Bikeway Study Part Two: Your Speed, Your Choice
Safer Motorists or Safer Bicyclists?
Last week, in the first of a three-part series describing my research on bikeway crash risk, I gave relative estimated risks for three types of motorist-caused crashes when bicyclists rode with the flow…
Travel Lane Edge – 61,000 Miles Between Crashes (Highest Risk)
Bike Lane – 75,000 Miles
Sidewalk – 122,000 Miles (Lowest Risk)
… and asked:
Why would the risk be lowest for bicyclists riding on sidewalks?
All crashes occurred at intersections and driveways with no sort of “protection” for the bicyclist. Don’t experienced bicyclists avoid using sidewalks — and sometimes even bike lanes — because riding in the street is supposed to help them avoid such conflicts?
Stay with me while I show you where the data led me as I pondered this question.
There are two key possibilities why these risks are lower: (1) Motorists might be more likely to yield to bike lane and sidewalk cyclists for some reason, and (2) Bike lane and sidewalk cyclists might be better able to avoid a motorist-caused crash.
It’s often argued and assumed that striping a bicycle lane leads motorists to look for and yield to cyclists. But the sidewalk cyclists had lower risk for these motorist-caused crashes than even the bike lane users, though sidewalks are neither designated nor designed for use by cyclists.
When we split up the crashes further, we see that only the risk of drive-outs is higher for bike-lane and sidewalk cyclists. Right-hook and left-cross crashes are less likely:
Right Hook & Left Cross
Travel Lane Edge – 73,000 Miles (Highest Risk)
Bike Lane – 100,000 Miles
Sidewalk – 300,000 Miles (Lowest Risk)
Sidewalk cyclists have the highest risk of drive-out crashes:
Drive-Out
Travel Lane Edge – 367,000 Miles (Lowest Risk)
Bike Lane – 292,000 Miles
Sidewalk – 245,000 Miles (Highest Risk)
Would motorists be more likely to yield to bike lane and sidewalk cyclists during overtaking right turns and opposing left turns, but not during drive-outs? I think that this is not more likely.
People Make Mistakes
News Flash: Humans make mistakes, whether they’re walking, bicycling, or driving motor vehicles.
We expect slower automobile and motorcycle drivers to be better able to react more quickly than faster ones to avoid conflicts caused by other motorists.
Why would we assume differently for bicyclists? [1]
Recall the average speeds we found for cyclists using the three positions. A slower cyclist needs less perception and reaction time and less braking distance:
Bicyclist Position | Bicyclist Average (Mean) Speed | Bicyclist 85th Percentile Speed | Stopping Distance at 85th Percentile Speed |
---|---|---|---|
Travel Lane | 14.5 MPH | 18.4 MPH | 104 Feet |
Bike Lane | 11.8 MPH | 15.7 MPH | 83 Feet |
Sidewalk | 9.3 MPH | 12.4 MPH | 60 Feet |
Facing an impending motorist-caused crash, the bike-lane user riding at the 85th-percentile speed would need an additional 23 feet of stopping distance compared to the sidewalk rider — about the width of the typical two-lane street. The difference is about the same between the travel lane and the bike lane. The total difference between travel lane and sidewalk is about the width of four traffic lanes, or 44 feet.
85th Percentile Bicyclist Speed and Stopping Distance
With right-hook and left-cross crashes, the motorist is coming from the bicyclist’s left, so the farther right the bicyclist is, the longer it takes for the motorist to reach the bicyclist’s path. This means that the bike lane or sidewalk bicyclist gets more reaction time. With drive-out crashes, the travel lane cyclist gets the most reaction time.
Rather than bikes lanes or sidewalks improving the safety of bicyclists, bicyclists are improving the safety of bike lanes or sidewalks by riding slower.
The Takeaway
If you’re going to ride on the sidewalk, bike lane or edge of the travel lane, you must ride slower.
I can’t tell you how slow is slow enough. But if you’re having more close calls with turning and crossing vehicles than you’d like, you need to either slow down, or use lane control. [2]
Based on the data, each additional mile per hour of bicyclist’s speed increases the risk of a motorist-caused crash about 9 percent.
With lane control, you give yourself much more reaction time for drive-out crashes, you eliminate the right hook crash, and you get more options for avoiding the left cross.
To put it in the simplest of terms: The faster you go, the more important it is for you to control your travel lane.
Sidepaths and Motorist-Caused Crash Risks
We performed the same type of analysis for five sidepaths which have been in place for over ten years. I was curious as to whether they would have better safety performance than a regular sidewalk. They did — about 68 percent better, on average — but it varied widely. Looking closer, I found that three of the paths had few intersections and commercial driveways — 4.6 per mile — while the other two had 11.6 per mile. Low-conflict paths had 64 percent lower motorist-caused crash rates compared to the other two paths; 51 percent lower than for ordinary sidewalks.
