S U D A A N Software for the Statistical Analysis of Correlated Data Copyright Research Triangle Institute July 2004 Release 9.0.0 ****************************************************************** SUDAAN Release 9.0.0 (Windows Network StandAlone version) Serial Number A0002933, for 1 user(s) is licensed to UGA (Mary Meyer). It expires on December 06, 2005. ****************************************************************** 1 PROC LOGISTIC DATA="h:\Papers\airbags\sascode\nass9702" FILETYPE=SASXPORT DESIGN =WOR; 2 NEST PSUSTRAT PSU STR2; 3 TOTCNT PSTRN _ZERO_ _MINUS1_; 4 WEIGHT NIF; 5 SUBGROUP DVCAT DOFCAT; 6 LEVELS 5 4; 7 REFLEVEL DVCAT=2 DOFCAT=1; 8 MODEL DEAD2= SEATBELT DVCAT OLD DOFCAT ABNSBLSP SMCARHSP AIRBAG ; 9 SETENV COLWIDTH=8 DECWIDTH=4 COLSPCE=2 LINESIZE=100; 10 PRINT; 11 PRINT / HLTEST=ALL; /* this is the subset of FARS example*/ /* PROC LOGISTIC data="h:\Papers\airbags\sascode\str19702" filetype=sasxport design=wor; NEST psustrat psu str2; totcnt pstrn _zero_ _minus1_; WEIGHT nif; subgroup dvcat dofcat abcat; levels 5 4 3; REFLEVEL dvcat=2 dofcat=1 abcat=1; MODEL dead2=seatbelt dvcat old dofcat abcat smcarhsp ; SETENV COLWIDTH=8 DECWIDTH=4 COLSPCE=2 LINESIZE=100; print; Opened SASXPORT data file h:\Papers\airbags\sascode\nass9702.XPT for reading. Unable to open file h:\Papers\airbags\sascode\LEVFILE.XPT PRINT / HLTEST=ALL;*/ Number of zero responses : 21735 Number of non-zero responses : 908 Independence parameters have converged in 10 iterations DATA WARNING: One or more parameters are approaching infinity. The data may have singularities for the model you are trying to fit. Number of observations read : 22643 Weighted count: 10597721 Number of observations skipped : 161 (WEIGHT variable nonpositive) Observations used in the analysis : 22643 Weighted count: 10597721 Denominator degrees of freedom : 15 Maximum number of estimable parameters for the model is 13 File h:\Papers\airbags\sascode\nass9702.XPT contains 133 Clusters 133 clusters were used to fit the model Maximum cluster size is 630 records Minimum cluster size is 2 records Sample and Population Counts for Response Variable DEAD2 0: Sample Count 21735 Population Count 10550521 1: Sample Count 908 Population Count 47200 R-Square for dependent variable DEAD2 (Cox & Snell, 1989): 0.021640 -2 * Normalized Log-Likelihood with Intercepts Only : 1293.21 -2 * Normalized Log-Likelihood Full Model : 797.85 Approximate Chi-Square (-2 * Log-L Ratio) : 495.37 Degrees of Freedom : 12 Note: The approximate Chi-Square is not adjusted for clustering. Refer to hypothesis test table for adjusted test. Date: 01-11-2005 Research Triangle Institute Page : 1 Time: 21:45:32 The LOGISTIC Procedure Table : 1 Variance Estimation Method: Taylor Series (WOR) SE Method: Robust (Binder, 1983) Working Correlations: Independent Link Function: Logit Response variable DEAD2: dead2 by: Independent Variables and Effects. -------------------------------------------------------------------------------- Independent Lower Upper Variables and 95% 95% P-value Effects Beta Limit Limit T-Test T-Test Coeff. SE Beta Beta Beta B=0 B=0 -------------------------------------------------------------------------------- Intercept -8.1815 0.2701 -8.7572 -7.6057 -30.2886 0.0000 seatbelt -1.1008 0.1396 -1.3983 -0.8032 -7.8853 0.0000 dvcat 1 -0.6818 0.5386 -1.8297 0.4661 -1.2660 0.2248 2 0.0000 0.0000 0.0000 0.0000 . . 