@drgothmania See also Nguyen et al. (2017) [https://t.co/FwvwooeGi0] who find that regression after matching doesn't reduce bias when the covariate SMDs are small, which is consistent with theory. If your SMDs aren't small, then you should try other matchi
3,686 followers
3,686 followers
@ehsan7x It's generally best to include all, but when you can't, those rules can be good heuristics. 1. Nguyen et al (2017) <https://t.co/FwvwoowhGA> find little improvement in bias when adjusting for covariates with <.1 SMD. 1/2
6,069 followers
"If covariate balance is not achieved [by PSM], we recommend reiterating PS modeling until standardized differences below 0.10 are achieved on most covariates. In case of remaining imbalance, a double adjustment might be worth considering." https://t.co/B
5,360 followers
Looking for a lunchtime read? - Check out this fantastic paper! https://t.co/C1LXGfVfPb