Multiple Imputation Regression Analysis
mira "regression command" using filename [, m(#) var(string)]
mira computes Rubin's (1987) measures for Multiple Imputation (MI) regression analysis using numbered datasets. For example, pretend that you have the following datasets generated by some MI procedure: imp1.dta, imp2.dta, etc. Then mira runs the regression command in each dataset up to dataset m (default is 5) and analyzes the results using Rubin's formulas.
In case that all the MIs are included in one dataset, var option should be used to indicate the name of variable that describes the number of imputation.
Finally, mira reports the arithmetic (simple) and geometric r2's, along with the number of groups, and sample periods, whenever the information is available.
. mira "reg y x1 x2" using imp
. mira "reg y x1 x2" using imp, m(10)
. mira "reg y x1 x2" using imp, m(5) var(impnum)
Rodrigo A. Alfaro, Central Bank of Chile email@example.com
Rubin, D.B. (1987) Multiple Imputation for Nonresponse in Surveys. J. Wiley & Sons, New York.