n rT kPKHߍYus/W4?L rT$[>EwK*5}kE{¦zV,{I<,Bzz YE(77#M5jͯ5qfTĄ>F X{{EtƆsԫXĬ1[7XNNKQ7we'iVE/ڶ,VDS 0<efvF!D3iٌj_l:OZTwY]Tk7,.2P~[ eҵv *{隫>\vC>d:+,F%:G3? ٸC1lzFOW5<K:gRK\"x05;T, i6  } /bmӧN\uD;zƇHj[J:~YOL>2[=ъ0{ ȂΖۀ5:y/E04_F.pyE_|,UϛH^R88'93 &ϩhfP?qѷR346.MAv^9РʴxQ^ o5M{&9C^P<U󨷘pz~3|[! S$c+g0)} Ĉ7\i;-O⯒g2v54[riqrJo)O vlOɂe뤅 /vmoH*u*NqJܦU^皍Xg 3qjhZo}sJ|b9NSY pcatcFyyhN<FxKJ݇:7`tKtIj]:4k\`ANFh1e߈J5-w;q[6<&JvC':zIg$4+ճ(I"+9ĐѲDO"=X0{ A%͕FGn <1}jApUɕgJZB u/P`(SYcZW|tQ=_ZH/ВurMxRC-ѧש|?k]sT2m;+8v-͆T1;(%)9k1fwLR9#Xgc v}Qȩ{u)OLQ];ya<ij]x*ɁYxoU "iLSR +R{K^{ѣMאG42' Nl?TmWIzc:;儼kuu9,[# c'eU-R5lEQZ:pb&EWb SP!ՐcӇH.{|7"}^5;ZjXb H{sO,DŽ;?ImtKY5%l_rzx)eH"Q)ޖڏcB~N5p`⍼"0l4Ӕ؅ s-!_IO |4ZnҳES0q^:U]"x- jU^J2%L"N7%uLd)]lP"3М T_K,I:M(;ogr75:HpBW P-<֛+^Uu,Fքͨ)&o23hs4lU{~҂G:rg1GKQ""qwP!;n->î&ۚ \9F4hѥ2=M`?F*G|1G=70KsV" ;nO/w#XVE~Ӗ峊oP,TI ؞:-! k+)%oH (5R#G IgpBgwo5=gL+֦d2z7]/0XpF>}v*2Ie| 'KW:]&kI/WkrQ>6:Uхe_q ÙĔCѥ@. BiR̷W-Rzˣ3{Ԃ2dɌ? >ﷻ\rh9ދRf4 PJ%W!aumnh$@FǙqؽ4Я+30#Hu3$U^D|Edp@=?^=M_K#p r7GlחNwiYw mg9_ooÓ5e8u9QjrF.236#fnl(0y Ơ!Ȍ^lm ,lw"XlsXoWL1&ދE?Dף RJRa6yE7۟g%X5p5Ećc@UCMr'+I!ԁ66 >(&ȉ)]ThJS D(@9l· *~ہq,Wdw!JO_b*˗sOV&DH-5o*'"+#;A>+!mưPƟ9r"Wj U2Nw P`3h?Y_6hW` e9n(Vg*gǖImDIE6 @z }ק g2p.·mnYgv[XO97 `7`˫EZyTg~FZ*f^\y 1LGʺ6W5K|p~T.׏Mӈ);ڿfBW =^Bxazಔ IZ{_ToƷqOlVƿQâ8`̿/%kF<4jB犆/u X$gd[kg@AE:t|k ⿵ ]-}YIav M)=l*EȀ! ucBآ2\E-c:j PZ*7[3;ơ`:-eg8xؘ"o%D3 ։ɵeS1YҮ?+cs*RKk_X#/^闄!1{6[*ٔRbWkGƯN@.JgȎEWCt;8ꌢkd9FDo7S9CMgI#.=2 @:[qHdUӅ#Z1pO$ۏSKܾV7k5_1zJgIV*mZo!"\5~]NzHq "YX`=K5*9@&UǬAdk'B!WlMd\vG]4#@jgto]Gr N]ήsbMh"ckac:,r'w!i~kE-kMPQ0A)v*:>@EV}Bt% K ]a=ۑuڦ,z//x,8g -99]B O`n4J5w&ᅁxA̍2&9),kEZ]#-L&BNYWP+_K3­>U `trA턺w*gϾ!u,{h}=D2܌&R0,Ze_u,s;d\X"YV e?KqNT %+%.Zۈg^g4%n[}o^Z/4ȆSEEcvOa3BI݁9U5x91vK&ŚxV/?l薐\JOEb.K jB jX'r:,V43kȒ>? pis=;rKqG[5w]F$U'XLs0vL{!;+8Dר]|WG)=aۄY$W^l|U џOO,~Ax/l*b=Fy3|O5* c-G։ߑfc;Q~m_>?~tl#'0\Z1* ~M[.c1{K̵xM  ՓR+T@ Y 'J;Ե M' BVw۲FY. }[yBR6oT|pfXSlJṢOu*8x( Y'GȟA tߦ<7Q[PHdI$i)R̔z/t =Fdݒw0zfX0O/-D9%+mBr?g0> 0؅w F8"c۬]&MRK-7 -,?9[kO <=`jԳXI(ǴƩ!Τ4.GTO\_z7l-n"j4-9>Ce-mV1( hDg>.oEdkBm.]5ޝZh͟X2NVD5FrS(tYM/fMC%n s~5OV)P ʥSJZŝo[@2]`k^:!#|bم֮ yO!pqaEZ0ʤtKU1 Dٙ_sm&ä6>^ZYo<$$U<\*ve_*]ްg7 baA.R4 *X #V&e1bDE?f\='t1([20`FxAƸI-\<5c !5r#a"_5M揞2uf]̭718X @[E<׉drѓ'gNb1:;𱪧5DL?}mQF^/%'oVB\D 2. iV e5NDw\rBHG0j( { ?̗ o#.FFp.ZSm#^8RoVo03Y U-ô"uؘ%݌}r#پhoMW Bi6F3}$8|Ϣx;ʿd=6ϕ%<:b܍R.>aeȓmUœMsI a`=[A LLUR*HM64oJԥwL5{Ye"LvpV .gZu:, Vea>=q7'6^#q/Gc =^1^x΃v,fUb1QʨOeo?{y7J2mh [Y9ȗ[:5Ss g][R\:7RdyTgPA9ʇJ'%=_ Jģfɞ[.L= z) t[XH\LVZ5r@o_ dsy W.x׮uTi5LX86]W6?!gȉ?8t¡9ʂy*ٓkvJEOr=⦞7[;N.j QP5/cq\9X2vgɻ{%h"gD1ՃuC8IW|(AZ=)N|eUgB8t9ze2*' pn\`$4::0@@ 3?0O`l= eqyfhz<ԯ݊4_YMg49rBؔ`A/HAuj*)哄edp Pg݊0GlDZ#&R1#%îuն^{%Vr3&{(!B~; xj/H+vPAG;qAnvGkE3KtCTtE"w@ZjS'…CN^CSei5)"Y*  <1B#Ar?jAs'B 1%V3sv4؅zDZ(F׷pk۽&l}$>~j♄$a*S{,fʪs4Tg>p?i\E^ژh׮6^gM6 xޗ M#:0wT~+gmt#Hl5KeE8JQx-5HRB3&2TV-<ӨŌ:nkc=FVQXYr8ZYԖ>nJ% `%Xhfhh/J)M g޻k$ {'Еy9"EAMZ'GqR./٦ $ܽL7Ǜ#Kv1`4rl9U؈Xuw;p^fA>Fq"/\ 2ZO su^l]zڟU0l@k۞]}.%[9Q PDYřw<,73Dr&VO2PR 'h2-B?"Rv֪R@D *t^P*46*"f =,Sϩ8@+-oƮXZC UK]Rex/ :-Wb ~b糿s-6K|)7{62|>@o[FėpqʬMB?.#L8s-2l؈VK˜=xʭtXy.{ZuB;?a>[t_ :46DNL{CEKdTJNyc F^)e#i3~2 R!fHI+!t>["C[7XK 8l! 13rB¸YS!~Pl!oB3IMimsaiwn^b@_A&*{&r}VXkY H2>d'|,x<03elXKk`h̩ 9w.˹ٵ;n#G<r.[DblK='O%Yl[0Y1O' "r&-y<\B!~)Ei~QN~wluC`zzi#D0FܧIfE&R7dNBD߁=axzo2_=ha?<ܔR;8?; ?l1ȓ`0>O:dvqGg䄦oQR?{pz p7zKNu_[ rJ o \R<[WO tU+BHv.(l!% %UW|.f#~> P&, L㫲;ڑhhqbP1M|OAdzж ov!(CJhr|@l0?":c'vaF7NJkaP>z{X.@ 2_Kf_NɈBpѬ}$GL,[v΃'s݅Ta>AU*pS$AuEkQ\^>KEH|+ }O@Ĩ'|-J{O> ~{# OAo\?J:Q?BW3Pݦ!k)\^^;ۨzel0Ci<Р4ϞU~t5+̘b N9d7 g,r4,a^ 7Wl!Ȗn7mr6d1,Iﷺ6}(k4Nr;C)ofC72"zҒ!՗>f9_EqMW T=]&%/md(UK<: K;R3yypT7zu@ }\R.RYvƈ&Ę1 Pk4n]s4+v\*DnܻAMEėF.d׾'fV.a)2﯒#LF ›<+qwˉ=iAAU&,Sĸ7,P3n}6GB–L7Ḃ΂3s y wޭKgC(|N6 ,5V A33)#BSVP$]%`^Mz`wo PAfw ._zYǍ ÀFFy;fp[p|#TtU՟]O.|5OوSte[&ӅhdMz*&cwc6wXV+7&EvmCͰG,}ֶ3{>y)9',ňufPE\ä9sw0!ö*ҞgxPX}n`̻\_wP /5D! E:IY%ܘEdtQdۗe˥ڮЃ\!-ɩ_ǻD_)Mꓴ'h8V 6'!EpzS{۪'?hh:}S;.HsEKUY!Oքu> sGY0mkS9jpyZyJe/9e>u>nIWET*vKy$4 W9,P:HD񟝡xX`t"_dEygR8۟Ka iާ+s^7*>μ_}0q¢!>2Ħz.w r hn?1*&:p\bwXo+,__ e_Nz+=`~7g&LjAEݻ^ ~GTȨp 2^ 7*_' X1Y۰K;*:I *gv47/nq3_eMLrw}bQ4dqckep²aj @<"0D9 P>2˽,D*z~[ڦҩ$#CK3wWm:QЎ Vr:+ R6xd;k=D1$tk1g!;߲֒o']{>W,r]ZxJ)0BCLYGiqyjY6|^% >1\aA>O{qU0@3|fJZK)Ƕ)Tz S:ӶAM8ךXCʼnf'yIH,Jˁ0r\8FJLOu#14溁 *|S9yՐIN ?ʙT,I@C~;FY cTX焗+riq]S&դT0;[V C^n2xAM= vRa˹ɀ`; :k,Df7zƯv\_x)t(?wBw$ >ؙkCE`$,G|+>K?FnQ9_پ%YV:*qH4w9 ]C Hmk>'90/~w|>/|,jԠ!WЦ^MǩfAXkqmfi5j07A I`| (z8w\Ws4BD  ^(*7UF1Jq~j=K Sf(_əMVn9m/|!-?zG*Ǣ2/s5Q/&vtOwɕ1|l8<"r*,:|1b #dDž:rzoն l:r3$'߰!HO^; oЙF-f8j{ޯ̈{/vfҚÙ=ίSB\xpZRn?n8+W@Ctq@|8 !eMuTcM,hJNE+E\(:pş{RSLۘ]=ZIq\#8LnXswrdZFB iQ?'J(lM2"`IAYCg0q+p͠vT5Rk/ys-Cg0%eɚ5Ɲ̧Z> W`4Ff뛐1]k7jwwjWHDn^@to1qJG\Rm6PʙO1 6u\uOq}ECB``hpge M;+j]dW{B? WUFD(/+2<^c$؊3)5aA!1.R;e $t ThhfF=;K;s¥%3{Χg)]'08gKtZR(8S1SoV R0s^Tt)ΆSuQEk|5ՆBOOlbKKIEwbD՗¶& ,EydLY`CHdkYl3KSppLή2jO,]ߏ/Ŷ*S4nT7L{* !5!5puCDX̋E@߼X~,Od?ݮ daR $@su8HwS%nHFU7BG?ǘb$Ioh},ޘq"ACaߒw E_ #ɛ'btN&=rlXy&3ڶW/ ՏT`p׻q C c!$$vcp6XF!=f|V8 htlv撷g"qGO*yW[ߨX CzCӓLY±7'i|ʙ6zYy \+\iOmDGO8/zSݛ?h~ILA02QNl[a7jٺ%LkqKw c^ۙ.];.[H}6\~!p8oYPH$0+'lj#J~.fU^v󛐷\-TVҤ]D ݨ pJ "._>k#W闬,ufU :F/\gfޥƚڸh9R`AA&0rx~_p\|-SnJ۪cBŴ^n022:ڜ\UNf>(&8nE,]1gd)V(8e~.+z-d"Ai0gj j81W7+xTچ6J4m[i9ijuAҩk=@BD#C9Zq 7;۽v{I(QiwJ85p 6JY߃t իK%#q-,6z(>}blhƷfX8);,ѱt%MC?Hq'XG?<]k<7Z'~HP40wЊ9 ̧A\4 EFL\nݣQUflJbc@xhDeH eGgX`H-rݓosaW}#6c6M@J*4-dJj&GWs%mrIs3h3rO;(8( Bf1SGvA }K/WP=n ^˽;)+Dz׬ڰі)`[V/@uҩ&ymji#-O5%wa'b_)0Y,10l_t(AL& 0SC Y LDX;Q|~5^?&4dapˣQ@Uͭ+}=T돜p\Ptb.__ Uԛ:J+(׮-xOFkqBy}O7 {> /Nss]4v9Ƣ~MA3V8JWy5d;Y? r94IZL9 [8'ާiZ>xKdq_`^o}sw<\6 uUޔ0λ]j2MyNS>]w-u(׊( IT (f-ՙJ~_&PiQ3( Vb3Nm6)Z_NDnҭ9O p6HFm*FGVё%D;э61/9vj‡7e]n?1^F(E;PMv[V\_[iMxPukEspeeEE*sO63-·h|فYھ؊xGG o`Mp\]~f>5hPiq1Z@a|Z@i<# Tùx0[otD$t>ƜDwX F⊆_![_G L[s4 nr[gdKL78m7@d69K@LÑiz.hJ '5픗$gZDūk)ӓ@XQGCq_zA)}T\-˭Fe HևH7"O߁*õN{g!J2,3ur,>}' 5oT{}w8H&)ZmtZ)~Zг1yAʏ yƗ:(/`;g ߄[ uJL4n#NYj.5zAd9 UP_nuI4w9KKYlW(P]8Cq$-;ig1fUP:Z{r>}*$mJ>*"}]z4#j[rN stbiHY  Tĭhb %+.4vĄi|T#2 P2k?ܷTGus ߎs$E< L&􌨠0L4R] l ݫAAsp񄊵xYG̮΅[&`i=\B,.~{q:&k. 4a51 b>[*VZJ9QGW[pHDfr<"ж3 T Nܸn`IڪQcuP@ahz4/"}2gԹ A@(NKFj3)sw+_feqT3=v|WpW'HzQEk {5vƵ'5!y?MLEjYYJHj)QI6݈e{IF'x.4ˢu2h벺@ŜcN^n2zUPh-+ft Lil5Igԯ"~Y}sL~f @?HGS@?X57FS|1$4TA͞:kl׳m,צ&l -Fq8hhpl*Fv qHǾzM2PƜ>75iqV+Iq^FN jjm]5t7>EXwqM8!!f7"LxOuIЎD==|cVu 3*8Z՜rHevzaTʝ ;hI@;,\oΞ G,8Fצ Qvk$F_0\NvCb2okQQ?ҁ 5A 43g}ϯ.cMV9sI!cAg{{["E加⎵$R]wwlH1t]B PoBɡJ' G#K;x=AoVa tP%t ޿HǕqvAv0 ü,U|jy۸[_edp4Ǹ'URd9F'NvGG-+>ЭG^UqZꇇQ!I^O9Z75pɺdIP!գP,2=CL>rUPCc*¡w1cAv3}:/}a7yqxWYx &+>k'' _kL%Uשv1W؂w?c8or 3~C_ [Wu/=<``(`lt1z+N ]\~4%f_ynyGT/pϡ)'VE/WrLf46]fyR(~)vx=l+)z$/z"eD+MJn7W b~ 9/Q"&Zr"5 iS$Έ <^JOS3ƕ*)h!5MySks =욊QFaf8bL=W0tHȚ&ݓ~-;/ZDF WR0z4v3u2FdR lYͽgKzB+\T[~% 3#H++eWtS<@PjK&In ZTn5:׈[:C?&MOhPP*4H=0yGws/iB<'x@]֋iи`1~5'6/~"=O1Ot5tgmq뺋(9Ǘg5Ɋr;1 ;o-n@!;*iߞ+·/cGނ SQ R3 domoCU YtOI #2uT%V_b-;V?}̞v{&m@1/1ӬU qӲ,l^y},l8` SpUIr %3BƲ0E (W()0׌kf*V`EoFc Nq!i9ϒpE͸~v͌W>c AԆ@ L8:Q'avЗ3y͖lp|kW\WW){{ٞ E# ?Jw\Q> q ]>'wX?<̤0E DX"W GB~*E/뵕4/te;pczf9{Ne-[t%'^ ,ߙqDS^>(sM:eoyK̉PMbd|# $#.{}|cy1ZEݑ,H`;3Do n[ͭ?8g۶bp p sz๤9}aѦ T.tGnxu u#1?5dWF|ɣ#fIty◴pkNc)fnt/gGbz f̌nK\fjj:K`{|f`bUv%LED/U߄/w",Nu&C fm^F޹mAN Ad N 85Mm^jz~J3dgu5.fF 5u!;TBBYQ_Ol-ϰq!&MuQBWګK&7|Y= Ę^0Wtx8{[Q?LrD-ŽȮ{U\P՗(۽*c? 3Ŏ؍ ;IC@3Ur)W!׬؄W~"u"u2{kGZo0/ yv>4ofarzk +eOL>XנшXRKwI%r6WgE Y]=9R|#gQp'=ŞaB6ScJ#s$rVud^Ŏ6X b\&9@;PP_G_g#zdiՊ>suoY!d;X-\#JR(T4yx(#j'<!1"|EDR(X\r Y@>.WnǽW9- 0怉Ip=szT=VٓGwVdb|/<0QB3bSiYQ-LJ.`/7-zKg+L1+}jzw(G b!c+>t!x{8|ad#0-1n X 9S, =v5ۏc,q-1] cƨ2FYA~(iO9Jgy+ðX2s!ݕǸ#:_$=F(cmF #1[dD9|(̍ L \ 01@@iApe{ML{t ~ w$Z:K}$*tM3c&F3È! z1@|2ya0wqxӪ289t@!_ڇABJ?;&]cm}tY. ]r^_L,IJR/Tb!\͖}dڅyGiϜXe~ _t>'mT`v(p3 Fz145ARg:2qJO'ŃlZQв;k?AkL,]ꏍel9$&ONeMn$57%!~2:A_oӚҽxvgtuNH: QpfY9?D';7I;bVFQϑYlN( eRĴh[sP0޶CWLB?B1xǹ/ Fr״YH><.uGh=Ac8_ ٝYu4R߼:N]/[B/WQ6A"N1+,fwbZ7߲!{$ MP)DZ Q]'n|H,(FDX۞ 19aփ:QBMzֱ^SEuQb$]Wbf@M-9M ]^^1AJn1vgxgyleE.VCX<9?7-2-MIq$7 Õ׷^ 4p#dvܠT/fJ'JJD<¯sZ;U]ۂ"d_w-޽io= sU ;. T#uW dRUF*3Nsio% Du5z$!Ʀ)^_Ph ZW&PMb7$K~JV?Re U[DY!+ѱY O?ŃL j1NpysQ@"j4<p#d&NYR+uΗ",+Z}s,׎-\< F[tSuR;4]T?AB9DVɡB>j,Ez8HTmc/W Qym|VP)^UJ<ƶ`ۿxR:&tfkWZR /H8JR]SEKw/#Hw(qX)&a qfaZX/.wGRToωse+=H;0f\xQl8thȰ1"oE%7(/"}"Fє'z+]op-j4o x x.0x@/~x;p ީKO1+0ESi^YyafdL % $/TJ_:ہ&t?'s8ڳAgYr6c刨mbs٦c>?֭q5S,0` h-^ .M6>kG>/k8rǐ$#^e`d7KKq`t&ytkz*U.?'dP{OV3g:DaoA iV͏gV&Z 'JXrsY'fGc򧫋y:2qlsvPOrsEF{Tseo@Ujτqzs$2V+zx3&\3ܩj= 3 Hfop ";[Ee"_R=լq?'`efX1ݣAf;,Sh1׸ҩ*~b 0W qJD-^kj?8P^!xD:Z>Ȩy"*lj%aHb 0D"_+Sfmk׿ϞdOFKCk!tw\Oi CЪ(If8-k=÷\,]**+q-z֤_eiHwHhzK=$E,`Mm+^|Ƞ WG@J|3\QkrD?f{*%B]X*cjBFF| }w8*Nxq9LʣЍi5g9Yv+^g̚+o8orL|=EI8!ˍKlQ`@d"akW7N_U>TL{q)s rDЏX'C:pȲc@w9;+ O:qyCG%@ FZioKB!/$& 0M_[z}~wЎ[Ғ{"ޟ#qݏ7:>)SBܹ2s¢X1LQKN=С)DOeX2 AK?oA*\*tE$렆M T{X ;0oMVZ"vW]@BCheUx˷UpjC҂x'A/ĬXgHz嚨{V /W'Uyf5oQ4ИW'M3R/{ 7TaHoJ `o2;Xe^rϊP RK&P^Anl[Z3--Lx`IvmjmUgӆ_VM8,@S';a֌ApwOl BQ8"Iqh~0{ɐ7I,8)Lyd43˥W#;ť}pGiZDƦSvb)^#gvvCr V4^B:*0򮈏~JfxAG W1Hg51Ğf K#aȝe3`-z +g 9Cӟ;9=P׼럴4}l#rI=_1au}^~!B<bqL-[ҝ%aBvM=7dhat1J-XϫԚ;q'AEQ_=;PfS gʫ 8I< <2vi RljMw;=1me t-0}fu@*8W P@%Qo?ɘDy4ũxC*%7U#c&<"hbn ݤ6"F6 ia֍e8J#=,}t(]ǽwzZ+ 'Bzlaz(HG-h=7]&qF|'Vve*!+rr]03~Jm'sѹeSꈗ5׸wE/P= qA#R}xwDŽhP^ e<*pu-U-l?^ U/4Ʈ^HzdnḞb 8NT# CCzR'`/WN\Zy[<UI:0Rc5%%-MN4*q;%ł81-#95kp*E36ԢIeIoRMHD "sߜ]WYWivpwBe}^x꒛ O͔9 ݚQ$-TEcN$=f?L_e ~N07d\EzYG?[\vbav~b԰Y7MwBGy @>ׁ1Ͳ BKb+m>gថj8?F-uH&9)!_}WK))gZ/**2dŊ1TF70{Xua:+CN$_B/^wRw Ko"~#<D0qX?\pMhUd[T1 lDBNک-FdjNQcŅ:,˳*5 /d;`OHJ*tIGr;{F87Sp?c#̆j}f-Kjg!mMT_ F(OZ@umPPVc8ӜM;B|} SбSRnAxT6M.ѫѻg:ޅx WvhL?V#C.. nI6P&B"JT`ѮxGųVxZ;qAe:7I0 1QC-*ܢ06|ɛxjTyFn٥&v<z:!< z+h^" lvﴟzP8mGIduuao>e[;T t_ŒB%T~U\w6__rN0׳|(Q] 5` Y]!6JFv'XFV-CuLjв墪O =,\ml7/V7DFQ!X勺2R;ce RFa+jF̙LRjO@)-ȃ8fAgQO󬿅M"# +mͤa4]OLLUS{yԣ@ YC:tp&@ 4t$̍]KM|9q)P^wc:P5p[\TerT= >E6hxGGOWSEn(}5%{'@}%mKK;2+.ۅ[hc2Ms~^.W0kF<7ic8"Ym/NsuY'Rn +ziךqQqRb;GC [U3{3jhcśX}bE2P5bL8'c j0o񌐆R}Aְ#Zi:qpXÍz6S7ݡT%8F"韣Dbl"Ns@%" ΞЊ&7P- l)a}Y{!gBʺid-wMs:#Mzlji֜>!SЀhf"A14xFOmKN¯ e ?VUC H~pL9MJ>X85Ue$g͆ |F3[\+M^:^I Ne*d|2l*&A~f0X7_8,>qpiQw;VsC9}n#/hnN?TMM/njm!TG`0U>2\ZUjDXlGAY]T>;Y FɊ;,Rtyc34ghؠ<㦴R% D)j ٷe;p".nh5(61XGM"&iG_5쬾Ru`»ؽzҐ\/e# "²/pMyR>@6bKuػ u>e3SŚy#m!qIj)G)O#\!eBrF2A'*CFё>hiM2ʊ 1Y/H32 c:Gw%}Ӓu SzK֩elUR]̮ErjS0GT<'L,J\ϚT IRhK1 cp#7zˠ. yoI;Fvq1YlƻG'Mf8lEd-0Z,z:4@;gJɫͶe"tIH`G$~تXSaOU^̟R?Naa ?gX 얌vrD&Gx_މ;y88 Q.f.{(*{lIq6"9m"r,`ֱ7%GG8#0 sTǕ;fiD;z9KI~z͢mhg˗67JG.x\$2b pzMZ.$rT%PB en2A%u,̋nRx7^Wr =MR0g@6 J Wr*z"EFi'0~~͵rX\9._}9)M6 )Cr?[bPِVi mZ¯)S[p: R_D4xǹ@GpiI"@eBsޫo\ьAׂW83\Xs*(a̟v-CǑaep e=|(*'4~Pj9=N\L5 n ~Cʂ60"e_!Zw_KG<]Λ P#,C-ju? Du&J}@"GqPr E(#a$cz%HR^1q /U5i B~9 BϷNM0WU^"YAP"\I{W/r|k}s緑%ZdDn[JNe}9ݒ%S$KQveح06ةaIB%m~)z@յ;B&@Wjt>5v:AڱTfw27hҨ'tv/5*W,.u09}!6/#.iS=-TOh5QmG 2AovN9 _{b_kI>#ݶ(`IZ+cL;غg#E_CL8L< c Q`/?"P:vDL4,dFW2PJt6b;vN}IAnY*U I^oac#?-8 &๱(td7_ A$yL*A· 6gS@": pwX!dzwfB*2I.['2+9B\jok[L`;ۍa#_ydm/Sc̏V]/7U:ԕDŽp 'ݦ^v7_>\kiAq%T5}6ːyaGkr[o81OSztgdq&Q<6HKuWNA}LHrd"k-bl*I= cHۼCx.ꌕff?k!U:CZCƂv3~nxmTF2I Pr=w FbܾfofoX478 v|O-Cjˍ,t*a~"ii4+* -f-fy.X OfB-=Nd.zA;L]ʮH&5w2A|:fN1eY8Vs)ד뺭?Y uC! J~;'H#2VM1m_"D*R9Ӽo=JS< ׌n_wF@䕡WmBaU|We24EzoՇ*2(ÿ nr^FR?1LĐ1|y{I=;뚡25).&칢U%مd)-jJkПhN!{4ϲE-W _yX+Fuݡ,A$ǜDܝB}YʁmͅP;(쏢rmI,'. &b.1N/dɎSˊ\g(,/(& 3Phcy^dC{E#S/J =L YPKB(G>i<> @U9KZM#Rb"ⳖP_k8V.ѠZy/{ZPYd:^ݨ\H:H_);)~I%#UĘK!԰&$cpB|vR%T~R1wŊIt|׫r, շ S+ul@tyl` Ռv擒Ӵ]`7@2 :ױZ2y-zx!oGP:{ac%LK5ˣ>'fPVuFԢKp(IfU*z9%ݻR`2+ l8^)mB'\#-Cާ^y>%+Wዼ>Ŷ p~\>%|QZxDK@ # YQLj|@nZxM?#vcTtl\nYmp됪MhB8E]K9>wJƄ2|p$fUm5%lx5 )fš ;B_xp@dBrz>I3PoLڢT#eÂKDKaBlG8?Ic&s=~~Y‡/iI(M7[)=j1ϻvYZ$@vb$O(i*nQ1UVř7S!>]4XgT~K~qj]y1"P)LDDK2FW`r 픓IR'Sy:"뻞ܽ΄Ad"ܓ8b-ءG~8MU5z1ӿ&}dMMsY霭OSIXkU}H(~ 1q OH9>ٶ4r~^ h ]=9GUڢN3Җ>&S@nHN?K[/ZU=)|yr|p꫉3 p;Q,x~'ɠ5 !$ jR+Bka EǿgΨ\(5iP}4xPb[w{eC|C:[V7#6 5D'[<S[xȺs`ZusʨZzIJsp+gܡ~;~8scIg>^Btc@e"]_I[QC[קݹxgH"'H+<*0k Ey(p*$?[ӂ*"ю"mrYjja?|qQH-vWV{?} Kv M(a^m`kF*.ȎhՈzA(JG`D^7DOIJ7Mq,K Āۜ ,Pc_j<--?k;Ȋ,?&H%τENv +wV&0+'ex$ >D*es^r:¶n[E%[d-+ U;C \:K^23 :.7Jr n2=h/cO.K:nrPkU3Ƶl4X0"G 1(Ph'P?l}A,= S1 YN0|(|qBm3y~d6k +}4vUY L=FYB5/4m?IM`bp*|lYy9>ه1J3T͍9BO|k7JǺ*e>6f/0uf&D%'8\/ |r౉:3es^4[Z`25Ow1'efdǛzyP!iֳ9!9 ak֝Am1Δ"v.L /v5n+Eƥz-,ic5` NP;eq sJ=Me2"?3r %=q <߮ TfND HVƂ&-d.rQX_Y,j TB.O}RF? h݈N6d@R_l/g~DW+]2'K%T")>S˄Il Y'h'n$ oL=tn8'j͞,? zy4̉"JAL5b(Nv 4f~@YQu wo:^ ϾS:$FsJ{OigNBMStpL0WN5ߑiU=T:BG6Pq?^K=aRda:"5}r >s9eFRLOkY]LT'+3fW;&(C"?2s`58a+8:/2w\NǛkOiD4kq;oagr|3X#\j JlwÂтN=I?ۍ>(GHcCAI8 wC:z+o9)UK)dV~'L>{6 V IO"^8iÎzbf*i:ixLzEMЅzqb>վko]T3)DU!A8 rJ$ 24K/1]_*ړjBxЩWcrv2[l-繐b=pLzd. `G[Zg)ۖ}ء5>XmKZjC9ƜJ{_ } Es%v83XҁDk;84b#1>"agWEY(cb-2wZџ߅%ndn:4.X|"EO`#FcBD@~XM:h;sĄ୰}ZT,cB]nl 22t/بNQU-ݸN=I>ஃKkΨpWn8Vͤ~MױM<$$r;Ak0WZN 7çA*ߩXBȔMg̑FQB59{}ขScH3|+&HǙl<{od1F:x.-Rg}`Ƴ 4c)֕h,btwmoK )0*[n";X SC&X<`K{PΠp^r#li)>9C@JftlHw^&ӄb3/ONlTy9I}jm B[60͊ؾuˣM\[+ˣEAиx?ΔY`XgrOlbE^/S>F9W]906Q;(h4-S'ߥ1珃u ~Q9pMPiw˶Gu֧?LrfSq[S*mH ժ_2m`Gޒ5؝մ@)FHKpRoeB{\s@M-4(y1h﮶TOGbiujT4//)i_wKCծF_ `CWA (Yֻ(<([Ń:gv5ݭm#vRaK:#Zxo4l NP$׫ݶ,/br8MsEE)'s_ wC·s@P[wA1mZr8hbFj_Rm)xb Ugz.47iqϷsqm]-%>2x`LKMx0;,%N'@}5ϴdw;'5S_a(b$yæbk}QNUC%y@v_(Ru?|Y;93fO =S W7G 4A3lQ;Щ3]92"\RQ |VK@'@ioe+?1#@@޽"W=2@xiYJI$%s̸ Fpח=ѴYݤM^%6NlG@n%Ԣ}*pz͗RwXanvO3Ҿ3UjhrИi%"Bs ey1jX.pIc$20M@`:.^wuM@js$LV7P1/X~zM KH;-/٥W&b9e5G0)br}ۆ % ?3B4ZB}mTr3,xmسEhIy]HڪDs$,FpK{V0; yY2In"}jmN_nVZ-b/ x5šGC/Æ꯻{43QHֳ?Վ9JK0owek7Vag ;&K)ZB2R] N O=C̘'[{(sh2a͆g~Kk T+1USnϗ݉Fm [l]>&!/(ڮ4|SR;#ƞcc,tYc:3 K>EGm#JL@r4VxfNAF*b’;>w 6Y豈llVO2LX]+4*]0OT-*n_%RMSOd%?QTZ()P7c!Bef ƺ/!%7d:: MSaBHuu? dePDy5)/&gy{üma;K mX4W.S:JN;ZXdPb)^pk)nO`CÛ JVN5tAa2}y7e^XS5f0T[D f>fQ<^9)n.Z=L#mbj|Id!y}rnN͔ED}a1AGC 4-GBAM_nBd7y,DjĨzd[yS_nDo0t+$nq7 %8־ riBL۳قCs@>,δ֖;+3t(YɑR\TԄK?05 '#^Z)}ʮѩxC𶓏FWu._v2t/DͰx\h;HR3PYYgOdD{mPʽBPQ'zCt_4(ęI8g7GA8"&QyUn`?ESxh9V;@0Gd.\1} 12֘ x^sJd|H/@[9Ƃ;Ki ٥F*B[l ӫJ%z-%z fiS!&Z 1x>3p_Teu_'NjFY9 4tG_ңKlnQ]l@8rR}uPMKe:xZP,Z;^b9Ȍ-~0L²m@D8(#O+JAZtf۾VYE/WIvTP4sct `#)MxP0PTdUXmXjSGl 8_Ya¿sM~   mҺ-CIc(ȼChm٭w+^y̿xX=t)-#YӻcG&IFK&DG`L63cmLôgj՛jO?W(Z3'5 0K`(/n?s}Qa9¦!Žx-ܯ'Ĩpv8OK{N۫ѱ0E͑ōַPC1YG>dv;G d|ċz1`]yyAj-vIbX)'#u&Qإ K"kaN&n͕\e#v ~y r;xPee aUDWEYpP%)22=yպ0F4kح( ;Ъ6,iX#N2UP2&(dvAxn|VfRL]S~i78ـQBـٱ>KA{tp-D-FU$0ݫrt,wouF9KG0_z7ܞ'(=s}gjHsV^' f_f˶tYux x%Ux@Mx- `:o7L}Wy:KAriK}7JUNX*#2E,O%(@]p-RmapJEQ3>#| DC~-TU򛙇]BL` q7hxrɭ/8U6+%zZ\deq䯽C-(%9M查N~-w M+M,Zeeo:5`2f_ǜrB’mI`]JX m2AStVh|}ٖ$ Nf|]Gm)̥pp03jwM]CL&rڝ~6ԳȻJ2 a?3 &=|}=/Mao<ÿ`D ǙŠu8,9ఎF$NZM`YJOc<+rjK3֊y9dAOuQx4>Qg`}iBBa" >AqcAt/ SfXRfkVt9v|5uA.LBO^SBMdiܚߣ@c!;qh["3 5KcUkgZWl"wtrBݗ 9h]%4BDXm~`%D>ʓr6}laK9DU@†>!ѧTezzg"Km5[~EBYGN uLC05r$0R2ҍ4֞*oLE* Ljghĝ$a]~E+2wNС FyC0̵Hx:ia Z@0Mb+Op\(Eگ#M8}`~2Gg&XgHdzYv%M[|FzEHoX٪i8/PWEj?o;b>ɪrTwLUh nyzOl#*(cM84 4y|6{p_GfHY/]~5-)e!p.Sq CM<3?+Vp2ڦiM0!P /lݨ G*3?[5rtnֳu7Cm瓂 Yc7UܰjZ޹=Ft^A:ӉT&QJg& *QMZ?-/l!Fids{aZog 3~rTe@F~B(=nŇ2-zB eI]FtIc5 {p`I=z?<[T 1Zi=PR_!#|NݍHF(bmNu5(?1 ۠[6 M@ڪ,Zk:+$/4O>0 YZmultilevel/data/tankdat.rda0000644000176200001440000001011212750500645015462 0ustar liggesusers]i\U3tOϞLD$A 1`ٳQVC "ʦe-","* ˢ*)JJb$L2LOL_vhQu~S}y{ιsޛʊO_9=cXߺa"^'f&]iW_rks*R%!.D !h")DJ!h"+DNf!B*DBt)D-!b 1OBt @_8P q!B*B! qp! )!>-g8B 9!>/đB%B#ıBqKK8^8Q/ %! q_+B,)B,T!* !V J q_t!!B)YB-9B+yBbkX'BB\ ķ q q qb Q+RMB|WB\-Fk q?'B čB$B"O q] qw !~)]B-=B+įOx@xH-BZ1!~'BU/ݧn<}gZ+^sJvbN^n_p#Vk{1jWw3;֎↖W_jWog^i^޿ޫ{JV(iQoj[fS\ժۿv|]QzfV9[uߠj|fzG;ծkLJP 7mjLJ%e9smz|;϶n\|KZL ]׷GGv}UWGZEߵmԮh(hԮk7g9*ۧ:壝_Dt;ӿoϨ?(}HvV}ӮhOgzhG{Un/SOj Ϗf]֎KMmj/vwmjo>B˴#m??ho3- X{'iG#wWnx/f0Iv#>#>#>#>#>#>#>#>#U0c#~軥VYAb,wT\﫠pV=A 6PAJl>kc I4% o|㿂^73 Iz?mh;@MkUaI;DHb%Ⱦ ] ù2qrn4Á __㘃DZ urxe;8 γWȟ^ cRu//PO`CCVB[ǣ^  Ob8 ;z'Aߋ~1Ж!ȵm >uQ_Z(l!gg } AΟoe+BWeo1K} ݻ_&[ȀcZN:I<;}Mm6$p9Eveas hcIg!d`O4Яgǃ,W9 -b1=L,s|}:sA.ǠfN@&e<&h?-^Ry{O*kb$'׫ б\kYg/5Z~__7}~ zMkvⰃ鯸Fx>3C[ x~^9[OO~=aVݵL &Ȼ3m2/VҰ8.z33d8c=Q| |0iMos  F3:㺢'a;}u2s#܂1吀̝⚣ej>rm;ln:9qf2ȍE f K򓇾a?sL/G qerާt]7ߟg1? }Ƭs'@_h{a%'qNߜ Shgx'cB!S2=;!6=vszBN8:^2qcF1c+ k/ŤKc|3=>L=p< $d!2~O0yH: }"qvM9׿[kn뮠z}q/Kb.e5]kӐy_ c>蛙3>rc s/ꅬ'SC-N';MͿOQGjfgCDOCnMJ?Bw&0V=xM@u&3gklozNqkc0k8J`#gO`ݠkyYti!yA8|PY3sq3 iюqy+p~[5 /dڄ,^38eĵcǢ}C#h[=9c0&gȱΚ˝}rMJ_(6>-Џ !kօ>ż)sO{Žd Fq}kƀ8i|uƭc|6~;9tYGM@8ͯcepι\8Yȇw}q| a|X4scޥ-r=S~]@ -׹a~n|\ɼyvs-7>s:ڛ1~/ss>f-y'kIe͇A cg;ٻ5E?Z xO`j >ײh[oYY mP;M <2 u?ے{e?7 8gņ gw>q,sL-'rˠMgGϏo>g'3>5 ʸ|/m6`9 6A?J2ofMap/U@&}Mоg7_?0pwzE1e Srq}5{.1 "[b2+< >;c73+^j|a|NЁe|̜ų)&Ǭ@F2.a,;,R&țZy:bz8xa7̡Zwb;_"?;sY*_v5A'C3) ȋuN%Xo \ l/63 .ffwhqhg\s<xr3&,;6wo֗xsOk1~ʶZ45kn[Wsy?g ZGAGϱg[7c_̧CC0Nd}Nz!