Crash prone 'modern roundabouts'

[lordsutch]

The fundamental issue in this thread as I see it is that there is a wealth of scientific studies completed and in-process on the safety impacts of roundabouts, in the United States and elsewhere. I found dozens with a cursory Google Scholar search.

At this point I'd say either you need to do a systematic study of publishable caliber to demonstrate that the existing studies are wrong and that multilane roundabouts are really deathtraps, despite the accumulating mountain of evidence to the contrary produced by transportation research professionals around the world, or this thread has run its course.

And by "systematic study," I don't mean cherry-picked data from one year from a couple of cities. I mean a statistical analysis of a large volume of data, using a proper sample and appropriate control variables.

[Rothman]

The fundamental issue in this thread as I see it is that there is a wealth of scientific studies completed and in-process on the safety impacts of roundabouts, in the United States and elsewhere. I found dozens with a cursory Google Scholar search.

At this point I'd say either you need to do a systematic study of publishable caliber to demonstrate that the existing studies are wrong and that multilane roundabouts are really deathtraps, despite the accumulating mountain of evidence to the contrary produced by transportation research professionals around the world, or this thread has run its course.

And by "systematic study," I don't mean cherry-picked data from one year from a couple of cities. I mean a statistical analysis of a large volume of data, using a proper sample and appropriate control variables.

[DaBigE]

The fundamental issue in this thread as I see it is that there is a wealth of scientific studies completed and in-process on the safety impacts of roundabouts, in the United States and elsewhere. I found dozens with a cursory Google Scholar search.

At this point I'd say either you need to do a systematic study of publishable caliber to demonstrate that the existing studies are wrong and that multilane roundabouts are really deathtraps, despite the accumulating mountain of evidence to the contrary produced by transportation research professionals around the world, or this thread has run its course.

And by "systematic study," I don't mean cherry-picked data from one year from a couple of cities. I mean a statistical analysis of a large volume of data, using a proper sample and appropriate control variables.

I said nearly the same thing several pages/months ago. Interesting to see this thread still going in circles (pun intended).

[tradephoric]

This thread has been focused on 2x2 multi-lane roundabouts.  These 2x2 roundabouts are the ones experiencing high crash rates.  Citing studies skewed towards single-lane and 2x1 multi-lane roundabouts don't reveal the safety performance of 2x2 multi-lane roundabouts.  Find a study that focuses on the safety benefits of 2x2 multi-lane roundabouts and then we can talk.  Of course this point has already been made but some fail to recognize it:

The Insurance Institute for Highway Safety released a report in 2000 evaluating the safety of 24 intersections that were converted to roundabouts.  The report found the following:

The present study evaluated changes in motor vehicle crashes following conversion of 24 intersections from stop sign and traffic signal control to modern roundabouts. The settings, located in 8 states, were a mix of urban, suburban, and rural environments. A before-after study was conducted using the empirical Bayes approach, which accounts for regression to the mean. Overall, the empirical Bayes procedure estimated highly significant reductions of 39 percent for all crash severities combined and 76 percent for all injury crashes. Reductions in the numbers of fatal and incapacitating injury crashes were estimated to be about 90 percent.

http://safety.fhwa.dot.gov/intersection/resources/fhwasa09027/resources/Crash%20Reductions%20Following%20Installation%20of%20Roundabouts.pdf

The study was skewed towards evaluating single-lane roundabouts as only 9 of the 24 intersections evaluated were multi-lane roundabouts.  The 9 multi-lane roundabouts evaluated were all from Colorado and 6 of the 9 serviced interchanges along I-70.  The 3 non-interchange multi-lane roundabouts evaluated in the study were all along Avon Road in Avon, Colorado. When looking at aerials, there appears to be limited traffic generated along the side-street for 2 of the 3 roundabouts (IE. the side-street isn't a major route).  Here is a list of all the roundabouts evaluated in the Insurance Institute study:



In the press release for the proposed multi-lane roundabout at State & Ellsworth, the Washtenaw County Road Commission cited a study with familiar looking safety numbers:

