Doing it for the first time, I used the MI set command and I performed multiple Imputation on my data set. Then I tried to remove the MI set by deleting the new variables and imputed datasets. I didn't know the stata command for doing this.
Now my data has MI set but the variable _mi_m which I need to erase this setting has been deleted. Is there still any other way to clear the MI set setting? I tried adding a new _mi_m and performing erase but it didn't work that way.
Thanks in advanced for your guidance.
Regards,
Mehdi
Dear Mehdi,
I don't actually have a great answer for you other than to refer this to STATA. However, practically I would export and then reimport the file to avoid dealing with it. Saving the file in .dta will keep the messed up setting. Exporting to .csv or whichever will generate a new .dta with no such setting when you import.
Dear Bruce,
Thank you very much for your help. I exported and imported the data to a new stata file and the setting was gone!
Best regards,
Mehdi
also keep a master file of the data, and use do-files to manage and analyse will help with unwanted stuff happening. Just rerun with new settings and not save before you have a final analysis dataset.
Dear Kristian,
Thank you very much for your valuable advices. I am doing the same way as you described. It was just some complicated data cleaning which I performed on the last version of the data and I didnt adapt the do file for them yet.
Dear Mehdi,
If I know, there is a long time from your question, but you can use the following command after MI:
mi extract 0, clear
Best,
Well Shahab, Mehdi and Suleiman thanks for bringing this up.
For me Shahab's recommendation of "mi extract 0, clear" didn't work immediately.
I first had to re-generate _mi_miss and -mi_id which i had inadvertently deleted earlier. For some reason, STATA did not insist on the regeneration of the _mi_m variable.
I regenerated them as follows:
gen _mi_miss=0
gen _mi_id=0
Only then did I apply:
mi extract 0, clear
It worked.
Hope all the above suggestions stops people from panicking.
Thanks Danieal Epeh so much. I got the same problem and it worked well as per your command.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Regression