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

Practice Exercises

Practice Exercises

The following questions and computer information con-

sider the Evans Country dataset on 609 white males that

has been previously discussed and illustrated in earlier

chapters of this text. Recall that the outcome variable is

CHD status (1¼case, 0¼noncase), the exposure variable

of interest is CAT status (1¼high CAT, 0¼low CAT). In

this example, we consider only two categorical control

variables AGEG (1¼age>55, 0¼age55) and ECG

(1¼abnormal, 0¼normal). The dataset involving the

above variables is given as follows:

Cases Total CAT AGE ECG

17 274 0 0 0

15 122 0 1 0

759001

532011

1 8100

939110

317101

14 58 1 1 1

The SAS output provided below was obtained for the fol-

lowing logistic model:

Logit PðXÞ¼aþb 1 CATþg 1 AGEþg 2 ECG

Deviance and Pearson Goodness-of-Fit Statistics

Criterion Value DF Value/DF Pr>ChiSq

Deviance 0.9544 4 0.2386 0.9166

Pearson 0.9793 4 0.2448 0.9129

Number of unique profiles: 8

Model Fit Statistics

Criterion Intercept Only Intercept and Covariates

2 Log L 438.558 418.181

Analysis of Maximum Likelihood Estimates

Parameter DF Estimate

Std

Error

Wald

Chi-Sq Pr>ChiSq

Intercept 1 2.6163 0.2123 151.8266 <.0001

cat 1 0.6223 0.3193 3.7978 0.0513

age 1 0.6157 0.2838 4.7050 0.0301

ecg 1 0.3620 0.2904 1.5539 0.2126

334 9. Assessing Goodness of Fit for Logistic Regression