A Few Thoughts About ‘On a Crash Course,’ by Miller & Zaloshnja

I’ve finally gotten around to reading ‘On a Crash Course,’ a report by Ted Miller and Eduard Zaloshnja that’s been getting a lot of play in the media. As the Post summarizes:

Bad highway design and conditions are a factor in more than half the fatal crashes in the United States, contributing to more deaths than speeding, drunken driving or failure to use seat belts, according to Ted R. Miller, who co-wrote the 18-month study released yesterday.

Road-related conditions were a factor in 22,000 fatalities and cost $217.5 billion each year, the study concludes. By comparison, Miller said, similar crashes where alcohol was a factor cost $130 billion, speeding cost $97 billion and failure to wear a seat belt caused losses of $60 billion.

Despite being sponsored by a consortium of road-building concerns, who naturally have a vested interest in highway improvements, there are some interesting and commendable points raised, or at least implied. The first is, given that road crashes bear a larger societal cost than congestion, we should be focusing whatever stimulus dollars (too many, in my opinion) are going to roads on indeed bringing up deficient roadways to modern safety standards, rather than building new roads. Unfortunately, this does not seem to be the case.

Another thing that caught my eye was the high figure of deaths attributed to roadway condition: “Roadway condition is a contributing factor in more than half—52.7 percent—of the nearly 42,000 American deaths resulting from motor vehicle crashes each year and 38 percent of the non-fatal injuries. In terms of crash outcome severity, it is the single most lethal contributing factor—greater than speeding, alcohol or non-use of seat belts.”

This surprised me, as any number of previous studies, including the famous (and much more comprehensive) Indiana Tri-Level Study, as pictured below, paint a different picture of causality.

Of course, this could be due to the difference in talking about what causes crashes, and what causes crashes that turn out to be fatal (this study, it should be noted, also relies on the “Large Truck Crash Causation Study” for data; it’s a well-regarded study, but the obvious problem is that it studied large trucks, which obviously have different characteristics than cars, and do the bulk of driving in different environments). This gets at the philosophy of the “forgiving road,” i.e., that road users should not pay a fatal price for all-too human mistakes. And, as Leonard Evans observed in Traffic Safety, the factors involved in crashes do not often correlate to the best countermeasures; just because “human factors” cause most crashes, we should not always try to fix the problem at the level of human factors. “An effective analogy points out that finding that mailed items are damaged by the human factor of careless handling does not mean improved handling is the most effective countermeasure. Better packaging is far more effective.” Call it “forgiving packaging.” (Another analogy given to me was “fire-safe cigarettes”; e.g., we can spend a lot of money trying to educate people not to smoke in bed, or punish those who do, or we can simply make cigarettes that are less likely to lead to fires — yes this may lead to risk compensation but if the overall result is more safety it’s still socially useful).

But let us note that Evans adds: “It is changes in driver behavior that have the potential to make, by far, the largest improvements in traffic safety.” It is interesting in this regard to note that the states that the authors claim as having the worst road conditions (they’re mostly in the south) are the very same states that have some of the worst seat-belt wearing rates and some of the highest drunk-driving rates. Building safer roads isn’t going to change those numbers.

And yes, clearly there is room for making many of America’s roads safer; many is the time I’ve been driving down, say, Highway 17 in New York near Bear Mountain, and wishing there was a barrier between myself and the opposing lane of traffic, which might be carrying a texting teen or a nodding-off salesman. And as I noted recently from Montana, I’m all in favor of cheap solutions that don’t lead to riskier behavior (like rumble strips). But on the other extreme, is it really society’s responsibility to guard-rail every last rural mile of highway in America, tear down every roadside tree, so that the drunk driver doesn’t run off the road and strike an obstacle? (One might make the ethical argument that to keep the drunk driver alive in this case is simply to raise the risk that he will eventually crash into another driver).

Another problem (which Elana Schor already raised, partially in my behalf), is exactly what kinds of roads we’re talking about fixing, and how. As I note in Traffic, Eric Dumbaugh in particular has done work showing that in many cases, so-called “forgiving road” environments actually have the perverse effect of bringing more crashes than areas that, by traditional safety engineering standards, are presumed to be more safe. Tree-lined streets in this regard are not a hazard; rather, they are the safety device — if one drives at the contextually chosen speed (and again, if one does not, is it society’s burden to prevent the outcome?)

“Forgiving road” is a rather expansive term, in any case; one way to make a road more “forgiving,” for all its users and not just drivers, is to lower speeds. This gives drivers more time to react to their mistakes, gives them a more forgiving crash outcome, and also makes the crash more forgiving when they happen to strike a pedestrian or a bicyclist or indeed another car.

