1 The SAS System 11:43 Thursday, October 30, 2003 NOTE: Copyright (c) 1999-2001 by SAS Institute Inc., Cary, NC, USA. NOTE: SAS (r) Proprietary Software Release 8.2 (TS2M0) Licensed to UNC-CHAPEL HILL-SYSTEMWIDE-T/R, Site 0039466002. NOTE: This session is executing on the WIN_PRO platform. NOTE: SAS initialization used: real time 0.23 seconds cpu time 0.17 seconds NOTE: AUTOEXEC processing beginning; file is c:\local\sas 8x\autoexec.sas. NOTE: AUTOEXEC processing completed. 1 /* file: ~jpreisse/DIA/UI/diag_uigee_ind.sas */ 2 /* date: 06/28/03 */ 3 /* name: John Preisser */ 4 /* what: Logitstic Regression Diagnostics for UI data using GEE macro */ 5 /* Program shows that same diagnostics can be obtained using GEE macro */ 6 /* as compared to PROC LOGISTIC for independent data */ 7 /* input data files: ~/home/projects/gts/DG/UI.dat */ 8 9 10 *filename geed "/home/faculty/jpreisse/DIA/geedia_100.mac"; 11 filename geed "diag100.sas"; 12 options fmtsearch=(work.format, formats.formats, fmt.format); 13 *options ls=72 ps=58 Mprint; 14 options ls=72 ps=58; 15 16 data UI; 17 input pract_id doct_id pat_id bothered female ageyrs age 17 ! weekacc dayacc severe toilet mdage dobs rit hit; 18 cards; NOTE: The data set WORK.UI has 137 observations and 15 variables. NOTE: DATA statement used: real time 0.04 seconds cpu time 0.04 seconds 156 ; 157 158 /*********************************************/ 159 /* THIS PART WILL USE GENMOD TO FIND DFBETAS */ 160 /*********************************************/ 161 data UI; 162 set UI; 163 inter=1; 164 *if pat_id in (8,44,122) then delete; 165 166 /***********************************/ 167 /* IF MISSING THEN DELETE */ 168 /***********************************/ 2 The SAS System 11:43 Thursday, October 30, 2003 169 if (bothered=. | female=. |age=. | dayacc=. | severe=. | 169 ! toilet=.) then delete; 170 run; NOTE: There were 137 observations read from the data set WORK.UI. NOTE: The data set WORK.UI has 137 observations and 16 variables. NOTE: DATA statement used: real time 0.03 seconds cpu time 0.03 seconds 171 172 proc sort data=UI; 173 by pat_id; 174 run; NOTE: There were 137 observations read from the data set WORK.UI. NOTE: The data set WORK.UI has 137 observations and 16 variables. NOTE: PROCEDURE SORT used: real time 0.03 seconds cpu time 0.03 seconds 175 176 /************************************************************ 176 ! ******************/ 177 /* THIS PART OF THE SAS CODE OUTPUTS RESIDUAL VALUES USED TO 177 ! CALCULATE DFBETA */ 178 /************************************************************ 178 ! ******************/ 179 /* 180 ods output covB=covB ObStats=ObStats 180 ! ParameterEstimates=ParameterEstimates; 181 proc genmod data=UI descending; 182 model bothered = female age dayacc severe toilet/ dist=bin 182 ! link=logit r covB; 183 output out=gencook stdresdev=stdresdev resdev=resdev 183 ! stdxbeta=stdxbeta reschi=reschi stdreschi=stdreschi; 184 run; 185 */ 186 187 /************************************************************ 187 ! ******************/ 188 /* THIS PART OF THE SAS CODE USES PROC LOGISTIC TO CALCULATE 188 ! THE DFBETAS */ 189 /************************************************************ 189 ! ******************/ 190 191 192 proc logistic data=UI descending; 193 model bothered = female age dayacc severe toilet; 194 output out=logcook cbar=cbar dfbeta=_all_ H=H; 195 run; NOTE: PROC LOGISTIC is modeling the probability that bothered=1. 3 The SAS System 11:43 Thursday, October 30, 2003 NOTE: Convergence criterion (GCONV=1E-8) satisfied. NOTE: There were 137 observations read from the data set WORK.UI. NOTE: The data set WORK.LOGCOOK has 137 observations and 24 variables. NOTE: The PROCEDURE LOGISTIC printed pages 1-2. NOTE: PROCEDURE LOGISTIC used: real time 0.07 seconds cpu time 0.07 seconds 196 197 198 data logcook2; 199 set logcook; 200 cookdobs = round(cbar/6,.00001); 201 run; NOTE: There were 137 observations read from the data set WORK.LOGCOOK. NOTE: The data set WORK.LOGCOOK2 has 137 observations and 25 variables. NOTE: DATA statement used: real time 0.03 seconds cpu time 0.03 seconds 202 /************************************************************ 202 ! ******************/ 203 /* THIS PART OF THE SAS CODE PRINTS OUT THE DFBETAS GIVEN BY 203 ! PROC LOGISTIC */ 204 /************************************************************ 204 ! ******************/ 205 206 proc print data=logcook2; 207 title 'Logistic regression diagnostics produced by PROC 207 ! LOGISTIC for UI data'; 208 title2 'Cookdobs is given by Corollary 2.2 of Preisser and 208 ! Qaqish under R=I'; 209 title3 ' '; 210 var DFBETA_intercept DFBETA_female DFBETA_age DFBETA_dayacc 210 ! DFBETA_severe 211 DFBETA_toilet H cbar cookdobs; 212 run; NOTE: There were 137 observations read from the data set WORK.LOGCOOK2. NOTE: The PROCEDURE PRINT printed pages 3-8. NOTE: PROCEDURE PRINT used: real time 0.01 seconds cpu time 0.01 seconds 213 214 /************************************************************ 214 ! ******************/ 215 /* PRODUCE DFBETAS FOR LOGISTIC REGRESSION USING SAS GEE 215 ! MACRO */ 216 /************************************************************ 216 ! ******************/ 4 The SAS System 11:43 Thursday, October 30, 2003 217 218 %include geed; 1370 1371 /* these results confirm output of proc logistic */ 1372 1373 %gee( 1374 data=ui, yvar=bothered, xvar=inter female age dayacc severe 1374 ! toilet, 1375 id=pract_id, link=3, vari=3, corr=1, monitor=no, 1375 ! obsout=infobs, 1376 ncovout=naivecov, rcovout=robustcov 1377 ); Worksize = 30720000 Symbol size = 262128 NOTE: IML Ready NOTE: Module LINKID defined. NOTE: Module DERI defined. NOTE: Module VARFUN defined. NOTE: Module DEVI defined. NOTE: Module CANON defined. NOTE: Module INFO defined. NOTE: Module DESC defined. NOTE: Module FIRSTB defined. NOTE: Module CONSCHK defined. NOTE: Module INIT defined. NOTE: Module FINDINV defined. NOTE: Module GLIM defined. NOTE: Module ESTB defined. NOTE: Module ESTRPHI defined. NOTE: Module XCH defined. NOTE: Module XCHINV defined. NOTE: Module INVCORR defined. NOTE: Module VARIAN defined. NOTE: Module MAIN defined. 1378 1379 NOTE: Exiting IML. NOTE: The data set WORK.NAIVECOV has 6 observations and 9 variables. NOTE: The data set WORK.ROBUSTCOV has 6 observations and 9 variables. NOTE: The data set WORK.INFOBS has 137 observations and 15 variables. NOTE: The PROCEDURE IML printed pages 9-11. NOTE: PROCEDURE IML used: real time 0.18 seconds cpu time 0.18 seconds 1380 proc print data= naivecov; 1381 title 'Model-based covariance estimate'; 1382 title2 ' '; 1383 title3 ' '; 1384 run; NOTE: There were 6 observations read from the data set WORK.NAIVECOV. NOTE: The PROCEDURE PRINT printed page 12. NOTE: PROCEDURE PRINT used: The SAS System real time 0.00 seconds cpu time 0.00 seconds 1385 1386 proc print data= robustcov; 1387 title 'Empirical Sandwich covariance estimate'; 1388 title2 ' '; 1389 title3 ' '; 1390 run; NOTE: There were 6 observations read from the data set WORK.ROBUSTCOV. NOTE: The PROCEDURE PRINT printed page 13. NOTE: PROCEDURE PRINT used: real time 0.00 seconds cpu time 0.00 seconds 1391 1392 * in order to compare with PROC LOGISTIC, DFBETA are 1392 ! standardized; 1393 data naive2(keep = nse_inter nse_female nse_age nse_dayacc 1393 ! nse_severe nse_toilet); 1394 set naivecov(keep = variable naive_se); 1395 if variable='INTER' then nse_inter=naive_se; 1396 else if variable='FEMALE' then nse_female=naive_se; 1397 else if variable='AGE' then nse_age=naive_se; 1398 else if variable='DAYACC' then nse_dayacc=naive_se; 1399 else if variable='SEVERE' then nse_severe=naive_se; 1400 else if variable='TOILET' then nse_toilet=naive_se; 1401 retain nse_inter nse_female nse_age nse_dayacc nse_severe 1401 ! nse_toilet; 1402 if _n_ = 6 then do; Do i=1 to 137; output; End; end; 1403 run; NOTE: There were 6 observations read from the data set WORK.NAIVECOV. NOTE: The data set WORK.NAIVE2 has 137 observations and 6 variables. NOTE: DATA statement used: real time 0.01 seconds cpu time 0.01 seconds 1404 1405 data infobs2; 1406 merge naive2 infobs; 1407 DFBETAS_INTER = inter/nse_inter; 1408 DFBETAS_FEMALE = female/nse_female; 1409 DFBETAS_AGE = age/nse_age; 1410 DFBETAS_DAYACC = dayacc/nse_dayacc; 1411 DFBETAS_SEVERE = severe/nse_severe; 1412 DFBETAS_TOILET = toilet/nse_toilet; 1413 run; NOTE: There were 137 observations read from the data set WORK.NAIVE2. NOTE: There were 137 observations read from the data set WORK.INFOBS. NOTE: The data set WORK.INFOBS2 has 137 observations and 27 variables. The SAS System NOTE: DATA statement used: real time 0.03 seconds cpu time 0.03 seconds 1414 1415 proc print data=infobs2; 1416 title 'Logistic regression diagnostics for UI data from SAS 1416 ! IML macro'; 1417 title2 'Cookdobs is given by Corollary 2.2 of Preisser and 1417 ! Qaqish under R=I'; 1418 title3 'Cookdobs under R=I is equivalent to Williams 1418 ! (1987)'; 1419 title4 'DFBETA from SAS GEE macro is standardized to match 1419 ! PROC LOGISTIC output'; 1420 run; NOTE: There were 137 observations read from the data set WORK.INFOBS2. NOTE: The PROCEDURE PRINT printed pages 14-29. NOTE: PROCEDURE PRINT used: real time 0.03 seconds cpu time 0.03 seconds 1421 1422 NOTE: SAS Institute Inc., SAS Campus Drive, Cary, NC USA 27513-2414 NOTE: The SAS System used: real time 1.35 seconds cpu time 0.74 seconds The SAS System 11:43 Thursday, October 30, 2003 1 The LOGISTIC Procedure Model Information Data Set WORK.UI Response Variable bothered Number of Response Levels 2 Number of Observations 137 Model binary logit Optimization Technique Fisher's scoring Response Profile Ordered Total Value bothered Frequency 1 1 54 2 0 83 Probability modeled is bothered=1. Model Convergence Status Convergence criterion (GCONV=1E-8) satisfied. Model Fit Statistics Intercept Intercept and Criterion Only Covariates AIC 185.737 139.107 SC 188.657 156.627 -2 Log L 183.737 127.107 Testing Global Null Hypothesis: BETA=0 Test Chi-Square DF Pr > ChiSq Likelihood Ratio 56.6300 5 <.0001 Score 46.2975 5 <.0001 Wald 30.3651 5 <.0001 The SAS System 11:43 Thursday, October 30, 2003 2 The LOGISTIC Procedure Analysis of Maximum Likelihood Estimates Standard Wald Parameter DF Estimate Error Chi-Square Pr > ChiSq Intercept 1 -3.2929 1.1083 8.8276 0.0030 female 1 -0.6723 0.6116 1.2083 0.2717 age 1 -0.6405 0.5847 1.2000 0.2733 dayacc 1 0.4154 0.0958 18.8056 <.0001 severe 1 0.8285 0.3645 5.1652 0.0230 toilet 1 0.1108 0.0855 1.6788 0.1951 Odds Ratio Estimates Point 95% Wald Effect Estimate Confidence Limits female 0.511 0.154 1.693 age 0.527 0.168 1.658 dayacc 1.515 1.256 1.828 severe 2.290 1.121 4.679 toilet 1.117 0.945 1.321 Association of Predicted Probabilities and Observed Responses Percent Concordant 84.7 Somers' D 0.694 Percent Discordant 15.2 Gamma 0.695 Percent Tied 0.1 Tau-a 0.334 Pairs 4482 c 0.847 Logistic regression diagnostics produced by PROC LOGISTIC for UI data 3 Cookdobs is given by Corollary 2.2 of Preisser and Qaqish under R=I 11:43 Thursday, October 30, 2003 DFBETA_ DFBETA_ DFBETA_ DFBETA_ DFBETA_ Obs Intercept female age dayacc severe 1 -0.02287 0.00725 -0.02509 0.04035 0.02496 2 -0.02229 -0.00756 -0.00194 0.00858 0.02201 3 -0.02442 0.01160 -0.02687 0.05250 0.03570 4 0.00453 -0.01140 -0.03810 0.02879 0.00100 5 -0.04419 -0.02247 0.04425 0.00933 0.00909 6 -0.02903 -0.01518 0.01795 0.02356 0.00455 7 -0.00125 0.00009 0.00051 0.00100 0.00051 8 -0.39710 0.51162 0.37937 -0.62834 0.48629 9 -0.00931 -0.03146 -0.03881 -0.00149 0.00401 10 -0.00148 0.00036 -0.00147 0.00490 -0.00097 11 -0.08745 0.32891 -0.26315 -0.15520 -0.00202 12 0.03091 -0.07981 -0.16290 -0.12629 0.00171 13 0.18462 -0.03519 -0.09717 0.07024 -0.15365 14 0.09855 0.06530 -0.07589 -0.15025 -0.03972 15 -0.07982 0.06278 0.02039 0.02937 0.04559 16 -0.03779 -0.01536 0.04418 0.02631 0.00951 17 -0.06415 0.02972 -0.10509 -0.06811 0.16832 18 -0.01668 -0.01336 0.01534 0.03718 0.01078 19 0.12597 0.10509 0.12118 -0.23631 0.00494 20 -0.02592 -0.01273 0.01456 0.02608 0.00359 21 -0.05465 -0.03215 0.06470 -0.01908 0.01408 22 0.06185 0.09636 0.09215 -0.06733 -0.01261 23 -0.02239 -0.01572 0.01874 0.03010 0.00969 DFBETA_ Obs toilet H cbar cookdobs 1 0.00642 0.03186 0.00347 0.00058 2 0.01284 0.01654 0.00105 0.00018 3 -0.01291 0.03757 0.00547 0.00091 4 -0.00235 0.02307 0.00254 0.00042 5 0.04868 0.02414 0.00774 0.00129 6 0.03127 0.01813 0.00358 0.00060 7 0.00108 0.00308 0.00000 0.00000 8 -0.36007 0.08174 1.01855 0.16976 9 0.04400 0.02230 0.00736 0.00123 10 0.00284 0.00830 0.00004 0.00001 11 0.02850 0.11889 0.24051 0.04009 12 0.12406 0.08653 0.07364 0.01227 13 -0.15735 0.06761 0.07332 0.01222 14 -0.01388 0.01557 0.07367 0.01228 15 0.00112 0.05630 0.00773 0.00129 16 0.02287 0.02310 0.00581 0.00097 17 0.00564 0.05156 0.06472 0.01079 18 -0.01308 0.01755 0.00404 0.00067 19 -0.23340 0.01639 0.17730 0.02955 20 0.02653 0.01762 0.00291 0.00048 21 0.06784 0.02942 0.01437 0.00239 22 -0.14562 0.01697 0.07602 0.01267 23 0.00359 0.01535 0.00377 0.00063 Logistic regression diagnostics produced by PROC LOGISTIC for UI data 4 Cookdobs is given by Corollary 2.2 of Preisser and Qaqish under R=I 11:43 Thursday, October 30, 2003 DFBETA_ DFBETA_ DFBETA_ DFBETA_ DFBETA_ Obs Intercept female age dayacc severe 24 0.00108 -0.01399 -0.01769 0.04026 0.00968 25 0.07053 -0.03461 0.01054 0.03907 -0.14441 26 -0.02348 0.01668 -0.05008 0.06290 0.05571 27 -0.00966 0.00080 -0.00162 0.00796 0.00838 28 -0.03300 -0.00879 0.01139 0.01123 0.03553 29 -0.02829 -0.01133 0.02132 0.02748 0.00382 30 0.13601 -0.38189 0.04069 -0.16159 -0.03896 31 -0.00779 -0.01073 -0.01536 0.02811 0.00172 32 -0.03981 -0.01277 0.05015 0.03011 0.00910 33 -0.03447 -0.01202 0.03297 0.02730 0.00505 34 -0.04419 -0.02247 0.04425 0.00933 0.00909 35 -0.05781 0.10781 -0.09476 0.03750 -0.01062 36 -0.04563 0.03660 -0.07656 0.01555 0.14810 37 0.01032 -0.01184 -0.04165 0.03721 0.00430 38 -0.01734 -0.04476 -0.01879 -0.00268 0.04579 39 -0.00174 -0.02115 -0.04635 0.01731 0.00033 40 -0.03444 -0.01521 0.02922 0.02343 0.00558 41 -0.04129 -0.01516 0.06300 0.02949 0.01450 42 0.27361 -0.05576 0.15368 -0.26985 -0.22053 43 -0.16474 0.01295 0.02822 -0.04142 0.19933 44 0.32347 0.00668 0.14742 0.01141 0.22566 45 -0.00721 0.00638 0.00111 0.03714 -0.00351 46 -0.05849 -0.07228 0.15736 -0.23207 0.04711 DFBETA_ Obs toilet H cbar cookdobs 24 -0.02981 0.02004 0.00426 0.00071 25 0.02962 0.04069 0.03336 0.00556 26 -0.03559 0.04910 0.01149 0.00191 27 0.00526 0.01359 0.00020 0.00003 28 0.00511 0.02123 0.00205 0.00034 29 0.02581 0.01990 0.00312 0.00052 30 0.26927 0.11736 0.21848 0.03641 31 0.00853 0.01429 0.00170 0.00028 32 0.02139 0.02751 0.00621 0.00103 33 0.02965 0.02385 0.00424 0.00071 34 0.04868 0.02414 0.00774 0.00129 35 0.04420 0.08975 0.02673 0.00445 36 -0.07848 0.04985 0.04327 0.00721 37 -0.02333 0.02782 0.00386 0.00064 38 0.03145 0.06854 0.00548 0.00091 39 0.02611 0.02423 0.00465 0.00078 40 0.03391 0.02121 0.00447 0.00075 41 0.00755 0.02838 0.00855 0.00143 42 -0.22475 0.03489 0.26525 0.04421 43 0.07361 0.11498 0.04936 0.00823 44 -1.19649 0.27497 1.12853 0.18809 45 -0.00231 0.03613 0.00146 0.00024 46 0.09242 0.05019 0.09729 0.01622 Logistic regression diagnostics produced by PROC LOGISTIC for UI data 5 Cookdobs is given by Corollary 2.2 of Preisser and Qaqish under R=I 11:43 Thursday, October 30, 2003 DFBETA_ DFBETA_ DFBETA_ DFBETA_ DFBETA_ Obs Intercept female age dayacc severe 47 0.17024 -0.08998 -0.32930 -0.12795 0.02635 48 0.02644 -0.01771 -0.05049 0.04808 0.01567 49 -0.02413 -0.01511 0.00784 0.02363 0.00362 50 -0.07437 0.00015 -0.00748 -0.00416 0.06630 51 0.02056 0.07260 0.01980 0.07189 -0.02576 52 -0.02148 0.06421 0.05955 -0.06806 -0.04630 53 -0.06662 0.02603 0.02274 0.06442 0.07259 54 -0.05923 0.02410 0.06806 0.08550 -0.01515 55 -0.02662 0.04569 -0.00799 -0.12592 -0.06678 56 -0.00126 -0.00700 -0.00364 0.00854 0.00322 57 -0.02911 0.01176 -0.03897 0.05017 0.04240 58 -0.02543 -0.01085 0.02099 0.03131 0.00541 59 -0.00149 0.00032 -0.00045 0.00286 0.00153 60 -0.00432 -0.02246 -0.03451 0.01850 0.00421 61 -0.02784 -0.01264 -0.00924 0.00673 0.03981 62 -0.06690 0.16631 -0.03232 0.09470 -0.13265 63 -0.03041 -0.00691 0.01350 0.01304 0.03038 64 0.35649 -0.42911 -0.24238 -0.17697 -0.02731 65 -0.06484 -0.03211 0.08836 -0.02304 0.01633 