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. Given a fitted binary logistic model, describe or
   illustrate how a cut-point can be used to classify
   subjects as predicted cases (Y¼1) and predicted
   noncases (Y¼0).


2. Given a fitted binary logistic model, describe or
   illustrate how a cut-point can be used to form a
   misclassification (or diagnostic table).


3. Define and illustrate what is meant by true positives,
   false positives, true negatives, and false negatives.


4. Define and illustrate what is meant by sensitivity and
   specificity.


5. Define and illustrate “perfect discrimination.”


6. Describe what happens to sensitivity and specificity
   parameters when a cut-point used for discrimination
   of a fitted logistic model decreases from 1 to 0.


7. Describe what happens to (1specificity) when a cut-
   point used for discrimination decreases from 1 to 0.


8. State one or more uses of an ROC curve.


9. State and/or describe briefly how an ROC curve is
   constructed.


10. State and/or describe briefly how the area under an
    ROC curves is calculated.


11. Describe briefly how to interpret a calculated area
    under an ROC curve in terms of the discriminatory
    performance of a fitted logistic model.


12. Given a printout of a fitted binary logistic model,
    evaluate how well the model discriminates cases from
    noncases.

Objectives 347