Monday 7 April 2014

Where the crashes are: Shared Spaces and other poor junction designs which don't protect cyclists lead to crashes and injuries

A useful website shows where all crashes have occurred on Dutch roads. I've used it below to demonstrate the relative safety of different roads and cycle-paths in this country.

Is Shared Space safe ?
The Laweiplein Shared Space "squareabout" in the small town of Drachten, highlighted in red, has been the subject of much hype. Many claims are made for a low accident rate here but the evidence does not support this. In fact, this single intersection is the second most dangerous location in Drachten for cyclists. From the map data you can see immediately how more cyclist crashes and injuries occur here than the conventional Dutch roundabout a few metres to the east. The Laweiplein is not only more dangerous than the nearby roundabout, but far more dangerous than the safest Dutch roundabout design as used in the larger city of Assen a few kilometres away. Assen has twenty-one roundabouts which between them cause fewer cyclist injuries than the Laweiplein "squareabout" causes all on its own. (Blue flags indicate crashes, yellow flags are where injuries occurred due to crashes.)
Drachten is a small town of under 45000 people. The Laweiplein, highlighted in red, has had 10 crashes reported between 2007 and 2012. That's a larger number of reported collisions than any other location close to the city centre. It's also higher than the number of reported crashes (7) on the busier but more conventional (conventional for the Netherlands. i.e. with segregated cycle-path) roundabout a few metres to the east. Also compare with a truly safe roundabout design in Assen a little further down this page.

Cyclists using the pavement to ride around the Laweiplein.
Pavement cycling is a symptom of inadequate subjective safety
at this junction, a problem common to many shared spaces.
The nature of the crashes at the Laweiplein are also interesting. While safe roundabout design results in cyclists and pedestrians being isolated from crashes between cars, the Laweiplein seeks to mix cyclists with motor vehicles. Even though many cyclists ride on the pavement rather than the road because they know the road isn't safe, the Laweiplein still managed to involve four cyclists in crashes and injure three of them. The more conventional roundabout which serves much of the same traffic a few metres to the east involved only one cyclist in a crash and that cyclist was not injured.

It would appear that the Laweiplein is more dangerous for vulnerable road users - a problem which has often been noted with Shared Space. The safety record of the Laweiplein clearly does not justify the hype from Shared Space enthusiasts and this design should not serve as an example for future developments. It would likely have been safer had this junction been given a more normal roundabout treatment.

Drachten crashes, injuries and deaths 2007-2012. Does this
really look like it's been reduced to "1 per year" ?
Many sources, Wikipedia included, include a claim that "yearly accidents were reduced to 1" in the centre of Drachten due to the introduction of Shared Space. This claim does not stand up to much investigation. Even the Laweiplein on its own has double the claimed accident rate for the entire city centre, and that's just one junction. Look at the rest of the city centre, part of which is shown right, and you see many more. The claim of there being just one accident per year simple does not stand up to any analysis at all.

The Laweiplein "squareabout" on its own is the site of more cyclist injuries than all 21 roundabouts and all the Simultaneous Green traffic light junctions in Assen combined. Far from adding to cyclist safety, he Laweiplein is the second most dangerous location in Dracthen for cyclists. This should not be regarded as a safe design.

Update 2018: Unfortunately, people continue to make the claim that the Laweiplein is safe and this results in other places trying to repeat that "success". The actual result is unsurprisingly that more junctions are built in a similar style which actually are quite unsafe. Another "Shared Space" roundabout built in 2017 has been named as the most unsafe roundabout in the Netherlands.

The Shared Space area of Haren is highlighted in red. The Shared Space layout road through Haren is where the largest concentration of crashes has occurred in this town between 2007 and 2012.

We visit this town on our study tours, cycling into the town on the relatively safe cycle-path which leads from the South East in this image and leaving on the relatively safe cycle-path to the North West. You can see quite clearly how the road both South and North of the Shared Space area has far fewer crashes than does the Shared Space itself.

A six year old blog post discusses some of the many problems in Haren and none of my objections from that time have changed.

Shared Space does not work well. The number of crashes which have occurred should speak for themselves. Claims have been made for enhanced safety in Haren just as in Drachten, but these are all in comparison with an earlier road layout with higher speed limits. It is probably that a more normal modern Dutch street layout would be safer than the current shared space. For example, the nearly car free streets in the centre of Assen (a have a much better safety record even though Assen's population is more than three times greater than Haren.

