If you want to try having your insurance rates based on GPS data

Scott en Aztl‡n

Expanding on Scott's idea, data on crashes involving "speed" are suspect.

Using the 85th percentile speed to establish a posted speed implies that 15% are guilty of speeding. But what doesn't show in the statistics is the conditioned probability of that 15%. Given the condition that 15% of motorists are speeding, how does their crash history stack up against the 85% that are legal? Even if the 15% represent 15% or fewer of the crashes, do they represent 15% or less of the cost of crashes.

Or are "speeding" crashes more involved with some confounding factor such as drunk-drugged, running from the police, specific locations, lack of road friction? for instance, school zones? pedestrian conflicts, polished or contaminated surfaces, poor skills in matching speed to road conditions?

In a similar vein, 15% of drivers don't buckle up. What are the statistics regarding the crash rate of that 15%? Are they over represented in the number of crashes? Is the correlation one of atbreastude in that they are careless and don't respect the seriousness of driving? Given the condition that someone is not belted and is in a crash, the damages are usually much greater.

One factor of speeding that does not get much discussion is air pollution. Existing federal test standards are for max speeds of about 55 MPH. Yet at higher speeds, combustion chamber pressures and temperatures are such as to produce more NOx. Are the pollution abatement devices good enough to keep a "speeding" car from polluting?