Why the ate_interval produces same result for different alpha values?

[ukm5]

Also, to what alpha do the confidence intervals in ate_inference summary correspond to? I was trying to determine the latter using the ate_interval function when I noticed that the intervals returned are independent of alpha passed into the function.

[vsyrgkanis]

I believe this relates to an issue that was recently fixed and merged #476

You could either install from the master branch of the repo or wait for the next release.

[ukm5]

Thanks! So just to confirm, the current intervals outputted for CATE and ATE are at 90% confidence? Also, how do I pass an alpha when using the effect_inference function?
Finally, is there an ETA on the next release? Since I work on my company maintained cloud of libraries, I am unsure of how to install from the master branch, but anyways thanks for the info!

[vsyrgkanis]

yeap!
you call: effect_inference(X).summary_frame(alpha=..)
or effect_inference(X).conf_int(alpha=...)
release should happen by Friday.