| 对评分者信度检验中零假设的更正 |
| |
| 引用本文： | 胡咏梅.对评分者信度检验中零假设的更正[J].教育学报,2001(11):37-41. |
| |
| 作者姓名： | 胡咏梅 |
| |
| 作者单位： | 北京师范大学教育管理学院,北京,100875 |
| |
| 摘 要： | 我们通常采用Friedman双向秩次方差分析和Kendall和谐系数检验法来进行评分者信度检验.在绝大多数书籍中,此检验的零假设都被设置成"H0m个评分者对n个被评对象的评分是一致的".但是,这一零假设是错误的,本文将列举反例并给出正确假设的理论依据.
|
| 关 键 词： | 评分者信度 零假设 Friedman双向秩次方差分析 Kendall和谐系数 |
| 文章编号： | 1002-5308(2001)11-0037-05 |
| 修稿时间： | 2001年3月25日 |
The correction for the null hypothesis in the test of evaluator reliability |
| |
| HU Yong-mei.The correction for the null hypothesis in the test of evaluator reliability[J].Journal of Educational Studies,2001(11):37-41. |
| |
| Authors: | HU Yong-mei |
| |
| Abstract: | People usually use Friedman's rank test and Kendall's W test in the study of evaluator reliability test. In most books on SPSS or educational statistics the null hypothesis of the test of evaluator reliability is made as " H 0 :The evaluation results given by m evaluators are consistent". However, this null hypothesis is wrong as shown in the paper by counter examples. Moreover, we present the correct hypothesis as well as its theory basis. |
| |
| Keywords: | evaluator reliability null hypothesis Friedman's rank test Kendall's coefficient of concordance |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
