A Regression Model for Predicting Automobile Crashworthiness.

A stepwise regression analysis of automobile crashworthiness is developed. A sample of four hundred eleven Texas accidents were used in which the drive mix was held constant and all other factors were randomized. Nine predictor variables length, width, weight, height, steering ratio, wheelbase, wheel stance, brake horsepower and brake area were considered. Using regression analysis, a final equation utilizing these independent variables and explaining eleven percent of the variance in the dependent variable, was obtained for car body damage cost.