Lane-splitting came up in a forum I frequent recently, and someone quoted NAMS on the subject.
NAMS states (p51, and also in this NHTSA document):
There is evidence (Hurt, 1981) that traveling between lanes of stopped or slow-moving cars (i.e., lane splitting) on multiple-lane roads (such as interstate highways) slightly reduces crash frequency compared with staying within the lane and moving with other traffic.
I don't find this in Hurt (I have a copy of the full report), and I don't think they even collected data that would make this kind of conclusion possible.
Can anyone provide a page number or quote from Hurt on lane-splitting?
BTW, I believe the conclusion probably is true, and as a long-time lane-splitter in California I feel safer when splitting (and taking the appropriate precautions) than risking a rear-ender in stop-and-go traffic. But as a stickler for precision, I would like to see the proof.
A superior rider uses superior judgment to avoid problems that would demand his superior skill.