Categorical Data Analysis, Agresti

9.4 Consider the loglinear model selection for Table 6.3. GM b. Using forward selection starting with (GM, E, P), show that model e. Using backward elimination, show that (GM, GP, EMP) or d. The EMP interaction seems vital. To describe it, show that the (GM, GP, EG, EMP) seems reasonable. (GM, GP, EG, EMP) seems reasonable. effect of extramarital sex on divorce is greater for subjects who had no premarital sex e. Use residuals to describe the lack of fit of model (GM, EMP) 9.4 Consider the loglinear model selection for Table 6.3. GM b. Using forward selection starting with (GM, E, P), show that model e. Using backward elimination, show that (GM, GP, EMP) or d. The EMP interaction seems vital. To describe it, show that the (GM, GP, EG, EMP) seems reasonable. (GM, GP, EG, EMP) seems reasonable. effect of extramarital sex on divorce is greater for subjects who had no premarital sex e. Use residuals to describe the lack of fit of model (GM, EMP)