Hi all,
I am working on a project where I have time-varying exposure and covariates, so I am using the inverse probability of treatment weight approach. Since I have quite many missing for my covariates, I run the "mi impute mvn" command and generated 20 datasets. I did impute missing values for baseline covariates, but also for one my time-varying covariate. However, when I reshape the dataset in long format I get the message from STATA that the "imputed values for my time-varying variable at time 1 and 2 unregistered because not in m=0". I used the "mi reshape long" command. Has any of you any suggestion for how to overcome this? Thank you so much in advance for your help!
I am working on a project where I have time-varying exposure and covariates, so I am using the inverse probability of treatment weight approach. Since I have quite many missing for my covariates, I run the "mi impute mvn" command and generated 20 datasets. I did impute missing values for baseline covariates, but also for one my time-varying covariate. However, when I reshape the dataset in long format I get the message from STATA that the "imputed values for my time-varying variable at time 1 and 2 unregistered because not in m=0". I used the "mi reshape long" command. Has any of you any suggestion for how to overcome this? Thank you so much in advance for your help!
