摘要
本研究应用Caojing等人的Bayesian IRT Guessing系列模型,分析初中二年级学生在汉语词汇测验中的猜测行为,使用DIC3指标评价模型的拟合程度,并将参数估计结果与双参数Logistic模型进行了比较。研究发现:(1)猜测模型的拟合度优于双参数Logistic模型;(2)初中二年级测验数据最适合临界猜测模型(IRT-TG),约有3.5%的学生存在TG型猜测行为;(3)猜测者的存在会明显影响本身的能力估计与项目难度估计,但是对非猜测者的能力及区分度参数估计影响不大。
This study applied the Bayesian IRT guess models by Caojing to Chinese vocabulary test of grand 8, and compared them with the 2PL-IRT using Bayesian methodology. The results show that ： （ 1 ） all guess models perform better than the 2PL model on the data set; （2） 3.5% students of grand 8are detected guessers, they answer questions based on their knowledge up to a certain test item, and guess randomly thereafter; （3） the guessers will significantly influence the estimation of their ability parameters and the item＇s difficulty parameters, however it has little impact on non- guessers＇ ability and item＇s discrimination estimates.
出处
《考试研究》
2012年第5期19-28,共10页
Examinations Research