|児ow`Yk h6/%ڃo6ue|mc]I}Z_G1wt|_}%~9z'cًE~ȸ1כ>4k֥=NȯIhyqxVCfZav1x^KqXxc+NڋV_x1nՋo3;`51&F1-',5BDY LL9hXsa%&Yp8 )X@V0Gs apiQ | @@h> 01_©"a|9Ě H5I5#jRKM%Ev K',հ8;{"s9d)ǕXVt8 `f/#NFS 01,` K(2ddC$Smultilevel/data/sherifdat.rda0000644000176200001440000000075512750500645016021 0ustar liggesusersVN0QM| " Y/+nDFw>1Ͱ,kCB/$_6A6B(->|P>X!U:O B6׋|T٧x)5P@-h@+kf҃Ge;0Y #`{KaTy}J$?Wy f^GƱ:ԹqukP/bM-Oؿ]h6T~`zߘO\@G=vuQ #g+q#Dy($H!,gADS`NNt ua4iCgH4O~|*!HF'ݕAtUئUAZ[ ueh{b: B e)ЮJ{*=RJOTzҳ?5K ` OBdYm =œ~7~W k)1I ʷ multilevel/R/0000755000176200001440000000000012750500664012642 5ustar liggesusersmultilevel/R/multilevel.R0000644000176200001440000007764612750500664015173 0ustar liggesusersad.m<-function (x, grpid,type="mean") { NEWDAT <- data.frame(x, grpid = grpid) NEWDAT <- na.exclude(NEWDAT) DATSPLIT <- split(NEWDAT[, 1:(ncol(NEWDAT) - 1)], NEWDAT$grpid) # Code to estimate AD Mean on scales if(ncol(as.matrix(x))>1){ ans1 <- lapply(DATSPLIT, function(Q) { if (nrow(Q) > 1) { mean(apply(Q,2,function(AD){ sum(abs(AD-eval(call(paste(type),AD))))/length(AD)})) } else { NA } }) ans2 <- lapply(DATSPLIT, nrow) ans1 <- unlist(ans1) ans2 <- unlist(ans2) OUTPUT <- data.frame(grpid = names(DATSPLIT), AD.M = ans1, gsize = ans2) return(OUTPUT) stop() } #Code to estimate AD Mean on single items ans1<-lapply(DATSPLIT,function(AD){ sum(abs(AD-eval(call(paste(type),AD))))/length(AD)}) ans2 <- lapply(DATSPLIT, length) ans1 <- unlist(ans1) ans2 <- unlist(ans2) ans1[ans2==1]<-NA OUTPUT <- data.frame(grpid = names(DATSPLIT), AD.M = ans1, gsize = ans2) return(OUTPUT) } ######## ad.m.sim<-function (gsize, nitems = 1, nresp, itemcors = NULL, type = "mean", nrep) { OUT <- rep(NA, nrep) if (nitems == 1 & is.null(itemcors)) { for (i in 1:nrep) { OUT[i] <- ad.m(x = sample(1:nresp, gsize, replace = T), grpid = rep(1, gsize), type)[, 2] } } if (nitems > 1 & is.null(itemcors)) { for (i in 1:nrep) { OUT[i] <- ad.m(x = matrix(sample(1:nresp, gsize * nitems, replace = T), ncol = nitems), grpid = rep(1, gsize), type)[, 2] } } if (!is.null(itemcors)) { nitems <- ncol(itemcors) for (i in 1:nrep) { SIMDAT <- mvrnorm(n = gsize, mu = rep(0, nitems), itemcors) SIMDAT <- apply(SIMDAT, 2, cut, breaks = qnorm(c(0, (1/nresp) * 1:nresp)), include.lowest = T, labels = F) OUT[i] <- ad.m(x = SIMDAT, grpid = rep(1, gsize), type)[, 2] } } cumpct <- cumsum(table(OUT)/length(OUT)) lag1 <- c(NA, cumpct[1:length(cumpct) - 1]) TDAT <- matrix(c(as.numeric(names(cumpct)),cumpct, lag1,1:length(cumpct)),ncol=4) TR <- TDAT[TDAT[,2] > 0.05 & TDAT[,3] <= 0.05,4] ad.m.05 <- TDAT[TR - 1, 1] estout <- list(ad.m = OUT, gsize = gsize, nresp = nresp, nitems = nitems, ad.m.05 = ad.m.05, pract.sig = nresp/6) class(estout) <- "disagree.sim" return(estout) } ########## quantile.disagree.sim<-function(x, confint, ...) { out<-data.frame(quantile.values=confint,confint.estimate=rep(NA,length(confint))) cumpct<-cumsum(table(x[[1]])/length(x[[1]])) lag1<-c(NA,cumpct[1:length(cumpct)-1]) TDAT<-data.frame(dagree.val=as.numeric(names(cumpct)),cumpct,lag1) row.names(TDAT)<-1:nrow(TDAT) for(i in 1:length(confint)){ TR<-as.numeric(row.names(TDAT[TDAT$cumpct>confint[i]&TDAT$lag1<=confint[i],])) out[i,2]<-TDAT[TR-1,1] } return(out) } ########## summary.disagree.sim<-function(object, ...) { out<-list(summary(object[[1]]), object[[2]], object[[3]], object[[4]], object[[5]], object[[6]]) names(out)<-names(object) return(out) } ######################################################################################### awg<-function (x, grpid, range = c(1, 5)) { NEWDAT <- data.frame(x, grpid = grpid) NEWDAT <- na.exclude(NEWDAT) DATSPLIT <- split(NEWDAT[, 1:(ncol(NEWDAT) - 1)], NEWDAT$grpid) if (ncol(as.matrix(x)) > 1) { ans1 <- lapply(DATSPLIT, function(Q) { if (nrow(Q) > 1) { mean(apply(Q, 2, function(AW) { H <- range[2] L <- range[1] M <- mean(AW) k <- length(AW) A.WG <- 1 - ((2 * var(AW))/(((H + L) * M - (M^2) - (H * L)) * (k/(k - 1)))) if (M < ((L * (k - 1) + H)/k)) A.WG <- NA if (M > ((H * (k - 1) + L)/k)) A.WG <- NA if (M == H | M == L) A.WG = 1 A.WG }), na.rm = T) } else { NA } }) ans2 <- lapply(DATSPLIT, nrow) ans3 <- lapply(lapply(DATSPLIT, var),mean,na.rm=T) ans1 <- unlist(ans1) ans2 <- unlist(ans2) ans3 <- unlist(ans3) OUTPUT <- data.frame(grpid = names(DATSPLIT), a.wg = ans1, nitems = ncol(as.matrix(x)), nraters = ans2, avg.grp.var = ans3) return(OUTPUT) stop() } ans1 <- lapply(DATSPLIT, function(AW) { H <- range[2] L <- range[1] M <- mean(AW) k <- length(AW) A.WG <- 1 - ((2 * var(AW))/(((H + L) * M - (M^2) - (H * L)) * (k/(k - 1)))) if (M < ((L * (k - 1) + H)/k)) A.WG <- NA if (M > ((H * (k - 1) + L)/k)) A.WG <- NA if (M == H | M == L) A.WG = 1 A.WG }) ans2 <- lapply(DATSPLIT, length) ans3 <- lapply(DATSPLIT, var) ans1 <- unlist(ans1) ans2 <- unlist(ans2) ans3 <- unlist(ans3) ans1[ans2 == 1] <- NA OUTPUT <- data.frame(grpid = names(DATSPLIT), a.wg = ans1, nraters = ans2, grp.var = ans3) return(OUTPUT) } ################################################################################## boot.icc<-function(x, grpid, nboot, aov.est=FALSE){ if(aov.est){ data<-data.frame(grpid,x) #fixes data because code below uses the first column to select level 2 B.OUT<-rep(NA,nboot) ngrp<-length(unique(grpid)) for(i in 1:nboot) { #The code below creates an empty list, selects a sample of level 2 #units and then goes in and samples level-1 units for each level 2 unit ROUT<-list(NA) rgrps<-sample(grpid,ngrp,replace=T) for (k in 1:ngrp){ ROUT[[k]]<-data.frame(newgrp=k,data[is.element(data[,1],rgrps[k]),]) dindex<-sample(nrow(ROUT[[k]]),nrow(ROUT[[k]]),replace=T) ROUT[[k]]<-ROUT[[k]][dindex,] } ROUT<-(do.call(rbind,ROUT)) tmod<-aov(x~as.factor(newgrp),data=ROUT) B.OUT[i]<-ICC1(tmod) } return(B.OUT) } if(!aov.est){ data<-data.frame(grpid,x) #fixes data because code below uses the first column to select level 2 B.OUT<-rep(NA,nboot) ngrp<-length(unique(grpid)) for(i in 1:nboot) { #The code below creates an empty list, selects a sample of level 2 #units and then goes in and samples level-1 units for each level 2 unit ROUT<-list(NA) rgrps<-sample(grpid,ngrp,replace=T) for (k in 1:ngrp){ ROUT[[k]]<-data.frame(newgrp=k,data[is.element(data[,1],rgrps[k]),]) dindex<-sample(nrow(ROUT[[k]]),nrow(ROUT[[k]]),replace=T) ROUT[[k]]<-ROUT[[k]][dindex,] } ROUT<-(do.call(rbind,ROUT)) tmod<-lme(x~1, random=~1|newgrp,data=ROUT,control=list(opt="optim")) temp<-VarCorr(tmod) Tau<-as.numeric(temp[[1]]) Sigma.Sq<-(tmod$sigma)^2 B.OUT[i]<-Tau/(Tau+Sigma.Sq) } return(B.OUT) } } ################################################################################## cordif<-function(rvalue1,rvalue2,n1,n2){ zvalue1<-.5*((log(1+rvalue1))-(log(1-rvalue1))) zvalue2<-.5*((log(1+rvalue2))-(log(1-rvalue2))) zest<-(zvalue1-zvalue2)/sqrt(1/(n1-3)+1/(n2-3)) out<-list("z value"=zest) return(out) } ######################################################################################### cordif.dep<-function(r.x1y, r.x2y, r.x1x2, n) { # # This function tests whether two dependent correlations are significantly # different from each other. The formula is taken from Cohen & Cohen (1983) # p. 56 # rbar <- (r.x1y + r.x2y)/2 barRbar <- 1 - r.x1y^2 - r.x2y^2 - r.x1x2^2 + 2 * r.x1y * r.x2y * r.x1x2 tvalue.num <- ((r.x1y - r.x2y) * sqrt((n - 1) * (1 + r.x1x2))) tvalue.den <- sqrt(((2 * ((n - 1)/(n - 3))) * barRbar + ((rbar^2)) * (1 - r.x1x2)^3)) t.value <- tvalue.num/tvalue.den DF <- n - 3 p.value <- (1 - pt(abs(t.value), DF)) * 2 OUT <- data.frame(t.value, DF, p.value) return(OUT) } ######################################################################################### cronbach<-function(items) { items<-na.exclude(items) N <- ncol(items) TOTVAR <- var(apply(items, 1, sum)) SUMVAR <- sum(apply(items, 2, var)) ALPHA <- (N/(N - 1)) * (1 - (SUMVAR/TOTVAR)) OUT<-list(Alpha=ALPHA,N=nrow(items)) return(OUT) } ######################################################################################## GmeanRel<-function(object) { OUTFILE<-aggregate(object$group,object$group,length) names(OUTFILE)<-c("Group","GrpSize") temp<-VarCorr(object) Tau<-as.numeric(temp[[1]]) Sigma.Sq<-(object$sigma)^2 ICC<-Tau/(Tau+Sigma.Sq) OUTFILE$GmeanRel<-(OUTFILE[,2]*ICC)/(1+(OUTFILE[,2]-1)*ICC) estout<-list(ICC=ICC,Group=OUTFILE[,1],GrpSize=OUTFILE[,2],MeanRel=OUTFILE[,3]) class(estout)<-"gmeanrel" return(estout) } ######################################################################################### graph.ran.mean<-function(x, grpid, nreps, limits, graph=TRUE, bootci=FALSE) { if(bootci){ if (missing(limits)) limits <- quantile(x[is.na(x) == F], c(0.10, 0.90)) if (is.factor(grpid)) grpid <- grpid[, drop = TRUE] TDAT<-na.exclude(data.frame(x,grpid)) x<-TDAT[,1] grpid<-TDAT[,2] NGRPS <- length(tapply(x, grpid, length)) OUT <- matrix(NA, NGRPS, nreps) for (i in 1:nreps) { TOUT <- mix.data(x, grpid) OUT[, i] <- sort(tapply(TOUT[, 3], TOUT[, 1], mean, na.rm = T)) } REALGRP <- sort(tapply(x, grpid, mean, na.rm = T)) if (graph) { plot(c(REALGRP, max(REALGRP)), type = "h", ylim = limits, ylab = "Group Average") lines(c(REALGRP, max(REALGRP)), type = "s") PSEUDOMEAN <- apply(OUT, 1, mean) lines(PSEUDOMEAN, type = "l") PSEUDO.LCI <- apply(OUT, 1, quantile, 0.025) lines(PSEUDO.LCI, type = "l",lty=2) PSEUDO.HCI <- apply(OUT, 1, quantile, 0.975) lines(PSEUDO.HCI, type = "l",lty=2) } else { REALGRP <- sort(tapply(x, grpid, mean, na.rm = T)) GRPNAMES <- row.names(REALGRP) REALGRP <- as.vector(REALGRP) PSEUDOMEAN <- apply(OUT, 1, mean) PSEUDO.LCI <- apply(OUT, 1, quantile, .025) PSEUDO.HCI <- apply(OUT, 1, quantile, .975) OUT <- data.frame(GRPNAMES, GRPMEAN = REALGRP, PSEUDOMEAN, PSEUDO.LCI, PSEUDO.HCI) return(OUT) } } if(!bootci){ if (missing(limits)) limits <- quantile(x[is.na(x) == F], c(0.10, 0.90)) if (is.factor(grpid)) grpid <- grpid[, drop = TRUE] TDAT<-na.exclude(data.frame(x,grpid)) x<-TDAT[,1] grpid<-TDAT[,2] NGRPS <- length(tapply(x, grpid, length)) OUT <- matrix(NA, NGRPS, nreps) for (i in 1:nreps) { TOUT <- mix.data(x, grpid) OUT[, i] <- sort(tapply(TOUT[, 3], TOUT[, 1], mean, na.rm = T)) } REALGRP <- sort(tapply(x, grpid, mean, na.rm = T)) if (graph) { plot(c(REALGRP, max(REALGRP)), type = "h", ylim = limits, ylab = "Group Average") lines(c(REALGRP, max(REALGRP)), type = "s") PSEUDOGRP <- apply(OUT, 1, mean) lines(PSEUDOGRP, type = "l") } else { REALGRP <- sort(tapply(x, grpid, mean, na.rm = T)) GRPNAMES <- row.names(REALGRP) REALGRP <- as.vector(REALGRP) PSEUDOGRP <- apply(OUT, 1, mean) OUT <- data.frame(GRPNAMES = GRPNAMES, GRPMEAN = REALGRP, PSEUDOMEAN = PSEUDOGRP) return(OUT) } } } ######################################################################################### ICC1<- function(object) { MOD <- summary(object) MSB <- MOD[[1]][1, 3] MSW <- MOD[[1]][2, 3] GSIZE <- (MOD[[1]][2, 1] + (MOD[[1]][1, 1] + 1))/(MOD[[1]][1, 1] + 1) # print(GSIZE) OUT <- (MSB - MSW)/(MSB + ((GSIZE - 1) * MSW)) return(OUT) } ######################################################################################### ICC2 <-function(object) { MOD <- summary(object) MSB <- MOD[[1]][1, 3] MSW <- MOD[[1]][2, 3] OUT <- (MSB - MSW)/MSB return(OUT) } ######################################################################################## item.total<-function(items) { items<-na.exclude(items) N <- ncol(items) ans <- matrix(0, N, 3) ans[, 1] <- labels(items)[[2]] for(i in 1:N) { ans[i, 2] <- cor(items[, i], apply(items[, - i], 1, mean)) ans[i, 3] <- cronbach(items[, - i])[[1]] } OUT <- data.frame(Variable=ans[,1],Item.Total=as.numeric(ans[,2]), Alpha.Without=as.numeric(ans[,3]),N=nrow(items)) return(OUT) } ######################################################################################## make.univ<-function (x, dvs, tname="TIME", outname="MULTDV") { NREPOBS <- ncol(dvs) UNIV.DAT <- x[rep(1:nrow(x), rep(NREPOBS, nrow(x))), 1:ncol(x)] FINAL.UNIV <- data.frame(timedat = rep(0:(NREPOBS - 1), nrow(x)), outdat = as.vector(t(dvs))) names(FINAL.UNIV)<-c(tname,outname) FINAL.DAT <- data.frame(UNIV.DAT, FINAL.UNIV) return(FINAL.DAT) } ######################################################################################## mix.data<-function (x, grpid) { TDAT <- cbind(rnorm(length(grpid)), grpid, x) TDAT <- TDAT[is.na(grpid) == F & grpid != "NA", ] TDAT <- TDAT[order(TDAT[, 1]),1:ncol(TDAT)] TMAT <- tapply(TDAT[, 2], TDAT[, 2], length) NGRPS <- length(TMAT) newid <- rep(1:NGRPS, TMAT) OUT <- cbind(newid, TDAT[, 2:ncol(TDAT)]) return(OUT) } ####################################################################################### mult.icc<-function (x, grpid) { ans <- data.frame(Variable = names(x[, 1:ncol(x)]), ICC1 = as.numeric(rep(NA, ncol(x))), ICC2 = as.numeric(rep(NA, ncol(x)))) GSIZE <- mean(aggregate(grpid, list(grpid), length)[,2]) for (i in 1:ncol(x)) { DV <- x[, i] tmod <- lme(DV ~ 1, random = ~1 | grpid, na.action = na.omit, control=list(opt="optim")) TAU <- as.numeric(VarCorr(tmod)[, 1][1]) SIGMASQ <- tmod$sigma^2 ICC1 <- TAU/(TAU + SIGMASQ) ICC2 <- (GSIZE * ICC1)/(1 + (GSIZE - 1) * ICC1) ans[i, 2] <- ICC1 ans[i, 3] <- ICC2 } return(ans) } ######################################################################################## mult.make.univ <- function(x,dvlist,tname="TIME",outname="MULTDV") { NREPOBS <- length(dvlist[[1]]) UNIV.DAT <- x[rep(1:nrow(x), rep(NREPOBS, nrow(x))), 1:ncol(x)] FINAL.UNIV <- data.frame(timedat = rep(0:(NREPOBS - 1), nrow(x)), as.data.frame(lapply(dvlist,function(cols) {as.vector(t(x[,cols]))}))) if (is.null(names(dvlist))) { names(FINAL.UNIV) <- c(tname,paste(outname,1:(ncol(FINAL.UNIV)-1),sep='')) }else{ names(FINAL.UNIV) <- c(tname,names(dvlist)) } FINAL.DAT <- data.frame(UNIV.DAT,FINAL.UNIV) return(FINAL.DAT) } ######################################################################################## sam.cor<-function(x,rho) { y <- (rho * (x - mean(x)))/sqrt(var(x)) + sqrt(1 - rho^2) * rnorm(length(x)) cat("Sample corr = ", cor(x, y), "\n") return(y) } ######################################################################################## rmv.blanks<-function (object) { OUT <- lapply(object, function(xsub) { ANY.BLNK <- grep(" +$", xsub) if (length(ANY.BLNK) < length(xsub)) xsub <- xsub else xsub <- sub(" +$", "", xsub) }) return(data.frame(OUT)) } ######################################################################################## rgr.agree<-function (x, grpid, nrangrps) { GVARDAT <- tapply(x, grpid, var) NGRPS <- length(GVARDAT) GSIZE <- tapply(grpid, grpid, length) if(min(GSIZE)<2){ print("One or more groups has only one group member.") stop("There must be at least two group members per group to estimate rgr.agree.") } NREPS <- round((nrangrps/NGRPS), digits = 0) ans <- rep(0, (length(GSIZE) * NREPS)) for (i in 1:NREPS) { ans[((i * length(GSIZE)) - (length(GSIZE)) + 1):(i * length(GSIZE))] <- ran.group(x, grpid, var) } AVGRPVAR <- mean(GVARDAT) NGRPS <- length(GVARDAT) RGRVAR <- mean(ans) RGRSD <- sqrt(var(ans)) ZVALUE <- (AVGRPVAR - RGRVAR)/(RGRSD/sqrt(NGRPS)) estout <- list(NRanGrp =length(ans), AvRGRVar = RGRVAR, SDRGRVar = RGRSD, AvGrpVar = AVGRPVAR, zvalue = ZVALUE, RGRVARS =ans) class(estout)<-"rgr.agree" return(estout) } ran.group<-function(x, grpid, fun, ...) { if(!is.null(ncol(x))) { GSIZE <- tapply(grpid, grpid, length) ans <- rep(0, length(GSIZE)) if(length(x[, 1]) != sum(GSIZE)) stop("The sum of group sizes does not match the number of observations" ) for(i in 1:length(GSIZE)) { GID2 <- c(1:length(x[, 1])) SAM <- sample(GID2, size = GSIZE[i]) ans[i] <- mean(apply(x[SAM, ], 2, fun)) x <- x[ - SAM, ] } return(ans) } GSIZE <- tapply(grpid, grpid, length) ans <- rep(0, length(GSIZE)) if(length(x) != sum(GSIZE,na.rm=T)) stop("The sum of group sizes does not match the number of observations" ) for(i in 1:length(GSIZE)) { GID2 <- c(1:length(x)) SAM <- sample(GID2, size = GSIZE[i]) ans[i] <- fun(x[c(SAM)]) x <- x[ - SAM] } ans } summary.rgr.agree<-function(object, ...) { Table <- data.frame(object$NRanGrp, object$AvRGRVar, object$SDRGRVar, object$AvGrpVar, object$zvalue) names(Table) <- c("N.RanGrps", "Av.RanGrp.Var", "SD.Rangrp.Var", "Av.RealGrp.Var", "Z-value") object$Table <- as.matrix(Table) object$lowercis<-quantile(object$RGRVARS,c(.005,.01,.025,.05,.10)) object$uppercis<-quantile(object$RGRVARS,c(.90,.95,.975,.99,.995)) OUT<-list(object$Table,object$lowercis,object$uppercis) names(OUT)<-c("Summary Statistics for Random and Real