Quote:
The third analysis looks at the 16,000 fatal head on collisions of cars of two different model years on the Fatal Accident Reporting System (FARS) to see in each collision which driver is more likely to be killed - the one in the older car or in the newer car - taking into account such other factors such as the difference in vehicle weights, the drivers' ages, etc. The individual comparisons are combined into a model which predicts the unrestrained driver's fatality risk index as a function of model year, controlling for vehicle weight - and the decrease of this index from model year 1964 to 1984 estimates the cumulative reduction in frontal fatality risk, as a result of vehicle modifications (other than weight changes) during those years. The model is then extended to right front passengers. This approach using head on collisions eliminates most of the sources of bias that have often been present in earlier analyses to estimate fatality risk by model year: reporting biases, effects of factors other than vehicle modifications. When cars of two different model years collide head on, but with the same car weight, driver age, etc. and the fatalities occur consistently more often in the older car than in the newer one, the only conclusion is that the newer car is safer.
