Causal Inference — Reproducibility

Direct answer

Causal Inference — Reproducibility is implemented as a source-bounded reproducibility guide.

Answer contract

Apply the reproducibility lens only to the exact inspected source scope; do not infer authority from adjacent topics.

Evidence and exact locators

t22-randomized-contrast — lib/federation/readiness-tranche-22.ts — randomizedMeanContrast. Supports: Computes the exact sample-mean contrast for two declared randomized synthetic arms.

src-causal-what-if — Chapter 1, p. 12, “A definition of causal effect”. Supports: Defines causal effects as contrasts under alternative treatments in a target population and distinguishes causal from associational quantities.

What the evidence does not establish

Randomization is caller-declared; this fixture does not verify assignment, estimate population effects, handle attrition, or justify transportability.

Does not establish identification for a particular dataset, the validity of a causal model, or an effect estimate.

Rights and reuse

project-owned-reference-only

reference-only

Dependencies and related concepts

applies-to: urn:maha:concept:computation:causal-inference

governed-by: urn:maha:concept:governance

evidence-for: urn:maha:concept:computation

Authority or implementation

This authority or implementation section is constrained to the same inspected scope: Computes the exact sample-mean contrast for two declared randomized synthetic arms. Defines causal effects as contrasts under alternative treatments in a target population and distinguishes causal from associational quantities.

It must preserve the recorded boundary: Randomization is caller-declared; this fixture does not verify assignment, estimate population effects, handle attrition, or justify transportability. Does not establish identification for a particular dataset, the validity of a causal model, or an effect estimate.

Method or mechanism

This method or mechanism section is constrained to the same inspected scope: Computes the exact sample-mean contrast for two declared randomized synthetic arms. Defines causal effects as contrasts under alternative treatments in a target population and distinguishes causal from associational quantities.

It must preserve the recorded boundary: Randomization is caller-declared; this fixture does not verify assignment, estimate population effects, handle attrition, or justify transportability. Does not establish identification for a particular dataset, the validity of a causal model, or an effect estimate.

Verification and uncertainty

This verification and uncertainty section is constrained to the same inspected scope: Computes the exact sample-mean contrast for two declared randomized synthetic arms. Defines causal effects as contrasts under alternative treatments in a target population and distinguishes causal from associational quantities.

It must preserve the recorded boundary: Randomization is caller-declared; this fixture does not verify assignment, estimate population effects, handle attrition, or justify transportability. Does not establish identification for a particular dataset, the validity of a causal model, or an effect estimate.

What this does not establish

This what this does not establish section is constrained to the same inspected scope: Computes the exact sample-mean contrast for two declared randomized synthetic arms. Defines causal effects as contrasts under alternative treatments in a target population and distinguishes causal from associational quantities.

It must preserve the recorded boundary: Randomization is caller-declared; this fixture does not verify assignment, estimate population effects, handle attrition, or justify transportability. Does not establish identification for a particular dataset, the validity of a causal model, or an effect estimate.

Dependencies

This dependencies section is constrained to the same inspected scope: Computes the exact sample-mean contrast for two declared randomized synthetic arms. Defines causal effects as contrasts under alternative treatments in a target population and distinguishes causal from associational quantities.

It must preserve the recorded boundary: Randomization is caller-declared; this fixture does not verify assignment, estimate population effects, handle attrition, or justify transportability. Does not establish identification for a particular dataset, the validity of a causal model, or an effect estimate.

Questions this page can answer