Many of the things recommended in the report — wider shoulders, clear zones, etc. — have little place in many urban and suburban residential environments, unless one wants to engage in the old dream of the modernist architects (which too many traffic engineers have tried to do) — i.e., to “kill the street,” as Le Corbusier said, fully segregating all uses, allowing a dense city with fast highways, as pictured below.

Futurama was a car-company sponsored world’s fair folly, of course, and cities that tried to approximate it are still paying the price (or rushing to fix their mistakes). But the results coming from places like Kensington High Street in London are that many of the so-called traffic safety improvements simply led to worse behavior, and in their absence safety numbers for all modes have unexpectedly improved. Commenting on Swedish traffic safety, Neil Pearce notes that their approach is built around different strategies for different environments:

How did the Swedes do it? Tough seat belt and helmet laws, to be sure. But they’ve also begun to remake their roadways. Red lights at intersections (which encourage drivers to accelerate dangerously to “beat the light”) are being replaced with traffic circles. Four-foot high barriers of lightweight but tough mylar are being installed down the center of roadways to prevent head-on collisions, and as side barriers at critical locations. On local streets, narrowed roadways and speed bumps, plus raised pedestrian crosswalks, limit speeds to a generally non-lethal 20 miles an hour.

We, on the other hand, try to engineer safety onto our lower speed roads, which then play host to routine excessive speeds. Is this is a big deal on roads with lower-speed limits? The following graph notes that a surprisingly high amount of traffic fatalities due to speeding occur on lower speed roads (typically home to the most pedestrians and cyclists). Is the problem here lack of safety engineering, or lack of behavioral engineering?

There’s another procedural item that bothers me about the study, which is the list of factors by which “road conditions” would be blamed for a crash. Among this list are predictable things like “surface defect” (although again, in New York City at least, these act as traffic calming safety devices) and “bad lane marking,” but then one also sees “narrow road” (what does this mean, exactly, that it wasn’t wide enough for two vehicles to pass; or that someone was driving too fast and clipped an edge, or too narrow for a large truck?), “icy conditions” (are we to build radiant heating under all roads?), “insufficient sight” (a number of studies have shown that drivers merely offset increased sight distance with increased speed), and so on — I could go on, and many of these are valid, to a point.

But where does “road factors” causality end? Does every left-turning crash that happens at an intersection without a protected left arrow come down to road factors? Better yet, does every intersection need to be replaced with a flyover, to prevent conflicts? Do we eliminate curves to prevent crashes, or keep roads straight and raise the risk of fatigue? Do we illuminate roads at night, thus putting more hazardous obstacles in the road by day? Once we finish all this work, the construction lobby can then move to every backyard in the U.S., and pave over the swimming pools which represent a very real health risk.

But in the list of road factors in this report, weirdly, one sees “congestion” coded as a road condition. So a perfectly engineered road is safe when empty but then mysteriously becomes dangerous when loads of cars decide to use it? Was this report merely a fig leaf for more building of new roads? Congestion is a social condition — rarely an engineering one — and one that could be easily avoided through any number of means (e.g., taking one out of five drivers out of a rush hour car and putting them in a carpool). In any case, Montana, a state with very little congestion, hosts the most fatalities per mile driven in the country — should we count “lack of congestion” as a causal factor as well?

It all goes back to what I like to call the “four Es” of traffic safety — engineering, education, enforcement, and exposure (on this last point, we largely have the poor economy, and the reduction in driving — and in certain types of driving — to thank for the lowest level of road fatalities since 1961). They are like legs of a table — pull one away, and things start to get wobbly.

This entry was posted on Tuesday, July 7th, 2009 at 3:46 pm and is filed under Risk, Roads, Traffic Culture, Traffic Engineering, Traffic safety, Uncategorized. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.

7 Responses to “A Few Thoughts About ‘On a Crash Course,’ by Miller & Zaloshnja”

  1. Rich in CO Says:

    This is a commisioned, non-peer reviewed “report”, which concluded that only 19.1% of fatal accidents were related to “road conditions” until the researchers decided that in any fatal accident where a bridge, a pole or a tree was involved the bridge, pole or tree had a 100% probability of contirbuting to the fatality. (Letting the fact that many of these things are stuck with non-fatal results slide). Even if we buy their argument that it should be impossible to strike a tree or a pole (those traffic signal poles are going to be a bit difficult to remove – they are often “involved” in fatal crashes when the crash occurs in an intersection) the authors assertion that properly designed bridges would not be involved in crashes is silly on its face.

    It is just as valid to point out that 79.9% of fatal crashes (according to the same report, in table A-11, page 33) occured in the absence of “road-related conditions”. Unless we accept (I don’t) that at least 79.9% of travel occurs on roadways meeting all standards (better than that, not just a forgiving roadside, but one where hitting trees or posts is actaully impossible) then apparently we can conclude that roads built to high standards are over-represented in fatal accidents.

