.

Why Pedestrians (And Bicyclists) Die

Marc Green


Abstract: While vehicle occupant fatalities have fallen steadily over the last decade, pedestrian deaths have remained constant despite attempts at remediation. This failure is rooted in lack of attention to the actual causes of vehicle-pedestrian collisions. These causes are evident in three types of research. One examines police reports and accident data to determine fault attribution. A second goes further by specifying the actual events that caused the collisions. The third, and most detailed method, is a case-by-case human factors analysis of documents, pictorial evidence and the collision scene itself. All three methods arrive at the same conclusion: the primary cause of pedestrian collisions is pedestrian behavior. Death rates have not fallen because attention has focused the burden of remedial measures exclusively on the wrong party - the driver. Instead, preventative steps would be more productively and more fairly aimed at changing pedestrian behavior. Lastly, the early evidence on bicyclist fatalities shows similar trends and leads to similar conclusions.

"The excessive speed of automobiles costs the lives of many persons; and that scarcely a week, sometimes scarcely a day passes without chronicling from one to a dozen deaths occasioned by the reckless driving of these machines. Fortunately, the chauffeur and his guests are the usual sufferers, and in their misfortunes as lawbreakers, the general public do not much concern themselves. Our sympathies are rather reserved for…the bystander who is run down and crushed by the enormous weight of these engines." -H. B. Brown, 1908. The status of the automobile. Yale Law Journal 17, 223-231.

I'm not sure he's wrong about automobiles," he said. "With all their speed forward they may be a step backward in civilization -- that is, in spiritual civilization. It may be that they will not add to the beauty of the world, nor to the life of men's souls. Booth Tarkington, The Magnificent Ambersons

[T]hey say although the traffic police are communists the drivers are fascists, which would be all right if it were not that the pedestrians are anarchists. (Len Deighton, 1964. Funeral in Berlin).

Background

Pedestrian safety is receiving increased attention. The fatality rate of drivers and their passengers has declined over the last decade while pedestrian fatalities have remained constant [1]2. They constitute an increasing percentage of overall road deaths, so authorities are focusing more than ever on means to reduce the number.

The failure to reduce pedestrian fatalities shows that previous strategies have reached a limit. The obvious question is "Why?" The short answer is that too much attention has focused on outcome to the exclusion of causation. An understanding of what actually causes vehicle-pedestrian collisions is an important precursor to improved safety.

The reasons for failure to seek solutions in the real causes lie in Brown's quote, which expresses an ideology that focuses solely on outcome. Pedestrians (and bicyclists) are "vulnerable road users" (VRU's). The term evokes images of the weak and the defenseless in a David vs. Goliath confrontation where David is bound to lose. In this story, cars and drivers are, of course, the Philistines.

The scientific literature and detailed collision investigation, however, tell a very different story. Moreover, I have performed in well over a hundred vehicle-pedestrian collision investigations during the last eighteen years. These investigations examined police reports and accident reconstructions, witness and participant statements, and photographic and video evidence. Moreover, they generally included first-hand scene inspections with time, space and lighting measurements, sometimes made during collision re-enactments.

Analysis

1. Fault Attribution Research

Police reports attribute greater blame to pedestrians than to drivers [e.g. 3], perhaps because pedestrians commit over 3.5 times more road violations [4]. Other research has similarly confirmed that pedestrians are the prime collision cause, putting their fault at 75 percent, [6], 80 percent [7] and even 90 percent [8]. The remaining cases were due to the driver or shared responsibility. For example, more recent research [9] found 59 percent pedestrian compared to 32 percent driver fault with the rest shared.

However, fault attribution changes somewhat with circumstances. Police blamed pedestrians more for midblock collisions and cases where the pedestrian darted out at an intersection [e.g. 10]. Drivers received more blame during turns, especially to the left. This is expected since left turns create a high driver workload and tax attention [3]. Drivers are very seldom the collision cause when traveling straight. Conversely, pedestrians gain more blame in low illumination [11], presumably in recognition that drivers cannot see as well in dim light. As discussed below, this is primarily due to pedestrian clothing choice. There are fewer data for vehicle-bicycle collisions, but the pattern is the same. Bicyclists are faulted in 50 percent of cases compared to 28 percent for drivers, with the remaining cases shared or uncertain fault [5].

2. Causes From Accident Data

Other research has sought to identify causal factors. In most cases, the cause is some form of pedestrian behavior. Collision avoidance has two main mental components, detection and prediction. The road user must first detect the other party. The main factors in detection are visibility, which is determined largely by contrast, and conspicuity, which is determined largely by expectation. The road user must also be able to predict the present and future location of the other party. The present location is important because attention is limited, and drivers must know where to look. The future is important so the driver can gain situational awareness, i.e., determine what will happen if he maintains current course and speed. Pedestrians frequently confound driver detection and prediction. Ironically, pedestrians exhibit many of the behaviors used to fault drivers: lack of attention, using a phone or other distraction, failure to search, and even speeding (darting into the roadway) and perhaps most notably, drunkenness.

Pedestrians cross illegally.

Most road collisions occur when an "active" party intrudes on the right-of-way of a "reactive" party. In typical collisions, the pedestrian is the active party. Pedestrians routinely ignore traffic rules, so illegal crossing is a major cause of pedestrian collisions, with some calling it the major cause [e.g. 12]. The numbers vary from study to study, but the lawlessness of pedestrians is well documented. Typical results include: 80-93 percent of pedestrian fatalities occurred outside of marked crosswalks [13], 80 percent of pedestrians in fatal collisions had violated some law [14] and close examination of 40 collisions found unsafe pedestrian behavior in almost all cases [15]. The general finding is that pedestrians tend to cross the road when it serves their goals of convenience and saving time, and they are not greatly concerned with potential safety implications [16]. In sum, "Pedestrians often consider themselves outside the law" [17].

