Why Looking at Crash Stats Alone Doesn’t Tell the Whole Story About Pedestrian Safety | Streetsblog

Stephen Miller at Streetsblog writes Why Looking at Crash Stats Alone Doesn’t Tell the Whole Story About Pedestrian Safety: New research from Minneapolis shows that there is safety in numbers for pedestrians.

This map shows the number of crashes involving pedestrians in relation to the number of people who walk at a given location. Via Murphy, Levinson, and Owen

This map shows the number of crashes involving pedestrians in relation to the number of people who walk at a given location. Via Murphy, Levinson, and Owen

Some intersections are riskier to cross than others, but looking at the number of pedestrian injuries alone doesn’t tell the whole story. A new study from Minneapolis combines crash data with pedestrian counts to deliver a more nuanced picture of traffic dangers for people on foot. Among the findings: There’s safety in numbers for pedestrians.

Using data from the city government, University of Minnesota researcher Brendan Murphy and his co-authors looked at 448 intersections where both pedestrian counts and automobile counts were available, then cross-referenced that data with the city’s crash reports. They found a strong negative correlation between the number of pedestrians and the risk of being hit by a car.

While the study found people are less likely to be struck by a driver at locations where lots of people walk, it does not establish causation, Murphy says. “We don’t have good statistical evidence to show that if a place is safe, people will walk — or in the other direction, that if people are walking, they make the place safer,” he says. “I personally think it’s a bit of both.”

Per person, pedestrian-rich areas downtown and near the University of Minnesota pose a low risk for people walking, though they have a high absolute number of pedestrian crashes. Quieter intersections in more residential neighborhoods also pose a lower risk.

A few streets jump off the map as high-risk areas, like Lake Street, which runs east-west across South Minneapolis, and Penn Avenue in North Minneapolis. Both are used by a steady if not enormous number of pedestrians, but are meant first and foremost to move lots of cars. “We can ask, ‘How are those roads designed?’” Murphy says. “They are two lanes each way, no shoulder or bike lane.”

The study looked at all crashes involving pedestrians, not just injuries and fatalities, in order to include enough data points to reach reliable conclusions. It also looked at the stats from 2000 to 2013 in aggregate, rather than year-by-year, so it doesn’t take into account intersection redesigns or major changes like the opening of a light rail line. If there were enough data, Murphy says, “it would be really nice to do a year-by-year analysis.”

The study did not consider the relationship between pedestrian risk and income or race, but the authors say that needs attention. “Equity is a very big problem in terms of pedestrian safety and poor and minority people are getting killed by cars at much higher rates,” Murphy said.

The authors hope their research will lead to better measurements of pedestrian safety and methods to improve it. In 2016, the U.S. Department of Transportation’s four-year strategic plan set a goal of reducing fatalities for pedestrians and cyclists to 0.15 per 100 million vehicle miles traveled by 2016. But that’s the wrong way to look at the problem.

“If we frame pedestrian deaths in terms of VMT, we’re really framing it in terms of automobiles themselves and car traffic,” said Murphy. “We should be focused on reducing pedestrian deaths as a percentage of the pedestrian population.”

There’s also a need for better data collection. Cities and states regularly collect standardized data on car and truck traffic, but there’s no standard for non-motorized users. This data is often collected manually and its reliability varies from city to city. In Minneapolis, three counts throughout the day at each intersection were added together to create a six-hour total. Other cities have different methods.

“Ideally we would like to have our cities wired up and know how many pedestrians are crossing each intersection,” Murphy says. “We need to focus in on the pedestrian population and really ask ourselves, where are they really experiencing undue burdens of risk and what can we do about it?”

On Blood Alcohol Content

Candace Lightner, founder of Mothers Against Drunk Drivers recently published a counter-intuitive op-ed against lowering the blood alcohol content (BAC).

