View Single Post
Old July 17, 2013, 03:27 PM   #116
sigcurious
Senior Member
 
Join Date: May 25, 2011
Posts: 1,755
Well those certainly support a correlation, but not causation and certainly not your assertion that in the 70's/80's over half of crashes were due to intoxication.

However, based on the NJ one(the only one I found a copy of not behind a paywall), the correlation is pretty tenuous. First they select a factor which isn't actually DUI crashes(seems like it would be easy enough to actually use those numbers...), but only believed(no evidence offered as to why) a valid indicator, and then attempt to parse a result from a general trend based on that factor(single car nighttime crashes). While they attempt to compensate for the the educational program factors, they ignore a number of other factors, like changes in law etc that also occurred within a relevant timeframe.
sigcurious is offline  
 
Page generated in 0.04753 seconds with 7 queries