You might think that having cameras at traffic intersections to automatically ticket cars that run red lights would increase safety. But the author concludes that this doesn't increase safety. Why not? Although she concedes that the cameras would reduce side-impact collisions, she points out that they would result in a lot more rear-impact collisions.

This argument boils down to the net effect of the cameras on traffic safety. The author thinks the net effect won't be an increase in safety, because we'll get more rear-impact collisions, which are presumably dangerous.

One common way to weaken arguments based on net effect is to point out that the author might be overlooking factors that cut against the author's conclusion. For example, cameras at traffic intersections might have other positive effects on traffic safety, such as decreasing the speed of cars that get into collisions, which might decrease the severity of injuries.

Another way to weaken arguments based on net effect is to amplify/minimize the weight we should give to factors cited by the author. For example, we know there will be a lot more rear-impact collisions, which the author assumes will outweigh any safety increase that would result from having fewer side-impact collisions. But what if that assumption isn't true? Rear-impact collisions might not involve much danger (think fender-benders), whereas side-impact collisions might be far more dangerous. If so, then the net effect of the cameras can still be an increase in traffic safety.