> I'm not sure why you think it would be more expensive to study the safety. Because you'd have to run a much larger experiment. The end point you're measuring

> I'm not sure why you think it would be more expensive to study the safety. Because you'd have to run a much larger experiment. The end point you're measuring (RTAs or possibly deaths depending) is significantly rarer than just running a light. But you're still trying to collect enough data to achieve confidence, so you'll need a larger experiment. My very strong impression is that your priorities are messed up and what you're looking for isn't safety at all, so that no matter what studies are done you'll find another excuse as to why enforcement isn't a good idea, at which point of course any further study is a waste of their time. For example you suggest in this comment that somehow police data would be good enough, and Chicago has police data showing the same thing, but no doubt you'll say you have a different reason why Chicago's police data doesn't satisfy you.