高考英语完形填空题的测试任务研究——以2014-2018年全国卷I为例

A Study of Cloze Test Tasks in NMET——Taking an Example from Its National Version I During the Years 2014-2018

本研究以 《普通高中英语课程标准（2017 年版） 》 和高考英语考试大纲对完形填空题所考查语言能力的要求为指引, 根据 Bachman 和 Palmer （1996） 的任务特征框架和李筱菊（1997） 的考点效度理论, 构建了完形填空试题测试任务的文本特征、 考点特征、 选项设置 “三维度” 分析框架, 分别对 2014-2018 年全国卷 I 的完形填空样本展开历时分析,对 2017、2018 年两年各三份完形填空样本展开共时分析. 采用访谈法对典型难题作以应对策略分析; 并藉此对模拟试题考点设置的科学性进行探讨.总体上, 本研究发现虽然目前高考完形填空题的三个维度基本稳定, 給模拟题的设计提供了可供模仿的参照标准, 在文本特征（如选材多样性、文章难度等） 和选项设置维度上尤其存在值得商榷的问题.经过分析, 笔者认为试题设计应把语篇输入真实性作为首要原则考虑; 在综合语言能力培养上应设计立足于语篇、 立足于语用的教学活动.

This study sets guidelines on the requirements for evaluating English language proficiencythrough cloze tests in accordance with the Senior High School English Curriculum Standards （2017 ed.） andthe annual NMET testing syllabus. Referring to Bachman and Palmer’ s （1996） framework of test taskcharacteristics and Xiaoju Li’ s （1997） theory of testing point validity, this study presents a three-dimensionalanalytical framework which consists of text characteristics, testing point characteristics and distribution ofoptions, and conduct a diachronic analysis of the five cloze test samples in NMET （National Version I） duringthe years 2014-2018 accordingly. It also compares the data with those on the cloze test samples （NationalVersions II ＆ III） in 2017 and 2018 synchronically. The study then explores the test taking strategies of thesurveyed items with the method of interview and the scientificalness of distributing testing points for testwriters. Basically, the study finds that there remain problems especially in text characteristics （e.g. the inputdiversity and text difficulty, etc.） and distribution of options, although the three dimensions of NMET clozetests are sustainable, which sets the visible criteria for mock test writers. On analysis, the authors argue thatin cloze test design it is expected to consider authenticity as the number one principle of input; and in ELTinstructional design be discourse-based and pragmatics-based, hence promote learners’integrated languagecompetence.