We are concerned with the problem of parameter estimation in normal regression when some of the observations are missing. A Bayesian approach with vague prior distributions is taken. No assumption is made about the independent variables for which n...
