Home

More Bicycle Safety Studies Needed

Yesterday, Judd and I posted suggestions for several ways to improve bicycle safety, which met with responses ranging from seething hatred to what I’ll term “belligerent support”. People seem to have very visceral reactions to this issue, and in an attempt to explain our positions, Judd and I have decided to elaborate our positions in a few posts.

As gas prices rise, more and more people are converting to bicycles as a means of transportation. This trend will probably continue into the foreseeable future, and roads designed for cars may find themselves choked with bikes – a potentially deadly situation.

We’ve all heard the statistics about how many bicyclists meet their grisly demises at the hands of automobiles. We’ve all become so familiar with the narrative about carnage at the hands of reckless drivers that it has almost become reflexive to demand bicyclist rights and improved facilities to lessen the loss of life and limb. But in our rush to do something about this problem, we haven’t done enough research into the root causes of these accidents.

A reader named Rob, who disagreed with our suggestions to improve bicycle safety, highlighted the below three statistics to me yesterday in a comment.

1. More than 5,000 pedestrians and cyclists are killed each year by motor vehicles, and more than 100,000 are seriously injured.

2. Drivers are strictly responsible for 80% of bicyclist deaths.

3. The National Highway Traffic Study Administration (NHTSA) studied 8,000 pedestrian and bicyclist crashes and found motorists responsible for 54% of crashes with a known cause.

I apply skepticism to most things that people tell me, and this was no exception, particularly since I’ve been researching this area a lot recently. I have taken away much different conclusions from my research than Rob. (For the record, I omitted several other points he made that I also disagree with, which Judd and I will address in later posts).

When I first began examining the issue of safety, I was virtually shocked at the way most research is packaged and conducted. Nearly all of it has a pro-bicycle slant, ranging from mildly benign to cars but completely fawning over the “advantages and benefits” of bicycles, to openly hostile to cars from the get-go.

Before I begin to explain why this approach is utterly wrong, let me make one thing clear – this post is not an attempt to make light of the death of any bicyclist. A life lost is tragic. But if we squander an opportunity to truly understand the causes and prevent further loss of life, then the dying, and studying, is in vain.

Safety is a complicated issue with very broad boundaries. It is hard to tell where you should start any analysis. But, let me start with an explanation why I think Rob’s concise bullets are misinformed.

I do not dispute Rob’s first point that “more than 5,000 pedestrians and cyclists are killed each year by motor vehicles, and more than 100,000 are seriously injured.” You can check the facts for yourself on the NHTSA website.

Rob’s second point that “drivers are strictly responsible for 80% of bicyclist deaths” is derived from two different “studies” of bicycle fatalities in New York City in the mid to late 1990’s: Charles Komanoff’s “Killed By Automobile Death in the Streets in New York City 1994-1997″ (March 1999) and Charles Komanoff and Michael J. Smith’s “The Only Good Cyclist: NYC Bicycle Fatalities - Who’s Responsible?” (May 2000).

LACK OF OBJECTIVITY

My problem with using these studies as “authoritative analysis” begins with bias.

My spider sense started pinging off the chart when I noticed the link on the first study - “Cars-Suck.org”. Seems like they have an agenda. Google this website, and you’ll see the following tag line appear: “Cars Suck! - Dedicated to the overthrow of car tyranny.”

Komanoff’s study begins with the following sentence: “Right of Way is a grassroots organization in New York City asserting the right of pedestrians and cyclists to travel without endangerment by motor vehicles.” Unfortunately for Right of Way, the U.S. Constitution guarantees you no such right. You have a right to use the roads, but you assume the risk in doing so, just like everybody else.

Moreover, Komanoff’s study suffers from methodological problems – specifically, problems with causation. First, it automatically assumes the blame rests with a car if a pedestrian gets hit, regardless of fault. Second, it defies accepted categorization conventions for causal factoring in crashes. Causal factors are things that are assigned to help determine what caused an accident.

In the 1970s, the NHTSA developed conventions for typing pedestrian and bicycle crashes with motor vehicles. These factors have become widely accepted. Komanoff and Right of Way, however, completely disregard these commonly accepted factors, and assign new ones that they made up which are ambiguously defined and blatantly hostile to cars.

I acknowledge that it is nearly impossible to eliminate all bias from a study. The analysis will always be tempered by the researchers’ personal opinions. But such blatant, open bias immediately calls into question the objectivity of any analysis in these studies, which is my next problem.

BIZARRE ANALYSIS

In an analysis better described as bizarre than academic, Right of Way includes a section titled “Ten Key Findings from Death by Automobile” (Page 12) in which we learn, “Automobiles were equal-opportunity threats, killing New Yorkers of every income level and ethnic group roughly in proportion to the group’s share of population.” So what about this is informative in any way about what causes pedestrian deaths? Bicyclists kill people too (more on that later in this post), what difference does it make if they are “equal opportunity” about it.

More from Ten Key Findings - “New Yorkers age 65 and older were more than twice as likely to be killed by an automobile as to be murdered during 1994 – 1997.” Again, what does this have to do with the causation of pedestrian fatalities or how to prevent them? Statements like this exist for one reason – fear mongering. It makes for a scary sound bite, but means little. During the same period as this study, approximately 465 people were killed in fires in New York City. That’s two and a half times more people killed in fires than all pedestrians and bicyclists killed in NYC by cars. Are we going to launch an initiative to ban ovens in kitchens to combat this menace next?

Remember, those last two paragraphs represent two of the ten key findings.

But by far the most bizarre of the analysis comes when the study admits to pedestrians and bicyclists breaking laws (which in many cases leads to their death), but rationalizes that motor vehicles are more dangerous so pedestrians and bicyclists should be allowed to violate the laws. Wrong. Suppose that you could replace all automobiles with horse drawn carriages – would this logic result in a decrease in pedestrian fatalities?

History instructs otherwise. Mayor Bloomberg’s office issued a press release this year saying that pedestrian fatalities were at the lowest they have ever been since record keeping began. In fact, the pedestrian death rate was 41% higher in 1910 – before cars were kings of the road.

But the grand finale comes when this study admits it basically makes up its own data. The main source of information for this “study” was police reports. But on page 40 of the report, we learn, “more surprising and disturbing is how frequently the second-order narrator, the police officer, is an accomplice to these fabrications. A pedestrian is flung 60 feet after impact, but there is no reason to suspect excessive speed. A driver is making a left turn when a pedestrian walks into her vehicle. A cyclist runs a red light and then he strikes a car (man bites dog?), killing himself.”

We’re supposed to get the idea that these are preposterous lies, and this justifies a new coding system that re-writes the accidents as we would like them to have happened so we can prove a point. But what is impossible about any of these scenarios? All of these are very plausible without much imagination, and possible according to the laws of physics. This study invented a new coding methodology to make the data show what they want, not what the data really shows. We should ask ourselves, if the police officer on the scene is not to be believed, why should someone using his report, years after the fact, be believed either? Answer – they shouldn’t.

MISINTERPRETATION OF FACTS

This kind of “analysis” leads to misinterpretation of facts, which is my final problem with Rob’s comment and the subject of his third point, that “the NHTSA studied 8,000 pedestrian and bicyclist crashes and found motorist responsible for motorists responsible for 54% of crashes with a known cause.”

The website Rob links to, Bicyclinginfo.org, has a headline that lists “a motorist failing to yield” as the predominant cause of crashes (21.7%). But if you look at the definitions of the crash types from that study, you’ll see these designations aren’t meant to assign blame. You will see are 14 different crash types, all of which have factors that can be assigned to bicyclists to establish causation. What’s more – two types of crashes don’t assign any causal factors to motorists. The most common factors for bicyclists include, disregarding traffic signals, riding the wrong way and riding on the sidewalk. Another interesting aspect of this study is that the main recommendations focus on controlling bicyclists.

Now, causation is not the same as fault, both a driver and a bicyclist can cause an accident, but one can cause it more and be at fault. But we don’t assign fault solely because one person is in a car either, and we certainly shouldn’t if we’re really interested in safety. This is a safety study, and it does not assign fault, but the headlines people pull from it, including Rob, Bicyclinginfo.org, Right of Way, and Cars Suck make it look like this study does assign fault.

Similar problems plague all of the available research in this field – it’s incomplete and insufficient for the purposes that people attempt to use it. Approaching a study with the notion that the automobile is automatically the bad guy because a bicyclist is hurt or injured won’t help us to arrive at the best possible solution to stop the deaths. Neither will approaching a study with the intent to cheerlead for the benefits of bicycles.

I have seen various studies on bicycles ranging from this one from the University of Texas to this one from a professor at the University of Washington. By far the best I have seen is this one from New York City - a ten year comprehensive study of all bicyclist fatalities in the city.

Even the best of these studies makes elementary mistakes for the purposes of trying to arrive at conclusions about safety recommendations. They all assume bicyclists are the victims. It’s true that bicyclists are almost always the one killed, but that doesn’t mean they cannot be at fault for their own death. None of the studies seriously examines this issue, they rely on the coding from the NHTSA. However, as we saw with the definitions above, these are just categories of crashes under which either drivers or bikes could be at fault.

There is no data on vehicle drivers hurt as a result of bicyclists. For instance, the University of Texas study suggests that drivers actually receive a safety benefit from bike lanes because they are less likely to swerve into oncoming traffic on roads that have bike lanes. This implies that there is potential for this swerving to cause accidents, and it is not hard to imagine other scenarios – slamming on breaks to avoid hitting cyclist and being rear-ended, slamming on breaks to avoid cyclist who runs light and getting t-boned in the intersection. No one bothers to look to see if this data exists, but I would be willing to bet there are crashes like this that bicyclists cause.

These studies often don’t even collect data about bicycle accidents when only other bicyclists or pedestrians are involved – again revealing the anti-car slant. No one looks at these issues to determine who is really to blame, and thus who we should focus on to decrease the accidents.

Let’s look at the best of the bunch - “Bicyclist Fatalities and Serious Injuries in New York City 1996-2005,” a joint report by New York City Departments of Health and Mental Hygiene, Parks and Recreation, Transportation, and the New York City Police Department - and see how bias creeps in to cause misinterpretation of findings.

Once again, we start with a biased statement, “Bicycling is good for New York City!” No discussion of the fact that this hypothesis may prove untrue when we look at the data.

Next we see methodological problems. The study combines walking and biking numbers together – in fact walkers outnumber bikers 16.5 to 1. This seems trivial, but it makes it seem that there are many more bicyclists on the roads than there are. This distorts the findings and makes bicyclists’ actions look safer in proportion to their numbers than they are. According to census data, on a given day there are 21,585 bicyclists in New York City commuting.

Despite being a bike friendly city – NY actually has a bicyclist death rate that is higher than the national average – 2.8 compared to 2.7 per 1,000,000. New York added 213 miles to basically double its bike lanes during this period (1996-2005) – we hear that this makes bicyclists safer, but we don’t see a corresponding drop in the number of fatalities

According to the study “changes did not suggest an upward or downward trend,” and in fact, more cyclists were killed in 2005 than in 1996 – 26% more. An average of 23 cyclists per year were killed according to the study, and after the improvements to bicycle infrastructure began, there was a spike to 40 fatalities in 1999.

There is no historical data to compare this to, but we should still expect to see a downward trend. Since we don’t, there is no way to tell if this increased bike infrastructure represents progress.

Injuries did go down by roughly one-third, but this is not directly tied to bike lane usage in the study – so again, we have no way to know if it helps to build more bike lanes. The study references the fact that only one fatality occurred inside the bike lanes, and all fatalities on bike lane roads occurred outside the lanes, but this doesn’t tell me that the lanes make bicyclists safe – it tells me that people still die on these roads, and the bicyclists aren’t using the lanes that are supposed to be helping them. Now it may be that in some of these cases a bicyclist swerves out of the lane to avoid being doored and gets hit by a car, but without accurate data on each accident, there is no way to know and we shouldn’t be assuming.

What we do actually know about these crashes is this: helmet use would save most of these lives.

“Of bicyclists with documented injury types, nearly three quarters involved a head injury. Also - Among the fatalities with documented helmet use, 97% of the bicyclists were not wearing a helmet at the time of the crash. Only 4 bicyclists who died (3%) were wearing a helmet.”

This also holds true in serious injury cases – where 87% of crash victims were not wearing a helmet at the time of crash.

Even more interesting than this statistic is the scant data on non-traffic related fatalities – basically instances where a bicyclist is killed because he hits a stationary object or another bicyclist. There were an average of 1.8 non-traffic related fatalities per year during the New York study. This represents a rate of .83 cyclist deaths per 10,000 bicycles on road. Compare that to cars, who kill 23 bicyclists on average per year – there are roughly 2,868,807 vehicles in city (1,820,000 vehicles crossed bridges + 82, 745 driving commuters in Manhattan who are already there). That equals .08 cyclist deaths per 10,000 cars on road. That means that bicyclists are ten times more dangerous to themselves as cars are when we consider how many of them are on the road. Deaths per mile won’t improve this rate because cars travel many more miles than bicycles, so this will only make the car’s relative danger to bicyclists seem smaller the more miles they drive.

Pedestrian fatalities have odd results too. In 2005 there were 156 pedestrian deaths – 155 of which belong to cars. There were an average of 1.1 pedestrians hit by bicycles per year during the study. That results in a rate of .51 pedestrian deaths per 10,000 bicycles on the road. By contrast, cars have a death rate of .54 pedestrian deaths per 10,000 cars on the road. So bicycles are virtually as dangerous to pedestrians as cars are, proportionate to their numbers. This is before we even consider that there is no data available for serious injuries to pedestrians hit by bicycles. My guess is that these numbers are significantly higher.

For fatalities, bicyclists have causal factors assigned fully or partially in 78% of all crashes (42% just bicyclist – largest category). Again, these factors do not assign blame, but this contradicts the studies suggesting that autos cause three quarters of bike accidents. It also confirms police assertions that bicyclists cause 75% of all crashes (which led to the Right of Way studies that began this post). It is fair to consider that the bicyclist isn’t present to give his side of the story in these incidents, but there are the reports of witnesses and insurance examiners who examine the crash data to see what happens. For serious injuries, bicycles still represent 19% of causal factors to vehicles’ 25%, but the largest portion of factors, 45%, is undocumented. Both people are present in these situations to tell their stories, but the study admits that the missing data makes it “difficult to confirm” if this is the correct trend. Regardless, it is clear either way that bicycle licensing and training could eliminate a large portion of serious injuries and deaths.

As gas prices increase, and we aggressively move to encourage more bikes to get on the streets, it would be wise for us to consider that bicyclists statistically pose more danger to themselves and pedestrians than cars pose to either group. This makes a convincing case for the adoption of mandatory licensing, registration, and bicycle education for bicyclists – if only to protect them.

My findings also make a convincing case that this is a woefully understudied area of the bicycle movement, and transportation policy in general. More thorough academic study is needed to determine who actually is at fault in these accidents, and what dangers bicyclists pose to themselves and to others. Only by fully understanding the whole picture can we arrive at the correct solution that will save the most lives.

The New York City study concludes with four recommendations for autos, and six for bicycles – wear helmets, follow traffic rules, be alert at intersections, stay in designated bike lanes, be aware at all times, and stay off sidewalks. While some actions are voluntary for both groups, only autos have mandatory education – on top of mandatory licensing and registration, and ticketing to enforce compliance. Bicycles should have the same rules and regulations. If we’re serious about safety, it’s time to put our money where our mouths are.

Related Posts

• Toward A New Bicycle Safety in America