As stated by a 2011 report from the Insurance Institute of Highway Safety, 23 intersections that were converted into roundabouts experienced significant declines in accidents. Total crash frequency fell by approximately 40 percent; injury crash frequency fell by approximately 80 percent; and fatal crash frequency fell by approximately 90 percent.

http://www.wcroads.org/node/529

Citing a study that is skewed towards single-lane roundabouts when a major multi-lane roundabout is being proposed seems disingenuous.  Even the multi-lane roundabouts evaluated in the 2000 study are dissimilar to the roundabout that was being proposed for State & Ellsworth.  In the first year of operation, State & Ellsworth experienced roughly a 10x increase in crashes (not a 40% reduction).  Agencies shouldn't be implying that large multi-lane roundabouts will see a reduction in total crashes (and by citing studies that are skewed towards single-lane roundabouts, that's exactly what they are doing).

[lordsutch]

This thread has been focused on 2x2 multi-lane roundabouts.  These 2x2 roundabouts are the ones experiencing high crash rates.  Citing studies skewed towards single-lane and 2x1 multi-lane roundabouts don't reveal the safety performance of 2x2 multi-lane roundabouts.  Find a study that focuses on the safety benefits of 2x2 multi-lane roundabouts and then we can talk.  Of course this point has already been made but some fail to recognize it:

No, the point is that the burden of proof that 2x2 multi-lane roundabouts are inherently unsafe is on you. You have not demonstrated this using evidence that would be accepted by the transportation research community, using a systematic study (not "I found another 2x2 roundabout with a few crashes").

Not to mention you keep moving the goalposts on this thread. First it was all modern roundabouts were deathtraps (which was demonstrated to be false). Then it was all multi-lane roundabouts are deathtraps, regardless of the lanes on each leg (ditto). Now it's all multi-lane roundabouts that have 2 or more lanes on each entering leg are deathtraps (where it's apparently "Carmel had 2 fatality accidents in roundabouts in 2014, so nyeh!").

[tradephoric]

No, the point is that the burden of proof that 2x2 multi-lane roundabouts are inherently unsafe is on you. You have not demonstrated this using evidence that would be accepted by the transportation research community, using a systematic study (not "I found another 2x2 roundabout with a few crashes").

Take a look at the 2x2 roundabouts that have been included in these published studies and see how they have performed.  A 2011 roundabout study funded by the Wisconsin Department of Transportation was cited previously on this thread.  Of the 24 roundabouts analyzed, there was only one 2x2 multi-lane roundabout (STH 32/ STH 57 Broadway in De Pere).   According to the study, 11 of the roundabouts had increases in crashes.  Here's an excerpt from the study found on page 23:

Among the 11 locations with increased crash record, STH 32/STH 57 Broadway, Canal St/25th Ave, Thompson and Commercial, contribute 38 percent, 15 percent and 18 percent of all increases in the total number of crashes, respectively. All combined, the three locations contribute 71 percent of the crash increases.
http://www.topslab.wisc.edu/programs/safety/projects/roundabouts/WI%20Roundabout%20Evaluation%20Volume%202%20Safety.pdf

The 2x2 multi-lane roundabout analyzed in the Wisconsin study (Sth 32/STH 57 Broadway) accounted for 38% of the crash increases in the study.  Put another way, if this 2x2 roundabout hadn't been included in the study, the Wisconsin safety numbers would have looked a lot better.  Shortly after the release of the study, a circulating lane inside the 32/57 roundabout was removed.  They waved the white flag and converted it to a 2x1 roundabout to address the high crash rate.

Not to mention you keep moving the goalposts on this thread. First it was all modern roundabouts were deathtraps (which was demonstrated to be false). Then it was all multi-lane roundabouts are deathtraps, regardless of the lanes on each leg (ditto). Now it's all multi-lane roundabouts that have 2 or more lanes on each entering leg are deathtraps (where it's apparently "Carmel had 2 fatality accidents in roundabouts in 2014, so nyeh!").

Quote where i said this.  Fair enough? 

