标准参照测验及格线设定研究中的模拟实验法  

Simulation Experiment Method in the Research of Standard Setting for Criterion-referenced Tests

作　　者：汪存友[1] 余嘉元[1] 

机构地区：[1]南京师范大学心理系,南京210097

出　　处：《心理学探新》2009年第2期81-85,共5页Psychological Exploration

基　　金：国家医学考试中心2007年科研基金项目(KY200702)

摘　　要：针对标准参照测验及格线设定研究中存在的问题,该文首先分析了Monte Carlo模拟实验法的基本原理,然后提出及格线设定研究的模拟实验思路,即:将专家的主观判断视为概率事件,对专家判断过程中的各种误差的概率分布做出合理假设,依据项目反应理论模型模拟得到专家的主观判断,然后采用重复抽样手段获得专家设定及格线的误差分布,据此衡量及格线的修复能力。文中给出了实例演示,文末讨论了模拟实验法的优点与不足,并展望了后续研究。In order to solve problems occurring in standard setting for Criterion-referenced Tests,the principle of Monte Carlo simulation experiment method was first analyzed here,and then, a new simulation experiment method for standard setting was proposed, which can be described like this：by considering judges＇ subjective judgments as a probability occurrence,errors influencing judgments can be assumed to appear like certain probubility distribution,then according to Item Response Theory models, simulated judgments can be generated; next, by sampling repeatedly, the distribution of standard setting error could be calculated, and until now, the ability of recovering the true standard can be evaluated.An example was given here to demonstrate the simulation experiment method.After discussing beth advantages and disadvantages,the paper ended with advices for future researches.

关 键 词：标准参照测验 及格线设定 MONTE Carlo 模拟实验 

分 类 号：B841.2[哲学宗教—基础心理学]