Bicyclists are even greater scofflaws. They routinely ignore traffic controls [18]. Failing to come to a stop at STOP signs is especially common. At low speeds, 80 percent do not stop while 34 percent don't even bother to slow [19]. Further, 70-75 percent of bicyclists ignore red lights while 40 percent travel the wrong way down traffic lanes [20]. When they do stop, it is usually only because heavy traffic leaves them no choice [21]. They also feel free to switch back and forth between being vehicles (in the roadway) and pedestrians (on the sidewalk) as a means of avoiding having to stop at intersections. Bicyclist rule violations are probably more dangerous pedestrian illegal behavior. The bicycles' greater speeds leave less time for drivers to react. They also pose a small threat to pedestrians in general, but a large threat to older pedestrians. Many cases of bicyclists killing elderly pedestrians have already occurred3

The law violation is a major cause of vehicle-bicycle collisions. A large proportion of collisions occur when the bicycle is traveling inconsistently with normal traffic flow, generally an illegal direction [5]. Such behavior is highly dangerous because of the reduced detectability of small, low probability objects in unexpected locations. The purpose of traffic laws is to allow road users to anticipate each other's actions. Since perception depends heavily on expectation, the best way for pedestrians and bicyclists to be safe is to be in a place and at a time that drivers expect them. Of course, the same applies to drivers.

Despite all this, some bicycle advocates encourage even more rule violation. They endorse red light and STOP sign running as a way to avoid vehicle exhaust behavior [22], but the real reason is probably that they don't want the work and inconvenience of having to slow or to stop. Many bicyclists apparently believe that they should be able to ignore traffic controls at their own discretion, which is a dangerous precedent. There is even now a push in some states for legislation to allow bicyclists to run STOP signs. This is sure to be a legal nightmare by removing any protection that drivers have from misbehaving bicyclists.

Pedestrians dart into the roadway.

Pedestrians also frequently suddenly run into the roadway, sometimes from behind an obstacle. For example, one study [23] found that 44 percent of pedestrian crashes were due to midblock dart-outs or dashing out at intersections compared to seven percent caused by a driver failing to see the pedestrian. Another observational study [24] found that pedestrians darted into the roadway from behind obstructing vehicles or buildings in 46 percent of near misses. However, other research [5] put "running into road" at only fifteen percent. All agree, however, that midblock-dart outs are a particular problem in child collisions, accounting for as much as 69 percent of injuries [see 2].

Another large study [25] classified 2100 vehicle-pedestrian collisions and divided them into six classes. Dart outs accounted for 42 percent [c.f., 23]. Most of the six scenarios have one thing in common: a pedestrian's sudden appearance and fast movement created a situation where the driver had little distance/time to respond. In other scenarios, physical constraints such as low lighting, background clutter, obstructions, bad weather, etc. further limited driver ability to respond. The only major scenario where pedestrian behavior was not a large factor occurred during a turn, where task demands tax the driver's limited attentional capacity. There is no mention of drivers on a straight path failing to yield as a major collision cause.

Pedestrians are unpredictable.

Collision avoidance depends on two factors, detection and prediction. In addition to darting, pedestrians often behave unpredictably by sudden changes in speed while crossing, often without searching for traffic [26]. The variation in speed as well as in trajectory makes driver prediction of their future location difficult. Bicyclist turns are also difficult to predict [27]. Although required by law, few bicyclists signal turns and passing.

Pedestrians don't check for traffic.

Pedestrian failure to search for approaching traffic is frequently cited [e.g., 24] as a major cause of collision and in some cases [cited in 16] as the major cause. Studies generally find that pedestrians don't look in both directions or don't look at all before crossing. Recent data [28], for example, put pedestrian "failed to look properly" as a factor in 48 percent of crashes. They are sometimes overconfident or distracted.

Search failure is a common explanation for increased collision rate, even in crosswalks More specifically, pedestrians tend to move their head to search for traffic as they approach the curb but then reduce head movements once the curb is reached [29]. The likely interpretation is that pedestrians make their decision to cross before reaching the curb and cease looking for traffic once the decision is made because they assume crossing is safe, - an example of confirmation bias. This explanation is supported by increased pedestrian search in riskier situations. For example, pedestrians search more for traffic when crossing during the "Don't Walk" phase of a crossing signal, but do not search during the "Walk" phase [30].

Pedestrians are drunk.

Pedestrian drunkenness, along with the use of other drugs, is a major factor in pedestrian collisions. Studies routinely find that anywhere from a third to as many as a half or more of pedestrians killed in road accidents were intoxicated at or above the .08 BAC level that is illegal for driving. Table 1 shows some of the results. The NTSB and earlier data suggest that the drunken pedestrian fatality rate has remained constant over the years

% BAC Population Source
32 0.15>= Pedestrians All Haddon, Valien, McCarroll, & Umberger (1961)
33 0.08<= Pedestrians All Holubowycz, Kloeden, & McLean (1994)
33 0.15<= Pedestrians All 50 intoxicated, Leaf & Presusser (1997)
35 0.08<= Pedestrians All National Highway Transportation Safety Admin. (2005)
35 0.08<= Pedestrians All National Highway Transportation Safety Admin. (2014)
35 0.10<= Pedestrians All Dickman (1992)
38 0.08<= Pedestrians All Holubowycz (1994)
46 0.09<= Pedestrians All Martin (2006)
51 0.08<= Pedestrians All Fontaine & Gourlet (1997)
70 Legal<= Pedestrians Male Australian Transportation Safety Board (2003)
74 0.0> Pedestrians 47% 0.05<=, 32%<=0.15 Haddon, et al., (1961)
19-26 0.08<= Bicyclists 16+ Insurance Institute for Highway Safety (2019)
21-28 0.08<= Bicyclists 16+ Eichelberger, McCartt, & Cicchino (2018)
23 Legal<= Bicyclists 15+ Li & Baker (1994)
29 0.18-0.34 Bicyclists Fife, Davis, Tate, Wells, Mohan, & Williams (1983)
27 0.12<= Bicyclists Airaksinen, Nurmi-Luthje, Kataja, Koger & Luthje (2018)
50 0.0<= Bicyclists Adult Frank, Frankel, Mullins, & Taylor (1995)
Table 1 BAC of pedestrian and cyclist roadway fatalities. See [2] for complete references.

