Likelihood Ratio Tests for Homogeneity of Two Normal Distributions under Small Sample Sizes
-
- ARIZONO Ikuo
- Osaka Prefecture University
-
- KATSUYAMA Shigeru
- Osaka Prefecture University
-
- OHTA Hiroshi
- Osaka Prefecture University
Bibliographic Information
- Other Title
-
- 小サンプルによる正規母集団分布に関する2標本尤度比検定
- ショウ サンプル ニヨル セイキ ボシュウダン ブンプ ニ カンスル 2 ヒョ
Search this article
Abstract
This paper deals with the null hypothesis test for homogeneity of two normal populations. The likelihood ratio test is a usual method for such a test. It is well-known that the distribution of the log-likelihood ratio statistics as the test statistics in likelihood ratio test is asymptotically explained based on the chi-square distribution with 2 degrees of freedom under the null hypothesis as sample size tends to infinity. Therefore, the null hypothesis test for homogeneity of two normal populations is generally designed based on the asymptotic chi-square distribution. However, when sample size is not sufficiently large, the accuracy of the approximation based on the asymptotic chi-square distribution is not necessarily assured. In this paper, the stochastic properties of the log-likelihood ratio statistics for homogeneity of two normal populations are investigated. Furthermore, we consider the null hypothesis test for homogeneity of two normal populations with high accuracy in the case that sample size is not sufficiently large, that is, small.
Journal
-
- Journal of Japan Industrial Management Association
-
Journal of Japan Industrial Management Association 47 (6), 367-372, 1997
Japan Industrial Management Association
- Tweet
Details 詳細情報について
-
- CRID
- 1390282680481649792
-
- NII Article ID
- 110003945681
-
- NII Book ID
- AN10561806
-
- ISSN
- 21879079
- 13422618
-
- NDL BIB ID
- 4134386
-
- Text Lang
- ja
-
- Data Source
-
- JaLC
- NDL
- CiNii Articles
-
- Abstract License Flag
- Disallowed