In the depression data set, define Y = the square root of total depression score (CESD), X1 = log(income), X2 = Age, X3 = Health and X4 = Bed days. Set X1 = missing whenever X3 = 4 (poor health). Also...

In the depression data set, define Y = the square root of total depression score (CESD), X1 = log(income), X2 = Age, X3 = Health and X4 = Bed days. Set X1 = missing whenever X3 = 4 (poor health). Also set X2 = missing whenever X2 is between 50 and 59 (inclusive). Are these data missing at random? Try various imputation methods and obtain the regression of Y on X1,...,X4 with the imputed values as well as using the complete cases only. Compare these results with those obtained from the original data (nothing missing).