Small-sample corrections for score tests in Birnbaum-Saunders regressions

TL;DR

This paper develops a Bartlett-type correction for score tests in Birnbaum-Saunders regression models to improve small-sample inference accuracy, outperforming traditional tests in simulations.

Contribution

It introduces a novel Bartlett-type correction for score tests specifically tailored for Birnbaum-Saunders regression models, enhancing small-sample reliability.

Findings

The corrected score test outperforms the usual score test in small samples.

The corrected test is more reliable than the likelihood ratio test and its Bartlett correction.

Simulation results demonstrate improved accuracy of the proposed correction.

Abstract

In this paper we deal with the issue of performing accurate small-sample inference in the Birnbaum-Saunders regression model, which can be useful for modeling lifetime or reliability data. We derive a Bartlett-type correction for the score test and numerically compare the corrected test with the usual score test, the likelihood ratio test and its Bartlett-corrected version. Our simulation results suggest that the corrected test we propose is more reliable than the other tests.