I dunno. When I was commuting 20 miles a day with a shower at each end, the “vehicular cycling” approach worked well for me. Now that I putter around town with kids on a cargo bike, or their own bikes, and need to go slow to avoid arriving sweaty. I’d like a bike lane: “bicycle driving” is less reasonable when you’re going 10mph, and no amount of “learning” is going to change those conditions.
Truth is the that both types of cyclists exist but growth in the total number of cyclists is likely achieved by appealing to those that would use bike lanes. Of course, that doesn’t mean that cities will manage to build good ones…
Thank you, Tadzik. Your two use cases are genuinely different, and I don't want to suggest that one approach fits all cyclists in all situations.
That said, the intersection conflicts I describe in the essay — right hooks, left crosses, pull-outs — don't disappear for a slow cargo bike with children. A poorly designed bike lane that positions you at the edge of the roadway, out of a turning motorist's sightline, is problematic even for slow cyclists.
Regarding lower bicyclist speed — I don't think it's as much of a problem as you suggest. I have found that, while I sometimes go faster, when riding around town I spend most of my time cycling in the range of 12 to 14 MPH (19–23 km/hr), and I have no problems. Your cargo-bike speed of 10 MPH isn't dramatically different, especially on calm side streets.
On the question of growth, I'd argue that adding a smaller number of genuinely capable, self-empowered cyclists does more for cycling safety and culture than adding a larger number of unprepared ones who depend on infrastructure that may not protect them when it matters most.
Where I think we agree: the infrastructure that actually gets built in the U.S. rarely lives up to the promise. My argument is that skill development is the more reliable path to safety — and the one most consistently ignored.
Love your analogies. Makes sense to anyone who is willing to have a rational discussion on the topic. Unfortunately, it is not a sexy subject, except for those who advocate for infrastructure. Money is always sexy.
Along with your learning to swim analogy, vehicle drivers need to learn how to operate according to the rules of the road. Many have forgotten or never completely understood what to do in different situations. Take for example who goes first at a four way stop, or at a flashing red light intersection. Confusion reigns on the streets at all times.
This is especially true when new things are introduced on roads (such as e-scooters, skaters and wheelchairs) and the "drivers" of those vehicles never took driver's ed or choose to ignore the rules as not applying to them (because they erroneously believe they are pedestrians). Some alternative vehicle drivers even believe, to their potential peril, that it is incumbent on the drivers of heavier vehicles to avoid hitting them even if the rules of the road are not followed.
I believe it is time for society to re-set what the rules are.
Money is indeed always sexy, and you're right that the driver-education deficit is by no means limited to cyclists: motorist education is widely neglected as well.
Just began bike commuting here in Norfolk, VA. We have the typical urban mish-mash of protected bike lines, unprotected bike lanes, sharrows, and nothing. For the most part, I have enjoyed “bicycle driving” even on those roads where there are no special designations. I try to make myself highly visible with fluorescent yellow safety “suspenders,” hi-vis tape, and blinking lights. I ride an e-bike, so it helps that I can edge up to 28 mph if I feel it it necessary to keep with the flow of traffic, but I’m usually averaging around 15-17.
Congratulations on beginning your bike-commuting journey! Where have you learned about bicycle-driving technique? Somewhere online? Have you taken a class?
Honestly, I think I got comfortable riding in traffic years ago when I rode a motorcycle. A lot of things carry over, e.g., look where you want to go, anticipate what the other guy is going to do, be visible, etc.
It's apt that you should mention motorcycles — one of the sources of inspiration in the development of the CyclingSavvy curriculum was the Motorcycle Safety Foundation's safety courses.
Thank you, Justin, for this excellent article. I often struggle to find the right words to explain these issues to fellow cyclists. I'm seeing more and more e-bike riders on our roads—as well as on multi-use paths and even sidewalks—and it reinforces my belief that we need to place far greater emphasis on education rather than simply building more infrastructure.
Pastor, I also enjoy riding in Norfolk. Personally, I'd much rather ride on roads like W. Olney Road or Llewellyn Avenue, despite their divided four-lane design, than use some of the newer bike infrastructure where the lanes are too narrow to feel comfortable. I also think the sharrows on Colley Avenue are a good example of infrastructure that works well without unnecessarily restricting cyclists.
I can't stand riding in curb-adjacent, parking-"protected" bike lanes. I've ridden tens of thousands of urban miles as a vehicular cyclist in the Los Angeles area in recent decades, entirely without any issues. The two times I ever rode in "protected" bike lanes - just to sample them - I had two near-collisions: one a right-hook turn in front of me by a guy turning into a parking lot, and the other by a delivery person who suddenly darted in front of me with his dolly from behind his parked delivery van.
In general, six-foot-wide barrier-"protected" bike lanes simply aren't suitable or safe for faster bicyclists traveling 25 mph and up; the lateral empty space just isn't sufficient (they also have a high propensity to accumulate debris and other junk). Typically in L.A. I see low numbers of generally low-skilled bicyclists and scooterists in such lanes going 8-10 mph. And I mean *low* numbers. In my hometown of Glendale, we were able to rescind the North Brand Boulevard road diet/PBL which was drawing about three users per hour. One of the most "popular" PBLs, the eastern two miles of Hollywood Boulevard, typically draws about twenty users per hour. (These are the two PBLs where I had my near-misses.) I've surveyed Hollywood Boulevard, and the ratio of motor vehicle users (including bus riders) to PBL users is typically about 100/1. From my survey of Hollywood Boulevard, I would also estimate that at least half of all such users could capably get around, by either bike or scooter, without needing a PBL (and of course, nearly all such users got to the PBL in the first place using streets without requiring "protection", some no doubt traveling considerable distances to do so.)
The main advocate for bicycle "infrastructure" in L.A. is Streetsblog's Joe Linton, who's filed over fifty lawsuits against the City of Los Angeles that are essentially about forcing the City to install PBLs whenever certain street-paving jobs are done, per the terms of the "infamous" (in my opinion) 2024 HLA-1 ballot measure. Now what I find most telling about Streetsblog L.A.'s website is that even though PBL advocates claim widespread support for their policies, despite all the efforts Joe Linton puts into the website, daily articles, it appears to draw very little traffic, judging by the fact that it would go months at a time without a single comment were it not for own scathing comments which I pose with regularity (as biker99) and which I get very little pushback to. At least Streetsblog SF will often draw four or five comments to an article, and New York 15 or 20.
The most recent controversial bicycling issue in California has been SB 569, which would make it virtually impossible for a city to remove "protected" bicycle infrastructure, once installed, for *twenty years*. About the only good thing I can see coming from passage of the bill is that it will cause cities to think long and hard before allowing such infrastructure to be forced upon them (usually stealthily) in the first place. It's noteworthy that in her presentation to the California Senate, Catherine Blakespear, the bill's sponsor, immediately referenced "conflated" numbers of pedestrian and bicyclist injuries, when in fact (1) the number of pedestrian injuries routinely far exceeds the number of bicyclist injuries, and (2) pedestrian injuries have virtually nothing to do with PBL-installation. But PBL advocates seem to always use this underhanded tactic when trying to "sell" them to the public (they've done that with the absurd South Huntington Drive project in the South Pasadena area, an area I've ridden safely many times).
The most important thing for me at any intersection is *clarity*, which is best achieved with no "barrier" or "route-channeling" infrastructure at all.
Greg, thank you for this — genuinely one of the most substantive comments this piece has gotten. Your Hollywood Blvd numbers (20 users/hour, ~100:1 against motor vehicle traffic) track with what I keep hearing anecdotally elsewhere, though I'd flag those as your own field observations rather than a published count, worth keeping in mind if you ever cite them further afield.
Your point about 6-foot PBLs not being built for 25+ mph riders is worth backing with a number I came across recently: NACTO's own current design guide states the typical design speed for protected bike lanes and similar facilities is just 10-15 mph — and in the very same document, NACTO states typical e-bike operating speeds run 12-18 mph. By the design guide's own numbers, ordinary e-bike riders already exceed the speed the infrastructure was built for, before you even get to someone riding 25+. That's not a marginal mismatch; it's baked into the design assumptions themselves.
On SB 569 — I looked into this after reading your comment. The "20 years" language checks out, at least for ATP-funded projects specifically: Blakespear's own office states directly that state-funded speed-calming/bike lane improvements "can't be removed for 20 years or until the end of its useful life." Worth noting the bill's actual mechanism is a demanding public/safety-review process rather than a flat prohibition — "virtually impossible" is a fair characterization of how high that bar is, even if it's not a literal absolute ban.
