Dear list
I would like to report standardized average treatment effects (ATEs) based on teffects ipwra. I have achieved this in linear models by standardizing the outcome variable prior to running teffects, but this is not an option if the outcome is Poisson distributed. Can I standardize the ATE by dividing by the standard deviation of the outcome variable? Or a weighted pooled standard deviation for treatment and control groups?
Barth
I would like to report standardized average treatment effects (ATEs) based on teffects ipwra. I have achieved this in linear models by standardizing the outcome variable prior to running teffects, but this is not an option if the outcome is Poisson distributed. Can I standardize the ATE by dividing by the standard deviation of the outcome variable? Or a weighted pooled standard deviation for treatment and control groups?
Barth