Low-Conflict Sidepaths | High-Conflict Sidepaths | Regular Sidewalks | |
---|---|---|---|
Intersections and Commercial Driveways per Mile | 4.6 | 11.6 | 10.5 |
Cyclist Miles Between Motorist-Caused Crashes | 81,000 (Lowest Risk) | 29,000 (Highest Risk) | 40,000 |
Notice that high-conflict paths had 10 percent more intersections and driveways per mile than the sidewalks, but 38 percent higher motorist-caused crash risk, likely due to the higher bicyclist speeds on the paths (16.3 MPH versus 12.4 MPH for the sidewalks).
If we replace a sidewalk with a sidepath without somehow reducing the turning and crossing conflicts, the risk for the bicyclists will likely increase. This information helps you as a cyclist to decide when a sidepath might be reasonable to use — such as going with the flow on a low-conflict path — and when to avoid one — for example, when going against the flow, or going with the flow fast on a high-conflict path.
Bikeway designers can’t foresee every situation for you, so don’t expect them to.
Next time: Safety in numbers, Vision Zero, and “Preparing the Child.”
Footnotes
[1] The 1976 Bikecentennial study found that 38% of crashes occurred on downgrades, though the constituted only 15% of the route. This was for all types of crashes.
[2] This short video explains lane control and the need for it.
Good Part 2 Mighk! I appreciate your research and sharing of it!
As a retired LEO, longtime bike cop, and continuing instructor and presenter, I eat this kind of stuff up!
Thanks
This is really important stuff, especially in the age of E-bikes, which boost the average (and neophyte) cyclist’s speed into ranges they aren’t experienced with. Speed and its risks definitely get more emphasis in my classroom sessions and tours!
Amen, as e-Bikes and their use will continue to increase!
Since late 2016, I’ve been instructing and presenting to law enforcement and the community on this very issue! Now mixed group riding safety and etiquette too!
I am wondering where the average speed cyclists was learned. It seems very fast from my experience. Since it plays such a large part in the conclusions, can you help me understand how this was learned? Also, is the entire study published somewhere so we can read the actual study? Thanks
First off, note that it’s not the average speed, but the 85th percentile. So it means 85% are going below that speed. The 85th percentile speeds are roughly 3 mph higher than the mean average.
We estimated speeds with the same video used for the counting. Lines were marked on the lanes, sidewalks and paths, and cyclists were timed between those lines to calculate their speeds.
The full report is still being reviewed internally. A video of a presentation to one of our committees can be found here: https://youtu.be/kHZh_RNeI80?t=1980
Thanks for your fast response and the video link. It was helpful in gaining a better understanding. I know you are only presenting a summary and further analysis is happening now, but I was wondering about some things. Please do not take any of this as criticism. Just a search to understand. I hope I am reading it right this time so my questions are relevant.
-Could a conclusion for higher sidewalk rider accidents for drive-outs be the distance from the car vs the speed of the cyclist? Consequently, sidewalk riders are involved in more of these types of accidents. I suppose it depends on direction of the car and where contact is made. Also, were there any trends over the 10 years – did these accidents increase/decrease over the years as awareness about biking increased, SRTS became more popular or …?
– Did accidents increase/decrease with addition of bike lanes (perhaps they were always there)?
-When looking at the sidewalk accidents , side-path vs the street, was there any difference? Or do driveways intersecting sidewalks make for dangerous riding, at the same level no matter all the different driver distractions?
-Are you planning to have the research design reviewed? I expect there is a local university that might be interested in providing some input. The last thing you would want is someone reviewing this later and suggesting the design was weak or conclusions were not considered in all their context, similar to the Montreal study.
Thanks again for doing this. It is so important to our safety. It makes me feel better as a new member to CyclingSavvy and as an advocate promoting CS to our community,
Good questions David.
I tried to get across that idea about proximity to the drive-out. The vast majority of drive-outs involve a vehicle coming from the cyclist’s right, so yes, a sidewalk rider is closer to that vehicle and therefor gets less perception/reaction/braking time. Hard to tease out how much risk varies by speed versus lateral position.
I think ultimately the number of motorist turning and crossing movements (which typically correlates to the number of intersections and commercial driveways) is going to be the primary factor in crash rate, and then the bicyclist’s direction, speed and position will either aggravate or mitigate that base risk.
These bike lanes have been in place for over ten years, so the intent from the beginning was to compare to control streets. Before/after studies are difficult because it takes so long to get enough crashes to do useful analysis.
One notable trend within the data was that motorist-caused crashes for bike lane users increased (by number) on the bike lane streets (comparing the last five years to the first five years), and travel lane crashes decreased on the control streets.
That’s good advice about getting a university review. I’ll look into that.
There is something troubling me that you likely have thought about. I still may be having trouble internalizing the data.