Groups","Lower Confidence Intervals (one-tailed)", "Upper Confidence Intervals (one-Tailed)") OUT } ####################################################################################### rgr.OLS<-function(xdat1, xdat2, ydata, grpid, nreps) { # # The number of columns in the output matrix has to correspond to # the number of mean squares you want in the output. # This function does RGR on a two IV OLS hierarchical OLS model. # OUT <- matrix(0, nreps, 4) NEWDAT <- cbind(grpid, xdat1, xdat2, ydata) for(k in 1:nreps) { TDAT <- mix.data(NEWDAT, grpid) Y <- tapply(TDAT[, 6], TDAT[, 1], mean) X1 <- tapply(TDAT[, 4], TDAT[, 1], mean) X2 <- tapply(TDAT[, 5], TDAT[, 1], mean) MOD <- lm(Y ~ X1 * X2) #print(anova(MOD,test="F")) TOUT <- anova(MOD, test = "F")[, 3] OUT[k, ] <- TOUT } return(OUT) } ######################################################################################## rgr.waba<-function(x, y, grpid, nrep) { # # Create Matrix and sort it by Group ID # SMAT <- cbind(grpid, x, y) SMAT <- SMAT[order(SMAT[, 1]), 1:3] GID.S <- SMAT[, 1] X.S <- SMAT[, 2] Y.S <- SMAT[, 3] # # # Create a matrix in which to put the random WABA elements # ans <- matrix(1, nrep, 7) # # # WABA random group loop # for(i in 1:nrep) { # # Generate a random number and sort x and y by it # TR <- rnorm(length(X.S)) T.DAT <- cbind(TR, X.S, Y.S) T.DAT <- T.DAT[order(T.DAT[, 1]), 1:3] # # # Create a Matrix in which to put the WABA elements # tmat <- matrix(0, length(X.S), 6) # # # Split up the x observations by the Group ID and make WABA elements # TX <- split(T.DAT[, 2], GID.S) TX.M <- unlist(lapply(TX, mean)) TX.L <- unlist(lapply(TX, length)) tmat[, 1] <- T.DAT[, 2] tmat[, 2] <- rep(TX.M, TX.L) tmat[, 3] <- (T.DAT[, 2] - tmat[, 2]) # # # Split up the y observations by the Group ID and make WABA elements # TY <- split(T.DAT[, 3], GID.S) TY.M <- unlist(lapply(TY, mean)) TY.L <- unlist(lapply(TY, length)) tmat[, 4] <- T.DAT[, 3] tmat[, 5] <- rep(TY.M, TY.L) tmat[, 6] <- T.DAT[, 3] - tmat[, 5] # # # Calculate WABA parameters and put them in a Matrix format # ans[i, 1] <- cor(tmat[, 1], tmat[, 4]) ans[i, 2] <- cor(tmat[, 1], tmat[, 2]) ans[i, 3] <- cor(tmat[, 4], tmat[, 5]) ans[i, 4] <- cor(tmat[, 2], tmat[, 5]) ans[i, 5] <- cor(tmat[, 1], tmat[, 3]) ans[i, 6] <- cor(tmat[, 4], tmat[, 6]) ans[i, 7] <- cor(tmat[, 3], tmat[, 6]) } estout <- data.frame(ans) names(estout) <- c("RawCorr", "EtaBx", "EtaBy", "CorrB", "EtaWx", "EtaWy", "CorrW") class(estout) <- "rgr.waba" return(estout) } summary.rgr.waba<-function(object, ...) { T.DAT <- rep(0, 3) object2<-data.frame(object$RawCorr,object$EtaBx,object$EtaBy, object$CorrB,object$EtaWx,object$EtaWy,object$CorrW) ans <- data.frame(RawCorr = T.DAT, EtaBx = T.DAT, EtaBy = T.DAT, CorrB = T.DAT, EtaWx = T.DAT, EtaWy = T.DAT, CorrW = T.DAT, row.names = c("NRep", "Mean", "SD")) for (i in 1:7) { ans[1, i] <- length(object2[, i]) ans[2, i] <- mean(object2[, i]) ans[3, i] <- sqrt(var(object2[, i])) } return(ans) } quantile.rgr.waba<-function (x, confint, ...) { object2 <- data.frame(x$EtaBx, x$EtaBy, x$CorrB, x$EtaWx, x$EtaWy, x$CorrW) names(object2)<-c("EtaBx","EtaBy","CorrB","EtaWx","EtaWy","CorrW") ans<-apply(object2,2,quantile,confint) return(ans) } ######################################################################################## rtoz<-function(rvalue){ zest<-.5*((log(1+rvalue))-(log(1-rvalue))) out<-list("z prime"=zest) return(out) } ######################################################################################## rwg<-function(x, grpid, ranvar=2) { NEWDAT<-data.frame(x=x,grpid=grpid) NEWDAT<-na.exclude(NEWDAT) DATSPLIT <- split(NEWDAT$x, NEWDAT$grpid) ans1 <- lapply(DATSPLIT, function(Q) { if (length(Q) > 1) { V <- var(Q) if (V > ranvar) V <- ranvar out <- 1 - (V/ranvar) out} else {out<-NA out} }) ans2<-lapply(DATSPLIT,length) ans1 <- unlist(ans1) ans2<-unlist(ans2) OUTPUT <- data.frame(grpid=names(DATSPLIT),rwg = ans1, gsize = ans2) return(OUTPUT) } rwg.j<-function(x, grpid,ranvar=2) { NEWDAT<-data.frame(x,grpid=grpid) NEWDAT<-na.exclude(NEWDAT) DATSPLIT <- split(NEWDAT[,1:(ncol(NEWDAT)-1)], NEWDAT$grpid) ans1 <- lapply(DATSPLIT, function(Q) { J <- ncol(Q) if (nrow(Q) > 1) { S <- mean(apply(Q, 2, var,na.rm=T)) if (S > ranvar) S <- ranvar out <- (J * (1 - (S/ranvar)))/((J * (1 - (S/ranvar))) + (S/ranvar)) out } else {out<-NA out } }) ans2<-lapply(DATSPLIT,nrow) ans1 <- unlist(ans1) ans2 <-unlist(ans2) OUTPUT <- data.frame(grpid=names(DATSPLIT),rwg.j = ans1, gsize = ans2) return(OUTPUT) } rwg.j.lindell<-function (x, grpid, ranvar = 2) { NEWDAT<-data.frame(x,grpid=grpid) NEWDAT<-na.exclude(NEWDAT) DATSPLIT <- split(NEWDAT[,1:(ncol(NEWDAT)-1)], NEWDAT$grpid) ans1 <- lapply(DATSPLIT, function(Q) { if (nrow(Q) > 1) { S <- mean(apply(Q, 2, var)) out <- 1-(S/ranvar) out } else {out<-NA out } }) ans2<-lapply(DATSPLIT,nrow) ans1 <- unlist(ans1) ans2 <-unlist(ans2) OUTPUT <- data.frame(grpid=names(DATSPLIT),rwg.lindell = ans1, gsize = ans2) return(OUTPUT) } ########################################################################## # Following functions pertain to simulating various agreement values # ########################################################################## rwg.sim<-function (gsize, nresp, nrep) { OUT <- rep(NA, nrep) for (i in 1:nrep) { OUT[i] <- rwg(x = sample(1:nresp, gsize, replace = T), grpid = rep(1, gsize), ranvar = (nresp^2 - 1)/12)[,2] } cumpct <- cumsum(table(OUT)/length(OUT)) lag1 <- c(NA, cumpct[1:length(cumpct) - 1]) lag2 <- c(NA, lag1[1:length(lag1) - 1]) TDAT <- matrix(c(as.numeric(names(cumpct)),cumpct,lag1,lag2),ncol=4) rwg.95 <- TDAT[TDAT[,2] > 0.95 & TDAT[,3] >= 0.95 & TDAT[,4] < 0.95, 1] estout <- list(rwg = OUT, gsize = gsize, nresp = nresp, nitems = 1, rwg.95 = rwg.95) class(estout) <- "agree.sim" return(estout) } ####### rwg.j.sim<-function(gsize, nitems, nresp, itemcors = NULL, nrep) { OUT <- rep(NA,nrep) if (is.null(itemcors)) { for (i in 1:nrep) { OUT[i] <- rwg.j(x = matrix(sample(1:nresp, gsize * nitems, replace = T), ncol = nitems), grpid = rep(1, gsize), ranvar = (nresp^2 - 1)/12)[, 2] } } if (!is.null(itemcors)){ for (i in 1:nrep) { nitems <- ncol(itemcors) SIMDAT <- mvrnorm(n = gsize, mu = rep(0, nitems), itemcors) SIMDAT <- apply(SIMDAT, 2, cut, breaks = qnorm(c(0, (1/nresp) * 1:nresp)), include.lowest = T, labels = F) OUT[i] <- rwg.j(SIMDAT, grpid = rep(1, gsize), ranvar = ((nresp^2 - 1)/12))[, 2] } } cumpct <- cumsum(table(OUT)/length(OUT)) lag1 <- c(NA, cumpct[1:length(cumpct) - 1]) lag2 <- c(NA, lag1[1:length(lag1) - 1]) TDAT <- matrix(c(as.numeric(names(cumpct)),cumpct,lag1,lag2),ncol=4) rwg.95 <- TDAT[TDAT[,2] > 0.95 & TDAT[,3] >= 0.95 & TDAT[,4] < 0.95, 1] estout <- list(rwg = OUT, gsize = gsize, nresp = nresp, nitems = nitems, rwg.95 = rwg.95) class(estout) <- "agree.sim" return(estout) } ######## summary.agree.sim<-function(object, ...) { out<-list(summary(object[[1]]), object[[2]], object[[3]], object[[4]], object[[5]]) names(out)<-names(object) return(out) } ######## quantile.agree.sim<-function(x, confint, ...) { out<-data.frame(quantile.values=confint,confint.estimate=rep(NA,length(confint))) cumpct<-cumsum(table(x[[1]])/length(x[[1]])) lag1<-c(NA,cumpct[1:length(cumpct)-1]) lag2<-c(NA,lag1[1:length(lag1)-1]) TDAT<-data.frame(agree.val=as.numeric(names(cumpct)),cumpct,lag1,lag2) for(i in 1:length(confint)){ out[i,2]<-TDAT[TDAT$cumpct>confint[i]&TDAT$lag1>=confint[i]&TDAT$lag2 Description: The functions in this package are designed to be used in the analysis of multilevel data by applied psychologists. The package includes functions for estimating common within-group agreement and reliability indices. The package also contains basic data manipulation functions that facilitate the analysis of multilevel and longitudinal data. Depends: R (>= 2.10), nlme, MASS License: GPL (>= 2) NeedsCompilation: no Packaged: 2016-08-03 23:43:59 UTC; Paul Repository: CRAN Date/Publication: 2016-08-04 11:45:37 multilevel/man/0000755000176200001440000000000012750500663013213 5ustar liggesusersmultilevel/man/ad.m.sim.Rd0000644000176200001440000000727512750500646015124 0ustar liggesusers\name{ad.m.sim} \alias{ad.m.sim} \title{Simulate significance of average deviation around mean or median} \description{ This function uses procedures detailed in Dunlap, Burke, and Smith-Crowe (2003) and Cohen, Doveh, and Nahum-Shani (2009) to estimate the significance of the average deviation of the mean or median (AD.M). Dunlap et al. proposed a strategy to use Monte Carlo techniques to estimate the significane of single item AD.M measures. Cohen et al., (2009) expanded these ideas to cover multiple item scales, ADM(J), and account for correlations among items. The ad.m.sim function is flexible and covers single item or multiple item measures. In the case of multiple item measures, correlations among items can be included (preferred method) or excluded. If item correlations are provided, the MASS library must also be attached. In the Monte Carlo simulations conducted by both Dunlap et al. (2003) and Cohen et al., (2009), 100,000 repetitions were used. In practice, it will require considerable time to perform 100,000 repititions and in most cases 10,000 should suffice. The examples use 1,000 repetitions simply to speed up the process. } \usage{ ad.m.sim(gsize, nitems=1, nresp, itemcors=NULL, type="mean",nrep) } \arguments{ \item{gsize}{Simulated group size.} \item{nitems}{Number of items to simulate. The default is 1 for single item measures. If itemcors are provided, this is an optional argument as nitems will be calculated from the correlation matrix, thus it is only necessary for multiple item scales where no correlation matrix is provided.} \item{nresp}{The number of response options on the items. For instance, nresp would equal 5 for a 5-point response option of strongly disagree, disagree, neither, agree, strongly agree.} \item{itemcors}{An optional matrix providing correlations among items.} \item{type}{A character string with either "mean" or "median".} \item{nrep}{The number of simulation repetitions.} } \value{ \item{ad.m}{Simulated estimates of AD.M values for each of the nrep runs.} \item{gsize}{Simulated group size.} \item{nresp}{Simulated number of response options.} \item{nitems}{Number of items. Either provided in the call (default of 1) or calculated from the correlation matrix, if given.} \item{ad.m.05}{Estimated p=.05 value. Observed values equal to or smaller than this value are considered significant.} \item{pract.sig}{Estimate of practical significance calculated as nresp/6 (see ad.m).} } \author{ Paul Bliese \email{paul.bliese@moore.sc.edu} } \references{ Cohen, A., Doveh, E., & Nahum-Shani, I. (2009). Testing agreement for multi-item scales with the indices rwg(j) and adm(j). Organizational Research Methods, 12, 148-164. Dunlap, W. P., Burke, M. J., & Smith-Crowe, K. (2003). Accurate tests of statistical significance for rwg and average deviation interrater agreement indices. Journal of Applied Psychology, 88, 356-362. } \seealso{ \code{\link{ad.m}} \code{\link{rgr.agree}} \code{\link{rwg.sim}} \code{\link{rwg.j.sim}} } \examples{ #Example from Dunlap et al. (2003), Table 3. The listed significance #value (p=.05) for a group of size 5 with a 7-item response format is #0.64 or less SIMOUT<-ad.m.sim(gsize=5, nitems=1, nresp=7, itemcors=NULL, type="mean", nrep=1000) summary(SIMOUT) #Example with a multiple item scale basing item correlations on observed #correlations among 11 leadership items in the lq2002 data set. Estimate #in Cohen et al., (2009) is 0.99 library(MASS) data(lq2002) SIMOUT<-ad.m.sim(gsize=10, nresp=5, itemcors=cor(lq2002[,3:13]), type="mean", nrep=1000) summary(SIMOUT) quantile(SIMOUT,c(.05,.10)) } \keyword{attribute}multilevel/man/rwg.j.sim.Rd0000644000176200001440000000734112750500657015330 0ustar liggesusers\name{rwg.j.sim} \alias{rwg.j.sim} \title{Simulate rwg(j) values from a random null distribution} \description{ This function is based on the work of Cohen, Doveh and Eick (2001) and Cohen, Doveh and Nahum-Shani (2009). The function draws data from a random uniform null distribution, and calculates the James, Demaree and Wolf (1984) within group agreement measure rwg(j) for multiple item scales. By repeatedly drawing random samples, a distribution of the rwg(j) is generated. The sampling distribution can be used to calculate confidence intervals for different combinations of group sizes and number of items (J). Users provide the number of scale response options (A) and the number of random samples. By default, items (J) drawn in the simulation are independent (non-correlated); however, an optional argument (itemcors) allows the user to specify a correlation matrix with relationships among items. Cohen et al. (2001) show that values of rwg(j) are primarily a function of the number of items and the group size and are not strongly influenced by correlations among items; nonetheless, assuming correlations among items is more realistic and thereby is a preferred model (see Cohen et al., 2009). If item correlations are provided, the MASS library also needs to be attached. } \usage{ rwg.j.sim(gsize, nitems, nresp, itemcors=NULL, nrep) } \arguments{ \item{gsize}{Group size used in the rwg(j) simulation.} \item{nitems}{The number of items (J) in the multi-item scale on which to base the simulation. If itemcors are provided, this is an optional argument as nitems will be calculated from the correlation matrix.} \item{nresp}{The number of response options for the J items in the simulation (e.g., there would be 5 response options if using Strongly Disagree, Disagree, Neither, Agree, Strongly Agree).} \item{itemcors}{An optional argument containing a correlation matrix with the item correlations.} \item{nrep}{The number of rwg(j) values to simulate. This will generally be 10,000 or more, but only 1,000 are used in the examples to increase the speed.} } \value{ \item{rwg.j}{rwg(j) value from each of the nrep simulations.} \item{gsize}{Simulation group size.} \item{nresp}{Simulated number of response options.} \item{nitems}{Number of items in the multiple item scale. Either provided in the call or calculated from the correlation matrix, if given.} \item{rwg.j.95}{95 percent confidence interval estimate associated with a p-value of .05. Values greater than or equal to the rwg.j.95 value are considered significant.} } \author{ Paul Bliese \email{paul.bliese@moore.sc.edu} } \references{ Cohen, A., Doveh, E., & Nahum-Shani, I. (2009). Testing agreement for multi-item scales with the indices rwg(j) and adm(j). Organizational Research Methods, 12, 148-164. Cohen, A., Doveh, E., & Eick, U. (2001). Statistical properties of the rwg(j) index of agreement. Psychological Methods, 6, 297-310. James, L.R., Demaree, R.G., & Wolf, G. (1984). Estimating within-group interrater reliability with and without response bias. Journal of Applied Psychology, 69, 85-98. } \seealso{ \code{\link{rwg.j}} \code{\link{rwg}} \code{\link{rwg.sim}} \code{\link{rwg.j.lindell}} \code{\link{rgr.agree}} } \examples{ #An example assuming independent items RWG.J.OUT<-rwg.j.sim(gsize=10,nitems=6,nresp=5,nrep=1000) summary(RWG.J.OUT) quantile(RWG.J.OUT, c(.95,.99)) #A more realistic example assuming correlated items. The #estimate in Cohen et al. (2006) is .61. data(lq2002) library(MASS) RWG.J.OUT<-rwg.j.sim(gsize=10,nresp=5, itemcors=cor(lq2002[,c("TSIG01","TSIG02","TSIG03")]), nrep=1000) summary(RWG.J.OUT) quantile(RWG.J.OUT,c(.95,.99)) } \keyword{attribute}multilevel/man/cordif.dep.Rd0000644000176200001440000000214512750500650015515 0ustar liggesusers\name{cordif.dep} \alias{cordif.dep} \title{Estimate whether two dependent correlations differ} \description{ This function tests for statistical differences between two dependent correlations using the formula provided on page 56 of Cohen & Cohen (1983). The function returns a t-value, the DF and the p-value. } \usage{ cordif.dep(r.x1y,r.x2y,r.x1x2,n) } \arguments{ \item{r.x1y}{The correlation between x1 and y where y is typically the outcome variable.