Hopefully everyone agrees that if there were fewer drunk drivers on the road, there would be fewer deaths from drunk driving. Hopefully everyone also agrees that BAC is correlated with impairment. The blood alcohol content limit, currently 0.08 in the US, is 0.05 in many other countries of the world. Should the US lower the BAC?1280px-Ljubljana_car_crash_2013

The argument against is that pulling over safe drivers (say in a police screenline, where all drivers on a road are pulled over and briefly tested) takes police resources that could be better spent pulling over observed dangerous drivers. Lightner writes: “Every dollar spent enforcing DUI laws against sober drivers is one not spent on getting the worst offenders off our roads.” Perhaps 2 drivers at 0.05 BAC are less dangerous than 1 driver at 0.10, so spend the time finding that driver.

But such police screenlines have the effect not just immediately about arresting people in violation of the law, and also as warning, reminder, and deterrent against future alcohol (and drug) impaired driving. To say the resources are a waste misses a major point.

International experience shows most other developed countries have significantly lower crash and fatality rates than the US, and they have 0.05 or lower BAC. Perhaps the US should just copy their traffic laws lock, stock, and barrel. Researchers have estimated ‘an additional 538 lives could be saved each year if the United States reduced the limit to 0.05,’ (Wagenaar et al. 2007)

Casual drinkers are a problem. Social drinking is a problem. I don’t care if you drink at home and don’t bother anyone (aside from the health insurance claims you impose on society from the damage you do to yourself), but when you drive a car, you endanger others. And because you are impaired, you don’t have the reasoning abilities to realise this.

The rules of the road should not only punish, but also provide a strong deterrent, which includes arrest and punishment even if you didn’t actually kill someone this time. Until robots fully rule the roads in 25 years, possibly another million Americans will be killed in car crashes. We can avoid tens of thousands of them with lower BAC limits.


The scientific evidence on this is fairly clear: Fell and Voas (2006) write:

Purpose

This scientific review provides a summary of the evidence regarding the benefits of reducing the illegal blood alcohol concentration (BAC) limit for driving and providing a case for enacting a .05 BAC limit.

Results

Fourteen independent studies in the United States indicate that lowering the illegal BAC limit from .10 to .08 has resulted in 5–16% reductions in alcohol-related crashes, fatalities, or injuries. However, the illegal limit is .05 BAC in numerous countries around the world. Several studies indicate that lowering the illegal per se limit from .08 to .05 BAC also reduces alcohol-related fatalities. Laboratory studies indicate that impairment in critical driving functions begins at low BACs and that most subjects are significantly impaired at .05 BAC. The relative risk of being involved in a fatal crash as a driver is 4 to 10 times greater for drivers with BACs between .05 and .07 compared to drivers with .00 BACs.

Summary

There is strong evidence in the literature that lowering the BAC limit from .10 to .08 is effective, that lowering the BAC limit from .08 to .05 is effective, and that lowering the BAC limit for youth to .02 or lower is effective. These law changes serve as a general deterrent to drinking and driving and ultimately save lives.

 

Evaluating the Safety In Numbers effect for pedestrians at urban intersections

Recently published:

Average annual 6-hour pedestrian count by location, Minneapolis
Average annual 6-hour pedestrian count by location, Minneapolis

Highlights

  • Collision risk at 448 intersections in the city of Minneapolis, MN was assessed.
  • The Safety In Numbers phenomenon was observed for both pedestrians and cars.
  • Maps of per-pedestrian crash rates inform discussion of safe vs. unsafe city areas.

Abstract

Assessment of collision risk between pedestrians and automobiles offers a powerful and informative tool in urban planning applications, and can be leveraged to inform proper placement of improvements and treatment projects to improve pedestrian safety. Such assessment can be performed using existing datasets of crashes, pedestrian counts, and automobile traffic flows to identify intersections or corridors characterized by elevated collision risks to pedestrians. The Safety In Numbers phenomenon, which refers to the observable effect that pedestrian safety is positively correlated with increased pedestrian traffic in a given area (i.e. that the individual per-pedestrian risk of a collision decreases with additional pedestrians), is a readily observed phenomenon that has been studied previously, though its directional causality is not yet known. A sample of 488 intersections in Minneapolis were analyzed, and statistically-significant log-linear relationships between pedestrian traffic flows and the per-pedestrian crash risk were found, indicating the Safety In Numbers effect. Potential planning applications of this analysis framework towards improving pedestrian safety in urban environments are discussed.

