ABSTRACT
This chapter applies the semiparametric odds ratio model to handle missing covariate data by either the maximum semiparametric likelihood approach or the imputation approach. Two modeling approaches are used. One is based on the consecutive conditional odds ratio model, the other is based on the joint semiparametric odds ratio model. Another application is to model data with a special not missing at random missing data. The modeling approach makes the challenging problem of analysis of such data trackable.
