Crash prone 'modern roundabouts'

[tradephoric]

After reading through this thread from start to finish, I will agree that there does seem to be discrepancies from what the engineers say to what reality says. But, I'm not sure what the fix is. I'm still convinced that roundabouts get safer with time (which is hard to measure though, since AADT numbers almost always rise with time, so naturally the number of collisions will rise), but do you think we are building too many roundabouts? Would you prefer more signals? Honest question, no premeditation here. I just want to hear what you think the solution to this problem is.

I don't believe America is building too many roundabouts.  The majority of single-lane and 2x1 multi-lane roundabouts function well and don't have high crash rates.  What has happened is more complex multi-lane (2x2) roundabouts have been constructed in recent years, and these are the roundabouts that are seeing high crash rates.  The good thing is these roundabouts are being monitored closely.  Agencies are addressing these high crash rate roundabouts by eliminating circulating lanes and experimenting with signage/pavement marking changes (many examples of this have already been mentioned on this thread).   

[jakeroot]

Drivers are having trouble navigating the higher capacity multi-lane roundabouts in Carmel.   The roundabout at 116th & Keystone Pkwy averaged 64.3 crashes a year since 2011.  The AADT of 116th Street at Keystone was 20,463 according to Hamilton County's 2011 traffic count data.  This equates to a crash rate of 8.7 which is off the charts (an intersection crash rate over 2.0 is often considered "˜critical' and warrants further investigation).

I'm sorry trade, but I just don't think anyone cares. Carmel has basically reduced their traffic fatalities to zero (so far as I know), and they don't have any traffic congestion. If the trade-off is a slight increase in collisions, by all means, most people will take the increase in collisions. But you'll be hard-pressed to find anyone in Carmel who would prefer a traffic signal, with its backups and t-bone collisions. I mean, of course traffic signals have less collisions. They're really fucking easy to navigate (red means stop, green means go), but that doesn't automatically mean it's the preferred choice for multi-lane intersections, just because there's fewer collisions.

It should be noted that Carmel's population has increased by 62% since 2000. Perhaps traffic engineers take a different approach, but personally, I've always viewed the number of collisions at an intersection as a factor of the amount of cars going through it. It should be well established that there are many gaffe-prone drivers out there, so intersections with more people are likely to have to more of these gaffe-prone drivers going through them at any given moment. Given this, it should be no surprise that the busier intersections have more collisions -- there's just more cars.

I don't believe America is building too many roundabouts.  The majority of single-lane and 2x1 multi-lane roundabouts function well and don't have high crash rates.  What has happened is more complex multi-lane (2x2) roundabouts have been constructed in recent years, and these are the roundabouts that are seeing high crash rates.  The good thing is these roundabouts are being monitored closely.  Agencies are addressing these high crash rate roundabouts by eliminating circulating lanes and experimenting with signage/pavement marking changes (many examples of this have already been mentioned on this thread).

If you take into account my point above, there are more cars in a two or three lane roundabout at any given time than a single lane roundabout. It should be no surprise that there are more collisions -- there's more cars!

[tradephoric]

The crash rate measures the number of crashes at an intersection per million entering vehicles.  Engineers like to see crash rates below 2.0 MEV.  What are the crash rates for the following intersections Jake?

[jakeroot]

The crash rate measures the number of crashes at an intersection per million entering vehicles.  Engineers like to see crash rates below 2.0 MEV.  What are the crash rates for the following intersections Jake?

Fuck if I know. I'm horrible at math. But I do know that the Troy junction does not permit left or U-turns, so comparing the two is like comparing apples to oranges. The bottom intersection is not necessarily a full-fledged four-way junction like the roundabout.

Then again, my theory from my previous post is just that: a theory. I didn't attempt to back it up with evidence. It's just a hunch. I can't imagine it's always true (that is, busier intersections have more collisions).

But, still, I maintain that the crash rate is far less important than the death rate. Very few of your stats ever show fatalities (and if they do, it's usually zero for both before and after the roundabout). At that point, it should come down to A) when possible, avoiding the intersection design which has the likelihood of creating the greatest potential for death, and B) which intersection design has the best vehicle through-put.

