ACCIDENT RATIOS ARE ANALYSED WITH REGARD TO THE VARIABLES ROAD-SURFACE SKIDDING RESISTANCE AND HOURLY TRAFFIC VOLUME. IN A FIRST ANALYSIS THE ADDITIVE CONJOINT MEASUREMENT MODEL (ACM) IS USED TO INVESTIGATE TO WHAT EXTENT THE ACCIDENT RATIOS CAN BE DESCRIBED AS A RESULT OF INDEPENDENT CONTRIBUTIONS OF SKIDDING RESISTANCE AND TRAFFIC VOLUME. FURTHERMORE IT IS CONSIDERED WHETHER THESE CONTRIBUTIONS HAVE TO BE COMBINED IN AN ADDITIVE OR MULTIPLICATIVE WAY. BASED ON THE RESULTS OF THIS INVESTIGATION A SECOND ANALYSIS TAKES PLACE IN WHICH A STOCHASTIC INTERPRETATION OF THE DATA IS COMBINED WITH THE MULTIPLICATIVE MODEL. THIS WEIGHTED POISSON MODEL (WPM) IS IN FACT A GENERALISATION OF THE LOG-LINEAR MODEL, RECENTLY PROPOSED FOR THE ANALYSIS OF CONTINGENCY TABLES. IT IS CONCLUDED THAT THE MULTIPLICATIVE MODEL DESCRIBES THE DATA BETTER THAN THE ADDITIVE MODEL. MOREOVER, THERE IS NO INTERACTION BETWEEN SKIDDING RESISTANCE AND TRAFFIC VOLUME IN THEIR EFFECT ON ACCIDENT RATIOS. THE RELATION BETWEEN ACCIDENT RATIOS AND BOTH VARIABLES ARE SHOWN, AND THE STATISTICS REGARDING THE CONTRIBUTIONS OF THE VARIABLES ARE GIVEN.(Author/publisher).
Abstract