All models may be wrong -- but that is not necessarily a problem for inference. Consider the standard -test for the significance of a variable for predicting response whilst controlling for other covariates in a random design linear model. This yields correct asymptotic type~I error control for the null hypothesis that is conditionally independent of given under an \emph{arbitrary} regression model of on , provided that a linear regression model for on holds. An analogous robustness to misspecification, which we term the "double-estimation-friendly" (DEF) property, also holds for Wald tests in generalised linear models, with some small modifications. In this expository paper we explore this phenomenon, and propose methodology for high-dimensional regression settings that respects the DEF property. We advocate specifying (sparse) generalised linear regression models for both and the covariate of interest ; our framework gives valid inference for the conditional independence null if either of these hold. In the special case where both specifications are linear, our proposal amounts to a small modification of the popular debiased Lasso test. We also investigate constructing confidence intervals for the regression coefficient of via inverting our tests; these have coverage guarantees even in partially linear models where the contribution of to can be arbitrary. Numerical experiments demonstrate the effectiveness of the methodology.
View on arXiv