Refuting causal estimates

As with scientific theories, the result of a causal analysis based on observational data cannot be proved correct, but we can try to refute it using multiple robustness tests. An analysis that fails any one of the tests is incorrect and needs to be fixed. These tests can also be used to rank competing estimation methods: we should choose the estimation procedure that fails the least number of refutation tests.

• Refuting Effect Estimates
  • Refutations based on negative control
  • Refutations based on sensitivity analysis