Bivariate symmetry tests for complete and competing risks data: a saddlepoint approach

Abstract

A class of bivariate symmetry tests for complete data and competing risks data is considered. Saddlepoint approximation for the exact p-values of the underlying permutation distribution of these tests is derived. Several simulation studies are conducted to evaluate the performance of the saddlepoint approximation and the asymptotic approximation. The saddlepoint approximation was found to be highly accurate and superior to the asymptotic approximations in replicating the exact permutation significance.

KEYWORDS:

• Bivariate symmetry tests
• competing risks
• permutation distribution
• saddlepointapproximations

MSC2010:

• 62G10
• 62E17
• 62P10
• 92B15