Appreciate the detailed addition to the conversation — this is exactly the kind of comment that makes writing about this stuff worthwhile.
Catherine Blakespear wrote SB569 because the city of Encinitas is working to eliminate her failed designs on Santa Fe Drive and Southbeach 101, both of which include the use of Class IV cycle tracks with bollards and/or wheel blocks. Steve Linke, Serge Issakov and I wrote a case report on the dramatic increase in cycle crashes on Southbeach 101 that occurred during its first 57 months of existence. If you DM me I’ll be happy to send it to you.
wow…I’ve ridden a bike since I was a young kid, I’m now 73. I teach bicycle safety to middle school children, and I ride an average of 100 miles a week if not more. Other than a rare occasion when I might ride in the right lane, forcing cars around me, I would never choose to so, and would never teach kids to do so. I only have my kids “take a lane” when we are going a block or two on back roads with a speed limit of 25 MPH & a clear view ahead. Even then I ask them to hug the right side of the lane because a distracted driver my hit a kid too near the median.
My very presence on a road seems to piss off some drivers, and “taking a lane” seems to only force some of them to do even worse behavior, like passing me on a blind curve or hill. Add to that that the speed of most cars is vastly faster than mine and I can’t begin to support this idea. In Seattle, where I lived for 20 years & still ride, there are good bike lane infrastructure, is it the end all? No. Accidents still happen. But if I thought of “taking a lane” on Westlake Ave. (look it up on Google) where cars are driving at least 30 if not a lot faster, I would be crazy. Given that there is a separate bike lanes just 30 feet away, well clear of any cars, your idea seems like lunacy.
My cousin, a bike rider, also lives in Houston and we talked about it just recently. He told me he rides on a bike path routinely on the ‘edge’ of town, and transports his bike to the ride rather than ride it there. I asked him about riding around town and he said, “it’s far too dangerous”. He won’t even bike to his local grocery store because he is too afraid of the streets to get there.
Gemini agrees, pointing out that “while Houston has built excellent recreational paths, crossing between those paths requires navigating high-speed road gaps.
Gemini also shows that 60 cyclists have been killed in Houston over the last 5 years. You average 11 to 13 deaths a year. The larger Harris county area averages 22 to 26 a year (a 120+ deaths in 5 years).
It goes on to say that “since Houston adopted its comprehensive Bike Plan, essentially zero fatalities have occurred on streets with fully protected bike lanes. Instead the vast majority of severe or fatal crashes happen on standard city streets and high speed arterials where the bike networks abruptly drop off. “
Houston is in the middle to high risk tier of similar sized cities. It is significantly more dangerous than dense, bike forward metros but is not the worse seen in the deep south and Florida. New York and LA have comparable size bu lower per capita deaths due to agressive “protected” bike lanes.
Highest per capita deaths in the country are Baton Rouge and Tampa. both with minimal bike infrastructure.
It goes on to point out that “Houston has an exceptionally low commuter cycling compared to cities like New York or Chicago. Well under 1% commutes.” Wonder why?
It also points out that “the Design Speed of Houston is far faster than other cities. Its roads are designed for 40 or 50 MPH traffic which means that bikers face a much higher chance of death or serious injury.”
I also think this may be a ‘guy’ thing. I’m much more likely to take risks like yours than my wife. My wife already thinks I’m insane riding far to the right on wide shoulders on highways where the speed limit is 45 or 50 mph. She would never consider taking a lane and would never send her grandchildren out to do such a thing. You might ask some women how they feel about it to round out your sample stats.
So I guess I would also have voted you out of the BikeHouston because the results I’ve discussed above show that they were right and you were not.
Hi Al — thank you for your detailed comment, and for your work teaching bicycle safety to young people. Here are a few responses:
On the Gemini citations: I'd be cautious about treating AI-generated statistics as reliable sources. Gemini and similar tools are known to produce plausible-sounding figures that don't always hold up to scrutiny. If those specific claims about Houston fatalities and protected bike lanes can be traced to a published, peer-reviewed source, I'd be glad to engage with them directly.
On Houston fatalities: a proper rebuttal would require a detailed crash-by-crash analysis that I'm not in a position to conduct here. But from what I know of urban cycling fatalities generally, the overwhelming majority involve circumstances unrelated to vehicular cycling — riding at night without lights, wrong-way riding, intoxication, crossing arterials essentially on foot mid-block, and so on. Attributing those fatalities to the absence of protected infrastructure is a claim requiring evidence that, to my knowledge, is not available.
Regarding your Westlake Avenue example: having looked it up, I see that the separated cycletrack along Lake Union sounds like exactly the kind of genuinely separated facility I praise in my essay — and using it instead of the main roadway makes perfect sense. That's entirely consistent with my position.
On the gender point: the reluctance to ride in traffic applies to everyone who hasn't been properly educated, regardless of gender. CyclingSavvy courses routinely transform people — including women — who previously wouldn't have considered riding in traffic into confident, capable cyclists.
All the best — and keep up the good work with those middle schoolers.
if you don't accept the Gemini information you should go do the research yourself on whatever Ai you want to and then look at the underlying reference references. I did the work you should do the work too.
you say that from what you know you have all of these ideas about what causes the actual crash, but if you really have facts to back up your ideas please give me some real data to back up your assumptions because as I have heard many times “to assume makes an ass of you and me”
On crash types — the foundational research is Kenneth Cross's 1977 NHTSA study (DOT HS-803-315) [1], which established that intersection and driveway conflicts dominate bicycle crash causation. It was updated by Hunter et al. in the 1996 FHWA report (FHWA-RD-95-163) [2]. More recently, Mighk Wilson's MetroPlan Orlando research [3] found that 95% of urban and suburban bicycle crashes involve turning and crossing conflicts at intersections and driveways — that study is also linked directly in my essay.
On nighttime riding — NHTSA's 2021 Traffic Safety Facts [4] documents that 52% of cyclist fatalities occur in darkness, despite nighttime riding representing a small fraction of total cycling activity.
These are decades of consistent research. The claims in my essay are not assumptions.
Al, there is fundamentally no difference in riding a bicycle, an e-bike, a moped, and a motorcycle in traffic. The only difference is the maximum speeds at which they are capable of traveling.
Auto drivers already have a duty to slow down when approaching a vehicle traveling slower (even much slower). I used to live in a rural area with a large Mennonite community (the Walmart had designated horse and buggy parking). It was not uncommon to encounter a horse and buggy on local roads and even on state highways. Auto drivers had a duty to slow down when approaching/following and only pass when it was legally allowed and could be done safely. The buggies were occupying the entire lane and everyone just accepted it whether they liked it or not.
I think Justin is making the same case about urban cycling. The more cyclists are confident thinking of themselves as a full lane user (despite their slower speeds) and the more auto drivers encounter lane-taking by cyclists, the more everyone will just get used to it and it will become normalized and, hopefully, safer for everyone.
The stats do not prove that. I too have driven in Pennsylvania Dutch country where the Mennonites and the Amish are and when you're in that area you were driving with an attitude that you might come over the next hill and they'll be a Horse buggy going 25 miles an hour. But even there you don't come over the hill thinking there might be a bicycle in my lane going 25 miles an hour.
Al, I actually lived in rural VA. I don’t know what they feed the horses in PA, but I’ve never seen a horse and buggy do 25 mph.
But you actually are making Justin’s case stronger. As you noted with the PA drivers and the horse and buggies, the more cyclists are riding in traffic, the more car drivers with become normalized to encountering slower (2-wheeled) vehicles.
well have at it Pastor you can take a lane anytime you want. As to me proving his assumptions, of course, if everybody rode in the right hand lane it probably would be a somewhat different world, inhabited by even angrier drivers and there are today.
We have a stretch of road not 1000 yards from my house, two lanes and the city has instead of widening it and putting a decent shoulder on it for both pedestrians and bicycles, have instead put a sign up saying that bicycles have a right to take the lane . This is on a blind short Hill. I have seen numerous times when cars pull up behind a couple of bicyclists taking the lane and then doing what, passing them on the blind hill, in a no passing zone, roaring around them and cutting in front of them. It is just a matter of time before somebody is going to be killed on that road. I refuse to even ride on it it's so dangerous.
His assumptions are not proven by the facts sorry. I can see why he was aousted from his position. If he was unwilling to even go fact check himself that's a pretty damning attitude.
Virtually none of the statistics in modern "studies" on cycling safety are valid. It's well known that cyclists ride more slowly on cycle lanes than the road. Since stopping distance is proportional to the square of the speed, that makes it a pretty important confounding variable, yet it is almost never accounted for.