Does the stopping distance, relating to speed mean that the motorist decides to make their turn when cyclists are in certain position on the road? We all do that as motorist/cyclist when we think we can make the turn without getting hit by a bike or car. I am not sure if there is a lesson or conclusion to make but if we looked at the incidents from the viewpoint of being able to make the turn (motorist) without contact or able to cross the street (cyclist) without being hit, aren’t both making the wrong decision at the same time? What influences the motorist that they believe they are turning in a safe manner? Is there a way to look at that? It seems everyone has the same goal but something is getting in the way of the process working. But even if the cyclist does slow down, does it change the decision making by the motorist? For example does the cyclist get hit by the second motorist in line?
Yes David, I’m sure the cyclist speed also has an effect on how well the motorist yields. It’s been both my personal experience that motorist left cross close calls happen at higher cyclist speeds, and that many of the publicized left cross and drive-out crashes involve fitness riders on road bikes. Misjudging the speed of faster vehicles is not unique to bicycling; it’s a problem that runs across all modes, including for pedestrians judging the speeds of approaching vehicles.
Since both parties can be affected by the speed — the perception of it by the motorist, and the stopping distance by the bicyclist — I’d assume both are important. But the perception part is much harder to measure, and can’t be with just the data I have access to.
I can’t imagine why a motorist would do a better job of judging the approach speeds of sidewalk or bike lane cyclists than they would of travel lane cyclists if their relative average speeds weren’t different.
I, too, was intrigued by the speed data.
On the average vs. 85th percentile , this is probably a typo — and the word average probably doesn’t belong here:
“Facing an impending motorist-caused crash, the bike lane user riding at the 85th percentile average speed would need an additional…”
Good catch, Ed. I’ll change that.
Good info, Mighk. I’m not a fan of bikes on sidewalks for adults, but I understand that people *feel* safer there. But what you seem to be suggesting is that the reason they may *be* safer there is because sidewalk riders are going much slower. Moreover, I think cyclists on sidewalks are being discourteous to peds. I’m OK with riding on shared use paths but only if riders always yield to and respect peds.
Otherwise, lane control is always best.
Right Bob, and the same goes for bike lanes; the lower risk (compared to travel lanes) is mostly because the average user is slower (but only for those going with the flow).
The focus of my study was primarily to determine relative risks of motorist-caused crashes for travel lanes, bike lanes, sidewalks, and sidepaths. Courtesy towards pedestrians is an entirely different matter.
How are the stopping distances being determined? Certainly this must be factoring in a substantially delayed reaction time? How softly and ineffectively are the brakes being applied? The figures given do not resonate with my experience or common sense.
A very quick search found this online “stopping distance calculator” tool which appears to differ dramatically from the assumed figures provided in this post: https://www.random-science-tools.com/physics/stopping-distance.htm
The stopping distance is the combined perception/reaction/braking distance. Perception/reaction time is the standard 2.5 seconds; that is what is typically used in traffic engineering. (The time for the link you gave is not described, but it must be less than one second based on the outputs; that’s ridiculously fast.)
Braking distance for cars is actually much shorter than for bicyclists given the same speed. The normal limiting factor for a bicyclist is pitch-over. Stronger brakes can’t stop you better, they just increase the pitch-over potential. With a car you can use hydraulics (and anti-lock technology) to provide very high braking power without concern for pitch-over.
The reference I used for braking distance is here: https://iamtraffic.org/wp-content/uploads/2013/01/speed-infographic-01.png I used the .3g range, for untrained cyclist with dual hand brakes.
All that aside, the real matter is the difference between average/typical speeds. The difference between the bike lane and travel lane for example is that the travel lane speed is 17% faster, but the stopping distance is 25% farther.
It seems that the perception/reaction time used in that calculator (ie “Thinking time used in The Highway Code” in UK) is 1 second, which does explain a lot of the discrepancies. It also seems to me that 2.5sec is quite a bit slower than a person riding a bike is able to react. This article I think does a much better job of explaining the numerous factors that affect perception-reaction time (PRT): https://www.visualexpert.com/Resources/reactiontime.html
” There is no such thing as the human perception-reaction time. Time to respond varies greatly across different tasks and even within the same task under different conditions. It can range from .15 second to many seconds. It is also highly variable. In many cases, the very concept of perception-reaction time simply doesn’t apply[…]
Reaction times are greatly affected by whether the driver is alert to the need to brake. I’ve found it useful to divide alertness into three classes:
Expected: the driver is alert and aware of the good possibility that braking will be necessary. This is the absolute best reaction time possible. The best estimate is 0.7 second. Of this, 0.5 is perception and 0.2 is movement, the time required to release the accelerator and to depress the brake pedal.
