Originally Posted by
mr_bill
FARS and NASS have merged into GES.
To start.
"In order for a crash to be eligible for the GES sample a police accident report (PAR) must be completed, it must involve at least one motor vehicle traveling on a traffic way, and the result must be property damage, injury, or death."
Follow up things to consider. What is threshold of property damage value where a PAR must be completed? (Varies by jurisdiction.) What is the threshold of injury where a PAR must be completed? (Varies by jurisdiction.)
What is injury that rolls up into the report that you cited? (It excludes possible injury, and believe it or not a "momentary loss of consciousness" and "limps" are possible injuries.)
So, how much does that effect your numerator?
-mr. bill
For a list of the data files
http://www-nrd.nhtsa.dot.gov/Pubs/811853.pdf, and you will find the answer you seek.
It's clear that you have no credible objection to the math since you're scanning through the web sites looking for something to object to, but not enough to check your objections. It's not my intent to argue with you whether the National Highway Transportation Safety Agency knows what they're doing in gathering statistics. I've identified the sources of their data sufficiently to know that it's applicable.
The numerator, as you put it, is fine.
Highway traffic statistics involve road traffic. We haven't made any bones about that.
But, if you want to make
sure that doorings are included at least once, feel free to multiply the number of accidents by 1.008 from the Cross-Fisher study. It won't change the answers given the low degree of precision.