The reality is even worse than Table 1 suggests. Children and older pedestrians are seldom intoxicated, so the rate for non-senior adults is even higher than the table suggests. For example, the NTSB's 35 percent for the entire population increases to near fifty percent for pedestrians aged 35-44. Since fewer females are intoxicated, the rate for adult males must be extremely high. One study found that 70 percent of adult male pedestrian fatalities were intoxicated.

Data on bicyclists are sparser, but Table 1 shows that alcohol also plays a major role in their crashes. The odds ratio of an intoxicated bicyclist having a crash was 4.5 compared to a sober one [56]. However, the bicycle data are more difficult to interpret than for pedestrians because studies sometimes mix vehicular and nonvehicular crashes.

When most think of alcohol involvement in a crash, they usually assume that the alcohol was in the driver. NHTSA likes to promote this misconception with the deceptive "alcohol related" crash statistic. In fact, NHTSA data show that a collision between a drunken pedestrian and a sober driver is three times more likely than between a drunken driver and sober pedestrian (24 vs. 8 percent). A particularly thorough study [29] examined five years of pedestrian fatality data in London, UK and found that alcohol in the "driver/rider" was "only" a factor in 2.5 percent. Note that the driver was not necessarily intoxicated or the cause in all these cases.

Amazingly, these data receive virtually no media attention. There is MADD, but no MOPP (Mothers Opposing Polluted Pedestrians). This is unfortunate because drunken walking (and possibly bicycling) is more dangerous than drunken driving.

Pedestrians are overly confident.

In most vehicle-pedestrian collisions, the pedestrian sees the vehicle approaching [2]. This begs the question of why the collision still occurs. One answer lies in the repeated finding that pedestrians drastically overestimate their visibility to oncoming drivers at night [e.g. 31]. They believe that because the approaching vehicle is easy to see, the vehicle driver can see them. Pedestrians often simply walk in front of the approaching vehicle and assume that the driver will stop. Bicyclists also exhibit a similar overconfidence in their visibility at night [e.g. 32]. Overconfidence may lead to accepting risky gaps. Pedestrians and bicyclists, for example, often leave very small margins for error, 1-2 seconds [33].

Pedestrians choose to make themselves invisible.

Many pedestrian collisions occur at night, with risk increasing by a factor of seven [34]. In 87 percent of nighttime collisions, the driver either saw the pedestrian only at the last split second or never saw the pedestrian. This is partly due to geometry, since a pedestrian emerges from the side, outside the headlamp beams and in the driver's peripheral vision [2].

However, the lack of nighttime visibility is also due to pedestrian clothing. Visual detection depends on contrast, the difference in reflectance between an object and its background, which is dark at night. Many [e.g., 35] have shown that visibility distances to pedestrians wearing dark clothing is often too short for normal younger and especially older drivers to avoid collision on typical roads. A good rule of thumb is that recognition distance for a dark pedestrian is only a hundred feet [e.g., 36], but a police officer has testified that forty feet is more realistic. Moreover, virtually all studies show that some younger and many older attentive and alerted drivers completely fail to see unexpected pedestrians in dark clothing.

Pedestrian clothing choice makes a great difference in detectability. Winter clothing has an average reflectance of only 5-6 percent, i.e., essentially black [2]. White clothing can increase visibility distance by 470 percent [31]. The use of "biological motion", attaching retroreflectors to joints so that the human form stands out from the background, is even more effective. I have made many nighttime pedestrian light measurements and can confirm that clothing choice is often the major difference between life and death, exceeding even road lighting in importance. Still, pedestrians fail to appreciate the beneficial effects of clothing reflectance and "biological motion" [e.g. 31].

Bicyclists would also benefit greatly from clothing that is more conspicuous. Bicycle lights are a step in the right direction, but they are not very conspicuous by day and camouflaged by urban background clutter at night. In contrast, a large-scale study [37] of 6,800 bicyclists found that visibility clothing decreased bicyclist collision rates by more than 70 percent.

In sum, dark clothing and invisibility is a choice. Whether it is due to fashion or to avoid dirty clothing, it is still a choice that the VRU can easily change. The driver who is put in a difficult position by having to detect and recognize a dark object at night has no choice.

Pedestrians are distracted.

While there is enormous research on distracted drivers, the problem of distracted pedestrians is only now beginning to receive attention. Studies have found that about 30-47 percent [38, 39] of pedestrians crossing the street are distracted. Cell phone conversations and texting were by far the major detractors. The problem is certain to grow due to the dramatic increase of cell phones for a variety of increasingly attention-absorbing tasks.

Distracted pedestrians exhibit the same impairments as distracted drivers. The following list4 shows some of the impairments just for merely talking on a cell phone while crossing.

  • Failed to detect objects in peripheral vision, i.e., tunnel vision and reduced field of view (Barkana, Zadok, Morad, & Avni, 2004);
  • Were less likely to look left and right for traffic when crossing (Bungum, Day, & Henry, 2005; Hatfield & Murphy, 2007; Neider, McCarley, Crowell, Kaczmarski, & Kramer, 2010);
  • Ignored traffic signals (Bungum, Day, & Henry, 2005; Peši?, Anti?, Glavi?, & Milenkovi?, 2016);
  • Were less aware of their surroundings (Kuzel, Heller, Gray, DiJorio, & Straughn, 2008; Nasar, Hecht, & Wener, 2008; Hyman, Boss, Wise, McKenzie, & Caggiano 2010; Lamberg & Muratori, 2012);
  • Walked and crossed more slowly (Murray, 2006; Hyman, Boss, Wise, McKenzie, & Caggiano 2010; Neider, McCarley, Crowell, Kaczmarski, & Kramer, 2010; Neider, Gaspar, McCarley,
  • Crowell, Kaczmarski, & Kramer, 2011; Lamberg & Muratori, 2012);
  • Deviated their paths more often (Hyman, Boss, Wise, McKenzie, & Caggiano 2010);
  • Did not wait for traffic to stop before crossing (Pešic, Antic, Glavic, & Milenkovic, 2016);
  • Waited longer to start crossing (Stavrinos, Byington, & Schwebel, 2009);
  • Selected smaller gaps and had more close calls (Stavrinos, Byington, & Schwebel, 2011);
  • Were poorer at judging gaps and time to arrival (Murray, 2006);
  • Had longer perception-response times (Murray, 2006; Tapiro, Oron-Gilad, & Parmet, 2016); and
  • Children were less attentive, chose less safe gaps and had more close calls (Stavrinos, Byington, & Schwebel, 2009).

