Among the U.C. Berkeley findings:
* Lane-splitting is safe if done in traffic moving at 50 mph or less, and if motorcyclists do not exceed the speed of other vehicles by more than 15 mph;
* 69 percent of lane-splitting motorcyclists were exceeding the traffic speed by 15 mph or less; speed differentials up to 15 mph were not associated with changes in the frequency of injury;
* Compared to riders who were not splitting lanes, lane-splitting motorcyclists were markedly less likely to suffer head injury (9 percent vs. 17 percent), torso injury (19 percent vs. 29 percent) or fatal injury (1.2 percent vs. 3 percent);
* Lane-splitting riders were significantly less likely to be rear-ended than non-lane-splitting riders (2.6 percent vs. 4.6 percent);
* Lane-splitting motorcyclists were more likely to be wearing a full-face helmet than other motorcyclists (81 percent vs. 67 percent);
* Compared to other motorcyclists, lane-splitting riders were more often riding on weekdays and during commuting hours, were using better helmets and were traveling at slower speeds;
* Lane-splitting riders were less likely to have been using alcohol.
The first one is probably the most important one of the list. The full-face helmet bit is probably important, too.