3-4年级汉语单字词阅读测验的编制

发布时间：2018-01-02 06:04

  本文关键词：3-4年级汉语单字词阅读测验的编制　出处：《湖南师范大学》2015年硕士论文　论文类型：学位论文

  更多相关文章： 单字词 阅读测验 3—4年级 IRT 信效度 等值

【摘要】：目的:本研究是根据汉字的特性,在前人编制的阅读测验工具的基础上,采用IRT理论编制了三套针对3—4年级的单字词的阅读测验并建立标准化的单字词项目库,为学生提供了单字词测评工具,为以后的汉语阅读及阅读障碍等研究提供有效的,标准化的原材料。方法:本研究首先通过前人对于汉字特性以及对汉字认知加工的研究,从小学课本中按汉字的语音、结构、字义以及使用频度进行归纳总结;其次对学生在日常学习过程中常见的字词错误进行收集和总结,对语文教师进行访谈,在其指导下完成项目编制。最后,通过个别施测与正式施测,根据数据分析结果对项目进行筛选或者修正。在本研究中,一共形成A、B、C三套平行题本,题本项目分别是108,102和103。正式施测样本抽取了长沙市城镇和农村地区6所3—4年级学生,共获得有效样本1067份。其中,三、四年级分别是596和471人男生556人,女生511人。结果:1.根据经典测验项目分析:三个题本的测验的项目难度总体偏易,测验难度在0.5—1之间所占的比例为76.8%。三个测验的项目的鉴别力指数尚可,鉴别力指数≥0.2的项目占总项目数的56%。2.单维性假设分析:采用探索性因素分析,三套测验的第一特征值与第二特征值的比值接近或者大于3,表明三个平行测验均符合IRT理论假设中的单维性假设。3.IRT项目分析:各个项目的三参模型的难度参数b在[-5.429,4]之间,其中在[-3,3]之间项目数为260个,占总项目的83%,其中在[-3,0]之间的占74%,三套测验的平均难度系数在[-1.155—-0.484]之间,总体看来,三套测验整体项目偏容易;项目区分度a在[0.296,2.608]之间,其中a"g0.5的项目占总项目的88.5%,三套测验的平均区分度[0.817,0.915]之间,其区分度比较好;项目的猜测系数c在[0.096,0.492]之间,其中c≤0.4的项目占总项目的92%,三套测验的平均猜测系数在[0.297,0.321]之间。项目特征曲线的峰值偏左并且难度参数表明,项目的难度大部分偏易。4.信度:A、B、C三个版本的测验信息最大函数值在16.32—28.20之间,测验信度较高。5.效度:通过分析结构效度采用验证性因素分析,证明本量表与前人所研究的结构框架一样,并且各种拟合指数都符合心理测量学指标。6.等值性:采用外锚的形式,使用RUMM软件将ABC三个平行卷进行了等值。7单字词分类:本次一共选取了704个字进行研究,根据各个项目的结果将字分成四类,一类字290个,将其归为一类字。二类字202个,三类字200个,四类字祥、详、的、地等在学生学习过程中容易混淆的易错字。结论:1.三套测验的项目质量优良并且题本信效度良好。2.采用外锚设计,将ABC三套试卷进行了等值。3.建立了标准化的3—4年级阅读测验的量表,为以后的汉语阅读及阅读障碍等研究提供有效的,标准化的原材料。
[Abstract]:Objective: according to the characteristics of Chinese characters, this study is based on the reading test tools developed by predecessors. Using IRT theory, three sets of single word reading tests for 3-4 grades are compiled and standardized single word item database is established, which provides students with a single word test tool. To provide effective and standardized raw materials for the further study of Chinese reading and dyslexia. Methods: firstly, this study is based on the previous studies on the characteristics of Chinese characters and the cognitive processing of Chinese characters. From the primary school textbooks according to the pronunciation, structure, meaning and frequency of use of Chinese characters to sum up; Secondly, it collects and summarizes the common mistakes in the students' daily learning, interviews the Chinese teachers, and completes the project under their guidance. Finally, through the individual test and formal test. According to the results of data analysis, the items were screened or corrected. In this study, three sets of parallel questions were formed, which were 108. 102 and 103.The official test samples were collected from 6 students of grade 3-4 in Changsha city, town and rural areas. A total of 1067 valid samples were obtained. The fourth grade was 596 and 471 male 556 and female 511 respectively. Results: 1. According to the analysis of classical test items: the difficulty of the three test items is relatively easy. The proportion of difficulty was between 0.5-1 and 76.8.The discriminant index of the three items was fair. One dimensional hypothesis analysis: using exploratory factor analysis, the ratio of the first eigenvalue to the second eigenvalue of the three sets of tests was close to or greater than 3. The results show that the three parallel tests are consistent with the one-dimensional hypothesis of the IRT hypothesis. 3.IRT item analysis: the difficulty parameter b of the three-parameter model of each item is obtained in the. [Between 5.429 and 4], in which the. [The number of projects between-3 and 3 is 260, accounting for 833 of the total. [The average difficulty coefficient of the three sets of tests was at. [-1.155-0.484], overall, the overall items of the three sets of tests are relatively easy; [0.296 / 2.608], in which a "g 0.5 items account for 88.5% of the total items, and the average differentiation of the three sets of tests is significant." [In the range of 0.817 ~ 0.915], the degree of differentiation is better; the guess coefficient c of the project is as follows:. [0. 096 / 0. 492], where c 鈮，

本文编号：1367963