Research map / Probabilistic and scientific modeling / Treatment effects
Control Function Instrumental Variable Estimation of Nonlinear Causal Effect Models
Zijian Guo, Dylan S. Small
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 changed | Where | Measured | Effect |
|---|---|---|---|
| use control function estimator instead of 2SLS when its assumptions hold | nonlinear IV regression · Satisfied Model (26) | WRMSE ratio, beta2: 0.139 | Improved |
| use control function estimator instead of 2SLS under moderately violated assumption (cubic first stage) | nonlinear IV regression · Cubic Model (27) | WRMSE ratio, beta3: 0.635 | Improved |
| use control function estimator instead of 2SLS under exponential first stage | nonlinear IV regression · Exp Model (28) | WRMSE ratio, beta3: 0.049 | Improved |
| use control function estimator instead of 2SLS under drastically violated assumptions | nonlinear IV regression · Drastically Violated Model (29) | WRMSE ratio, beta2: 6.900; bias ratio 0.128 | Hurt |
| use pretest estimator instead of control function under drastic violation | pretest estimator · Drastically Violated Model (29) | WRMSE ratio, beta2: 6.900 -> 1.000 | Improved |
| use control function estimator under non-normal double exponential errors | nonlinear IV regression · Model (30), double exponential | WRMSE ratio, beta2: 0.154 | Improved |
| use control function estimator under log-normal errors | nonlinear IV regression · Model (30), log-normal | WRMSE ratio, beta2: 0.077 | Improved |
| use control function estimator under absolute-normal errors | nonlinear IV regression · Model (30), absolute normal | WRMSE ratio, beta2: 0.964 | No clear effect |
| add pretest to control function estimator under absolute-normal errors (ablation) | pretest estimator · Model (30), absolute normal | WRMSE ratio, beta2: 0.964 -> 1.063 | Hurt |
| use control function instead of 2SLS on Burundi data | nonlinear IV regression · Burundi field experiment | variance of delta(y2) | Improved |
| use control function instead of 2SLS on Burundi data | nonlinear IV regression · Burundi field experiment | SD of beta2: 7.033 -> 3.479 | Improved |
| use control function instead of 2SLS on food demand data | nonlinear IV regression · Philippines food demand | SD of beta2: 22.353 -> 3.728 | Improved |
Closest papers by meaning
- Valid Causal Inference with (Some) Invalid Instruments (ICML 2021 · similarity 0.90)
- Data-Driven Selection of Instrumental Variables for Additive Nonlinear, Constant Effects Models (ICML 2025 · similarity 0.88)
- Conditional Instrumental Variable Regression with Representation Learning for Causal Inference (ICLR 2024 · similarity 0.87)
- Automating the Selection of Proxy Variables of Unmeasured Confounders (ICML 2024 · similarity 0.86)
- An Analysis of Causal Effect Estimation using Outcome Invariant Data Augmentation (NeurIPS 2025 · similarity 0.86)
- Instrumental Variable Estimation for Causal Inference in Longitudinal Data with Time-Dependent Latent Confounders (AAAI 2024 · similarity 0.86)
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.
