E shows that the US Department of Transportation's data is biased towards traffic fatalities that occur at speeds of under 50 miles per hour and within 25 miles of home.
Let's take a number example. Total population under study is 10000.
Say you have 9000 who drive at a speed under 50 miles per hour and who live within 25 miles of home. --> Category 1
Then 1000 people under study do not fall under this category because they either drive faster than 50miles/hr or they live beyond 25 miles of home. --> Category 2
The above 2 categories are consistent with E which claims that 90% of all driving time is registered within a 25-miles radius of the driver's home and at speeds less than 50 miles per hour
If we say that there are 100 accidents and 80% of those accidents come from category 1, as the argument claims, then 80/9000 = .88% likelihood of having an accident.
Remainder 20%, 20 persons, will represent a likelihood of accident of: 20/1000 = 2% for category 2. Thus, does the conclusion that you are safer in a car if you are driving at a speed over 50 miles per hour and not within a 25-mile radius of your home (category 2) stand? The proof that there are more data "logged" under category 1, as E says, will clearly weaken the above conclusion