目的通过探讨结构方程模型等同性检验方法在分组比较中的应用,说明等同性检验的必要性和重要性。方法结合抑郁症病例-对照临床研究中特质应对方式量表数据,利用LISREL软件实现各种等同性检验及参数估计。结果通过等同性考察,证实TCSQ量表在病例-对照研究中具有等同性。结论对量表测量数据在进行分组比较时,首先须进行各种等同性检验,并在此基础上,进一步进行潜变量均值结构的分析,在此前提下研究结果更可靠。
Objective To illustrate the necessity and importance of measurement invariance test for structure equation model in multiple comparison studies.Methods Trait Coping Style Questionnaire(TCSQ) data in depression case-control study were analyzed and LISREL was used to test invariance and estimate parameter.Results Measurement invariance test verified that TCSQ questionnaire had invariance in case and control group of depression data.Conclusion Measurement invariance should be tested first,then analyzed latent mean structure in multi-group comparison for questionnaire data.The results were more reliable in this precondition.