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

Answers to
Practice
Exercises

1. Exchangeable correlation structure


2. Stationary 1-dependent correlation structure


3. Independent correlation structure


4. Autoregressive (AR1) correlation structure


5. Unstructured correlation structure


6. T


7. F: the likelihood is never formulated in a GEE model


8. F: the estimation of parameters is more efficient [i.e.,
   smaller varð^bÞ] if the correct correlation structure is
   specified


9. T


10. F: the converse is true (i.e., ifXandYare independent,
    then the correlation is 0). The correlation is a measure
    of linearity.XandYcould have a nonlinear depen-
    dence and have a correlation of 0. In the special case
    whereXandYfollow a normal distribution, then a
    correlation of 0 does imply independence.

538 14. Logistic Regression for Correlated Data: GEE