Closed feliuserra closed 3 years ago
Hi, this issue is caused by two few treated units. I have "fixed" this issue by using an approximation based on a normal distribution with bootstrapped SE instead of using the percentile method to obtain CIs. The updates are reflected in v.1.1.7.
Thanks for the suggestions!
Thank you! I have already tried the v.1.1.7
and now the confidence intervals are symmetric. Great job!
You're welcome.
On Fri, Nov 20, 2020 at 1:25 AM Feliu Serra Burriel < notifications@github.com> wrote:
Thank you! I have already tried the v.1.1.7 and now the confidence intervals are symmetric. Great job!
— You are receiving this because you modified the open/close state. Reply to this email directly, view it on GitHub https://github.com/xuyiqing/gsynth/issues/49#issuecomment-731052207, or unsubscribe https://github.com/notifications/unsubscribe-auth/AB2PKGG3QWX5FQJW4VVIT2TSQYYZHANCNFSM4S7F75OA .
-- Yiqing Xu
Assistant Professor Department of Political Science Stanford University http://yiqingxu.org/
Hi,
Usually when computing the Confidence intervals around the ATT, the bands shoud be "estimated_effect +- 2*se_bootstrap" and therefore these bands are symmetric around the estimated effects. However, even in some of your examples from: https://yiqingxu.org/software/gsynth/gsynth_examples.html do not have symmetric confidence intervals (I am looking at the MC example).
Could you please provide some help or further explanation on why this happens?
Thanks