读写结合写作任务思辨能力评价效度研究  

Validation Study on Critical Thinking Evaluation in Reading-to-Writing Writing Test

作　　者：邹润[1] 

机构地区：[1]南方医科大学外国语学院,广东广州510515

出　　处：《考试与评价》2022年第3期11-18,共8页TESTING AND EVALUATION

基　　金：广东省哲学社会科学“十三五”规划2020年度项目外语信息化专项“基于中国英语能力等级量表的大学英语口语信息化评价模式构建研究”(GD20WZX01-01)的阶段性成果;2021年广东省本科高校教学质量与教学改革工程项目“新文科背景下大学英语分级教学改革与教学评价”研究阶段性成果。

摘　　要：本文运用多面Rasch模型,从考生、评分员、评分标准三个维度,对新构建的读写结合写作任务思辨能力评分标准进行效度验证。通过该模型研究发现,包含清晰性、相关性、全面性、逻辑性和灵活性5个评分维度的评分标准能够有效区分不同思辨能力的受试,评分员的一致性和内部稳定性在可接受的范围之内,且分值的使用较为理想,模型拟合比较理想,评分结果是可靠的。评分员对逻辑性和完整性两个评分维度的严厉度把握不稳定,需进一步完善其操作化定义。研究结果对读写结合写作任务中的思辨能力培养与评价具有启示意义。A many-facet Rasch model was used to analyze the validity of critical thinking evaluation in reading-to-writing test from test takers,raters and rating scales.The results showed that,the rating scale with five dimensions,namely,clarity,relevance,completeness,logic and flexibility,could sufficiently and reliably discern the test takers in terms of traits measured;raters could consistently use the rating scale though they showed a significant difference in severity.The results also showed that raters might have inconsistent understanding concerning two of the dimensions in the rating scale,logic and completeness,which needed improving.The findings have pedagogical and evaluating implications for the critical thinking in the reading-to-writing writing task.

关 键 词：多面RASCH模型 读写结合写作任务 思辨能力 评价标准 效度验证 

分 类 号：G63[文化科学—教育学]