That's not a refutation, it's a contradiction. The evidence offered is not sufficient to actually prove the assertions. In particular, those crash frequencies need to be normalized against the reduced capacity in order to be a meaningful comparison. If narrowing the road decreases crashes because fewer people are using the road, you don't have a safer road, just less road, and you may have merely shifted the crashes to alternative routes. On the other hand, if narrowing the lanes makes people drive more carefully without significantly constricting flow, that's a real result that deserves to be stated clearly.
I chose my words carefully. The quote in the article is in turn a quote from http://trb.metapress.com/content/x7854w1160551331/ . The article links to that publication. The abstract of that publication says:
> This research investigated the relationship between lane width and safety for roadway segments and intersection approaches on urban and suburban arterials. The research found no general indication that the use of lanes narrower than 3.6 m (12 ft) on urban and suburban arterials increases crash frequencies. This finding suggests that geometric design policies should provide substantial flexibility for use of lane widths narrower than 3.6 m (12 ft). The inconsistent results suggested increased crash frequencies with narrower lanes in three specific design situations. Narrower lanes should be used cautiously in these three situations unless local experience indicates otherwise.
It it turn builds on, for example, results by Hauer, et al, Strathman et al. which appear.
You then raised another objection, which is, I believe, that a 10' lane causes people to use alternate routes because of decreased capacity on those lanes, so there are simply fewer people on the 10' lane roads to cause accidents.
This may well be. It's a subtle network effect that is hard to analyze, and not covered in this article. (The article does comment that capacity is unchanged, but I think its literature citations are weak. It quotes Petritsch who in turn quotes a summary of an unpublished literature search.)
However, your objection is not what was refutated. AnthonyMouse proposes that narrowing lanes lead to a higher accident rate since it "reduces the amount of space available to avoid an obstruction, the amount of reaction time available to avoid a collision, etc." While true for country roads, those above papers show that the same correlation can not be identified on city roads.
Now, what I know is only from this article, and it may be that the author cherry-picked the few papers which show that the 'reaction time' hypothesis is unsupported by the evidence. But "refute" means "to deny the accuracy or truth of", and certainly the article refuted that hypothesis.
