Paul L.

Ok, well here it is, a recent comment from Brian on the thread which shall remain nameless has got me thinking. Assuming we had money to run such a study, what would we do to make an objective study of Bike Lanes Vs. Wide Outside Lanes vs. Narrow shoulder lanes Vs. Share the road sections.

I'll start.

We would need to conduct the study in such a way as the participants would not realize to what end the data was being gathered so as not to skew their input.
A large group of beginning cyclists would need to be used to determine what help facilities actually gave them if any.
The study would have to include many different cities but categorize the bike lanes in their degree of VC friendlieness (ending before an intersection, crossing over a right turn lane, staying to the right of the right turn lane).
The study would need to be able to ascertain the cyclists knowledge and comfort with cycling in traffic.
The average daily vehicle number would need to be taken into account to compare roads and cities as well as speed limits.

Ok, that is my start on it. I have to go but please submit your input. Not that it will go anywhere but wouldn't it be fun to send to League of American Bicyclists or find a safety organization that might show some interest in it? Or at least get us all thinking together for a few minutes about what we really should expect from a bonafide study on bicyle facilities and methods in the current day and system.

Tom Stormcrowe

I hadn't seen this thread, I just proposed something similar on the JF thread! It could be done on a volunteer basis with a spreadsheet and volunteer reporting from members collecting data in their municipalities. I'll be glad to to the computer work and perhaps we can arrive at some current data.

Paul L.

We might be able to attack this two ways. My first proposition was kind of a "Fantasy Football" type idea that we could design a study that would prove this once and for all assuming money and time.

Your idea is good too. I am going to email some of the local city traffic engineers and see if they have any data on this.

Tom Stormcrowe

I made the offer on the other thread and everybody seems more interested in fighting!

Let me know what you get and if nothing else, we can start up the DB! Let's see what we find out.

randya

You could set up video cameras at various locations where each type of facility exists and analyze the motorist-bicyclist interactions at each location. This is how San Francisco and Portland 'tested' Sharrows.

Tom Stormcrowe

Good idea, but my concept is a $0.00 budget effort from Law enforcement stats and volunteer collection. I'm willing to donate the time to crunch the data.

I-Like-To-Bike

I'm confused. What metrics are you measuring? Database of what kind of data? Number of cyclists/motorists, age, traffic density, accident counts, near misses, distance from passing cars, etc? Presumably you need to figure out what it is you want to know before going out to measure/record data.

Tom Stormcrowe

Yes, that's why I stated that we would need to set the parameters of the database. It's just an idea at this stage.

I-Like-To-Bike

Suggestion: if the law enforcement stats and volunteer collection does not include some sort of measure of the severity of the accidents, save your time and don't bother. Because such raw "crash data" cannot be accurately crunched.

Tom Stormcrowe

Parameters I was thinking of:

• Frequency (Total number of incidents in annual period to be determined)
• Weather
• Severity
• Personal injury Y/N
• Population of the area
• AVG traffic count at the incident site
• Bike facility available Y/N
• Facility used Y/N

Tom Stormcrowe

Data source citing as well (ie: Police Statistics, city, etc to be cited in a sources page. The study will have to be peer able to stand peer revue).

randya

I think I've got about four years of the City of Portland's crash data for 'pedal cycles'; trouble is I'm pretty sure the raw data I've got is all in hard copy. I previously performed my own evaluation which I placed in a couple of excel spreadsheets. If you want my analysis in excel PM me with an email address and I'll send it to you. But I'm not sure the data or the evaluation says much about the safety of facilities vs. VC/no facilities, it is mostly just an analysis of the causes of the most common types of crashes (e.g. overtaking, failure to yield when turning, wrong way riding, failure to obey traffic control device, etc.) and assignment of responsibility to either the motorist or the cyclist.

I-Like-To-Bike

Now your talkin'! Especially with severity and traffic count at the incident site. Don't forget time of day or at least light vs. darkness.

chipcom 

In the context of the facilities debate, a key piece of data would be simply car-bike collisions by:

1. occuring on the roadway, outside of any bikeway
2. occuring on bikeways by type - bike lanes, bike paths, sharrows, sidepaths, MUPs (we'd have to flesh this out)
3. total

I think by limiting the scope to car-bike collisions, rather than accidents, we can get more meaningful data in regards to the debate and avoid the problems ILTB mentioned when it comes to identifying severity of injuries. Of course determining if a car-bike collision occured in or out of a bikeway might require more effort, based on how a locality reports it's data.

I think the KISS principle applies - and this data is the simplest we can gather to provide some meat to the facilities debate one way or the other.

