There's nothing contradictory about it. They both claim that rear-end collisions increase when red light cameras go up. The only difference is in the authors' attitudes towards the cameras. As for the choice you present, we can reduce both without the red light cameras. You can have your cake and eat it too. I refer you to posts #6 and 48 in this thread.
I beg to disagree. They contradict in that the Canadian study shows that collisions of all types increase whereas the NC study shows that only rear collisions increase and that injuries decrease.