[DaBigE]

Among the 11 locations with increased crash record, STH 32/STH 57 Broadway, Canal St/25th Ave, Thompson and Commercial, contribute 38 percent, 15 percent and 18 percent of all increases in the total number of crashes, respectively. All combined, the three locations contribute 71 percent of the crash increases.
http://www.topslab.wisc.edu/programs/safety/projects/roundabouts/WI%20Roundabout%20Evaluation%20Volume%202%20Safety.pdf

The 2x2 multi-lane roundabout analyzed in the Wisconsin study (Sth 32/STH 57 Broadway) accounted for 38% of the crash increases in the study.  Put another way, if this 2x2 roundabout hadn't been included in the study, the Wisconsin safety numbers would have looked a lot better.  Shortly after the release of the study, a circulating lane inside the 32/57 roundabout was removed.  They waved the white flag and converted it to a 2x1 roundabout to address the high crash rate.

To be fair, it wasn't just the crash rate they were addressing. They were also addressing queuing and delays that developed counter to the pre-construction forecasts. It could also be said that the queuing had a part in the increase in crashes...the longer you're waiting, the more risks/smaller gaps you'll try and take.

[tradephoric]

After two months of operation, a double-lane roundabout in Lakeville, Minnesota has tallied up 44 crashes.  Traffic counts at the roundabout are about 30,000 vehicles per day.  This equates to a crash rate of 24.1 MEV.  The crash rate should drop as drivers become more familiar with the roundabout - but even still - this is a horrible initial result.  It will be interesting to track the crashes over time at this roundabout. 


http://sunthisweek.com/2015/11/05/number-of-lakeville-roundabout-accidents-drop/?utm_source=feedburner&utm_medium=feed&utm_campaign=Feed%3A+Thisweeklivecom+%28ThisweekLive.com%29

[Tom958]

...produced by transportation research professionals around the world...

That's what I want to know: is this questionable accident experience inherent to roundabouts in general, or is it an American thing? As I understand it, roundabouts were invented and first built in the UK, then spread to other places after their advantages had been proven through years of experience. They didn't make it over here until the third or fourth wave, at which point the accumulated evidence made it increasingly indefensible not to adopt them. If said evidence was a hoax (!), then we've got a concrete (and asphalt) problem. But if it's a matter of somehow inducing Americans to drive like other people... it's just a software issue. 

[DaBigE]

New story regarding the Lakeville, MN roundabout: http://bringmethenews.com/2015/11/05/lakeville-drivers-are-starting-to-get-used-to-the-roundabout/

The newspaper says following its opening, there were 28 accidents on the roundabout in September, according to Lakeville Police Chief Jeff Long. However, this number dropped to 16 in October.

After it opened, the Minnesota Department of Transportation said most of the accidents occur on roundabouts because people don't yield to other vehicles when they should. And that is still a problem in Lakeville, with local Steve Bakke telling the Sun he sees drivers sometimes accelerate rather than slow down as they approach the roundabout. Nonetheless, he admitted: "Every day it's getting better."

It is the first multi-lane roundabout to open in Dakota County

[tradephoric]

The newspaper says following its opening, there were 28 accidents on the roundabout in September, according to Lakeville Police Chief Jeff Long. However, this number dropped to 16 in October.

If the roundabout averages 16 crashes a month, that equates to 192 crashes per year.  It's premature to write an article with the headline "drivers are starting to get use to the roundabout"  when the roundabout is on track to experience nearly 200 crashes in the first year of operation. 

After it opened, the Minnesota Department of Transportation said most of the accidents occur on roundabouts because people don't yield to other vehicles when they should. And that is still a problem in Lakeville, with local Steve Bakke telling the Sun he sees drivers sometimes accelerate rather than slow down as they approach the roundabout. Nonetheless, he admitted: "Every day it's getting better."

Let's hope it gets better every day.  It needs to get a lot better for multiple days.