Now, I've basically created my own criteria by which I judge safety (which may or may not be a straw-man). But, I do more than likely represent the majority of people, who would take an uptick in collisions if it meant basically eliminating traffic fatalities.

[jakeroot]

Can you please analyze some of Washington's roundabouts? As usual, I have no idea where to find those AADT numbers, nor the crash rates. I know back on the first or second page, you pulled some numbers for some roundabouts in Washingotn, but the numbers were over a decade old, and I'm interested to see whether or not any new developments have occurred since then.

[tradephoric]

The Bluffton roundabout has a crash rate of 3.41 which is 4X the average crash rate of signalized intersections.  This is more than just a "˜slight increase in collisions'.  The injury crash rate at the Bluffton roundabout is 0.54 which is approaching the total crash rate of signalized intersections.  It's not a good result.


https://www.massdot.state.ma.us/highway/Departments/TrafficandSafetyEngineering/CrashData/CrashRates/Intersection.aspx

I'm sorry trade, but I just don't think anyone cares. Carmel has basically reduced their traffic fatalities to zero (so far as I know), and they don't have any traffic congestion. If the trade-off is a slight increase in collisions, by all means, most people will take the increase in collisions. But you'll be hard-pressed to find anyone in Carmel who would prefer a traffic signal, with its backups and t-bone collisions. I mean, of course traffic signals have less collisions. They're really fucking easy to navigate (red means stop, green means go), but that doesn't automatically mean it's the preferred choice for multi-lane intersections, just because there's fewer collisions.

Carmel has seen a drop in injury accidents and city officials attribute the drop in to its roundabout-building initiative.  I'm not surprised as the vast majority of Carmel roundabouts function well.  However, there are a handful of Carmel roundabouts with high crash rates.  These problematic roundabouts are the 2x2 multi-lane ones that have been cited so prevalently throughout this thread.  Should we ignore the Carmel roundabouts with high crash rates just because most of the roundabouts in the city work well?    There are lessons to be learned from these high crash rate roundabouts.

The generic roundabout safety statistics should maybe read like this...

-40% reduction in total crashes....
-80% reduction in injury crashes...
-90% reduction in fatal crashes.....
 
*except for 2x2 multi-lane roundabouts.  These will see a significant increase in total crashes and the reduction in injury crashes will be questionable at best.  Signage and pavement marking tweaks will be made to the roundabout in a vain attempt to reduce the high crash rate.  Engineers, perplexed with why there are so many crashes, will eventually say "screw it"  and eliminate circulating lanes inside the roundabout as a last ditch effort to reduce the high crash rate.

[tradephoric]

Can you please analyze some of Washington's roundabouts? As usual, I have no idea where to find those AADT numbers, nor the crash rates. I know back on the first or second page, you pulled some numbers for some roundabouts in Washingotn, but the numbers were over a decade old, and I'm interested to see whether or not any new developments have occurred since then.

Here's one that hasn't been discussed yet.  In 2002, the City of Bellingham constructed a roundabout at the intersection of Cordata Parkway & Kellogg Road near Whatcom Community College.  In 2008, after 6 years of operation, the roundabout had the highest accident rate in Bellingham with 17 total crashes.  In 2009 there were 18 crashes.  According to the Whatcom County traffic counts, the roundabout has an AADT of 15,200.  This equates to a crash rate of 3.15.

http://wcog.org/wp-content/uploads/documents/2012TrafficCountManual.pdf
http://www.whatcomhorizon.com/2010/04/roundabout-confusion/

[ScottRAB]

The one at US-101/OR-202 in Astoria has, over the long term, resulted in a 150% increase in crashes, without any reduction in accident severity

Even worse is the one at MLK Pkwy, Pioneer Parkway, and Hayden Bridge Way in Springfield.  It cost almost $10 million to build, and it had almost 200 accidents between 2009 and 2013, with volumes only running 15,000-20,000 ADT. 

Can you provide us a reference for your data?  One we can independently verify?

[tradephoric]

The one at US-101/OR-202 in Astoria has, over the long term, resulted in a 150% increase in crashes, without any reduction in accident severity

Even worse is the one at MLK Pkwy, Pioneer Parkway, and Hayden Bridge Way in Springfield.  It cost almost $10 million to build, and it had almost 200 accidents between 2009 and 2013, with volumes only running 15,000-20,000 ADT. 

