Research map / Probabilistic and scientific modeling / Treatment effects
Causal Estimation for Text Data with (Apparent) Overlap Violations
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 changed | Where | Measured | Effect |
|---|---|---|---|
| Use the TI-estimator instead of outcome-only estimation | CDE estimation · Amazon Reviews semi-synthetic data | average absolute bias: 0.376 -> 0.059; 0.076 -> 0.114 | Mixed |
| Use the TI-estimator instead of outcome-only estimation | CDE confidence interval estimation · Amazon Reviews semi-synthetic data | 95% CI coverage: 0% -> 84%; 0% -> 87% | Helped |
| Use TI-estimation rather than the unadjusted estimator | CDE estimation · CFPB consumer complaints | CDE: -0.038 -> 0.200 | Mixed |
| Use GPR with Dot Product+White Noise rather than GPR with RBF for the propensity score (ablation) | TI-estimator · Amazon Reviews semi-synthetic data | 95% CI coverage: 84% vs 0% | Helped |
| Use KNN rather than GPR with Dot Product+White Noise for the propensity score (ablation) | TI-estimator · Amazon Reviews semi-synthetic data | 95% CI coverage: 84% vs 0% | Hurt |
| Use AdaBoost rather than GPR with Dot Product+White Noise for the propensity score (ablation) | TI-estimator · Amazon Reviews semi-synthetic data | 95% CI coverage: 84% vs 0% | Hurt |
| Use logistic regression rather than GPR with Dot Product+White Noise for the propensity score (ablation) | TI-estimator · Amazon Reviews semi-synthetic data | 95% CI coverage: 84% vs 84% | No clear effect |
| Use ATT AIPTW rather than ATT BMM for double machine learning (ablation) | TI-estimator · Amazon Reviews semi-synthetic data | 95% CI coverage: 84% vs 4% | Helped |
| Use ATT AIPTW rather than ATT TMLE for double machine learning (ablation) | TI-estimator · Amazon Reviews semi-synthetic data | 95% CI coverage: 84% vs 22%; 57% vs 75% | Mixed |
| Improve conditional outcome model accuracy (ablation) | TI-estimator · Amazon Reviews semi-synthetic data | 95% CI coverage: 95% | Helped |
Closest papers by meaning
- Proximal Causal Inference With Text Data (NeurIPS 2024 · similarity 0.88)
- End-To-End Causal Effect Estimation from Unstructured Natural Language Data (NeurIPS 2024 · similarity 0.88)
- Data Fusion for Partial Identification of Causal Effects (NeurIPS 2025 · similarity 0.87)
- Doubly robust identification of treatment effects from multiple environments (ICLR 2025 · similarity 0.86)
- An Analysis of Causal Effect Estimation using Outcome Invariant Data Augmentation (NeurIPS 2025 · similarity 0.85)
- Conditional Instrumental Variable Regression with Representation Learning for Causal Inference (ICLR 2024 · similarity 0.85)
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.