Syndicate content

Add new comment

Submitted by Simon H. Heß on

Dear Timo,

It depends a bit on how you specify your regression exactly and I have to guess here because I can't see your code. But your problem seems to be that you're instructing -ritest- to permute the treatment-period-interaction---instead of the treatment-assignment---which you shouldn't do. This interaction is of course not constant within clusters. If that's the case reformulating your Stata-statement following this example should solve your problem:

ritest treatment_village _b[c.treatment_village#c.period], cluster(village): reg y c.treatment_village#c.period c.treatment_village c.period