• 主题
高级搜索  >
历史搜索 清空历史
首页> 美国政府科技报告 >Note on the Robustness of Dixon's Ratio Test in Small Samples
【24h】

Note on the Robustness of Dixon's Ratio Test in Small Samples

机译：关于小样本中狄克逊比率检验的鲁棒性的注记

获取原文
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

Dixon's ratio test is a simple procedure for testing for outliers in statistical data that has been particularly effective when the data come from an underlying normal distribution. The critical values for the normal case are compared to the uniform, exponential, and Cauchy cases with samples of size three, four, and five. When the sample size is three, Dixon's ratio is fairly robust to various departures from normality (i.e., skewness and kurtosis). For symmetric distributions that are not heavy tailed, it appears to be robust in larger sample sizes than just three. However, for the Cauchy case and for very skewed distributions such as the exponential, the critical values are very different even when the sample size is only four. The fact that Dixon's ratio is robust in samples of size three is a valuable practical result, since many times the underlying distribution is not well understood. In a data validation study of a monthly power plant report Dixon's test was applied by grouping the data in sets of three. That application is discussed in this paper. 1 figure, 2 tables. (ERA citation 05:025111)

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号