高考作文改革与评分误差控制:基于测量学的视角

College Entrance Essay Reform and Scoring Error Control:From the Perspective of Psychometrics

从测量学角度来看,高考作文因其评分主观性强影响了对考生写作能力甚至是语文能力的测量。如何改革作文才能进一步减小评分误差、提高考试的公平性,是落实此次考试招生制度改革的一项具体任务。研究一表明,与西方采用的小评分量表相比,我国高考采用的60分制大评分量表评分趋中效应更为严重,评分标准更为宽松,不同评分者对评分标准的掌握一致性较差,据此建议改革高考作文评分量表的设计,将目前的大评分量表改为小评分量表,成绩单独报告。研究二表明,增加作文任务数量有助于明显提高评分信度,据此建议将高考作文由一个大作文变为一大一小两个作文。

From the perspective of psychometrics, college entrance essay scoring is so subjective that much impact on measure of writing ability, as well as measure of the candidates＇ language ability. In order to further reduce the scoring error and improve the fairness of the examination, essay reform is to implement a specific task in the system reform of the Entrance Examination and Enrollment. The first study shows that, compared with small scoring scale in western countries, large scale such as using 60-point rating in China turns out to be more serious effect of tending toward the middle, scoring criterion is more relaxed, and different raters show poor consistency on scoring. So the first suggestion about essay reform is that scoring scale should be designed to be smaller than present large scale, and report the essay scores independently. The second study shows that increasing the number of writing tasks will improve the reliability of scoring. So it is proposed to design two writing tasks instead only one in college entrance essay.