Since neither program is perfect, I want to use both. PSMATCH2 for drawing the PS distributions and testing for covariate balance, and TEFFECTS PSMATCH for getting the correct standard errors. If I put LOGIT and TIES on PSMATCH2. I replicate the TEFFECTS PSMATCH results. However, I usually use the option COMMON on PSMATCH2 and when I do this, the answers differ. I cannot figure out how to impose common support in TEFFECTS PSMATCH2. If I try to get to it indirectly via CALIPER, the program just quits if there isn't a match inside the stated interval. Either I'm missing something, or TEFFECTS PSMATCH2 is messed up. Any help or advice would be appreciated.
cheers
Kevin
cheers
Kevin