Keywords

  • Pedestrians;
  • Safety;
  • Collisions;
  • Urban planning

Accepting risks

People died in a barbaric terrorist act recently at [insert terrorist act here]. That is terrible news. I wish it didn’t happen. It happens far too frequently.

We ask collectively “Could it have been prevented?” This question is more than idle curiosity, as it informs the follow-up “Can future terrorist acts be prevented?”

I highly doubt both of these propositions. They assume some fantastical superhero-like state, with the strength of an all-seeing and all-knowing Allfather. As much money as the state security apparatus gets, … and this is enough to monitor a lot, it still doesn’t bug my office, and would be very bored if it did. As much computer power as it has, it still cannot predict where I am going this afternoon.

Since we don’t have such a state, some people are proposing giving the state more powers so that we can naively feel comfortable in our security, foregoing our freedoms. The state will be at least a more-seeing and more-knowing Most-father.

Strategies proffered, like banning encryption (if encryption is outlawed, only outlaws, and the government, will have encryption), or registering people based on their religion or ethnicity, or building a wall, or prohibiting refugees are unenforceable or miss the point entirely. These are mostly nonsense ideas which might sound good if you live your life in fear because you watch too much news on television. Will ever more power for the security state really make a difference in our security? Like everything else, there are diminishing returns to investments in security.

Someone with a modicum of skill, who is determined to kill you, and is willing to lose their own life, will kill you.

No – kill them first.

Leaving aside the constitutional problem of killing people who have not actually committed a crime in the United States, is the physical problem of precognition. The state cannot actually monitor everyone all the time, with humans. Imagine 50% of the people were guardian/spies listening in on the other 50%. Do you trust the first 50%? This is such an old problem, there is a Latin phrase for this: Quis custodiet ipsos custodes?

Even the most successful state security systems of fascist and communist countries still faced assassination attempts and revolutions from time to time. Perhaps this omni-state can reduce the likelihood of success of terrorist attempts, but it cannot ever eliminate the possibility. All we are arguing about is degree of security. Frankly, we are fairly safe now. Total global violence is near an all-time low. We should aim for zero deaths from violence, but efforts to reach zero are not without costs, which reduce the possibilities of other kinds of improvements.

Then there are the costs in human life of such a security state, which are often higher then the actual costs of barbaric terrorist acts that justify them. While these are not directly comparable numbers, I will compare them anyway, as getting a sense of the magnitudes of risks is important. In the US, the police killed over 1000 people so far this year. Terrorists  killed 3.  The September 11 attacks killed nearly 3000 people.

Certainly there is the risk terrorists will kill more. Maybe they will get the bomb, or poison the water supply. My 15 year old self was fairly sure that some city would have been nuked by now.

We want to avoid these mega-risks. But we also want to avoid run-of-the-mill non-political intentional killing (US homicides 16121 in 2013 and suicides  41149) and car crashes (33804).  We conspicuously don’t confiscate guns or cars. We accept certain risks as the cost of doing business.

There is also the risk that in an enhanced superstate, the police and military will kill more Americans than they do now. Increasing deportation rates of undocumented residents will increase the mistaken deportation of legal residents and citizens. Increasing the level of policing will likely increase the number of innocent people killed or jailed by the police.

Bring the fight to them, so we don’t have to fight them here.

We have of course brought the fight to them.

The number of coalition soldiers lost in the Afghanistan war (Operation Enduring Freedom) (3506) and Iraq wars (Operation Iraqi Freedom) (4814), much less the number of Aghanis and Iraqis.

