Stop and give way are danger signs for cyclists

[Tequestra]

Chart of the day: Stop and give way are danger signs for cyclists
By Ben Spraggon Updated 41 minutes ago

Cyclists are more likely to be injured in crashes with motor vehicles at intersections with stop or give way signs than at intersections without signage or with traffic signals, according to a new study by researchers at Queensland University of Technology.

ABC Story Link 1

Related: What happens when cyclist and car collide?

[AdelaidePeter]

I'm confused. I can't see the graphic from the news article in the actual academic paper (here: http://journals.sagepub.com/doi/10.1177 ... 8118773576 ), nor can I even see those numbers in the paper. In fact, in the paper, the injury numbers are very similar, and the hospitalisation numbers percentages are lowest for Stop / Give Way:



What am I missing?

Anyway, wouldn't it be more about the fact that Stop and Give Way signs are put on the more high risk intersections?

[AdelaidePeter]

Replying to my own question...

These simply use the absolute numbers from the first two rows. i.e. out of the fatal, hospitalised and medical treatment rows, 46% occurred at a stop/give way, 37% at no traffic control, and 18% at traffic lights.

Which is pretty meaningless. You might as well say that the safest place to cycle is against the flow of traffic, middle lane of the freeway, at 5pm on a weeknight... because 0% of cyclist injuries occur that way.

So we're not more likely to be injured if a crash is at a Stop / Give Way; it just that happens more often - which isn't the same thing.

[Tequestra]

I'm confused. I can't see the graphic from the news article in the actual academic paper (here: http://journals.sagepub.com/doi/10.1177 ... 8118773576 ), nor can I even see those numbers in the paper. In fact, in the paper, the injury numbers are very similar, and the hospitalisation numbers are lowest for Stop / Give Way:



What am I missing?

Anyway, wouldn't it be more about the fact that Stop and Give Way signs are put on the more high risk intersections?

This is all I could find in conclusion of the source report:

Despite similar distributions of injury severity across the three types of traffic control, more factors were identified as influencing cyclist injury severity at Stop/Give-way controlled intersections than at signalized intersections or intersections with no traffic control.

I'd not actually bothered to look at it, because I assumed that the ABC wouldn't have any reason to lie about something so non-political, or pro-green/left if anything. The link I found most interesting was last year's story about the five riders in QLD who got pulled out on by a silver Magna, in which there was no stop sign pictured.

In my experience, car drivers enjoy pulling out of side streets through stop & giveway signs in front of cyclists much more than they like to run through red lights to massace their political enemies. The reason is that they expect the cyclist to stop, (like pedestrians on footpaths stop at side streets where cars think they have priority (and probably do)), to let them pull out, and they know that they have that easy excuse of blaming the cyclist for riding faster than the offending driver was expecting.

There is no part of the driving test that measures a driver's ability to calculate the speed of approaching vehicles, and so it is acceptable for car drivers to presume that the approaching cyclist must be riding as fast as a pedestrian walks. It is also expected that cyclists keep out the way of the cars because car drivers supposedly pay rego and bike riders assumedly do not, which is more misunderstood codswallop from the nuisances of Australia.

Anyway, I suppose since you've done the proper research whereas I just took the ABC's word for it then we can mark another one down in the ABC fake news file. I didn't think that they'd resort to such tactics on a subject like this one. Sorry on behalf of my public broadcaster.

[human909]

Thanks to going to the source.

What am I missing?

I believe you are missing the fact that the percentages given are pecerntage of injury type. (As in the percentages sum to 100% vertically in the above graph)

Giveway and Stop signs do seem to have the most accidents.

Which is pretty meaningless. You might as well say that the safest place to cycle is against the flow of traffic, middle lane of the freeway, at 5pm on a weeknight... because 0% of cyclist injuries occur that way.

In the language of Yes Minister... Yes and No.

It is far from meaningless. It does give direct indication of where accidents are occuring. This information is very useful. It allows us to focus on the places and the reason why and where accidents are occuring.

You might as well say that the safest place to cycle is against the flow of traffic, middle lane of the freeway, at 5pm on a weeknight... because 0% of cyclist injuries occur that way.

That is true but that would not be a helpful statistic. The previous ones are.


The stated statistics could be improved by weighting them on frequency. But that does not make them irrelevent.

So we're not more likely to be injured at a Stop / Give Way; it just that happens more often - which isn't the same thing.

Yes and no. Cyclists as a population are more likely to be injured at a Stop / Give Way. It is right there in the statistics. The same applies to individual cyclists if their cycling patterns don't differ too much from the cycling population patterns. So even individuals are more likely to be injured at a Stop / Give Way.

What it doesn't touch upon is likelihood of injury independent of frequency. AKA the riskiness of a location/action.


Ah statistics!

[ColinOldnCranky]

To paraphrase quite correct argument made by some here: Where stop and give way signs are is where it is more dangerous for cyclists. The signs do not necessarily make those places more dangerous.

Then again, the Idaho Stop makes sense and implies that indeed the existence of stop signs does introduce a hazard.

