Authors: Tom Vanderbilt
This happens all the time in traffic, in which reneging on the queue is often impossible. It’s why I changed lanes and became a “late merger,” and why people get frustrated with late mergers. (I will explain shortly why they shouldn’t.) Sometimes, changing lanes (i.e., moving to a different queue) is actually a useful strategy. Often, however, it gets us no real gain. A Canadian television news program had two drivers commute along the same route on a highway at the same time. One was told to make as many lane changes as possible, the other to avoid changing lanes. The chronic lane changer saved a mere four minutes
out of an eighty-minute drive,
which hardly seems worth it. The stress involved in making all those changes probably took more than four minutes off the driver’s life.
One reason why many people constantly change lanes was demonstrated in a fascinating experiment conducted by Donald Redelmeier and Robert Tibshirani, a clinical epidemiologist in Toronto and a statistician at Stanford University, respectively. Using a simple computer simulation of two lanes’ worth of congested traffic obeying typical traffic behavior, as well as a video of an actual congested highway, the researchers found an illusion when looking at a sample driver: Even though the subject car had as many “passing events” as it had “overtaking events”—meaning it was maintaining the same overall relative pace as the next lane—the car spent more
time
being passed by cars than it did passing them.
Traffic, for reasons I will later explain, tends to act like an accordion: As traffic slows in a jam, it compresses; as congestion eases, the accordion “opens” and cars begin to speed up. Because of the uneven nature of stop-and-go traffic, these shifts happen in different lanes at different times. A driver in a temporarily opening lane may very quickly pass a cluster of compressing cars in the next lane. But then he will find himself in the compressing lane. And what happens? He spends more time watching those vehicles zip by in the next lane. To make matters worse, the researchers found that the closer a driver drove to the car in front of him, and the more glances he made to the next lane, the worse the illusion seemed.
Something else might also be helping to create the illusion. Drivers spend most of their time—anywhere from 80 percent to over 90 percent, studies have found—looking at the forward roadway. This includes, of course, the adjacent lane; estimates are that for every two glances we make at our own lane, we make one glance at the next lane—simply so we can actually stay in our lane. This means we are highly aware of vehicles passing us. We spend only about 6 percent of our driving time looking in the rearview mirror. In other words, we’re much more aware of what is passing us than what we have passed.
The fact that we spend more time seeing losses than gains while driving in congestion plays perfectly into a well-known psychological theory called “loss aversion.” Any number of experiments have shown that humans register losses more powerfully than gains. Our brains even seem rigged to be more sensitive to loss. In what psychologist Daniel Kahneman has called the “endowment affect,” once people have been given something, they are instantly more hesitant to give it up.
Do you remember the childlike glee you felt the last time you found a parking spot at the mall on a crowded day? You may have left the spot with a certain reluctance, particularly if someone else was waiting for it. Studies have shown that people take longer to leave a parking spot when another driver is waiting, even though they predict they will not. It’s as if the space suddenly becomes more valuable once another person wants it. In strict terms it does, even though it is no longer of intrinsic value to the person leaving it. This sensitivity to loss might also help explain the late-merger dilemma described in the Prologue. What really triggers the decision to change lanes is not so much the coolly rational assessment of underused transportation capacity but the fact that people kept passing while the early mergers stood still. The late merger’s gain is perceived as the early merger’s loss.
But what’s the harm in merely changing lanes, anyway? One study, by the National Highway Traffic Safety Administration, found that almost 10 percent of all crashes involved lane changes. How many of those lane changes were necessary, and how many were discretionary? Do we really understand what is involved in the choices we are making? It is this last question that was at the heart of Redelmeier and Tibshirani’s lane-changing study, for Redelmeier, a soft-spoken, sober doctor who spends a third of his time seeing patients at the Sunnybrook Health Sciences Centre in Toronto, has a privileged window on to the consequences of the decisions we make while driving.
