Output from logist1.sas	Source 0 Graphs

                     Predicting Improved vs. Not Improved                    1

                            The LOGISTIC Procedure

    Data Set: WORK.ARTHRIT
    Response Variable: BETTER
    Response Levels: 2
    Number of Observations: 84
    Link Function: Logit

                               Response Profile

                          Ordered
                            Value  BETTER     Count

                                1       1        42
                                2       0        42

               Deviance and Pearson Goodness-of-Fit Statistics

                                                            Pr >
          Criterion        DF       Value    Value/DF    Chi-Square

          Deviance          1      0.2776      0.2776        0.5983
          Pearson           1      0.2637      0.2637        0.6076

                         Number of unique profiles: 4

     Model Fitting Information and Testing Global Null Hypothesis BETA=0

                              Intercept
                Intercept        and
  Criterion       Only       Covariates    Chi-Square for Covariates

  AIC             118.449       104.222         .
  SC              120.880       111.514         .
  -2 LOG L        116.449        98.222       18.227 with 2 DF (p=0.0001)
  Score              .             .          16.797 with 2 DF (p=0.0002)

                   Analysis of Maximum Likelihood Estimates

              Parameter Standard    Wald       Pr >    Standardized     Odds
  Variable DF  Estimate   Error  Chi-Square Chi-Square   Estimate      Ratio

  INTERCPT 1    -1.9037   0.5982    10.1286     0.0015            .     .
  _SEX_    1     1.4687   0.5756     6.5092     0.0107     0.372433    4.343
  _TREAT_  1     1.7817   0.5188    11.7961     0.0006     0.493956    5.940

                     Predicting Improved vs. Not Improved                    2

                            The LOGISTIC Procedure

        Association of Predicted Probabilities and Observed Responses

                  Concordant = 61.7%          Somers' D = 0.480
                  Discordant = 13.8%          Gamma     = 0.635
                  Tied       = 24.5%          Tau-a     = 0.243
                  (1764 pairs)                c         = 0.740

                     Predicting Improved vs. Not Improved                    3
                           Testing Sex * Treatment

                            The LOGISTIC Procedure

    Data Set: WORK.ARTHRIT
    Response Variable: BETTER
    Response Levels: 2
    Number of Observations: 84
    Link Function: Logit

                               Response Profile

                          Ordered
                            Value  BETTER     Count

                                1       1        42
                                2       0        42

     Model Fitting Information and Testing Global Null Hypothesis BETA=0

                              Intercept
                Intercept        and
  Criterion       Only       Covariates    Chi-Square for Covariates

  AIC             118.449       105.944         .
  SC              120.880       115.667         .
  -2 LOG L        116.449        97.944       18.505 with 3 DF (p=0.0003)
  Score              .             .          16.822 with 3 DF (p=0.0008)

                   Analysis of Maximum Likelihood Estimates

              Parameter Standard    Wald       Pr >    Standardized     Odds
  Variable DF  Estimate   Error  Chi-Square Chi-Square   Estimate      Ratio

  INTERCPT 1    -2.3026   1.0488     4.8200     0.0281            .     .
  _SEX_    1     1.9231   1.1088     3.0079     0.0829     0.487673    6.842
  _TREAT_  1     2.3026   1.1772     3.8262     0.0505     0.638372   10.000
  SEXTRT   1    -0.6703   1.3151     0.2598     0.6103    -0.173637    0.512

        Association of Predicted Probabilities and Observed Responses

                  Concordant = 61.7%          Somers' D = 0.480
                  Discordant = 13.8%          Gamma     = 0.635
                  Tied       = 24.5%          Tau-a     = 0.243
                  (1764 pairs)                c         = 0.740

                     Predicting Improved vs. Not Improved                    4
                      Testing all interactions stepwise

                            The LOGISTIC Procedure

    Data Set: WORK.ARTHRIT
    Response Variable: BETTER
    Response Levels: 2
    Number of Observations: 84
    Link Function: Logit

                               Response Profile

                          Ordered
                            Value  BETTER     Count

                                1       1        42
                                2       0        42

                         Forward Selection Procedure

Step  0. The following variables were entered:
         INTERCPT  _SEX_     _TREAT_   AGE

     Model Fitting Information and Testing Global Null Hypothesis BETA=0

                              Intercept
                Intercept        and
  Criterion       Only       Covariates    Chi-Square for Covariates

  AIC             118.449       100.063         .
  SC              120.880       109.786         .
  -2 LOG L        116.449        92.063       24.386 with 3 DF (p=0.0001)
  Score              .             .          22.005 with 3 DF (p=0.0001)


              Residual Chi-Square = 4.0268 with 4 DF (p=0.4024)



Step  1. Variable AGESEX entered:

                     Predicting Improved vs. Not Improved                    5
                      Testing all interactions stepwise

                            The LOGISTIC Procedure

     Model Fitting Information and Testing Global Null Hypothesis BETA=0

                              Intercept
                Intercept        and
  Criterion       Only       Covariates    Chi-Square for Covariates

  AIC             118.449        98.640         .
  SC              120.880       110.794         .
  -2 LOG L        116.449        88.640       27.809 with 4 DF (p=0.0001)
  Score              .             .          24.472 with 4 DF (p=0.0001)


              Residual Chi-Square = 0.2903 with 3 DF (p=0.9618)