Can you provide us a reference for your data?  One we can independently verify?

According to a City of Springfield report, the multi-lane roundabout at MLK Parkway & Hayden Bridge Way had a crash rate of 2.91; nearly 3x higher than the next crash rate intersection in the city.  Refer to pages 30-32:


http://www.centrallanertsp.org/sites/default/files/Volume%203,%20Appendix%20B,%2001.09.14-web_0.pdf

Oregon's crash data system can be found in the link below if you want to independently verify the crashes that are occurring. 
https://zigzag.odot.state.or.us/uniquesig08615cf883bed667d26bcec3a7dc5c6b/uniquesig0/SecurezigzagPortalHomePage/

[tradephoric]

Here's a list of multi-lane roundabouts sorted by crash rate.  Most of the roundabouts on this list have already been discussed and the data can be verified digging through the links in this thread.  Rows highlighted in yellow have had circulating lanes removed in an attempt to address the high crash rate.  The average crash rate at signalized intersections is about 0.8 MEV.  The crash rate at these multi-lane roundabouts exceeds the average crash rate at signalized intersections by a big margin.



Not all multi-lane roundabouts have high crash rates.  In the graph below, the 2x1 roundabouts have low crash rates while the 2x2 roundabouts have high crash rates. The exception is the 2x1 roundabout at Diffley & Rahn.  This roundabout has a central island diameter of 80' which is the smallest CID of any of the Minnesota roundabouts analyzed.  It's my belief that if a roundabout is too small, drivers have difficulty judging gaps to safely enter the roundabout.  The small CID may be a plausible reason why the 2x1 roundabout at Diffley & Rahn has a high crash rate.

[tradephoric]

The one at US-101/OR-202 in Astoria has, over the long term, resulted in a 150% increase in crashes, without any reduction in accident severity

Even worse is the one at MLK Pkwy, Pioneer Parkway, and Hayden Bridge Way in Springfield.  It cost almost $10 million to build, and it had almost 200 accidents between 2009 and 2013, with volumes only running 15,000-20,000 ADT. 

Can you provide us a reference for your data?  One we can independently verify?

ScottRAB, what's your take on these 2x2 multi-lane roundabouts?  The generic safety stats say there will be a 40% drop in total crashes but that's not what we are seeing with these multi-lane roundabouts.  More concerning, several of the 2x2 multi-lane roundabouts have seen injury accidents increase. 

[Tarkus]

The one at US-101/OR-202 in Astoria has, over the long term, resulted in a 150% increase in crashes, without any reduction in accident severity

Even worse is the one at MLK Pkwy, Pioneer Parkway, and Hayden Bridge Way in Springfield.  It cost almost $10 million to build, and it had almost 200 accidents between 2009 and 2013, with volumes only running 15,000-20,000 ADT. 

Can you provide us a reference for your data?  One we can independently verify?

These are taken directly from ODOT's crash reporting system.  Plugging in the same dates, the data is replicable.

Here's all the source data from US-101/OR-202 Astoria.  The Pre-Roundabout Period runs from 10/1/1993 to 09/30/2002, while the Post-Roundabout Period runs 10/1/2002 to 09/30/2011.

Astoria Pre-Roundabout KABCO
Astoria Pre-Roundabout Comprehensive
Astoria Pre-Roundabout Truck Included
Astoria Post-Roundabout KABCO
Astoria Post-Roundabout Comprehensive
Astoria Post-Roundabout Truck Included

The data showing over 180 crashes between 2009 and 2013 at the Springfield Pioneer/MLK/Hayden Bridge roundabout is below.  Because there's multiple road names involved, ODOT has the data divided by approach, requiring one to add everything up to get the full picture.

Pioneer Pkwy W and Hayden Bridge KABCO
Pioneer Pkwy E and Hayden Bridge KABCO
MLK and Hayden Bridge KABCO

Neither Springfield nor Lane County has good volume data on the intersection beyond the earlier Springfield TSP figures.  The most recent Springfield data is from 2008 and only shows the approaches.

