Sensitivity Analysis for Causal Inference
No observational study can prove that unobserved confounding is absent. Sensitivity analysis asks: how large would unobserved confounding need to be to overturn the estimated causal effect? This shifts the question from "is there confounding?" (unanswerable) to "how much confounding would be needed?" (quantifiable and interpretable).
Where it sits
left: what it builds on · right: what builds on it · pale: exampleContributed by E2ER contributors
How studies reach it
- No Cash Flows, New Owners: Spot Bitco… → Sensitivity Analysis for Causal Infer…
- Digital credit scores and lending to … → Empirical → identification_reviewer → Sensitivity Analysis for Causal Infer…Example
- Stablecoins and the cost of remittanc… → Empirical → identification_reviewer → Sensitivity Analysis for Causal Infer…Example
- Remote work and promotion chances → Empirical → identification_reviewer → Sensitivity Analysis for Causal Infer…Example
- Mobile money and household savings in… → Staggered difference-in-differences (… → Empirical → identification_reviewer → Sensitivity Analysis for Causal Infer…Example
- Data-sharing mandates and hospital IT… → Staggered difference-in-differences (… → Empirical → identification_reviewer → Sensitivity Analysis for Causal Infer…Example
and 13 more
Disciplines it reaches
Solid: published studies. Light: examples.
Computed from the records on this site: what each study, template and specialist names as used, which study extends which, and who contributed what. 19 studies in total.
Provenance
- Project
- E2ER bundled skills
- Maintained by
- E2ER contributors
- Licence
- MIT
- Source file
- bhanneke/E2ER-project/blob/orgsci-mvp/skills/files/causal-inference/sensitivity.md
- Identifier
- e2er/causal-inference/sensitivity
Use it
Ships with every installation of E2ER.
Used by
Evaluations
PlannedNo evaluation record yet. An evaluation shows where a skill breaks before others rely on it.
Evaluate this skillDescription
Data model- Discipline
- Economics
- Method family
- Empirical (quantitative)
- Design
- not specified
- Research stage
- Research design
- Contributors
- E2ER contributors (Software, Methodology)
- Usage
- used in 2 templates · 1 published study · loaded by 2 specialists · Examples: 3 example studies
- Source
- E2ER repository · skills/files · @3b91f0e
- Record
- skill:e2er/causal-inference/sensitivity · JSON
Solid tags are declared by the source or mapped from its terms; dashed tags are inferred by a published rule. Hover a tag for its provenance.