1              EXAMPLE 11. LOGISTIC REGRESSION: HEALTH DATA             1
                                          16:37 Sunday, December 30, 2001

                          The LOGISTIC Procedure

                            Model Information

              Data Set                      WORK.ESR
              Response Variable             RESPONSE
              Number of Response Levels     2
              Number of Observations        32
              Link Function                 Logit
              Optimization Technique        Fisher's scoring


                             Response Profile

                    Ordered                      Total
                      Value     RESPONSE     Frequency

                          1            1             6
                          2            0            26


                         Model Convergence Status

              Convergence criterion (GCONV=1E-8) satisfied.


                          Model Fit Statistics

                                               Intercept
                                Intercept         and
                 Criterion        Only        Covariates

                 AIC               32.885         28.971
                 SC                34.351         33.368
                 -2 Log L          30.885         22.971


                 Testing Global Null Hypothesis: BETA=0

         Test                 Chi-Square       DF     Pr > ChiSq

         Likelihood Ratio         7.9138        2         0.0191
         Score                    8.2067        2         0.0165
         Wald                     4.7561        2         0.0927


                Analysis of Maximum Likelihood Estimates

                                  Standard
   Parameter    DF    Estimate       Error    Chi-Square    Pr > ChiSq

   Intercept     1    -12.7920      5.7964        4.8704        0.0273
   FIBRIN        1      1.9104      0.9710        3.8708        0.0491
   GLOBULIN      1      0.1558      0.1195        1.6982        0.1925
1              EXAMPLE 11. LOGISTIC REGRESSION: HEALTH DATA             2
                                          16:37 Sunday, December 30, 2001

                          The LOGISTIC Procedure

                           Odds Ratio Estimates

                              Point          95% Wald
               Effect      Estimate      Confidence Limits

               FIBRIN         6.756       1.007      45.308
               GLOBULIN       1.169       0.924       1.477


      Association of Predicted Probabilities and Observed Responses

            Percent Concordant     80.1    Somers' D    0.609
            Percent Discordant     19.2    Gamma        0.613
            Percent Tied            0.6    Tau-a        0.192
            Pairs                   156    c            0.804