Hi All,

        Do we need to use weights and adjust for all the covariates used in CEM again to avoid possible residual confounding? There is no need for adjust for the covariates if the CEM makes perfect/Multivariate L1->0 match?

        Thanks!

Fang