Accident involvement rate and car size

Abstract In order to determine if accident involvement rate depends on car size, accidents per unit distance of travel were examined as a function of car mass using police reported crashes for North Carolina. When disaggregated by driver age, accident involvement rate is found to increase with car mass. Without such disaggregation, the trend is in the opposite direction because of a confounding relationship between driver age and car mass. Similar effects are found using data from two other states (New York and Michigan). All nine subsets of data examined (three states times three age groups) consistently indicate that accident involvement rates are lower for small cars than they are for larger cars driven by drivers of similar age. For example, the rate for 900 kg cars is estimated to be, on average, 28% lower than for 1800 kg cars. This is interpreted to be due to driver behavior changes related to how drivers perceive protection to vary with car size. The result fits, both qualitatively and quantitatively, into an increasingly complete description of interactions in the complex traffic system that has an one output total national traffic fatalities. For example, combining the present result that the accident involvement rate for a 900 kg car is 0.72 that for a 1800 kg car with an earlier result that given a single car crash, the driver of a 900 kg car is 2.6 times as likely to be killed as is a driver of a 1800 kg car gives 0.72 × 2.6 = 1.9 as an estimate of the number of driver fatalities per registered car in 900 kg cars to those in 1800 kg cars. This estimate may be compared to the previously reported directly estimated value of this ratio of 1.7.