Also note the cluster of two blue and two yellow flags on top right of the image. These are residential roads which serve as a rat-run in Haren. They're over-used by speeding through traffic at rush hour times because drivers as well as cyclists seek to avoid the Shared Space and it was on these roads at rush-hour that I experienced the closest thing to real road rage that I've ever seen in the Netherlands.

Update: Shifting goal-posts - three more examples
No discussion of the shortcomings of Shared Space is ever complete without the sound of goal-posts being shifted. There are as many opinions about what Shared Space "really is" as there are advocates for it. Naturally, this blog post has resulted in the same attempts to paint any given examples as "not actually shared space" or as unusual "bad" examples.

I didn't pick the examples above because they were particularly bad, but merely because they've been discussed often. The wikipedia page claim about Drachten has come up many times and I've written about Haren before. Actually, neither of these two examples are particularly dangerous as Shared Spaces go.

Here are three extra examples showing the frequency of collisions at other shared spaces in the Netherlands, including those designed by Hans Monderman himself:
Quite near the Laweiplein in Dracthen is "De Kaden", another much lauded Shared Space. The Noordkade and Zuidkade run on either side of car parking and trees from left to right on this map. The junctions on this street, at either end and in the middle, have shared space designs and each junction has seen incidents. In total, there have been more than 40 incidents between 2007 and 2012. This is a worse record than the Laweiplein, illustrated above. At an average of eight per year, this short road, an example of Shared Space, has as many crashes each year as some Shared Space advocates claim was the total for the entire city centre before Shared Space supposedly reduced the total. And these are not just crashes without consequence: there have been four injuries to cyclists and two to pedestrians. The total number of incidents is similar to that for the most dangerous junction in the Netherlands, shown a little further down this page.
The main street through Muntendam, a village of only 4500 people, is a shared space. There's a line of incidents stretching the entire length of the shared space from top left to bottom right on the map.
Oosterwolde is a town of less than 10000 people. The shared space follows the red line on the map. Other streets in Oosterwolde also lack segregated infrastructure, particularly the west-east route at the bottom of the map. In my experience, it's not a very pleasant place to cycle.
There are many other examples. For instance, both the small villages of Onnen (population 430) and Boornbergum (less than 1500) have had fatalities within their shared spaces. However, I don't consider that we can find very much significance in such a small amount of data from villages like this. There is too much variation over time. The same applies, of course, for claims of a miraculous lack of incidents in other small villages.

I think it's sufficient to say that exaggerated claims of safety are made for shared space. Shared space should not be considered to be a magic bullet which removes danger. The one thing shown repeatedly to remove danger is slowing, restricting and removing motorized vehicles from where people walk and cycle. This applies especially to through traffic. Of course, when you take those vehicles away, the space is no longer shared. Read other blog posts about Shared Space, including video.

Not all dangerous roads and intersections in the Netherlands are Shared Space. Read on for other examples:

Not all Dutch roundabouts are created equal. We visit this roundabout in Groningen on study tours as an example of something not to emulate. No fewer than 38 collisions have taken place here between 2007 and 2012, including several injuries to both cyclists and pedestrians.

The design of roundabout shown above is not one which we recommend. It doesn't work well in the Netherlands and given the relative lack of experience of non-Dutch drivers with this layout I'd expect it to work even less well in other nations. For that reason, we recommend instead taking note of the safest design of roundabout design used in Assen.

Genuinely safe design: At this example of the safest design of roundabout used in the Netherlands there were just four "fender benders" between cars. No injuries to anyone and no cyclists or pedestrians were involved at all.
This roundabout in Assen is exceptionally well designed and has proven to be safe. Cyclists are segregated properly from drivers on this roundabout and the design makes it easy for everyone to see what they should be doing.

Just four incidents occurred at this roundabout between 2007 and 2012. All of them involved only motor vehicles and there were no injuries. We will look closely at the features of this roundabout on this year's study tours.

Read more about safe roundabout design.

Simultaneous Green junctions
Where a roundabout is not the right solution, traffic light junctions may be required instead. The safest traffic light junction design uses Simultaneous Green lights for cyclists. Read more information at that link about this very safe design.

The most dangerous junction in the Netherlands
Another site that we visit on the study tours in order to see how and why it doesn't work well is this complex junction, a gyratory which is the most dangerous road junction in the whole of the Netherlands. 29 collisions have occured at one side of this gyratory and another 10 at the other. While there is segregated cycling infrastructure at this junction, it's not well designed.

Every user of this junction has too much to take in. Cyclists, pedestrians and drivers all make mistakes all the time. You can see this in a video. We use this junction to demonstrate what not to do.