“I mostly look at individuals that get seriously damaged in the aftermath of a crash,” he told me in his office. “For many of them, their lives are ruined forever. For many of them, there’s also this tremendous sense of remorse or chagrin—you know, if only they had behaved slightly differently, they would have never ended up in the hospital. There’s a real element of almost counterfactual thinking that goes on in the aftermath of a crash. When someone comes down with pancreatic cancer there’s a lot of suffering that’s going on, but they usually don’t start second-guessing themselves about how things could have been done differently in order to avoid this terrible predicament, whereas with motor vehicle crashes it’s a very strong theme. That got me thinking how complicated driving is.”
We may be doing things in traffic for reasons we do not even understand as we are acting. But how we can resist things like the next-lane-is-faster illusion? Redelmeier suggests, if not completely seriously, that we might feel better if we spent more time looking in the rearview mirror. Then we could make “downward comparisons,” as with the poor saps in the Hong Kong post office, and not feel so bad. But we would also quite likely collide with the vehicle in front of us, and then cars in the next lane really would be going faster. The very nature of driving, posited as a constant progress along an endless queue, defeats us. Traffic messes with our heads in a strangely paradoxical way: We act too human, we do not act human enough.
People are afraid to merge on freeways in Los Angeles.
—Bret Easton Ellis,
Less Than Zero
We humans have achieved great things. We have unlocked the once-unfathomable human genetic sequence, sent space probes to the far reaches of the solar system, and even managed to freeze a beam of light. But there’s one scientific conquest that has largely eluded us. It’s all the more puzzling because, on the face of it, it seems so mundane: We have not found a way to make drivers merge with the most efficiency and safety on the highway.
The situation described in the Prologue that I encountered on the Jersey highway is known in the traffic-engineering world as a “work-zone merge.” Work zones, it turns out, are among the most complex and dangerous areas on the highway. Despite the signs often warning of large penalties for striking a worker (or pleas like
SLOW DOWN, MY DADDY WORKS HERE
), they are much more dangerous for the drivers passing through them than for the workers—some 85 percent of people killed in work zones are drivers or passengers. The reasons are not difficult to imagine. Drivers moving from an incredibly fast, free-flowing environment are suddenly being asked, sometimes unexpectedly, to come to a crawl or even a full stop, perhaps change lanes, and pass through a narrow, constricted space filled with workers, heavy machinery, and other objects of visual fascination.
And then there’s the inevitable point at which two lanes of traffic will be forced to become one (or three to become two, etc.), when the early mergers, the late mergers, and everyone in between are suddenly introduced to one another. This can get sticky. It seems that even though (or maybe because) we’re all tossed together on the road, drivers are not all that comfortable with interacting; a survey undertaken by the Texas Transportation Institute found that the single most common cause of stress on the highway was “merging difficulties.”
Traffic engineers have spent a lot of time and money studying this problem, but it is not as simple as you might think. The “conventional merge” site, the sort I experienced on the highway in New Jersey, works reasonably well when traffic is light. Drivers are warned in advance to move into the correct lane, and they do so at a comfortable distance and speed, without a “conflict” with a driver in the other lane. But the very nature of a work zone means that traffic is often
not
light. A highway going from two lanes to one, or experiencing a “lane drop,” loses at least half of its capacity to process cars—even more if drivers are slowing to see what is going on in the work zone itself. Because the capacity is quickly exceeded by the arriving cars, a “queue” soon forms. The queue, inevitably, is longer in the lane that will remain open, probably because signs have told drivers to move there.
This causes more problems. As the queue grows, it may move far back up the highway—engineers call this “upstream”—perhaps even past the signs warning of the lane closure. This means that newly arriving drivers will be encountering an unexpected queue of cars. Seeing no reason for it, they will be unaware that they’re in a lane that is due to close. Once they learn this, they will have to “force” their way into the queued line, whose drivers may view the new arrivals, fairly or not, as “cheaters.” As the entering drivers slow or even stop to merge, they create a temporary second queue. Drivers who grow frustrated in the queued line might similarly force their way into the faster open lane. This is all a recipe for rear-end collisions, which, as it happens, are among the leading types of crashes in work zones.