Al, any AI source is only as good as the sources they're "trained" on and what an AI gives you the user also has a lot to do with an individual's prompt.
None of these AI systems, just like humans, are some sort of absolute source of truth.
Key local example in my area - "local" "news" was doing a report on speed cameras.
They claimed the segment they were covering went from a posted speed limit of 65 mph directly to 45 mph. That's made it into every major AI summary.
But the actual signage on that segment, ironically confirmed by some of their video footage, indicated that the speed limit dropped from 65 to 55, then to 45 with ample signage showing the speed limit reduction ahead and signage as mandated by state law to warn motorists of the upcoming speed camera.
But the article's copy basically left it at "the speed limit reduced from 65mph to 45 mph and there was no adequate warning of the speed camera, therefore motorists were victimized." AI isn't being "trained" to watch for the minor details of the "news" footage or what signage is present in actual reality, therefore it's going to be "trained" by whatever the copy of the news article says happened.
AI systems also tend to want you to keep using them so the tendency to push back on anything you say, unless it borders violating the AI provider's the terms of service, is unlikely.
I worked in the computer field for almost 30 years. I am well aware of A.I. And it’s limitations that’s why I make sure I check it’s underline sources when I’m going to quote them. I don’t know what type of assumptions you wanna make of me trying to teach me about A.I. . I’ll use it as I see fit and I’ll make sure that it’s giving me valid information you should too.
As someone who does a fair amount of vehicular cycling myself and who has completed a few CyclingSavvy courses, I always think of "people of all ages and abilities" when I discuss bike safety advocacy. Will parents with their children on the back seats feel safe and comfortable riding in a 30-mph street mixed with cars? How about elderly people? Do I want my 6-year-old daughter to ride in the street? All of them would be able to ride safely where a connected network of cycling infrastructure is available.
Hi Daigoro — we met when I assisted a CyclingSavvy class in Seattle in October 2017. Small world!
On the substance: I want to be careful not to dismiss the 'people of all ages and abilities' concern, but I'd push back gently on the six-year-old example. We don't design roads, kitchens, swimming pools, or staircases for unsupervised six-year-olds — at some point, adult supervision and skill development are the answer, not infrastructure redesign. A six-year-old riding unsupervised in any urban environment is a parenting decision, not an infrastructure question.
More broadly: the issue isn't whether we want truly separated, well-designed cycling facilities that serve all ages and abilities — of course we do. It's whether what American cities are actually building delivers that safety, or merely the illusion of safety. Painted stripes, flexible posts, and 'armadillos' that disappear at intersections — and in some configurations make things worse — are not the connected network you're describing. That gap between aspiration and execution is precisely what my essay addresses. Infrastructure cannot be a substitute for know-how.
Thank you for the reply, Justin. I wondered if you were in the course I took—you looked familiar! It makes me feel at ease knowing you, like me, are for safe cycling infrastructure that serves people of all ages and abilities. I agree that a lot of what we have in the US today isn't anywhere as safe as it should be—painted lanes, flex posts and unmanaged intersections. OTOH, there are cities moving in the right direction, and I'm hopeful that more elected officials, working with safety advocates like us, will begin making effort to build more "truly" safe bike facilities.
As for children riding unsupervised, that's not what I had in mind. When my daughter rides her bike, I'm near her. And I'm fairly comfortable riding with her on separated bike paths, but not in the street. Hope I've clarified my point.
Daigoro — what a small world indeed! And yes, I think we're very much on the same page. Here's to more cities moving in the right direction. Safe riding to you and your daughter.
You are entirely too kind in your assessment of "protected" intersections. These things force cyclists out of their way, even further from view by motorists, and into awkward bends that destroy momentum, and cyclists thrive on momentum. The only reason they might work is that they slow cyclists down. A speed limit sign would accomplish the same thing. Here we see a "protected" intersection doing it's job so well: https://www.facebook.com/watch/?ref=saved&v=1280309327449266 Segregated intersections are dogshit.
Justin, I read all of these studies you posted to back up your assumptions and beliefs. . Let’s look at them one at a time to see if your assertion that bike infrastructure is not worth investing in, because, as you say: “Do cyclists become safer when the infrastructure around them improves, or when their own skills and confidence improve?” Your thought that investing in bike safety infrastructure is a waste of time and money. Since your post could be used by anti-bicycle groups out there, i.e. see Toronto’s Rob Ford,I feel it’s important to make sure that your assertions are challenged.
My first question is, “why can’t both be true? Why not do both?” We'll leave that question as is.
As to your foundational research is Kenneth Cross's 1977 NHTSA study (DOT HS-803-315) [1], which established that intersection and driveway conflicts dominate bicycle crash causation, the first thing that jumped out at me was the date. This study was done in an era that predated bike safety infrastructure and even the widespread use of helmets. Not even the Tour de France riders used helmets then. I knew no one that rode with helmets back then, myself included. It also commented on the fact that many bicycle riders were riding against traffic not with it. This was a common misconception of the era. I remember doing this too until I understood it’s flaw. So that issue is moot today for virtually all riders. (obviously a few still do it).
This document is 287 pages long, why don’t you use Ai or your own time to summarize this documents findings. But let me point out a few things.
Age: the majority of crashes in 1977 study were from 12 to 18. While education could help with some of these, if these children were not wearing helmets it could easily be part of the statistical sample worth more than whether they had bicycle education and used it. They also did not have a safe bicycle infrastructure. So your claims fail this data set.
Alcohol use: “Alcohol was found that less than one percent of the bicyclists were impaired by alcohol. However, evidence that the motorist had been drinking was found in 3.5% of the non-fatal accidents and 16.9% of the fatal accidents. Alcohol was judged contributory in nearly every case in which it was found present. “
So putting bicyclists into roads with alcohol impaired drivers contributes to their injury or death? Of course. Could safe bike infrastructure get them away from these drunks? Sure. Again, your claims fail this data set. By the way, this impairment is the reason I don’t bike in the late afternoon or evening. I’m worried about drunks that don’t tend to be out in the morning hours.
“About 17% of all accident trips were made during darkness. However, it was found that a significantly greater proportion of fatal (30%) than non-fatal (10%) accidents occurred during darkness. These findings provide strong support for the contention that likelihood of sustaining fatal injuries from a bicycle/motor-vehicle accident is significantly greater when the accident occurs at night.”
Again, if we had safe bicycle infrastructure the likelihood of getting hit at night is reduced dramatically. So again, your data doesn’t make sense to your conclusions.
“Most accidents occurred in close proximity to the operator's residence. The median distance between the accident site and the operator's residence was .6 mile for bicyclists and 2.6 miles for motorists.” Again, this supports the idea of safe bike infrastructure especially near residential areas. If these bicyclists would have been off the streets on safe bike lanes or even had bike lanes on their streets the accident rate would almost certainly be lower. See Amsterdam or Paris to validate this assumption.
Your second reference is only a technical summary, which references the 1977 study. It points out the following:
Examples of bicycle-motor vehicle crash types include:
• Motorist left turn facing the bicyclist. (safe infrastructure can cut down on this kind of accident.)
• Bicyclist left turn in front of traffic. (safe infrastructure can cut down on this kind of accident.)
• Motorist drive-out from a driveway or alley. (getting bicycles onto bike paths can substantially help with this, and having green or otherwise clear understanding that that motorist is pulling out into a bike lane would be helpful)
• Bicyclist ride-out from a stop sign or flashing red signal. (this is clearly educational)
The purpose of the study is documented as:
The purpose of this research was to apply the basic NHTSA pedestrian and bicyclist typologies to a sample of recent crashes, and to refine and update the crash-type distributions, paying particular attention to roadway and locational factors.
The specific study aims were to:
1. Identify and code, according to NHTSA typologies, a recent sample of approximately 5,000 pedestrian-motor vehicle and 3,000 bicycle-motor vehicle crashes.
2. Determine the specific roadway, locational, and other factors associated with each crash type.
3. Identify situations where engineering and/or educational or regulatory countermeasures might be employed effectively to reduce the frequency of pedestrian or bicycle crashes.
I asked Gemini to summarize the findings, since the reference you used did not do that. Here’s the findings.
2. Major Bicycle Crash Findings
Bicycle crashes were divided into two main categories depending on how the bike and the car were moving relative to each other:
Crossing Paths (57% of crashes)
This was the most common setup—the car and the bicycle were traveling on roads that crossed each other.
• Driver at fault: The single most common trigger was a motorist failing to yield the right-of-way to a cyclist.
• Cyclist at fault: The second most common trigger was a bicyclist failing to yield—either riding out past a stop sign/flashing red light at an intersection, or pulling out blindly into the road from a driveway or alleyway.
