Refer to the clinical trial in Problem 3.10 with Table 3.6

 a. Fit logit model (7.1.1). Using the fitted values, compute the estimated odds ratio between group and response (i) for each center, (ii) for the marginal table. Why do they differ?

b. Fit the simpler logit model deleting the center effects to this 223 table. Using the fitted values, compute the estimated odds ratio between group and response (i) for each center, (ii) for the marginal table. When this model fits well, can we collapse over centers? Compare the fit to the model in (a). Is the simpler model adequate?