An Analysis of Variables to Predict Traffic Fatalities Based on the Characteristics of Road Transport Infrastructure In Indonesia

摘要

The Andreassen's prediction model (1985) with the variables of population and vehicles to predict actual fatalities has been used in Indonesia. These variables need to be investigated and re-examined to suit the demographic characteristics of Indonesia with its vast area, large population, the largest number of vehicles among ASEAN countries, long road infrastructure, and poor driver discipline. This study aimed to determine which variables, out of the eight analyzed, namely population, vehicle, area, road length, driver behavior, licensed drivers, mobility, and accessibility, can be used to predict the number of actual fatalities in Indonesia. Employing a criterion-referenced test, linear correlation coefficient between variables, and correlation analysis of each variable with fatalities, the findings show: (1) The variables of 'population', 'vehicle', and 'accessibility' are the selected variables that have strong and significant effects on actual fatality number prediction; (2) The variable of licensed drivers has no positive effect on driver behavior; (3) The variable of poor driver behavior has a negative effect on actual fatality number prediction in Indonesia and to the other variables under research. Based on our research, we conclude that Andreassen model (1985) is not suitable for fatality prediction in Indonesia. It is recommended that road safety researchers in Indonesia develop fatality prediction models by considering the three selected variables in this study; thus, a new fatality prediction model which is in accordance with the characteristics of areas and road transport infrastructures in Indonesia can be obtained. [EH1] (231).