Interesting... I'm curious to know more about the details of that experiment. I could think of a few hypotheses for why that might've happened.

My first thought is maybe that particular road's speed limit was just lower than what it should be? If the road design and location allow for a similar level of safety at a higher speed, the limit itself might just be a bit too conservative.

Or, perhaps the previous cycle timings weren't as effective at slowing down drivers as first thought? I don't know the previous circumstances, though.

There's also the possibility that speed-reducing measures just don't have much of an impact on overall safe driving behaviors, and that dangerous drivers are going to drive dangerously no matter how fast they're able to go. I could even see the possibility of a negative feedback loop in which being constantly being slowed down by red light cycles frustrates aggressive drivers into engaging in other dangerous behaviors, i.e. running reds.

If it's that last case... Yeah, I don't really know what you do with that one. I would hope that safer driving could be encouraged through positive reinforcement, because the alternatives for negative reinforcement are either a massive ramp-up in police traffic enforcement, or aggressive use of speed cameras and automated ticketing. I do not think that making people have more encounters with US police officers would do anything good for public safety, and while speed cameras could work in principle, they would be massively unpopular, and career suicide for any politician that tries to make a push for it.