R-RUM的参数估计及性能评价研究

认知诊断作为21世纪一种新的测量范式,在国内外越来越受到重视。该文运用MCMC算法实现了R-RUM的参数估计,并采用Monte Carlo模拟方法探讨其性能。研究结果表明:(1)R-RUM参数估计方法可行,估计精度较高;(2)Q矩阵复杂性和模型参数水平对模型参数估计精度有较大影响,随着r_(jk)*值的增大和Q矩阵复杂性的增加,项目参数和被试参数估计精度逐渐下降;(3)在特定情形下,R-RUM具有一定的稳健性。

A Research on Parameter Estimation and the Properties of R-RUM

Cognitive diagnosis(CD)is the new testing paradigm in the 21st century.Recently it has attracted increasing importance internationally.The Reduced Reparameterized Unified Model(R-RUM)is the reduced version of the Reparameterized Unified Model that is a cognitive diagnostic model developed by Hartz(2002).The paper introduces the MCMC method with the higher-order model given by Jimmy & Douglas(2004)and discusses its application in the parameter estimation of R-RUM.Research shows that:(1)The estimation method of MCMC algorithm holds fairly robustness,and its precision of items and ability parameters are preferably great,which indicates the MCMC algorithm method is feasible.(2)The complexity of the Q-matrix and the level of item parameters influence the estimation of items and ability parameters greatly.Its precision of items and ability parameters decrease as the complexity of the Q-matrix and the value of item parameters increase.(3)The R-RUM has robustness under special circumstances.