[b4igo]

But if we pass through 10x more intersections with stop/give way signs compared to controlled intersections.... that would actually make them safer??

[AdelaidePeter]

So we're not more likely to be injured at a Stop / Give Way; it just that happens more often - which isn't the same thing.

Yes and no. Cyclists as a population are more likely to be injured at a Stop / Give Way. It is right there in the statistics. The same applies to individual cyclists if their cycling patterns don't differ too much from the cycling population patterns. So even individuals are more likely to be injured at a Stop / Give Way.

What it doesn't touch upon is likelihood of injury independent of frequency. AKA the riskiness of a location/action.

Indeed. The more relevant statistic is injuries per cyclist-who-rides-through-that-intersection, which is very hard to measure.

But there is something else interesting in the statistics: Stop / Give ways had a 36.9% hospitalisation rate, compared to 41.4% for both of the other types. i.e. more severe injuries happen at traffic lights, or at intersections with no control; i.e. if you're going to be hit at one of the latter, you're going to be hit harder, on average. But that could still mean one of two opposite conclusions:
1. Drivers go slower through a Stop / Give Way, meaning injuries are less severe, so they are safer.
2. All intersections have their fair share of serious collisions, but the uncertainty of a Stop/Give Way means there are a lot of additional minor collisions, making them less safe.

You really can't know for sure without like-for-like comparison, and that is very hard when comparing different intersections on different routes.

[uart]

In the language of Yes Minister... Yes and No.

Well after all, the paper was written by Nigel Hawthorne. Ar whoops, my mistake, it was Narelle Haworth.

Regarding the relative risks versus the raw numbers, I agree that it's hard to draw too much meaning when we don't the the relative number of intersections of different type traversed by cyclists.

One thing however that does stand out in the data, is that the percent of accidents where the motorists is at fault are far higher at stop/give-way controlled intersections than either non controlled or traffic light controlled. And yes I checked the stats, and those differences are very significant ( statistically).

[human909]

Also this definition this definition on 'no traffic control' or 'uncontrolled' intersections confuses me.

My skimming of the article couldn't find what was considered this. But I would presume that T-intersections and roundabout fit their definition. These are hardly uncontrolled intersections. I don't see how give way and stop signs are in any way more controlled than T-intersections, roundabouts or driveways. They are all uncolled by traffic light but have clear rules under our road rules.

[AdelaidePeter]

One thing however that does stand out in the data, is that the percent of accidents where the motorists is at fault are far higher at stop/give-way controlled intersections than either non controlled or traffic light controlled. And yes I checked the stats, and those differences are very significant ( statistically).

Interesting. (Motorists at fault: 72.69% at Stop / Give Way, 58.68% at uncontrolled, 50.84% at traffic lights).

I'm surprised that motorists were at fault much less often at "uncontrolled" intersections than at Stop / Give Way, because I assume "uncontrolled" mainly means T-junctions, which are an implied Give Way... and we probably all know that motorists can be pretty bad to cyclists at T junctions. The only explanation I can think of is that cyclists are more liable to sail through a T junction, in the case where they're the one who should give way, if there's no Give Way or Stop sign. Or to put it another way: cyclists are better at altering their behaviour when there IS a Stop / Give Way sign, than motorists are.

(In answer to human909, I suspect a roundabout is classed as Give Way if it has one of those Give-way-shaped roundabout signs and dashed give way lines; but the article should have clarified that).

[fat and old]

Ah statistics!

Have belief, will confirm!

[fat and old]

Also this definition this definition on 'no traffic control' or 'uncontrolled' intersections confuses me.

My skimming of the article couldn't find what was considered this. But I would presume that T-intersections and roundabout fit their definition. These are hardly uncontrolled intersections. I don't see how give way and stop signs are in any way more controlled than T-intersections, roundabouts or driveways. They are all uncolled by traffic light but have clear rules under our road rules.

It comes down to the terminology. Give way and Stop intersections have signs. They "controlled" through those. The colour of the sign Human! Before you tell me Roundabouts have a sign....

Traffic Management 101

[Tequestra]

Thanks for spurring me to read through the actual report in full at last. One stat' I hope might help to shed some light on the situation was in the Empirical Results -> Descriptive Statistics paragraph.

almost all (94.4%) occurred while the cyclist was riding straight ahead.

This fits my anecdotal lived-experience that cars seem to like to pull out from stop / giveway signs in front of or onto passing cyclists travelling straight ahead, although back up at the end of the abstract, they state that:

cyclists at uncontrolled intersections can misjudge the distance and speed of a vehicles on the through approach [sic]

Considering almost 95% of cyclists were travelling along the 'through' approach, I think that it could be concluded that 19 times out of 20 it is the car driver at the stop sign that misjudges the distance and speed of a cyclist on the through approach.

It is also sometimes a gamble to enter a main road after turning left from a side street when there is a car approaching from say, for example 100m away, and trust that they will see the cyclist 5 or 10 m down the road and veer a little to the right to pass them, as they would be expected to do if the cyclist had been riding straight through along main road prior to the side street. Who really is at fault in that kind of accident? Does it depend on the number of seconds that the cyclist has been visible on that main road before the car runs into them from behind?