Unexpected: the driver detects a common road signal such as a brake from the car ahead or from a traffic signal. Reaction time is somewhat slower, about 1.25 seconds. This is due to the increase in perception time to over a second with movement time still about 0.2 second.
Surprise: the drive encounters a very unusual circumstance, such as a pedestrian or another car crossing the road in the near distance. There is extra time needed to interpret the event and to decide upon response. Reaction time depends to some extent on the distance to the obstacle and whether it is approaching from the side and is first seen in peripheral vision. The best estimate is 1.5 seconds for side incursions and perhaps a few tenths of a second faster for straight-ahead obstacles. Perception time is 1.2 seconds while movement time lengthens to 0.3 second.”
Additionally, I question the assertion that stopping distances are greater for bicycles compared to motor vehicles. It was difficult to find reliable/objective measures for this, but the theoretical analyses I found estimated between equal to half the relative stopping distance for bikes vs cars, given the same PRTs.
I like a lot of things about the graphic from Dan / IAT.
Overall, I am still quite skeptical about the relative importance of stopping distances as an explanation for differences in crash risk, upon which the main conclusions seem to rest heavily.
Thanks for the research Mighk. The crash rate/speed connection makes sense to me.
“I can’t imagine why a motorist would do a better job of judging the approach speeds of sidewalk or bike lane cyclists than they would of travel lane cyclists if their relative average speeds weren’t different.”
Mighk, I might still mis-understand but I do not think they judge the speed better, I think they have more time to stop with bike lane/sidewalk riders. Perhaps as they enter their turn, they see the road differently? They definitely stop looking straight ahead and begin to focus on the road they are turning. During this part of the turn, the motorist no longer worries about being broadsided and they place all their attention on the road. Looking directly down the street they are turning, they see the bike lane/sidewalk cyclists in their path. Certainly the slower speed of the bikers contributes to the fewer accidents as perhaps does the motorists redirected attention to the road/obstacles in front of them. Perhaps the tapes show both motorist and cyclist stopping/slowing down to prevent the accident. If so, are both contributing to the lack of a collision? And, is their learning for the motorist when turning, like do not accelerate too much until passing the cross walk?
But, given we do not know the reason motorists do not judge speed well for bikes (lighting, buildings, ?) at least at some intersections , perhaps one of your considerations for the left cross accidents, at streets with traffic lights, would be to introduce a left turn on a green arrow turn light. It seems for an accident prone intersection, that would be a fairly inexpensive design change. That removes the judgement from everyone.
Mighk, just one other thought that may/may not contribute to your analysis. My previous comments stem from a terrible accident corner in Milwaukee, WI. The crazy thing was the cyclist broadsided the motor vehicle, flying over the hood and receiving sprains and broken bones. The accidents at this freeway exit ramp went on for years until my wife gathered and analyzed the data. The exit is now redesigned.
Here is what happened. Motorists would take the last expressway exit in Milwaukee County. Nearly all people exiting there were turning left at the ‘T’ on a 4 lane county road. As motorists exited, they had an excellent view of the motor vehicles traveling south on the county road, even at 55 MPH. The exit ramp was longer than the typical ones and it angled north, so the view of the road was longer/better. By the time the motorist reached the stop sign, at the T with the county road, their mind was made up that they could proceed to their left hand turn, on the county road. The turn was also made easier since the left lane had a concrete barrier forcing northern bound traffic to the curb lane.
As you would imagine, a cyclist was not as easy to see, as a vehicle. And while this was a popular recreational biking road, cycling traffic was still not very high. However, the ramp was very busy.
So, what would happen. Once the MV reached the stop sign, the motorist was really confident that they did not need to check back to their left again. The motorist would proceed and the unlucky cyclist, traveling at the ‘perfect’ speed (the perfect speed was wrong place, wrong time) would broadside the motor vehicle. The cyclist would go flying over the hood. No deaths, lots of injuries.
Once we learned this as bicycle riders, our tactic to prevent being hit was to move to the middle of the road so the motorist who may/may not have seen us, will have time to move their foot from gas to brake before contact. They had about 12 feet from a dead stop. Our speed also allowed us to quickly move out of the ‘contact zone’ as the motorist was stopping to avoid us. To be clear, we never had a close call before or after our avoidance technique. We assumed that by moving in front of the motorist, they would see us either before or just after accelerating. It was like taking the lane as the motorist was going straight, across the southern lanes, before turning.
An alternative approach, would have been to treat the motorist stop sign as our stop sign. Wait for the motorist to proceed before we proceed. Perhaps the better choice but either we did not think of it or did not want to do it.
Given my experience at this intersection in Milwaukee, I now wonder if there is an inherent design flaw that results in the intersection collisions. We definitely should be practicing techniques that make us safer but I continue to wonder if there is a root cause that might provide a physical change solution or perhaps a motorist training one. You know better than me that a solution to intersection accidents would be monumental.