| 信度与信度的估计 |
| |
| 引用本文： | 罗冠中.信度与信度的估计[J].华南师范大学学报(自然科学版),1989,0(1):1. |
| |
| 作者姓名： | 罗冠中 |
| |
| 作者单位： | 华南师范大学教育科学研究所 |
| |
| 摘 要： | 信度的实质是二元随机变量(x,t)的一个数字特征.在不同的条件下,信度可用稳定性系数或内部一致性系数来估计.本文对这些估计的性态进行讨论,指出;1)不能把总体方差与样本方差混为一谈;2)所有“平行测量”之间有相同的相关系数(并不一定为1);3)K-R_(20)公式中因子N/N-1的运用值得商榷.
|
| 关 键 词： | 测量 信度 估计 总体方差 样本方差 平行测量 K-R_(20)公式 |
RELIABILITY AND ITS ESTIMATION |
| |
| Luo Guanzhong.RELIABILITY AND ITS ESTIMATION[J].Journal of South China Normal University(Natural Science Edition),1989,0(1):1. |
| |
| Authors: | Luo Guanzhong |
| |
| Institution: | Educational Science Inst. |
| |
| Abstract: | Reliability is essentially a numeral characteristic of bi-variable (x,t). Reliability can be estimated by stability coefficient or internal consistency coefficient under various conditions. This paper discusses the properties of these estimators and points out. I. Sample variance must not be taken as the same as population variance;II . all the parallel measurements have the same coyariance coefficient ( not always be 1.) ; II. In formula K-R20, the factor
N/N-1 is not being applied properly. |
| |
| Keywords: | measurement reliability estimation population variance sample variance parallel measurement formula K-R_(20) |
| 本文献已被 CNKI 等数据库收录！ |
| 点击此处可从《华南师范大学学报(自然科学版)》浏览原始摘要信息 |
| 点击此处可从《华南师范大学学报(自然科学版)》下载免费的PDF全文 |