Step  2. Variable SEXTRT entered:

     Model Fitting Information and Testing Global Null Hypothesis BETA=0

                              Intercept
                Intercept        and
  Criterion       Only       Covariates    Chi-Square for Covariates

  AIC             118.449       100.369         .
  SC              120.880       114.954         .
  -2 LOG L        116.449        88.369       28.080 with 5 DF (p=0.0001)
  Score              .             .          24.633 with 5 DF (p=0.0002)


              Residual Chi-Square = 0.0305 with 2 DF (p=0.9849)



Step  3. Variable AGE2 entered:

                     Predicting Improved vs. Not Improved                    6
                      Testing all interactions stepwise

                            The LOGISTIC Procedure

     Model Fitting Information and Testing Global Null Hypothesis BETA=0

                              Intercept
                Intercept        and
  Criterion       Only       Covariates    Chi-Square for Covariates

  AIC             118.449       102.350         .
  SC              120.880       119.366         .
  -2 LOG L        116.449        88.350       28.099 with 6 DF (p=0.0001)
  Score              .             .          24.647 with 6 DF (p=0.0004)


              Residual Chi-Square = 0.0110 with 1 DF (p=0.9166)



Step  4. Variable AGETRT entered:

     Model Fitting Information and Testing Global Null Hypothesis BETA=0

                              Intercept
                Intercept        and
  Criterion       Only       Covariates    Chi-Square for Covariates

  AIC             118.449       104.339         .
  SC              120.880       123.785         .
  -2 LOG L        116.449        88.339       28.110 with 7 DF (p=0.0002)
  Score              .             .          24.649 with 7 DF (p=0.0009)

NOTE: All explanatory variables have been entered into the model.

                    Summary of Forward Selection Procedure

                    Variable    Number       Score         Pr >
            Step    Entered         In    Chi-Square    Chi-Square

               1    AGESEX           4        3.6874        0.0548
               2    SEXTRT           5        0.2588        0.6110
               3    AGE2             6        0.0194        0.8893
               4    AGETRT           7        0.0110        0.9166

                     Predicting Improved vs. Not Improved                    7
                      Testing all interactions stepwise

                            The LOGISTIC Procedure

                   Analysis of Maximum Likelihood Estimates

              Parameter Standard    Wald       Pr >    Standardized     Odds
  Variable DF  Estimate   Error  Chi-Square Chi-Square   Estimate      Ratio

  INTERCPT 1    -1.3492   5.4280     0.0618     0.8037            .     .
  _SEX_    1    -2.3273   2.9165     0.6367     0.4249    -0.590162    0.098
  _TREAT_  1     1.9655   3.0666     0.4108     0.5216     0.544916    7.138
  AGE      1    -0.0329   0.1869     0.0310     0.8602    -0.231651    0.968
  AGESEX   1     0.0797   0.0501     2.5241     0.1121     1.178622    1.083
  AGETRT   1    0.00541   0.0517     0.0110     0.9166     0.086555    1.005
  SEXTRT   1    -0.6324   1.3744     0.2117     0.6454    -0.163804    0.531
  AGE2     1   0.000276  0.00177     0.0243     0.8761     0.190087    1.000

        Association of Predicted Probabilities and Observed Responses

                  Concordant = 81.0%          Somers' D = 0.629
                  Discordant = 18.0%          Gamma     = 0.636
                  Tied       =  1.0%          Tau-a     = 0.318
                  (1764 pairs)                c         = 0.815

                     Proportional Odds Model for IMPROVE                     8

                            The LOGISTIC Procedure

    Data Set: WORK.ARTHRIT
    Response Variable: IMPROVE
    Response Levels: 3
    Number of Observations: 84
    Link Function: Logit

                               Response Profile

                          Ordered
                            Value  IMPROVE     Count

                                1        0        42
                                2        1        14
                                3        2        28

               Score Test for the Proportional Odds Assumption

                   Chi-Square = 1.8833 with 2 DF (p=0.3900)

     Model Fitting Information and Testing Global Null Hypothesis BETA=0

                              Intercept
                Intercept        and
  Criterion       Only       Covariates    Chi-Square for Covariates

  AIC             173.916       158.029         .
  SC              178.778       167.753         .
  -2 LOG L        169.916       150.029       19.887 with 2 DF (p=0.0001)
  Score              .             .          17.868 with 2 DF (p=0.0001)

                   Analysis of Maximum Likelihood Estimates

              Parameter Standard    Wald       Pr >    Standardized     Odds
  Variable DF  Estimate   Error  Chi-Square Chi-Square   Estimate      Ratio

  INTERCP1 1     1.8128   0.5566    10.6072     0.0011            .     .
  INTERCP2 1     2.6672   0.5997    19.7809     0.0001            .     .
  _SEX_    1    -1.3187   0.5292     6.2102     0.0127    -0.334418    0.267
  _TREAT_  1    -1.7973   0.4728    14.4493     0.0001    -0.498287    0.166

                     Proportional Odds Model for IMPROVE                     9

                            The LOGISTIC Procedure

        Association of Predicted Probabilities and Observed Responses

                  Concordant = 58.8%          Somers' D = 0.438
                  Discordant = 15.0%          Gamma     = 0.593
                  Tied       = 26.2%          Tau-a     = 0.271
                  (2156 pairs)                c         = 0.719