Multiple Imputation - How to compute restrictions for the imputation?

[Mona Wald]

Dear all,

I think my problem should be easy to solve, but I just cannot figure it out, so I hope someone here might help me.

I am applying multiple imputation on a variable whose values can only range from 0 to 0.99. I just cannot figure out where to and how to compute that restriction into my "mi impute regress depvar indepvar" command.

Would be great if someone could help me!

Thank you and best wishes
Mona

[daniel klein]

Use pmm instead of regress; pmm is a regression-based approach but instead of using the predicted values from the regression model directly it chooses one of the k observed values that are closest to the predicted values.

Best
Daniel