Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: Re-using multiple imputation data

From   Mosi Ifatunji <>
Subject   Re: st: Re-using multiple imputation data
Date   Sat, 7 May 2011 17:00:12 -0400


Problem solved.

Thanks! :-)


On May 7, 2011, at 2:17 PM, daniel klein wrote:

"[...] to generate 20 datasets in the wide format"

First of all, there is a difference between -wide- (as used in -xt-,
-reshape- or the like context) and -mi wide- (see -help mi set-, help
mi style-, -help mi convert-).

Second, if you have more than one dataset after imputing missing
values your dateset  is not in -wide- format, but -flongsep- format.
That's fine, but if you want -mi wide- format (that is: one dataset)
-mi set- your data -wide-. You may switch between mi formarts easily
with -mi convert-.

However, no matter what -mi style- your data is, there is no need for
a special command to load it. -use- will do the job, thanks to
characteristics (for more start here: -help char-) saved along with
the dataset.

Try the following example

(change to a new directory so it is easy to delete the hole thing
after the example)

. clear all
. sysuse auto.dta
. mi set flongsep some_name
. mi register imputed rep78
. mi impute reg rep78 price mpg weight ,add(5)
. save some_name.dta ,replace
. dir
. clear all
. use some_name.dta
. mi est : reg price mpg weight rep78

*   For searches and help try:

*   For searches and help try:

© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index