Parallel Paths (36% of crashes)
This occurred when the car and the bicycle were traveling in the same or opposite directions on the same street.
• Turning and Merging (34.4% of parallel crashes): The driver turned left or right directly into the path of a cyclist traveling next to them.
• Overtaking (24.2% of parallel crashes): A driver hitting a cyclist from behind while trying to pass them.
• Swerving (20.6% of parallel crashes): A cyclist suddenly veering or merging into the path of a car.
(The remaining 6% of bicycle crashes were attributed to random unique circumstances, like a car backing up or a car door opening into a cyclist).
The Big Takeaways & Solutions
The authors concluded that preventing these crashes required a system-wide approach rather than just blaming one side or the other. They recommended three types of "countermeasures":
• Engineering changes: Better road design, including dedicated bike lanes, painted or raised crosswalks, better street lighting, and modifying intersection layouts to give pedestrians more time to cross safely.
• Education: Teaching drivers to look for bikes before turning, and training younger cyclists on how to ride predictably with traffic rather than riding against it or darting out from driveways.
• Regulation: Adjusting laws around right-turns-on-red and enforcing speed limits in high-pedestrian zones.
So again, in addition to Education, engineering changes which include bike lanes, better street lighting and modifying intersection layouts are a primary recommendation.
So again, your data only somewhat supports your assumption.
Your third reference is from Orlando in 2021. It’s core finding according to the video on the front page you referenced stated that, “the speed of the bicyclist affected the severity of the crash..” and went on to say that bike accidents on bike paths are less serious because the cyclists is usually going much slower than on the highway.”
It also states:
Among the key research findings:
• Direction: Bicyclists going against (facing) the flow of normal vehicle movement – on sidepaths, sidewalks, bike lands or travel lanes – are 5 times more likely to be involved in a crash.
• Intersections: Since 95% of urban and suburban bicycle crashes involve turning and crossing conflicts at intersections and driveways, the frequency of those conflict points is a key contributor. (so getting the bikes off roads onto safe infrastructure is obviously going to reduce these crashes). See Paris’ infrastructure improvements.
• Speed: It is generally acknowledged that motorists’ speed can contribute to crashes of all kinds. The research showed that cyclists’ speed can also be a factor. Bicyclists traveling on sidepaths, sidewalks, and bike lanes travel slower than those using regular travel lanes. Slower cyclists need less time and distance to stop when a motorist violates the cyclist’s right-of-way.
It concludes: In short, the data in this study strongly suggests that bicycle lanes, sidewalks and shared use sidepaths do not make bicyclists significantly safer from motorist-caused crashes; rather, the context in which the facility exists (land use, numbers of intersections and commercial driveways).
This can clearly be read as to say that putting in sidepaths that cross car intersections as infrequently as possible is going to lower the deaths and injury of cyclists. This is why so many European cities have implemented these dedicated lanes as well as Seattle, Portland and other cities are doing.
There is no statement that those ideas should not be done, as you are saying.
Finally, your reference to the 2021 Traffic Study facts.
The fact you point out that 52% of the cycle fatalities occurred at night means to me that we need to do a number of things. One is to make sure we drive home the use of lights on bikes at night. Secondly, why shouldn’t we promote dedicated bike lanes and bike paths that can possibly keep bicyclists off the roads with cars and trucks? I don’t understand how that is something you would vote against. The study also does not have any conclusions, it is only a statement of facts. It does point out an interesting fact:
The highest fatality rate per 100,000 population was in Florida (0.90 fatalities per 100,000 people).
I could conclude that Florida, with a very low rate of safe bicycle infrastructure is a good example proving my point. Other cities that have implemented safe bicycle infrastructure, including your hometown of Houston, have seen dramatically lower deaths and injuries.
Now go do the work on my references that you don’t accept.
Al, a few final points, since this has gone on for a while and some of what you're describing isn't what I've argued.
First, nobody in this thread has proposed putting eight-year-olds alone in a traffic lane. That's not my position, CyclingSavvy's position, or anything I've said all week. I'd ask you to engage with what I've actually written rather than a fictional revision of it.
Second, on the Orlando study you cited yourself: its own conclusion states that "bicycle lanes, sidewalks and shared use sidepaths do not make bicyclists significantly safer from motorist-caused crashes; rather, the context in which the facility exists (land use, numbers of intersections and commercial driveways)" is what matters. That's my argument, not yours — you quoted the study that makes my case.
Third, on Houston specifically: 2023 was Harris County's worst year for cyclist deaths in at least a decade — 26 killed, per TxDOT data. That's the opposite of 'dramatically lower deaths,' and it happened during the exact period Houston was expanding its bike infrastructure.
I've engaged with your challenges in good faith all week, including doing real research to answer your citation requests. At this point I think we've each made our case as clearly as we're going to. I'm happy to leave it there.
The sources that Gemini used for the answer that I quoted you were the following:
The statistics provided regarding bicycle crashes and fatalities are compiled from several tracking databases, municipal initiatives, and local safety advocacy reports.
The primary underlying sources for this data include:
### 1. Public Safety Databases & Incident Reporting
* **TxDOT CRIS (Texas Department of Transportation Crash Records Information System):** This is the master repository for all police-reported traffic accidents across Texas. It provides the foundational data for both the City of Houston and the greater Harris County totals. Public safety entities pull micro-data from CRIS to track crashes involving "pedalcyclists."
* **Houston Police Department (HPD) Records:** Local data on fatal traffic incidents is initially processed through HPD's vehicular crimes division before being mapped into city safety registries.
### 2. City Planning & Safety Initiatives
* **Houston Vision Zero Action Plan:** Launched in 2020, this municipal initiative explicitly tracks traffic fatalities to eliminate deaths and serious injuries on Houston streets by 2030. The data identifying Houston's **High Injury Network (HIN)**—the 6% of streets that account for 60% of severe or fatal crashes—is derived directly from their reports.
* **Houston Bike Plan Monitor:** The ongoing tracking of high-comfort versus unprotected street miles managed by the Houston Planning & Development Department.
### 3. Federal Metrics & National Comparisons
* **NHTSA FARS (National Highway Traffic Safety Administration Fatality Analysis Reporting System):** This federal census provides the standardized national numbers used to calculate and compare per-capita fatality rates (such as the numbers cited for Phoenix, New York, and Los Angeles).
* **U.S. Census Bureau American Community Survey (ACS):** Provides the "journey-to-work" data used by safety analysts to calculate the commuter "exposure rate" (the incredibly low <1% mode-share percentage of daily cyclists in Houston).
### 4. Non-Profit Advocacy & Research Reports
* **BikeHouston:** The local cycling advocacy group regularly compiles and contextualizes city crash data. They are the primary source for spatial analysis proving the safety gap—specifically how fatalities heavily correlate with high-speed commercial arterials rather than the city's expanding network of protected lanes.
* **LINK Houston:** A non-profit focused on equitable transportation that co-authored early documentation on the intersections most hostile to vulnerable road users in Houston.
I appreciate you taking the time to dig into this further, Al. That said, I'd note that what's listed here is the *category* of sources Gemini likely drew from — organizations like TxDOT, Vision Zero, and BikeHouston — rather than a specific report, dataset, or study I could actually look at and evaluate. I'd also gently flag that BikeHouston is listed as a source, and BikeHouston's director has publicly made this same "zero fatalities on protected infrastructure" claim — so if that's where it originates, it may not be independent confirmation so much as the same claim being repeated self-referentially.
If you're able to track down the actual underlying CRIS query or report, I'd genuinely welcome the chance to look at it directly. I just want to make sure we're working from primary data rather than a secondhand summary.
once again I think that you ought to do some research yourself. You make a vast array of statements without any underlying data to back it up. I’ve just shown you the sources that my reference used to come up with these “facts” but if you’re gonna want me to continue to read your articles I’m gonna need to understand that you’re actually doing the underlying work instead of just posture a bunch of assumptions on your part based on your one data point which is you.
I dunno. When I was commuting 20 miles a day with a shower at each end, the “vehicular cycling” approach worked well for me. Now that I putter around town with kids on a cargo bike, or their own bikes, and need to go slow to avoid arriving sweaty. I’d like a bike lane: “bicycle driving” is less reasonable when you’re going 10mph, and no amount of “learning” is going to change those conditions.
Truth is the that both types of cyclists exist but growth in the total number of cyclists is likely achieved by appealing to those that would use bike lanes. Of course, that doesn’t mean that cities will manage to build good ones…
Thank you, Tadzik. Your two use cases are genuinely different, and I don't want to suggest that one approach fits all cyclists in all situations.