Pedestrians engrossed with their smart phones exhibit the same deficits but to an even higher degree. As with drivers, texting and emailing are the perfect storm of distraction - intense foveal concentration to a small display requiring good acuity while performing a task that requires deep cognitive processing for an extended time. Sooner or later, looking away from where you are going leads to trouble. This is as true for pedestrians as it is for drivers.

The growth of dangers posed by such distraction can only increase from here. I saw no technology-distracted pedestrian collisions ten years ago. A few began appearing about five years ago, most involved listening to music or a cell phone conversation. Now they are common, primarily pedestrians looking at their smart phones. In one recent case, the pedestrian reading a text message was also crossing against a red light. She never looked up or noticed the bus traveling only 15-20 mph that collided with her. In another, a pedestrian reading his email wandered into an intersection, stopped and was struck by a car.

Similar distraction effects have been found in cell phone using bicyclists, who exhibit even more unsafe behavior [40]. When using a cell phone, they engaged in more risky behavior and more often forced a vehicle driver to perform an avoidance response [41].

Pedestrians walk with their backs to traffic.

While most collisions occur during crossings, many happen when pedestrians walk in the road. In most jurisdictions, the law requires pedestrians to walk facing traffic when they are in the roadway because it is far safer. It is easier to see a vehicle approach from the front than to see it by occasionally looking over the shoulder. Drivers are better able to see faces than the dark backs of pedestrians. This is why pedestrians facing traffic have 77 percent fewer collisions [42]. When collisions do occur, moreover, pedestrians facing traffic suffer milder injuries. Pedestrians facing traffic also choose larger and safer gaps, possibly because they are a population of lower risk takers.

Again, the published data are consistent with my investigation experience. In several collisions, pedestrians (often multiple pedestrians strung out across the roadway) walked with their backs toward traffic. No cases had a pedestrian walking toward the vehicle. When a collision occurred late at night, the pedestrian walking in the road was almost invariably drunk and often on drugs.

Pedestrians don't know the traffic rules.

A large percentage of pedestrians have an incorrect understanding of when an approaching vehicle is required to stop for them. A questionnaire study [43] of attitudes found that both pedestrians and cyclists are prone to self-serving interpretations of safety and responsibility, especially for nighttime conditions.

Lastly, another was to look at pedestrian collisions is to examine exactly who is killed. If it were all the drivers' fault, then it would be expected that the fatal pedestrian collisions would be evenly distributed through out the population. As already discussed, it is in fact heavily skewed toward intoxicated pedestrians. Further, the distribution is heavily skewed toward males. For example, one study [55] found that males were 2.3 times more likely than females to be killed by a vehicle-pedestrian collision. Unlike most such studies which report gender effects without accompanying exposure data, this study found that only one percent of the difference could be attributed to exposure. Unless drivers target male pedestrians, these data clearly demonstrate that pedestrian behavior is a crucial factor.

3. Detailed Human Factors Collision Investigation

These research data match the analyses of over 100 detailed collision investigations. The pedestrian behavior described above caused almost all collisions. The most common collision scenario occurred when the pedestrian violated driver expectation, especially in low visibility where safety margins shrink. In these cases, the pedestrian had usually violated the driver's right-of-way by ignoring a red traffic signal or crossing midblock. Sometimes they walked in the wrong direction in the roadway. Collisions due to the erratic pedestrians absorbed in their smart phones are increasingly common.

Driver behavior was seldom the primary cause. For example, I have seen numerous cases of pedestrians crossing illegally, but only four cases where a vehicle collided with a pedestrian in a marked crosswalk. Two cases occurred in the nighttime at intersections where the streetlamp was burned out and visibility was a significant factor. One of these was a left turn and the other had a driver with BAC=.12. The two other collisions occurred when a car in the near lane stopped to allow a pedestrian to cross. As the pedestrian began walking, a vehicle in the center lane approached. The stopped car blocked the sightline between the pedestrian and the center lane driver, so both continued to the collision. The scenario is so common that it has its own name, "multiple threat". Limited visibility, not driver behavior, was the primary cause.

DISCUSSION

1. What About Drivers?

The discussion has drawn a heavily one-sided view of collisions, painting pedestrians as the primary authors of their own misfortune. There is good reason. First, it is an "inconvenient truth" that pedestrian behavior is the primary cause of the collisions. Second, the bias against vehicles is so strong that it is hardly news to say that drivers sometimes cause pedestrian fatalities. Feeding the anti-car confirmation bias is unproductive. Many already apparently believe that it is the driver's responsibility to perfectly compensate for even the riskiest pedestrian and bicyclist behavior. In the case mentioned above, the pedestrian who was sending a text was also crossing against a red light. The driver's attention had been diverted by yet a second illegal pedestrian running midblock across the road. The driver was still criminally charged. In another case, a driver spent three years in jail for colliding with a .24 BAC pedestrian who wandered into the middle of the road.

Certainly, the driver controls greater energy and the pedestrian suffers the greater consequences. That does not necessarily make the driver responsible for the collision. Examination of human factors shows that the driver may control the energy, but the pedestrian has greater control over whether a collision will occur:
  • Visibility: Vehicles are far more visible to pedestrians than pedestrians are to drivers, especially at night. Drivers cannot change this. Pedestrians can [2];

  • Predictability: Cars travel highly constrained paths. Pedestrians know where to look for possible hazards - down the road. For a driver, pedestrians disregard crossing rules, so they can appear from almost anywhere at any time. Bicyclists also routinely disregard traffic rules such as red lights, STOP signs and one-way arrows;

  • Time. Pedestrians have virtually unlimited time to cross the road since they can wait as long as necessary for a safe gap. Once a pedestrian starts crossing, a driver may have only a few seconds to respond. Of course, signal times must be long enough for safe crossing;

  • Reaction and maneuverability: Pedestrians can react, stop and turn faster and in a shorter distance than a moving vehicle;

  • Attention: Drivers have to control their vehicles and watch for other vehicles, a task which often directs and consumes much attention. Pedestrians need only be concerned with crossing the street. There is no concurrent task - unless they are on a cell phone.

