Samenvatting
THE ANALYSIS OF ACCIDENT COUNT DATA FOR A NUMBER OF AUCKLAND, NEW ZEALAND INTERSECTIONS IS DESCRIBED. THE RESULTS REVEAL CONSIDERABLE VARIATION IN THE VARIABILITY OF ACCIDENT COUNTS, INCONSISTENT WITH THE "POISSON ASSUMPTION", UPON WHICH ACCIDENT DATA ANALYSIS IS GENERALLY BASED. THE PRACTICAL IMPLICATIONS OF THE RESULTS AND A PROCEDURE FOR TAKING PROPER ACCOUNT OF VARIATIONS IN THE VARIABILITY OF ACCIDENT COUNTS ARE DISCUSSED.(Author/publisher).