What causes crashes ?
The line made up by blue and yellow flags on the left is the motorway. No cycling is allowed on the motorway and all these crashes were between motor vehicles. The line highlighted in red is the equivalent main North-South route by bicycle, a busy cycle-route here in Assen. This cycle-path supports riding at very high speeds and is well used by racing cyclists as well as by recreational riders and the occasional dog walker.

No reported crashes have occurred on the cycle-route between 2007 and 2012. Crashes occur predominantly where cars are. Unravelling of routes keeps cyclists from danger.

In another example, rural roads are also punctuated regularly by blue and yellow flags. Only one of the flags on this map shows a collision between a car and a bike all of the others are between cars or between cars and inanimate objects. For example, the blue flag at the bottom involved a tree "in collision with" a car.

Cycling takes place almost entirely on segregated cycle-paths in this area and these have a very good safety record in comparison with roads. This is most easily observed on the map from the cycle-route along the red line, a very peaceful and popular car free recreational cycle-path which is entirely separated from the road. On that route, where cars are excluded, there have been no collisions at all.

We demonstrate good practice on study tours
as well as showing what not to do
Just because something is "Dutch", that doesn't mean it's good. The Netherlands has many excellent examples, but you have to be very selective about what serves as a model.

Cyclists fare best where their interactions with motor vehicles are limited and controlled. They fare best where infrastructure ensures that minor mistakes do not result in injuries.

Anywhere that we rely upon everyone behaving perfectly but where we do not protect the most vulnerable, there will be injuries.

Good design takes human nature into account and removes the causes of danger from those who are most vulnerable.

More updates on Shared Space and danger
16 April 2014
The BBC recently reported on Blackett Street, a Shared Space in Newcastle in the UK. They say that there is an "area of concern is Blackett Street, where four accidents have taken place in two years". People world-wide need to start to take note of how dangerous Shared Space has proven to be.

December 2014
Poynton in the UK has been the subject of much hype and many claims of improved behaviour and safety. As with other claims about improved safety of Shared Space, which are often made before the real results are known, the claims about Poynton also do not seem to stand up to scrutiny.

In the five years between July 1, 2005 to July 1, 2010, before the Poynton scheme was introduced, there were nine accidents, one of which involved a pedestrian (i.e. 1.8 per year, 0.2 per year involving a pedestrian). Between March 31, 2012 and March 31, 2014, after the Poynton scheme was completed, there have been six accidents, four involving a pedestrian (i.e. 3 per year, 2 per year involving a pedestrian).

This data suggests that the accident rate in Poynton for all users after shared space was introduced is 67% higher than it was before. The accident rate for pedestrians is now 10x greater than before.

Shared Surfaces. A video showing the problems that blind people have with negotiating the Shared Space in Poynton. Note the incident with a frustrated pedestrian at 3:50.

The Shared Space in Ashford, Kent (UK) opened in 2008. According to a local source, the heralded improvement in safety has not been seen. The exact same number of collisions and injuries occured in the four years after the shared space redesign of a street in that town as happened in the four years before. This may sound almost benign, but while there were no serious injuries in the four years leading up to 2008, there were two serious injuries in the four years after 2008. The numbers are small, but they point to the situation becoming worse in Ashford rather than better.

Who was Shared Space ever for ?
A document from 2007 claimed a dramatic improvement in safety on the Laweiplein in the first two years after it was converted to shared space, but as you can see from the figures at the top of this article, the claimed short term benefit does not appear to have continued in the longer term.

It's also interesting to read some of the text in the 2007 document. "Public perceptions of traffic safety have declined since 2000, with around 45% responding “poor” or “bad” to opinions on traffic safety compared to 3 0% in 2000. A significantly higher proportion of elderly people are negative about traffic safety now( 47% compared to 3 8% before)." It is also interesting to note that "Public perceptions of the ability of traffic to flow freely have changed dramatically from 2000 to 2005 . In 2000, 66 % rated congestion “bad”; by 200 this proportion had fallen to 5 %."

It has been my contention for many years now that the better flow of motorized traffic claimed for Shared Space is achieved at the expense of vulnerable road users. Both the convenience and safety of vulnerable road users are compromised by Shared Space. This is what I have observed in every Shared Space that I have visited. It can be observed quite dramatically in two of my videos, of a Shared Space junction in Assen and of Exhibition Road in London.

Where are the advocates ?
I've been writing about the problems with Shared Space since 2008. In the last six years there have been many hand-waving attempts to "prove" safety but no actual statistics have turned up.