    What’s the old saw about liars and figures?

  2. Christopher Monnier Says:

    Is the problem here lack of safety engineering, or lack of behavioral engineering?

    I think the problem is in the perspective of whom roads should serve. In Sweden (and Europe, in general), it seems that the default road user (in urban areas, at least) is pedestrians, and cars must play second fiddle. In the US, the bias is skewed far in favor of cars over pedestrians. The American mentality is that roads are merely conduits that exist to shuffle vehicles through as quickly as possible. Pedestrians, congestion, traffic calming measures…these are all obstructions to what seems to be a God-given right for cars to get from point A to point B as soon as possible.

  3. Bossi Says:

    As a traffic engineer, I approach my job with the firm belief that people *want* to drive in a safe & legal manner, but road design encourages them to disobey & disregard. That is, road design leads into the human factors. However, unlike the report, I hardly believe that straightening curves & widening clear zones is the answer… that’s the reason we’re in our predicament to begin with. In my opinion, the late Hans Monderman was spot-on with more context-sensitive designs reflecting the class of the roadway.

  4. Dr. Leonard Evans Says:

    Potholes and rumble strips are similar safety devices that work in similar ways. Removing either reduces safety. There are many compelling reasons to keep the roads in good repair – but safety is not one of them. Replacing a rural two-lane road with an Interstate does reduce crash risk — but there are convincing reasons to not do this. The overwhelmingly dominant factor in safety is speed. Factors that increase speeds, like removing congestion or potholes, INCREASE fatality risk. A 1% increase in travel speed leads to a more than 4% increase in fatality risk.
    Leonard Evans, author of “Traffic Safety” []

  5. Lee Says:

    If you reduce the risk in the environment people will just do risk compensation.

    After they spent stimulus money paving over lots of brick roads (which were as perfect as brick can be) here in East Baltimore, we started getting lots of accidents where we never had any traffic accidents whatsovever before… in response to the complaints about speeding cars, they are now installing speed bumps everywhere. So now, instead of complaining about the brick roads, people are complaining about all the speed bumps, and the cars that speed from one speed bump to the next. Plus now they say they have to install more stop signs, lights, etc. However on the streets that are still brick, everything is just fine.

  6. Ted Miller Says:

    Someone just brought this commentary to our attention. We may reply further to some of the many thoughtful points raised here later. But I wanted to put a few thoughts out there quickly. First, a version of this study has been peer-reviewed; it was in press before we released the estimates. Second, the Tri-Level study has long been criticized for over-stressing the human side and underemphasizing vehicles and environment. Nevetheless, if we looked only at crash causation and not at crashes made more severe by road deficiences, we would get roughly 35% as Tri-Level did. What tilts the equation is that 46% of fatal crashes and 40% of serious to critical injury crashes (but only 10% of other crashes) involve harmful contact with a non-breakaway medium or large pole or with a large tree or first harmful contact with a bridge. That was not true in 1990, but cars have become safer, drivers more sober, and the roadside therefore has emerged as a more important factor. The relative risk for those three kinds of crashes being serious-to-fatal is around 4 so we attributed 75% of the serious-to-fatal cost to roadside deficiencies (being careful not to double-count crashes we already had said were related to roadside deficiencies. I was quite surprised at what a large factor non-breakaway poles have become. Those are totally fixable hazards.
    Finally, I wanted to give you my take on our study, which is not that dissimilar to Tom’s. The driving environment is very forgiving. Drivers often make minor errors. They also speed, they get distracted, they drive drowsy, or they take one drink too many. When the roadway is deficient, those errors are more likely to cause a crash and crashes that occur are more likely to result in serious injury or in death. Although behavioral factors are involved in most crashes, avoiding those crashes through driver improvement and enforcement requires reaching millions of individuals and getting them to sustain best safety practices. It is far more practical to make the roadway environment more forgiving and protective. Safer drivers and safer cars remain vitally important, but it also is critical to make the roads, bridges, and shoulders safer.
    Ted Miller

  7. Bike Accidents Says:

    Its a catch 22 kind of situations here … we cannot have better roads overnight and maybe the need is for safer corridors on existing infrastructure, and better implementation of a rigid and strict monitoring system which helps in reducing error prone driving and also acts as a deterrent for those considering to try out living a video game kind of scenario in real life. After all safety is a must and needs improvement globally.

Leave a Reply

Traffic Tom Vanderbilt

How We Drive is the companion blog to Tom Vanderbilt’s New York Times bestselling book, Traffic: Why We Drive the Way We Do (and What It Says About Us), published by Alfred A. Knopf in the U.S. and Canada, Penguin in the U.K, and in languages other than English by a number of other fine publishers worldwide.