66 -0.07138 -0.02120 0.04488 -0.01461 0.09207 67 -0.00001 -0.01874 0.05455 0.06203 -0.09628 68 0.08174 -0.18858 0.03503 0.06025 -0.00954 69 -0.00996 0.01221 -0.00931 0.06551 -0.00937 DFBETA_ Obs toilet H cbar cookdobs 47 -0.10937 0.11844 0.15000 0.02500 48 -0.07791 0.04195 0.01197 0.00199 49 0.02882 0.01614 0.00298 0.00050 50 0.07516 0.08855 0.01155 0.00193 51 -0.06656 0.02043 0.03359 0.00560 52 0.11974 0.01920 0.05191 0.00865 53 -0.01703 0.04371 0.01337 0.00223 54 0.05515 0.08565 0.01701 0.00283 55 0.24806 0.03835 0.10639 0.01773 56 0.00355 0.01259 0.00017 0.00003 57 -0.00234 0.03794 0.00703 0.00117 58 0.01471 0.01834 0.00302 0.00050 59 -0.00045 0.00453 0.00001 0.00000 60 0.01669 0.01662 0.00402 0.00067 61 0.00452 0.02169 0.00246 0.00041 62 0.11154 0.13082 0.06026 0.01004 63 0.00448 0.02071 0.00168 0.00028 64 0.01216 0.10120 0.26399 0.04400 65 0.07295 0.03593 0.01883 0.00314 66 -0.00259 0.04463 0.01207 0.00201 67 0.06977 0.06017 0.02399 0.00400 68 0.01084 0.07971 0.04783 0.00797 69 0.00269 0.04450 0.00466 0.00078 Logistic regression diagnostics produced by PROC LOGISTIC for UI data 6 Cookdobs is given by Corollary 2.2 of Preisser and Qaqish under R=I 11:43 Thursday, October 30, 2003 DFBETA_ DFBETA_ DFBETA_ DFBETA_ DFBETA_ Obs Intercept female age dayacc severe 70 -0.07128 -0.02056 0.04233 -0.01428 0.08194 71 -0.03237 -0.01022 0.01402 0.02249 0.06777 72 0.16383 -0.04819 0.02120 -0.03782 -0.19347 73 -0.00891 -0.02260 -0.02425 0.01767 0.00533 74 0.04988 -0.02649 -0.09273 0.03842 0.01876 75 0.17744 0.09753 0.00541 0.12283 -0.32722 76 -0.01097 -0.02739 0.14719 -0.02492 0.06292 77 -0.02852 -0.00637 0.01181 0.01000 0.02337 78 -0.02953 -0.00665 0.01264 0.01143 0.02666 79 -0.01178 -0.01264 0.01432 0.04258 0.01326 80 0.00015 -0.00950 -0.03113 0.02188 -0.00120 81 -0.04765 -0.02257 0.06485 0.00959 0.01458 82 -0.07287 0.06033 0.15115 -0.23440 -0.06028 83 -0.00777 -0.01081 -0.01863 0.02442 0.00017 84 0.06453 -0.14320 -0.03568 0.04979 -0.01757 85 -0.08504 0.04851 0.07878 -0.10831 -0.06790 86 -0.00846 0.02506 -0.00697 0.10589 -0.01201 87 0.01495 0.10182 0.22111 -0.17665 -0.01053 88 0.26080 0.11722 0.16016 -0.18938 -0.36322 89 -0.13472 0.35728 -0.10496 -0.15029 0.01734 90 0.14448 -0.20220 -0.11695 0.04617 -0.01928 91 -0.03393 -0.02281 0.04790 0.01370 0.01680 92 -0.01138 -0.00544 -0.01022 0.01142 0.01626 DFBETA_ Obs toilet H cbar cookdobs 70 0.01442 0.042532 0.01027 0.00171 71 -0.05489 0.048810 0.00844 0.00141 72 -0.06656 0.037688 0.07384 0.01231 73 0.01850 0.014494 0.00373 0.00062 74 -0.09939 0.054802 0.02112 0.00352 75 0.01884 0.048181 0.14840 0.02473 76 -0.16259 0.059390 0.06613 0.01102 77 0.01409 0.019880 0.00129 0.00022 78 0.00983 0.019909 0.00144 0.00024 79 -0.03066 0.021998 0.00535 0.00089 80 0.01028 0.022298 0.00174 0.00029 81 0.03304 0.025713 0.00983 0.00164 82 0.29628 0.037870 0.18169 0.03028 83 0.01554 0.015882 0.00169 0.00028 84 0.04966 0.074576 0.02760 0.00460 85 0.30888 0.054291 0.13045 0.02174 86 -0.02369 0.059412 0.01275 0.00212 87 -0.06676 0.017913 0.12380 0.02063 88 -0.03130 0.015649 0.26101 0.04350 89 -0.02209 0.082485 0.19355 0.03226 90 0.00010 0.092813 0.05467 0.00911 91 0.00810 0.018996 0.00756 0.00126 92 0.00109 0.013935 0.00064 0.00011 Logistic regression diagnostics produced by PROC LOGISTIC for UI data 7 Cookdobs is given by Corollary 2.2 of Preisser and Qaqish under R=I 11:43 Thursday, October 30, 2003 DFBETA_ DFBETA_ DFBETA_ DFBETA_ DFBETA_ Obs Intercept female age dayacc severe 93 0.04779 0.03610 -0.14926 0.07626 -0.04336 94 -0.01039 -0.01095 -0.01377 0.02498 0.00063 95 -0.07553 0.06732 0.13773 0.04468 -0.03623 96 -0.01612 0.03100 -0.02965 0.09845 -0.02951 97 -0.00174 0.00050 0.00024 0.00396 0.00074 98 -0.00279 0.01008 -0.01327 0.05814 -0.00623 99 0.08962 0.06090 -0.12622 0.03535 -0.03566 100 0.29868 -0.35775 -0.18785 -0.06314 -0.01844 101 -0.08281 0.06014 0.02723 0.02664 0.04305 102 -0.03351 -0.00914 0.01194 0.01292 0.04042 103 -0.02310 -0.01096 0.02199 0.03517 0.00762 104 -0.00892 -0.00758 -0.01753 0.02943 0.04831 105 -0.05902 -0.05965 0.05506 -0.11489 0.01626 106 -0.00428 -0.01217 -0.01410 0.03629 0.00618 107 -0.11107 0.15801 -0.07091 0.01147 -0.01282 108 -0.02192 -0.01128 0.03928 0.04457 0.01579 109 0.03025 -0.04644 -0.09842 -0.02708 0.01230 110 -0.02193 -0.04491 0.01100 -0.04992 0.01842 111 -0.05076 0.11832 -0.07732 0.06071 -0.00231 112 -0.00620 -0.02021 -0.04127 0.01508 -0.00129 113 -0.06456 -0.03024 0.01216 -0.03530 0.10759 114 -0.05090 0.02137 -0.00444 0.06165 0.06347 115 -0.06928 0.02085 -0.02924 0.03218 0.08497 DFBETA_ Obs toilet H cbar cookdobs 93 0.01805 0.04253 0.03904 0.00651 94 0.01693 0.01499 0.00170 0.00028 95 0.09991 0.03577 0.05193 0.00865 96 0.04169 0.03850 0.01773 0.00295 97 0.00016 0.00572 0.00002 0.00000 98 -0.01096 0.04691 0.00359 0.00060 99 -0.07990 0.02425 0.04694 0.00782 100 -0.04904 0.09767 0.16703 0.02784 101 0.01162 0.05750 0.00755 0.00126 102 -0.00270 0.02302 0.00249 0.00041 103 0.00284 0.01849 0.00340 0.00057 104 -0.05982 0.05533 0.00652 0.00109 105 0.14150 0.04911 0.04424 0.00737 106 -0.01234 0.01606 0.00270 0.00045 107 0.10016 0.09598 0.04730 0.00788 108 -0.03081 0.02882 0.00756 0.00126 109 0.00042 0.03371 0.01873 0.00312 110 0.04490 0.02065 0.01396 0.00233 111 -0.01532 0.07875 0.02424 0.00404 112 0.03645 0.02672 0.00460 0.00077 113 -0.00534 0.04582 0.01551 0.00258 114 -0.01735 0.03920 0.01050 0.00175 115 0.02778 0.03745 0.01620 0.00270 Logistic regression diagnostics produced by PROC LOGISTIC for UI data 8 Cookdobs is given by Corollary 2.2 of Preisser and Qaqish under R=I 11:43 Thursday, October 30, 2003 DFBETA_ DFBETA_ DFBETA_ DFBETA_ DFBETA_ Obs Intercept female age dayacc severe 116 -0.01160 -0.01796 -0.05134 0.00315 0.04493 117 -0.08740 0.05742 0.13869 0.10881 -0.02702 118 0.01225 -0.01178 -0.04626 0.03627 0.00356 119 -0.00417 -0.02963 -0.04915 0.00372 0.00249 120 -0.22561 0.06870 0.26609 -0.03830 0.20298 121 -0.00329 -0.00986 -0.02541 0.02314 -0.00063 122 0.13366 -0.02953 0.02065 0.13116 -0.34967 123 -0.01700 0.00928 -0.03743 0.04897 0.03096 124 -0.00562 -0.01275 -0.02628 0.02366 -0.00004 125 0.00314 -0.04232 -0.08639 -0.02252 -0.00072 126 0.02636 0.06760 -0.00409 0.01507 -0.03457 127 0.00739 -0.01534 -0.04840 0.02869 0.00179 128 0.01908 0.02745 -0.05971 0.12056 -0.01547 129 -0.01052 -0.00500 -0.00972 0.00866 0.01219 130 0.04346 -0.02100 -0.01673 0.07086 -0.09498 131 -0.00284 -0.00960 -0.00092 0.01174 0.00518 132 -0.01576 -0.05244 0.02953 0.03219 0.03599 133 0.16145 -0.22594 -0.08608 0.04772 -0.01357 134 0.02297 -0.01829 -0.04619 0.04901 -0.06365 135 0.21957 0.12890 0.07670 0.18350 -0.31382 136 -0.01685 -0.00921 0.05891 0.05827 0.02573 137 0.04682 0.03524 -0.13524 0.10632 -0.03450 DFBETA_ Obs toilet H cbar cookdobs 116 -0.00759 0.04203 0.00560 0.00093 117 0.07180 0.06589 0.04890 0.00815 118 -0.02266 0.03079 0.00402 0.00067 119 0.03886 0.02465 0.00721 0.00120 120 0.03253 0.08491 0.11635 0.01939 121 0.01230 0.01849 0.00163 0.00027 122 0.13671 0.16068 0.13618 0.02270 123 -0.01264 0.04020 0.00516 0.00086 124 0.01699 0.01798 0.00214 0.00036 125 0.07053 0.04826 0.01831 0.00305 126 -0.01327 0.01349 0.02801 0.00467 127 -0.00257 0.02520 0.00374 0.00062 128 -0.05208 0.06937 0.01971 0.00328 129 0.00675 0.01332 0.00049 0.00008 130 0.01489 0.04345 0.01745 0.00291 131 0.00385 0.01584 0.00031 0.00005 132 0.00773 0.05682 0.00687 0.00114 133 -0.03353 0.08597 0.06383 0.01064 134 0.04831 0.05301 0.01060 0.00177 135 -0.18718 0.05985 0.21987 0.03665 136 -0.08040 0.05277 0.01781 0.00297 137 -0.02236 0.05055 0.03450 0.00575 Logistic regression diagnostics produced by PROC LOGISTIC for UI data Cookdobs is given by Corollary 2.2 of Preisser and Qaqish under R=I GEE with deletion diagnostics ( Ver - 1.00 ) ============================================ Data File: UI Outcome variable: BOTHERED Covariates: INTER FEMALE AGE DAYACC SEVERE TOILET Link: 3 (Logit) Variance: 3 (Binomial) Correlation: 1 (Independent) Total number of records read: 137 binary responses: scale parameter is fixed at 1 Logistic regression diagnostics produced by PROC LOGISTIC for UI data Cookdobs is given by Corollary 2.2 of Preisser and Qaqish under R=I Convergence after 5 iteration(s). Working exchangeable correlation is 0 Logistic regression diagnostics produced by PROC LOGISTIC for UI data Cookdobs is given by Corollary 2.2 of Preisser and Qaqish under R=I Estimate, s.e. and z-score: Estimate s.e.