May 24, 2018
By Rachel Nethery, Fabrizia Mealli, and Francesca Dominici
Most causal inference studies rely on the assumption of positivity, or overlap, to identify population or sample average causal effects. When this assumption is violated, these estimands are unidentifiable without some degree of reliance on model specifications, due to poor data support. Existing methods to address non-overlap, such as trimming or down-weighting data in regions of poor support, all suffer from the limitation of changing the estimand. The change in estimand may diminish the impact of the study results, particularly for studies intended to influence policy, because estimates may not be representative of effects in the population of interest to policymakers. Researchers may be willing to make additional, minimal modeling assumptions in order to preserve the ability to estimate population average causal effects.