Please send tips, news, research papers, links, photos (bad road signs, outrageous bumper stickers, spectacularly awful acts of driving or parking or anything traffic-related), or ideas for my Transport column to me at:

For publicity inquiries, please contact Kate Runde at Vintage:

For editorial inquiries, please contact Zoe Pagnamenta at The Zoe Pagnamenta Agency:

For speaking engagement inquiries, please contact
Kim Thornton at the Random House Speakers Bureau:

Order Traffic from:

Amazon | B&N | Borders
Random House | Powell’s

U.S. Paperback UK Paperback
Traffic UK
Drive-on-the-left types can order the book from

For UK publicity enquiries please contact Rosie Glaisher at Penguin.

Upcoming Talks

April 9, 2008.
California Office of Traffic Safety Summit
San Francisco, CA.

May 19, 2009
University of Minnesota Center for Transportation Studies
Bloomington, MN

June 23, 2009
Driving Assessment 2009
Big Sky, Montana

June 26, 2009
PRI World Congress
Rotterdam, The Netherlands

June 27, 2009
Day of Architecture
Utrecht, The Netherlands

July 13, 2009
Association of Transportation Safety Information Professionals (ATSIP)
Phoenix, AZ.

August 12-14
Texas Department of Transportation “Save a Life Summit”
San Antonio, Texas

September 2, 2009
Governors Highway Safety Association Annual Meeting
Savannah, Georgia

September 11, 2009
Oregon Transportation Summit
Portland, Oregon

October 8
Honda R&D Americas
Raymond, Ohio

October 10-11
INFORMS Roundtable
San Diego, CA

October 21, 2009
California State University-San Bernardino, Leonard Transportation Center
San Bernardino, CA

November 5
Southern New England Planning Association Planning Conference
Uncasville, Connecticut

January 6
Texas Transportation Forum
Austin, TX

January 19
Yale University
(with Donald Shoup; details to come)

Monday, February 22
Yale University School of Architecture
Eero Saarinen Lecture

Friday, March 19
University of Delaware
Delaware Center for Transportation

April 5-7
University of Utah
Salt Lake City
McMurrin Lectureship

April 19
International Bridge, Tunnel and Turnpike Association (Organization Management Workshop)
Austin, Texas

Monday, April 26
Edmonton Traffic Safety Conference
Edmonton, Canada

Monday, June 7
Canadian Association of Road Safety Professionals
Niagara Falls, Ontario

Wednesday, July 6
Fondo de Prevención Vial
Bogotá, Colombia

Tuesday, August 31
Royal Automobile Club
Perth, Australia

Wednesday, September 1
Australasian Road Safety Conference
Canberra, Australia

Wednesday, September 22

Wisconsin Department of Transportation’s
Traffic Incident Management Enhancement Program
Statewide Conference
Wisconsin Dells, WI

Wednesday, October 20
Rutgers University
Center for Advanced Infrastructure and Transportation
Piscataway, NJ

Tuesday, March 8, 2011
Ontario Injury Prevention Resource Centre
Injury Prevention Forum

Monday, May 2
Idaho Public Driver Education Conference
Boise, Idaho

Tuesday, June 2, 2011
California Association of Cities
Costa Mesa, California

Sunday, August 21, 2011
American Association of Motor Vehicle Administrators
Milwaukee, Wisconsin

Wednesday, October 26, 2011
Attitudes: Iniciativa Social de Audi
Madrid, Spain

April 16, 2012
Institute for Sensible Transport Seminar
Gardens Theatre, QUT
Brisbane, Australia

April 17, 2012
Institute for Sensible Transport Seminar
Centennial Plaza, Sydney
Sydney, Australia

April 19, 2012
Institute for Sensible Transport Seminar
Melbourne Town Hall
Melbourne, Australia

January 30, 2013
University of Minnesota City Engineers Association Meeting
Minneapolis, MN

January 31, 2013
Metropolis and Mobile Life
School of Architecture, University of Toronto

February 22, 2013
ISL Engineering
Edmonton, Canada

March 1, 2013
Australian Road Summit
Melbourne, Australia

May 8, 2013
New York State Association of
Transportation Engineers
Rochester, NY

August 18, 2013 “Ingenuity” Conference
San Francisco, CA

September 26, 2013
TransComm 2013
(Meeting of American Association
of State Highway and Transportation
Officials’ Subcommittee on Transportation
Grand Rapids MI



July 2009
« Jun   Aug »

No, you probably won be compensated one million dollars; however, with the right blend of negotiating skills and patience, your efforts will be substantially rewarded!I have seen up to forty thousand dollars added to starting compensation through diligent negotiations. It is a way to significantly raise your standard of living and sense of self, simply by