Your Cart

Your cart is empty

Home  >  volume 36 (no1)

Large Sample Approximation for Two Sample Problems: How Large is Large? y S.M. Ogbonmwan and F. Ewere (Pages 169-172)
Sale price: $5.00
Quantity:

Abstract

In statistical inference, researchers have not agreed on the actual sample size before recommending large sample approximations for various tests of hypotheses. Many suggestions for sample sizes have been given by different researchers. A compilation of the minimum sample sizes for large sample approximations for twenty nonparametric statistics exist in the literature. The report of this study is an attempt at setting the actual sample sizes for some nonparametric tests for two sample experiments. The reported sample sizes for large sample approximations were obtained through unconditional permutation and combinatorial approaches. Both approaches generate exact p-values and the probability of type I error is exactly  . Thus, the results obtained in the study reported here are the actual sample sizes for large sample approximations. This study has therefore resolved the conflicts in the choice of sample sizes for the application of asymptotic results for some tests for two sample experiments.

Keywords: Exact test, large sample approximation, Permutation test, Combinatorics, Nonparametric test

Close
Loading...