1
$\begingroup$

I'm using the CBPS package in R to calculate covariate balancing propensity scores following Imai and Ratkovic 2014. However, I'm a bit confused. The scores that I get from the CBPS package -- are those both weights and propensity scores or just weights? Do I use those scores to weight my covariates only or do I match on closest scores and then weight my covariates?

$\endgroup$
1
$\begingroup$

The CBPS object that is the output of a call to CBPS() contains both $fitted.values, which are the estimated propensity scores, and $weights, which are the estimated weights. You can match on the estimated propensity scores or perform some kind of weighting (e.g., IPW for the ATE, weighting by the odds for the ATT) using the propensity scores or the weights. In theory, calculating weights from the estimated propensity scores yourself should yield the same weights that are supplied by CBPS() using the corresponding formula, but I've encountered some issues where this is not true, and so I trust the estimated propensity scores more.

Glad to see this package being used! I find it vastly outperforms twang and is much faster. If you want to check balance on your covariates, I recommend the cobalt package, which has methods for CBPS objects.

$\endgroup$
  • $\begingroup$ Thanks for the reply. When doing IPW, do you use the weights ($weights) or propensity scores ($fitted.values)? $\endgroup$ – user3210369 May 16 '17 at 22:48
  • $\begingroup$ It really shouldn't matter. If you're estimating the ATT, try with(cbps.out, all.equal(weights, ifelse(treat == 1, 1, fitted.values/(1-fitted.vales)). If TRUE, it doesn't matter whether you use the produced weights or manually create weights with the propensity scores. (cbps.out is the output of a call to CBPS(), and treat is your treatment variable.) $\endgroup$ – Noah May 17 '17 at 19:18
1
$\begingroup$

You should decide first if you want to use matching or weighting. the command of $fiited.values will provide you the propensity scores but you can also get weights by using $weights. However, you should keep in mind that you cannot get ATT in case of multiple treatments. CBPS package currently only provides ATE for multiple treatment and ATT is only available in case of binary treatments.

$\endgroup$
  • 1
    $\begingroup$ Before that, figure out why covariate adjustment is not adequate, ignoring the propensity score. $\endgroup$ – Frank Harrell Jun 16 '17 at 11:21

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.