Note: Here’s the fourth in a trove of unpublished blog posts I salvaged from an old laptop. The news story referenced is now a few years out of date, but it’s still interesting…at least to me.
In the Vancouver Sun a few weeks ago (April 2013) there was an article about how crosswalk timers – the walk signals at many intersections that count down the number of seconds till the light turns yellow – reduce collisions between vehicles and pedestrians, but increase collisions between vehicles.
Arvind Magesan, the University of Calgary researcher who discovered this phenomenon, theorizes that the extra accidents are caused by drivers speeding up to beat the red light:
“If a road is really busy and it’s slow-moving traffic, you can’t really use a countdown to decide to drive faster to get through a light. In places like that, it seems to have a positive effect. It reduces the rates of accidents,” Magesan said Tuesday.
“In places where a driver does have the opportunity to react to this thing – which they are not supposed to be reacting to in the first place – they use the information and accelerate,” Magesan said.
I never realized that drivers “are not supposed to be reacting to” these timers. When I took driver training, not so very many years ago, the instructor explicitly advised us to keep an eye on crosswalk signals to get a sense of when the light was about to change. The idea was that when we saw the blinking red hand, we should be prepared to slow down. But a more reckless driver will obviously take the red hand as a cue to make haste through the intersection.
It’s not only crosswalk signals that are interpreted in ways not intended by the designers. Recall that scene in the movie Starman, where Jeff Bridges’ stranded alien, having learned to drive by observing Karen Allen, interprets the meaning of the traffic signals as “red light, stop – green light, go – yellow light, go very fast.”
A crosslight timer is just a more explicit blinking red hand, which in turn is a kind of yellow light advance-warning. They all serve the same purpose; to tell us roughly how long until the light turns red. What we do with that information is up to us. Reckless drivers use it one way, cautious drivers another.
I was skeptical of Arvind Magesan’s crosswalk timer study, so I tracked it down online. Co-authored with Sacha Kapoor, the study is disguised under the unrevealing title Paging Inspector Sands: The Costs of Public Information. In supporting their findings, Magesan and Kapoor assert (citing earlier research) that “providing drivers with information about the time until a light change causes drivers to approach traffic lights more aggressively on average.” That on average is key. Some of us use the information the way my driving instructor wanted us to – to give us a little extra warning that we need to slow down. Others use the information to determine how much they need to speed up to avoid wasting time at a red light.
I would have thought the two effects would balance out – cautious drivers would avoid crashes by slowing down, reckless drivers would get into more crashes by speeding up, and the net effect would be a wash. But when you think about it, it makes sense that accidents would increase even if crosswalk timers didn’t, as the authors claim, make drivers more aggressive on average. Because you need both a reckless and a cautious driver to create the conditions for the kind of crash they describe.
Let me explain. After I read the news story, but before I read the study, I assumed that the higher accident rate was caused by vehicles speeding through yellow lights and colliding head-on with vehicles attempting to turn left. This was consistent with Magesan’s assertion that it was at less busy intersections where the increase in accidents occurred. At busy, slow-moving intersections, vehicles are unable to work up a head of steam, so the crosswalk timer makes no difference. But at less busy intersections, drivers who see the timer from fifty or a hundred yards out might stomp on the gas to try and beat the light. Since less busy intersections tend not to get designated left turn signals, I reasoned, there are more likely to be vehicles waiting there to turn left on yellow, hence, more opportunities for smashes.
My assumption was wrong. The authors broke down the accidents by type, and it seems that collisions involving a “turning movement” increased only negligibly when the crosswalk timers were installed. The greatest increase was in rear-end collisions. What’s more, speeding wasn’t the major cause of the accidents. Tailgating was.
Consider that, in any given pair of vehicles approaching an intersection, there are four possible combinations:
1. Cautious driver following cautious driver.
2. Cautious driver following reckless driver.
3. Reckless driver following reckless driver.
4. Reckless driver following cautious driver.
In scenarios 1 and 2, the crosswalk timer ought to help cautious drivers more accurately gauge when to start slowing down. This can be useful, especially in slippery winter conditions (the study was conducted in Toronto) where braking distance might be three or four times longer than usual. You’d think some number of rear-end collisions would be avoided here.
In scenario 3, where a reckless driver follows a reckless driver, they should both zoom safely through the intersection.
The typical crosswalk timer-induced accident involves scenario 4, where a reckless driver follows too closely behind a cautious driver. The reckless driver sees the timer and concludes that if he sticks closely to the vehicle ahead, he can sneak through the light just as it changes. He reasons that if he can make it through the intersection on time, the driver ahead, who obviously can see the timer just as well as he can, has even less reason to stop. What he doesn’t realize is that the cautious driver ahead isn’t interpreting the signal the same way. The cautious driver hits the brakes just as the reckless driver hits the gas.
In this case, the added information provided by the crosswalk timer is making things worse by facilitating both the recklessness of the reckless driver and the caution of the cautious driver. The accident would have been avoided if the reckless driver had slowed down or if the cautious driver had maintained his speed.
Whatever behavioural mechanism is at work, it appears that offering too much information can in fact make traffic signals more dangerous. Does it follow that less information makes signals less dangerous? According to Wikipedia, the very first modern traffic light included a warning that the light was about to change – a buzzer, back then, rather than a yellow light. How would drivers behave at an intersection where there was no warning at all – no buzzer, no yellow light, just red and green, stop and go? It might lead to more running of red lights. Or it might make everyone more cautious as they approached an intersection, for precisely that reason.
If it turns out having some warning is in fact safer than having none, what’s the sweet spot, safety-wise, between not enough warning and too much? And what’s the tradeoff between that optimally safe arrangement and the optimally efficient flow of traffic? After all, the safest traffic condition is probably gridlock, where no-one can build up enough speed to hurt anybody else.
Recently a few blogs have featured this video of Poynton, England, where they removed all traffic signals from the central intersection. In order to navigate the new uncontrolled double-roundabout safely, drivers are forced to actually pay attention to pedestrians and their fellow drivers. The result, apparently, is a much safer and smoother flow of traffic than you might expect.
Predictably, most of the discussion of this innovation has occurred on crunchy-leaning urban design blogs, and is uniformly uncritical. On some forums you can find a few skeptical comments from cyclists and drivers who claim that the new configuration is simply shifting traffic problems to routes that miss the town. (Not that this is necessarily a bad thing.)
Me, I’m agnostic. I’m glad planners are trying out new ideas, but I wonder if the current mania for walkability, shared roadways, and related New Urbanist tropes will in the long run suffer a backlash just as the previous mania for Brutalist skyscrapers, vast concrete plazas, and neighbourhood-gutting freeways did. My impression of that bizarre double-roundabout in the Poynton video is that it looks pleasant enough for pedestrians, but as a driver I’d gladly go a few miles out of my way to avoid it. But then, I’m a North American, and we’re famously flummoxed by roundabouts.
Incidentally, Magesan and Kapoor’s policy recommendation is that crosswalk timers be replaced with audible countdowns that only pedestrians, and not drivers, can hear. If I were a policymaker I’d be reluctant to act on this until the study has been replicated a few times by other researchers in other places. Since many cities seem to be installing these timers, there should be plenty of data available.
As a fairly cautious driver, I’d prefer to keep the visible timers – I find them useful for regulating my speed when conditions are slippery, as they so often are in Canada. I suppose I’ll feel differently if I’m ever rear-ended at one of these intersections.