Causal InferenceΒΆ

Estimating treatment effects is a core problem in causal inference. WhyNot provides a powerful framework that generates data to stress-test methods for estimating both Average Treatment Effects and Heterogeneous Treatment Effects in a wide variety of simulated environments.

All of the Simulators implemented in WhyNot come equipped with experiments in both of these settings. Moreover, WhyNot provides a clean framework to allow the user to implement new experiments to explore issues not addressed by the fixed set of benchmarks. See Creating New Experiments for a detailed discussed of framework and API for creating new benchmarks.