制定汉语作为第二语言的能力标准的主要任务之一是使用“能做(call—do)”描述来刻画语言水平。通过能做描述语的任务难度研究可以把语言能力标准中的描述语分派到不同的语言水平等级中去,从而完成水平等级与描述语难度之间的对应。本文以中级口语能力标准中的描述语任务难度研究为例,阐述了描述语任务难度研究的理论、方法与过程。我们请教师依据量表对443名中级水平留学生的口语能力进行评价,并使用项目反应理论中的多级计分Rasch模型(RSM)对52条描述语的任务难度进行估算。结果表明,描述语任务难度适中且分布均匀,区分性良好,但难度分布范围略小于学生能力值分布范围,出现了一定的“天花板”效应。我们依据难度估算结果,将52条描述语划分为中级高和中级低两档,初步实现了与欧洲语言共同参考框架(CEFR)的兼容。
One of the main tasks to develop the Chinese as a second language proficiency guidelines is to use ' can-do' descriptions to portray language proficiency levels. Through a study of the difficulties of the tasks, we can put the descriptors into different levels. In turn we can make correspondence between the levels of learners' tasks. This article concerns the theory, vite teachers to evaluate 443 students' language proficiency and the difficulties of method and process of studies on task difficulties. We in- speaking ability based on the oral proficiency scale of the intermediate level, and use the RSM model to estimate the 52 descriptors. The results show that the difficulties of these descriptors are moderate and evenly distributed, but the range of descriptor difficulties is slightly narrower than that of students' ability. According to this finding, we divide the 52 descriptors into two ranks, respectively named intermediate-high and intermediate-low, so as to render the scale compatible with the CEFR.