基于IRT框架采用Angoff法进行合格标准设置的探索  

Exploration of Qualified Standard Setting with Angoff Method Based on IRT Framework

作　　者：杨观惠 王晓慧 Yang Guanhui;Wang Xiaohui(Shanghai Municipal Educational Examinations Authority,Shanghai,200433)

机构地区：[1]上海市教育考试院,上海200433

出　　处：《考试研究》2023年第4期59-66,共8页Examinations Research

摘　　要：学业水平考试分界分数的划定关系到教育决策的科学性和公平性,涉及标准设定技术。为探究在项目反应理论(IRT)框架下采用Angoff法进行合格标准设置,以某地区初中学业水平考试某学科的模拟考数据为分析对象,将专家第三轮打分结果作为“最低合格能力考生”的作答数据,与真实考生的数据汇总,进行IRT参数估计。最终以“最低合格能力考生”能力分数的平均值-0.85作为合格的分界分数,该合格分数线的信息量为28.13,专家判断与实际结果的Pearson相关系数为0.9,表明采用该方法进行标准设定具有较高的质量和推广性。Setting cut-off scores for the academic proficiency test is related to the scientificity and fairness of the educational decision-making,which involves standard setting techniques.This study explored the adoption of Angoff method for the qualified standard setting according to Item Response Theory(IRT),taking the mock test data of a subject in the junior high school proficiency test in a certain area as an example.The expert's third round scoring results were used as the response data of the"minimum qualified candidates",and the data of real candidates were put together for parameter estimation by R.The average of the ability score of the"minimum qualified candidates"is-0.85.Taking it as the qualified cut score,the information content of this point is 28.13.Additionally,the Pearson correlation coefficient between expert judgment and actual result was 0.9.The results showed that this method had high quality in standard setting.

关 键 词：标准设定 分界分数 合格分数线 Angoff法 学业水平考试 IRT 

分 类 号：G424.74[文化科学—课程与教学论]