Comment by lolinder
What you are describing has a major sampling bias: most pedestrian fatalities will be at large intersections with many lanes crossing each other. Those intersections are on busy streets where drivers are going fast and where there are an insane number of conflict points. Yes, they're invariably controlled by a signal, but that's because a four-way stop is totally out of the question. The signal didn't cause the fatalities, it was necessary to install it because of the same factors that lead to fatalities.
Using that data doesn't remotely begin to predict what happens when you take a small four-way stop and add a signal to control it. Adding a signal does not create new conflict points, it does not increase the speed limit on the road, all it does is control the intersection in a more aggressive way.
> What you are describing has a major sampling bias: most pedestrian fatalities will be at large intersections with many lanes crossing each other. Those intersections are on busy streets where drivers are going fast and where there are an insane number of conflict points.
That's not what the point plot of the Toronto data shows. Many of our fatalities are on city streets with 40 or 50 km/h speed limits.
Anyway, I was responding to the OP who was claiming that they would rather deal with stop lights than 4 way stops. There is nothing that shows that 4 way stops are dangerous at all, let alone more dangerous than light controlled stops in similar situations.