That said, the intersection conflicts I describe in the essay — right hooks, left crosses, pull-outs — don't disappear for a slow cargo bike with children. A poorly designed bike lane that positions you at the edge of the roadway, out of a turning motorist's sightline, is problematic even for slow cyclists.
Regarding lower bicyclist speed — I don't think it's as much of a problem as you suggest. I have found that, while I sometimes go faster, when riding around town I spend most of my time cycling in the range of 12 to 14 MPH (19–23 km/hr), and I have no problems. Your cargo-bike speed of 10 MPH isn't dramatically different, especially on calm side streets.
On the question of growth, I'd argue that adding a smaller number of genuinely capable, self-empowered cyclists does more for cycling safety and culture than adding a larger number of unprepared ones who depend on infrastructure that may not protect them when it matters most.
Where I think we agree: the infrastructure that actually gets built in the U.S. rarely lives up to the promise. My argument is that skill development is the more reliable path to safety — and the one most consistently ignored.
Love your analogies. Makes sense to anyone who is willing to have a rational discussion on the topic. Unfortunately, it is not a sexy subject, except for those who advocate for infrastructure. Money is always sexy.
Along with your learning to swim analogy, vehicle drivers need to learn how to operate according to the rules of the road. Many have forgotten or never completely understood what to do in different situations. Take for example who goes first at a four way stop, or at a flashing red light intersection. Confusion reigns on the streets at all times.
This is especially true when new things are introduced on roads (such as e-scooters, skaters and wheelchairs) and the "drivers" of those vehicles never took driver's ed or choose to ignore the rules as not applying to them (because they erroneously believe they are pedestrians). Some alternative vehicle drivers even believe, to their potential peril, that it is incumbent on the drivers of heavier vehicles to avoid hitting them even if the rules of the road are not followed.
I believe it is time for society to re-set what the rules are.
Money is indeed always sexy, and you're right that the driver-education deficit is by no means limited to cyclists: motorist education is widely neglected as well.
Just began bike commuting here in Norfolk, VA. We have the typical urban mish-mash of protected bike lines, unprotected bike lanes, sharrows, and nothing. For the most part, I have enjoyed “bicycle driving” even on those roads where there are no special designations. I try to make myself highly visible with fluorescent yellow safety “suspenders,” hi-vis tape, and blinking lights. I ride an e-bike, so it helps that I can edge up to 28 mph if I feel it it necessary to keep with the flow of traffic, but I’m usually averaging around 15-17.
Congratulations on beginning your bike-commuting journey! Where have you learned about bicycle-driving technique? Somewhere online? Have you taken a class?
Honestly, I think I got comfortable riding in traffic years ago when I rode a motorcycle. A lot of things carry over, e.g., look where you want to go, anticipate what the other guy is going to do, be visible, etc.
It's apt that you should mention motorcycles — one of the sources of inspiration in the development of the CyclingSavvy curriculum was the Motorcycle Safety Foundation's safety courses.
I took that course!
Thank you, Justin, for this excellent article. I often struggle to find the right words to explain these issues to fellow cyclists. I'm seeing more and more e-bike riders on our roads—as well as on multi-use paths and even sidewalks—and it reinforces my belief that we need to place far greater emphasis on education rather than simply building more infrastructure.
Pastor, I also enjoy riding in Norfolk. Personally, I'd much rather ride on roads like W. Olney Road or Llewellyn Avenue, despite their divided four-lane design, than use some of the newer bike infrastructure where the lanes are too narrow to feel comfortable. I also think the sharrows on Colley Avenue are a good example of infrastructure that works well without unnecessarily restricting cyclists.
Many thumbs up for this article.
I can't stand riding in curb-adjacent, parking-"protected" bike lanes. I've ridden tens of thousands of urban miles as a vehicular cyclist in the Los Angeles area in recent decades, entirely without any issues. The two times I ever rode in "protected" bike lanes - just to sample them - I had two near-collisions: one a right-hook turn in front of me by a guy turning into a parking lot, and the other by a delivery person who suddenly darted in front of me with his dolly from behind his parked delivery van.
In general, six-foot-wide barrier-"protected" bike lanes simply aren't suitable or safe for faster bicyclists traveling 25 mph and up; the lateral empty space just isn't sufficient (they also have a high propensity to accumulate debris and other junk). Typically in L.A. I see low numbers of generally low-skilled bicyclists and scooterists in such lanes going 8-10 mph. And I mean *low* numbers. In my hometown of Glendale, we were able to rescind the North Brand Boulevard road diet/PBL which was drawing about three users per hour. One of the most "popular" PBLs, the eastern two miles of Hollywood Boulevard, typically draws about twenty users per hour. (These are the two PBLs where I had my near-misses.) I've surveyed Hollywood Boulevard, and the ratio of motor vehicle users (including bus riders) to PBL users is typically about 100/1. From my survey of Hollywood Boulevard, I would also estimate that at least half of all such users could capably get around, by either bike or scooter, without needing a PBL (and of course, nearly all such users got to the PBL in the first place using streets without requiring "protection", some no doubt traveling considerable distances to do so.)
The main advocate for bicycle "infrastructure" in L.A. is Streetsblog's Joe Linton, who's filed over fifty lawsuits against the City of Los Angeles that are essentially about forcing the City to install PBLs whenever certain street-paving jobs are done, per the terms of the "infamous" (in my opinion) 2024 HLA-1 ballot measure. Now what I find most telling about Streetsblog L.A.'s website is that even though PBL advocates claim widespread support for their policies, despite all the efforts Joe Linton puts into the website, daily articles, it appears to draw very little traffic, judging by the fact that it would go months at a time without a single comment were it not for own scathing comments which I pose with regularity (as biker99) and which I get very little pushback to. At least Streetsblog SF will often draw four or five comments to an article, and New York 15 or 20.
The most recent controversial bicycling issue in California has been SB 569, which would make it virtually impossible for a city to remove "protected" bicycle infrastructure, once installed, for *twenty years*. About the only good thing I can see coming from passage of the bill is that it will cause cities to think long and hard before allowing such infrastructure to be forced upon them (usually stealthily) in the first place. It's noteworthy that in her presentation to the California Senate, Catherine Blakespear, the bill's sponsor, immediately referenced "conflated" numbers of pedestrian and bicyclist injuries, when in fact (1) the number of pedestrian injuries routinely far exceeds the number of bicyclist injuries, and (2) pedestrian injuries have virtually nothing to do with PBL-installation. But PBL advocates seem to always use this underhanded tactic when trying to "sell" them to the public (they've done that with the absurd South Huntington Drive project in the South Pasadena area, an area I've ridden safely many times).
The most important thing for me at any intersection is *clarity*, which is best achieved with no "barrier" or "route-channeling" infrastructure at all.
Greg, thank you for this — genuinely one of the most substantive comments this piece has gotten. Your Hollywood Blvd numbers (20 users/hour, ~100:1 against motor vehicle traffic) track with what I keep hearing anecdotally elsewhere, though I'd flag those as your own field observations rather than a published count, worth keeping in mind if you ever cite them further afield.
Your point about 6-foot PBLs not being built for 25+ mph riders is worth backing with a number I came across recently: NACTO's own current design guide states the typical design speed for protected bike lanes and similar facilities is just 10-15 mph — and in the very same document, NACTO states typical e-bike operating speeds run 12-18 mph. By the design guide's own numbers, ordinary e-bike riders already exceed the speed the infrastructure was built for, before you even get to someone riding 25+. That's not a marginal mismatch; it's baked into the design assumptions themselves.
On SB 569 — I looked into this after reading your comment. The "20 years" language checks out, at least for ATP-funded projects specifically: Blakespear's own office states directly that state-funded speed-calming/bike lane improvements "can't be removed for 20 years or until the end of its useful life." Worth noting the bill's actual mechanism is a demanding public/safety-review process rather than a flat prohibition — "virtually impossible" is a fair characterization of how high that bar is, even if it's not a literal absolute ban.
Appreciate the detailed addition to the conversation — this is exactly the kind of comment that makes writing about this stuff worthwhile.
Catherine Blakespear wrote SB569 because the city of Encinitas is working to eliminate her failed designs on Santa Fe Drive and Southbeach 101, both of which include the use of Class IV cycle tracks with bollards and/or wheel blocks. Steve Linke, Serge Issakov and I wrote a case report on the dramatic increase in cycle crashes on Southbeach 101 that occurred during its first 57 months of existence. If you DM me I’ll be happy to send it to you.