-Naive s.e.-Robust z-robust INTER -3.2930 1.1083 0.9663 -3.408 FEMALE -0.6723 0.6116 0.6230 -1.079 AGE -0.6406 0.5847 0.5740 -1.116 DAYACC 0.4154 0.0958 0.0980 4.239 SEVERE 0.8285 0.3646 0.3665 2.261 TOILET 0.1108 0.0855 0.1007 1.100 Scale parameter is 1 Pseudo deviance 127.13438 Pseudo chi-square 136.55053 (c) J. S. Preisser, 2003 Department of Biostatistics University of North Carolina Chapel Hill, N.C. Model-based covariance estimate Obs VARIABLE estimate naive_se INTER FEMALE 1 INTER -3.29305 1.10833 1.22839 -0.29766 2 FEMALE -0.67227 0.61159 -0.29766 0.37405 3 AGE -0.64057 0.58475 -0.23578 0.02960 4 DAYACC 0.41544 0.09580 -0.01774 0.00307 5 SEVERE 0.82854 0.36455 -0.25454 -0.00854 6 TOILET 0.11083 0.08553 -0.03461 -0.00377 Obs AGE DAYACC SEVERE TOILET 1 -0.23578 -0.017742 -0.25454 -0.034607 2 0.02960 0.003071 -0.00854 -0.003772 3 0.34193 -0.008241 0.01604 0.004056 4 -0.00824 0.009177 -0.00192 -0.000462 5 0.01604 -0.001923 0.13290 -0.003912 6 0.00406 -0.000462 -0.00391 0.007316 Empirical Sandwich covariance estimate robust_ Obs VARIABLE estimate se INTER FEMALE 1 INTER -3.29305 0.96627 0.93368 -0.22014 2 FEMALE -0.67227 0.62298 -0.22014 0.38811 3 AGE -0.64057 0.57401 -0.12637 0.04446 4 DAYACC 0.41544 0.09800 0.01030 -0.01694 5 SEVERE 0.82854 0.36652 -0.26435 0.02603 6 TOILET 0.11083 0.10072 -0.01818 -0.01824 Obs AGE DAYACC SEVERE TOILET 1 -0.12637 0.010299 -0.26435 -0.018183 2 0.04446 -0.016945 0.02603 -0.018237 3 0.32949 -0.011248 0.02605 -0.011252 4 -0.01125 0.009604 -0.00867 0.001287 5 0.02605 -0.008674 0.13434 -0.010424 6 -0.01125 0.001287 -0.01042 0.010144 Logistic regression diagnostics for UI data from SAS IML macro Cookdobs is given by Corollary 2.2 of Preisser and Qaqish under R=I Cookdobs under R=I is equivalent to Williams (1987) DFBETA from SAS GEE macro is standardized to match PROC LOGISTIC output nse_ nse_ nse_ nse_ nse_ Obs inter female nse_age dayacc severe toilet ID 1 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 8 2 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 8 3 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 8 4 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 24 5 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 24 6 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 27 7 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 27 8 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 27 9 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 27 Obs I IJ NI FIT RES SRES QWOBS COOKDOBS INTER 1 1 1 3 0.90473 0.09527 0.32450 0.03186 0.00060 -0.02534 2 1 2 3 0.05879 -0.05879 -0.24992 0.01653 0.00018 -0.02470 3 1 3 3 0.87709 0.12291 0.37434 0.03757 0.00095 -0.02706 4 2 4 2 0.09717 -0.09717 -0.32806 0.02307 0.00043 0.00502 5 2 5 2 0.23827 -0.23827 -0.55929 0.02414 0.00132 -0.04898 6 3 6 5 0.16224 -0.16224 -0.44006 0.01813 0.00061 -0.03218 7 3 7 5 0.99899 0.00101 0.03174 0.00308 0.00000 -0.00139 8 3 8 5 0.91964 -0.91964 -3.38289 0.08173 0.18488 -0.44011 9 3 9 5 0.24396 -0.24396 -0.56805 0.02230 0.00125 -0.01032 DFBETAS_ Obs FEMALE AGE DAYACC SEVERE TOILET INTER 1 0.00443 -0.01467 0.003865 0.00910 0.00055 -0.02286 2 -0.00462 -0.00113 0.000822 0.00802 0.00110 -0.02228 3 0.00710 -0.01571 0.005029 0.01301 -0.00110 -0.02441 4 -0.00697 -0.02228 0.002758 0.00036 -0.00020 0.00452 5 -0.01374 0.02587 0.000894 0.00331 0.00416 -0.04419 6 -0.00928 0.01050 0.002257 0.00166 0.00267 -0.02903 7 0.00005 0.00030 0.000095 0.00019 0.00009 -0.00125 8 0.31291 0.22184 -0.060196 0.17728 -0.03080 -0.39709 9 -0.01924 -0.02270 -0.000143 0.00146 0.00376 -0.00931 DFBETAS_ DFBETAS_ DFBETAS_ DFBETAS_ DFBETAS_ Obs FEMALE AGE DAYACC SEVERE TOILET 1 0.00725 -0.02509 0.04035 0.02496 0.00642 2 -0.00756 -0.00194 0.00858 0.02201 0.01284 3 0.01160 -0.02687 0.05249 0.03569 -0.01291 4 -0.01140 -0.03810 0.02879 0.00100 -0.00235 5 -0.02247 0.04425 0.00933 0.00909 0.04868 6 -0.01518 0.01795 0.02356 0.00455 0.03126 7 0.00009 0.00051 0.00100 0.00051 0.00107 8 0.51163 0.37938 -0.62836 0.48629 -0.36009 9 -0.03147 -0.03881 -0.00149 0.00401 0.04400 Logistic regression diagnostics for UI data from SAS IML macro Cookdobs is given by Corollary 2.2 of Preisser and Qaqish under R=I Cookdobs under R=I is equivalent to Williams (1987) DFBETA from SAS GEE macro is standardized to match PROC LOGISTIC output nse_ nse_ nse_ nse_ nse_ Obs inter female nse_age dayacc severe toilet ID 10 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 27 11 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 41 12 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 41 13 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 41 14 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 41 15 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 45 16 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 45 17 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 45 18 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 45 Obs I IJ NI FIT RES SRES QWOBS COOKDOBS INTER 10 3 10 5 0.99577 0.00423 0.06515 0.00830 0.00001 -0.00164 11 4 11 4 0.64062 -0.64062 -1.33512 0.11889 0.04550 -0.09692 12 4 12 4 0.43737 -0.43737 -0.88169 0.08653 0.01344 0.03426 13 4 13 4 0.50276 -0.50276 -1.00554 0.06762 0.01311 0.20463 14 4 14 4 0.17671 0.82329 2.15850 0.01557 0.01247 0.10923 15 5 15 4 0.11475 -0.11475 -0.36003 0.05630 0.00137 -0.08846 16 5 16 4 0.19738 -0.19738 -0.49591 0.02310 0.00099 -0.04188 17 5 17 4 0.45653 0.54347 1.09107 0.05156 0.01137 -0.07110 18 5 18 4 0.18431 -0.18431 -0.47535 0.01755 0.00068 -0.01849 DFBETAS_ Obs FEMALE AGE DAYACC SEVERE TOILET INTER 10 0.00022 -0.00086 0.000469 -0.000353 0.00024 -0.00148 11 0.20117 -0.15388 -0.014868 -0.000736 0.00244 -0.08745 12 -0.04881 -0.09526 -0.012099 0.000625 0.01061 0.03091 13 -0.02153 -0.05682 0.006729 -0.056014 -0.01346 0.18463 14 0.03994 -0.04438 -0.014394 -0.014481 -0.00119 0.09855 15 0.03839 0.01192 0.002813 0.016618 0.00010 -0.07982 16 -0.00939 0.02583 0.002521 0.003467 0.00196 -0.03779 17 0.01818 -0.06145 -0.006525 0.061364 0.00048 -0.06415 18 -0.00817 0.00897 0.003562 0.003931 -0.00112 -0.01668 DFBETAS_ DFBETAS_ DFBETAS_ DFBETAS_ DFBETAS_ Obs FEMALE AGE DAYACC SEVERE TOILET 10 0.00036 -0.00147 0.00490 -0.00097 0.00284 11 0.32892 -0.26315 -0.15520 -0.00202 0.02850 12 -0.07981 -0.16290 -0.12630 0.00171 0.12406 13 -0.03520 -0.09717 0.07024 -0.15365 -0.15735 14 0.06530 -0.07589 -0.15025 -0.03972 -0.01388 15 0.06277 0.02039 0.02937 0.04558 0.00112 16 -0.01536 0.04418 0.02631 0.00951 0.02287 17 0.02972 -0.10509 -0.06811 0.16833 0.00564 18 -0.01336 0.01534 0.03718 0.01078 -0.01308 Logistic regression diagnostics for UI data from SAS IML macro Cookdobs is given by Corollary 2.2 of Preisser and Qaqish under R=I Cookdobs under R=I is equivalent to Williams (1987) DFBETA from SAS GEE macro is standardized to match PROC LOGISTIC output nse_ nse_ nse_ nse_ nse_ Obs inter female nse_age dayacc severe toilet ID 19 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 55 20 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 55 21 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 56 22 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 56 23 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 56 24 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 60 25 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 60 26 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 60 27 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 60 Obs I IJ NI FIT RES SRES QWOBS COOKDOBS INTER 19 6 19 2 0.08590 0.91410 3.26212 0.01639 0.03004 0.13963 20 6 20 2 0.13945 -0.13945 -0.40255 0.01762 0.00049 -0.02872 21 7 21 3 0.32153 -0.32153 -0.68841 0.02942 0.00247 -0.06057 22 7 22 3 0.18506 0.81494 2.09851 0.01697 0.01289 0.06855 23 7 23 3 0.19466 -0.19466 -0.49164 0.01535 0.00064 -0.02482 24 8 24 7 0.17239 -0.17239 -0.45641 0.02004 0.00072 0.00120 25 8 25 7 0.44026 -0.44026 -0.88688 0.04069 0.00580 0.07817 26 8 26 7 0.81802 0.18198 