Re: st: mi estimate with wide data - "no imputations" error

[Vanessa Harbin <vharbin@childtrends.org>]

Thanks Daniel. I tried setting M by using char _dta[_mi_M] 5 and it now will run the mi estimate command, however, it seems to be re-creating the imputation rather than using the imputed values I set in mi register. I get this message:

(44280 values of imputed variable plev in m>0 updated to match values in m=0) 
(variable _1_plev in m=1 recreated using m=0 because not found) 
(variable _1_plev in m=2 recreated using m=0 because not found) 
(variable _1_plev in m=3 recreated using m=0 because not found) 
(variable _1_plev in m=4 recreated using m=0 because not found) 
(variable _1_plev in m=5 recreated using m=0 because not found) 
(variable _2_plev in m=1 recreated using m=0 because not found) 
(variable _2_plev in m=2 recreated using m=0 because not found) 
(variable _2_plev in m=3 recreated using m=0 because not found) 
(variable _2_plev in m=4 recreated using m=0 because not found) 
(variable _2_plev in m=5 recreated using m=0 because not found) 
(variable _3_plev in m=1 recreated using m=0 because not found)
(variable _3_plev in m=2 recreated using m=0 because not found) 
(variable _3_plev in m=3 recreated using m=0 because not found) 
(variable _3_plev in m=4 recreated using m=0 because not found) 
(variable _3_plev in m=5 recreated using m=0 because not found) 
(variable _4_plev in m=1 recreated using m=0 because not found) 
(variable _4_plev in m=2 recreated using m=0 because not found) 
(variable _4_plev in m=3 recreated using m=0 because not found) 
(variable _4_plev in m=4 recreated using m=0 because not found) 
(variable _4_plev in m=5 recreated using m=0 because not found) 
(variable _5_plev in m=1 recreated using m=0 because not found) 
(variable _5_plev in m=2 recreated using m=0 because not found) 
(variable _5_plev in m=3 recreated using m=0 because not found) 
(variable _5_plev in m=4 recreated using m=0 because not found) 
(variable _5_plev in m=5 recreated using m=0 because not found)

Perhaps this is because there is no variable for M to indicate the number of the imputation? I thought mi set wide and mi register imputed would eliminate the need for that variable. 

Any other thoughts? 

Thanks!!

Vanessa


*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/