多级评分题组模型在《基于实物操作的老年人日常问题解决能力测验》中的应用

发布时间：2018-04-14 13:14

本文选题：题组 + 局部试题依赖　；参考：《赣南师范学院》2014年硕士论文

【摘要】：随着现代心理与教育测量的发展，特别是自二十世纪五十年代以来项目反应理论的发展，测验编制的科学性越来越高。项目反应理论是建立在局部试题独立的强假设下，但随着具有局部试题依赖的题组题型的运用越来越广泛，项目反应理论对题组类测验的参数估计会出现偏差，此时应运用题组反应理论进行测验质量分析。目前国内外对题组反应理论的研究多是模拟研究，实证研究非常少。本研究为考察基于实物操作的老年人日常问题解决能力测验的项目性能，采用基于实物操作的问卷调查法对450名60至84周岁的老年人进行实证研究，运用题组反应模型及传统项目反应模型进行数据分析。结果如下：（1）《基于实物操作的老年人日常问题解决能力测验》质量优良，该测验具有良好的区分度、难度；（2）《基于实物操作的老年人日常问题解决能力测验》符合单维性假设条件，即测验只测量老年人日常问题解决能力；（3）《基于实物操作的老年人日常问题解决能力测验》违背局部试题独立前提，Q3统计量及题组效果γ结果显示项目间存在依赖性；（4）测验长度一定时，不管是4个题组的分类还是更精细的10个题组的分类，题组反应模型参数估计结果基本一致，特别是被试能力参数估计结果几乎一样；（5）与题组反应模型相比，忽略项目间依赖性的传统项目反应模型会高估测验区分度、难度，而且明显高估被试能力，低估被试能力估计值的估计标准误。
[Abstract]:With the development of modern psychological and educational measurement, especially the development of project response theory since 1950s, the scientific nature of test compilation is becoming more and more scientific.The item response theory is based on the strong assumption that local test questions are independent. However, with the more and more extensive application of item response theory to the parameter estimation of problem group test, the item response theory will deviate from the parameter estimation of item group test.At this time, we should use the theory of group reaction to analyze the quality of the test.At present, the research on the theory of group reaction is mostly simulated, but the empirical research is very few.In order to investigate the project performance of daily problem-solving ability test based on physical manipulation, 450 aged 60 to 84 years old were investigated by questionnaire based on physical operation.The problem group response model and the traditional item response model were used to analyze the data.The results were as follows: (1) < the Test of Daily problem solving ability based on physical manipulation > > excellent quality, the test has a good degree of distinction, the difficulty is 2) < the Test of Daily problem solving ability based on physical manipulation > meets the condition of one dimensional hypothesis.That is, the test only measures the daily problem-solving ability of the elderly (3) < Daily problem-solving ability Test based on physical manipulation > contrary to the independent premise of local test questions / Q3 statistics and the results of item group 纬 show that there is a dependency between items.When the length of the test is fixed,Whether it is the classification of 4 problem groups or the classification of more refined 10 problem groups, the results of estimating the parameters of the problem group response model are basically the same, especially the result of estimating the ability parameters of the subjects is almost the same as that of the test group response model.The traditional project-response model which neglects the inter-item dependence overestimates the degree of distinction and difficulty of the test and obviously overestimates the ability of the subjects and underestimates the standard of estimating the ability of the subjects.
【学位授予单位】：赣南师范学院
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：B844.4

【参考文献】