} \item{r.x2y}{The correlation between x2 and y where y is typically the outcome variable.} \item{r.x1x2}{The correlation between x1 and x2 (the correlation between the two predictors).} \item{n}{The sample size.} } \value{ Returns three values. A t-value, DF and p-value. } \author{ Paul Bliese \email{paul.bliese@moore.sc.edu} } \references{ Cohen, J. & Cohen, P. (1983). Applied multiple regression/correlation analysis for the behavioral sciences (2nd Ed.). Hillsdale, nJ: Lawrence Erlbaum Associates. } \seealso{ \code{\link{cordif}} } \examples{ cordif.dep(r.x1y=.30,r.x2y=.60,r.x1x2=.10,n=305) } \keyword{htest}multilevel/man/summary.disagree.sim.Rd0000644000176200001440000000155112750500661017550 0ustar liggesusers\name{summary.disagree.sim} \alias{summary.disagree.sim} \title{S3 method for class 'disagree.sim'} \description{ This function provides a concise summary of objects created using the function ad.m.sim. } \usage{ \method{summary}{disagree.sim}(object,\dots) } \arguments{ \item{object}{An object of class 'disagree.sim'.} \item{\dots}{Optional additional arguments. None used.} } \value{A summary of all the output elements in the disagree.sim class object.} \author{ Paul Bliese \email{paul.bliese@moore.sc.edu} } \seealso{ \code{\link{ad.m.sim}} } \examples{ #Example from Dunlap et al. (2003), Table 3. The listed significance #value for a group of size 5 with a 7-item response format is 0.64 or less SIMOUT<-ad.m.sim(gsize=5, nitems=1, nresp=7, itemcors=NULL, type="mean", nrep=1000) summary(SIMOUT) } \keyword{programming}multilevel/man/rwg.sim.Rd0000644000176200001440000000452212750500657015076 0ustar liggesusers\name{rwg.sim} \alias{rwg.sim} \title{Simulate rwg values from a random null distribution} \description{ This function is based on the work of Dunlap, Burke & Smith-Crowe (2003). The function draws data from a random uniform null distribution, and calculates the within group agreement measure rwg for single item measures as described in James, Demaree & Wolf (1984). By repeatedly drawing random samples, a distribution of the rwg is generated. The sampling distribution can be used to calculate confidence intervals for different combinations of group sizes and number of response options (A). } \usage{ rwg.sim(gsize, nresp, nrep) } \arguments{ \item{gsize}{Group size upon which to base the rwg simulation.} \item{nresp}{The number of response options (e.g., there would be 5 response options if using Strongly Disagree, Disagree, Neither, Agree, Strongly Agree).} \item{nrep}{The number of rwg values to simulate. This will generally be 10,000 or more, although the examples use nrep of 1000 to make the calculations fast.} } \value{ \item{rwg}{rwg value from each simulation.} \item{gsize}{Group size used in the rwg simulation.} \item{nresp}{Simulated number of response options.} \item{nitems}{Will always be 1 for an rwg estimate.} \item{rwg.95}{Estimated 95 percent confidence interval. Values greater than or equal to rwg.95 are considered significant, p<.05.} } \author{ Paul Bliese \email{paul.bliese@moore.sc.edu} } \references{ Cohen, A., Doveh, E., & Eick, U. (2001). Statistical properties of the rwg(j) index of agreement. Psychological Methods, 6, 297-310. Dunlap, W. P., Burke, M. J., & Smith-Crowe, K. (2003). Accurate tests of statistical significance for rwg and average deviation interrater agreement indices. Journal of Applied Psychology, 88, 356-362. James, L.R., Demaree, R.G., & Wolf, G. (1984). Estimating within-group interrater reliability with and without response bias. Journal of Applied Psychology, 69, 85-98. } \seealso{ \code{\link{ad.m}} \code{\link{rwg.j}} \code{\link{rwg}} \code{\link{rwg.j.sim}} \code{\link{rgr.agree}} } \examples{ #An example from Dunlap et al. (2003). The estimate from Dunlap #et al. Table 2 is 0.53 (p=.05) RWG.OUT<-rwg.sim(gsize=10,nresp=5,nrep=1000) summary(RWG.OUT) quantile(RWG.OUT, c(.95,.99)) } \keyword{attribute}multilevel/man/awg.Rd0000644000176200001440000000524212750500646014264 0ustar liggesusers\name{awg} \alias{awg} \title{Brown and Hauenstein (2005) awg agreement index} \description{ This function calculates the awg index proposed by Brown and Hauenstein (2005). The awg agreement index can be applied to either a single item vector or a multiple item matrix representing a scale. The awg is an analogue to Cohen's kappa. Brown and Hauenstein (pages 177-178) recommend interpreting the awg similarly to how the rwg (James et al., 1984) is commonly interpreted with values of .70 indicating acceptable agreement; values between .60 and .69 as reasonable agreement, and values less than .60 as unacceptable levels of agreement. } \usage{ awg(x, grpid, range=c(1,5)) } \arguments{ \item{x}{A vector representing a single item or a matrix representing a scale of interest. If a matrix, each column of the matrix represents a scale item, and each row represents an individual respondent.} \item{grpid}{A vector identifying the groups from which x originated.} \item{range}{A vector with the lower and upper response options (e.g., c(1,5)) for a five-point scale from strongly disagree to strongly agree.} } \value{ \item{grpid}{The group identifier.} \item{a.wg}{The awg estimate for each group.} \item{nitems}{The number of scale items when x is a matrix or dataframe representing a multi-item scale. This value is not returned when x is a vector.} \item{nraters}{The number of raters. Given that the awg estimate is based on the sample estimate of variance with N-1 in the denominator, Brown and Hauenstein (2005) contend that awg can be estimated on as few as A-1 raters where A represents the number of response options specified by the range option (5 as the default). Note that in many situations nraters will correspond to group size.} } \author{ Paul Bliese \email{paul.bliese@moore.sc.edu} } \references{ Brown, R. D. & Hauenstein, N. M. A. (2005). Interrater Agreement Reconsidered: An Alternative to the rwg Indices. Organizational Research Methods, 8, 165-184. Wagner, S. M., Rau, C., & Lindemann, E. (2010). Multiple informant methodology: A critical review and recommendations. Sociological Methods and Research, 38, 582-618. } \seealso{ \code{\link{rwg}} \code{\link{rwg.j}} \code{\link{ad.m}} } \examples{ data(lq2002) #Examples for multiple item scales awg.out<-awg(lq2002[,3:13],lq2002$COMPID,range=c(1,5)) summary(awg.out) #Example for single item measure awg.out<-awg(lq2002$LEAD05,lq2002$COMPID,range=c(1,5)) summary(awg.out) } \keyword{attribute}multilevel/man/rwg.j.lindell.Rd0000644000176200001440000000513212750500656016156 0ustar liggesusers\name{rwg.j.lindell} \alias{rwg.j.lindell} \title{Lindell et al. r*wg(j) agreement index for multi-item scales} \description{ This function calculates the Lindell et al r*wg(j) within-group agreement index for multiple item measures. It is similar to the James, Demaree and Wolf (1984) rwg and rwg(j) indices. The r*wg(j) index is calculated by taking the average item variability as the Observed Group Variance, and using the average item variability in the numerator of the rwg formula (rwg=1-(Observed Group Variance/ Expected Random Variance)). In practice, this means that the r*wg(j) does not increase as the number of items in the scale increases as does the rwg(j). Additionally, the r*wg(j) allows Observed Group Variances to be larger than Expected Random Variances. In practice this means that r*wg(j) values can be negative. } \usage{ rwg.j.lindell(x, grpid, ranvar=2) } \arguments{ \item{x}{A matrix representing the scale of interest upon which one is interested in estimating agreement. Each column of the matrix represents a separate scale item, and each row represents an individual respondent.} \item{grpid}{A vector identifying the groups from which x originated.} \item{ranvar}{The random variance to which actual group variances are compared. The value of 2 represents the variance from a rectangular distribution in the case where there are 5 response options (e.g., Strongly Disagree, Disagree, Neither, Agree, Strongly Agree). In cases where there are not 5 response options, the rectangular distribution is estimated using the formula \eqn{\mathtt{ranvar}=(A^{2}-1)/12}{ranvar=(A^2-1)/12} where A is the number of response options. Note that one is not limited to the rectangular distribution; rather, one can include any appropriate random value for ranvar.} } \value{ \item{grpid}{The group identifier} \item{rwg.lindell}{The r*wg(j) estimate for the group} \item{gsize}{The group size} } \author{ Paul Bliese \email{paul.bliese@moore.sc.edu} } \references{ James, L.R., Demaree, R.G., & Wolf, G. (1984). Estimating within-group interrater reliability with and without response bias. Journal of Applied Psychology, 69, 85-98. Lindell, M. K. & Brandt, C. J. (1999). Assessing interrater agreement on the job relevance of a test: A comparison of CVI, T, rWG(J), and r*WG(J) indexes. Journal of Applied Psychology, 84, 640-647. } \seealso{ \code{\link{rwg}} \code{\link{rwg.j}} \code{\link{rgr.agree}} } \examples{ data(lq2002) RWGOUT<-rwg.j.lindell(lq2002[,3:13],lq2002$COMPID) RWGOUT[1:10,] summary(RWGOUT) } \keyword{attribute}multilevel/man/mult.icc.Rd0000644000176200001440000000271612750500653015225 0ustar liggesusers\name{mult.icc} \alias{mult.icc} \title{Multiple ICCs from a dataset} \description{ Given a data frame and a group identifier, this function will estimate ICC(1) and ICC(2) values for each column in the dataframe. Note that this function depends upon the nlme package, and it only works with one level of nesting (e.g., students within schools). The dependent variable is assumed to be gaussian. } \usage{ mult.icc(x, grpid) } \arguments{ \item{x}{A data frame containing the variables of interest in each column.} \item{grpid}{A vector identifying the groups from which the variables originated.} } \value{ \item{Variable}{The variable name.} \item{ICC1}{The intraclass correlation coefficient 1.} \item{ICC2}{The group mean reliability or intraclass correlation coefficient 2.} } \author{ Paul Bliese \email{paul.bliese@moore.sc.edu} } \references{ Bliese, P. D. (2000). Within-group agreement, non-independence, and reliability: Implications for data aggregation and Analysis. In K. J. Klein & S. W. Kozlowski (Eds.), Multilevel Theory, Research, and Methods in Organizations (pp. 349-381). San Francisco, CA: Jossey-Bass, Inc. Bartko, J.J. (1976). On various intraclass correlation reliability coefficients. Psychological Bulletin, 83, 762-765. } \seealso{ \code{\link{ICC2}} \code{\link{ICC1}} } \examples{ library(nlme) data(bh1996) mult.icc(bh1996[,c("HRS","LEAD","COHES")],grpid=bh1996$GRP) } \keyword{attribute}multilevel/man/tankdat.Rd0000644000176200001440000000306412750500662015132 0ustar liggesusers\name{tankdat} \docType{data} \alias{tankdat} \title{Tank data from Bliese and Lang (in press)} \description{ This data set is a partial sample of data collected by Lang and reported in Lang and Bliese (2009). The tankdat sub-sample was used as an example of discontinuous growth modeling in Bliese and Lang (in press). The data set is in long (univariate) format, and contains performance data from 184 participants over 12 repeated measures on a complex tank simulation task. In the research paradigm, the task was unexpectedly changed after the first six performance episodes. Discontinuous growth models were used to examine participants' reactions to the unexpected change. The data set contains the person-level predictor of conscientiousness. } \usage{data(tankdat)} \format{A dataframe with 4 columns and 2208 observations \tabular{llll}{ [,1] \tab ID \tab numeric \tab Participant ID\cr [,2] \tab CONSC \tab numeric \tab Participant Conscientiousness\cr [,3] \tab TIME \tab numeric \tab Time\cr [,4] \tab SCORE \tab numeric \tab Task Performance\cr } } \references{ Bliese, P. D., & Lang, J. W. B. (in press). Understanding relative and absolute change in discontinuous growth models: Coding alternatives and implications for hypothesis testing. Organizational Research Methods. Lang, J. W. B., & Bliese, P. D. (2009). General mental ability and two types of adaptation to unforeseen change: Applying discontinuous growth models to the task-change paradigm. Journal of Applied Psychology, 92, 411-428. } \keyword{datasets}multilevel/man/rwg.j.Rd0000644000176200001440000000445612750500657014545 0ustar liggesusers\name{rwg.j} \alias{rwg.j} \title{James et al., (1984) agreement index for multi-item scales} \description{ This function calculates the within group agreement measure rwg(j) for multiple item measures as described in James, Demaree & Wolf (1984). James et al. (1984) recommend truncating the Observed Group Variance to the Expected Random Variance in cases where the Observed Group Variance was larger than the Expected Random Variance. This truncation results in an rwg.j value of 0 (no agreement) for groups with large variances. } \usage{ rwg.j(x, grpid, ranvar=2) } \arguments{ \item{x}{A matrix representing the scale items. Each column of the matrix represents a separate item, and each row represents an individual respondent.} \item{grpid}{A vector identifying the group from which x originated.} \item{ranvar}{The random variance to which actual group variances are compared. The value of 2 represents the variance from a rectangular distribution in the case where there are 5 response options (e.g., Strongly Disagree, Disagree, Neither, Agree, Strongly Agree). In cases where there are not 5 response options, the rectangular distribution is estimated using the formula \eqn{\mathtt{ranvar}=(A^2-1)/12}{ranvar=(A^2-1)/12} where A is the number of response options. While the rectangular distribution is widely used, other random values may be more appropriate.} } \value{ \item{grpid}{The group identifier} \item{rwg.j}{The rwg(j) estimate for the group} \item{gsize}{The group size} } \author{ Paul Bliese \email{paul.bliese@moore.sc.edu} } \references{ Bliese, P. D. (2000). Within-group agreement, non-independence, and reliability: Implications for data aggregation and analysis. In K. J. Klein & S. W. Kozlowski (Eds.), Multilevel Theory, Research, and Methods in Organizations (pp. 349-381). San Francisco, CA: Jossey-Bass, Inc. James, L.R., Demaree, R.G., & Wolf, G. (1984). Estimating within-group interrater reliability with and without response bias. Journal of Applied Psychology, 69, 85-98. } \seealso{ \code{\link{ad.m}} \code{\link{rwg}} \code{\link{rgr.agree}} \code{\link{rwg.j.lindell}} \code{\link{rwg.j.sim}} } \examples{ data(lq2002) RWGOUT<-rwg.j(lq2002[,3:13],lq2002$COMPID) RWGOUT[1:10,] summary(RWGOUT) } \keyword{attribute}multilevel/man/chen2005.Rd0000644000176200001440000000342012750500647014727 0ustar liggesusers\name{chen2005} \docType{data} \alias{chen2005} \title{Data from Chen (2005)} \description{ This data set contains the complete data used in Chen (2005). Chen (2005) examined newcomer adaptation in 65 project teams. The level of analysis was the team-level. In the study, team leaders assessed the initial team performance (TMPRF) at time 1 and then assessed newcomer performance over three additional time points (NCPRF.T1, NCPRF.T2, NCPRF.T3). Initial team expectations (TMEXP) and initial newcomer empowerment (NCEMP) were also assessed and modeled, but were not analyzed as repeated measures. To specify the Table 2 model in Chen (2005), these data need to be converted to univariate or stacked form (see the make.univ function). Using the default values of make.univ and creating a dataframe called chen2005.univ, the specific lme model is lme(MULTDV~NCEMP*TIME+TMEXP*TIME+TMPRF*TIME,random=~TIME|ID,chen2005.univ) } \usage{data(chen2005)} \format{A data frame with 7 columns and 65 team-level observations \tabular{llll}{ [,1] \tab ID \tab numeric \tab Team Identifier\cr [,2] \tab TMPRF \tab numeric \tab Initial Team Performance (time 1 in article)\cr [,3] \tab TMEXP \tab numeric \tab Team Expectations (time 1 in article)\cr [,4] \tab NCEMP \tab numeric \tab Initial Newcomer Empowerment(time 2 in article)\cr [,5] \tab NCPRF.T1 \tab numeric \tab Newcomer Performance Time 1 (time 2 in article)\cr [,6] \tab NCPRF.T2 \tab numeric \tab Newcomer Performance Time 2 (time 3 in article)\cr [,7] \tab NCPRF.T3 \tab numeric \tab Newcomer Performance Time 3 (time 4 in article)\cr } } \references{ Chen, G.