A Simple Test for the Absence of Covariate Dependence in Hazard Regression Models

TY - UNPB

T1 - A Simple Test for the Absence of Covariate Dependence in Hazard Regression Models

AU - Bhattacharjee, Arnab

PY - 2004

Y1 - 2004

N2 - This paper extends commonly used tests for equality of hazard rates in a two-sample or k-sample setup to a situation where the covariate under study is continuous. In other words, we test the hypothesis that the conditional hazard rate is the same for all covariate values, against the omnibus alternative as well as more specific alternatives, when the covariate is continuous. The tests developed are particularly useful for detecting trend in the underlying conditional hazard rates or changepoint trend alternatives. Asymptotic distribution of the test statistics are established and small sample properties of the tests are studied. An application to the e¤ect of aggregate Q on corporate failure in the UK shows evidence of trend in the covariate e¤ect, whereas a Cox regression model failed to detect evidence of any covariate effect. Finally, we discuss an important extension to testing for proportionality of hazards in the presence of individual level frailty with arbitrary distribution.

AB - This paper extends commonly used tests for equality of hazard rates in a two-sample or k-sample setup to a situation where the covariate under study is continuous. In other words, we test the hypothesis that the conditional hazard rate is the same for all covariate values, against the omnibus alternative as well as more specific alternatives, when the covariate is continuous. The tests developed are particularly useful for detecting trend in the underlying conditional hazard rates or changepoint trend alternatives. Asymptotic distribution of the test statistics are established and small sample properties of the tests are studied. An application to the e¤ect of aggregate Q on corporate failure in the UK shows evidence of trend in the covariate e¤ect, whereas a Cox regression model failed to detect evidence of any covariate effect. Finally, we discuss an important extension to testing for proportionality of hazards in the presence of individual level frailty with arbitrary distribution.

KW - Covariate dependence

KW - Continuous covariate

KW - Two-sample tests

KW - Trend tests

KW - Proportional hazards

KW - Frailty

KW - Linear transformation model

M3 - Working paper

T3 - MPRA

BT - A Simple Test for the Absence of Covariate Dependence in Hazard Regression Models

PB - University Library of Munich, Germany

CY - Munich

ER -