Natural Experiments

“Sometimes a rule, border, or event assigns exposure without researchers controlling the allocation.”

01 / PLAINLY

What it means, plainly

A natural experiment uses a difference created by an outside rule, border, or event rather than by the research team. That variation enables comparison but does not guarantee equivalent groups.

02 / CONTEXT

A little more

A natural experiment uses external variation that creates exposed and unexposed groups. It can strengthen causal inference when allocation is credible, but remains vulnerable to bias and concurrent changes.

03 / WHY IT MATTERS

Why it matters

It offers a way to study effects when no designed experiment occurred, provided the allocation process is examined closely.

04 / EXAMPLE

A familiar example

A policy begins in one region before another, creating a time comparison the research team did not design.

05 / LIMIT

What it does not mean

A natural event does not make assignment random or guarantee comparable groups.

06 / NOTICE

Notice it in your day

Imagine a policy beginning earlier in one region. Make two lists: why the timing could support comparison, and which simultaneous changes might confuse it.

FINAL NOTE

The idea worth keeping

Investigate why units were exposed and what else changed at the same time.

QUESTIONS / 02

Questions people still have

What creates variation in a natural experiment?

An outside rule, border, or event rather than controlled assignment by the research team.

Does ‘natural’ mean randomly assigned?

No. Comparability and possible biases still require investigation.