More background on SB569...
https://principledbicycling.substack.com/p/an-urgent-statewide-issue-or-a-plot
Good post. I cross-posted.
wow…I’ve ridden a bike since I was a young kid, I’m now 73. I teach bicycle safety to middle school children, and I ride an average of 100 miles a week if not more. Other than a rare occasion when I might ride in the right lane, forcing cars around me, I would never choose to so, and would never teach kids to do so. I only have my kids “take a lane” when we are going a block or two on back roads with a speed limit of 25 MPH & a clear view ahead. Even then I ask them to hug the right side of the lane because a distracted driver my hit a kid too near the median.
My very presence on a road seems to piss off some drivers, and “taking a lane” seems to only force some of them to do even worse behavior, like passing me on a blind curve or hill. Add to that that the speed of most cars is vastly faster than mine and I can’t begin to support this idea. In Seattle, where I lived for 20 years & still ride, there are good bike lane infrastructure, is it the end all? No. Accidents still happen. But if I thought of “taking a lane” on Westlake Ave. (look it up on Google) where cars are driving at least 30 if not a lot faster, I would be crazy. Given that there is a separate bike lanes just 30 feet away, well clear of any cars, your idea seems like lunacy.
My cousin, a bike rider, also lives in Houston and we talked about it just recently. He told me he rides on a bike path routinely on the ‘edge’ of town, and transports his bike to the ride rather than ride it there. I asked him about riding around town and he said, “it’s far too dangerous”. He won’t even bike to his local grocery store because he is too afraid of the streets to get there.
Gemini agrees, pointing out that “while Houston has built excellent recreational paths, crossing between those paths requires navigating high-speed road gaps.
Gemini also shows that 60 cyclists have been killed in Houston over the last 5 years. You average 11 to 13 deaths a year. The larger Harris county area averages 22 to 26 a year (a 120+ deaths in 5 years).
It goes on to say that “since Houston adopted its comprehensive Bike Plan, essentially zero fatalities have occurred on streets with fully protected bike lanes. Instead the vast majority of severe or fatal crashes happen on standard city streets and high speed arterials where the bike networks abruptly drop off. “
Houston is in the middle to high risk tier of similar sized cities. It is significantly more dangerous than dense, bike forward metros but is not the worse seen in the deep south and Florida. New York and LA have comparable size bu lower per capita deaths due to agressive “protected” bike lanes.
Highest per capita deaths in the country are Baton Rouge and Tampa. both with minimal bike infrastructure.
It goes on to point out that “Houston has an exceptionally low commuter cycling compared to cities like New York or Chicago. Well under 1% commutes.” Wonder why?
It also points out that “the Design Speed of Houston is far faster than other cities. Its roads are designed for 40 or 50 MPH traffic which means that bikers face a much higher chance of death or serious injury.”
I also think this may be a ‘guy’ thing. I’m much more likely to take risks like yours than my wife. My wife already thinks I’m insane riding far to the right on wide shoulders on highways where the speed limit is 45 or 50 mph. She would never consider taking a lane and would never send her grandchildren out to do such a thing. You might ask some women how they feel about it to round out your sample stats.
So I guess I would also have voted you out of the BikeHouston because the results I’ve discussed above show that they were right and you were not.
All in good spirits!
Hi Al — thank you for your detailed comment, and for your work teaching bicycle safety to young people. Here are a few responses:
On the Gemini citations: I'd be cautious about treating AI-generated statistics as reliable sources. Gemini and similar tools are known to produce plausible-sounding figures that don't always hold up to scrutiny. If those specific claims about Houston fatalities and protected bike lanes can be traced to a published, peer-reviewed source, I'd be glad to engage with them directly.
On Houston fatalities: a proper rebuttal would require a detailed crash-by-crash analysis that I'm not in a position to conduct here. But from what I know of urban cycling fatalities generally, the overwhelming majority involve circumstances unrelated to vehicular cycling — riding at night without lights, wrong-way riding, intoxication, crossing arterials essentially on foot mid-block, and so on. Attributing those fatalities to the absence of protected infrastructure is a claim requiring evidence that, to my knowledge, is not available.
Regarding your Westlake Avenue example: having looked it up, I see that the separated cycletrack along Lake Union sounds like exactly the kind of genuinely separated facility I praise in my essay — and using it instead of the main roadway makes perfect sense. That's entirely consistent with my position.
On the gender point: the reluctance to ride in traffic applies to everyone who hasn't been properly educated, regardless of gender. CyclingSavvy courses routinely transform people — including women — who previously wouldn't have considered riding in traffic into confident, capable cyclists.
All the best — and keep up the good work with those middle schoolers.
if you don't accept the Gemini information you should go do the research yourself on whatever Ai you want to and then look at the underlying reference references. I did the work you should do the work too.
you say that from what you know you have all of these ideas about what causes the actual crash, but if you really have facts to back up your ideas please give me some real data to back up your assumptions because as I have heard many times “to assume makes an ass of you and me”
Al, you asked for citations. Here are several.
On crash types — the foundational research is Kenneth Cross's 1977 NHTSA study (DOT HS-803-315) [1], which established that intersection and driveway conflicts dominate bicycle crash causation. It was updated by Hunter et al. in the 1996 FHWA report (FHWA-RD-95-163) [2]. More recently, Mighk Wilson's MetroPlan Orlando research [3] found that 95% of urban and suburban bicycle crashes involve turning and crossing conflicts at intersections and driveways — that study is also linked directly in my essay.
On nighttime riding — NHTSA's 2021 Traffic Safety Facts [4] documents that 52% of cyclist fatalities occur in darkness, despite nighttime riding representing a small fraction of total cycling activity.
These are decades of consistent research. The claims in my essay are not assumptions.
[1] https://rosap.ntl.bts.gov/view/dot/1172
[2] https://www.fhwa.dot.gov/publications/research/safety/pedbike/issues/srd95163.htm
[3] https://metroplanorlando.gov/safety/possible-causes-of-cyclist-crashes-are-focus-of-metroplan-orlando-research/
[4] https://crashstats.nhtsa.dot.gov/Api/Public/Publication/813484
Al, there is fundamentally no difference in riding a bicycle, an e-bike, a moped, and a motorcycle in traffic. The only difference is the maximum speeds at which they are capable of traveling.
Auto drivers already have a duty to slow down when approaching a vehicle traveling slower (even much slower). I used to live in a rural area with a large Mennonite community (the Walmart had designated horse and buggy parking). It was not uncommon to encounter a horse and buggy on local roads and even on state highways. Auto drivers had a duty to slow down when approaching/following and only pass when it was legally allowed and could be done safely. The buggies were occupying the entire lane and everyone just accepted it whether they liked it or not.
I think Justin is making the same case about urban cycling. The more cyclists are confident thinking of themselves as a full lane user (despite their slower speeds) and the more auto drivers encounter lane-taking by cyclists, the more everyone will just get used to it and it will become normalized and, hopefully, safer for everyone.
so you're proposing putting eight-year-olds on their bikes out into the middle of a lane of traffic and telling them to take the lane. Really?
Al, I said nothing about anything goes. Age appropriateness, wisdom, and training are all good and necessary.
Great example of pedophrasty there, Al.
https://medium.com/incerto/pedophrasty-bigoteering-and-other-modern-scams-c84bd70a29e8
The stats do not prove that. I too have driven in Pennsylvania Dutch country where the Mennonites and the Amish are and when you're in that area you were driving with an attitude that you might come over the next hill and they'll be a Horse buggy going 25 miles an hour. But even there you don't come over the hill thinking there might be a bicycle in my lane going 25 miles an hour.
Al, I actually lived in rural VA. I don’t know what they feed the horses in PA, but I’ve never seen a horse and buggy do 25 mph.
But you actually are making Justin’s case stronger. As you noted with the PA drivers and the horse and buggies, the more cyclists are riding in traffic, the more car drivers with become normalized to encountering slower (2-wheeled) vehicles.
well have at it Pastor you can take a lane anytime you want. As to me proving his assumptions, of course, if everybody rode in the right hand lane it probably would be a somewhat different world, inhabited by even angrier drivers and there are today.
We have a stretch of road not 1000 yards from my house, two lanes and the city has instead of widening it and putting a decent shoulder on it for both pedestrians and bicycles, have instead put a sign up saying that bicycles have a right to take the lane . This is on a blind short Hill. I have seen numerous times when cars pull up behind a couple of bicyclists taking the lane and then doing what, passing them on the blind hill, in a no passing zone, roaring around them and cutting in front of them. It is just a matter of time before somebody is going to be killed on that road. I refuse to even ride on it it's so dangerous.
