摘要
目的探讨如何利用均值结构模型实现潜变量数据的分组比较。方法结合抑郁症病例-对照临床研究量表数据,利用LISREL软件实现参数估计,并以Δχ2(Δd)、RMSEA、NNFI、CFI等拟合优度指标评价拟合效果。结果社会支持量表的验证性因子分析模型拟合优度不理想,但其均值结构模型拟合效果有所改善;特质应对方式问卷模型拟合优度检验比较理想。结论采用均值结构模型进行抑郁症病例-对照分组比较,不仅获得量表结构效度及关联性结果,而且得到病例组与对照组潜变量均值的差异,结果符合专业解释。
Objective Discussed the method of multi - group comparison with mean structure equation models for latent variable data. Methods Depression case-control data were analyzed and using LISREL to estimate parameter,△x^2 （ △d）, RMSEA, NNFI, CFI were used as fitting goodness indicator. Results The goodness of fit about CFA mean strueture model of social support scale is not very good, but mean structure models obtained a more meaningful measure of goodness of fit. The goodness of fit about mean structure model for TCSQ is perfect. Conclusion Mean structural model not only obtained the results of structural validity and relevancy, but also obtained the mean differences of latent variables.
出处
《中国卫生统计》
CSCD
北大核心
2009年第4期352-354,共3页
Chinese Journal of Health Statistics
基金
山西省自然科学基金资助项目
山西省创新拔尖人才基金项目
