When Natural Experiments Are Neither Natural Nor Experiments : Lessons from the Use of Redistricting to Estimate the Personal Vote

@inproceedings{Sekhon2010WhenNE,
  title={When Natural Experiments Are Neither Natural Nor Experiments : Lessons from the Use of Redistricting to Estimate the Personal Vote},
  author={Jasjeet S. Sekhon and Rocı́o Titiunik},
  year={2010}
}
Natural experiments offer a path to overcome the severe obstacles to causal inference in the social sciences, and for that reason they have become increasingly prominent. But in the quest for as-if random interventions, it is all too easy to overlook crucial methodological issues. Even if the intervention is actually random, some of the treatment–control comparisons made available by the natural experiment may not be valid, and those comparisons that are valid may not be directly related to the… CONTINUE READING