0.47166 0.04910 0.00201 -0.02602 27 8 27 7 0.98559 0.01441 0.12090 0.01359 0.00003 -0.01071 DFBETAS_ Obs FEMALE AGE DAYACC SEVERE TOILET INTER 19 0.06427 0.07086 -0.022639 0.001799 -0.01996 0.12598 20 -0.00779 0.00852 0.002498 0.001308 0.00227 -0.02592 21 -0.01966 0.03783 -0.001828 0.005132 0.00580 -0.05465 22 0.05893 0.05389 -0.006450 -0.004598 -0.01246 0.06185 23 -0.00961 0.01096 0.002884 0.003532 0.00031 -0.02239 24 -0.00856 -0.01035 0.003856 0.003528 -0.00255 0.00108 25 -0.02117 0.00616 0.003743 -0.052648 0.00253 0.07053 26 0.01020 -0.02929 0.006026 0.020310 -0.00304 -0.02348 27 0.00049 -0.00095 0.000763 0.003054 0.00045 -0.00966 DFBETAS_ DFBETAS_ DFBETAS_ DFBETAS_ DFBETAS_ Obs FEMALE AGE DAYACC SEVERE TOILET 19 0.10509 0.12118 -0.23632 0.00493 -0.23340 20 -0.01273 0.01456 0.02608 0.00359 0.02653 21 -0.03215 0.06470 -0.01908 0.01408 0.06784 22 0.09636 0.09215 -0.06733 -0.01261 -0.14562 23 -0.01572 0.01874 0.03010 0.00969 0.00359 24 -0.01399 -0.01769 0.04025 0.00968 -0.02981 25 -0.03461 0.01054 0.03907 -0.14442 0.02962 26 0.01668 -0.05008 0.06290 0.05571 -0.03558 27 0.00080 -0.00162 0.00796 0.00838 0.00526 Logistic regression diagnostics for UI data from SAS IML macro Cookdobs is given by Corollary 2.2 of Preisser and Qaqish under R=I Cookdobs under R=I is equivalent to Williams (1987) DFBETA from SAS GEE macro is standardized to match PROC LOGISTIC output nse_ nse_ nse_ nse_ nse_ Obs inter female nse_age dayacc severe toilet ID 28 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 60 29 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 60 30 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 60 31 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 65 32 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 65 33 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 65 34 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 89 35 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 89 36 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 89 Obs I IJ NI FIT RES SRES QWOBS COOKDOBS INTER 28 8 28 7 0.08632 -0.08632 -0.30737 0.02123 0.00035 -0.03657 29 8 29 7 0.13305 -0.13305 -0.39176 0.01990 0.00053 -0.03136 30 8 30 7 0.37834 0.62166 1.28184 0.11736 0.04126 0.15074 31 9 31 3 0.10498 -0.10498 -0.34248 0.01429 0.00029 -0.00863 32 9 32 3 0.17992 -0.17992 -0.46839 0.02751 0.00106 -0.04412 33 9 33 3 0.14795 -0.14795 -0.41670 0.02385 0.00072 -0.03820 34 10 34 4 0.23827 -0.23827 -0.55929 0.02414 0.00132 -0.04898 35 10 35 4 0.21326 -0.21326 -0.52065 0.08975 0.00489 -0.06408 36 10 36 4 0.54801 0.45199 0.90818 0.04985 0.00759 -0.05057 DFBETAS_ Obs FEMALE AGE DAYACC SEVERE TOILET INTER 28 -0.00537 0.00666 0.001076 0.01295 0.00044 -0.03300 29 -0.00693 0.01247 0.002632 0.00139 0.00221 -0.02829 30 -0.23357 0.02379 -0.015480 -0.01420 0.02303 0.13601 31 -0.00656 -0.00898 0.002693 0.00063 0.00073 -0.00779 32 -0.00781 0.02933 0.002884 0.00332 0.00183 -0.03981 33 -0.00735 0.01928 0.002615 0.00184 0.00254 -0.03447 34 -0.01374 0.02587 0.000894 0.00331 0.00416 -0.04419 35 0.06594 -0.05541 0.003593 -0.00387 0.00378 -0.05781 36 0.02238 -0.04477 0.001490 0.05399 -0.00671 -0.04563 DFBETAS_ DFBETAS_ DFBETAS_ DFBETAS_ DFBETAS_ Obs FEMALE AGE DAYACC SEVERE TOILET 28 -0.00879 0.01139 0.01123 0.03553 0.00511 29 -0.01133 0.02132 0.02748 0.00383 0.02581 30 -0.38190 0.04069 -0.16160 -0.03896 0.26927 31 -0.01073 -0.01536 0.02811 0.00172 0.00853 32 -0.01277 0.05015 0.03011 0.00910 0.02138 33 -0.01202 0.03297 0.02730 0.00505 0.02965 34 -0.02247 0.04425 0.00933 0.00909 0.04868 35 0.10781 -0.09476 0.03750 -0.01062 0.04420 36 0.03660 -0.07656 0.01555 0.14810 -0.07848 Logistic regression diagnostics for UI data from SAS IML macro Cookdobs is given by Corollary 2.2 of Preisser and Qaqish under R=I Cookdobs under R=I is equivalent to Williams (1987) DFBETA from SAS GEE macro is standardized to match PROC LOGISTIC output nse_ nse_ nse_ nse_ nse_ Obs inter female nse_age dayacc severe toilet ID 37 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 89 38 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 102 39 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 102 40 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 102 41 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 102 42 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 102 43 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 107 44 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 107 45 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 107 Obs I IJ NI FIT RES SRES QWOBS COOKDOBS INTER 37 10 37 4 0.11891 -0.11891 -0.36737 0.02782 0.00066 0.01144 38 11 38 5 0.93069 0.06931 0.27289 0.06854 0.00098 -0.01922 39 11 39 5 0.15781 -0.15781 -0.43287 0.02423 0.00079 -0.00193 40 11 40 5 0.17113 -0.17113 -0.45439 0.02121 0.00076 -0.03817 41 11 41 5 0.22656 -0.22656 -0.54122 0.02838 0.00147 -0.04576 42 11 42 5 0.88007 -0.88007 -2.70889 0.03489 0.04581 0.30327 43 12 43 3 0.72466 0.27534 0.61641 0.11498 0.00930 -0.18258 44 12 44 3 0.74849 -0.74849 -1.72510 0.27496 0.25944 0.35853 45 12 45 3 0.96249 0.03751 0.19741 0.03612 0.00025 -0.00799 DFBETAS_ Obs FEMALE AGE DAYACC SEVERE TOILET INTER 37 -0.00724 -0.02435 0.003565 0.00157 -0.00200 0.01032 38 -0.02737 -0.01099 -0.000257 0.01669 0.00269 -0.01734 39 -0.01293 -0.02710 0.001658 0.00012 0.00223 -0.00175 40 -0.00930 0.01709 0.002245 0.00204 0.00290 -0.03444 41 -0.00927 0.03684 0.002825 0.00529 0.00065 -0.04129 42 -0.03410 0.08987 -0.025853 -0.08040 -0.01922 0.27362 43 0.00792 0.01650 -0.003967 0.07267 0.00630 -0.16474 44 0.00409 0.08621 0.001092 0.08227 -0.10234 0.32348 45 0.00390 0.00065 0.003558 -0.00128 -0.00020 -0.00721 DFBETAS_ DFBETAS_ DFBETAS_ DFBETAS_ DFBETAS_ Obs FEMALE AGE DAYACC SEVERE TOILET 37 -0.01184 -0.04165 0.03721 0.00430 -0.02333 38 -0.04476 -0.01879 -0.00268 0.04579 0.03145 39 -0.02115 -0.04635 0.01731 0.00033 0.02611 40 -0.01521 0.02922 0.02343 0.00558 0.03391 41 -0.01516 0.06300 0.02949 0.01450 0.00755 42 -0.05576 0.15369 -0.26987 -0.22054 -0.22476 43 0.01295 0.02822 -0.04142 0.19933 0.07361 44 0.00668 0.14742 0.01140 0.22566 -1.19653 45 0.00638 0.00111 0.03714 -0.00351 -0.00231 Logistic regression diagnostics for UI data from SAS IML macro Cookdobs is given by Corollary 2.2 of Preisser and Qaqish under R=I Cookdobs under R=I is equivalent to Williams (1987) DFBETA from SAS GEE macro is standardized to match PROC LOGISTIC output nse_ nse_ nse_ nse_ nse_ Obs inter female nse_age dayacc severe toilet ID 46 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 108 47 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 108 48 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 108 49 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 108 50 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 111 51 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 111 52 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 113 53 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 113 54 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 118 Obs I IJ NI FIT RES SRES QWOBS COOKDOBS INTER 46 13 46 4 0.64803 -0.64803 -1.35690 0.05020 0.017074 -0.06482 47 13 47 4 0.52751 -0.52751 -1.05662 0.11845 0.028361 0.18869 48 13 48 4 0.21467 -0.21467 -0.52283 0.04195 0.002082 0.02931 49 13 49 4 0.15372 -0.15372 -0.42619 0.01614 0.000505 -0.02674 50 14 50 2 0.89373 0.10627 0.34482 0.08854 0.002112 -0.08242 51 14 51 2 0.38303 0.61697 1.26915 0.02043 0.005715 0.02279 52 15 52 2 0.27387 0.72613 1.62831 0.01920 0.008822 -0.02381 53 15 53 2 0.77373 0.22627 0.54078 0.04371 0.002330 -0.07383 54 16 54 4 0.84635 0.15365 0.42609 0.08565 0.003100 -0.06564 DFBETAS_ Obs FEMALE AGE DAYACC SEVERE TOILET INTER 46 -0.04421 0.09202 -0.022233 0.01717 0.007905 -0.05849 47 -0.05503 -0.19256 -0.012258 0.00961 -0.009356 0.17024 48 -0.01083 -0.02953 0.004606 0.00571 -0.006664 0.02644 49 -0.00924 0.00458 0.002263 0.00132 0.002465 -0.02413 50 0.00009 -0.00437 -0.000398 0.02417 0.006429 -0.07437 51 0.04440 0.01158 0.006887 -0.00939 -0.005693 0.02056 52 0.03927 0.03482 -0.006520 -0.01688 0.010242 -0.02148 53 0.01592 0.01330 0.006171 0.02646 -0.001456 -0.06661 54 0.01474 0.03979 0.008190 -0.00552 