Natural experiment
A study design where real-world circumstances (like birthdays or policy) randomly divide people into groups, allowing causal inference without a designed randomized trial.
As defined in et_1003
Other definitions
A situation where an external event or policy assigns treatment in a way that resembles random assignment, allowing causal estimation.
Robert Townsend on Development, Poverty, and Financial Institutions 03/14/2011
Where it comes up · 28
- Noah Smith on Whether Economics is a Science8 claims
- Anupam Bapu Jena on Random Acts of Medicine6 claims
- 367 | Jared Diamond on the Course of History and the Role of leaders6 claims
- Adam Ozimek on the Power of Econometrics and Data4 claims
- Nature, Nurture, and Identical Twins (with David Bessis)3 claims
- 84 | Suresh Naidu on Capitalism, Monopsony, and Inequality3 claims
- Valerie Ramey on Stimulus and Multipliers 10/24/20112 claims
- Conversations with Tyler - Jared Diamond2 claims
and 20 more
Most important claims · 10 of 57Search all claims for “Natural experiment” →
- Because leaders cannot be randomly assigned or removed experimentally, social scientists must use natural experiments, comparing outcomes where leaders died naturally, were assassinated, or were absent with matched cases where they were not, which Diamond argues offers the most insight into leaders across politics, business, sports, and religion.367 | Jared Diamond on the Course of History and the Role of leaders
- Hurricane Katrina destroyed not only New Orleans' school buildings but the power of the vested interests: with no schools, no enrollments, and no money, the school board had to lay off all teachers, which left the union with virtually no members, no dues, and no power—creating, for the first time in modern American history, a reform process with the power of vested interests gone.Terry Moe on Educational Reform, Katrina, and Hidden Power
- Luria's 1931 natural experiment in Uzbekistan found that subsistence farmers reasoned only about things they had directly experienced and refused to transfer knowledge to unseen situations, while those exposed to modernity (school, vocational training, collective farming) could abstract, use formal logic, classify by shared qualities, and transfer knowledge—not better, but adapted to modern demands for knowledge transfer.David Epstein on Mastery, Specialization, and Range
- The natural experiment of UN diplomats' parking tickets — where pre-2002 diplomatic immunity meant no enforcement — showed diplomats from corrupt countries parked illegally far more than those from honest countries (culture matters), but once Mayor Bloomberg began impounding cars, even Egyptians 'parked like the Norwegians' almost overnight (enforcement matters), while the Norwegians' behavior had actually worsened over time the longer they could get away with it.189 | Brian Klaas on Power and the Temptation of Corruption
- A Danish registry study of 78,000 companies found that when a CEO's only child or spouse dies, profits drop for a year or two, with sibling deaths having a smaller effect and parent deaths smaller still, showing that CEO distraction measurably affects firm performance.367 | Jared Diamond on the Course of History and the Role of leaders
- Pre-Columbian Mexican cities, which developed independently of Eurasia and Africa, show the same superlinear scaling when assessed via archaeological surveys, serving as a natural experiment for universality, with monument construction used as a proxy for socioeconomic output because the Aztecs lacked money proper.362 | Luis Bettencourt on the Universal Properties of Self-Organizing Cities
- Dividing an ethnically, culturally, and geographically homogeneous society with a border and imposing different laws and incentives on each side creates a near natural experiment—as with Nogales AZ vs Nogales Sonora and North vs South Korea—revealing that institutions, not culture or geography, drive the stark differences in income, schooling, and health.Daron Acemoglu on Why Nations Fail
- Roughly 15 years of identification-based studies using lotteries, distance to universities, and similar instruments converge on a return to an additional year of schooling of about 9% to 12% in additional earnings, especially in the high-school-to-college range, after controlling for ability and family background.84 | Suresh Naidu on Capitalism, Monopsony, and Inequality
- Natural experiments (quasi-experiments) exploit the real-world policy-making process to achieve clean causal identification; they cannot be replicated like lab experiments since each happens only once, making them inferior to lab science, but they are substantially better than merely looking at correlations, and this 'credibility revolution' has greatly improved economists' ability to understand policy effects.Noah Smith on Whether Economics is a Science
- James Robinson taught Diamond the power of natural experiments, and Diamond adapted this approach from his bird studies in New Guinea, comparing mountains with and without a particular bird species, though his own natural experiments were cruder than those of Acemoglu and Robinson.Conversations with Tyler - Jared Diamond