Had there been no Operation Iraqi Freedom, there would likely be no Daesh. And we continue to pummel them with air strikes and drone attacks, day after day. We could increase these numbers, but face the same problem. We target the “best” targets first (the one with the greatest likelihood of getting the bad guys without mistakenly getting the innocents and creating future terrorists in a multi-generational war), the next best targets second, and so on, until the last target we have little confidence we won’t be mostly killing innocents. If we killed everyone in the country, sure we would kill all the barbarians, but we would kill the innocents too. That is not the American way.

Small amounts of terrorism are among the unfortunate costs of living in the  modern world. Responding by sacrificing our freedoms and the opportunities for others to live in America is a self-inflicted wound that fails to treat the disease it was supposed to cure.

Evaluating the “Safety In Numbers” Effect With Estimated Pedestrian Activity

Recent working paper:

Pedestrian risk vs. PM pedestrian flow
Pedestrian risk vs. PM pedestrian flow

 

Pedestrian and bicyclist collision risk assessment offers a powerful and informative tool in urban planning applications, and can greatly serve to inform proper placement of improvements and treatment projects. However, sufficiently detailed data regarding pedestrian and bicycle activity are not readily available for many urban areas, and thus the activity levels and collision risk levels must be estimated. This study builds upon other current work by Murphy et al. (1) regarding pedestrian and bicycle activity estimation based on centrality and accessibility metrics, and extends the analysis techniques to estimation of pedestrian collision risk. The Safety In Numbers phenomenon, which refers to the observable effect that pedestrians become safer when there are more pedestrians present in a given area, i.e. that the individual per-pedestrian risk of a collision decreases with additional pedestrians, is a readily observed phenomenon that has been studied previously. The effect is investigated and observed in acquired traffic data, as well as estimated data, in Minneapolis, Minnesota.

Rising gas prices might make us safer

Tim Harlow at the Star Tribune Drive column reports on our research. Rising gas prices might make us safer

Death, prices correlated

Researcher Guangqing Chi of the Department of Sociology and Rural Studies at South Dakota State University looked at the correlation between gas prices and traffic safety. In a study examining crash data in Minnesota from 1998 to 2007, Chi found that a 20-cent drop in gas prices resulted in 15 more fatalities a year. Conversely, he found that a 20-cent increase would bring a decrease of 15 deaths annually.

The study also found that as gas prices rise, the crash rate per million miles traveled dropped in urban and rural areas. It found higher gasoline prices also have significant effects in reducing property damage and injury crashes.

In another study using data from Alabama and Mississippi, Chi found higher gas prices had the biggest effect on teens. With their lower incomes, teens are discouraged from driving by high gas prices and that reduces their crash rate. That makes the roads safer for other drivers, he said.

When fuel prices skyrocketed to more than $4 a gallon in 2008, many drivers drove less frequently and perhaps less aggressively, which reduced their chances of having a crash, the study said.

The bottom line is that when gas prices go up, “we suspect that people drive more carefully,” Chi said.

Read the paper here: Chi, Guangchi, Mohammed Quddus, Arthur Huang and David Levinson (2013) Gasoline Price Effects on Traffic Safety in Urban and Rural Areas: Evidence from Minnesota, 1998–2007. Safety Science 59: pp. 154-162

Safety is a shared responsibility

A shared space street in Bern, allowing only bikes, pedestrians and transit. http://pedbikesafetyinternationalscan.blogspot.com/2009/05/shared-space-street-in-bern-allowing.html

In the constant exhortations to pedestrians around cars and trains, we hear “Safety is a shared responsibility“. This of course is true. Many crashes are the product of a chain of failures. The driver was too fast for conditions. The driver did not pay attention. The pedestrian did not pay attention. Someone else did not pay attention and braked sharply, and someone behind them swerved because they were following too close and hit a third person. And so on. Yet authorities are often quick to blame the victim, rather than the system design.

The more you hear the exhortation, the less effective it becomes (diminishing returns set in), much like the security threat warnings of the Bush Administration, telling us at the airport we were at threat level orange, constantly.