(2005). Newcomer adaptation in team: Multilevel antecedents and outcomes. Academy of Management Journal, 48, 101-116. } \keyword{datasets}multilevel/man/bh1996.Rd0000644000176200001440000000350412750500646014427 0ustar liggesusers\name{bh1996} \docType{data} \alias{bh1996} \title{Data from Bliese and Halverson (1996)} \description{ This dataset contains the complete data used in Bliese and Halverson (1996). The dataset contains 4 variables. These variables are Cohesion (COHES), Leadership Climate (LEAD), Well-Being (WBEING) and Work Hours (HRS). Each of these variables has two variants -- a group mean version that replicates each group mean for every individual, and a within-group version where the group mean is subtracted from each individual response. The group mean version is designated with a G. (e.g., G.HRS), and the within-group version is designated with a W. (e.g., W.HRS). } \usage{data(bh1996)} \format{A data frame with 13 columns and 7,382 observations from 99 groups \tabular{llll}{ [,1] \tab GRP \tab numeric \tab Group Identifier\cr [,2] \tab COHES \tab numeric \tab Cohesion\cr [,3] \tab G.COHES \tab numeric \tab Average Group Cohesion\cr [,4] \tab W.COHES \tab numeric \tab Group-Mean Centered Cohesion\cr [,5] \tab LEAD \tab numeric \tab Leadership \cr [,6] \tab G.LEAD \tab numeric \tab Average Group Leadership\cr [,7] \tab W.LEAD \tab numeric \tab Group-Mean Centered Leadership\cr [,8] \tab HRS \tab numeric \tab Work Hours\cr [,9] \tab G.HRS \tab numeric \tab Average Group Work Hours\cr [,10] \tab W.HRS \tab numeric \tab Group-Mean Centered Work Hours\cr [,11] \tab WBEING \tab numeric \tab Well-Being\cr [,12] \tab G.WBEING \tab numeric \tab Average Group Well-Being\cr [,13] \tab W.WBEING \tab numeric \tab Group-Mean Centered Well-Being } } \references{ Bliese, P. D. & Halverson, R. R. (1996). Individual and nomothetic models of job stress: An examination of work hours, cohesion, and well-being. Journal of Applied Social Psychology, 26, 1171-1189. } \keyword{datasets} multilevel/man/ICC1.Rd0000644000176200001440000000237412750500651014164 0ustar liggesusers\name{ICC1} \alias{ICC1} \title{Function to Estimate Intraclass Correlation Coefficient 1 or ICC(1) from an aov model} \description{This function calculates the Intraclass Correlation Coefficient 1 or ICC(1) from an ANOVA model. This value is equivalent to the ICC discussed in the random coefficient modeling literature, and represents the amount of individual-level variance that can be "explained" by group membership. } \usage{ ICC1(object) } \arguments{ \item{object}{An ANOVA (aov) object from an one-way analysis of variance.} } \value{Provides an estimate of ICC(1) for the sample.} \references{ Bliese, P. D. (2000). Within-group agreement, non-independence, and reliability: Implications for data aggregation and Analysis. In K. J. Klein & S. W. Kozlowski (Eds.), Multilevel Theory, Research, and Methods in Organizations (pp. 349-381). San Francisco, CA: Jossey-Bass, Inc. Bartko, J.J. (1976). On various intraclass correlation reliability coefficients. Psychological Bulletin, 83, 762-765.} \author{ Paul Bliese \email{paul.bliese@moore.sc.edu} } \seealso{ \code{\link{ICC2}} \code{\link{aov}} } \examples{ data(bh1996) hrs.mod<-aov(HRS~as.factor(GRP),data=bh1996) ICC1(hrs.mod) } \keyword{attribute} multilevel/man/lq2002.Rd0000644000176200001440000000655112750500652014427 0ustar liggesusers\name{lq2002} \docType{data} \alias{lq2002} \title{Data used in special issue of Leadership Quarterly, Vol. 13, 2002} \description{ This dataset contains the complete data used in a special issue of Leadership Quarterly edited by Paul Bliese, Ronald Halverson and Chet Schriesheim in 2002 (Vol 13). Researchers from several universities analyzed this common dataset using various multilevel techniques. The three scales used in the analyses are Leadership Climate (LEAD), Task Significance (TSIG) and Hostility (HOSTILE). The data set contains each item making up these scales. These items were used by Cohen, Doveh and Nahum-Shani (2009). } \usage{data(lq2002)} \format{A data frame with 27 columns and 2,042 observations from 49 groups \tabular{llll}{ [,1] \tab COMPID \tab numeric \tab Army Company Identifying Variable\cr [,2] \tab SUB \tab numeric \tab Subject Number\cr [,3] \tab LEAD01 \tab numeric \tab Officers Get Cooperation From Company (EXV01)\cr [,4] \tab LEAD02 \tab numeric \tab NCOs Get Cooperation From Company (EXV02)\cr [,5] \tab LEAD03 \tab numeric \tab Impressed By Leadership (EXV04)\cr [,6] \tab LEAD04 \tab numeric \tab Go For Help Within Chain of Command (EXV05)\cr [,7] \tab LEAD05 \tab numeric \tab Officers Would Lead Well In Combat (EXV07)\cr [,8] \tab LEAD06 \tab numeric \tab NCOs Would Lead Well In Combat (EXV08)\cr [,9] \tab LEAD07 \tab numeric \tab Officers Interested In Welfare (EXV11)\cr [,10] \tab LEAD08 \tab numeric \tab NCOs Interested In Welfare (EXV13)\cr [,11] \tab LEAD09 \tab numeric \tab Officers Interested In What I Think (EXV14)\cr [,12] \tab LEAD10 \tab numeric \tab NCOs Interested In What I Think (EXV15)\cr [,13] \tab LEAD11 \tab numeric \tab Chain Of Command Works Well (EXV16)\cr [,14] \tab TSIG01 \tab numeric \tab What I Am Doing Is Important (MIS05)\cr [,15] \tab TSIG02 \tab numeric \tab Making Contribution To Mission (MIS06)\cr [,16] \tab TSIG03 \tab numeric \tab What I Am Doing Accomplishes Mission (MIS07)\cr [,17] \tab HOSTIL01 \tab numeric \tab Easily Annoyed Or Irritated (BSI09)\cr [,18] \tab HOSTIL02 \tab numeric \tab Temper Outburst That You Cannot Control (BSI18)\cr [,19] \tab HOSTIL03 \tab numeric \tab Urges To Harm Someone (BSI47)\cr [,20] \tab HOSTIL04 \tab numeric \tab Urges To Break Things (BSI49)\cr [,21] \tab HOSTIL05 \tab numeric \tab Getting Into Frequent Arguments (BSI54)\cr [,22] \tab LEAD \tab numeric \tab Leadership Climate Scale Score\cr [,23] \tab TSIG \tab numeric \tab Task Significance Scale Score\cr [,24] \tab HOSTILE \tab numeric \tab Hostility Scale Score\cr [,25] \tab GLEAD \tab numeric \tab Leadership Climate Scale Score Aggregated By Company\cr [,26] \tab GTSIG \tab numeric \tab Task Significance Scale Score Aggregated By Company\cr [,27] \tab GHOSTILE \tab numeric \tab Hostility Scale Score Aggregated By Company } } \references{ Bliese, P. D., & Halverson, R. R. (2002). Using random group resampling in multilevel research. Leadership Quarterly, 13, 53-68. Bliese, P. D., Halverson, R. R., & Schriesheim, C. A. (2002). Benchmarking multilevel methods: Comparing HLM, WABA, SEM, and RGR. Leadership Quarterly, 13, 3-14. Cohen, A., Doveh, E., & Nahum-Shani, I. (2009). Testing agreement for multi-item scales with the indices rwg(j) and adm(j). Organizational Research Methods, 12, 148-164. } \keyword{datasets} multilevel/man/rgr.OLS.Rd0000644000176200001440000000277712750500655014746 0ustar liggesusers\name{rgr.OLS} \alias{rgr.OLS} \title{Random Group Resampling OLS Regression} \description{This function uses Random Group Resampling (RGR) within an Ordinary Least Square (OLS) framework to allow one to contrast actual group results with pseudo group results. The number of columns in the output matrix of the function (OUT) has to correspond to the number of mean squares you want in the output which in turn is a function of the number of predictors. This specific function does RGR on an OLS hierarchical OLS model with two predictors as in Bliese & Halverson (2002). To run this analysis on data with more predictors, the function will have to be modified.} \usage{ rgr.OLS(xdat1,xdat2,ydata,grpid,nreps) } \arguments{ \item{xdat1}{The first predictor.} \item{xdat2}{The second predictor.} \item{ydata}{The outcome.} \item{grpid}{The group identifier.} \item{nreps}{The number of pseudo groups to create.} } \value{A matrix containing mean squares. Each row provides mean square values for a single pseudo group iteration} \references{Bliese, P. D., & Halverson, R. R. (2002). Using random group resampling in multilevel research. Leadership Quarterly, 13, 53-68.} \author{ Paul Bliese \email{paul.bliese@moore.sc.edu}} \seealso{\code{\link{mix.data}}} \examples{ data(lq2002) RGROUT<-rgr.OLS(lq2002$LEAD,lq2002$TSIG,lq2002$HOSTILE,lq2002$COMPID,100) #Compare values to those reported on p.62 in Bliese & Halverson (2002) summary(RGROUT) } \keyword{attribute} multilevel/man/ran.group.Rd0000644000176200001440000000207012750500655015415 0ustar liggesusers\name{ran.group} \alias{ran.group} \title{Randomly mix grouped data and return function results} \description{This function is called by rgr.agree (and potentially other functions). The ran.group function randomly mixes data and applies a function to the pseudo groups. Pseudo group IDs match real group IDs in terms of size.} \usage{ ran.group(x,grpid,fun,...) } \arguments{ \item{x}{A matrix or vector containing data to be randomly sorted.} \item{grpid}{A vector containing a group identifier.} \item{fun}{A function to be applied to the observations within each random group.} \item{...}{Additional arguments to fun.} } \value{A vector containing the results of applying the function to each random group.} \references{Bliese, P. D., & Halverson, R. R. (2002). Using random group resampling in multilevel research. Leadership Quarterly, 13, 53-68.} \author{ Paul Bliese \email{paul.bliese@moore.sc.edu}} \seealso{\code{\link{rgr.agree}}} \examples{ data(bh1996) ran.group(bh1996$HRS,bh1996$GRP,mean) } \keyword{programming} multilevel/man/cordif.Rd0000644000176200001440000000205612750500650014747 0ustar liggesusers\name{cordif} \alias{cordif} \title{Estimate whether two independent correlations differ} \description{ This function tests for statistical differences between two independent correlations using the formula provided on page 54 of Cohen & Cohen (1983). The function returns a z-score estimate. } \usage{ cordif(rvalue1,rvalue2,n1,n2) } \arguments{ \item{rvalue1}{Correlation value from first sample.} \item{rvalue2}{Correlation value from second sample.} \item{n1}{The sample size of the first correlation.} \item{n2}{The sample size of the second correlation.} } \value{ Produces a single value, the z-score for the differences between the correlations. } \author{ Paul Bliese \email{paul.bliese@moore.sc.edu} } \references{ Cohen, J. & Cohen, P. (1983). Applied multiple regression/correlation analysis for the behavioral sciences (2nd Ed.). Hillsdale, NJ: Lawrence Erlbaum Associates. } \seealso{ \code{\link{rtoz}} \code{\link{cordif.dep}} } \examples{ cordif(rvalue1=.51,rvalue2=.71,n1=123,n2=305) } \keyword{htest}multilevel/man/mult.make.univ.Rd0000644000176200001440000000516412750500653016364 0ustar liggesusers\name{mult.make.univ} \alias{mult.make.univ} \title{Convert two or more variables from multivariate to univariate form} \description{ Longitudinal data is often stored in multivariate or wide form. In multivariate form, each row contains data from one subject, and repeated measures variables are indexed by different names (e.g., OUTCOME.T1, OUTCOME.T2, OUTCOME.T3). In the case of repeated measures designs and growth modeling, it is necessary to convert the data to univariate or stacked form where each row represents one of the repeated measures indexed by a TIME variable and nested within subject. In univariate form, each subject has as many rows of data as there are time points. The make.univ function in the multilevel library will convert a single item to univariate form while the mult.make.univ function converts two or more variables to univariate form. The mult.make.univ function was developed by Patrick Downes at the University of Iowa, and was recommended for inclusion in the multilevel library in January of 2013. } \usage{ mult.make.univ(x,dvlist,tname="TIME", outname="MULTDV") } \arguments{ \item{x}{A dataframe in multivariate form.} \item{dvlist}{A list containing the repeated measures. Note that each element of the list must be ordered from Time 1 to Time N for this function to work properly.} \item{tname}{An optional name for the new time variable. Defaults to TIME.} \item{outname}{An optional name for the outcome variable name. Defaults to MULTDV1 to MULTDV(N).} } \value{ Returns a dataframe in univariate (i.e., stacked) form with a TIME variable representing the repeated observations, and new variables representing the time-indexed variables (MULTDV1, MULTDV2, etc.). The TIME variable begins with 0. } \author{ Patrick Downes \email{pat-downes@uiowa.edu} Paul Bliese \email{paul.bliese@moore.sc.edu} } \references{ Bliese, P. D., & Ployhart, R. E. (2002). Growth modeling using random coefficient models: Model building, testing and illustrations. Organizational Research Methods, 5, 362-387. } \seealso{ \code{\link{make.univ}} } \examples{ data(univbct) #a dataframe in univariate form for job sat TEMP<-univbct[3*1:495,c(22,1:17)] #convert back to multivariate form names(TEMP) #use the column names to find the column numbers #Create a list of DV's - each DV should have the same number of obs dvlist <- list(c(10,13,16),c(11,14,17)) names(dvlist) <- c("JOBSAT","COMMIT") #names for univariate output #Transform the data into univariate form with multiple level-1 variables mldata <- mult.make.univ(x=TEMP,dvlist=dvlist) } \keyword{manip}multilevel/man/summary.agree.sim.Rd0000644000176200001440000000142512750500661017050 0ustar liggesusers\name{summary.agree.sim} \alias{summary.agree.sim} \title{S3 method for class 'agree.sim'} \description{ This function provides a concise summary of objects created using the functions rwg.sim and rwg.j.sim. } \usage{ \method{summary}{agree.sim}(object,\dots) } \arguments{ \item{object}{An object of class 'agree.sim'.} \item{\dots}{Optional additional arguments. None used.} } \value{A summary of all the output elements in the agree.sim class object.} \author{ Paul Bliese \email{paul.bliese@moore.sc.edu} } \seealso{ \code{\link{rwg.sim}} \code{\link{rwg.j.sim}} } \examples{ #An example from Dunlap et al. (2003). The estimate from Dunlap et al. #Table 2 is 0.53 RWG.OUT<-rwg.sim(gsize=10,nresp=5,nrep=1000) summary(RWG.OUT) } \keyword{programming}multilevel/man/sobel.Rd0000644000176200001440000000361212750500661014606 0ustar liggesusers\name{sobel} \alias{sobel} \title{Estimate Sobel's (1982) Test for Mediation} \description{Estimate Sobel's (1982) indirect test for mediation. The function provides an estimate of the magnitude of the indirect effect, Sobel's first-order estimate of the standard error associated with the indirect effect, and the corresponding z-value. The estimates are based upon three models as detailed on page 84 of MacKinnon, Lockwood, Hoffman, West and Sheets (2002).} \usage{ sobel(pred,med,out) } \arguments{ \item{pred}{The predictor or independent variable (X).} \item{med}{The mediating variable (M).} \item{out}{The outcome or dependent variable (Y).} } \value{ \item{Mod1: Y~X}{A summary of coefficients from Model 1 of MacKinnon et al., (2002).} \item{Mod2: Y~X+M}{A summary of coefficients from Model 2 of MacKinnon et al., (2002).} \item{Mod3: M~X}{A summary of coefficients from Model 3 of MacKinnon et al., (2002).} \item{Indirect.Effect}{The estimate of the indirect mediating effect.} \item{SE}{Sobel's (1982) Standard Error estimate.} \item{z.value}{The estimated z-value.