首页 | 本学科首页   官方微博 | 高级检索  
     

4参数GRM对猜测现象和失误现象的纠正
引用本文:简小珠,戴海琦. 4参数GRM对猜测现象和失误现象的纠正[J]. 江西师范大学学报(自然科学版), 2016, 40(2): 116-119
作者姓名:简小珠  戴海琦
作者单位:1.井冈山大学教师教育研究中心,江西 吉安 343009; 2.江西师范大学心理学院,江西省心理与认知科学重点实验室,江西 南昌 330022
摘    要:将c、γ参数加入到Samejima等级反应模型中形成4参数等级反应模型(4参数GRM),该模型包含了两级记分1、2、3、4参数Logistic模型、Samejima等级反应模型.4参数GRM适合测验中的多级和两级记分试题,也可以适合两级记分试题的猜测现象和失误现象.Samejima等级反应模型下,被试作答的猜测现象会导致能力高估现象,失误现象会导致能力低估现象.在4参数GRM下,被试能力高估现象和低估现象均得到了有效的纠正.

关 键 词:项目反应理论  等级反应模型  4参数等级反应模型  猜测现象  失误现象

Four-Parameter GRM and the Countermeasure to Sleeping and Guessing Phenomena
JIAN Xiaozhu,DAI Haiqi. Four-Parameter GRM and the Countermeasure to Sleeping and Guessing Phenomena[J]. Journal of Jiangxi Normal University (Natural Sciences Edition), 2016, 40(2): 116-119
Authors:JIAN Xiaozhu  DAI Haiqi
Affiliation:1.Education Research Centre,Jinggangshan University,Ji'an Jiangxi 343009,China; 2.School of Psychology,key Laboratory of Psychology and Cognition Science of Jiangxi Jiangxi Normal University,Nanchang Jiangxi 330022,China
Abstract:There exist the sleeping phenomena that the high-ability examinees make wrong response on the easy item,and the guessing phenomena that the low-ability examinees make correct response on the difficult item,in the Paper and Pencil Test and CAT.The authors add c parameter and γ parameter into the Samejima's graded-response model,and get a new model,4P-GRM.Rasch model,two-parameter logistic model,three-parameter logistic model,four-parameter logistic model all are the special case of the 4P-GRM.A Polytomous test is designed,and an average examinee have been arranged to tested on the test.After the examinee finish the test and can get the estimated ability θ using the program of MULTILOG.An extra item with difficulty parameters arranged from big to small in different test case was given to the examinee..The author estimates the new ability θ'.The author calculates the ability steplength(θ-θ').According the outcome of the ability steplength:(i)When getting full score,the examinee will be overestimated when the examinee makes correct responses on the difficulty items.(ii)When getting zero score,the examinee will be underestimated when the examinee makes wrong responses on the easy items.(iii)When getting middle score,there is the ability overestimation phenomenon and underestimation phenomenon meanwhile.Furthermore,under the 4P-GRM,the result is that:(i)When the score is full score,the overestimation phenomenon on the low ability examinee can be rectified.(ii)When the score is zero,the underestimation phenomenon on the high ability examinee can be rectified.(iii)When the score is middle score,the overestimation phenomenon and underestimation phenomenon can also be rectified
Keywords:IRT  GRM  4P-GRM  guessing phenomenon  sleeping phenomenon
本文献已被 CNKI 等数据库收录!
点击此处可从《江西师范大学学报(自然科学版)》浏览原始摘要信息
点击此处可从《江西师范大学学报(自然科学版)》下载全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号