Logistic Regression Model for Probability of Remaining in the Study
Clinical Trial of Contracepting Women

The GENMOD Procedure

Model Information
Data Set WORK.CONTRACEP
Distribution Binomial
Link Function Logit
Dependent Variable r

Number of Observations Read 2902
Number of Observations Used 2902
Number of Events 2465
Number of Trials 2902

Class Level Information
Class Value Design Variables
time 1 0 0
  2 1 0
  3 0 1

Response Profile
Ordered
Value
r Total
Frequency
1 1 2465
2 0 437


PROC GENMOD is modeling the probability that r='1'.

Criteria For Assessing Goodness Of Fit
Criterion DF Value Value/DF
Log Likelihood   -1206.5443  
Full Log Likelihood   -1206.5443  
AIC (smaller is better)   2425.0885  
AICC (smaller is better)   2425.1175  
BIC (smaller is better)   2460.9274  

Algorithm converged.

Analysis Of Maximum Likelihood Parameter Estimates
Parameter   DF Estimate Standard Error Wald 95% Confidence Limits Wald Chi-Square Pr > ChiSq
Intercept   1 1.6681 0.1044 1.4635 1.8727 255.43 <.0001
time 2 1 0.1367 0.1190 -0.0965 0.3700 1.32 0.2507
time 3 1 0.7286 0.1439 0.4466 1.0106 25.65 <.0001
dose   1 0.0680 0.1313 -0.1893 0.3253 0.27 0.6046
prevy   1 -0.4514 0.1619 -0.7687 -0.1341 7.77 0.0053
dose*prevy   1 -0.2381 0.2196 -0.6685 0.1923 1.18 0.2782
Scale   0 1.0000 0.0000 1.0000 1.0000    

Note: The scale parameter was held fixed.