Right now, the package produces ATTs by treatment time (e.g., t+1, t+2). Is it possible to return a single estimate of the ATT pooled across treatment time? I imagine this would have to take into account a) differing number of treated units at different treatment times and b) within-unit covariance over time.
Right now, the package produces ATTs by treatment time (e.g., t+1, t+2). Is it possible to return a single estimate of the ATT pooled across treatment time? I imagine this would have to take into account a) differing number of treated units at different treatment times and b) within-unit covariance over time.