0.004717 -0.05922 DFBETAS_ DFBETAS_ DFBETAS_ DFBETAS_ DFBETAS_ Obs FEMALE AGE DAYACC SEVERE TOILET 46 -0.07229 0.15737 -0.23208 0.04711 0.09242 47 -0.08998 -0.32931 -0.12796 0.02635 -0.10938 48 -0.01771 -0.05049 0.04808 0.01567 -0.07791 49 -0.01511 0.00784 0.02363 0.00362 0.02882 50 0.00015 -0.00748 -0.00416 0.06629 0.07516 51 0.07260 0.01980 0.07189 -0.02576 -0.06656 52 0.06421 0.05955 -0.06806 -0.04630 0.11974 53 0.02603 0.02274 0.06442 0.07259 -0.01702 54 0.02410 0.06805 0.08550 -0.01514 0.05515 Logistic regression diagnostics for UI data from SAS IML macro Cookdobs is given by Corollary 2.2 of Preisser and Qaqish under R=I Cookdobs under R=I is equivalent to Williams (1987) DFBETA from SAS GEE macro is standardized to match PROC LOGISTIC output nse_ nse_ nse_ nse_ nse_ Obs inter female nse_age dayacc severe toilet ID 55 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 118 56 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 118 57 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 118 58 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 124 59 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 124 60 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 124 61 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 124 62 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 125 63 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 125 Obs I IJ NI FIT RES SRES QWOBS COOKDOBS INTER 55 16 55 4 0.27262 0.72738 1.63342 0.03835 0.018439 -0.02951 56 16 56 4 0.98697 0.01303 0.11492 0.01259 0.000028 -0.00140 57 16 57 4 0.84873 0.15127 0.42218 0.03793 0.001218 -0.03226 58 17 58 4 0.13910 -0.13910 -0.40196 0.01834 0.000513 -0.02819 59 17 59 4 0.99750 0.00250 0.05006 0.00453 0.000002 -0.00165 60 17 60 4 0.19221 -0.19221 -0.48780 0.01662 0.000682 -0.00479 61 17 61 4 0.09974 -0.09974 -0.33284 0.02169 0.000419 -0.03085 62 18 62 8 0.28590 -0.28590 -0.63275 0.13082 0.011555 -0.07415 63 18 63 8 0.07361 -0.07361 -0.28188 0.02071 0.000286 -0.03370 DFBETAS_ Obs FEMALE AGE DAYACC SEVERE TOILET INTER 55 0.02794 -0.00467 -0.012063 -0.02434 0.021218 -0.02662 56 -0.00428 -0.00213 0.000818 0.00117 0.000304 -0.00126 57 0.00719 -0.02279 0.004806 0.01546 -0.000200 -0.02911 58 -0.00663 0.01227 0.002999 0.00197 0.001258 -0.02543 59 0.00019 -0.00026 0.000274 0.00056 -0.000039 -0.00149 60 -0.01374 -0.02018 0.001772 0.00154 0.001427 -0.00432 61 -0.00773 -0.00540 0.000644 0.01451 0.000387 -0.02784 62 0.10171 -0.01890 0.009072 -0.04836 0.009541 -0.06690 63 -0.00423 0.00789 0.001249 0.01108 0.000383 -0.03041 DFBETAS_ DFBETAS_ DFBETAS_ DFBETAS_ DFBETAS_ Obs FEMALE AGE DAYACC SEVERE TOILET 55 0.04569 -0.00799 -0.12592 -0.06678 0.24806 56 -0.00700 -0.00364 0.00854 0.00322 0.00355 57 0.01176 -0.03897 0.05016 0.04240 -0.00234 58 -0.01085 0.02099 0.03131 0.00541 0.01471 59 0.00032 -0.00045 0.00286 0.00153 -0.00045 60 -0.02246 -0.03451 0.01850 0.00421 0.01669 61 -0.01264 -0.00924 0.00673 0.03981 0.00452 62 0.16631 -0.03232 0.09470 -0.13265 0.11154 63 -0.00691 0.01350 0.01304 0.03038 0.00448 Logistic regression diagnostics for UI data from SAS IML macro Cookdobs is given by Corollary 2.2 of Preisser and Qaqish under R=I Cookdobs under R=I is equivalent to Williams (1987) DFBETA from SAS GEE macro is standardized to match PROC LOGISTIC output nse_ nse_ nse_ nse_ nse_ Obs inter female nse_age dayacc severe toilet ID 64 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 125 65 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 125 66 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 125 67 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 125 68 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 125 69 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 125 70 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 127 71 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 127 72 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 127 Obs I IJ NI FIT RES SRES QWOBS COOKDOBS INTER 64 18 64 8 0.29897 0.70103 1.53127 0.10120 0.048954 0.39512 65 18 65 8 0.33566 -0.33566 -0.71082 0.03593 0.003255 -0.07187 66 18 66 8 0.20528 -0.20528 -0.50824 0.04464 0.002105 -0.07911 67 18 67 8 0.27257 -0.27257 -0.61212 0.06017 0.004254 -0.00001 68 18 68 8 0.64422 0.35578 0.74315 0.07971 0.008663 0.09059 69 18 69 8 0.90901 0.09099 0.31638 0.04450 0.000813 -0.01104 70 19 70 4 0.18779 -0.18779 -0.48084 0.04253 0.001788 -0.07900 71 19 71 4 0.14121 -0.14121 -0.40551 0.04881 0.001478 -0.03588 72 19 72 4 0.65344 -0.65344 -1.37315 0.03769 0.012790 0.18159 DFBETAS_ Obs FEMALE AGE DAYACC SEVERE TOILET INTER 64 -0.26244 -0.14173 -0.016954 -0.00996 0.001040 0.35650 65 -0.01964 0.05167 -0.002208 0.00595 0.006239 -0.06484 66 -0.01297 0.02624 -0.001400 0.03357 -0.000221 -0.07138 67 -0.01146 0.03190 0.005942 -0.03510 0.005968 -0.00001 68 -0.11533 0.02048 0.005772 -0.00348 0.000927 0.08174 69 0.00747 -0.00544 0.006275 -0.00342 0.000230 -0.00996 70 -0.01258 0.02475 -0.001368 0.02987 0.001233 -0.07128 71 -0.00625 0.00820 0.002155 0.02471 -0.004695 -0.03237 72 -0.02947 0.01240 -0.003624 -0.07053 -0.005693 0.16384 DFBETAS_ DFBETAS_ DFBETAS_ DFBETAS_ DFBETAS_ Obs FEMALE AGE DAYACC SEVERE TOILET 64 -0.42912 -0.24238 -0.17698 -0.02731 0.01216 65 -0.03211 0.08836 -0.02305 0.01633 0.07295 66 -0.02120 0.04488 -0.01461 0.09207 -0.00259 67 -0.01874 0.05455 0.06203 -0.09628 0.06977 68 -0.18858 0.03503 0.06025 -0.00954 0.01084 69 0.01221 -0.00931 0.06550 -0.00937 0.00269 70 -0.02056 0.04233 -0.01428 0.08194 0.01442 71 -0.01022 0.01402 0.02249 0.06777 -0.05489 72 -0.04819 0.02120 -0.03782 -0.19348 -0.06656 Logistic regression diagnostics for UI data from SAS IML macro Cookdobs is given by Corollary 2.2 of Preisser and Qaqish under R=I Cookdobs under R=I is equivalent to Williams (1987) DFBETA from SAS GEE macro is standardized to match PROC LOGISTIC output nse_ nse_ nse_ nse_ nse_ Obs inter female nse_age dayacc severe toilet ID 73 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 127 74 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 130 75 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 132 76 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 132 77 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 132 78 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 137 79 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 137 80 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 137 81 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 137 Obs I IJ NI FIT RES SRES QWOBS COOKDOBS INTER 73 19 73 4 0.20236 -0.20236 -0.50368 0.01449 0.000631 -0.00987 74 20 74 1 0.26705 -0.26705 -0.60361 0.05480 0.003725 0.05529 75 21 75 3 0.25434 0.74566 1.71223 0.04818 0.025987 0.19667 76 21 76 3 0.51158 -0.51158 -1.02344 0.05939 0.011719 -0.01216 77 21 77 3 0.05985 -0.05985 -0.25231 0.01988 0.000220 -0.03161 78 22 78 4 0.06640 -0.06640 -0.26669 0.01991 0.000246 -0.03272 79 22 79 4 0.19223 -0.19223 -0.48783 0.02200 0.000912 -0.01306 80 22 80 4 0.07082 -0.07082 -0.27607 0.02230 0.000296 0.00016 81 22 81 4 0.27145 -0.27145 -0.61040 0.02571 0.001682 -0.05281 DFBETAS_ Obs FEMALE AGE DAYACC SEVERE TOILET INTER 73 -0.01382 -0.01418 0.001693 0.00194 0.001583 -0.00891 74 -0.01620 -0.05422 0.003681 0.00684 -0.008501 0.04988 75 0.05965 0.00316 0.011767 -0.11929 0.001612 0.17745 76 -0.01675 0.08607 -0.002388 0.02294 -0.013908 -0.01097 77 -0.00389 0.00691 0.000958 0.00852 0.001205 -0.02852 78 -0.00407 0.00739 0.001094 0.00972 0.000841 -0.02952 79 -0.00773 0.00837 0.004079 0.00483 -0.002623 -0.01178 80 -0.00581 -0.01820 0.002096 -0.00044 0.000879 0.00015 81 -0.01381 0.03792 0.000919 0.00532 0.002826 -0.04765 DFBETAS_ DFBETAS_ DFBETAS_ DFBETAS_ DFBETAS_ Obs FEMALE AGE DAYACC SEVERE TOILET 73 -0.02260 -0.02425 0.01767 0.00533 0.01850 74 -0.02649 -0.09273 0.03842 0.01876 -0.09939 75 0.09753 0.00541 0.12283 -0.32722 0.01884 76 -0.02739 0.14720 -0.02493 0.06292 -0.16260 77 -0.00637 0.01181 0.01000 0.02337 0.01409 78 -0.00665 0.01264 0.01142 0.02666 0.00983 79 -0.01264 0.01432 0.04258 0.01326 -0.03066 80 -0.00950 -0.03112 0.02188 -0.00120 0.01028 81 -0.02258 0.06485 0.00959 0.01458 0.03304 Logistic regression diagnostics for UI