causal-inference-in-R icon indicating copy to clipboard operation
causal-inference-in-R copied to clipboard

Published 20 hours ago • verified publisher icon malcolmbarrett

14.01: Calculating propensity scores for categorical exposures

Open malcolmbarrett opened this issue 3 years ago • 1 comments

Probably using continuous -> quintiles approach rather than finding a third example. Good chance to discuss https://pubmed.ncbi.nlm.nih.gov/24487212/

  • [ ] ATO: https://arxiv.org/pdf/1808.05339.pdf
  • [ ] https://www.sciencedirect.com/science/article/abs/pii/S0378375822000337?casa_token=WGu_Vq3gqqkAAAAA:vdjjde8G_Kq3-aikLfRCrj2ROk0SU1hlTDVbrjGjCu3km5QNuMzS-KzUXJOG7pSRH-jaP1vTenP3
  • [ ] https://pubmed.ncbi.nlm.nih.gov/33462862/

Might want to consider the differences in the distribution of bootstrapped estimates for the dnorm() vs quantile approaches

malcolmbarrett avatar Aug 10 '22 15:08 malcolmbarrett

First use ordinal, then show multinomial because ordinal (usefully) assumes dose response, but maybe it's really non-linear and multinomial is better

malcolmbarrett avatar Mar 03 '23 15:03 malcolmbarrett