Dear Statalist,
I am currently using Stata 16 to perform multiple imputation using the mi impute chain command.
For binary variables I have been using the logit function. However, when entering several binary variables, I always seem to have an issue with convergence.
Is using predictive mean matching (i.e. mi impute chain (pmm)) a suitable/reasonable alternative?
In the Stata manual is states the pmm is for continuous variables. However, I have seen some state that pmm may also be used for binary variables too (although I could find very little information about this topic on the internet)?
Thanks in advance for your help,
James
I am currently using Stata 16 to perform multiple imputation using the mi impute chain command.
For binary variables I have been using the logit function. However, when entering several binary variables, I always seem to have an issue with convergence.
Is using predictive mean matching (i.e. mi impute chain (pmm)) a suitable/reasonable alternative?
In the Stata manual is states the pmm is for continuous variables. However, I have seen some state that pmm may also be used for binary variables too (although I could find very little information about this topic on the internet)?
Thanks in advance for your help,
James