data from SAS IML macro Cookdobs is given by Corollary 2.2 of Preisser and Qaqish under R=I Cookdobs under R=I is equivalent to Williams (1987) DFBETA from SAS GEE macro is standardized to match PROC LOGISTIC output nse_ nse_ nse_ nse_ nse_ Obs inter female nse_age dayacc severe toilet ID 82 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 146 83 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 146 84 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 146 85 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 153 86 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 156 87 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 156 88 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 156 89 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 182 90 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 182 Obs I IJ NI FIT RES SRES QWOBS COOKDOBS INTER 82 23 82 3 0.17806 0.82194 2.14851 0.03787 0.031475 -0.08076 83 23 83 3 0.09461 -0.09461 -0.32326 0.01588 0.000286 -0.00861 84 23 84 3 0.74490 0.25510 0.58520 0.07458 0.004970 0.07151 85 24 85 1 0.30559 0.69441 1.50744 0.05429 0.022991 -0.09426 86 25 86 3 0.83209 0.16791 0.44922 0.05941 0.002259 -0.00938 87 25 87 3 0.12840 0.87160 2.60541 0.01791 0.021011 0.01657 88 25 88 3 0.05741 0.94259 4.05211 0.01565 0.044196 0.28906 89 26 89 6 0.68285 -0.68285 -1.46732 0.08249 0.035161 -0.14931 90 26 90 6 0.65174 0.34826 0.73100 0.09281 0.010044 0.16013 DFBETAS_ Obs FEMALE AGE DAYACC SEVERE TOILET INTER 82 0.03690 0.08839 -0.022455 -0.02198 0.025342 -0.07287 83 -0.00661 -0.01089 0.002340 0.00006 0.001329 -0.00777 84 -0.08758 -0.02086 0.004769 -0.00641 0.004247 0.06452 85 0.02967 0.04606 -0.010376 -0.02475 0.026420 -0.08504 86 0.01533 -0.00407 0.010144 -0.00438 -0.002026 -0.00846 87 0.06227 0.12929 -0.016923 -0.00384 -0.005711 0.01495 88 0.07169 0.09366 -0.018143 -0.13242 -0.002677 0.26081 89 0.21852 -0.06137 -0.014399 0.00632 -0.001890 -0.13472 90 -0.12366 -0.06838 0.004423 -0.00703 0.000008 0.14448 DFBETAS_ DFBETAS_ DFBETAS_ DFBETAS_ DFBETAS_ Obs FEMALE AGE DAYACC SEVERE TOILET 82 0.06033 0.15115 -0.23440 -0.06028 0.29628 83 -0.01081 -0.01863 0.02442 0.00017 0.01554 84 -0.14320 -0.03568 0.04979 -0.01757 0.04966 85 0.04851 0.07878 -0.10831 -0.06790 0.30888 86 0.02506 -0.00697 0.10589 -0.01201 -0.02369 87 0.10182 0.22111 -0.17666 -0.01054 -0.06677 88 0.11722 0.16017 -0.18939 -0.36323 -0.03130 89 0.35729 -0.10496 -0.15030 0.01733 -0.02210 90 -0.20219 -0.11695 0.04617 -0.01927 0.00010 Logistic regression diagnostics for UI data from SAS IML macro Cookdobs is given by Corollary 2.2 of Preisser and Qaqish under R=I Cookdobs under R=I is equivalent to Williams (1987) DFBETA from SAS GEE macro is standardized to match PROC LOGISTIC output nse_ nse_ nse_ nse_ nse_ Obs inter female nse_age dayacc severe toilet ID 91 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 182 92 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 182 93 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 182 94 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 182 95 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 185 96 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 185 97 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 185 98 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 185 99 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 185 Obs I IJ NI FIT RES SRES QWOBS COOKDOBS INTER 91 26 91 6 0.28080 -0.28080 -0.62484 0.01900 0.001284 -0.03760 92 26 92 6 0.04312 -0.04312 -0.21227 0.01393 0.000108 -0.01261 93 26 93 6 0.53228 0.46772 0.93740 0.04253 0.006795 0.05297 94 26 94 6 0.10024 -0.10024 -0.33378 0.01499 0.000287 -0.01151 95 27 95 5 0.41669 0.58331 1.18317 0.03577 0.008975 -0.08371 96 27 96 5 0.69313 0.30687 0.66539 0.03850 0.003073 -0.01786 97 27 97 5 0.99690 0.00310 0.05574 0.00572 0.000003 -0.00193 98 27 98 5 0.93207 0.06793 0.26997 0.04690 0.000627 -0.00309 99 27 99 5 0.34617 0.65383 1.37431 0.02425 0.008019 0.09933 DFBETAS_ Obs FEMALE AGE DAYACC SEVERE TOILET INTER 91 -0.01395 0.02801 0.001312 0.00612 0.000693 -0.03393 92 -0.00333 -0.00598 0.001094 0.00593 0.000093 -0.01138 93 0.02208 -0.08728 0.007305 -0.01581 0.001544 0.04779 94 -0.00670 -0.00805 0.002393 0.00023 0.001448 -0.01039 95 0.04117 0.08054 0.004280 -0.01321 0.008546 -0.07553 96 0.01896 -0.01734 0.009431 -0.01076 0.003566 -0.01612 97 0.00031 0.00014 0.000380 0.00027 0.000014 -0.00174 98 0.00617 -0.00776 0.005569 -0.00227 -0.000938 -0.00279 99 0.03725 -0.07381 0.003387 -0.01300 -0.006834 0.08962 DFBETAS_ DFBETAS_ DFBETAS_ DFBETAS_ DFBETAS_ Obs FEMALE AGE DAYACC SEVERE TOILET 91 -0.02281 0.04790 0.01370 0.01680 0.00810 92 -0.00544 -0.01022 0.01142 0.01626 0.00109 93 0.03610 -0.14926 0.07626 -0.04336 0.01805 94 -0.01095 -0.01377 0.02498 0.00063 0.01693 95 0.06732 0.13773 0.04468 -0.03623 0.09991 96 0.03100 -0.02965 0.09845 -0.02950 0.04169 97 0.00050 0.00024 0.00396 0.00074 0.00016 98 0.01008 -0.01327 0.05814 -0.00623 -0.01096 99 0.06090 -0.12622 0.03535 -0.03566 -0.07990 Logistic regression diagnostics for UI data from SAS IML macro Cookdobs is given by Corollary 2.2 of Preisser and Qaqish under R=I Cookdobs under R=I is equivalent to Williams (1987) DFBETA from SAS GEE macro is standardized to match PROC LOGISTIC output nse_ nse_ nse_ nse_ nse_ Obs inter female nse_age dayacc severe toilet ID 100 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 195 101 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 195 102 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 195 103 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 201 104 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 201 105 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 206 106 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 206 107 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 206 108 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 206 Obs I IJ NI FIT RES SRES QWOBS COOKDOBS INTER 100 28 100 3 0.39321 0.60679 1.24224 0.09767 0.030852 0.33104 101 28 101 3 0.11008 -0.11008 -0.35171 0.05750 0.001334 -0.09177 102 28 102 3 0.09547 -0.09547 -0.32489 0.02302 0.000424 -0.03714 103 29 103 2 0.15291 -0.15291 -0.42486 0.01849 0.000577 -0.02560 104 29 104 2 0.10012 -0.10012 -0.33356 0.05533 0.001150 -0.00989 105 30 105 4 0.46138 -0.46138 -0.92552 0.04911 0.007754 -0.06541 106 30 106 4 0.14210 -0.14210 -0.40698 0.01606 0.000458 -0.00475 107 30 107 4 0.30819 -0.30819 -0.66745 0.09599 0.008721 -0.12310 108 30 108 4 0.20313 -0.20313 -0.50488 0.02882 0.001298 -0.02429 DFBETAS_ Obs FEMALE AGE DAYACC SEVERE TOILET INTER 100 -0.21880 -0.10985 -0.006048 -0.00672 -0.004195 0.29869 101 0.03678 0.01592 0.002552 0.01569 0.000994 -0.08280 102 -0.00559 0.00698 0.001238 0.01473 -0.000231 -0.03351 103 -0.00671 0.01286 0.003369 0.00278 0.000243 -0.02310 104 -0.00464 -0.01025 0.002819 0.01761 -0.005117 -0.00893 105 -0.03648 0.03220 -0.011007 0.00593 0.012103 -0.05902 106 -0.00744 -0.00825 0.003477 0.00225 -0.001056 -0.00428 107 0.09664 -0.04146 0.001098 -0.00467 0.008567 -0.11107 108 -0.00690 0.02297 0.004270 0.00576 -0.002635 -0.02192 DFBETAS_ DFBETAS_ DFBETAS_ DFBETAS_ DFBETAS_ Obs FEMALE AGE DAYACC SEVERE TOILET 100 -0.35775 -0.18785 -0.06313 -0.01844 -0.04904 101 0.06014 0.02723 0.02664 0.04305 0.01162 102 -0.00914 0.01194 0.01292 0.04041 -0.00270 103 -0.01096 0.02199 0.03517 0.00762 0.00284 104 -0.00758 -0.01753 0.02943 0.04831 -0.05982 105 -0.05965 0.05506 -0.11490 0.01626 0.14150 106 -0.01217 -0.01410 0.03629 0.00618 -0.01234 107 0.15801 -0.07091 0.01147 -0.01282 0.10016 108 -0.01128 0.03928 0.04457 0.01579 -0.03081 Logistic regression diagnostics for UI data from SAS IML macro Cookdobs is given by Corollary 2.2 of Preisser and Qaqish under R=I Cookdobs under R=I is equivalent to Williams (1987) DFBETA from SAS GEE macro is standardized to match PROC LOGISTIC output nse_ nse_ nse_ nse_ nse_ Obs inter female nse_age dayacc severe toilet ID 109 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 207 110 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 207 111 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 207 112 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 