This last point deserves amplification. Even Ian Walker [44] an extreme anti-car ideologue (see below), stated that his review of the scientific evidence forced the conclusion that:

"to compensate for limited information-processing capacity, drivers form expectations of what they will encounter on the road and use these to guide their attention. These expectations often predict that only cars and other such vehicles will be encountered (which, statistically, is a reasonable assumption in many environments). As a result, drivers attend only to motor vehicles and regions of space in which motor vehicles are found... it is too demanding for drivers to attend to regions that might contain motor vehicles and regions that might contain VRUs."

Moreover, simply telling drivers to pay attention is futile. Attentional allocation is usually not under conscious control [2]. Conscious attention may temporarily prevail, but it won't last long. Humans are not constructed to notice everything in a scene. Human nature is like gravity; it always pulls us back into our innate predisposition to attend what is probable, relevant and harmful. Drivers will always unconsciously give priority to avoiding collisions with other vehicles.

In sum, drivers control the energy, but pedestrians and bicyclists control the circumstances. Drivers have a much greater workload. Intersections have 32 potential conflict points, including eight with pedestrians. He also must steer while searching and judging gaps in two or three road directions. The task demands on the pedestrian are minimal. The greater control and lower task demands makes pedestrians the better medium for reducing fatalities.

2. Why Pedestrian Fatalities Have Not Declined

Recent attempts to reduce fatalities have focused on steps such as bigger and more signs, tighter rules on yielding and lower speed limits. These have predictably failed to have much effect. The lack of sign effectiveness is no surprise given the limitations of driver attention. There is also only weak correlation between the presence of the sign and the presence of the pedestrian [2]. The "cry wolf" effect is inevitable. After repeatedly passing a sign warning of an event that seldom occurs, the sign takes on the information value of a tree. This problem is the reason for variable message boards.

Another obvious step is tougher yielding laws. Yet, no causation study finds that collisions often occur because a driver failed to yield for a pedestrian that he saw in time to respond. Moreover, such laws can produce unintended results. Sweden passed a law forcing drivers to "give the entire road" and yield whenever a pedestrian stepped off the curb. A 27 percent increase in pedestrian collisions ensued [45]. The study authors attributed the increase to pedestrian overconfidence. Another explanation is increased number of multiple threat scenarios.

Speed reduction has obvious face validity. Brown was already blaming excessive speed in 1908. In fact, the accusation that drivers travel too fast is as old as pedestrian fatalities. At the inquest following the world's first pedestrian fatality in 1896, an eyewitness accused the driver of acting recklessly and speeding. The car was traveling at four mph speed. "Excessive speed" is a highly subjective judgment.

Since then, studies have often reported that lower speed results in fewer collisions. However, most simply correlate speed limit with collision rate and ignore other situational variables. The studies provide no proof that the higher speeds actually caused the collision. Many are just statistical regression artifacts. Speed reductions and other safety measures are often introduced immediately after a spate of collisions. The collision rate then falls back to normal giving the false impression that the safety change caused the decline.

Moreover, studies often find little or no effect of reducing speed. A Virginia Commonwealth University study noted, "excessive speed does not appear to play a major role in pedestrian crash causation." Similarly, a detailed analysis [15] of 40 pedestrian collisions found no evidence of excess speed. A review [46] of thirteen traffic calming studies found that the reduced speed had failed to lower pedestrian collision rates. A Toronto Police Services reported that cars were speeding in 8% of fatal pedestrian crashes. However, pedestrian behavior may have been a causal factor in some of these. Even NHTSA [47] describes the effects of speed reduction on vehicle-pedestrian collisions as "limited." Moreover, drivers are least likely to be the cause of a collision when they are driving straight, the very time that their speed is highest. Conversely, street parking slows driver speed but increases pedestrian collision rate. In sum, the speed results are likely mixed because other factors are more important. Changing speed limits by a five kph-ten is unlikely to have much effect. In fact, it may backfire as frustrated, traffic-bound drivers accept smaller gaps and take more risks.

Of course, large, draconian speed reductions would at the least reduce injury severity, but this brings up another issue. Safety never comes for free. There is nothing in the world that can't be made safer by costing someone more money, inconvenience and inefficiency. The only real question is "Who is going to pay the price?" I would argue that the cost should be paid at least partially by those who cause the problem and who are coincidentally in a better position to solve it. That would be pedestrians and bicyclists.

The narrow focus on speed reflects a simplistic geocentric viewpoint, the third-party, god's eye view of collisions that appears in accident reconstruction diagrams. It focuses solely on physical factors. Pedestrian fatality rates have asymptoted, suggesting that any low hanging blame-the-driver fruit has been long ago picked. Further declines will require a different strategy, especially given a likely upsurge in fatalities due to 1) more pedestrian (and driver) impairment resulting from imminent legalization of marijuana, 2) more pedestrian (and driver) distraction from smart phones and other technology, 3) the aging driver and pedestrian population, 4) the increased automation of driving, which reduces driver vigilance and slows response and 5) the growing size number of SUVs with their higher, pedestrian-killing bumpers. This is yet another reason to change focus to the pedestrian and an egocentric viewpoint that sees the situation through road user eyes and behavior. Unfortunately, appreciating and applying the egocentric view requires reading the human factors research literature, which seldom interests decision makers, bureaucrats, most engineers and, of course, pedestrian and especially bicycle zealots.

3. Reducing Fatalities

Safety efforts should be directed toward changing the causal factors, pedestrian and bicyclist behavior. However, humans are the least mutable component of any system, so changing behavior is difficult. Theoretically, the simplest and cheapest solution is to educate road users to behave themselves. Such programs feel good, but they have little influence because they conflict with real-world experience. Psychology's basic rule is Thorndike's "Law Of Effect" - people act to promote desirable and to avoid undesirable consequences. People learn the contingencies, the relation between behavior and outcome. Convincing people to behave in the face of catastrophic but very rare outcomes is a major safety problem. Humans exhibit "practical drift" [48], a gradual moving away from the safety rules because they are rewarded with convenience and efficiency while suffering no catastrophic bad outcomes. This "contingency trap" teaches people to act "unsafely". For pedestrians, it currently is perfectly rational to cross illegally, dart out, "drink walk", etc. given the low probability of harm or negative consequence. Given enough pedestrians, however, this bound to mount up to a considerable number.

