当前位置:主页 > 社科论文 > 心理论文 >

题组反应理论及其在中学英语考试中的应用研究

发布时间:2018-08-30 10:31
【摘要】:题组作为众多测验中的一种常见题型,由于项目间存在一定程度的依赖性而违背了局部独立性假设,若用项目反应模型进行参数估计将会出现较大的偏差。题组反应理论将被试与题组的交互作用纳入到模型中,解决了项目间相依性的问题。笔者对题组反应理论的发展、基本原理及其相关研究进行了综述,并将其应用在中学英语考试中。与项目反应理论相对比,结果发现:(1)题组反应模型与项目反应模型在各参数估计值的相关系数较强,尤其是能力参数和难度参数;(2)在置信区间宽度的比较上,题组反应模型在各个参数上均窄于项目反应模型,即题组反应模型的估计精度优于项目反应模型。
[Abstract]:As a common item type in many tests, the problem group violates the assumption of local independence because of the dependence between items. If the item response model is used for parameter estimation, there will be a big deviation. The interaction between subjects and problem groups is incorporated into the model by the theory of problem group reaction, which solves the problem of item dependence. In this paper, the author summarizes the development, basic principles and related research of group reaction theory, and applies it to the middle school English examination. Compared with the item response theory, the results show that: (1) the correlation coefficient between the item response model and the item response model is strong in each parameter estimation, especially the ability parameter and difficulty parameter; (2) in the confidence interval width comparison, it is found that: (1) the correlation coefficient between the item response model and the item response model is strong, especially the ability parameter and the difficulty parameter; The problem group response model is narrower than the item response model in each parameter, that is, the estimation accuracy of the problem group response model is better than that of the item response model.
【作者单位】: 华南师范大学心理应用研究中心;
【基金】:教育部重点课题(GFA111009) 广州卓越教育培训中心项目
【分类号】:B841.2

【参考文献】

相关期刊论文 前4条

1 叶萌;;对IRT局部独立性假设问题的认识[J];考试研究;2010年02期

2 涂冬波;漆书青;蔡艳;戴海琦;丁树良;;IRT模型参数估计的新方法——MCMC算法[J];心理科学;2008年01期

3 涂冬波;蔡艳;漆书青;丁树良;戴海琦;;项目反应理论新进展——题组模型及其参数估计的实现[J];心理科学;2009年06期

4 刘s,

本文编号:2212804


资料下载
论文发表

本文链接:https://www.wllwen.com/shekelunwen/xinlixingwei/2212804.html


Copyright(c)文论论文网All Rights Reserved | 网站地图 |

版权申明:资料由用户5255a***提供,本站仅收录摘要或目录,作者需要删除请E-mail邮箱bigeng88@qq.com