摘要
目的:探讨隐变量非线性效应分析的建模方法及其在SAS软件上的实现。方法:阐述隐变量非线性效应建模的原理,通过拟合一个非线性效应分析示例,对加入非线性项的统计分析结果与无非线性项的统计分析结果进行对比。结果:示例中非线性项引起的变异在结果变量的总变异占到63.99%。结论:在应用结构方程模型分析隐变量产生的效应时,有时候不进行隐变量非线性效应的分析可能会产生偏倚。
Object: Exploring the method of modeling latent variable nonlinear effect analysis in structural equation modeling and the technique of latent variable nonlinear effect analysis by SAS software. Method: Explicating the basic theory of modeling latent variable nonlinear effect analysis in structural equation modeling, then demonstrating an artificial example of latent variable nonlinear effect analysis, comparing the statistical result of the model including quadratic with that of no quadratic model. Results: The given example showed that the quadratic could lead to 63.99% variance of the criterion variable. Conclusion: It may give rise to bias not to complement nonlinear effect analysis in SEM.
出处
《数理医药学杂志》
2005年第3期193-195,共3页
Journal of Mathematical Medicine
