It seems to me that you can’t conclude that speeding causes accidents from these statistics. If people speed 30% of the time, then accidents will involve speeding 30% of the time if speeding doesn’t influence the risk at all. I would also expect the amount of speeding to matter a lot.
I’ve always had this issue with speeding statistics: the assertion that “speeding causes accidents” usually sounds like a textbook case of base rate neglect. Especially considering that nearly everyone in the US drives 5-15 mph above the posted speed limit.
You can't out statistic physics. Energy goes up with velocity squared. A small reduction in velocity means a lot less energy involved which delivers better outcomes for all involved.
But on a freeway accidents generally happen between cars traveling in the same direction, so the effective velocity is much lower than the velocity of travel.