It is the first multi-lane roundabout to open in Dakota County

That inaccurate.   There are several multi-lane roundabouts in Dakota County.  In fact, there is a multi-lane roundabout just two miles down the road from the new Lakeville roundabout.

https://www.google.com/maps/@44.6904743,-93.2376639,140m/data=!3m1!1e3
https://www.google.com/maps?ll=44.80465,-93.20761&z=18&t=h
https://www.google.com/maps/@44.8373943,-93.0853549,140m/data=!3m1!1e3

[DaBigE]

The newspaper says following its opening, there were 28 accidents on the roundabout in September, according to Lakeville Police Chief Jeff Long. However, this number dropped to 16 in October.

If the roundabout averages 16 crashes a month, that equates to 192 crashes per year.  It's premature to write an article with the headline "drivers are starting to get use to the roundabout"  when the roundabout is on track to experience nearly 200 crashes in the first year of operation.

It's also premature to make yearly crash predictions based on two months of operation, especially when a decreasing trend is beginning to appear. Given the timeframe the article was referencing, their assumption is accurate.

[tradephoric]

The newspaper says following its opening, there were 28 accidents on the roundabout in September, according to Lakeville Police Chief Jeff Long. However, this number dropped to 16 in October.

If the roundabout averages 16 crashes a month, that equates to 192 crashes per year.  It's premature to write an article with the headline "drivers are starting to get use to the roundabout"  when the roundabout is on track to experience nearly 200 crashes in the first year of operation.

It's also premature to make yearly crash predictions based on two months of operation, especially when a decreasing trend is beginning to appear. Given the timeframe the article was referencing, their assumption is accurate.

Let's just wait to see if monthly crashes drop to 2 per month at this roundabout.  Even then, the roundabout would have a crash rate above 2.0 MEV. 

[Tom958]

...produced by transportation research professionals around the world...

That's what I want to know: is this questionable accident experience inherent to roundabouts in general, or is it an American thing?

I found this study, which compares intersections in the city of Las Vegas, and which breaks the results down by ADT into minor, medium and major intersections. In the major intersections category there are three roundabouts and three signalized intersections, with ADT ranging from 21k to 28k (which really isn't that much, is it?). It says, among other things:

"However, the roundabouts that fall into the major intersections category have more crashes than the signalized intersections. Even though the total number of crashes was more for the roundabouts, the number of injury crashes was in the same range for both roundabouts and intersections with traditional controls."

And, later in the report:

"A higher number of crashes at the roundabouts creates an impression that the signalized intersections are safer compared to roundabouts at the candidate major intersections. This contradicts the findings from Europe that the roundabouts are safer than the signals even when the daily traffic entering the roundabouts is high.." (my italics)

[tradephoric]

Thanks for the link to that Las Vegas study Tom.  Of the 3 roundabouts classified as major, only Town Center Dr. and Hualapai Way would be considered a 2x2 multi-lane roundabout.  Interestingly, this 2x2 roundabout had the worst crash rate and the worst injury crash rate of any intersection analyzed in the study (excluding stop-controlled intersections):



This published study further questions the safety benefits of 2x2 multi-lane roundabouts.  There were significantly more crashes at the Town Center Dr and Hualapai Way roundabout with no reductions in injury crashes when compared to similar ADT signalized intersections in Las Vegas.  To put things into perspective, there were nearly as many injury crashes at the Town Center / Hualapai Way roundabout than total crashes at the Tenaya St / Vegas Dr. signalized intersection (3 injury crashes vs. 4.6 total crashes). 

[jeffandnicole]

Pretty funny when tradeph is very quick to note a high volume of accidents within a few weeks of a roundabout's opening and details how unsafe they are, but very quick to dispute someone's report of a reduction in accidents, noting that more time is needed to see the big picture.

Btw, you are still cherry-picking data.

[tradephoric]

Pretty funny when tradeph is very quick to note a high volume of accidents within a few weeks of a roundabout's opening and details how unsafe they are, but very quick to dispute someone's report of a reduction in accidents, noting that more time is needed to see the big picture.

Btw, you are still cherry-picking data.

If I'm cherry picking data, then there must be plenty of 2x2 modern roundabouts that have low crash rates.  Cite these 2x2 roundabouts with low crash rates or your accusation rings hollow.

