节点文献

IRT框架下题组测验的等值方法在中文篇章阅读测验中的应用

Based on IRT Framework Using Equating Methods of Testlet in Chinese Passage Reading Testing

  • 推荐 CAJ下载
  • PDF下载
  • 不支持迅雷等下载工具,请取消加速工具后下载。

【作者】 罗照盛; 张飘;

【Author】 Zhang Piao;Luo Zhaosheng;School of Psychology, Jiangxi Normal University;

【机构】 江西师范大学心理学院;

【摘要】 随着社会的进步和教育的发展,现代教育测验越来越重视测验的公平性,同时也意味着测验等值问题在现实中应用的范围也越来越广。题组测验是教育测验中一种常用的测验形式,篇章阅读测验作为题组测验的一种典型,在语文学科考试等测验和语文教育质量等评估中应用地也越来越频繁。但由于语文学科的特殊性、被试理解的主观性,在中文篇章阅读测验中的等值问题也显得更为复杂。本研究立足于该研究背景,以项目反应理论(IRT)和题组反应理论(TRT)为基础,采用标准的三参数逻辑斯蒂模型(3PLM)与三参数题组模型(3PTM),用IRT项目特征曲线法进行锚测验等值,采用不同的等值准则,以等值系数误差的大小为衡量标准,分别从项目参数、随机误差、题组内项目间的相依程度等方面比较篇章阅读题组测验的等值效果,以探索适用于题组测验尤其是篇章阅读测验等值的模型和方法,并探讨在三参数情况下,题组测验中的猜测参数对等值结果的影响。本研究先进行大量蒙特卡洛(MCMC)模拟实验,再对某省语文教育质量监测中篇章阅读测验A、B、C、D卷的作答数据进行等值分析,应用并验证研究结果。研究发现,考虑题组效应影响的3PTM的等值误差比忽略题组影响的3PLM的等值误差更小,且具有显著性差异,得出3PTM更加适用于含中文篇章阅读测验的题组等值。研究进一步考察了不同的等值准则对中文篇章阅读测验等值的有效性和适用性,旨在对含有篇章阅读测验的语文类科目考试和相关教育质量评估的等值问题带来可借鉴的应用方法。

【Abstract】 Along with the progress of society and the development of education, modern education test pay more and more attention to the fairness of the test, this also means that the test equating problem in the practical the scope of application is becoming more and more widely. The testlet test is a kind of commonly used test form in the educational test. The Chinese passage reading test is a typical example of testlets, and it is also enioying more and more populairty in Chinese subject tests and Chinese education quality evaluation. However, due to the particularity of Chinese subject, the participants understand the subjectivity, the equating problem in Chinese passage reading test is also appears more complicated. Based on the background of the Item Response Theory(IRT) and the Testlet Response Theory(TRT), this study used the standard 3 Parameters Logistic Model(3 PLM) and the 3 Parameters Testlet model(3 PTM), the item characteristic curve method is used to test the equivalence of the anchor test, the different equating criteria are used to measure the size of the error of the equating coefficient.The article compares the item parameters, the random error, the degree of dependency between the items in the testlet, in order to explore the model and method applicable to the testlet tests, especially the Chinese passage reading test, and to explore the influence of the guessing parameters on the equivalence results in the 3 parameter case. In this study, a large number of Monte Carlo(MCMC) simulation experiments were carried out, and then the equivalence analysis was performed on the test data of A, B, C and D Chinese passage reading volumes in the education quality assessment of some province. The results were applied and verified. It is found that the equivalent error of 3 PTM, which is influenced by the testlet effect, is smaller than that of the 3 PLM which is affected by the neglected group, and has significant differences mostly. It is found that 3 PTM is more suitable for the equating of the Chinese passage reading test. The study further examines the validity and applicability of the different equating criteria to the Chinese passage reading test’s equate, and aims to provide a reference method for the equat of Chinese subject examinations and some related educational quality assessments with Chinese passage reading tests.

  • 【会议录名称】 第二十届全国心理学学术会议--心理学与国民心理健康摘要集
  • 【会议名称】第二十届全国心理学学术会议--心理学与国民心理健康
  • 【会议时间】2017-11-03
  • 【会议地点】中国重庆
  • 【分类号】G633.3
  • 【主办单位】中国心理学会
节点文献中: 

本文链接的文献网络图示:

本文的引文网络