Analyzing Vigilance Rate of Motor Vehicle Driver Using Regression and Structural Equation Modeling (SEM)

Abstract

This study focuses on the impact of risky driving behavior, monotonous road and fatigue factor on vigilance of motor vehicle drivers. Accident case growth in Indonesia each year were 3.3%. Therefore it need to be conducted study to minimize the risk of accidents. The samples consist of 100 respondents. And then the data were analyzed using regression and Structural Equation Modeling (SEM). Regression analysis shows that latent variable risky driving behavior, latent variable monotonous road and latent variable fatigue can explain latent variable vigilance by 57.6%. Meanwhile, the result  of SEM analysis show that latent variable risky driving behavior, and the latent variable fatigue can explain the latent variable vigilance by 75.3%. The value R square of SEM analysis are higher than regression analysis. There are several cause of differences of R square between regression and SEM analysis, namely: 1) the multicollinearity is not allowed in regression analysis, while it is allowed in the SEM analysis; 2) there is no latent variable in the regression analysis, while there are latent variables and indicators of latent variables in the SEM analysis; 3) the regression analysis is explanatory, while the SEM analysis is confirmatory.