作　　者： ; ;

机构地区： 广州大学教育学院心理学系

出　　处： 《心理科学》 2013年第1期202-208,共7页

摘　　要： 方差分量估计是概化理论的必用技术,但受限于抽样,需要对其变异量进行探讨。采用Monte Carlo数据模拟技术,探讨非正态数据分布对四种方法估计概化理论方差分量变异量的影响。结果表明:(1)不同非正态数据分布下,各种估计方法的"性能"表现出差异性;(2)数据分布对方差分量变异量估计有影响,适合于正态分布的方差分量变异量估计方法不一定适合于非正态分布。 Estimating variance component, which is the essential technique of generalizability theory, is constrained by sampling. Different sampling may cause different estimated variance component. Therefore, estimating the variability of estimated variance components needs to be further explored. The variability of estimated variance components mainly includes standard error and confidence interval. Past studies have these problems. First, it is often the case that some researchers only focused on normal distribution data and neglected non-normal distribution data. In fact, non-normal distribution data could be always seen in such tests as the Test of English as a Foreign Language （TOEFL）. Second, previous studies did not compare the variability of estimated variance components using traditional, bootstrap, jackknife and Markov Chain Monte Carlo method （MCMC） at the same time. This study adopts the Monte Carlo data simulation technique to compare the variability of the estimated variance components of generalizability theory based on three waves of non-normal distribution data using four methods that include the traditional method, the bootstrap method, the jackknife method and the MCMC method. As for the traditional method, ANOVA is used to estimate the variance components, and their standard errors and TBJGLs are used to estimate the confidence intervals. As for the bootstrap method, 12 bootstrap strategies are considered, which include unadjusted bootstrap strategies and adjusted bootstrap strategies. But the jackknife method only considered three strategies, namely, jack-p, jack-i and jack-pi. Moreover, two strategies, informative priors and noninformative priors, are considered in MCMC method. Three waves of non-normal distribution data are simulated using these techniques ： normal distribution data, dichotomous distribution data, polytomous distribution data, and skewed distribution data. To compare these four methods in two variabilities, i.e. , standard error and 80% confidence interval, the criterion

关 键 词： 概化理论 非正态分布 方差分量变异量 蒙特卡洛模拟

领　　域： [文化科学]