Using the covariates in the population average model obtained in problem 5

(a) explore alternative ways of including history of a low birth weight baby. For reach model compute and interpret the esti-mate of the odds ratio and a 95 percent confidence interval for the previous low birth weight covariate. Recall that we fit these models using the usual logistic regression model.

(a) Fit the model that adds the outcome of the previous birth to the model. In this problem explore the use of two versions of this covariate: one that assigns a missing value for the first birth and one that assigns a value of zero to the first birth.

(b) Fit the model that includes a dichotomous covariate indicating if any previous birth was of low weight. In this problem explore the use of two versions of this covariate: one that assigns a missing value for the first birth and one that assigns a value of zero to the first birth.