sensitivity analysis following multiple imputation.

[Mike Golding]

Hi all,

I'm working with a dataset with a fair bit of missing data that I have addressed using MI. While the MAR assumption is reasonable for the majority of the imputed variables, I'm skeptical that it holds for the missing income values. Therefore, I'd like to run a simple sensitivity analysis using the delta adjustment method described by van Buuren and others; however, I'm relatively new to Stata and I'm not sure how to go about making the adjustment. If I understand van Buuren's description correctly, you simply adjust the imputed values by some plausible constant.

Just wondering if there is a simple way to apply an adjustment to the income variable across all of the imputed data sets, without manually adjusting each dataset one at a time.

thanks,
-Mike

[Rich Goldstein]

see

help mi passive

note however, that you might need to re-register your variable as passive if it is now registered as imputed; again, see the help file

[Mike Golding]

Great! Works perfectly. Thank you Rich for your quick and helpful response.

[Rich Goldstein]

you're welcome and thanks for the closure