高中英语词汇知识双阶自适应测试系统的开发与应用

发布时间：2018-04-19 15:09

本文选题：项目反应理论 + 双阶自适应测试　；参考：《山东师范大学》2017年硕士论文

【摘要】：英语作为一门在国际上比较通用的语言,在我国基础教育阶段就备受重视。而词汇是语言学习中的基础性分支,其重要性不言而喻。作为言语产出的源动力,词汇在日常交际中发挥着不可替代的作用,对词汇知识的掌握会直接影响到个体其他语言能力分项的水平。因此,英语词汇知识的多寡在很大程度上表征了一个人英语水平的高低,对词汇的测试就成为英语教学环节中非常关键的一部分。然而,目前的英语词汇测试系统虽然在测试形式、词汇的设置上展示出了优良特性,但仍然存在诸多问题。对词汇知识维度的测试不够完善,只关注到词汇广度知识而忽略了词汇深度知识;此外,传统的纸笔测验形式或简单的基于计算机的测试已不能满足人们对个性化测试的需求。因此,开发一个便于学生全面了解自身词汇知识水平的个性化测试系统势在必行。研究中采用了文献研究法、数学模型法、问卷调查法、统计分析法等多种研究方法,在文献查阅的基础上明确了系统研发的基础方向和终极目标。在系统设计阶段,针对目前英语词汇测试系统存在的问题以及人们对个性化测试的呼吁,以高中英语词汇为例,设计了一个基于项目反应理论的英语词汇知识双阶自适应测试系统。秉承了在词汇广度知识测试的基础上再进行深度知识测试的思想,综合考量词汇知识的广度与深度两个维度,精确评估用户的词汇知识水平,为其带来全新的测试体验。在系统开发阶段,采用DreamWeaver+ASP+Access的方式完成系统研发,其中最核心的部分是两个词汇知识题库的建设。在广度知识题库建设中,选用单参数逻辑斯蒂模型拟合试题,并从英语词汇的词频、长度、语音书写和谐程度三个维度综合定义了3366个高中英语大纲词汇的难度值。经统计,词汇难度值的频次分布大体拟合正太分布。在深度知识题库建设中,选用了双参数逻辑斯蒂模型。采用锚测验的设计模式,通过试测收集被试作答深度测试题的真实数据,借助国际上流行的参数估计软件BILOGMG 3.0估计出初始参数值,然后利用平均数/标准差法进行等值,从而获取了298道深度测试题的最终项目参数值。系统研发成功后进行了小规模测试,通过测试所得数据评估了系统的信度与效度水平,结果显示系统信度、效度水平良好。此外,由于个体对自适应测试系统的满意度在很大程度上影响着自适应测试的推广,因此通过实证研究分析了计算机自我效能感、测验焦虑两大个体特质变量以及培训对自适应测试系统满意度的影响。借助问卷调查所得数据,采用双因素方差分析、配对样本T检验的方式进行统计分析。研究结果表明,计算机自我效能感高的个体比计算机自我效能感低的个体对自适应测试系统的满意度高;测验焦虑对个体自适应测试系统的满意度并没有显著影响,但测验焦虑与计算机自我效能感的交互作用对系统满意度影响显著;培训后个体对自适应测试的满意度得到显著改善。因此,今后在自适应测试推广的过程中应重视培训的作用,针对不同用户群体合理地设计培训内容及形式;并且,提高个体的计算机自我效能感同样有助于自适应测试的普及。
[Abstract]:As a common language in the world, English is highly valued at the stage of basic education in our country. And vocabulary is the basic branch of language learning, its importance is self-evident. As the source of speech output, vocabulary plays an irreplaceable role in daily communication, and the mastery of vocabulary knowledge will directly affect the language knowledge. Therefore, the amount of English vocabulary knowledge to a large extent represents the level of one's English, and the test of vocabulary is a very important part of the English teaching link. However, the present English vocabulary testing system has shown excellent results in the form of test and the setting of vocabulary. There are many problems, but there are still a lot of problems. The test of the vocabulary knowledge dimension is not perfect enough to pay attention to vocabulary breadth and neglect the depth of vocabulary. In addition, the traditional paper pen test form or simple computer based test can not meet the needs of people's individualized testing. It is imperative to understand the personalized test system of the knowledge level of its own vocabulary. In the study, many research methods such as literature research, mathematical model, questionnaire and statistical analysis are adopted, and the basic direction and ultimate target of the system R & D are clarified on the basis of literature review. In the case of high school English vocabulary, a two order self-adaptive test system of English vocabulary knowledge based on the theory of project response is designed, which is based on the test of vocabulary breadth to test the depth of knowledge, and take a comprehensive consideration of the breadth of vocabulary knowledge. And the depth of the two dimensions, accurate evaluation of the level of the user's vocabulary knowledge, to bring a new test experience for them. In the system development stage, the system is completed by DreamWeaver+ASP+Access, the most important part of which is the construction of two vocabulary knowledge database. The difficulty value of the vocabulary of 3366 high school English syllabuses is defined by fitting the test questions and the three dimensions of word frequency, length and phonetic writing harmony of English vocabulary. By testing and collecting the real data of the subjects to answer the test questions, the initial parameter values are estimated by the international popular parameter estimation software BILOGMG 3, and then the average number / standard deviation method is used to obtain the final item parameter values of the 298 depth test questions. The reliability and validity of the system are evaluated by the test data, and the results show that the reliability and validity of the system are good. In addition, because the satisfaction of the individual to the adaptive test system has a great influence on the popularization of self-adaptive test, the self-efficacy of the computer is analyzed by the empirical study, and the test anxiety is two big. The effect of the body trait variables and the training on the satisfaction of the adaptive test system. By using the data from the questionnaire survey, the two factor variance analysis and the paired sample T test are used to carry out statistical analysis. The results show that the individual with high self-efficacy of computer is better than the self-efficacy of the self-adaptive test system. The satisfaction of test anxiety has no significant influence on the satisfaction of individual self-adaptive testing system, but the interaction of test anxiety and computer self-efficacy has significant influence on system satisfaction. The satisfaction of self-adaptive test after training is greatly improved. Therefore, it should be paid more attention to in the process of adaptive test promotion in the future. The role of training is to rationally design training content and form for different user groups, and improving the self-efficacy of the individual is also helpful to the popularization of self-adaptive testing.

【学位授予单位】：山东师范大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：G633.41

【参考文献】