TESENSITIVITY: Stata module for assessing sensitivity to the unconfoundedness assumption

This module provides a set of tools for analyzing the sensitivity of treatment effect estimates to relaxations of the unconfoundedness assumption, also known as selection on observables, exogenous selection, ignorability, or conditional independence. Specifically, it calculates bounds on treatment effect parameters by relaxing the unconfoundedness assumption to a conditional partial independence assumption. Bounds are calculated for a range of alternative assumptions about the magnitude of selection on unobservables. The breakdown point for conclusions about these treatment effects is also calculated. This is defined as the maximum level of selection on unobservables under which a specific conclusion still holds. The module also provides tools to visualize the results, compare results from multiple datasets, and interpret the the scale of the sensitivity parameter.