To improve things, North American engineers have responded in two basic ways. First, there is the school of Early Merge. To tackle the “forced merge” problem, Early Merge spreads out the whole merging zone. Drivers are warned by a sign several miles in advance of the “taper” that a lane drop is coming, rather than the twelve hundred feet or so in the conventional merge. “No Passing Zones” signs are often placed in the lane that will close. The earlier notice, in theory, means drivers will merge sooner and with less “friction,” as engineers politely say, and will be less surprised by a sudden queue of stopped cars. Indeed, a 1997 study of an Indiana construction site using this system showed very few forced merges, few “traffic conflicts,” and few rear-end collisions.
Early Merge suffers from a critical flaw, however. It has not been shown to move vehicles through the work zone more quickly than the conventional merge. One simulation showed that it actually took vehicles
longer
to travel through the work zone, perhaps because faster-moving cars were being put behind slower-moving cars in a single lane sooner than they might naturally have gotten there, thus creating an artificial rolling traffic jam. An Early Merge system would also seem to require some kind of active law enforcement presence to make sure drivers do not violate the concept. As we all know, the presence of a police car on the highway has its own unique effects on traffic.
The second school, Late Merge, was rolled out by traffic engineers in Pennsylvania in the 1990s in response to reports of aggressive driving at merge locations. In this system, engineers posted a succession of signs, beginning a mile and a half from the closure. First came
USE BOTH LANES TO MERGE POINT,
then a
ROAD WORK AHEAD
or two, and finally, at the lane drop:
MERGE HERE TAKE YOUR TURN
.
The beauty of the Late Merge system is that it removes the insecurity or anxiety drivers may feel in choosing lanes, as well as their annoyance with a passing “cheating” driver. The Late Merge compresses what may normally be thousands of feet of potential merging maneuvers to a single point. There is, presumably, no lane jumping or jockeying, as the flow or speed should be no better or worse in one lane than another—hence there are fewer chances for rear-end collisions. Because cars are using both lanes to the end point, the queue is cut in half.
The most surprising thing about the Late Merge concept is that it showed
a 15 percent improvement in traffic flow
over the conventional merge. It turns out that the Live Free crowd was right. Merging late, that purported symbol of individual greed, actually makes things better for everyone. As one of my Live Free responders had succinctly put it: “Isn’t it obvious that the best thing to do is for both lanes to be full right up to the last moment, and then merge in turn? That way, the full capacity of the road is being used, and it’s fair on everyone, rather than a bunch of people merging early and trying to create an artificial one-lane road earlier than necessary.” (Note: This does
not
apply to people “late-merging” their way to the head of queues at off-ramps and the like, as those late mergers may temporarily block an otherwise free-flowing lane of traffic, not to mention greatly irritating those already queued.)
It’s not just North Americans who have problems with merging. The United Kingdom’s Transport Research Laboratory, in an internal report looking at new work-zone merging treatments, noted the “poor utilization of the closed lane well in advance of the taper,” which it partially attributed to “vehicles blocking this lane deliberately to prevent others from ‘queue jumping.’” In the 1990s, U.K. road authorities began experimenting with new signs and the so-called zipper merge, used in Germany since the 1970s. Rather than simply warn of an impending lane closure, the signs, beginning well in advance of the lane drop, advised drivers,
WHEN QUEUING USE BOTH LANES
and
MERGE IN TURN
. But the TRL, in trials on Scottish motorways, found that while the system reduced queue lengths, it didn’t make traffic flow any more smoothly through the work zone. (Part of the problem may be that drivers are often still unclear about exactly where to merge: where the sign tells them to, or where the two lanes become one, or somewhere in between?) Most European traffic engineers try to avoid merging problems wherever possible by simply eliminating the need to merge. Instead, they carve out extra lanes by making the remaining lanes much smaller; this not only preserves multiple lanes, it forces drivers to slow, which is also safer.