People are imperfect. They feel they have better things to be doing then looking out for lurking dangers around every corner. They did not evolve to operate in a city with multi-ton machines operating at speeds faster than the fastest land animals. They see meaningless signs and signals and learn to ignore them. Breaking traffic and pedestrian laws may be illegal, but it is hardly immoral – we don’t feel guilty when we conscientiously don’t allow ourselves to be governed by degrading light bulbs implemented by unthinking bureaucracies.

A shared space street in Bern, allowing only bikes, pedestrians and transit. http://pedbikesafetyinternationalscan.blogspot.com/2009/05/shared-space-street-in-bern-allowing.html
A shared space street in Bern, allowing only bikes, pedestrians and transit. http://pedbikesafetyinternationalscan.blogspot.com/2009/05/shared-space-street-in-bern-allowing.html

Designs for systems that involve people should consider human imperfection. Ideally systems are forgiving of human error.  Light rail trains, e.g, are much more dangerous than buses. (Cars are too). They are far less tolerant of imperfection, as they can neither brake quickly (due to mass)  nor swerve (due to tracks), and are more deadly on impact (again due to mass).

There are two good strategies for multi-modal travel within a finite space:

  • keep them separated and
  • mix them slowly

(The third strategy: mix them quickly will lead to tragedy as long as people are making decisions, rather than machines.)

The “keep them separated” strategy is why safety on interstate highways is much better than other streets. High speed vehicles are interacting with other high speed vehicles, but low speed vehicles are prohibited. It is far from perfect, but better than the previous alternative. So much so that when speed limits were raised in the 1980s, overall safety went up as drivers were attracted off much more dangerous roads onto the interstate, which was only marginally more dangerous with the higher speed limit, and now less likely to result in a speeding ticket. Security theater that deters people from flying and encourages them to drive is more dangerous than the original threat.

Newer subway systems (such as the shuttle at  MSP airport between the terminal and the LRT station/parking ramps), have glass barriers preventing people from accidentally falling on the track. Despite running one train every 90 seconds or so for over 10 years, I have not heard of any incidents with this system.

The Minneapolis-St. Paul region has chosen, for the most part, not to build grade separated transit systems. They are certainly more expensive, even if more beneficial (safer and faster). That leaves the strategy of “mix them slowly.”

Proponent of shared spaces, the late Hans Monderman has a famous quote “When you treat people like idiots, they’ll behave like idiots.”

As the song goes “Signs, signs, everywhere signs”. Each sign and signal degrades  the effectiveness of all the others. Monderman went for a sign and signal-free approach, using design to guide people and vehicles through town centers. In this scheme pavements give people the guidance they need.

While peer-reviewed evaluations of shared spaces have been limited, Kaparias et al. from Imperial College, evaluating Exhibition Road, say: “The results suggest that pedestrians feel most comfortable in shared space under conditions which ensure their presence is clear to other road users – these conditions include low vehicular traffic, high pedestrian traffic, good lighting and pedestrian-only facilities. Conversely, the presence of many pedestrians and, in particular, children and elderly, makes drivers feel uneasy and, therefore, enhances their alertness.”  Subsequent research by the team finds “The results of the comparative analysis indicated a general decrease in traffic conflict rates as a result of the redesign but also highlighted specific issues that may require additional analysis”

Karndacharuk et al. write: “A comparative analysis of the data after implementation highlights the importance of the active frontage in enabling a lower (vehicular) speed environment in relation to the number of pedestrians within the shared space.”

In short, design matters. Over-engineering can be as great or even greater sin than under-engineering. The best design is not necessarily more gadgets, instructions, rat runs, prohibitions on actions, closing of desire lines, or other devices constraining people from their intuitions.

Rather it is running with and shaping travelers natural instincts, so the environment is not chafing but accommodating. Safety is a shared responsibility, and those who diminish the effectiveness of safety tools such as signs and signals by their misuse and excessive exhortations which loosely spend people’s scarce attention are culpable as well.

