Asymptotically Robust Permutation-Based Randomization Confidence Intervals for Parametric Ols Regression

Randomization inference provides exact finite sample tests of sharp nulls which precisely specify the outcomes associated with randomized treatment. A growing literature shows that when based upon studentized statistics these same tests, despite incorrectly assuming an absence of heterogeneity, can share the asymptotic validity of conventional tests of heterogeneous treatment effects. This paper extends this work to the consideration of interactions between treatment variables and covariates, a common feature of published regressions, as well as issues in the construction of confidence intervals and testing of subsets of treatment effects