The robustness of test statistics to nonnormality and specification error in confirmatory factor analysis
The robustness of test statistics to nonnormality and specification error in confirmatory factor analysis
复制标题
DOI：
10.1037//1082-989x.1.1.16
复制
发表时间：
1996-03-01
影响因子：
7
通讯作者：
Finch, JF
中科院分区：
文献类型：
--
作者：
Curran, PJ;West, SG;Finch, JF
采用MonteCarlo计算机模拟的方法研究了三种XY检验统计量在验证性因素分析中的应用。对于适当指定的模型，在不同的样本量、模型规格和多变量分布条件下检查了正态理论最大似然chi（2）（ML）、Browne渐近分布自由chi（2）（ADF）和Satorra-Bentler重新标度的psi（2）（SB）。ML和SB在所有样本量的正态分布下均未显示偏倚证据，而ADF在所有但最大样本量下均存在偏倚。随着非正态性的增加，ML被越来越高估，但SB（在所有样本量下）和ADF（仅在大样本量下）均未显示偏倚证据。对于错误指定的模型，ML再次膨胀，增加非正态性，但SB和ADF被低估，增加非正态性。看来SB和ADF检验统计量检测模型误设的能力在给定非正态分布数据的情况下会减弱。
Monte Carlo computer simulations were used to investigate the performance of three XY test statistics in confirmatory factor analysis (CFA). Normal theory maximum likelihood chi(2) (ML), Browne's asymptotic distribution free chi(2) (ADF), and the Satorra-Bentler rescaled psi(2) (SB) were examined under varying conditions of sample size, model specification, and multivariate distribution, For properly specified models. ML and SB showed no evidence of bias under normal distributions across ail sample sizes, whereas ADF was biased at ail but the largest sample sizes. ML was increasingly overestimated with increasing nonnormality, but both SB (at all sample sizes) and ADF (only at large sample sizes) showed no evidence of bias. For misspecified models, ML was again inflated with increasing nonnormality, but both SB and ADF were underestimated with increasing nonnormality. It appears that the power of the SB and ADF test statistics to detect a model misspecification is attenuated given nonnormally distributed data.