On 2004-11-19, at 11.37, chris wallace wrote:
Quite possibly the imputation would be not much different if I used 15
rather
than 17 predictor variables; I will investigate using regress. But as
all
these biochemical measures relate to the same biochemical pathway, I
am not
clear how to select "the best" 15 out of 17.
You might want to try multiple imputation (MI) instead. MI has better
statistical properties than "single" imputation because it recognises
the uncertainty in the imputed values by making multiple imputations in
multiple datasets. After imputation you have to run your analyses on
each of the datasets and combine the result into a single estimate. MI
has been made available in Stata by Patrick Royston's -mice- package
(st0067) and routines for estimating models and combining results are
available within the -mice- package as well as in a package by Li,
Greenwood & Carlin (st0042).
. ssc install mice
. whelp mvis
. whelp micombine
. whelp misplit
. net install st0042.pkg
. whelp miset
. whelp mifit
------------------------------------------------
Michael Ingre , PhD student & Research Associate
Department of Psychology, Stockholm University &
National Institute for Psychosocial Medicine IPM
Box 230, 171 77 Stockholm, Sweden
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/