[aetherson]

Obviously it's hard to judge studies by one sentence summaries, but some of those papers sound like extraordinary claims. "Speeds declined by 5 km/h. Fatal crashes declined by 12%." Come on, really? We're going to imagine that a 3mph change in speed (at speeds around 80mph) could possibly be causal to a 12% change in fatalities?

[cpncrunch]

I haven't looked at the refs, but you're welcome to. I suspect wikipedia is more reliable than priceonomics.com, although you have to look at the sources yourself to be sure.