The only solution is enforcement, which changes the contingencies by creating smaller, but higher probability, negative outcomes, such as fines, etc. This requires enforcing current jaywalking laws and adding new ones on distracted and drink walking and drink pedaling. Australia has already instituted "sweeper patrols", to look for drunken pedestrians. Spain is considering classifying pedestrians as "road users" subject to the same drinking laws as drivers. No one pays much attention to laws that create inconvenience but are never enforced.

A third possibility is changing the environment to physically prevent unwanted pedestrian behavior. The ideal solution would be segregation of pedestrians from vehicles, but this is impractical as a general solution. Tunnels and overpasses are an option, but they are expensive and pedestrians frequently ignore them due to their inconvenience. The steps are also a problem for older pedestrians. Barriers are also a limited possibility. Some streets in Singapore are lined with fences that prevent mid-block crossing. Las Vegas has used barriers to force pedestrians onto overpasses. Mid road safety islands would aid older and slower pedestrians who cannot cross in one light cycle.

For bicyclists, a few dedicated lanes are the obvious solution. They should also be on secondary, not primary roads, to minimize interference and vehicle conflicts and to keep speed differentials low. However, they are not a panacea any more than the current dedicated pedestrian lanes called "sidewalks". Moreover, not all studies find that bicycle lanes improve safety [e.g., 49]. The lanes likely induce higher speed, which discourages bicyclist traffic sign/signal compliance, leaves less driver response time and endangers pedestrians. If provided, bicyclists using the lanes should be subject to licensing, strict law enforcement (contingencies) and required to wear high visibility clothing. France already requires bicyclists to wear high-viz clothing. Some other countries and US states are considering similar laws.

Lastly, increasing road lighting would be a simple, but highly effective measure. Seeing dark pedestrians in dim light is simply too difficult [50, 51]. Moreover, the increased number of older drivers with their cloudy ocular media will magnify the visibility problem in coming years. Pedestrians are unlikely to start wearing more conspicuous clothing anytime soon and glare concerns limit headlamp effectiveness. Higher street lighting is the only currently viable solution.

Many will balk at the energy cost and light pollution, but safety is never free. The only real issues are "How much is it worth to save one more pedestrian?" and "Who is going to pay?"

CONCLUSION

For many, the analysis of pedestrian collisions begins and ends with the fact that the pedestrians suffer the injuries. While true, the bad outcomes for pedestrians do not necessarily imply bad behavior by drivers. It takes two to make a collision, an important but often ignored fact. The scientific research shows that of the two, the pedestrian is the primary cause of pedestrian fatalities. Moreover, first-hand human factors investigation that has examined what is actually happening out on the roadway confirms the research findings.

This conclusion will, of course, be disputed by anti-car ideologues. "VRU" is really just a code for "victim road user", as if they have no control over their fates. Saying that the pedestrians cause the collisions is tantamount to blaming the victim. Someone else must be to blame, so a bull's eye is drawn around drivers. However, the concepts of "victim" and "blame" are not scientific or productive. Blame and responsibility are not inherent in events but are rather socially constructed and the result of the language used to describe the event [52]. They do not help in understanding why collisions occur or suggest effective ways of reducing them. Only examining causation and human factors can.

The pedestrian causation will be especially disputed by the legal blame industry that profits from the "reverse onus", the reversal of normal Western jurisprudence that presumes a person innocent until proven guilty. A Driver striking a pedestrian is presumed guilty until proven innocent. The law apparently believes that threats of blame will make them more careful. No evidence supports this assumption. The reverse onus law is misguided and ineffective and simply unfair.

However, the anti-car ideology runs even deeper than victimhood and blame. Cars are just plain evil. Just below the surface of pedestrian and especially bicyclist safety discussions is a thinly disguised hint that cars are inhuman and unnatural. Cars "encourage its owner to get fat and unhealthy" and "even if they never crashed", they still require "obliging healthy and harmless walkers to yield priority to inactive and polluting drivers" [53]. Even this is just a rationalization. The real ideology roots are expressed in the Tarkington quote above and in the "return to Eden" mentality that similarly trumpets organic food, homeopathic remedies, paleo diets, etc., etc. Levinovitz [54] noted that this ideology leads to arguments characterized by three attributes:

  • based on a paradise past narrative;
  • promising a simple cure for multiple problems; and
  • grounded in a narrative of conspiracy and of good and evil.

    • These arguments are as old as the bible, so the quasi-religious nature of the anti-car campaign shows through. As is typical of all such arguments, they are designed to appeal to the unanalytical "reptilian" brain rather the analytical, scientific and logical "mammalian" brain. They bear no relation to safety, but they are there in the background demonizing drivers and steering the search for solutions.

    In sum, the answer to the title's question, "Why pedestrians die?" is ideology. Despite the overwhelming evidence that previous strategies have reached an asymptote and that pedestrians cause most collisions, safety measures continue the failed strategy of focusing on drivers and outcome to the almost complete exclusion of pedestrians and causation. Why?

    The entire issue of pedestrian and bicyclist safety is clouded by ideology at the expense lower fatality rates. The ideology sees pedestrians as victims and drivers as Philistines, who should be blamed for the unnatural act of driving a car. Perhaps a more apt biblical metaphor is that drivers and their cars are serpents who have despoiled the Garden of Eden where mankind lived in a perfect natural state, in tune with nature, eating only organic food and drinking only fair trade coffee.

    Some may object that an issue such as ideology has no place in a discussion of road fatalities. On the contrary, it is unavoidable and essential. It is unavoidable because the anti-car zealots have already made it the overriding plank in their anti-car campaign. It is essential because virtually all understanding of the world is filtered through an ideology, and this includes pedestrian fatalities and their possible remedies. For pedestrian safety, as well as for most problems, such ideology is dangerous because it promotes unreality and, as Charles Reich observed, "Unreality is the true source of powerlessness. What we do not understand, we cannot control."

