I'd like some help to perform multiple imputation on my database. I
have four categorical variables (yes/no): malepart, d10y_cp, e10y_np,
d6msex_male and three continuous variables: numcp10y3, numnp10y2,
elife_lp. All of them have around 20% missing values. I read in some
papers that the best option for imputation would be the multiple
imputation. I found the Stata's command for MI but I'm not sure
whether I've been running it correctly. I followed all instructions on
the Stata library, but something is still wrong. I typed:
"ice d10y_cp e10y_np numcp10y3 numnp10y2 elife_lp d6msex_male malepart
cp10y1 cp10y2 np10y1 np10y2 male6m1 male6m2 malever1 malever2 using
impnew, m(5) passive(cp10y1:d10y_cp==1 \cp10y2:d10y_cp==2
\np10y1:e10y_np==1 \np10y2:e10y_np==2 \male6m1:d6msex_male==1
\male6m2:d6msex_male==2 \malever1:malepart==1 \malever2:malepart==2)
sub(d10y_cp: cp10y1 cp10y2, e10y_np: np10y1 np10y2, d6msex_male:
male6m1 male6m2, malepart: malever1 malever2) replace"
Does this command make sense?
The Stata's output was:
"Error running -uvis- I detected a problem with running uvis with
command logit on response e10y_np
and covariates numcp10y3 numnp10y2 elife_lp cp10y1 cp10y2 male6m1
male6m2 malever1 malever2 r(198)"
I have no idea which error was exactly detected. Could you give me a
clue to what I have been doing wrongly?
Thanks in advance.
Regards,
Adriana
*
* 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/
