Due: February 8, 2006

1.  Use SAS to perform multiple imputations on a dataset of your choice, using the MCMC method.  Impute at least 3 datasets.
 a) Display and interpret the time series plot for the worst linear function.  At approximately what iteration would you judge the iterations to have converged to a stable distribution?
 b) Display and interpret the autocorrelation plot for the worst linear function.  What appears to be the minimum amount of iterations between imputations if we want correlations below 0.05?
 c) Run basic descriptives (means and variances) on three of the imputed datasets.  Please only run them on one variable that had missing data initially!
 d) Would you have been able to use a monotone missing data method on your dataset?

Notes:
If your PROC MI output data set is called "Imputed" in the library "lib", the following code creates a subset of only the first imputed dataset:
data lib.Imputed_subset;
set lib.Imputed;
where _Imputation_ = 1;
run;