    REFERENCES

    [1] NATIONAL HIGHWAY TRANSPORTATION SAFETY ADMINISTRATION. Traffic Safety Facts 2014 Data: Pedestrians. Washington, D. C.: NHTSA, 2016.

    [2] GREEN, M. Roadway Human Factors: From Science To Application, secondedition. Forthcoming 2022.

    [3] PREUSSER, D., J. WELLS, A. WILLIAMS, AND WEINSTEIN, H.. Pedestrian Crashes In Washington, DC and Baltimore. Accident Analysis and Prevention 703-710, 2002.

    [4] CINNAMON, J., SCHUURMAN, N., & HAMEED, S. M. Pedestrian injury and human behaviour: observing road-rule violations at high-incident intersections. PloS 1, 6, e21063, 2011.

    [5] HUNTER, W., STUTTS, J., PEIN, W. & COX, C. 1995. Pedestrian And Bicycle Crash Types Of The Early 1990's. Report FHWA-RD-95-193, 1995.

    [6] SPAINHOUR, L., I. WOOTTON, J. SOBANJO, AND P. BRADY. 2006. Causative Factors and Trends in Florida Pedestrian Crashes. Transportation Research Record 1982, 90-98, 2006.

    [7] LEE, C., & ABDEL-ATY, M. 2005. Comprehensive analysis of vehicle-pedestrian crashes at intersections in Florida. Accident Analysis & Prevention, 37(4), 775-786, 2005.

    [8] SIMON, M. C., HERMITTE, T., & PAGE, Y. 2009. Intersection road accident causation: A European view. In Proceedings of 21st (ESV) International Technical Conference on the Enhanced Safety of Vehicles, 1-10, 2009.

    [9] ULFARSSON, G. & BOOTH K. 2010. Analyzing fault in pedestrian-motor vehicle crashes in North Carolina. Accident Analysis And Prevention, 42,1805-1813, 2010.

    [10] HARRUFF, R.C., AVERY, A. & ALTER-PANDYA, A.S. Analysis of circumstances and injuries in 217 pedestrian traffic fatalities. Accident Analysis and Prevention, 30, 11-20, 1998.

    [11] MCCARROLL, J. R., BRAUNSTEIN, P. W., COOPER, W., HELPERN, M., SEREMETIS, M., WADE, P. A., & WEINBERG, S. B. Fatal pedestrian automotive accidents. JAMA: The Journal of the American Medical Association, 180, 127-133, 1962.

    [12] SHINAR, D. Psychology of the Road: New York: Wiley, 1978.

    [13] YANAGISAWA M, SWANSON E & NAJM W. G. Target Crashes and Safety Benefits Estimation Methodology for Pedestrian Crash Avoidance/Mitigation System. NHTSA DOT-VNTSC-NHTSA-13-02, 2014.

    [14] MCCARROLL, J. R., BRAUNSTEIN, P. W., COOPER, W., HELPERN, M., SEREMETIS, M., WADE, P. A., & WEINBERG, S. B. Fatal pedestrian automotive accidents. JAMA: The Journal of the American Medical Association, 180, 127-133, 1962.

    [15] MCLEAN, A. In depth study of pedestrian accidents. In Joint ARRB/DOT Pedestrian Conference, 1978, Sydney, Australia, 1978.

    [16] MARTIN, A. Factors Influencing Pedestrian Safety: A Literature Review. TRL, 2006.

    [17] DEWAR, R. Pedestrian and bicyclists. In R. Dewar and P. Olson (eds.) Human Factors In Highway Safety. Tucson, AZ: Lawyers & Judges Publishing, 2002.

    [18] AYRES, T. J., KELKAR, R., KUBOSE, T., & SHEKHAWAT, V. Bicyclist Behavior at Stop Signs. In Proceedings of the Human Factors and Ergonomics Society Annual Meeting (Vol. 59, No. 1, pp. 1616-1620), 2015.

    [19] TUCKEL, P. & MILCZARSKI, W. An Observational Study of biking behavior in lower and central Manhattan. University Of New York Hunter College, 2014.

    [20] LANGFORD, B. C., CHEN, J., & CHERRY, C. R. Risky riding: naturalistic methods comparing safety behavior from conventional bicycle riders and electric bike riders. Accident Analysis & Prevention, 82, 220-226, 2015.

    [21] THOM, R., & CLAYTON, A. Low-cost opportunities for making cities bicycle-friendly based on a case study analysis of cyclist behavior and accidents. Transportation Research Record, 1372, 90-101, 1992.

    [22] HERMAN, M. Cyclists and city streets. Bicycle blueprint: A plan to bring bicycling into the mainstream in New York City. New York: Transportation Alternatives, 26-29, 1993.

    [23] SHINAR, D. Traffic Safety and Human Behavior. Bingley UK: Emerald Group, 2007.

    [24] MATSUI, Y., TAKAHASHI, K., IMAIZUMI, R., & ANDO, K. Car-to-pedestrian contact situations in near-miss incidents and real-world accidents in Japan. In 22nd International Technical Conference on the Enhanced Safety of Vehicles, paper (No. 110164), 2011.

    [25] SNYDER, M. 1972. Traffic engineering for pedestrian safety. Some new data and solutions. Highway Research Record, 406, 21-27, 1972.

    [26] IRYO-ASANO, M., ALHAJYASEEN, W., ZHANG, X., & NAKAMURA, H. Analysis of pedestrian speed change behavior at signalized crosswalks. In 2015 Road Safety & Simulation International Conference, October 6th? 8th, Orlando, USA, 2015.

    [27] WESTERHUIS & DE WAARD, D. Reading cyclist intentions: Can a lead cyclist's behaviour be predicted? Accident Analysis And Prevention, 2014.

    [28] KNOWLES, J., SMITH, L., CUERDEN, R., & DELMONTE, E. Analysis Of Police Collision Files For Pedestrian Fatalities In London, 2006-10 (No. TfL2520), 2012.

    [29] GRAYSON, G. The Hampshire child pedestrian study. Department Of The Environment, Transport And Road Research Laboratory. Report 66. TRRL, 1975.