3 2.9013 0.1754 2.5275 3.2751 16.5429 0.0000 4 3.7214 0.2968 3.0887 4.3540 12.5380 0.0000 5 5.7982 0.2671 5.2289 6.3675 21.7082 0.0000 old 1.6305 0.2181 1.1656 2.0953 7.4758 0.0000 dofcat 1 0.0000 0.0000 0.0000 0.0000 . . 2 -1.7904 0.8061 -3.5086 -0.0722 -2.2210 0.0422 3 0.6564 0.1966 0.2373 1.0755 3.3383 0.0045 4 1.8474 0.1497 1.5282 2.1665 12.3376 0.0000 abnsblsp 1.5393 0.3387 0.8174 2.2611 4.5452 0.0004 smcarhsp 0.9294 0.1915 0.5213 1.3375 4.8543 0.0002 airbag 0.0275 0.1794 -0.3549 0.4099 0.1534 0.8801 -------------------------------------------------------------------------------- Date: 01-11-2005 Research Triangle Institute Page : 2 Time: 21:45:32 The LOGISTIC Procedure Table : 1 Variance Estimation Method: Taylor Series (WOR) SE Method: Robust (Binder, 1983) Working Correlations: Independent Link Function: Logit Response variable DEAD2: dead2 by: Contrast. -------------------------------------------------- Contrast Degrees of P-value Freedom Wald F Wald F -------------------------------------------------- OVERALL MODEL 13.0000 ******** 0.0000 MODEL MINUS INTERCEPT 12.0000 368.5878 0.0000 INTERCEPT . . . SEATBELT 1.0000 62.1780 0.0000 DVCAT 4.0000 199.5327 0.0000 OLD 1.0000 55.8882 0.0000 DOFCAT 3.0000 51.6805 0.0000 ABNSBLSP 1.0000 20.6591 0.0004 SMCARHSP 1.0000 23.5646 0.0002 AIRBAG 1.0000 0.0235 0.8801 -------------------------------------------------- Date: 01-11-2005 Research Triangle Institute Page : 3 Time: 21:45:32 The LOGISTIC Procedure Table : 1 Variance Estimation Method: Taylor Series (WOR) SE Method: Robust (Binder, 1983) Working Correlations: Independent Link Function: Logit Response variable DEAD2: dead2 by: Independent Variables and Effects. -------------------------------------------------- Independent Variables and Lower Upper Effects Odds 95% 95% Ratio Limit OR Limit OR -------------------------------------------------- Intercept 0.0003 0.0002 0.0005 seatbelt 0.3326 0.2470 0.4479 dvcat 1 0.5057 0.1605 1.5937 2 1.0000 1.0000 1.0000 3 18.1972 12.5216 26.4453 4 41.3202 21.9490 77.7872 5 329.7163 186.5924 582.6220 old 5.1063 3.2079 8.1282 dofcat 1 1.0000 1.0000 1.0000 2 0.1669 0.0299 0.9303 3 1.9278 1.2678 2.9314 4 6.3432 4.6100 8.7281 abnsblsp 4.6611 2.2647 9.5935 smcarhsp 2.5330 1.6842 3.8094 airbag 1.0279 0.7013 1.5067 -------------------------------------------------- Date: 01-11-2005 Research Triangle Institute Page : 4 Time: 21:45:32 The LOGISTIC Procedure Table : 1 Variance Estimation Method: Taylor Series (WOR) SE Method: Robust (Binder, 1983) Working Correlations: Independent Link Function: Logit Response variable DEAD2: dead2 ------------------------------------------------------------------------------------------ H-L H-L H-L Chi- H-L ChiSq P- H-L Wald H-L Wald Sattert- Square ChiSq DF value F H-L DF P-value hwaite F ------------------------------------------------------------------------------------------ 4.0075 8.0000 0.8564 3.0139 9.0000 0.0287 2.3049 ------------------------------------------------------------------------------------------ Date: 01-11-2005 Research Triangle Institute Page : 5 Time: 21:45:32 The LOGISTIC Procedure Table : 1 Variance Estimation Method: Taylor Series (WOR) SE Method: Robust (Binder, 1983) Working Correlations: Independent Link Function: Logit Response variable DEAD2: dead2 ---------------------------------------- H-L Sattert- H-L hwaite Sattert- Adjusted hwaite DF P-value ---------------------------------------- 3.1394 0.1164 ---------------------------------------- LOGISTIC used CPU time : 56.0 seconds Elapsed time : 57 seconds Virtual memory : 8.99 MB