} \item{N}{The number of observations used in model estimation.} } \references{MacKinnon, D. P., Lockwood, C. M., Hoffman, J. M., West, S. G., Sheets, V. (2002). A comparison of methods to test mediation and other intervening variable effects. Psychological Methods, 7, 83-104. Sobel, M. E., (1982). Asymptotic confidence intervals for indirect effects in structural equation models. In S. Leinhardt (Ed.), Sociological Methodology 1982 (pp. 290-312). Washington, DC: American Sociological Association.} \author{ Paul Bliese \email{paul.bliese@moore.sc.edu}} \examples{ data(bh1996) #A small but significant indirect effect indicates leadership mediates #the relationship between work hours and well-being. sobel(pred=bh1996$HRS,med=bh1996$LEAD,out=bh1996$WBEING) } \keyword{htest} multilevel/man/item.total.Rd0000644000176200001440000000161512750500651015562 0ustar liggesusers\name{item.total} \alias{item.total} \title{Item-total correlations} \description{This function calculates item-total correlations in multi-item scales.} \usage{ item.total(items) } \arguments{ \item{items}{A matrix or dataframe where each column represents an item in a multi-item scale.} } \value{ \item{Variable}{Variable examined in the reliability analyses.} \item{Item.Total}{The item-total correlation.} \item{Alpha.Without}{The Cronbach Alpha reliability estimate of the scale without the variable.} \item{N}{The number of observations on which the analyses were calculated.} } \references{Cronbach L. J. (1951) Coefficient Alpha and the internal structure of tests. Psychometrika, 16,297-334} \author{ Paul Bliese \email{paul.bliese@moore.sc.edu}} \seealso{\code{\link{cronbach}}} \examples{ data(bhr2000) item.total(bhr2000[,2:11])} \keyword{attribute} multilevel/man/cronbach.Rd0000644000176200001440000000136512750500650015262 0ustar liggesusers\name{cronbach} \alias{cronbach} \title{Estimate Cronbach's Alpha} \description{This function calculates the Cronbach's Alpha estimate of reliability for a multi-item scale.} \usage{ cronbach(items) } \arguments{ \item{items}{An matrix or data frame where each column represents an item in a multi-item scale.} } \value{ \item{Alpha}{Estimate of Cronbach's Alpha.} \item{N}{The number of observations on which the Alpha was estimated.} } \references{Cronbach L. J. (1951) Coefficient Alpha and the internal structure of tests. Psychometrika, 16,297-334} \author{ Paul Bliese \email{paul.bliese@moore.sc.edu}} \seealso{\code{\link{cronbach}}} \examples{ data(bhr2000) cronbach(bhr2000[,2:11]) } \keyword{attribute} multilevel/man/rtoz.Rd0000644000176200001440000000142712750500656014506 0ustar liggesusers\name{rtoz} \alias{rtoz} \title{Conducts an r to z transformation} \description{ This function transforms a correlation (r) to a z variate using the formula provided on page 53 of Cohen & Cohen (1983). The formula is z=.5*((log(1+r))-(log(1-r))) where r is the correlation. } \usage{ rtoz(rvalue) } \arguments{ \item{rvalue}{The correlation for which one wants the z transformation.} } \value{ Produces a single value, the z transformation. } \author{ Paul Bliese \email{paul.bliese@moore.sc.edu} } \references{ Cohen, J. & Cohen, P. (1983). Applied multiple regression/correlation analysis for the behavioral sciences (2nd Ed.). Hillsdale, NJ: Lawrence Erlbaum Associates. } \seealso{ \code{\link{cordif}} } \examples{ rtoz(.84) } \keyword{htest}multilevel/man/bhr2000.Rd0000644000176200001440000000477012750500647014571 0ustar liggesusers\name{bhr2000} \docType{data} \alias{bhr2000} \title{Data from Bliese, Halverson and Rothberg (2000)} \description{ This data set contains the complete data used in Bliese, Halverson & Rotheberg (2000). The data set contains 14 variables with individual ratings of US Army Company leadership, work hours, and the degree to which individuals find comfort from religion. The leadership and workhours variables are subsets of the Bliese and Halveson (1996) data set; however, in the case of leadership, the agree data set contains the 11 items that make up the scale whereas the bh1996 data set contains only the scale score. Most items are on a strongly disagree to strongly agree scale. The RELIG item is on a never to always scale. } \usage{data(bhr2000)} \format{A data frame with 14 columns and 5,400 observations from 99 groups \tabular{llll}{ [,1] \tab GRP \tab numeric \tab Group Identifier\cr [,2] \tab AF06 \tab numeric \tab Officers get willing and whole-hearted cooperation\cr [,3] \tab AF07 \tab numeric \tab NCOS most always get willing and whole-hearted cooperation\cr [,4] \tab AP12 \tab numeric \tab I am impressed by the quality of leadership in this company\cr [,5] \tab AP17 \tab numeric \tab I would go for help with a personal problem to the chain of command\cr [,6] \tab AP33 \tab numeric \tab Officers in this Company would lead well in combat\cr [,7] \tab AP34 \tab numeric \tab NCOs in this Company would lead well in combat\cr [,8] \tab AS14 \tab numeric \tab My officers are interested in my personal welfare\cr [,9] \tab AS15 \tab numeric \tab My NCOs are interested in my personal welfare\cr [,10] \tab AS16 \tab numeric \tab My officers are interested in what I think and feel about things\cr [,11] \tab AS17 \tab numeric \tab My NCOs are intested in what I think and fell about things\cr [,12] \tab AS28 \tab numeric \tab My chain-of-command works well\cr [,13] \tab HRS \tab numeric \tab How many hours do you usually work in a day\cr [,14] \tab RELIG \tab numeric \tab How often do you gain strength of comfort from religious beliefs } } \references{ Bliese, P. D. & Halverson, R. R. (1996). Individual and nomothetic models of job stress: An examination of work hours, cohesion, and well-being. Journal of Applied Social Psychology, 26, 1171-1189. Bliese, P. D., Halverson, R. R., & Rothberg, J. (2000). Using random group resampling (RGR) to estimate within-group agreement with examples using the statistical language R. } \keyword{datasets}multilevel/man/summary.rgr.agree.Rd0000644000176200001440000000202412750500662017047 0ustar liggesusers\name{summary.rgr.agree} \alias{summary.rgr.agree} \title{S3 method for class 'rgr.agree'} \description{ This function provides a concise summary of objects created using the function rgr.agree. } \usage{ \method{summary}{rgr.agree}(object,\dots) } \arguments{ \item{object}{An object of class 'rgr.agree'.} \item{\dots}{Optional additional arguments. None used.} } \value{ \item{Summary Statistics for Random and Real Groups}{Number of random groups, Average random group variance, Standard Deviation of random group variance, Actual group variance, z-value} \item{Lower Confidence Intervals (one-tailed)}{Lower confidence intervals based on sorted random group variances.} \item{Upper Confidence Intervals (one-Tailed)}{Upper confidence intervals based on sorted random group variances.} } \author{ Paul Bliese \email{paul.bliese@moore.sc.edu} } \seealso{ \code{\link{rgr.agree}} } \examples{ data(bh1996) RGROUT<-rgr.agree(bh1996$HRS,bh1996$GRP,1000) summary(RGROUT) } \keyword{programming}multilevel/man/cohesion.Rd0000644000176200001440000000204512750500647015314 0ustar liggesusers\name{cohesion} \docType{data} \alias{cohesion} \title{Five cohesion ratings from 11 individuals nested in 4 platoons in 2 larger units} \description{ This data set contains five cohesion measures provided by 11 individuals. The individuals providing the measures are members of four platoons further nested within two larger units. This data file is used for demonstative purposes in the document "Multilevel Modeling in R" that accompanies this package. } \usage{data(cohesion)} \format{A data frame with 7 columns and 11 observations \tabular{llll}{ [,1] \tab UNIT \tab numeric \tab Higher-level Unit Identifier\cr [,2] \tab PLATOON \tab numeric \tab Lower-level Platoon Identifier\cr [,3] \tab COH01 \tab numeric \tab First Cohesion Variable\cr [,4] \tab COH02 \tab numeric \tab Second Cohesion Variable\cr [,5] \tab COH03 \tab numeric \tab Third Cohesion Variable\cr [,6] \tab COH04 \tab numeric \tab Fourth Cohesion Variable\cr [,7] \tab COH05 \tab numeric \tab Fifth Cohesion Variable\cr } } \keyword{datasets}multilevel/man/simbias.Rd0000644000176200001440000000475712750500661015144 0ustar liggesusers\name{simbias} \alias{simbias} \title{Simulate Standard Error Bias in Non-Independent Data} \description{Non-independence due to groups is a common characteristic of applied data. In non-independent data, responses from members of the same group are more similar to each other than would be expected by chance. Non-independence is typically measured using the Intraclass Correlation Coefficient 1 or ICC(1). When non-independent data is treated as though it is independent, standard errors will be biased and power can decrease. This simulation allows one to estimate the bias and loss of statistical power that occurs when non-independent data is treated as though it is independent. The simulation contrasts a simple Ordinary Least Squares (OLS) model that fails to account for non-independence with a random coefficient model that accounts for non-independence. The simulation assumes that both the outcome (y) and the predictor (x) vary among individuals in the same group. } \usage{ simbias(corr,gsize,ngrp,icc1x,icc1y,nrep) } \arguments{ \item{corr}{The simulated true correlation between x and y.} \item{gsize}{The group size from which x and y are drawn.} \item{ngrp}{The number of groups.} \item{icc1x}{The simulated ICC(1) value for x.} \item{icc1y}{The simulated ICC(1) value for y.} \item{nrep}{The number of repetitions of simulated data sets.} } \value{ \item{icc1.x}{Observed ICC(1) value for x in the simulation.} \item{icc1.y}{Observed ICC(1) value for y in the simulation.} \item{lme.coef}{Parameter estimate from the lme model.} \item{lme.se}{Standard error estimate from the lme model.} \item{lme.tvalue}{t-value from the lme model.} \item{lm.coef}{Parameter estimate from the linear model (OLS).} \item{lm.se}{Standard error estimate from the linear model (OLS).} \item{lm.tvalue}{t-value from the linear model (OLS).} } \author{ Paul Bliese \email{paul.bliese@moore.sc.edu} } \references{ Bliese, P. D. & Hanges, P. J. (2004). Being both too liberal and too conservative: The perils of treating grouped data as though they were independent. Organizational Research Methods, 7, 400-417. } \seealso{ \code{\link{ICC1}} } \examples{ library(nlme) set.seed(15) SIM.OUTPUT<-simbias(corr=.15,gsize=10,ngrp=50,icc1x=0.05, icc1y=0.35, nrep=100) apply(SIM.OUTPUT,2,mean) 1-pnorm(1.96-3.39) #Power of the lme model (two-tailed, alpha=.05) 1-pnorm(1.96-2.95) #Power of the OLS model (two-tailed, alpha=.05) } \keyword{datagen}multilevel/man/mix.data.Rd0000644000176200001440000000176012750500652015211 0ustar liggesusers\name{mix.data} \alias{mix.data} \title{Randomly mix grouped data} \description{This function is called by graph.ran.mean (and potentially other functions) to randomly mix data and create new pseudo group ID variables. Pseudo group IDs match real group IDs in terms of size.} \usage{ mix.data(x,grpid) } \arguments{ \item{x}{A matrix or vector containing data to be randomly sorted.} \item{grpid}{A vector containing a group identifier.} } \value{ \item{newid}{ A pseudo group ID.} \item{grpid}{ The real group ID.} \item{x}{ The values in x arranged as belonging to newid.} } \references{Bliese, P. D., & Halverson, R. R. (2002). Using random group resampling in multilevel research. Leadership Quarterly, 13, 53-68.} \author{ Paul Bliese \email{paul.bliese@moore.sc.edu}} \seealso{\code{\link{graph.ran.mean}}} \examples{ data(bh1996) mix.data(x=bh1996[c(1:10,200:210,300:310),2:3], grpid=bh1996$GRP[c(1:10,200:210,300:310)]) } \keyword{programming} multilevel/man/GmeanRel.Rd0000644000176200001440000000317412750500650015175 0ustar liggesusers\name{GmeanRel} \alias{GmeanRel} \title{Group Mean Reliability from an lme model (nlme package)} \description{This function calculates the group-mean reliability from a linear mixed effects (lme) model. If group sizes are identical, the group-mean reliabilty estimate equals the ICC(2) estimate from an ANOVA model. When group sizes differ, however, a group-mean reliability estimate is calculated for each group based on the group size. The group-mean reliability estimate for each group is based upon the Spearman-Brown formula, the overall ICC, and group size for each group. } \usage{ GmeanRel(object) } \arguments{ \item{object}{A Linear Mixed Effect (lme) object.} } \value{ \item{ICC}{Intraclass Correlation Coefficient} \item{Group}{A vector containing all the group names.} \item{GrpSize}{A vector containing all the group sizes.} \item{MeanRel}{A vector containing the group-mean reliability estimate for each group.} } \author{ Paul Bliese \email{paul.bliese@moore.sc.edu} } \references{ Bliese, P. D. (2000). Within-group agreement, non-independence, and reliability: Implications for data aggregation and Analysis. In K. J. Klein & S. W. Kozlowski (Eds.), Multilevel Theory, Research, and Methods in Organizations (pp. 349-381). San Francisco, CA: Jossey-Bass, Inc. Bartko, J.J. (1976). On various intraclass correlation reliability coefficients. Psychological Bulletin, 83, 762-765. } \seealso{ \code{\link{ICC1}} \code{\link{ICC2}} \code{\link{lme}} } \examples{ data(bh1996) library(nlme) tmod<-lme(WBEING~1,random=~1|GRP,data=bh1996) GmeanRel(tmod) } \keyword{attribute}multilevel/man/rgr.agree.Rd0000644000176200001440000000460712750500655015366 0ustar liggesusers\name{rgr.agree} \alias{rgr.agree} \title{Random Group Resampling for Within-group Agreement} \description{ This function uses random group resampling (RGR) to estimate within group agreement. RGR agreement compares within group variances from actual groups to within group variances from pseudo groups. Evidence of significant agreement is inferred when variances from the actual groups are significantly smaller than variances from pseudo groups. RGR agreement methods are rarely reported, but provide another way to consider group level properties in data. } \usage{ rgr.agree(x, grpid, nrangrps) } \arguments{ \item{x}{A vector upon which to estimate agreement.} \item{grpid}{A vector identifying the groups from which x originated (actual group membership).} \item{nrangrps}{A number representing the number of random groups to generate. Note that the number of random groups created must be directly divisible by the number of actual groups to ensure that group sizes of pseudo groups and actual groups are identical. The rgr.agree routine will generate the number of pseudo groups that most closely approximates nrangrps given the group size characteristics of the data.} } \value{An object of class 'rgr.agree' with the following components: \item{NRanGrp}{The number of random groups created.} \item{AvRGRVar}{The average within-group variance of the random groups.} \item{SDRGRVar}{Standard deviation of random group variances used in the z-score estimate.} \item{zvalue}{Z-score difference between the actual group and random group variances.} \item{RGRVARS}{The random group variances.} } \author{ Paul Bliese \email{paul.bliese@moore.sc.edu} } \references{ Bliese, P. D., & Halverson, R. R. (2002). Using random group resampling in multilevel research. Leadership Quarterly, 13, 53-68. Bliese, P.D., Halverson, R. R., & Rothberg, J. (2000). Using random group resampling (RGR) to estimate within-group agreement with examples using the statistical language R. Walter Reed Army Institute of Research. Ludtke, O. & Robitzsch, A. (2009). Assessing within-group agreement: A critical examination of a random-group resampling approach. Organizational Research Methods, 12, 461-487. } \seealso{ \code{\link{rwg}} \code{\link{rwg.j}} } \examples{ data(bh1996) RGROUT<-rgr.agree(bh1996$HRS,bh1996$GRP,1000) summary(RGROUT) } \keyword{attribute}multilevel/man/sim.icc.Rd0000644000176200001440000000377712750500660015042 0ustar liggesusers\name{sim.icc} \alias{sim.icc} \title{Simulate 2-level ICC(1) values with and without level-1 correlation} \description{ICC(1) values play an important role influencing the form of relationships among variables in nested data. This simulation allows one to create data with known ICC(1) values. Multiple variables can be created both with and without level-1 correlation. } \usage{ sim.icc(gsize, ngrp, icc1,nitems=1,item.cor=FALSE) } \arguments{ \item{gsize}{The simulated group size.