摘要
采用模拟实验法比较研究了两种Angoff法——"概率法"和"对错法"——设定分数线的准确性和稳定性,结果表明:(1)当真能力低于测验的平均难度时,"概率法"高估分数线,而"对错法"低估分数线;反之,当真能力高于测验平均难度时,"概率法"低估,而"对错法"高估;(2)当真能力接近测验平均难度时,"概率法"比"对错法"更准确;反之,当真能力远高于或低于测验平均难度时,"对错法"更准确;(3)无论在何种实验条件下,"概率法"均比"对错法"更稳定。
In order to compare both the accuracy and stability of two Angoff methods: probability method and yes/no method, a simulation research was conducted. The research found that 1. When the true ability was less than the average test difficulty, the probability method would overestimate the passing score, while yes/no method would underestimate it; and when the true ability was greater than the average test difficulty, the probability method would underestimate the passing score, while the yes/no method would overestimate it; 3. When the true ability was close to the average test difficulty, the probability method was more accurate than yes/ no method; but when the true ability was far from the average test difficulty, yes/no method was more accurate; 3. In any case, the probability method was more stable than the yes/no method.
出处
《心理科学》
CSSCI
CSCD
北大核心
2010年第1期159-161,共3页
Journal of Psychological Science
关键词
Angoff法
标准设定
模拟研究
分数线
Angoff method
standard setting
simulation research
passing score