基于在线英语口语诊断测试的语篇表现典型特征分析

An Analysis of the Distinguishing Features of Discourse Performance in an Online Diagnostic Speaking Test

语篇能力是口语能力的重要组成部分,以往研究较少从语篇表现层面关注不同水平英语学习者,尤其是大学生群体的口语能力。本研究基于口语语篇衔接和连贯表现,使用某在线诊断测试任务中的真实语料,结合量化分析和质性分析,探究可区别大学生口语水平的语篇表现典型特征。结果表明,在衔接表现方面:补充词和前指照应词可以明显区分高、中、低水平学生,水平更高的学生更多地使用补充词和前指照应词;对比词可以明显区分高水平学生,高水平学生更多地使用对比词;因果词、重复和近义词可以明显区分低水平学生,低水平学生更少地使用这些手段。在连贯表现方面,除产出语篇类型外,C-单元间的联系和话题与任务关联度均能明显区分不同水平学生:低水平学生的观点间联系松散,缺少具体说明和解释;中等水平学生的观点间有一定联系但准确性欠佳,能适当推进话题发展;高水平学生的观点间联系紧密,能有效推进话题发展。本研究为英语口语测试评分标准的优化、诊断反馈指标的确定及测试效度论证提供一定的参考。

This empirical study seeks to identify the criterial features of coherence and cohesion that distinguish levels(indicated by test scores)of candidates’performances in an oral task of an online lowstakes diagnostic speaking test for college students.By adopting both quantitative methods(descriptive analysis and the analysis of variance)and qualitative methods(discourse analysis and rater interviews),discourse features that can distinguish levels of test candidates and how they make the meaning conveyance possible were identified.The results suggest that additives can distinguish three levels of candidates(low,mid,and high)significantly whereas most other linking devices could distinguish levels between low and high but fail to differentiate adjacent levels(i.e.,low-mid,mid-high).Findings from discourse analysis and rater interviews indicate that candidates of the three levels do differ in their coherence performance in terms of the structure,relevance of supporting details and thesis,and the richness of content.This study will shed light on the construction of rating criteria for the coherence and cohesion dimension of speaking tests and provide evidence for the validity argument of this speaking test.