AI Paper Atlas

Research map / Probabilistic and scientific modeling / Treatment effects

Falsification before Extrapolation in Causal Effect Estimation

Zeshan Hussain, Michael Oberst, Ming-Chieh Shih, David Sontag

NeurIPS 2022 · Probabilistic and scientific modeling › Treatment effects · approach: Latent-variable causal inference · 1 citations (OpenAlex)

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
Falsify studies against RCT validation effects, then conservatively take the union of accepted-study confidence intervalsObservational effect estimation on WHI held-out subgroups · WHI held-out subgroupsCoverage: 0.39 -> 0.45Improved
Falsify studies against RCT validation effects, then conservatively take the union of accepted-study confidence intervalsObservational effect estimation on WHI held-out subgroups · WHI held-out subgroupsLength: 0.416 -> 0.081Improved
Exclude falsification and meta-analyze all observational studiesObservational effect estimation on WHI held-out subgroups · WHI held-out subgroupsCoverage: 0.45 vs. 0.03Hurt
After falsification, meta-analyze the remaining studies instead of conservatively taking their interval unionExPCS · WHI held-out subgroupsCoverage: 0.45 vs. 0.28Hurt
Select only the original unbiased observational studyExPCS · WHI held-out subgroupsCoverage: 0.45 vs. 0.44No clear effect
Select only the original unbiased observational studyExPCS · WHI held-out subgroupsLength: 0.081 vs. 0.068Hurt
Use falsification followed by conservative interval combination rather than random-effects meta-analysisObservational effect estimation on semi-synthetic IHDP data · IHDP extrapolated subgroupsCoverageImproved
Use ExPCS rather than ExOCS when observational sample sizes are smallObservational effect estimation on semi-synthetic IHDP data · IHDP extrapolated subgroupsCoverageImproved
Use ExPCS rather than Simple Union when falsification has adequate powerObservational effect estimation on semi-synthetic IHDP data · IHDP extrapolated subgroupsLengthImproved

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