From Science 20 June 2008 (subscription needed)
When asked which improvement in fuel economy will reduce gas usage most, Americans using miles per gallon make more mistakes than those given the same numbers in gallons per hundred miles, GPM. (Europeans already use liters per hundred km.)
In one test, one of two cars, each driven 10,000 miles per year, will be replaced. Does it make sense to replace the 15 mpg car with a 19 mpg model, or upgrade the 34 mpg car to a 44 mpg version? (Yes, it makes most sense to upgrade the 15 mpg model to 44 mpg, or to drive the car with worse fuel economy less.) Three quarters recommended upgrading the 34 mpg car.
When also given the gallons per hundred miles figure (upgrade from 6.67 to 5.26 GPM? or from 2.94 to 2.27 GPM?), 64% chose the first upgrade.
Since that saves more than 140 gallons/year, while the second upgrade saves less than 70 gallons/year, it appears that shifting to GPM terminology helps make the discussion more understandable to Americans.
A more dramatic example: upgrading from 12 mpg (8.33 gpm) to 14 mpg (7.14 gpm) reduces gas consumption more than upgrading from 28 mpg (3.57 gpm) to 40 mpg (2.5 gpm).
The use per year is grossly wrong! It would be hard to use 14,000 gallons while only driving 10,000 mile unless you only got .14 miles per gallon, much less save that much!
Oops, corrected, thanks for catching this. In the article, the savings was for 100 cars.