EVALUATING PERFORMANCE OF SURVIVAL REGRESSION MODELS WITH INTERVAL CENSORED DATA IN MOTORVEHICLE CRASH EXPERIMENTS