Causal Inference without Control Units
Episode details
Episode description
Randomized experiments are the gold standard for causal claims, yet randomization is not feasible or ethical for many questions in the social sciences. Researchers have thus devised methods that approximate experiments with observational data, usually via regression or matching. These techniques use nonexperimental control units to estimate counterfactuals. However, control units may be costly to obtain, incomparable to the treated units, or completely unavailable when all units are treated. We explore the possibility of making causal inferences without control units by using front-door and front-door difference-in-differences estimators that make use of post-treatment information. We illustrate the viability of this approach using multiple applications from the social sciences, including the evaluation of a job training program and the evaluation of an early voting program. Our statistical contributions have important implications for both experimental and observational research design and analysis.
Watch also from the same channel
More episodes from this channel
You might be interested in
