On Mon, 17 Aug 2015 04:35:41 -0700, trader_4 wrote:

" Data are estimated. Year-to-year comparisons should be made with caution."

Yet here he is, doing exactly that, making year to year comparisons
and refusing to even acknowledge this striking disclaimer.

You have a logic problem if you really believe that your entire premise
is that the answer is hidden inside of "estimation error".

I thank you for looking for a solution out of the conundrum, but, you're
not going to find it in accident rate estimation error.

You apparently have no concept of the powers of ten (hint: It's an extra zero or two
or three on the numbers, which no estimation error in the world is going
to hide).,

That your entire premise hinges on the estimation error being so large
as to greatly sway the numbers means you're simply grasping at straws.

I too am looking for *where* the conundrum is solved, but, it's not going
to be in the "estimation errors" of the US figures on year to year accident
rates.