Not quite, I think. But I was considering, as were the reviewers, reduction in crash rate produced by each kind of facility. You are considering more the existing crash rates. Yes, sidewalks and multi-use trails have high crash rates, major roads have a somewhat higher crash rate than do minor roads. Off-road bike paths are a mixed lot; what distinguishes them from multiple-use trails? And what distinguishes paths far from the roadway from paths alongside the roadway, which latter have very high crash rates, because they have the same topological relationship to the roadway as do sidewalks? And streets to be designated as bike routes have already been selected as safer than usual. And, nobody really expects that putting up bike route signs, which is the typical treatment, actually reduces crashes, though new traffic signals at some intersections would help. Much the same goes for streets on which bike lanes get built; mostly on streets that are already sufficiently wide for them. The reviewers did not find, and I know of none, studies that specifically did either before and after, or comparison of otherwise identical streets, to determine if it was the stripe itself that reduced crashes, and came up with the conclusion that the bike-lane stripe reduced crashes.