The phenomenon of health data presentation in the central statistical agency or health department, still presented by deskripitif statistical form. Health research is still limited by correlation analysis. Theories and models in the health departement generally can be formulating using which construct can’t be measured or observed directly, so there is variables unobserved and variables observed. Usually, the analysis used is multiple regression analysis. This analysis only observed the direct influence on the response variable. Structural equation modeling is an alternative option to solve the problem if the research variables is a unobserved variable. This research purpose use validity and reliability of the survey data that influence health status and the determination of the structural model and the goodness of fit of the survey data factors that affect health status. This research is non-reactive research by use of secondary data, that is data of East Java health department. Validity and reliability of the indicators can measure constructs significantly manifest latent exogenous variables and endogenous latent variable with loading factor / λ> 0.5 and construct reliability values> 0.6. Causality interaction variables formed by latent product is health care, health behaviors and health resources that influence the health status line parameter coefficient> 0.5, p-value <0.05 at significance level α = 5%. Result of this research, expected to contribute for the development of science, such as : (1) considerations to setting health policy and health promotion to preventive & improve health status (2) Guidance to creating health modeling.Keywords : structural equation modeling, health status.
Copyrights © 2013