Examining genre effects on test takers' summary writing performance
Examining genre effects on test takers' summary writing performance
复制标题
DOI：
10.1016/j.asw.2014.08.003
复制
发表时间：
2014-10-01
影响因子：
3.9
通讯作者：
Li Jiuliang
中科院分区：
文献类型：
--
作者：
Li Jiuliang
摘要的任务要求与源文本的特征密切相关，而体裁是一个必不可少的特征。本文报告了一项实证研究，探讨了文本类型（体裁）如何影响考生在摘要任务中的表现。从中国一所大学的本科项目中抽取了86名学生作为样本。学生们首先写一篇叙事性文章的摘要，然后写一篇说明性文章的摘要。从学生的总结分数、总结脚本以及问卷调查和测试后访谈中反映的对这些效果的感知三个角度来考察类型效应。MFRM分析显示，参与者在说明性写作上的表现优于叙事性文本摘要。题目组成部分的难度各不相同，在不同的任务中有几个难度差别很大。然而，从问卷调查和访谈的结果来看，参与者普遍认为叙事性文本摘要任务比说明性任务更容易。通过检查参与者的总结以及他们对任务难度和测试过程的描述，研究了导致表现和感知之间矛盾的因素。本文讨论了总结任务设计、总结教学以及类型效应在创建等效测试版本中的相关性。(C) 2014 Elsevier Ltd.版权所有。
The task demands of summarization are closely related to the characteristics of source texts, and genre is an essential characteristic. This paper reports an empirical study that examines how text type (genre) affects test takers' performance on summarization tasks. A sample of 86 students was drawn from an undergraduate program in a Chinese university. The students first wrote summaries of a narrative text and subsequently wrote summaries of an expository text. Genre effects were examined from three perspectives: students' summary scores, summary scripts, and perception of these effects as reflected in questionnaire surveys and post-test interviews. MFRM analysis showed that participants performed better on expository writing than on narrative text summarization overall. The difficulties of the rubric components differed, with several differing significantly across the tasks. However, the participants generally considered the narrative text summarization to be easier than the expository task according to the results of the questionnaire surveys and interviews. Factors that led to this contradiction between performance and perception were explored by examining the participants' summaries and their accounts of the task difficulty and test-taking processes. Implications are discussed with reference to summarization task design, summarization teaching, and the relevance of genre effects in the creation of equivalent versions of tests. (C) 2014 Elsevier Ltd. All rights reserved.