His assumptions are not proven by the facts sorry. I can see why he was aousted from his position. If he was unwilling to even go fact check himself that's a pretty damning attitude.
Virtually none of the statistics in modern "studies" on cycling safety are valid. It's well known that cyclists ride more slowly on cycle lanes than the road. Since stopping distance is proportional to the square of the speed, that makes it a pretty important confounding variable, yet it is almost never accounted for.
Al, any AI source is only as good as the sources they're "trained" on and what an AI gives you the user also has a lot to do with an individual's prompt.
None of these AI systems, just like humans, are some sort of absolute source of truth.
Key local example in my area - "local" "news" was doing a report on speed cameras.
They claimed the segment they were covering went from a posted speed limit of 65 mph directly to 45 mph. That's made it into every major AI summary.
But the actual signage on that segment, ironically confirmed by some of their video footage, indicated that the speed limit dropped from 65 to 55, then to 45 with ample signage showing the speed limit reduction ahead and signage as mandated by state law to warn motorists of the upcoming speed camera.
But the article's copy basically left it at "the speed limit reduced from 65mph to 45 mph and there was no adequate warning of the speed camera, therefore motorists were victimized." AI isn't being "trained" to watch for the minor details of the "news" footage or what signage is present in actual reality, therefore it's going to be "trained" by whatever the copy of the news article says happened.
AI systems also tend to want you to keep using them so the tendency to push back on anything you say, unless it borders violating the AI provider's the terms of service, is unlikely.
I worked in the computer field for almost 30 years. I am well aware of A.I. And it’s limitations that’s why I make sure I check it’s underline sources when I’m going to quote them. I don’t know what type of assumptions you wanna make of me trying to teach me about A.I. . I’ll use it as I see fit and I’ll make sure that it’s giving me valid information you should too.
Not making any assumptions about you, Al.
I simply read and responded to what you wrote.
As someone who does a fair amount of vehicular cycling myself and who has completed a few CyclingSavvy courses, I always think of "people of all ages and abilities" when I discuss bike safety advocacy. Will parents with their children on the back seats feel safe and comfortable riding in a 30-mph street mixed with cars? How about elderly people? Do I want my 6-year-old daughter to ride in the street? All of them would be able to ride safely where a connected network of cycling infrastructure is available.
Hi Daigoro — we met when I assisted a CyclingSavvy class in Seattle in October 2017. Small world!
On the substance: I want to be careful not to dismiss the 'people of all ages and abilities' concern, but I'd push back gently on the six-year-old example. We don't design roads, kitchens, swimming pools, or staircases for unsupervised six-year-olds — at some point, adult supervision and skill development are the answer, not infrastructure redesign. A six-year-old riding unsupervised in any urban environment is a parenting decision, not an infrastructure question.
More broadly: the issue isn't whether we want truly separated, well-designed cycling facilities that serve all ages and abilities — of course we do. It's whether what American cities are actually building delivers that safety, or merely the illusion of safety. Painted stripes, flexible posts, and 'armadillos' that disappear at intersections — and in some configurations make things worse — are not the connected network you're describing. That gap between aspiration and execution is precisely what my essay addresses. Infrastructure cannot be a substitute for know-how.
Thank you for the reply, Justin. I wondered if you were in the course I took—you looked familiar! It makes me feel at ease knowing you, like me, are for safe cycling infrastructure that serves people of all ages and abilities. I agree that a lot of what we have in the US today isn't anywhere as safe as it should be—painted lanes, flex posts and unmanaged intersections. OTOH, there are cities moving in the right direction, and I'm hopeful that more elected officials, working with safety advocates like us, will begin making effort to build more "truly" safe bike facilities.
As for children riding unsupervised, that's not what I had in mind. When my daughter rides her bike, I'm near her. And I'm fairly comfortable riding with her on separated bike paths, but not in the street. Hope I've clarified my point.
Daigoro — what a small world indeed! And yes, I think we're very much on the same page. Here's to more cities moving in the right direction. Safe riding to you and your daughter.
You are entirely too kind in your assessment of "protected" intersections. These things force cyclists out of their way, even further from view by motorists, and into awkward bends that destroy momentum, and cyclists thrive on momentum. The only reason they might work is that they slow cyclists down. A speed limit sign would accomplish the same thing. Here we see a "protected" intersection doing it's job so well: https://www.facebook.com/watch/?ref=saved&v=1280309327449266 Segregated intersections are dogshit.
Justin, I read all of these studies you posted to back up your assumptions and beliefs. . Let’s look at them one at a time to see if your assertion that bike infrastructure is not worth investing in, because, as you say: “Do cyclists become safer when the infrastructure around them improves, or when their own skills and confidence improve?” Your thought that investing in bike safety infrastructure is a waste of time and money. Since your post could be used by anti-bicycle groups out there, i.e. see Toronto’s Rob Ford,I feel it’s important to make sure that your assertions are challenged.
My first question is, “why can’t both be true? Why not do both?” We'll leave that question as is.
As to your foundational research is Kenneth Cross's 1977 NHTSA study (DOT HS-803-315) [1], which established that intersection and driveway conflicts dominate bicycle crash causation, the first thing that jumped out at me was the date. This study was done in an era that predated bike safety infrastructure and even the widespread use of helmets. Not even the Tour de France riders used helmets then. I knew no one that rode with helmets back then, myself included. It also commented on the fact that many bicycle riders were riding against traffic not with it. This was a common misconception of the era. I remember doing this too until I understood it’s flaw. So that issue is moot today for virtually all riders. (obviously a few still do it).
This document is 287 pages long, why don’t you use Ai or your own time to summarize this documents findings. But let me point out a few things.
Age: the majority of crashes in 1977 study were from 12 to 18. While education could help with some of these, if these children were not wearing helmets it could easily be part of the statistical sample worth more than whether they had bicycle education and used it. They also did not have a safe bicycle infrastructure. So your claims fail this data set.
Alcohol use: “Alcohol was found that less than one percent of the bicyclists were impaired by alcohol. However, evidence that the motorist had been drinking was found in 3.5% of the non-fatal accidents and 16.9% of the fatal accidents. Alcohol was judged contributory in nearly every case in which it was found present. “
So putting bicyclists into roads with alcohol impaired drivers contributes to their injury or death? Of course. Could safe bike infrastructure get them away from these drunks? Sure. Again, your claims fail this data set. By the way, this impairment is the reason I don’t bike in the late afternoon or evening. I’m worried about drunks that don’t tend to be out in the morning hours.
“About 17% of all accident trips were made during darkness. However, it was found that a significantly greater proportion of fatal (30%) than non-fatal (10%) accidents occurred during darkness. These findings provide strong support for the contention that likelihood of sustaining fatal injuries from a bicycle/motor-vehicle accident is significantly greater when the accident occurs at night.”
Again, if we had safe bicycle infrastructure the likelihood of getting hit at night is reduced dramatically. So again, your data doesn’t make sense to your conclusions.
“Most accidents occurred in close proximity to the operator's residence. The median distance between the accident site and the operator's residence was .6 mile for bicyclists and 2.6 miles for motorists.” Again, this supports the idea of safe bike infrastructure especially near residential areas. If these bicyclists would have been off the streets on safe bike lanes or even had bike lanes on their streets the accident rate would almost certainly be lower. See Amsterdam or Paris to validate this assumption.
Your second reference is only a technical summary, which references the 1977 study. It points out the following:
Examples of bicycle-motor vehicle crash types include:
• Motorist left turn facing the bicyclist. (safe infrastructure can cut down on this kind of accident.)
• Bicyclist left turn in front of traffic. (safe infrastructure can cut down on this kind of accident.)
• Motorist drive-out from a driveway or alley. (getting bicycles onto bike paths can substantially help with this, and having green or otherwise clear understanding that that motorist is pulling out into a bike lane would be helpful)
• Bicyclist ride-out from a stop sign or flashing red signal. (this is clearly educational)
The purpose of the study is documented as:
The purpose of this research was to apply the basic NHTSA pedestrian and bicyclist typologies to a sample of recent crashes, and to refine and update the crash-type distributions, paying particular attention to roadway and locational factors.
The specific study aims were to:
1. Identify and code, according to NHTSA typologies, a recent sample of approximately 5,000 pedestrian-motor vehicle and 3,000 bicycle-motor vehicle crashes.
2. Determine the specific roadway, locational, and other factors associated with each crash type.
3. Identify situations where engineering and/or educational or regulatory countermeasures might be employed effectively to reduce the frequency of pedestrian or bicycle crashes.
I asked Gemini to summarize the findings, since the reference you used did not do that. Here’s the findings.
