On Sun, 16 Aug 2015 16:20:57 -0400, Scott Dorsey wrote:

All we have data on are accidents..... we have no data at all on
accidents that didn't happen but would have under other circumstances.
And the data we do have aren't enough to tell us about what caused all
the accidents there were. This is what I mean by there being so many
different inputs.

Yes. All we have that is reliable is the data on *all* accidents, state
by state, and those are going down, year after year.

There isn't even a blip for the years that cellphones were starting to be
used. It's the same declining accident rate (give or take a few) with no
visible effect from cellphone use.

Hence the paradox.

I believe that if a huge number (essentially 100% of the drivers in the
USA) *own* a cellphone, then a certain percentage of those people will be
*using* that cellphone while driving, and a certain percentage of those
users will be *distracted* enough to cause accidents.

Since the numbers are so huge, and the numbers of accidents are so
constant, you'd expect a huge increase in the number of accidents, or, if
not huge, at least discernible.

But there is no increase.
Accidents are steadily going down.

Hence the paradox.
Where are the accidents?