[human909]

Also this definition this definition on 'no traffic control' or 'uncontrolled' intersections confuses me.

My skimming of the article couldn't find what was considered this. But I would presume that T-intersections and roundabout fit their definition. These are hardly uncontrolled intersections. I don't see how give way and stop signs are in any way more controlled than T-intersections, roundabouts or driveways. They are all uncolled by traffic light but have clear rules under our road rules.

It comes down to the terminology. Give way and Stop intersections have signs. They "controlled" through those. The colour of the sign Human! Before you tell me Roundabouts have a sign....

Traffic Management 101

Yep. I figured as such. There is no way that a T intersection is somehow less 'controlled' than a 4 way intersection with give way signs.

It is a somewhat arbitrary division that can readily skew the reporting of the statistics.

[AdelaidePeter]

Thanks for spurring me to read through the actual report in full at last. One stat' I hope might help to shed some light on the situation was in the Empirical Results -> Descriptive Statistics paragraph.

almost all (94.4%) occurred while the cyclist was riding straight ahead.

This fits my anecdotal lived-experience that cars seem to like to pull out from stop / giveway signs in front of or onto passing cyclists travelling straight ahead

I don't think that's what they mean. There's nothing in the tables about whether the cyclist was going straight ahead or turning. But there is this entry in table I:



Even though those numbers don't exactly match up (I get 94.78% total on "straight", not 94.4%), I think that is what they are referring to. So 94.4% of crashes occurred where the road(s) were not curved, and that 94.4% includes times when the cyclist was doing a 90 degree turn between two straight roads. (And the quoted sentence is really badly worded).

[AdelaidePeter]

Sorry to keep posting, but there is in fact more information on roundabouts etc. (First column is "Stop / Give way", 2nd column is "uncontrolled", 3rd column is traffic lights):



The study was done on Queensland stats, and I guess uncontrolled cross roads are a Queensland thing? I can't think of any cross roads in SA (or Vic for that matter) without either a traffic light, Stop / Give Way, or roundabout.

[Tequestra]

I don't think that's what they mean. There's nothing in the tables about whether the cyclist was going straight ahead or turning. But there is this entry in table I:



Even though those numbers don't exactly match up (I get 94.78% total on "straight", not 94.4%), I think that is what they are referring to. So 94.4% of crashes occurred where the road(s) were not curved, and that 94.4% includes times when the cyclist was doing a 90 degree turn between two straight roads. (And the quoted sentence is really badly worded).

You're having more luck with those tables than I did, but if that's what you have gathered by the meaning of 'straight ahead' then I'll leave you to it, Peter. Good luck.

[NASHIE]

Sorry to keep posting, but there is in fact more information on roundabouts etc. (First column is "Stop / Give way", 2nd column is "uncontrolled", 3rd column is traffic lights):



The study was done on Queensland stats, and I guess uncontrolled cross roads are a Queensland thing? I can't think of any cross roads in SA (or Vic for that matter) without either a traffic light, Stop / Give Way, or roundabout.

I think you have to head way out bush to find uncontrolled cross roads. Can't imagine there are many roundabouts thats not a "give way" ?

[find_bruce]

I think you have to head way out bush to find uncontrolled cross roads. Can't imagine there are many roundabouts thats not a "give way" ?

You need to get out more - here is a T-intersection in the inner-west of Sydney I can't think of a 4 way (or worse) intersection that is uncontrolled though

[NASHIE]

I think you have to head way out bush to find uncontrolled cross roads. Can't imagine there are many roundabouts thats not a "give way" ?

You need to get out more - here is a T-intersection in the inner-west of Sydney I can't think of a 4 way (or worse) intersection that is uncontrolled though

I get out plenty thanks. Yes agree, i can think of a few 'T-intersections' in Perth as well, but no 'cross roads' or 4 way if that helps.

[Tequestra]

I think you have to head way out bush to find uncontrolled cross roads. Can't imagine there are many roundabouts thats not a "give way" ?

You need to get out more - here is a T-intersection in the inner-west of Sydney I can't think of a 4 way (or worse) intersection that is uncontrolled though

I get out plenty thanks. Yes agree, i can think of a few 'T-intersections' in Perth as well, but no 'cross roads' or 4 way if that helps.

In WA at least, unsigned crossroads rely on the 'give way to the right' rule.

[Mububban]

The best advice anyone ever gave me when I was a teenager starting to ride on the roads was "imagine you are invisible, and that nobody can see you. Expect every car pulling out of a side street/stop sign/give way to hit you, then you'll be ready to put the brakes on."

This advice has saved me many times. I don't ride in a state of constant panic, more a state of constant vigilance. Expect to be disappointed by drivers, and you'll rarely be disappointed for long.

Nowadays I ride with flashing lights front and rear even in bright summer daytime, which I believe makes some positive difference to my visibility. It still didn't stop the last guy in a Hilux pulling out in front of me across 2 lanes, but as I was *expecting* him to do it, I was able to brake safely.