高考历史试卷的多元概化理论研究

A Multivariate Generalizability Study on the History Test of College Entrance Examination

运用多元概化理论及其方法,本研究分析了上海市高考历史试卷(2006年),结果发现:(1)在试卷结构,各部分试题具有较好的测量信度,但在赋分比例存在问题;(2)在主观题评分,评阅者对分数影响微弱,说明分数能较为真实地反映不同考生之间的学科知识和能力差异。此外,研究还为将来提高高考命题和评阅质量提供合理化建议。

Our multivariate generalizability study on the history test of the college entrance examination in Shanghai showed that on the one hand, parts of the test had higher reliability, but were poorer in the proportion of scoring; on the other hand, the rater had a weak effect on the scoring, which meant the real scores could explain the competence difference of the examinees. Some suggestions for future college entrance examinations are provided.