Natural Experiment

A design that borrows the RCT's identification logic from a real-world process — a policy, boundary, or lottery — judged plausibly as-good-as-random, where the as-if-random assumption must be substantively defended rather than guaranteed by protocol.

Core Idea

A natural experiment is a design in which assignment to treatment and control is determined by a real-world process — a policy change, an administrative or geographic boundary, a lottery, a biological coincidence — that the researcher judges plausibly as-good-as-random with respect to the outcome, though the researcher did not control it. The identification logic is the RCT's: if assignment is uncorrelated with potential outcomes, post-treatment differences are attributable to treatment. Its departure from an RCT is that the as-if-random assumption must be substantively defended from the generating process, not guaranteed by protocol. Estimates are typically local.

Scope of Application

The natural experiment is the identification workhorse of empirical causal inference; its reach is across the substantive subfields sharing that one methodological home — the method travels intact, the literature changes.

• Labour economics — the founding turf; minimum-wage boundaries, draft lotteries, immigration shocks.
• Epidemiology and public health — Mendelian randomization; the lineage back to Snow's Broad Street pump.
• Political science — close-election regression discontinuity, random ballot order, term-limit cutoffs.
• Education research — enrolment lotteries, grade-retention and class-size cutoffs.
• Development and environmental economics — rainfall instruments, protected-area boundary discontinuities.

Clarity

Naming the natural experiment dissolves the false dichotomy that credible causal claims require researcher-controlled randomisation. It relocates the design's central assumption — as-if-randomness — from a hidden premise to an explicit, defensible target, and sharpens the distinction between internal validity (the assumption holding for whoever the variation moved) and external validity (whom the estimate generalises to), keeping the local scope in view.

Manages Complexity

Empirical causal inference is a per-setting sprawl of different literatures, confounding structures, and estimators. The category collapses that onto one recurring question — where did treatment get assigned by a process unrelated to the outcome? — so the varied technique families become surface variants of one identification move, and the credibility of a claim reduces to a short, decidable checklist.

Abstract Reasoning

It supports an identification audit run in reverse (scan for where the world already randomized), the load-bearing move of articulating and defending the as-if-random assumption with a diagnostic enumeration of threats, a boundary-drawing move keeping internal and external validity apart, and a verdict-and-robustness move (placebo tests, donor-pool sensitivity, multiple strategies) that stress-tests the found randomization.

Knowledge Transfer

A natural experiment is a research method, not a structural pattern in the world. Within empirical causal inference it transfers as method, intact and exceptionally wide, across economics, epidemiology, political science, and more — the same activity applied to different literatures, sharing vocabulary and technique families. The wide spread is across substantive fields within one methodological home, not to a different substrate. The deeper cross-domain reach belongs to its constituents, randomization and causal_inference; borrowing the phrase outside empirical research is analogy.

Relationships to Other Abstractions

Current abstraction Natural Experiment Domain-specific

Parents (2) — more general patterns this builds on

• Natural Experiment is a kind of Causal Inference Domain-specific

  A Natural Experiment is Causal Inference specialized to found as-if-random assignment, a substantively defended identifying process, and a local effect.

• Natural Experiment is part of, conditional Randomization Prime

  A natural experiment contains randomization when its found assignment process is literally stochastic with known allocation probabilities; merely as-if-random policy and boundary branches are excluded.

Children (3) — more specific cases that build on this

• Difference-in-Differences Domain-specific is a kind of Natural Experiment

  Difference-in-differences is a natural experiment specialized to found treatment variation across groups and time whose causal warrant is the defended parallel-trends assumption.

• Instrumental variable Domain-specific is a kind of, typical Natural Experiment

  A found instrumental-variable design is a natural experiment whose real- world exogenous variation shifts treatment and reaches outcome only through it.

• Regression Discontinuity Design Domain-specific is a kind of Natural Experiment

  Regression discontinuity design is a natural experiment specialized to found as-if-random assignment at a sharp cutoff on a continuous running variable.

Hierarchy paths (12) — routes to 8 parentless roots

• Natural Experiment → Causal Inference → Statistical Inference → Inductive Reasoning

Show alternative paths (11)

Neighborhood in Abstraction Space

Natural Experiment sits in a sparse region of the domain-specific corpus (65^th percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.

Family — Unclustered & Miscellaneous (309 abstractions)

Nearest neighbors

• Instrumental variable — 0.86
• Selection on Observables — 0.85
• Gambler's Fallacy — 0.84
• Difference-in-Differences — 0.83
• Endogeneity — 0.83

Computed from structural-signature embeddings · 2026-07-12