基于IRT模型的数学试题类型测验效果的研究概率论与数理统计专业论文.docxVIP

基于 IRT 模型的数学试题类型测验效果的研究 中文摘要 基于 I RT 模型的数学试题类型测验效果的研究 中文摘要 项目反应理论是一种新型的心理和教育测试理论。本文运用此理论中的三参 数正态曲线模型分别得出同一批试题在选择题和填空题形式下的项目参数值，而 项目参数的估计可以通过 Bayesian 理论和 MCMC 方法得出。从 2008 年开始，江 苏高考数学不再有选择题，此改革引起一片质疑声。本文通过运用项目反应理论 并借助 WinBUGS 软件探讨如下问题：数学试卷上没有选择题是否妥当？ 本文通过实验分析出的数据得到如下结论：在保证选择题提供的有用信息较 少的条件下，选择题与填空题对学生能力区分的效果是没有显著差异的，试题的 难度参数与机遇参数也没有显著性差异，但填空题对学生能力提出的要求更高。 关键词：项目反应理论；MCMC 方法；三参数正态曲线模型；WinBUGS 作 者：陈佳程 指导老师：汪四水 Abstract The Research on the Effect of Different Types of Math Questions Based on… The Research on the Effect of Different Types of Math Questions Based on the IRT Model Abstract Item Response Theory is a new theory of psychological and educational testing. We use the three-parameter normal curve model of this theory to get the item parameter values by a number of questions in the form of multiple-choice and fill-in-the-blank questions, respectively, then the parameter values can be estimated by Bayesian theory and MCMC methods. From 2008, there are no multiple-choice questions in the college entrance math examination of Jiang Su. This change has brought a lot of controversy. In this paper, we use the item response theory and WinBUGS software to analyze the following issue: Is it reasonable that no multiple-choice questions are included in the Mathematical exams ? According to the data, the final conclusion draws like this: the difference between using multiple-choice and using fill-in-the-blank questions to distinguish the students ability is not much obvious without providing useful information by multiple-choice questions, and the item difficulty parameters and opportunities parameters also have no significant difference, however, the fill-in-the-blank questions raise the higher requirements for students’ abilities. Keywords: IRT; MCMC method; three-parameter normal ogive model; WinBUGS Written by Jiacheng Chen Supervised by Associate Prof Sishui Wang 目录 第一章 引言 ..............................................................................................