AI Paper Atlas

Research map / Probabilistic and scientific modeling / Treatment effects

Naive regression requires weaker assumptions than factor models to adjust for multiple cause confounding

Justin Grimmer, Dean Knox, Brandon Stewart

JMLR 2023 · Probabilistic and scientific modeling › Treatment effects · approach: Latent-variable causal inference

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use full probabilistic-PCA deconfounder instead of naive regressionnaive regression · Medical deconfounder simulation (orig: beta1=0, beta2=0.3)RMSE beta1: 0.125 -> 0.690Hurt
use PCA+CV-Ridge deconfounder instead of naive regression when true beta1 is zeronaive regression · Medical deconfounder simulation (orig: beta1=0, beta2=0.3)RMSE beta1: 0.125 -> 0.031Improved
use PCA+CV-Ridge deconfounder instead of naive regression when true beta1 is -0.3naive regression · Medical deconfounder simulation (ours: beta1=-0.3, beta2=0.3)RMSE beta1: 0.191Hurt
switch true effect of A1 from 0 to -0.3PCA+CV-Ridge · Medical deconfounder simulation (ours: beta1=-0.3, beta2=0.3)RMSE beta1: 0.031 -> 0.191Hurt
add more treatments (m=3 to m=200) for naive regressionnaive regression · Subset simulation, beta_j = 10avg RMSE: 0.333 -> 0.011Improved
add more treatments (m=3 to m=200) for subset deconfounder with constant effectssubset deconfounder · Subset simulation, beta_j = 10avg RMSE: 10.000 -> 10.000No clear effect
add more treatments for subset deconfounder with beta_j = 100subset deconfounder · Subset simulation, beta_j = 100avg RMSE: 100.000 -> 100.000No clear effect
add more treatments for subset deconfounder with beta_j ~ N(1,2)subset deconfounder · Subset simulation, beta_j ~ N(1,2)avg RMSE: 1.465 -> 1.026Mixed
add more treatments for subset deconfounder with beta_j = 1/msubset deconfounder · Subset simulation, beta_j = 1/mavg RMSE: 0.611 -> 0.033Improved
adjust for budget in actor regressiondeconfounder · TMDB 5000 actor effectsplausibility of actor effectsImproved
use factor-model substitute confounder instead of plain regression on actorsnaive regression · TMDB 5000 actor effectsplausibility of top actor effect: 831%Hurt
use subset deconfounder for single actorsnaive regression · TMDB 5000 actor effectsplausibility of top actor effect: 1,128%Hurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy