NHTSA Statistical Database Query - Fatal Crashes - Bike Lane/Paved Shoulder vs. Travel Lane
Neal
Hello All,
I am having a bit of a problem reconciling Vehicular Cycling ‘take the lane’ theory with a NHTSA statistical data base query.
Can you help me out?
http://www-fars.nhtsa.dot.gov/QueryTool/QuerySection/SelectCriteria.aspx
It is easy to query the data base for various statistical breakdowns of fatal motorist/bicycle crashes and submit different variables for Fatal Crash output.
I am probably doing something that causes an error …….. for instance using 2010 as a datum year the data base appears to show the safest place for avoiding a fatal motorist/bicycle crash is to ride is in the bike lane/paved shoulder – almost 5 times as safe for non-intersection crashes.
Showing 55 fatal crashes in the bike lane/paved shoulder as opposed to showing 266 fatal crashes in the travel lane for non-intersection crashes.
For intersection crashes the data shows 2 fatal crashes in bike lane/paved shoulder and 37 fatal crashes in travel lane …. Making it appear that it is over 18 times as safe to be in bike lane/paved shoulder rather than the travel lane in intersection crashes.
Is there an error in my query technique or are the police accident reports that make up the database in error in defining the travel lane or the bicycle lane/paved shoulder?
Even if these statistics are correct I realize that Vehicular Cycling is not all about safety. Getting from point A to B as efficiently as possible must be given some weight in a decision on where and how to cycle. There is a level of acceptable risk for each cyclist as shown on this graphic:
What do you think?
Neal
Eli Damon
they don't take into account how much cycling is done in travel lanes
versus how much is done in bike lanes or paved shoulders. Most roads
don't have bike lanes or paved shoulders, so I would imagine that most
cycling is done in travel lanes. This is the most immediate problem with
the comparison, although there are other problems lurking behind this
one. Eli
Trevor Bourget
For intersection crashes the data shows 2 fatal crashes in bike lane/paved shoulder and 37 fatal crashes in travel lane …. Making it appear that it is over 18 times as safe to be in bike lane/paved shoulder rather than the travel lane in intersection crashes.
John Forester
Second, while there is information on fatal collisions, there is no information on exposure, how much cycling is being done in each situation, so that there cannot be any calculation of rate, that is, of the risk for any one cyclist. That is, the information fails to provide any measure of the relative risk of the various cycling situations.
Third, you assume that turning and crossing movements occur only in intersections. That is not correct, either. The NHTSA statistics list place as either intersection or non-intersection, which may be fine for their purposes but it is inaccurate for purposes of cyclist safety.
Fourth, you quote the political guess on which Portland's bicycle transportation program is based. Your reason is that this shows: "There is a level of acceptable risk for each cyclist as shown on this graphic". Neal, you are assuming that the political propaganda guess published by Portland's bicycle advocates demonstrates something valid about the general public, and that this general public has reasonably accurate estimates of cycling risk. That guess accurate? I think Portland has a new bridge to sell you. That the general American public has any valid knowledge about the risks of cycling? The evidence for the forty years that that has been studied, since the start of scientific study of bicycle transportation, demonstrates the utter falsity of that claim.
Neal, it has become obvious to many of us that your arguments are driven by ideology rather than concern for the facts and reason. Pursuing the wrong path once, and being corrected, is entirely acceptable. But continuing with the same erroneous arguments and working up other erroneous arguments to support the first error is a sign of ideology rather than scientific process.
On 5/12/2012 7:19 PM, Neal wrote:
Hello All,
�
I am having a bit of a problem reconciling Vehicular Cycling �take the lane� theory with a NHTSA statistical data base query.
�
Can you help me out?
It is easy to query the data base for various statistical breakdowns of fatal motorist/bicycle crashes and submit different variables for Fatal Crash output.
�
I am probably doing something that causes an error ��.. for instance using 2010 as a datum year the data base appears to show the safest place for avoiding a fatal motorist/bicycle crash is to ride is in the bike lane/paved shoulder � almost 5 times as safe for non-intersection crashes.
�
Showing 55 fatal crashes in the bike lane/paved shoulder as opposed to showing 266 fatal crashes in the travel lane for non-intersection crashes.
�
For intersection crashes the data shows 2 fatal crashes in bike lane/paved shoulder and 37 fatal crashes in travel lane �. Making it appear that it is over 18 times as safe to be in bike lane/paved shoulder rather than the travel lane in intersection crashes.
�
Is there an error in my query technique or are the police accident reports that make up the database in error in defining the travel lane or the bicycle lane/paved shoulder?
�
Even if these statistics are correct I realize that Vehicular Cycling is not all about safety.� Getting from point A to B as efficiently as possible must be given some weight in a decision on where and how to cycle.� There is a level of acceptable risk for each cyclist as shown on this graphic:
�
�
What do you think?
�
�
Neal
+1 mph Faster�
--
To post: bicycle...@googlegroups.com
Only rule: no personal commentary (please comment about content, not people)
�
To unsubscribe: bicycledrivin...@googlegroups.com
�
Group website: http://groups.google.com/group/bicycledriving
Discussion archives: http://groups.google.com/group/bicycledriving/topics?hl=en
Glossary: http://groups.google.com/group/bicycledriving/web/glossary
Links: http://groups.google.com/group/bicycledriving/web/links
-- John Forester, MS, PE Bicycle Transportation Engineer 7585 Church St. Lemon Grove CA 91945-2306 619-644-5481 fore...@johnforester.com www.johnforester.com
Trevor Bourget
Happy Mother's Day, everyone.
-- trevor
Mark Ortiz
As Eli also points out, these statistics are not per bicycle mile based. In fact, as the recent NC State study on NC statistics (http://www.pedbikeinfo.org/pbcat/pdf/summary_bike_types05-09.pdf) points out, responsible researchers admit there is no reliable data on bicycle vehicle mileage in the state. Presumably this is true for most of the rest of the nation as well.
But just from the personal experience of those here, would it even be possible, where you ride, to do as much as 55/(55+266) = 17% of your travel on a paved shoulder or bike lane? Do anything like 17% of the roads you ride on have those features? The answer could be yes in some locations, but not where I live, and not nationally.
As for very few intersection crashes involving bicyclists on the shoulder or in a bike lane, the reason there should be readily apparent: generally, it’s physically impossible to be within an intersection and at the same time be on a shoulder or in a bike lane.
I guess this would vary some depending on how much of the approach to the intersection the researchers chose to include with the intersection, or when determining bicyclist position prior to collision. Probably the only way to avoid ambiguity here is to simply use the location per the accident report, and generally this would use the legal definition of the boundaries of the intersection. The NC study introduces a third location category, “intersection related”, in an attempt to address this. The breakdown they get then is:
· Intersection 43.5%
· Intersection related 2.5%
· Non-intersection 49.1%
· Non-roadway 4.6% (mainly parking lot collisions here)
· Unknown location 0.3%
There are all sorts of other breakdowns in the report. Overtaking collisions are the largest single category, but by no means a majority. They are 17% of crashes, and 39% of fatalities. Head-ons are less common but more deadly: 3% of collisions; 11% of fatalities.
It’s a bit hard to sort through the different breakdowns. For example, we see that “motorist overtaking” is 17% of crashes, but we don’t know for sure what percentage of the non-intersection crashes those are, since this type of crash can sometimes occur within an intersection, or so one might suppose. But if we assume that all motorist overtaking crashes happened between intersections, they are 16.8/49.1 = 34.2% of non-intersection crashes. Motorist driveouts and bicyclist rideouts between intersections account for about another third, with other categories making up the balance. Within the last third, “bicyclist left turn/merge” is fairly heavily represented, and this includes bicyclist rideouts where the bicyclist intends to go the same direction as other traffic. In other words, at least a plurality of the non-intersection crashes are in fact some form of cross-traffic collision.
They have a category for doorings. These are very uncommon: only 9 reported statewide over the five years surveyed.
Mark Ortiz
--
To post: bicycle...@googlegroups.com
Only rule: no personal commentary (please comment about content, not people)
To unsubscribe: bicycledrivin...@googlegroups.com
Neal
“Neal, you have been told, repeatedly, that the patterns for fatal car-bike collisions and non-fatal car-bike collisions are drastically different. You cannot properly use fatal data to extrapolate to safety. That's your first major error, and you have been warned before this.”
Hello John and All,
Thanks for your prompt and kind reply to my thread: NHTSA Statistical Database Query - Fatal Crashes - Bike Lane/Paved Shoulder vs. Travel Lane.
I find it difficult to agree with you on your point stated above since I think of avoiding Fatal Crashes as an important component of cycling safety. Perhaps you do not.
I agree that severe injuries and road rash are a consideration also but be advised that was not a part of this database query. Please refer to the title of the thread - '……… Fatal Crashes …….'
“Second, while there is information on fatal collisions, there is no information on exposure, how much cycling is being done in each situation, so that there cannot be any calculation of rate, that is, of the risk for any one cyclist. That is, the information fails to provide any measure of the relative risk of the various cycling situations.”
You are correct on this point. I did not add or subtract anything from the database query. It is what it is. What it shows is aggregate risk.
………..for instance using 2010 as a datum year the data base appears to show the safest place for avoiding a fatal motorist/bicycle crash is to ride is in the bike lane/paved shoulder – almost 5 times as safe for non-intersection crashes.
Showing 55 fatal crashes in the bike lane/paved shoulder as opposed to showing 266 fatal crashes in the travel lane for non-intersection crashes.
For intersection crashes the data shows 2 fatal crashes in bike lane/paved shoulder and 37 fatal crashes in travel lane …. Making it appear that it is over 18 times as safe to be in bike lane/paved shoulder rather than the travel lane in intersection crashes.
“Third, you assume that turning and crossing movements occur only in intersections. That is not correct, either. The NHTSA statistics list place as either intersection or non-intersection, which may be fine for their purposes but it is inaccurate for purposes of cyclist safety.”
I believe the NHTSA safety mission is quite clear and their statistics are relevant:
[As NHTSA proclaims in its mission statement, the agency's main focus is "to save lives, prevent injuries and reduce traffic-related health care and other economic costs." As such, NHTSA functions as both an information source and an investigatory body. Its responsibilities fall into three main areas.....]
I use the statistics as prepared by the NHTSA. If you have a better method for categorizing statistics (perhaps similar to North Carolina as Mark posted) you should notify the NHTSA and ask them to incorporate your improvements in their database reporting.
Neal, it has become obvious to many of us that your arguments are driven by ideology rather than concern for the facts and reason. Pursuing the wrong path once, and being corrected, is entirely acceptable. But continuing with the same erroneous arguments and working up other erroneous arguments to support the first error is a sign of ideology rather than scientific process.
Nope – I don’t think so. I believe the NHTSA data is reasonably accurate and while I value your thoughtful opinion, in this case I prefer the NHTSA database statistics.
John, do you have data that would invalidate the NHTSA database statistics?
I submit that the NHTSA database is a useful scientific tool although the data may cause conflict with closely held beliefs that are only supported by opinion.
There is an old saw in the law: If the facts are not on your side … argue the law …… and if the law is not on your side – discredit the witness – or as the Brits would say ‘Pound the Table’.
Neal
Neal
Neal
Trevor Bourget
"to save lives, prevent injuries and reduce traffic-related health care and other economic costs."
Sara R
630
51,000 cyclists injured in wrecks with motor vehicles
of the 630 fatalities 574 were not wearing helmets
only 53 of the fatalities were wearing helmets
does this mean wearing a helmet saves lives? maybe, maybe not..it
could simply be that the kind of people who wear helmets behave
differently than those who don't.
If you truly fear being killed by being overtaken by a motor vehicle
on a rural road, it would make sense for you to ride against traffic,
but no one recommends that. I wonder why??
Neal
I received the response belowfrom NHTSA.
There may be other methods to resolve the location of the crashes and determine the type of lane use/availability such as using Google Earth and GPS coordinates. I will have to check on this. There may be coordinates associated with each crash in the database.
I might be cumbersome to hand evaluate each crash but perhaps some statistical sampling might give some direction to an answer to the location of the crashes. That will not be an answer to your question as of the number of cyclists in each mode of course - travel lane vs bike lane/shoulder...... but such an evaluation might provide some unexpected insights.
Until we have bicycle traffic counters as in Europe can you think of currently available data that would provide an answer as to where most cyclists travel? A 'depiction' map .... http://www.depiction.com/ or BikeScore presentation of cyclist density would be useful .... http://www.walkscore.com/bike
The answer may not be obvious since most cyclist fatalities are urban and most bike lanes/paved shoulder markings are in urban areas.
The data would be useful for rational planning of improved facilities for cyclists.
==========================================
[From NHTSA]
The 2010 Bicyclists & Other Cyclists Traffic Safety fact sheet is not currently published. We are working on this publication at this time. The data contained in these fact sheets is the only data we can provide.
Lyn Cianflocco
National Highway Traffic Safety Administration
National Center for Statistics & Analysis
Data Reporting & Information Division
NVS-424 W53-127
Cheers,
Neal
Neal
Hello Sara and All,
Thanks for the interesting questions …..
Noted in Cross/Fisher NTHSA Study of 1977: …………. Another major conclusion of this study is that the causes of the vast majority of bicycle/motor-vehicle accidents are behavioral. In well over 60% of the cases, the bicyclist's pre-crash course was suboptimal, indicating that a predisposing or precipitating error was made before the other vehicle could have been observed.
Your supposition that helmet use/nonuse may be an indicator of fatalistic behavior may have some merit although I think the authors of the study were using ‘behavioral’ in a different context …. More about search/reaction/sub optimal pre-crash course ….. rather than psychological risk assessment. If I recall correctly it is mentioned and noted that cyclists in crashes did not think they were taking undue risks – but rather that they had made a faulty risk assessment.
“If you truly fear being killed by being overtaken by a motor vehicle
on a rural road, it would make sense for you to ride against traffic,
but no one recommends that. I wonder why??”
Wiki:
Excerpts:
Fear is a distressing negative sensation induced by a perceived threat. It is a basic survival mechanism occurring in response to a specific stimulus, such as pain or the threat of danger.
Additionally, fear is frequently related to the specific behaviors of escape and avoidance, whereas anxiety is the result of threats which are perceived to be uncontrollable or unavoidable. It is worth noting that fear almost always relates to future events, such as worsening of a situation, or continuation of a situation that is unacceptable.
I suspect it is important to ride with the normal traffic direction since that is the law and what is anticipated by motorists. Anticipation, expectation, and assumptions are used by motorists and cyclists to determine their course and prevent collisions.
There may be other reasons …. What do you think is the reason?
I do not think some cyclists ‘fear’ (in the sense that they have a physiological fear response) being killed by being overtaken by a motor vehicle. I think it is more that they are aware of the danger as posed by various reports and have a heightened awareness to the danger as they make a risk assessment.
Many cyclists take countermeasures - As Serge writes and as recommended by Cross/Fisher 1977 study …. Use a rear view mirror ……
Excerpt:
“Sounds reasonable, if you don't use a mirror. If you do use a mirror, all of that sound reasoning becomes irrelevant, because whatever the odds to-be-hit are of a mirror-less lane controlling cyclist, they are surely much lower, practically negligible, for the mirrored lane controller.
As anyone who regularly controls lanes with a mirror can attest, the mirror allows you to control lanes much more often than you would without a mirror, and allows you to learn to "read" traffic so well that you can easily identify those rare drivers who are not slowing or changing lanes soon enough to indicate that you've been noticed, giving you plenty of time and space to do something to grab their attention, and, if that doesn't work, to ditch.
Serge”
Some cyclists – like bike messengers get a high from drafting and hooking a tow from trucks, cutting in front of taxis, and generally riding a thin line.
PubMed:
RESULTS:
Most working couriers have suffered at least one injury resulting either in days lost from work (70%) and in visits to a health-care professional or hospital (55%). The annual incidence rate for injuries resulting in days away from work was 47/100-bike couriers. Bone fractures accounted for the most days lost from work, followed by dislocations, sprains, and strains. Collisions and avoiding collisions with motor vehicles, including being "doored," and collisions with pedestrians accounted for the majority (66%) of events leading to injury. Twenty-four percent of messengers reported wearing a helmet on a regular basis, and 32% have health insurance.
CONCLUSIONS:
Urban bicycle messengers are a poorly documented, largely unstudied workforce who suffer a very high rate of occupational injury.
What is great sport to one individual is a dangerous activity to another.
Cross 1977 page 20
Contrary to popular beliefs, bicycle/motor-vehicle accidents are seldom the direct or indirect result of roadway-surface defects, debris on the roadway surface, sewer grates, bicycle defects or failures, motor-vehicle defects or failures, riding double, bicycle too large or too small for the operator, bicycle-handling skill deficiencies, hostile acts by motorists, high risk acceptance by bicyclists, or the bicyclist's deficient knowledge of traffic laws and ordinances.
The authors pointed out the difference between faulty risk assessment and risk acceptance and importance of modifying bicyclists risk assessments and misjudgment of the risk associated with the critical action.
The non-behavioral factors that are the most important contributors to bicycle/motor-vehicle accidents include: visual obstructions, narrow roadways (selected locations), darkness, daytime and nighttime conspicuity of bicycles, and the vertical dimension of the bicycle/bicyclist unit.
Rather than assessing the emotion of fear the NHTSA studies provide information to assess relative danger levels which are based on crashes.
As the Cross/Fisher Study 1977 states:
There is virtually no doubt that off-street bicycle lanes would reduce the incidence of overtaking accidents, if such facilities were available and used by bicyclists who otherwise would be riding on roadways.
It is quite interesting to note that while some of the Cross/Fisher 1977 NHTSA paper findings may be dated – these have not changed:
Characteristics of the Accident Location
Cross/Fisher Study for NHTSA 1977
Urban vs. rural accidents. A proportionate sample of urban and rural accidents was not drawn for this study. However, based upon the findings of this study and data reported elsewhere, it is estimated that about 32% of all fatal accidents and 11% of all non-fatal accidents occur in a rural area.
Applying the estimates of the proportions of cases correctly classified to the National Safety Council's estimates of the distribution of incorporated and unincorporated accidents yields the following estimates: FATAL NON-FATAL URBAN 68% 89% RURAL 32% 11% It is believed that the above estimates are the best available. However, additional research should be conducted to verify these estimates.
NHTSA 2009
The majority of pedalcyclist fatalities in 2009 occurred in urban areas (70%). In respect to vehicle crash location in relation to an intersection, most pedalcyclist fatalities in 2009 occurred at non-intersections. Compared to 2008 these numbers increased by 5 percent.
NHTSA 2008
Pedalcyclist fatalities occurred more frequently in urban areas (69%), at non-intersection locations (64%), between the hours of 5 p.m. and 9 p.m. (28%), and during the months of June (9%) and September (12%).
It is worth noting as Jeff Gross pointed out in a previous post that rear-end collisions between motorists are the largest type of motorist crash. If motorists cannot stop before hitting another large automobile or truck do not be surprised if they do not see you in time to stop.
In a US study half of these drivers did not brake before the impact. This is found in statistical crash data collections as well. Analyzing UK National accident database (STATS19) from 2005, Grover et al. (2007) found that the drivers in 44% of the vehicles in the sample took no avoiding action prior to the impact.
Ride safe …… J