高风险测试中写作提示对学习者英语写作表现的影响研究

发布时间：2018-04-23 22:10

本文选题：任务条件 + 写作提示　；参考：《江西师范大学》2017年硕士论文

【摘要】：任务型语言测试自20世纪90年代末已广受关注。从那时起,许多研究人员开始对深入研究任务型测试。关于任务难度和任务类型的研究丰富,但深入研究任务条件及其对任务质量影响的研究较少。本研究以Skehan的测试模型为基础,从词汇多样性、衔接和连贯、句法复杂性和可读性的角度探究提示特征对学习者写作文本复杂性的影响。本文研究四种任务条件(中文提纲提示、图片提示、图表提示和情景提示)对学习者写作表现的影响。研究问题如下:1)不同的提示类型是否会影响学生的写作表现?2)提示特征是否会影响学习者写作的文本特征?3)学习者写作表现的主要预测因子有哪些?本研究以60个英语专业本科生为研究对象,主要采用测试和文本分析的方法,研究语料为受试的240篇作文的文本。受试在2016年9月的第一周到第四周分别完成4篇不同提示的写作任务。写作表现由3个有经验的老师依据五级评分表(总分15分)给出评价。文本特征(基础计数、衔接与连贯、词汇多样性、文本复杂性及可读性)通过软件Coh-Metrix3.0计算得出。学生写作表现及文本分析结果所得数据通过SPSS 16.0进行分析。数据分析显示:1)提示特征影响学生的整体写作。写作任务1(中文提纲提示)明显高于其他三个写作任务(p=0.000)。写作任务3学生的表现最差。其原因可能是学生从高中开始就大量的训练提纲式写作。学生觉得图表写作最难,因为图表提示给学生的文本提示信息太少。2)提示特征影响学习者写作的文本特征,尤其对写作产出的长度有显著影响。情景提示的写作文本最长,而中文提纲式提示的写作文本最短。就词汇多样性和文本复杂性而言,图片提示和情景提示的词汇和句法较为复杂。3)影响学习者写作表现的预测因子主要包括段落数,句子数和字数(p=0.000)。文字的长度是衡量写作能力的重要指标。这3个因素加起来可以解释写作任务1中52.7%的写作得分。写作任务2中的6个因素(段落数、句子数、字数、词汇多样性、所有关联词数、副词关联词数)可以解释62.7%的得分。写作任务3中的5个因素(段落数、句子数、字数、词汇多样性、词干重叠率、名词短语修辞词数)可以解释50.8%的写作表现。写作任务4中的7个因素(段落数、句子数、字数、词汇多样性、名词短语修辞词数和可读性指标)可解释学生写作74.3%的得分。鉴于以上的研究发现,本研究得出一些进行语言测试和培养学习者写作能力的启示:1)高风险英语测试中的写作提示形式宜多样化,以便体现学生的个性及创造性。测试中的中文提示可能会引导学生直接将提示翻译成英语,不能真正地检测他们潜在的写作能力。2)测试设计人员应该意识到写作提示对受试者写作表现的影响。为了保证考试的效度,对不同的写作提示进行前测是有必要的。3)教师在语言教学过程中应培养学生在不同的写作提示下提高英语写作的能力。
[Abstract]:Since the end of 1990s, task-based language testing has attracted much attention. From that time, many researchers began to study task based testing. The research on task difficulty and task type was rich, but the research on task conditions and the impact on task quality were less. This study was based on the Skehan test model. The effects of diversity, cohesion, coherence, syntactic complexity and readability on the complexity of the writer's composition are explored. This paper studies the effects of four task conditions (Chinese outline hints, picture tips, chart hints and situational hints) on learners' writing performance. Research questions are as follows: (1) whether different types of hints will be different 2) does the feature affect the text characteristics of the learners' writing? (3) what are the main predictors of the learners' writing performance? This study focuses on 60 English majors, mainly using the methods of testing and text analysis, and studying the text of the 240 compositions of the subjects. The subjects are 2016 4 different writing tasks were completed in the first week to the fourth week of September. The writing performance was evaluated by 3 experienced teachers based on the five grade score (total score 15). The text features (basic count, cohesion and coherence, lexical diversity, text complexity and readability) were calculated by software Coh-Metrix3.0. The data obtained from the results of the present and text analysis are analyzed by SPSS 16. Data analysis shows that: 1) the characteristics affect the overall writing of the students. The writing task 1 (the Chinese outline hints) is obviously higher than the other three writing tasks (p=0.000). The writing task 3 students perform the worst. The reason may be that the students have been training a lot from high school. Class writing. Students find graphic writing the hardest, because the graph hints that the text messages to students are too little.2) the hint features influence the text characteristics of the learners' writing, especially on the length of the writing output. In text complexity, the vocabulary and syntax of picture hints and situational hints are more complex.3). The predictors of the learners' writing performance mainly include the number of paragraphs, the number of sentences and the number of words (p=0.000). The length of the text is an important indicator of the writing ability. The addition of these 3 factors can explain the writing score of 52.7% of the writing task 1. 6 factors (paragraph number, sentence number, word number, lexical diversity, all associated words, adverbs associated words) can explain the score of 62.7% in writing task 2. The 5 factors in writing task 3 (paragraph number, sentence number, word number, lexical diversity, stem overlap, noun phrase number) can explain 50.8% writing performance. Writing task 4 7 factors (paragraph number, sentence number, word number, lexical diversity, noun phrase number, and readability index) can explain the score of student writing 74.3%. In view of the findings above, this study finds some inspiration for language testing and training of learners' writing ability: 1) there are more forms of writing tips in high risk English tests. The Chinese hints in the test may lead the students to directly translate the hints into English and not really detect their potential writing ability.2). The test designers should be aware of the effect of writing tips on the writing performance of the subjects. In order to ensure the validity of the test, the different writing of the test will be guaranteed. It suggests that it is necessary to carry out pretest..3) teachers should cultivate students' ability to improve their English writing skills under different writing tips in the process of language teaching.

【学位授予单位】：江西师范大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：H319.3

【参考文献】