摘要
本文在探讨测验构想效度验证和测验方法的基础上,使用等级聚类分析法(HCA)对HSK(初、中等)听力理解测验的项目纬度进行研究。基于测验构想效度和任务真实性考虑,并参照听力测验项目的实证聚类结构与句子、简短对话、讲话或长对话三种测验任务的操作结构对应关系,对听力理解测验任务形式的有效程度和改进方式,提出了自己的建议。
Based on the discussion of the construct validation of listening comprehension and test methods, this research, using hierarchical cluster analysis, addresses itself to the underlying dimensions of listening comprehension items. The empirical research results indicate that there is no correspondence between item dimensionality and the listening test operational structures. From the construct validity perspective, the research suggests that the statement-level items be replaced with other more authentic test tasks without compromising the validity.
出处
《语言文字应用》
CSSCI
北大核心
2011年第1期73-80,共8页
Applied Linguistics
基金
北京语言大学校级科研项目(项目编号:08YB01)
关键词
听力测验
效度
构想效度
汉语水平考试
效度研究
listening comprehension tests
validity
construct validity
language testing
validation