After a chat with Vonhosen on Pistonheads:
http://www.pistonheads.com/gassing/topi ... 351199&p=5 (requires registration) regarding this effect:
http://www.safespeed.org.uk/problem2.html it seems to me that there should by now be a smoking gun in the crash stats.
The essence of the hypothesis is that 'speed' is a stressor that promotes experience learning. With high emphasis on speed controls there should be damage to the acquistion of experience.
So if we look at crash rate changes with age in the 1980s, we should find that crashes fell faster with growing experience compared with this decade.
I'm off to see if I can find any suitable crash stats by age to start an investigation.