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

Objectives Upon completing this chapter, the learner should be able to:

1. State or recognize when to use unconditional vs.
   conditional ML methods.


2. State or recognize what is a likelihood function.


3. State or recognize that the likelihood functions for
   unconditional vs. conditional ML methods are
   different.


4. State or recognize that unconditional vs. conditional
   ML methods require different computer programs.


5. State or recognize how an ML procedure works to
   obtain ML estimates of unknown parameters in a
   logistic model.


6. Given a logistic model, state or describe two alternative
   procedures for testing hypotheses about parameters in
   the model. In particular, describe each procedure in
   terms of the information used (log likelihood statistic
   orZstatistic) and the distribution of the test statistic
   under the null hypothesis (chi square orZ).


7. State, recognize, or describe three types of information
   required for carrying out statistical inferences
   involving the logistic model: the value of the maximized
   likelihood, the variance–covariance matrix, and a
   listing of the estimated coefficients and their standard
   errors.


8. Given a logistic model, state or recognize how interval
   estimates are obtained for parameters of interest; in
   particular, state that interval estimates are large
   sample formulae that make use of variance and
   covariances in the variance–covariance matrix.


9. Given a printout of ML estimates for a logistic model,
   use the printout information to describe characteristics
   of the fitted model. In particular, given such a printout,
   compute an estimated odds ratio for an exposure–
   disease relationship of interest.

Objectives 105