How to do n regressions for n placebos?
Dear Stata Forum,
I am evaluating a policy outcome by means of a dummy that identifies the treated and the control groups (the dummy is 1 when X > 5 of May, for instance).
I'd like to give robustness to my findings by implementing a placebo test. So far I did a build a placebo by putting the dummy in a different date and it works well (X' > 3 of January, for instance). I 'd like to run n regressions with n different placebos (one for each day different from the real treatment) and estimate the average coefficient and sd of all placebos (so I can compare placebos to Treatment).
Which is the most efficient way to do this? (and avoiding doing it manually day by day) maybe simul?
thanks
jj
I am evaluating a policy outcome by means of a dummy that identifies the treated and the control groups (the dummy is 1 when X > 5 of May, for instance).
I'd like to give robustness to my findings by implementing a placebo test. So far I did a build a placebo by putting the dummy in a different date and it works well (X' > 3 of January, for instance). I 'd like to run n regressions with n different placebos (one for each day different from the real treatment) and estimate the average coefficient and sd of all placebos (so I can compare placebos to Treatment).
Which is the most efficient way to do this? (and avoiding doing it manually day by day) maybe simul?
thanks
jj
Comment