not sure the data you want for #5 is available anywhere or even tracked, since you would have to consider the speeds of all participants involved. I would argue that the given statistics might give an even stronger case because the speed-limited sections usually already have a higher probability of accidents (hence the limit), so being above that already says something.

But having written that, I have no drivers license and no real opinion either way. ofc I'm slightly leaning towards general speed limits (less externalities for me to pay) but overall I'm indifferent.