    [30] JENNINGS, R.D., M.A. BURKI & ONSTINE, B.W. Behavioural observations and the pedestrian accident. Journal of Safety Research, 9, 27-33, 1977.

    [31] TYRRELL, R., J. BROOKS, J. WOOD, AND CARBERRY, T. Nighttime Conspicuity from the Pedestrian's perspective Transportation Research Board Annual Meeting CD, 2004.

    [32] WOOD, J., LACHEREZ, P., MARSZALEK, R. & KING, M. Drivers' and cyclists' experiences of sharing the road: Incidents, attitudes and perceptions of visibility. Accident Analysis & Prevention, 41, 772-776, 2009.

    [33] GUTH, D., ASHMEAD, D., LONG, R., WALL, R., & PONCHILLIA, P. Blind and sighted pedestrians' judgments of gaps in traffic at roundabouts. Human Factors, 47(2), 314-331, 2005.

    [34] SULLIVAN, J. M., & FLANNAGAN, M. J. The role of ambient light level in fatal crashes Accident Analysis & Prevention, 34, 487-498, 2002.

    [35] OLSON, P. L. Visibility problems in nighttime driving. SAE Paper No. 870600, 1987.

    [36] LEIBOWITZ, H. W., & OWENS, D. A. We Drive by night; And when we do we often misjudge our visual abilities, courting disaster, Psychology Today, 20, 54-58, 1986.

    [37] LAHRMANN, H., & MADSEN, T. K. O. The safety impact of a yellow bicycle jacket. Safety Science, submitted for publication, 2016.

    [38] BRUMFIELD, R., & PULUGURTHA, S. S. When distracted road users cross paths. Public Roads, 75(3). https://www.fhwa.dot.gov/publications/publicroads/11novdec/01.cfm, 2011.

    [39] FITZPATRICK, C. D., GOMEZ, R. A., PIRES, M., & KNODLER, M. A. The prevalence of distracted walking and its effect on driver behavior. Advances in Transport Studies, (39), 2016.

    [40] WILBUR, M. H., & SCHROEDER, P. B. Distracted bike riding and its discontents: Findings from the 2012 National Survey of Bicyclist and Pedestrian Attitudes and Behavior. In Transportation Research Board 93rd Annual Meeting (No. 14-1765), 2014.

    [41] TERZANO, K. Bicycling Safety and Distracted Behavior in The Hague, the Netherlands. Accident Analysis & Prevention, 57, 87-90, 2013.

    [42] LUOMA, J., & PELTOLA, H. Does facing traffic improve pedestrian safety?. Accident Analysis & Prevention, 50, 1207-1210, 2013.

    [43] KING, M.J., WOOD, J.M., LACHEREZ, P.F. & MARSZALEK, R.P. Optimism about safety and group-serving interpretations of safety among pedestrians and cyclists in relation to road use in general and low light conditions. Accident Analysis & Prevention, 44(1), 154-159, 2012.

    [44] WALKER, I., Psychological factors affecting the safety of vulnerable road users http://people.bath.ac.uk/pssiw/traffic/Walkerreview.pdf. Downloaded 1 November 2016.

    [45] JOHANSSON, C., GARDER, P., LEDEN, L. The effect of change of code on safety and mobility for children and elderly as pedestrians at marked crosswalks-A Case Study Comparing Sweden to Finland. Paper no. 04-2659 in the 83rd Annual TRB Meeting, 2004.

    [46] BUNN, F., COLLIER, T., FROST, C., KER, K., ROBERTS, I., & WENTZ, R. Area-wide traffic calming for preventing traffic related injuries. Cochrane Database Review, 1(1), 2003.

    [47] GOODWIN, A, KIRLEY, B., SANDT, L., HALL, W., THOMAS, L., O'BRIEN, N. & SUMMERLIN, D. Countermeasures That Work: A Highway Safety Countermeasure Guide for State Highway Safety Offices, Seventh Edition. DOT HS 811 727, 2013. v [48] SNOOK, S. A. Friendly Fire: Princeton University Press, 2011.

    [49] WEE, J. H., PARK, J. H., PARK, K. N., & CHOI, S. P. A comparative study of bike lane injuries. Journal of Trauma and Acute Care Surgery, 72(2), 448-453. 2012.

    [50] GREEN, M., Allen, M.J., Abrams, B.S., & Weintraub, L. Forensic Vision: With Application To Highway Safety. Tucson: Lawyers & Judges Publishing Company, 2008.

    [51] GREEN, M. Seeing Pedestrians at Night. Impact - Journal Of The Institute of Traffic Accident Investigators UK, 20, 14-17, 2012.

    [52] DEKKER, S. (2009). Just culture: who gets to draw the line? Cognition, Technology & Work, 11(3), 177-185.

    [53] WALKER, I., https://thepsychologist.bps.org.uk/volume-25/edition-9/interview-vulnerable-road-users. Downloaded 3 January, 2017.

    [54] Levinovitz, A. The Gluten Lie: And Other Myths About What You Eat. Regan Arts. 2015.

    [55] ZHU, M., ZHAO, S., COBEN, J. H., & SMITH, G. S. Why more male pedestrians die in vehicle-pedestrian collisions than female pedestrians: a decompositional analysis. Injury Prevention, 19(4), 227-231, 2014. [56] ORSI, C., FERRARO, O. E., MONTOMOLI, C., OTTE, D., & MORANDI, A. (2014). Alcohol consumption, helmet use and head trauma in cycling collisions in Germany. Accident Analysis & Prevention, 65, 97-104.

    Endnotes

    1In response to some comments that I have received on this article, yes I am a pedestrian. In fact, our family has one car, which is 16 years old with only 110 km (68,000 miles) on it.

    2However, the latest statistics show an significant uptick in pedestrian fatalities. The likely causes are discussed in Why Are Pedestrian Fatalities On The Rise?

    3See for example: This, this and this.

    4This list is somewhat outdated and primarily reflects data on cell phone conversations. More recent data on texting, emailing etc. show even more profound distraction effects, as described in this page.

    A division of 2057949 Ontario, Inc.
    Copyright © 2024 Marc Green, Ph. D.
    Home Page
    Contact Us