Logistic Regression: A Self-learning Text, Third Edition (Statistics in the Health Sciences)

Model 2

Variable Coefficient

Empirical
Std Err
INTERCEPT 0.5625 0.4058
RX 0.3104 0.3992
SEQUENCE 0.7118 0.5060

Model 3

Variable Coefficient

Empirical

Std Err

INTERCEPT 0.2007 0.3178
RX 0.3008 0.3868

1. State the logit form of the model for Model 1, Model 2,
   and Model 3.


2. Use Model 1 to estimate the odds ratios and 95% confi-
   dence intervals for RX (active vs. standard treatment).
   Hint.Make use of the variance–covariance matrix for
   the parameter estimates.


3. In Model 1, what is the difference between the working
   covariance matrix and the covariance matrix for
   parameter estimates used to obtain the 95% confidence
   interval in the previous question?


4. Use Model 1 to perform the Wald test on the interaction
   term RX*SEQ at a 0.05 level of significance.


5. Use Model 1 to perform the Score test on the interac-
   tion term RX*SEQ at a 0.05 level of significance.


6. Estimate the odds ratio for RX using Model 2 and
   Model 3. Is there a suggestion that SEQUENCE is con-
   founding the association between RX and heartburn
   relief. Answer this question from a data-based perspec-
   tive (i.e., comparing the odds ratios) and a theoretical
   perspective (i.e., what it means to be a confounder).

Test 563