What there have been are attempts to move the goal posts. This takes the form of claims that there's some better "shared space" which really works, either some other site than whichever one it was that I last wrote about, or perhaps some future design.

There have also been emotive suggestions that Shared Space is safe, even including "appeals to authority". For example, I was recently informed that Hans Monderman was a genius, that I should respect him for that reason and therefore not disagree. As it happens, I have a lot of respect for the man and I think it very unfortunate that he is no longer here to defend and, where necessary, perhaps improve upon his ideas. However that does not mean any of us should take his work as beyond criticism and we certainly should not assume that other peoples' interpretations of his ideas are beyond criticism.

Actual evidence in favour of the safety of Shared Space, especially for vulnerable road users, seems to be seriously lacking. Available statistics do not support  the proposition that this is a safe way to design streets.

June 2014 update. Steve Melia
Please read Steve Melia's webpage and article about how Shared Space guidance lacks "evidence-based policy".

He says "The key message for transport planners and urban designers concerned about sustainability and the pedestrian experience is that sharing space with traffic is no substitute for traffic removal". I concur. While Shared Space does not achieve these objectives, Nearly Car Free streets do.


André said...

Looking at the site you mention, the most dangerous junction in the Netherlands is not the one you mention, but the traffic light controlled junction between Geldropseweg and ringroad in Eindhoven, found on In the period under consideration, 174 accidents happened here, one of them a car-car collision with 3 injured and 2 death and 12 others with injury.

David Hembrow said...

Andre: If you visit the link that I provided, that has a link through to the source of the claim that this is the most dangerous. The claim made was that there had been 14 injuries in the three years leading up to 2010.

André said...

David: Ok, then the difference between my count and that one is that I took the criterium 'most accidents' and the page you mention 'most accidents with injuries'.

André said...

Under that definition, it's not correct either with these data. They go 3 years further (2007-2012 rather than 2007-2009), and the Boterdiep had much fewer accidents in the second half of that period. In a list I now have it at #12, most dangerous being in Hoorn, where in 20 different accidents 25 people got injured among which 1 was killed.

My list is:
1. Hoorn, Provincialeweg/Lionestraat 20
2. Utrecht, Koningin Wilhelminalaan/Europalaan 19
3. Rotterdam, Vierambachtsstraat/Heemraadstraat 17
4. Schiphol, Kruisweg/Vuursteen 16
5. Amsterdam, IJburglaan/Zuiderzeeweg 15
6. Rotterdam, Claes de Vrieselaan/Nieuwe Binnenweg 15
7. Zeist, Scharewijdelaan/Oude Arnhemseweg 15
8. Rotterdam, Kreekhuizenlaan/Grote Kreek 14
9. Heerhugowaard, Zuidgtangent/Bevelandseweg 14
10. Eindhoven, Geldropseweg/ring road 14
11. Gouda, Nieuwe Gouwe/Hanzeweg 14
12. Groningen, Boterdiep 14
13. Groningen, Hoendiep/Atoomweg 13
14. Heerhugowaard, Zuidtangent/Middenweg 13
15. Rotterdam, Rosestraat/Vuurplaat 13
16. Den Haag, Prinses Beatrixlaan/Guntersteinweg 12
17. Rotterdam, 's Gravendijkwal/1e Middellandstraat 12
18. Rotterdam, Westblaak/Eendrachtsweg 12
19. Rotterdam, Schiekade/Bergweg 12
20. Amsterdam, Insulindeweg/Molukkenstraat 12

Unknown said...

Any data on the severity of accidents? Has the shared space slowed everyone down? ie when accidents occur is the the chance of severe injury or death is reduced?

David Hembrow said...

Jason: Follow the link at the top of the blog post and you can see the data yourself. Numbers in flags say how many at this location. Blue flags = crashes, yellow = at leasts one injury, red = at least one death.

Any of the rest is supposition. When presented next to data for nearby conventional (Dutch conventional, with cycle-paths etc.) junctions, and seeing that those other junctions have fewer crashes and cause fewer injuries to vulnerable road users than do the shared spaces, I can't see an advantage.

Compare the Assen roundabout above, with four fender-benders and no injuries since 2007 with the Drachten roundabout which in the same period of time was the site of ten crashes, involving 4 cyclists and sending one of those to hospital. Assen's a bigger, busier town than Drachten.

Ian Stark said...

Fascinating maps. Have you considered just editing the Wikipedia page for Drachten, giving the more recent Laweiplein numbers? Even if only to indicate that not all sources agree with the conclusion given there?
Any thoughts on why the table in the report Wikipedia references does list a much reduced number of crashes (although with only a couple of years of data following the reconstruction)?

