I dont have any code on hand to share
And my suggestionis based on the following
If you consider only not yet treated (so drop cases where the not yet have already been treated), then there should be no effect in the estimation, because there was no treatemnt.
Thus, you give them a treatment when none occurs, you should find that there is no effect.
And my suggestionis based on the following
If you consider only not yet treated (so drop cases where the not yet have already been treated), then there should be no effect in the estimation, because there was no treatemnt.
Thus, you give them a treatment when none occurs, you should find that there is no effect.
Comment