【摘要】 以大规模学业成就水平测验为背景,采用组间设计,按类别将专家分为五组,在反馈数据环节随机给予每个专家组未调整的考生真实作答数据和经过上调、下调0.5或1个标准差的调整数据后,采用单因素方差分析与项目反应理论两参数模型考查专家组在标准设定Angoff方法中如何参考使用题目作答数据。结果表明,反馈考生题目作答数据对Angoff标准划定结果有明显影响:反馈未经调整的真实考生题目作答数据影响较大;反馈调整后的高于真实考生题目作答数据影响相对较小,反馈低于真实考生题目作答数据影响相对较大。更多还原

【Abstract】 Based on the large- scale academic test,the study used between group design,divided the experts into five groups,and provided them with the examinee true performance data,up or down 0. 5 or 1 standard deviations of the true data. At the same time,the study used one- way ANOVA and two parameter model to test that how the experts read the performance data to set standard in the Angoff method. The results show that the feedback of the data had the significant influence on the setting of the qualified line and the excellent line; the feedback with unadjusted and true performance data had great influence on the standard setting. Compared with that,the feedback with adjusted and above the true performance data had smaller influence on the standard setting,and the feedback with adjusted and below the true performance data had greater influence on the standard setting.更多还原