[tradephoric]

This would be an eye opening study if it were published:

"The Safety Performance of Roundabouts with Average Daily Traffic exceeding 30,000"

The 30,000 ADT condition would weed out single-lane roundabouts and most 2x1 multi-lane roundabouts.  The focus would be nearly entirely on the safety performance of 2x2 (or higher) roundabouts in America. Ideally, the study would analyze interchange and non-interchange roundabouts separately and come up with two different sets of safety numbers.

[tradephoric]

According to a 2014 article, the OPP have responded to 52 collisions at the Herb Gray Parkway multi-lane roundabout in Windsor, Ontario.  All the accidents were referred to as "minor accidents" .  Does this mean only minor injury accidents occurred or no injury accidents occurred at all?  The article doesn't clarify. 
http://blackburnnews.com/chatham/chatham-news/2014/12/22/two-lane-roundabout-yields-accidents/

In addition, in April 2015 there were two separate tractor trailer rollovers inside the roundabout about a week apart. 
http://www.cbc.ca/news/canada/windsor/all-lanes-open-after-another-roundabout-rollover-at-highway-3-1.3043578

[tradephoric]

The overall crash rate benefits may be "null and void," but these high volume intersections that require a multilane solution are the ones that lead to the vast majority of fatality crashes when they're signalized or have multi-way stop conditions. Without getting into the weeds of cost-benefit analysis,* a few more (or even many more) fender benders and whiplash accidents are massively preferable to a few more deaths from people being T-boned by red-light runners and the like.

* If we take the value of a human life at somewhere between $8-10 million, even several hundred more accidents at $5-10k a pop would be "worth" it.

In a recent study, MDOT evaluated the performance and safety effectiveness of roundabouts in Michigan.  According to the cost-benefit crash analysis included in the study, triple-lane roundabouts saw a negative average benefit of -$122,778 annually.  The double-lane roundabout at 68th Ave & Randall St was the only 2x2 roundabout analyzed in the study; it too had a negative benefit.   


https://www.michigan.gov/documents/mdot/MDOT_Research_Report_RC1566_379286_7.pdf

The safety improvements of roundabouts are mitigated as you get into high capacity double-lane (specifically 2x2) and triple-lane roundabouts.  The triple-lane roundabouts analyzed in the Michigan study saw a 19.9% reduction in injury crashes while total crashes increased by 97.5%.  The reduction in injury crashes wasn't enough to offset the large increases in total crashes in the cost-benefit analysis.

[tradephoric]

Here's a link to a presentation given by the FHWA at the 4th Annual International Conference on Roundabouts.  It is titled "The Mutli-Lane Roundabout  PDO Dilemma" and summarizes a lot of the points made on this thread. 

http://teachamerica.com/RAB14/RAB1410AIsebrands/index.htm

A few comments...
In the first slide, the speaker states that "they [multi-lane roundabouts] are showing great improvement in terms of the injury and fatal crashes".  However, she doesn't cite any evidence to support this and it's up to the listener to take her claim at face value. 

The Bluffton, SC roundabout is highlighted in this presentation and favorable crash statistics are shown.  However, the presenter questions whether crashes will go back up as drivers get use to the new markings.  In 2014, the roundabout experienced 30 total crashes including 7 injury crashes.  This equates to 2.5 crashes per month (much higher than the 0.4/month cited in the presentation).  Unfortunately, crashes have gone back up.

http://www.islandpacket.com/news/local/traffic/article33716778.html

[tradephoric]

This roundabout is going to be a doozy (scheduled for 2017):

http://www.postcrescent.com/story/news/local/2014/07/24/roundabout-coming-richmond-northland/13119107/

Is there any evidence that triple-lane roundabouts have reduced injury crashes in Wisconsin?

[Rothman]

Alright...triple-laners make me twitch.  I'll agree with your concern here, tradey.

[AlexandriaVA]

Egads. That's no roundabout. That's a corn-fed abomination!

[cl94]

Alright...triple-laners make me twitch.  I'll agree with your concern here, tradey.

Um...yeah. There are very few places where a triple-lane roundabout makes sense. Basically only where it replaces a triple-lane traffic circle.