 

Speed Control in Israel (updated)

When drivers pass by, they almost stop completely due to the holes and to avoid car damages.

The following appeared across the email transom. The source is unknown. The forwarder stated:

One of my Internet Buddies sent it claiming it to be an Isreali invention. Who knows. Maybe it’s a joke, but none-the-less it appears to be a great idea, unless of course people swerve out of their lane and cause an accident!

When drivers pass by, they almost stop  completely due to the holes and to avoid car damages.
When drivers pass by, they almost stop
completely due to the holes and to avoid car damages.
This is a strategy currently used in Israel as a high-speed control. It is more economical than using cameras, radar, police officers, etc.
This is a strategy currently used in Israel as a high-speed control.
It is more economical than using cameras, radar, police officers, etc.
They move them around every day!
They move them around every day!

Update: part of an ad campaign by Pioneer Suspension.

Hi Dr. Levinson, I searched for the message that was written on the road in the pictures and came up with this:

http://www.hoax-slayer.com/fake-potholes.shtml

Keeping the Green Line safe – The Minnesota Daily

Things I would not be allowed to say were I a public official … in The Minnesota Daily Keeping the Green Line safe :

“…

“Light rail trains empirically kill more people than buses,” said David Levinson, a civil engineering professor and transportation studies expert.

Metro Transit spokesman John Siqveland said the Green and Blue Lines were built with safety in mind and that the Green Line will generally run slower since it will travel through more densely populated areas.

The vast majority of light-rail and streetcar systems around the country run on street level, Siqveland said.

Metro Transit has two train routes and 125 bus routes. Of the 81 million rides Metro Transit gave last year, 86 percent of riders use buses and 14 percent use trains.

Since the Blue Line opened in June 2004, the agency’s buses have had 6,979 incidents, which resulted in five deaths. Like the light rail, the majority of collisions involved motor vehicles. Of the 145 bus incidents involving pedestrians, four were fatal.

“Buses serve a lot more people and run a lot more miles … but have fewer fatalities,” Levinson said. “The [fatality] rates are obviously much higher for light-rail trains. Still, it’s a lot safer than driving a car.”

The 11-mile light-rail route connecting downtown Minneapolis to downtown St. Paul is decades in the making and cost $957 million.

“It’s been built, so it’s too late. It’s an at-grade light-rail facility — there’s nothing we can do about it without spending another billion dollars,” Levinson said. “We don’t operate in this world to maximize safety and only safety; we have trade-offs.””

The low-friction nature of train travel makes it efficient, but it also makes stopping a problem, said Stephen Zitzow, Minnesota Traffic Observatory laboratory manager.

Each train weighs about 300,000 pounds, and bringing one to a complete stop from 55 mph takes the length of two football fields. At 20 mph, it takes 81 feet, Siqveland said.

“It doesn’t have the option to swerve out of the way of someone in the way,” Zitzow said. “The difference here is that most vehicles can maneuver much more than a light rail, which is stuck.”

Central Corridor project spokeswoman Laura Baenen said the Green Line will begin running before the Major League Baseball All-Star game at Target Field in July.

Baenen said the Metropolitan Council has taken many safety considerations, including creating a pedestrian mall and using posters and community educators.

Baenen said a street-level light-rail track is less costly than an underground or elevated one, which was “prohibitively expensive.”

Siqveland said part of the Metro Transit safety campaign will have links to its website posted at every train stop.

When asked if he’d visit the website, graduate student Hill said he most likely wouldn’t.

“If I am an example,” he said, “[other students] probably will not.”

See also this (presented at TRB) (which is a bit dated), still I think it is largely accurate. Another interesting tidbit is that urban autos kill relatively few people (on a per mile basis). More auto deaths (on a per capital, per mile, etc. basis) are rural.

Also note that the APTA Public Transportation Fact Book (p.26) does not break out LRT deaths, lumping that with rail transit. I wonder why.