} \item{ngrp}{The simulated number of groups.} \item{icc1}{The simulated ICC(1) value.} \item{nitems}{The number of items (vectors) to simulate.} \item{item.cor}{An option to create level-1 correlation among items. Provided as a value between 0 and 1. If used, nitems must be larger than 1.} } \value{ \item{GRP}{The grouping designator.} \item{VAR1}{The simulated value. Multiple numbered columns if nitems>1} } \author{ Paul Bliese \email{paul.bliese@moore.sc.edu} } \references{ Bliese, P. D. (2000). Within-group agreement, non-independence, and reliability: Implications for data aggregation and analysis. In K. J. Klein & S. W. Kozlowski (Eds.), Multilevel Theory, Research, and Methods in Organizations (pp. 349-381). San Francisco, CA: Jossey-Bass, Inc. } \seealso{ \code{\link{ICC1}} } \examples{ \dontrun{ set.seed(1535324) ICC.SIM<-sim.icc(gsize=10,ngrp=100,icc1=.15) ICC1(aov(VAR1~as.factor(GRP), ICC.SIM)) # 4 items with no level-1 correlation set.seed(15324) ICC.SIM<-sim.icc(gsize=10,ngrp=100,icc1=.15,nitems=4) #items with no level-1 correlation mult.icc(ICC.SIM[,2:5],ICC.SIM$GRP) with(ICC.SIM,waba(VAR1,VAR2,GRP))$Cov.Theorem #Examine CorrW # 4 items with a level-1 correlation of .30 set.seed(15324) ICC.SIM<-sim.icc(gsize=10,ngrp=100,icc1=.15,nitems=4, item.cor=.3) #.30 level-1 item correlations mult.icc(ICC.SIM[,2:5],ICC.SIM$GRP) with(ICC.SIM,waba(VAR1,VAR2,GRP))$Cov.Theorem #Examine CorrW } } \keyword{datagen}multilevel/man/ICC2.Rd0000644000176200001440000000216312750500651014161 0ustar liggesusers\name{ICC2} \alias{ICC2} \title{Function to Estimate Intraclass Correlation Coefficient 2 or ICC(2) from an aov model} \description{This function calculates the Intraclass Correlation Coefficient 2 or ICC(2) from an ANOVA model. This value represents the reliability of the group means. } \usage{ ICC2(object) } \arguments{ \item{object}{An ANOVA (aov) object from an one-way analysis of variance.} } \value{Provides an estimate of ICC(1) for the sample.} \references{ Bliese, P. D. (2000). Within-group agreement, non-independence, and reliability: Implications for data aggregation and Analysis. In K. J. Klein & S. W. Kozlowski (Eds.), Multilevel Theory, Research, and Methods in Organizations (pp. 349-381). San Francisco, CA: Jossey-Bass, Inc. Bartko, J.J. (1976). On various intraclass correlation reliability coefficients. Psychological Bulletin, 83, 762-765.} \author{ Paul Bliese \email{paul.bliese@moore.sc.edu} } \seealso{ \code{\link{ICC1}} \code{\link{aov}} } \examples{ data(bh1996) hrs.mod<-aov(HRS~as.factor(GRP),data=bh1996) ICC2(hrs.mod) } \keyword{attribute} multilevel/man/rmv.blanks.Rd0000644000176200001440000000154212750500656015563 0ustar liggesusers\name{rmv.blanks} \alias{rmv.blanks} \title{Remove blanks spaces from non-numeric variables imported from SPSS dataframes} \description{When large SPSS datasets are imported into R, non-numeric fields frequently have numerous blank spaces prior to the text. The blank spaces make it difficult to summarize non-numeric text. The function is applied to an entire dataframe and removes the blank spaces. } \usage{ rmv.blanks(object) } \arguments{ \item{object}{Typically a dataframe created from an imported SPSS file.} } \value{Returns a new dataframe without preceeding } \author{ Paul Bliese \email{paul.bliese@moore.sc.edu} } \seealso{ \code{\link[foreign]{read.spss}} } \examples{ \dontrun{library(foreign) mydata<-read.spss(file.choose(),to.data.frame=T,use.value.labels=F) mydata<-rmv.blanks(mydata)} } \keyword{manip} multilevel/man/summary.rgr.waba.Rd0000644000176200001440000000162612750500662016705 0ustar liggesusers\name{summary.rgr.waba} \alias{summary.rgr.waba} \title{S3 method for class 'rgr.waba'} \description{ This function provides a concise summary of objects created using the function rgr.waba. } \usage{ \method{summary}{rgr.waba}(object,\dots) } \arguments{ \item{object}{An object of class 'rgr.waba'.} \item{\dots}{Optional additional arguments. None used.} } \value{A dataframe containing summary statistics in the form of number of repetitions (NRep), Mean and Standard Deviations (SD) for each parameter in the rgr.waba model. } \author{ Paul Bliese \email{paul.bliese@moore.sc.edu} } \seealso{ \code{\link{rgr.waba}} } \examples{ data(bh1996) #estimate the actual group model waba(bh1996$HRS,bh1996$WBEING,bh1996$GRP) #create 100 pseudo group runs and summarize results RWABA<-rgr.waba(bh1996$HRS,bh1996$WBEING,bh1996$GRP,100) summary(RWABA) } \keyword{programming}multilevel/man/univbct.Rd0000644000176200001440000000355712750500663015166 0ustar liggesusers\name{univbct} \docType{data} \alias{univbct} \title{Data from Bliese and Ployhart (2002)} \description{ This data set contains the complete data set used in Bliese and Ployhart (2002). The data is longitudinal data converted to univariate (i.e., stacked) form. Data were collected at three time points. } \usage{data(univbct)} \format{A data frame with 22 columns and 1485 observations from 495 individuals \tabular{llll}{ [,1] \tab BTN \tab numeric \tab BN Id\cr [,2] \tab COMPANY \tab numeric \tab Co Id\cr [,3] \tab MARITAL \tab numeric \tab Marital Status\cr [,4] \tab GENDER \tab numeric \tab Gender\cr [,5] \tab HOWLONG \tab numeric \tab Time in Unit \cr [,6] \tab RANK \tab numeric \tab Rank\cr [,7] \tab EDUCATE \tab numeric \tab Education\cr [,8] \tab AGE \tab numeric \tab Age\cr [,9] \tab JOBSAT1 \tab numeric \tab JOBSAT Time 1\cr [,10] \tab COMMIT1 \tab numeric \tab Commitment Time 1\cr [,11] \tab READY1 \tab numeric \tab Readiness Time 1\cr [,12] \tab JOBSAT2 \tab numeric \tab JOBSAT Time 2\cr [,13] \tab COMMIT2 \tab numeric \tab Commitment Time 2\cr [,14] \tab READY2 \tab numeric \tab Readiness Time 2\cr [,15] \tab JOBSAT3 \tab numeric \tab JOBSAT Time 3\cr [,16] \tab COMMIT3 \tab numeric \tab Commitment Time 3\cr [,17] \tab READY3 \tab numeric \tab Readiness Time 3\cr [,18] \tab TIME \tab numeric \tab 0 to 2 time maker\cr [,19] \tab JSAT \tab numeric \tab Jobsat in univariate form \cr [,20] \tab COMMIT \tab numeric \tab Commitment in univariate form\cr [,21] \tab READY \tab numeric \tab Readiness in univariate form \cr [,22] \tab SUBNUM \tab numeric \tab Subject number } } \references{ Bliese, P. D., & Ployhart, R. E. (2002). Growth modeling using random coefficient models: Model building, testing and illustrations. Organizational Research Methods, 5, 362-387.} \keyword{datasets}multilevel/man/boot.icc.Rd0000644000176200001440000000300612750500647015203 0ustar liggesusers\name{boot.icc} \alias{boot.icc} \title{Bootstrap ICC values in 2-level data} \description{ Implements a 2-level bootstrap. The bootstrap first draws a sample of level-2 units with replacement, and in a second stage draws a sample of level-1 observations with replacement from the level-2 units. Following each bootstrap replication, the Intraclass Correlation Coefficient 1 is estimated using the lme function. } \usage{ boot.icc(x, grpid, nboot, aov.est=FALSE) } \arguments{ \item{x}{A vector representing the variable upon which to estimate the ICC values.} \item{grpid}{A vector representing the level-2 unit identifier.} \item{nboot}{The number of bootstrap iterations. Computational demands underlying a 2-level bootstrap are heavy, so the examples use 100; however, the number of interations should generally be 10,000.} \item{aov.est}{An option to estimate the ICC using aov.} } \value{Provides ICC(1) estimates for each bootstrap draw.} \author{ Paul Bliese \email{paul.bliese@moore.sc.edu} } \references{ Bliese, P. D. (2000). Within-group agreement, non-independence, and reliability: Implications for data aggregation and analysis. In K. J. Klein & S. W. Kozlowski (Eds.), Multilevel Theory, Research, and Methods in Organizations (pp. 349-381). San Francisco, CA: Jossey-Bass, Inc.} \seealso{ \code{\link{ICC1}} \code{\link{ICC2}} } \examples{ \dontrun{ data(bh1996) ICC.OUT<-boot.icc(bh1996$WBEING,bh1996$GRP,100) quantile(ICC.OUT,c(.025,.975)) } } \keyword{attribute}multilevel/man/rgr.waba.Rd0000644000176200001440000000475312750500656015220 0ustar liggesusers\name{rgr.waba} \alias{rgr.waba} \title{Random Group Resampling of Covariance Theorem Decomposition} \description{ This routine performs the covariance theorem decomposition discussed by Robinson (1950) and Dansereau, Alutto and Yammarino (1984), but builds upon this work by incorporating Random Group Resampling or RGR. RGR is used to randomly assign individuals to pseudo groups. This creates sampling distributions of the covariance theorem components, and allows one to contrast actual group covariance components to pseudo group covariance components. Note that rgr.waba is a labor intensive routine. } \usage{ rgr.waba(x, y, grpid, nrep) } \arguments{ \item{x}{A vector representing one variable for the correlation.} \item{y}{A vector representing the other variable for the correlation.} \item{grpid}{A vector identifying the groups from which X and Y originated.} \item{nrep}{The number of times that the entire data set is reassigned to pseudo groups} } \value{ Returns an object of class rgr.waba. The object is a list containing each random run for each component of the covariance theorem. } \author{ Paul Bliese \email{paul.bliese@moore.sc.edu} } \references{ Bliese, P. D. & Halverson, R. R. (1996). Individual and nomothetic models of job stress: An examination of work hours, cohesion, and well- being. Journal of Applied Social Psychology, 26, 1171-1189. Bliese, P. D., & Halverson, R. R. (2002). Using random group resampling in multilevel research. Leadership Quarterly, 13, 53-68. Dansereau, F., Alutto, J. A., & Yammarino, F. J. (1984). Theory testing in organizational behavior: The varient approach. Englewood Cliffs, NJ: Prentice-Hall. Robinson, W. S. (1950). Ecological correlations and the behavior of individuals. American Sociological Review, 15, 351-357. } \seealso{ \code{\link{waba}} } \examples{ # This example is from Bliese & Halverson (1996). Notice that all of the # values from the RGR analysis differ from the values based on actual # group membership. Confidence intervals for individual components can # be estimated using the quantile command. data(bh1996) #estimate the actual group model waba(bh1996$HRS,bh1996$WBEING,bh1996$GRP) #create 100 pseudo group runs and summarize the model RWABA<-rgr.waba(bh1996$HRS,bh1996$WBEING,bh1996$GRP,100) summary(RWABA) #Estimate 95th percentile confidence intervals (p=.05) quantile(RWABA,c(.025,.975)) } \keyword{attribute}multilevel/man/sherifdat.Rd0000644000176200001440000000316012750500660015450 0ustar liggesusers\name{sherifdat} \docType{data} \alias{sherifdat} \title{Sherif (1935) group data from 3 person teams} \description{ This data set contains estimates of movement length (in inches) of a light in a completely dark room. Eight groups of three individuals provided three estimates for a total of 72 observations. In four of the groups, participants first made estimates alone prior to providing estimates as a group. In the other four groups participants started as groups. Lang and Bliese (forthcoming) used these data to illustrate how variance functions in mixed-effects models (lme) could be used to test whether groups displayed consensus emergence. Data were obtained from https://brocku.ca/MeadProject/Sherif/Sherif_1935a/Sherif_1935a_3.html } \usage{data(sherifdat)} \format{A dataframe with 5 columns and 72 observations \tabular{llll}{ [,1] \tab person \tab numeric \tab Participant ID within a group\cr [,2] \tab time \tab numeric \tab Measurment Occasion\cr [,3] \tab group \tab numeric \tab Group Identifier \cr [,4] \tab y \tab numeric \tab Estimate of movement length in inches \cr [,4] \tab condition \tab numeric \tab Experimental Condition for either starting individually (1) or as a group (0)\cr } } \references{ Sherif, M. (1935). A study of some social factors in perception: Chapter 3. Archives of Psychology, 27, 23- 46. https://brocku.ca/MeadProject/Sherif/Sherif_1935a/Sherif_1935a_3.html Lang, J. W. B., & Bliese, P. D. (forthcoming). A Temporal Perspective on Emergence: Using 3-level Mixed Effects Models to Track Consensus Emergence in Groups. } \keyword{datasets}multilevel/man/quantile.disagree.sim.Rd0000644000176200001440000000266712750500654017710 0ustar liggesusers\name{quantile.disagree.sim} \alias{quantile.disagree.sim} \title{S3 method for class 'disagree.sim'} \description{ This function provides a concise quantile summary of objects created using the function ad.m.sim. The simulation functions for the average deviation of the mean (or median) return a limited number of estimated values. Consequently, the normal quantile methods are biased. The quantile methods incorporated in this function produce unbiased estimates. } \usage{ \method{quantile}{disagree.sim}(x,confint,\dots) } \arguments{ \item{x}{An object of class 'disagree.sim'.} \item{confint}{The confidence intervals to return. The values of 0.05 and 0.01 return the approximate 5 percent and 1 percent confidence intervals. Values equal to or smaller than these values are significant (p=.05, p=.01).} \item{\dots}{Optional arguments. None used.} } \value{A dataframe with two columns. The first column contains the quantile value and the second contains the estimate based on the object. } \author{ Paul Bliese \email{paul.bliese@moore.sc.edu} } \seealso{ \code{\link{ad.m.sim}} } \examples{ #Example from Dunlap et al. (2003), Table 3. The listed significance #value (p=.05) for a group of size 5 with a 7-item response format is #0.64 or less. SIMOUT<-ad.m.sim(gsize=5, nitems=1, nresp=7, itemcors=NULL, type="mean", nrep=1000) quantile(SIMOUT, c(.05,.01)) } \keyword{programming}multilevel/man/quantile.agree.sim.Rd0000644000176200001440000000240312750500653017173 0ustar liggesusers\name{quantile.agree.sim} \alias{quantile.agree.sim} \title{S3 method for class 'agree.sim'} \description{ This function provides a concise quantile summary of objects created using the functions rwg.sim and rwg.j.sim. The simulation functions for rwg and rwg.j return a limited number of estimated values. Consequently, the normal quantile methods are biased. The quantile methods incorporated in this function produce unbiased estimates. } \usage{ \method{quantile}{agree.sim}(x,confint,\dots) } \arguments{ \item{x}{An object of class 'agree.sim'.} \item{confint}{The confidence intervals to return. The values of 0.95 and 0.99 return the approximate 95th and 99th percentile confidence intervals (p=.05 and p=.01).} \item{\dots}{Optional arguments. None used.} } \value{A dataframe with two columns. The first column contains the quantile value and the second contains the estimate based on the object. } \author{ Paul Bliese \email{paul.bliese@moore.sc.edu} } \seealso{ \code{\link{rwg.sim}} \code{\link{rwg.j.sim}} } \examples{ #An example from Dunlap et al. (2003). The estimate from Dunlap et al. #Table 2 is 0.53 RWG.OUT<-rwg.sim(gsize=10,nresp=5,nrep=1000) quantile(RWG.OUT, c(.95,.99)) } \keyword{programming}multilevel/man/graph.ran.mean.Rd0000644000176200001440000000513412750500650016300 0ustar liggesusers\name{graph.ran.mean} \alias{graph.ran.mean} \title{Graph Random Group versus Actual Group distributions} \description{This function uses random group resampling (RGR) to create a distribution of pseudo group means. The pseudo group means are then contrasted with actual group means to provide a visualization of the group-level properties of the data. It is, in essense, a way of visualizing an Intraclass Correlation Coefficient -- ICC(1).