Valid post-correction inference for censored regression problems

TL;DR

This paper develops a framework for valid post-correction inference in censored regression models, enabling accurate confidence intervals and significance tests for two-step estimators affected by censoring bias.

Contribution

It introduces a novel approach leveraging recent post-selection inference results to correct bias and perform valid inference in censored regression problems.

Findings

Provides valid confidence intervals for censored regression coefficients.

Enables significance testing after bias correction.

Improves inference accuracy in censored data models.

Abstract

Two-step estimators often called upon to fit censored regression models in many areas of science and engineering. Since censoring incurs a bias in the naive least-squares fit, a two-step estimator first estimates the bias and then fits a corrected linear model. We develop a framework for performing valid /post-correction inference/ with two-step estimators. By exploiting recent results on post-selection inference, we obtain valid confidence intervals and significance tests for the fitted coefficients.