Next, perform a multiple imputation analysis to estimate the mean BPXSY1, and compute a variance estimate for the estimated mean that incorporates the design-based standard errors from the analysis of each imputed data set in addition to the between-imputation variance in the estimates. (Use a command designed to analyze multiply imputed data sets and incorporate the complex sample variance corrections in addition to the variability introduced by imputing the missing values multiple times. In Stata Version 10+ this command would be mim, in SAS it would be PROC MIANALYZE, and in IVEware it would be %DESCRIBE.)
a. What is the overall estimated mean for systolic blood pressure for all imputed data sets combined?
b. What is the standard error of the estimated mean, and what has been accounted for in the standard error when using the mim modifier with the svy: mean command (in Stata) or a similar command in another software package?
c. Obtain the within- and between-imputation variance components by either calculating these statistics manually or requesting them from the software package of choice (see Section 11.6.1).