Generalized Structural Equations Approach in the of Elderly Self-rated Health

Abstract To model classified and ordinal data, the Generalized Structural Equation Model (GSEM), which is based on integration of two generalized linear (GLM) Modeling (SEM) algorithms, applied. Unlike SEM, this does not require normality assumption. The main purpose paper to introduce compare weighted (WLSMV) unweighted (ULSMV) least squares mean variance adjusted methods, most applicable estimators GSEM, for studying factors affecting elderly self-rated health in 2015 Tehran, Iran. 600 people aged 60 years above from 22 regions Tehran were selected using multi-stage sampling. Self-rated variable (a 5-point Likert scale) was analyzed as an modeled considering variables social support, financial environment security, spirituality, mental physical health, functional health-related behaviors by Mplus software. results showed that WLSMV outperformed ULSMV according smaller values RMSEA larger CFI TLI indexes (RMSEA = 0.04, 0.965, 0.936). prevent concluding invalid data due them a continues variable, it important selecting correct statistical method type variables.