Of course this is just my suggestion, I'll do what I can to assist using whatever everyone agrees is the proper format.

deputyjones

Weather, severity and personal injury are standard on most accident reports. Here in MI all law enforcement Officers use the same report which captures that data. Most of the rest could be found from traffic studies and census data except for bike facility and if the facility was used. I am not sure how you would capture that. Maybe in an area where bicycling is more common (like Portland) they capture this data?

Tom Stormcrowe

Yeppers, I agree, for the purpose of the study, a car/bike collision is all the study will cover parameter wise. I've started setting up the DB now and will be fine tuning it over the next few days. Right now I need to set up the macros, etc. I should be able to start accepting preliminary data by the weekend to start testing the macros and chart.

Helmet Head

Vehicular Cycling = Same Roads Same Rights Same Rules
Bike Lanes = Same Roads Different Rights Different Rules

No way to design around that. It's fundamental to the essence of a bike lane.

chipcom 

That's my concern...one might have to actually read each report and/or research the location of the accident.

Tom Stormcrowe

They should have an annual summary locally, if we are lucky!

Select cities by State? Limit the study to 50 cities, perhaps? Like I said, it's a Q&D survey study and that would fulfill a regional study category.

RobertHurst

Chipcom Man, even gathering info as simple as that from currently available sources would be extremely difficult. Any BL-related data that does happen to present itself (from currently available sources) should be viewed as highly suspect. Just a for-instance, I was involved in a car-bike collision on a street with a bike lane, and the bike lane wasn't mentioned in the police report, nor was my lateral position mentioned, nor my speed, etc. because none of those factors was deemed relevant to determine liability. We wouldn't be able to get good 'data' on collisions occurring on bike-laned streets vs. non bike-laned streets from police reports, and we sure couldn't get good info on whether or not someone was inside or outside the BL if there was one present.

No good BL-related data available, that means someone will have to go out and get good data -- somehow. Good luck. A lot more work than looking at some pdf's and plugging numbers into a spreadsheet. And even if someone here had tremendous data-gathering energy I very much doubt such a fresh study could provide much compelling evidence one way or the other, let alone put the issue to rest. Too many variables. A nice survey might work, you say. Note that William Moritz addressed the safety of different types of facilities in his survey of LAB members, "Adult Bicyclists in the United States"; he found bike-laned streets to have a much lower 'Relative Danger Index' than non-bike-lane streets. While an interesting result from a relatively robust survey, it is vulnerable to all sorts of valid, crushing criticism, at its very core, and we shouldn't get overly excited about it. People would almost certainly end up saying the same nasty things about any frankenstudy we could cook up here. And his survey cost (iirc) $6500.

Robert, wet blanket

Tom Stormcrowe

From my viewpoint, Robert, it gives me, if nothing else, real world experience "crunching numbers". My current career path is Psychology and there is a tremendous amount of statistical work involved in the field, so it will help me out if nothing else! It may have some impact in other ways as well with quantifying data on a regional basis and serve as a springboard for other studies. Isn't that part of advocacy?

RobertHurst

All I'm saying is that there are no BL-related numbers to crunch. Any numbers that appear to be crunchable are probably bogus. I wish it weren't true. If we had some numbers you would be the guy to crunch the hell out of those suckers.

Robert

SamHouston

Roberts got a wet, blankety point about gathering the data. In the past I've been in hospitals with recently injured cyclists & observed doctors/nurse practitioners asking some relevant questions (were you wearing a helmet? were you working at the time of your accident?) that they are supposed to ask with each accident they treat. Other times...not so much & I end up reminding the patient if it's a work injury to inform the doctor or the nurse with the paperwork the doc will sign so that the powers that be will get their info. That program, at least as far as work injuries go, works relatively well for such a large undertaking, and yet I've seen it flop firsthand.
Not that I'd bother an ER doc about that, even though it's important for the patients I'm there to see it can be rectified, and IMO ER docs have enough on the plate without someone busting their chops over a forgotten procedural query.

Tom Stormcrowe

I definitely understand your point and intend to apply a type II statistical error possibility on the facilities available numbers that I am able to gather. That allows for inaccuracy of up to 50% for error to allow for suspect data If even available). Other data will be type I of +/- 6%.

Tom Stormcrowe

States I need data for:
Alabama
Alaska
Arizona
Arkansas
Colorado
Connecticut
Delaware
District of Columbia
Florida
Georgia
Hawaii
Idaho
Illinois
Iowa
Kansas
Kentucky
Louisiana
Maine
Maryland
Massachusetts
Michigan
Minnesota
Mississippi
Missouri
Montana
Nebraska
Nevada
New Hampshire
New Jersey
New Mexico
New York
North Carolina
North Dakota
Northern Marianas Islands
Ohio
Oklahoma
Pennsylvania
Puerto Rico
Rhode Island
South Carolina
South Dakota
Tennessee
Texas
Utah
Vermont
Virginia
Washington
West Virginia
Wisconsin
Wyoming