208 113 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 208 114 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 211 115 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 211 116 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 211 117 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 211 Obs I IJ NI FIT RES SRES QWOBS COOKDOBS INTER 109 31 109 3 0.34930 -0.34930 -0.73268 0.03371 0.003231 0.03352 110 31 110 3 0.39828 -0.39828 -0.81357 0.02065 0.002375 -0.02430 111 31 111 3 0.22094 -0.22094 -0.53254 0.07875 0.004386 -0.05625 112 32 112 2 0.14363 -0.14363 -0.40954 0.02672 0.000788 -0.00687 113 32 113 2 0.24410 -0.24410 -0.56826 0.04582 0.002709 -0.07155 114 33 114 4 0.79537 0.20463 0.50723 0.03920 0.001821 -0.05641 115 33 115 4 0.70609 0.29391 0.64517 0.03745 0.002804 -0.07678 116 33 116 4 0.11322 -0.11322 -0.35732 0.04203 0.000975 -0.01286 117 33 117 4 0.59058 0.40942 0.83261 0.06589 0.008725 -0.09687 DFBETAS_ Obs FEMALE AGE DAYACC SEVERE TOILET INTER 109 -0.02840 -0.05755 -0.002594 0.00448 0.000036 0.03025 110 -0.02747 0.00643 -0.004782 0.00672 0.003841 -0.02193 111 0.07236 -0.04521 0.005816 -0.00084 -0.001311 -0.05076 112 -0.01236 -0.02413 0.001445 -0.00047 0.003117 -0.00620 113 -0.01850 0.00711 -0.003382 0.03922 -0.000456 -0.06456 114 0.01307 -0.00260 0.005906 0.02314 -0.001484 -0.05090 115 0.01275 -0.01710 0.003083 0.03097 0.002376 -0.06927 116 -0.01098 -0.03002 0.000301 0.01638 -0.000649 -0.01160 117 0.03512 0.08110 0.010423 -0.00985 0.006142 -0.08740 DFBETAS_ DFBETAS_ DFBETAS_ DFBETAS_ DFBETAS_ Obs FEMALE AGE DAYACC SEVERE TOILET 109 -0.04644 -0.09842 -0.02708 0.01230 0.00042 110 -0.04491 0.01100 -0.04992 0.01842 0.04490 111 0.11832 -0.07732 0.06071 -0.00231 -0.01532 112 -0.02021 -0.04127 0.01508 -0.00129 0.03645 113 -0.03024 0.01216 -0.03530 0.10759 -0.00533 114 0.02137 -0.00444 0.06165 0.06346 -0.01734 115 0.02085 -0.02924 0.03218 0.08497 0.02778 116 -0.01796 -0.05134 0.00315 0.04492 -0.00759 117 0.05742 0.13869 0.10881 -0.02701 0.07180 Logistic regression diagnostics for UI data from SAS IML macro Cookdobs is given by Corollary 2.2 of Preisser and Qaqish under R=I Cookdobs under R=I is equivalent to Williams (1987) DFBETA from SAS GEE macro is standardized to match PROC LOGISTIC output nse_ nse_ nse_ nse_ nse_ Obs inter female nse_age dayacc severe toilet ID 118 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 216 119 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 216 120 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 216 121 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 216 122 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 220 123 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 220 124 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 220 125 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 220 126 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 220 Obs I IJ NI FIT RES SRES QWOBS COOKDOBS INTER 118 34 118 4 0.11236 -0.11236 -0.35579 0.03079 0.000692 0.01358 119 34 119 4 0.22191 -0.22191 -0.53405 0.02465 0.001232 -0.00463 120 34 120 4 0.44367 0.55633 1.11979 0.08491 0.021191 -0.25005 121 34 121 4 0.07972 -0.07972 -0.29433 0.01849 0.000277 -0.00365 122 35 122 5 0.41566 -0.41566 -0.84340 0.16069 0.027042 0.14814 123 35 123 5 0.89025 0.10975 0.35111 0.04020 0.000897 -0.01884 124 35 124 5 0.10482 -0.10482 -0.34218 0.01798 0.000364 -0.00623 125 35 125 5 0.26529 -0.26529 -0.60090 0.04826 0.003206 0.00348 126 35 126 5 0.32800 0.67200 1.43136 0.01349 0.004732 0.02922 DFBETAS_ Obs FEMALE AGE DAYACC SEVERE TOILET INTER 118 -0.00720 -0.02705 0.003474 0.00130 -0.001938 0.01225 119 -0.01812 -0.02874 0.000356 0.00091 0.003324 -0.00417 120 0.04201 0.15560 -0.003669 0.07400 0.002783 -0.22561 121 -0.00603 -0.01485 0.002217 -0.00023 0.001052 -0.00329 122 -0.01806 0.01208 0.012565 -0.12748 0.011694 0.13366 123 0.00568 -0.02189 0.004691 0.01128 -0.001081 -0.01699 124 -0.00780 -0.01537 0.002266 -0.00001 0.001454 -0.00562 125 -0.02588 -0.05052 -0.002157 -0.00026 0.006033 0.00314 126 0.04134 -0.00239 0.001444 -0.01260 -0.001135 0.02636 DFBETAS_ DFBETAS_ DFBETAS_ DFBETAS_ DFBETAS_ Obs FEMALE AGE DAYACC SEVERE TOILET 118 -0.01178 -0.04625 0.03627 0.00356 -0.02266 119 -0.02963 -0.04915 0.00372 0.00249 0.03886 120 0.06870 0.26609 -0.03830 0.20298 0.03253 121 -0.00986 -0.02540 0.02314 -0.00063 0.01230 122 -0.02953 0.02065 0.13116 -0.34968 0.13671 123 0.00928 -0.03743 0.04897 0.03095 -0.01264 124 -0.01275 -0.02628 0.02365 -0.00004 0.01699 125 -0.04232 -0.08639 -0.02252 -0.00072 0.07053 126 0.06760 -0.00409 0.01507 -0.03457 -0.01327 Logistic regression diagnostics for UI data from SAS IML macro Cookdobs is given by Corollary 2.2 of Preisser and Qaqish under R=I Cookdobs under R=I is equivalent to Williams (1987) DFBETA from SAS GEE macro is standardized to match PROC LOGISTIC output nse_ nse_ nse_ nse_ nse_ Obs inter female nse_age dayacc severe toilet ID 127 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 228 128 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 228 129 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 228 130 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 232 131 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 232 132 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 232 133 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 232 134 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 235 135 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 235 Obs I IJ NI FIT RES SRES QWOBS COOKDOBS INTER 127 36 127 3 0.12648 -0.12648 -0.38051 0.02520 0.000640 0.00819 128 36 128 3 0.79091 0.20909 0.51416 0.06937 0.003529 0.02115 129 36 129 3 0.03469 -0.03469 -0.18956 0.01332 0.000082 -0.01166 130 37 130 4 0.27752 -0.27752 -0.61977 0.04346 0.003041 0.04817 131 37 131 4 0.98129 0.01871 0.13807 0.01584 0.000052 -0.00315 132 37 132 4 0.89769 0.10231 0.33760 0.05681 0.001213 -0.01747 133 37 133 4 0.59574 0.40426 0.82376 0.08598 0.011639 0.17894 134 38 134 4 0.15918 -0.15918 -0.43511 0.05301 0.001865 0.02546 135 38 135 4 0.22451 0.77549 1.85853 0.05985 0.038979 0.24337 DFBETAS_ Obs FEMALE AGE DAYACC SEVERE TOILET INTER 127 -0.00938 -0.02830 0.002748 0.00065 -0.000220 0.00739 128 0.01679 -0.03492 0.011549 -0.00564 -0.004455 0.01908 129 -0.00306 -0.00568 0.000830 0.00444 0.000577 -0.01052 130 -0.01284 -0.00978 0.006788 -0.03462 0.001273 0.04346 131 -0.00587 -0.00054 0.001124 0.00189 0.000329 -0.00284 132 -0.03207 0.01727 0.003084 0.01312 0.000661 -0.01576 133 -0.13818 -0.05033 0.004572 -0.00495 -0.002868 0.16145 134 -0.01119 -0.02701 0.004695 -0.02320 0.004132 0.02297 135 0.07883 0.04485 0.017579 -0.11441 -0.016011 0.21958 DFBETAS_ DFBETAS_ DFBETAS_ DFBETAS_ DFBETAS_ Obs FEMALE AGE DAYACC SEVERE TOILET 127 -0.01534 -0.04840 0.02869 0.00179 -0.00257 128 0.02745 -0.05971 0.12055 -0.01547 -0.05208 129 -0.00500 -0.00972 0.00866 0.01219 0.00675 130 -0.02100 -0.01673 0.07086 -0.09498 0.01489 131 -0.00960 -0.00092 0.01174 0.00518 0.00385 132 -0.05243 0.02953 0.03219 0.03599 0.00773 133 -0.22594 -0.08608 0.04772 -0.01357 -0.03353 134 -0.01829 -0.04619 0.04901 -0.06365 0.04831 135 0.12890 0.07670 0.18350 -0.31382 -0.18718 Logistic regression diagnostics for UI data from SAS IML macro Cookdobs is given by Corollary 2.2 of Preisser and Qaqish under R=I Cookdobs under R=I is equivalent to Williams (1987) DFBETA from SAS GEE macro is standardized to match PROC LOGISTIC output nse_ nse_ nse_ nse_ nse_ Obs inter female nse_age dayacc severe toilet ID 136 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 235 137 1.10833 0.61159 0.58475 0.095797 0.36455 0.085535 235 Obs I IJ NI FIT RES SRES QWOBS COOKDOBS INTER 136 38 136 4 0.24222 -0.24222 -0.56537 0.05277 0.003133 -0.01868 137 38 137 4 0.60681 0.39319 0.80497 0.05055 0.006056 0.05189 DFBETAS_ Obs FEMALE AGE DAYACC SEVERE TOILET INTER 136 -0.00563 0.03445 0.005582 0.00938 -0.006877 -0.01685 137 0.02155 -0.07908 0.010185 -0.01258 -0.001913 0.04682 DFBETAS_ DFBETAS_ DFBETAS_ DFBETAS_ DFBETAS_ Obs FEMALE AGE DAYACC SEVERE TOILET 136 -0.00921 0.05891 0.05827 0.02573 -0.08040 137 0.03523 -0.13524 0.10632 -0.03450 -0.02236