} \usage{ graph.ran.mean(x, grpid, nreps, limits, graph=TRUE, bootci=FALSE) } \arguments{ \item{x}{The vector representing the construct of interest.} \item{grpid}{A vector identifying the groups associated with x.} \item{nreps}{A number representing the number of random groups to generate. Because groups are created with the exact size characteristics of the actual groups, the total number of pseudo groups created may be calculated as nreps * Number Actual Groups. The value chosen for nreps only affects the smoothness of the pseudo group line -- values greater than 25 should provide sufficiently smooth lines. Values of 1000 should be used if the bootci option is TRUE although only 25 are used in the example to reduce computation time.} \item{limits}{Controls the upper and lower limits of the y-axis on the plot. The default is to set the limits at the 10th and 90th percentiles of the raw data. This option only affects how the data is plotted.} \item{graph}{Controls whether or not a plot is returned. If graph=FALSE, the program returns a data frame with two columns. The first column contains the sorted means from the actual groups, and the second column contains the sorted means from the pseudo groups. This can be useful for plotting results in other programs.} \item{bootci}{Determines whether approximate 95 percent confidence interval estimates are calculated and plotted. If bootci is TRUE, the nreps option should be 1000 or more.} } \value{Produces either a plot (graph=TRUE) or a data.frame (graph=FALSE)} \references{Bliese, P. D., & Halverson, R. R. (2002). Using random group resampling in multilevel research. Leadership Quarterly, 13, 53-68.} \author{ Paul Bliese \email{paul.bliese@moore.sc.edu}} \seealso{ \code{\link{ICC1}} \code{\link{mix.data}} } \examples{ data(bh1996) # with the bootci=TRUE option, nreps should be 1000 or more. The value # of 25 is used in the example to reduce computation time with(bh1996,graph.ran.mean(HRS,GRP,limits=c(8,16),nreps=25, bootci=TRUE)) GRAPH.DAT<-graph.ran.mean(bh1996$HRS,bh1996$GRP,limits=c(8,16),nreps=25, graph=FALSE) } \keyword{dplot}multilevel/man/make.univ.Rd0000644000176200001440000000450612750500652015402 0ustar liggesusers\name{make.univ} \alias{make.univ} \title{Convert data from multivariate to univariate form} \description{ Longitudinal data is often stored in multivariate or wide form. In multivariate form, each row contains data from one subject, and repeated measures variables are indexed by different names (e.g., OUTCOME.T1, OUTCOME.T2, OUTCOME.T3). In repeated measures designs and growth modeling, data often needs to be converted to univariate or stacked form where each row represents one of the repeated measures indexed by a TIME variable nested within subject. In univariate form, each subject has as many rows of data as there are time points. R has several functions to convert data from wide to long formats and vice versa including reshape. The code used in make.univ borrows heavily from code provided in Chambers and Hastie (1991). the } \usage{ make.univ(x,dvs,tname="TIME", outname="MULTDV") } \arguments{ \item{x}{A dataframe in multivariate form.} \item{dvs}{A subset dataframe of x containing the repeated measures columns. Note that the repeated measures must be ordered from Time 1 to Time N for this function to work properly.} \item{tname}{An optional name for the new time variable. Defaults to TIME.} \item{outname}{An optional name for the outcome variable name. Defaults to MULTDV.} } \value{ Returns a dataframe in univariate (i.e., stacked) form with a TIME variable representing the repeated observations, and a variable named MULTDV representing the time-indexed variable. The TIME variable begins with 0. } \author{ Paul Bliese \email{paul.bliese@moore.sc.edu} } \references{ Bliese, P. D., & Ployhart, R. E. (2002). Growth modeling using random coefficient models: Model building, testing and illustrations. Organizational Research Methods, 5, 362-387. Chambers, J. M., & Hastie, T. J. (1991). Statistical models in S. CRC Press, Inc.. } \seealso{ \code{\link{mult.make.univ}} \code{\link{reshape}} } \examples{ data(univbct) #a dataframe in univariate form for job satisfaction TEMP<-univbct[3*1:495,c(22,1:17)] #convert back to multivariate form #Transform data to univariate form TEMP2<-make.univ(x=TEMP,dvs=TEMP[,c(10,13,16)]) #Same as above, but renaming repeated variable TEMP3<-make.univ(x=TEMP,dvs=TEMP[,c(10,13,16)],outname="JOBSAT") } \keyword{manip}multilevel/man/waba.Rd0000644000176200001440000000577112750500663014426 0ustar liggesusers\name{waba} \alias{waba} \title{Covariance Theoreom Decomposition of Bivariate Two-Level Correlation} \description{ This routine performs the covariance theorem decomposition discussed by Robinson (1950) and Dansereau, Alutto and Yammarino (1984). Dansereau et al. have labeled the variance decomposition Within-And-Between-Analysis II or WABA II. The program decomposes a raw correlation from a two-level nested design into 6 components. These components are (1) eta-between value for X, (2) eta-between value for Y, (3) the group-size weighted group-mean correlation, (4) the within-eta value for X, (5) the within-eta value for Y, and (6) the within-group correlation between X and Y. The last value represents the correlation between X and Y after each variable has been group-mean centered. The program is designed to automatically perform listwise deletion on missing values; consequently, users should pay attention to the diagnostic information (Number of Groups and Number of Observations) provided as part of the output. Note that Within-And-Between-Analysis proposed by Dansereau et al. involves more than covariance theorem decomposition of correlations. Specifically, WABA involves decision rules based on eta-values. These are not replicated in the R multilevel library because the eta based decision rules have been shown to be highly related to group size (Bliese, 2000; Bliese & Halverson, 1998), a factor not accounted for in the complete Within-And-Between-Analysis methodology. } \usage{ waba(x, y, grpid) } \arguments{ \item{x}{A vector representing one variable in the correlation.} \item{y}{A vector representing the other variable in the correlation.} \item{grpid}{A vector identifying the groups from which x and y originated.} } \value{ Returns a list with three elements. \item{Cov.Theorem}{A 1 row dataframe with all of the elements of the covariance theorem.} \item{n.obs}{The number of observations used to calculate the covariance theorem.} \item{n.grps}{The number of groups in the data set.} } \author{ Paul Bliese \email{paul.bliese@moore.sc.edu} } \references{ Bliese, P. D. (2000). Within-group agreement, non-independence, and reliability: Implications for data aggregation and Analysis. In K. J. Klein & S. W. Kozlowski (Eds.), Multilevel Theory, Research, and Methods in Organizations (pp. 349-381). San Francisco, CA: Jossey-Bass, Inc. Bliese, P. D., & Halverson, R. R. (1998). Group size and measures of group-level properties: An examination of eta-squared and ICC values. Journal of Management, 24, 157-172. Dansereau, F., Alutto, J. A., & Yammarino, F. J. (1984). Theory testing in organizational behavior: The varient approach. Englewood Cliffs, NJ: Prentice-Hall. Robinson, W. S. (1950). Ecological correlations and the behavior of individuals. American Sociological Review, 15, 351-357. } \seealso{ \code{\link{rgr.waba}} } \examples{ data(bh1996) waba(bh1996$HRS,bh1996$WBEING,bh1996$GRP) } \keyword{attribute} multilevel/man/quantile.rgr.waba.Rd0000644000176200001440000000205112750500654017024 0ustar liggesusers\name{quantile.rgr.waba} \alias{quantile.rgr.waba} \title{S3 method for class 'rgr.waba'} \description{ This function provides a concise quantile summary of objects created using the function rgr.waba. } \usage{ \method{quantile}{rgr.waba}(x,confint,\dots) } \arguments{ \item{x}{An object of class 'rgr.waba'.} \item{confint}{The confidence intervals to return. The values of 0.025 and 0.975 return the approximate two-tailed 95th percentile confidence intervals (p=.05).} \item{\dots}{Optional arguments. None used.} } \value{A dataframe containing the confidence intervals for each parameter in the rgr.waba model. } \author{ Paul Bliese \email{paul.bliese@moore.sc.edu} } \seealso{ \code{\link{rgr.waba}} } \examples{ data(bh1996) #estimate the model based on actual group membership waba(bh1996$HRS,bh1996$WBEING,bh1996$GRP) #create 100 pseudo group runs and summarize RWABA<-rgr.waba(bh1996$HRS,bh1996$WBEING,bh1996$GRP,100) quantile(RWABA,confint=c(.025,.975)) } \keyword{programming}multilevel/man/rwg.Rd0000644000176200001440000000445212750500657014311 0ustar liggesusers\name{rwg} \alias{rwg} \title{James et al., (1984) agreement index for single item measures} \description{ This function calculates the within group agreement measure rwg for single item measures as described in James, Demaree and Wolf (1984). The rwg is calculated as rwg = 1-(Observed Group Variance/Expected Random Variance). James et al. (1984) recommend truncating the Observed Group Variance to the Expected Random Variance in cases where the Observed Group Variance was larger than the Expected Random Variance. This truncation results in an rwg value of 0 (no agreement) for groups with large variances. } \usage{ rwg(x, grpid, ranvar=2) } \arguments{ \item{x}{A vector representing the construct on which to estimate agreement.} \item{grpid}{A vector identifying the groups from which x originated.} \item{ranvar}{The random variance to which actual group variances are compared. The value of 2 represents the variance from a rectangular distribution in the case where there are 5 response options (e.g., Strongly Disagree, Disagree, Neither, Agree, Strongly Agree). In cases where there are not 5 response options, the rectangular distribution is estimated using the formula \eqn{\mathtt{ranvar}=(A^{2}-1)/12}{ranvar=(A^2-1)/12} where A is the number of response options. While the rectangular distribution is widely used, other random values may be more appropriate.} } \value{ \item{grpid}{The group identifier} \item{rwg}{The rwg estimate for the group} \item{gsize}{The group size} } \author{Paul Bliese \email{paul.bliese@moore.sc.edu} } \references{ Bliese, P. D. (2000). Within-group agreement, non-independence, and reliability: Implications for data aggregation and analysis. In K. J. Klein & S. W. Kozlowski (Eds.), Multilevel Theory, Research, and Methods in Organizations (pp. 349-381). San Francisco, CA: Jossey-Bass, Inc. James, L.R., Demaree, R.G., & Wolf, G. (1984). Estimating within-group interrater reliability with and without response bias. Journal of Applied Psychology, 69, 85-98. } \seealso{ \code{\link{ad.m}} \code{\link{rwg.j}} \code{\link{rwg.sim}} \code{\link{rgr.agree}} \code{\link{rwg.j.lindell}} } \examples{ data(lq2002) RWGOUT<-rwg(lq2002$LEAD,lq2002$COMPID) RWGOUT[1:10,] summary(RWGOUT) } \keyword{attribute}multilevel/man/ad.m.Rd0000644000176200001440000000422212750500646014322 0ustar liggesusers\name{ad.m} \alias{ad.m} \title{Average deviation around mean or median} \description{ This function calculates the average deviation of the mean or median as a measure of within-group agreement as proposed by Burke, Finkelstein and Dusig (1999). A basic rule for interpreting whether or not the results display practically significant levels of agreement is whether the AD value is smaller than A/6 where A represents the number of response options. For instance, A would be 5 on a five-point response option format of strongly disagree, disagree, neither, agree, strongly agree (see Dunlap, Burke & Smith-Crowe, 2003). To estimate statistical significance see the ad.m.sim function and help files. } \usage{ ad.m(x, grpid, type="mean") } \arguments{ \item{x}{A vector representing a single item or a matrix representing a scale of interest. If a matrix, each column of the matrix represents a scale item, and each row represents an individual respondent.} \item{grpid}{A vector identifying the groups from which x originated.} \item{type}{A character string for either the mean or median.} } \value{ \item{grpid}{The group identifier} \item{AD.M}{The average deviation around the mean or median for each group} \item{gsize}{Group size} } \author{ Paul Bliese \email{paul.bliese@moore.sc.edu} } \references{ Burke, M. J., Finkelstein, L. M., & Dusig, M. S. (1999). On average deviation indices for estimating interrater agreement. Organizational Research Methods, 2, 49-68. Dunlap, W. P., Burke, M. J., & Smith-Crowe, K. (2003). Accurate tests of statistical significance for rwg and average deviation interrater agreement indices. Journal of Applied Psychology, 88, 356-362. } \seealso{ \code{\link{ad.m.sim}} \code{\link{rwg}} \code{\link{rwg.j}} \code{\link{rgr.agree}} \code{\link{rwg.sim}} \code{\link{rwg.j.sim}} } \examples{ data(bhr2000) #Examples for multiple item scales AD.VAL<-ad.m(bhr2000[,2:12],bhr2000$GRP) AD.VAL[1:5,] summary(AD.VAL) summary(ad.m(bhr2000[,2:12],bhr2000$GRP,type="median")) #Example for single item measure summary(ad.m(bhr2000$HRS,bhr2000$GRP)) } \keyword{attribute}multilevel/man/klein2000.Rd0000644000176200001440000000265212750500651015110 0ustar liggesusers\name{klein2000} \docType{data} \alias{klein2000} \title{Data from Klein, Bliese, Kozlowski et al., (2000)} \description{ This data set contains the complete data used in Klein et al. (2000). The Klein et al. chapter uses a simulated data set to compare and contrast WABA, HLM, and Cross-Level Operator Analyses (CLOP). The simulated data set was created by Paul Bliese. } \usage{data(klein2000)} \format{A data frame with 9 columns and 750 observations from 50 groups \tabular{llll}{ [,1] \tab GRPID \tab numeric \tab Group Identifier\cr [,2] \tab JOBSAT \tab numeric \tab Job Satisfaction (DV)\cr [,3] \tab COHES \tab numeric \tab Cohesion\cr [,4] \tab POSAFF \tab numeric \tab Positive Affect\cr [,5] \tab PAY \tab numeric \tab Pay \cr [,6] \tab NEGLEAD \tab numeric \tab Negative Leadership\cr [,7] \tab WLOAD \tab numeric \tab Workload\cr [,8] \tab TASKSIG \tab numeric \tab Task Significance\cr [,9] \tab PHYSEN \tab numeric \tab Physical Environment } } \references{ Klein, K. J., Bliese, P.D., Kozlowski, S. W. J, Dansereau, F., Gavin, M. B., Griffin, M. A., Hofmann, D. A., James, L. R., Yammarino, F. J., & Bligh, M. C. (2000). Multilevel analytical techniques: Commonalities, differences, and continuing questions. In K. J. Klein & S. W. Kozlowski (Eds.), Multilevel Theory, Research, and Methods in Organizations (pp. 512-553). San Francisco, CA: Jossey-Bass, Inc} \keyword{datasets}multilevel/man/sam.cor.Rd0000644000176200001440000000165312750500660015046 0ustar liggesusers\name{sam.cor} \alias{sam.cor} \title{Generate a Sample that Correlates with a Fixed Set of Observations} \description{This function will generate a vector (y) with a known correlation to a given vector (x). The degree of correlation between x and y is determined by the parameter rho (the population correlation). Observed sample correlations between x and y will vary around rho, but this variation will decrease as the size of x increases.} \usage{ sam.cor(x,rho) } \arguments{ \item{x}{The given vector.} \item{rho}{Population correlation.} } \value{The function prints the sample correlation for the specific set of numbers generated. \item{y}{A vector of numbers correlated with x.} } \author{ Paul Bliese \email{paul.bliese@moore.sc.edu}} \seealso{\code{\link{simbias}}} \examples{ data(bh1996) NEWVAR<-sam.cor(x=bh1996$LEAD,rho=.30) cor(bh1996$LEAD,NEWVAR) } \keyword{programming}