Ian Perry (Cardiff, UK) said...

I'm not sure that you are comparing like with like. Drachten is an urban roundabout giving priority to cyclists.

The Assen example appears to be more rural and cyclists must yield to motorists.

David Hembrow said...

Ian: The Assen example is anything but rural. It's right in the city, in the centre of this map.

The fact is that I've looked at injury figures for many roundabouts in the Netherlands.

Few roundabouts in the Netherlands have caused so many injuries as the "squareabout" in Drachten. Indeed, as I point out above, even the next roundabout along in Drachten, conventionally designed by Dutch standards, has a considerably safer record.

David Hembrow said...

As for the comment about giving way, you've obviously not understood at all. This roundabout is on my route between home and the centre of the city. This video shows how much I have to give way. i.e. not at all.

I've started a serious of posts about good junction design. At the moment, I'm focussing on traffic light junctions, showing the safest traffic light junction design possible for cyclists. A post on safe roundabout design will follow shortly.

Unknown said...

Hello David, I am pleased to see the challenge to shared space in your blog but I'm not sure its relevant to the UK where us poor cyclists share the carriageway with vehicles all the time. I do generally support shared space concept but only if it is designed properly and in the right place. It looks to me like in Assen this is not a shared space as we would define it in the UK as the carriageway is defined and pedestrians areas limited to the pavements. This is a normal road but with flush kerbs. Without knowing the place properly its difficult to judge but for a space to be fully shared the I think it should be designed primarily as a pedestrian space which cars are allowed to pass through.

David Hembrow said...

David: Thank you for your contribution, but sadly yours would simply appear to be yet another example of the "No True Scotsman" argument as applied to shared space. i.e. you're claiming that whatever example is illustrated, it's somehow lacking compared to some other theoretical example.

As it happens, the blog post under which you commented refers to five different shared spaces in the Netherlands, none of which are in Assen.

At other times I've also written about a shared space in Assen and about one in London. The last in particular is clearly shared space "as we would define it in the UK" because it is in the UK. Nevertheless, this also has clearly delineated spaces for pedestrians apart from the road for cars and it shares the problems of bullying by motorists which we observe at all shared spaces.

It's just not good enough to claim that all of these shared spaces, several of which have been the subject of much hype by shared space advocates, suddenly become less valid examples when their weaknesses are exposed.

Shared Space does not create a pleasant space and does not result in a safe environment as its advocates claim. The statistics above show the facts: crashes and injuries are concentrated at shared spaces. What we do know works to reduce this problem are nearly car free streets, pedestrianization and other means of removing motor traffic from where cyclists and pedestrians need to go.

r s thompson said...

......I think it's sufficient to say that exaggerated claims of safety are made for shared space..............

if they are lying about shared space safety is ti sufficient to say that is an exagerated claim?

Dave Nunn said...

Poynton doesn't work because it should have been combined with a bypass to remove non-essential traffic from the village crossroads. It has however improved traffic flow compared to the previous traffic lights.

Anonymous said...

Surely comparing to other roundabouts is less valid than comparing to the same junction itself before Laweiplein was converted? I.e. see if traffic levels are still the same but accidents fell.

Comparing to nearby roundabouts is simply too difficult simply because they have far different conditions.

David Hembrow said...

Anonymous: Your argument quite convincing initially because if traffic levels are the same now as before we can of course gain an impression of how well this new layout compares with the situation beforehand. However in my view this is not a very useful comparison to make. Why ? Because by making only this comparison we can compare only two designs. The first design we are comparing is the old layout. Note that this is the old design which everyone already knew was bad and which everyone wanted to replace. By making your suggested comparison, we compare that known poor design with only one of the many possible ways that the junction could have been redesigned.

Given that the old layout was poor, it's no surprise that the new layout is better. But how much better is it in reality ? Is this as good as it could be ? By making your comparison, we can never know.

The problem is that by comparing only before and after in one location we gain no knowledge at all about whether the new design at the Laweiplein is actually a good design relative to other good new designs. We learn nothing about it compares with best practice elsewhere.

We can't keep rebuilding the Laweiplein so the only way of comparing modern designs is to make comparisons between new designs at different locations. When we do this, we find that while the Laweiplein may be slightly safer than the old design in that location, it falls a very long way short of better designs which have been built elsewhere in comparible locations. We have stats to prove this. It would now be a good idea to change the Laweiplein layout so that it is more similar to proven better designs of junctions elsewhere. perhaps including the safest roundabout design.