2. Major Bicycle Crash Findings
Bicycle crashes were divided into two main categories depending on how the bike and the car were moving relative to each other:
Crossing Paths (57% of crashes)
This was the most common setup—the car and the bicycle were traveling on roads that crossed each other.
• Driver at fault: The single most common trigger was a motorist failing to yield the right-of-way to a cyclist.
• Cyclist at fault: The second most common trigger was a bicyclist failing to yield—either riding out past a stop sign/flashing red light at an intersection, or pulling out blindly into the road from a driveway or alleyway.
Parallel Paths (36% of crashes)
This occurred when the car and the bicycle were traveling in the same or opposite directions on the same street.
• Turning and Merging (34.4% of parallel crashes): The driver turned left or right directly into the path of a cyclist traveling next to them.
• Overtaking (24.2% of parallel crashes): A driver hitting a cyclist from behind while trying to pass them.
• Swerving (20.6% of parallel crashes): A cyclist suddenly veering or merging into the path of a car.
(The remaining 6% of bicycle crashes were attributed to random unique circumstances, like a car backing up or a car door opening into a cyclist).
The Big Takeaways & Solutions
The authors concluded that preventing these crashes required a system-wide approach rather than just blaming one side or the other. They recommended three types of "countermeasures":
• Engineering changes: Better road design, including dedicated bike lanes, painted or raised crosswalks, better street lighting, and modifying intersection layouts to give pedestrians more time to cross safely.
• Education: Teaching drivers to look for bikes before turning, and training younger cyclists on how to ride predictably with traffic rather than riding against it or darting out from driveways.
• Regulation: Adjusting laws around right-turns-on-red and enforcing speed limits in high-pedestrian zones.
Part 1 of 2
Part 2 of 2
So again, in addition to Education, engineering changes which include bike lanes, better street lighting and modifying intersection layouts are a primary recommendation.
So again, your data only somewhat supports your assumption.
Your third reference is from Orlando in 2021. It’s core finding according to the video on the front page you referenced stated that, “the speed of the bicyclist affected the severity of the crash..” and went on to say that bike accidents on bike paths are less serious because the cyclists is usually going much slower than on the highway.”
It also states:
Among the key research findings:
• Direction: Bicyclists going against (facing) the flow of normal vehicle movement – on sidepaths, sidewalks, bike lands or travel lanes – are 5 times more likely to be involved in a crash.
• Intersections: Since 95% of urban and suburban bicycle crashes involve turning and crossing conflicts at intersections and driveways, the frequency of those conflict points is a key contributor. (so getting the bikes off roads onto safe infrastructure is obviously going to reduce these crashes). See Paris’ infrastructure improvements.
• Speed: It is generally acknowledged that motorists’ speed can contribute to crashes of all kinds. The research showed that cyclists’ speed can also be a factor. Bicyclists traveling on sidepaths, sidewalks, and bike lanes travel slower than those using regular travel lanes. Slower cyclists need less time and distance to stop when a motorist violates the cyclist’s right-of-way.
It concludes: In short, the data in this study strongly suggests that bicycle lanes, sidewalks and shared use sidepaths do not make bicyclists significantly safer from motorist-caused crashes; rather, the context in which the facility exists (land use, numbers of intersections and commercial driveways).
This can clearly be read as to say that putting in sidepaths that cross car intersections as infrequently as possible is going to lower the deaths and injury of cyclists. This is why so many European cities have implemented these dedicated lanes as well as Seattle, Portland and other cities are doing.
There is no statement that those ideas should not be done, as you are saying.
Finally, your reference to the 2021 Traffic Study facts.
The fact you point out that 52% of the cycle fatalities occurred at night means to me that we need to do a number of things. One is to make sure we drive home the use of lights on bikes at night. Secondly, why shouldn’t we promote dedicated bike lanes and bike paths that can possibly keep bicyclists off the roads with cars and trucks? I don’t understand how that is something you would vote against. The study also does not have any conclusions, it is only a statement of facts. It does point out an interesting fact:
The highest fatality rate per 100,000 population was in Florida (0.90 fatalities per 100,000 people).
I could conclude that Florida, with a very low rate of safe bicycle infrastructure is a good example proving my point. Other cities that have implemented safe bicycle infrastructure, including your hometown of Houston, have seen dramatically lower deaths and injuries.
Now go do the work on my references that you don’t accept.
Al, a few final points, since this has gone on for a while and some of what you're describing isn't what I've argued.
First, nobody in this thread has proposed putting eight-year-olds alone in a traffic lane. That's not my position, CyclingSavvy's position, or anything I've said all week. I'd ask you to engage with what I've actually written rather than a fictional revision of it.
Second, on the Orlando study you cited yourself: its own conclusion states that "bicycle lanes, sidewalks and shared use sidepaths do not make bicyclists significantly safer from motorist-caused crashes; rather, the context in which the facility exists (land use, numbers of intersections and commercial driveways)" is what matters. That's my argument, not yours — you quoted the study that makes my case.
Third, on Houston specifically: 2023 was Harris County's worst year for cyclist deaths in at least a decade — 26 killed, per TxDOT data. That's the opposite of 'dramatically lower deaths,' and it happened during the exact period Houston was expanding its bike infrastructure.
I've engaged with your challenges in good faith all week, including doing real research to answer your citation requests. At this point I think we've each made our case as clearly as we're going to. I'm happy to leave it there.
as I said before I would've definitely voted you out of the Houston bike club.
The sources that Gemini used for the answer that I quoted you were the following:
The statistics provided regarding bicycle crashes and fatalities are compiled from several tracking databases, municipal initiatives, and local safety advocacy reports.
The primary underlying sources for this data include:
### 1. Public Safety Databases & Incident Reporting
* **TxDOT CRIS (Texas Department of Transportation Crash Records Information System):** This is the master repository for all police-reported traffic accidents across Texas. It provides the foundational data for both the City of Houston and the greater Harris County totals. Public safety entities pull micro-data from CRIS to track crashes involving "pedalcyclists."
* **Houston Police Department (HPD) Records:** Local data on fatal traffic incidents is initially processed through HPD's vehicular crimes division before being mapped into city safety registries.
### 2. City Planning & Safety Initiatives
* **Houston Vision Zero Action Plan:** Launched in 2020, this municipal initiative explicitly tracks traffic fatalities to eliminate deaths and serious injuries on Houston streets by 2030. The data identifying Houston's **High Injury Network (HIN)**—the 6% of streets that account for 60% of severe or fatal crashes—is derived directly from their reports.
* **Houston Bike Plan Monitor:** The ongoing tracking of high-comfort versus unprotected street miles managed by the Houston Planning & Development Department.
### 3. Federal Metrics & National Comparisons
* **NHTSA FARS (National Highway Traffic Safety Administration Fatality Analysis Reporting System):** This federal census provides the standardized national numbers used to calculate and compare per-capita fatality rates (such as the numbers cited for Phoenix, New York, and Los Angeles).
* **U.S. Census Bureau American Community Survey (ACS):** Provides the "journey-to-work" data used by safety analysts to calculate the commuter "exposure rate" (the incredibly low <1% mode-share percentage of daily cyclists in Houston).
### 4. Non-Profit Advocacy & Research Reports
* **BikeHouston:** The local cycling advocacy group regularly compiles and contextualizes city crash data. They are the primary source for spatial analysis proving the safety gap—specifically how fatalities heavily correlate with high-speed commercial arterials rather than the city's expanding network of protected lanes.
* **LINK Houston:** A non-profit focused on equitable transportation that co-authored early documentation on the intersections most hostile to vulnerable road users in Houston.
I appreciate you taking the time to dig into this further, Al. That said, I'd note that what's listed here is the *category* of sources Gemini likely drew from — organizations like TxDOT, Vision Zero, and BikeHouston — rather than a specific report, dataset, or study I could actually look at and evaluate. I'd also gently flag that BikeHouston is listed as a source, and BikeHouston's director has publicly made this same "zero fatalities on protected infrastructure" claim — so if that's where it originates, it may not be independent confirmation so much as the same claim being repeated self-referentially.
If you're able to track down the actual underlying CRIS query or report, I'd genuinely welcome the chance to look at it directly. I just want to make sure we're working from primary data rather than a secondhand summary.
once again I think that you ought to do some research yourself. You make a vast array of statements without any underlying data to back it up. I’ve just shown you the sources that my reference used to come up with these “facts” but if you’re gonna want me to continue to read your articles I’m gonna need to understand that you’re actually doing the underlying work instead of just posture a bunch of assumptions on your part based on your one data point which is you.