search

ronikobrosly / causal-curve

A python package with tools to perform causal inference using observational data when the treatment of interest is continuous.
MIT License
271 stars 18 forks source link

balanced correlations #44

Closed juandavidgutier closed 2 years ago

juandavidgutier commented 2 years ago

Hello @ronikobrosly

I am new with Generalized Propensity Score (GPS), and I am reproducing the script of the example named NHANES_BLL_example.ipynb, of the module causal-curve. My question is how I can get the adjusted (balanced) correlations between the treatment (Blood lead levels) and the potential confounders in the example (Age, Sex_Male, Race_NH_Black, Race_NH_White, Race_Other, Edu_HS, Edu_LT_HS, Smoke_Home_Yes, Baby_NICU_Yes, Food_Often_bad, Food_Sometimes_bad), or how can I evaluate the performance of covariates balance?

I will appreciate a lot your cooperation.

Regards,

ronikobrosly commented 2 years ago

Hi @juandavidgutier , thanks for reaching out. This isn't something I've implemented yet, but this is a very good idea. I did a little googling and came across this paper: https://arxiv.org/pdf/1812.06575.pdf It seems to discuss approaches to assessing covariate balance. I'm really busy these days but can eventually get to it. If you discover an approach and want to contribute here I would love the help!

juandavidgutier commented 2 years ago

Hi @ronikobrosly Thanks a lot for your answer.