User:Jimad/Sandbox

The issue of the correlation of Traffic Safety vs. MPG had been contentious for many years. The below plots demonstrate these correlations for readers who want to consider these issues for themselves. The plots are based on EPA gas mileage data 2004 vs. Insurance Institute for Highway Safety data on Driver Death rates per each vehicle model in model years 2001-2004. The IIHS Data in turn is based on the NHTSA FARS database. The Insurance Institute for Highway Safety gives permission to republish their data if acknowlegement is given [and as such is given here]. Click on each plot for more details.

This plot shows the variation by vehicle model in Driver Death rates vs MPG for vehicles involved in multiple car crashes. Each grey dot represents one engineering design tradeoff in vehicle safety vs MPG by one vehicle design team. For example one of these grey dots represents the Ford Taurus design model, which in turn has about a million individual vehicles on the road. The scattering of these dots does not represent random variation. Rather, the scattering represents the wide variation in engineering design tradeoffs between vehicle safety and MPG chosen by differing engineering design teams. Clearly, there are more issues involved here than just safety and MPG. For example, one engineering team may choose to design a four door sedan, another a two door pickup, both vehicles could have the same MPG, but greatly varying safety performance. However, even comparing two similar four door sedans, with similar gas mileage, one can easily see a 10X variability in design safety. Crash statistics are often analyzed on a Per Driver basis, rather than on a Per Passenger basis, since each crashed vehicle has exactly one driver, but may have varying numbers of passengers. The linear regression trendline increases with increasing MPG, showing how heavier vehicles with lower MPG tend to transfer risk to lighter vehicles with higher MPG.

This plot shows the variation by vehicle model in Driver Death rates vs MPG for vehicles involved in single car crashes. Historically, researchers have believed that low MPG vehicles fare better in single car crashes by "mowing down" the objects that they crash into. For example, a large pickup truck with low MPG might be able to crash right through a small tree, whereas a small sedan with high MPG would be stopped. This plot shows little or no such effect. What it shows instead is that at any given MPG there is a large range of engineering safety design tradeoffs, with the safest vehicle designs more than 10X less likely to cause a fatality than the least safe engineering designs.

This plot shows the variation by vehicle model in Driver Death rates vs MPG for vehicles involved in rollover fatalities. Historically, researchers have believed that low MPG (bigger) vehicles fare better in regards to rollover in the belief that larger vehicles are more stable because they have a larger stability "footprint". This plot demonstates on the contrary a negative correlation where lower MPG (bigger) vehicles actually roll over more often, or are more likely to suffer roof collapse based on their increased weight. It further demonstates that engineers tend to make bigger vehicles higher (higher center of gravity) faster than they make the track width wider, leading to decreased stability in larger vehicles. For example, a large SUV tends to have a much higher center of gravity than a Mini Cooper. The negative correlation in this plot demonstates how design engineers ought to be able to easily improve rollover safety AND MPG -- there is no design conflict.

By adding together the causes of Driver Deaths shown in the previous three plots we can arrive at an overall plot demonstrating the correlation between Death Rates and MPG. This plot shows little or no correlation between Death Rates and MPG. What it shows instead is the very large, greater than 10X variability in Driver Death depending on the engineering design choices made by different vehicle model design teams. This plot would imply that what matters is the quality of the engineering design, not a vehicle's size or MPG.