Data Set 2 Appendix: Computer Data Sets 255 Data set name: bronchitis.dat Variable 1 = number of bronchitis cases-BRC 2 = total number at risk in stratum 38 48 28 40 84 102 47 59 3 = smoking status (0 = never, 1 = ever)-SMK 4 = socioeconomic status (0 = low, 1 = high)-SES 5 = age (0 = below 40, 1 = 40-59)-AGE 111 0 1 0 134 0 1 1 95 0 0 0 124 0 0 1 173 1 1 0 148 1 1 1 143 1 0 0 112 1 0 1 SampIe Questions A. Fit a logistic model that predicts BRC from the predictors SMK, SES, andAGE.

B. Fit a logistic model that predicts BRC from the predictors SMK, SES, AGE, SMK x SES, and SMK x AGE.

C. Determine the estimated odds ratio for the effect of SMK adjusted for SES and AGE based on the model fit in Ouestion A.

D. Determine the formula for the estimated odds ratio for the effect of SMK adjusted for SES and AGE based on the model fit in Question B.

Use this odds ratio formula to derive individual odds ratio estimates for various choices of SES and AGE.

E. Fit the same model as used in Question A but code the SMK variables as -1 = low, + 1 = high. Compare your results in Ouestions A and C with the corresponding results using the new coding.

F. Test for interaction using a likelihood ratio test.

G. Test for the significance of SMK adjusting for SES and AGE using both Wald and LR tests.

H. Obtain 95% confidence intervals for the effect of SMK adjusted for SES andAGE.

I. Use a hierarchical backward elimination strategy to obtain the "best"

model.