Declare datasets to be multiple imputed datasets
miset [using filename-prefix] [, mimps(#) clear]
Description
miset creates temporary copies of imputed datasets so that subsequent mi (multiple imputation) commands can be executed on these data. The other mi commands can only be used after the multiple datasets have been declared by miset. The imputed datasets are assumed to be created from an original dataset by a "proper" imputation method as in Schafer (1997). These datasets must have the same variables and the same number of observations.
The following naming convention must be used: the filenames of the imputed datasets must consist of a common word followed by a consecutive number and followed by the normal extension .dta; for example, foo1.dta, foo2.dta, ..., foo5.dta. Only the common word or filename-prefix ("foo" in this case) is to be specified after the using command.
miset creates temporary files mitemp1.dta, mitemp2.dta, and so on by copying the using files. All subsequent commands will be executed on these temporary files, with the original using files left unchanged. The temporary files remain in the working directory until mireset is issued. To save the updated temporary files after a series of mi commands, see misave.
Options
mimps(#) specifies the number, m, of datasets to be used. The default is 5. A minimum of 2 and a maximum of 9 datasets can be specified. If there exist more than m datasets, Only the first m datasets are used. As mentioned earlier, usually as few as 3 or 5 multiple imputed datasets is sufficient.
clear permits the data to be loaded even if there is a dataset already in memory and even if that dataset has changed since the data were last saved.
Remarks
miset is one of the suite of mi commands. The other mi commands can only be used after the multiple datasets have been declared by miset.
Examples
. miset using mydata . miset using mydata, mimps(3)
Authors
Ning Li, Philip Greenwood, and John Carlin, Clinical Epidemiology & Biostatistics Unit, Murdoch Children's Research Institute and University of Melbourne. Email jbcarlin@unimelb.edu.au
References
Rubin, D. B. 1987. Multiple Imputation for Nonresponse in Surveys. New York: John Wiley & Sons.
Schafer, J. L. 1997. Analysis of Incomplete Multivariate Data. London: Chapman & Hall.
Also see
Online: help for miappend, mimerge, misave, mido, mici, mifit, milincom, mireset