Exact Bahadur Slope for Combining Independent Tests in Case of Laplace Distribution

Abstract

Combining n independent tests of simple hypothesis, vs one-tailed alternative as n approaches infinity, in case of Laplace distribution L(γ , 1) is proposed. Four free-distribution ”nonparametric” combination procedures namely; Fisher, logistic, sum of P-values and inverse normal were studied. Several comparisons among the four procedures using the exact Bahadur’s slopes were obtained. Results showed that the sum of p-values procedure is better than all other procedures under the null hypothesis, and the inverse normal procedure is better than the other procedures under the alternative hypothesis.

Key words and phrases. Laplace distribution, combining independent tests, Bahadur efficiency.

2000 Mathematics Subject Classification. 40H05, 46A45