search

py-why / dowhy

DoWhy is a Python library for causal inference that supports explicit modeling and testing of causal assumptions. DoWhy is based on a unified language for causal inference, combining causal graphical models and potential outcomes frameworks.
https://www.pywhy.org/dowhy
MIT License
7.01k stars 922 forks source link

Refutation tests of ATE and heterogeneous effect #956

Closed juandavidgutier closed 1 year ago

juandavidgutier commented 1 year ago

Hi @amit-sharma I want to estimate the effect of the human footprint (exposure) on malaria incidence (outcome). I developed an instrumental variable model using the DMLIV, to estimate the ATE I used the method ate. Additionally, I estimated the heterogeneous effect of the human footprint on malaria with the method effect. The value of the ATE=7.15. However, when I test the validity of the estimation with a placebo, I get a value of 243.18 with p-value=0.27. I interpret that the ATE's estimation is incorrect, but is the estimation of the heterogeneous effect also incorrect?

github-actions[bot] commented 1 year ago

This issue is stale because it has been open for 14 days with no activity.

emrekiciman commented 1 year ago

Hi @juandavidgutier , the EconML package is adding a CATE validation suite: https://github.com/py-why/EconML/pull/777 though the pull request is still under development

juandavidgutier commented 1 year ago

Hi @emrekiciman, thank for the info!!!

erikcs commented 1 year ago

Hi @juandavidgutier, if you are comfortable with R and don’t want to wait for the microsoft team, then GRF already has built-in functionality to validate heterogeneity with the “RATE”

juandavidgutier commented 1 year ago

Hi @erikcs, Thank you for the info, and it sounds nice!!! I will explore GRF.