I. Analyze the data in Table, using logistic regression to analyze sets of 2 2 tables.

(a) Report the conditional odds ratios for each department and their 95% confidence intervals.

(b) Perform the tests using logistic regression equivalent to the Cochran-MantelHaenszel and Breslow-Day tests. Write down their null and alternative hypotheses and interpret the results.

(c) If you delete the data for Department 1, do you think a common odds ratio is reasonable? If so, report an estimate for the common odds ratio using computer output and report a 90% confidence interval for this quantity.

Table 5.7 Data for Exercise 5.9 on admissions to Berkeley. Department Yes 1 512 2 353 3 120 4 138 5 53 6 22 Total 1 198 Note: Based on data in P. Hicks] 9! (11., Science 187: 398403 (19315). Admitted. Male No 313 207 205 279 138 351 1493 Yes 89 17 202 131 94 24 557 Admitted, Female No 1 9 8 39 l 244 299 3 1 '1' 1278