ceg wrote:
On Sun, 16 Aug 2015 11:49:26 -0400, Scott Dorsey wrote:

It's hard to get good data, though, when there are just so many
different inputs into the system.

The accident data for the USA is as reliable as any data you'll ever get,
particularly because the police report it, the insurance companies report
it, and in many states (such as mine), both individuals involved in even
a minor accident are required to report it.

Reliable but not very complete. How many accidents were caused by
distracted driving? How many were not caused by distracted driving?
How many accidents would have happened if cars didn't have ABS? How
many additional accidents happened only because cars had ABS? How
many accidents would have been avoided if drivers had been able to
see past the enlarged rear pillars on newer cars?

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.
--scott



--
"C'est un Nagra. C'est suisse, et tres, tres precis."