We discuss estimating population-averaged parameters when some of the data are missing. In particular, we show how to use gmm to estimate population-averaged parameters for a probit model when the process that causes some of the data to be missing is a function of observable covariates and a random process that is independent of the […]
↧