[tradephoric]

Some 2x2 multi-lane roundabouts were constructed along Shiloh Rd in Billings, Montana in 2010.  According to the Billings Police Department, the roundabout at Shiloh & Grand is currently the 8th most dangerous intersection in the city and the roundabout at Shiloh & King is the 4th most dangerous intersection in the city. 

http://billingsgazette.com/news/local/crime-and-courts/crime-stats-billings-most-dangerous-intersections/collection_97fee293-e4d2-5dac-8854-ceb50a22f030.html

[tradephoric]

Here's an interesting case study.  Pre-2008 Cony Circle in Augusta, Maine averaged over 100 crashes per year and was the highest crash rate intersection in the state.  In 2008, the circle was reconfigured with modern roundabout standards.  After the conversion accidents dropped by 50%.  That may sound impressive, but as of 2014 the roundabout has a crash rate of 11.87 and is still the 3rd highest crash location in Maine.  Is that really a victory?  It's similar to an 800 pound person losing 50% of their body weight.  They are still morbidly obese. 

http://www.centralmaine.com/2011/08/11/cony-circle-accidents-down_2011-08-10/
http://www.wgme.com/images/DANGEROUS%20ROADWAY%20INTERSECTIONS.pdf

Cony Circle:  Pre-2008


Cony Circle:  Post-2008

[jakeroot]

Some 2x2 multi-lane roundabouts were constructed along Shiloh Rd in Billings, Montana in 2010.  According to the Billings Police Department, the roundabout at Shiloh & Grand is currently the 8th most dangerous intersection in the city and the roundabout at Shiloh & King is the 4th most dangerous intersection in the city. 

http://billingsgazette.com/news/local/crime-and-courts/crime-stats-billings-most-dangerous-intersections/collection_97fee293-e4d2-5dac-8854-ceb50a22f030.html

Any idea why the most dangerous intersections are signals?

Here's an interesting case study.  Pre-2008 Cony Circle in Augusta, Maine averaged over 100 crashes per year and was the highest crash rate intersection in the state.  In 2008, the circle was reconfigured with modern roundabout standards.  After the conversion accidents dropped by 50%.  That may sound impressive, but as of 2014 the roundabout has a crash rate of 11.87 and is still the 3rd highest crash location in Maine.  Is that really a victory?  It's similar to an 800 pound person losing 50% of their body weight.  They are still morbidly obese.

I'm not sure there are any other effective ways to intersect five roads that wouldn't cause outrageous congestion.

What are the top two intersections?

Here's one that hasn't been discussed yet.  In 2002, the City of Bellingham constructed a roundabout at the intersection of Cordata Parkway & Kellogg Road near Whatcom Community College.  In 2008, after 6 years of operation, the roundabout had the highest accident rate in Bellingham with 17 total crashes.  In 2009 there were 18 crashes.  According to the Whatcom County traffic counts, the roundabout has an AADT of 15,200.  This equates to a crash rate of 3.15.
http://www.whatcomhorizon.com/2010/04/roundabout-confusion/

This got me thinking:

Is the crash rate an effective means of studying intersection safety? By your measure, the roundabout featured here would be considered dangerous, but that's only because of the "3.15" number. In real life, 17 or 18 collisions just isn't that many (that's one every three weeks, give or take, and there's no telling how severe the collisions were). An intersection with an AADT of 50,000 with a crash rate of 3.15 would have fewer collisions per number of cars entering, but still way more collisions than 17 or 18 a year.

Now, if that's okay simply because more cars = more collisions, that's been my point all along. Multi-lane roundabouts are, obviously, capable of handling far more cars than single-lane roundabouts, simply because there are more circulating lanes, thus more cars in the intersection at any given point capable of whacking into each other. Considering that, I propose that three lane roundabouts are no more dangerous than two or one lane roundabouts. They just handle an enormous amount of cars relative to the smaller roundabouts, so they appear to be much more dangerous. If a traffic engineers' only method of making an intersection safer is by reducing the theoretical through-put of an intersection, simply to lessen the chances of a collision by decreasing the amount of cars in the intersection that could theoretically collide, you're not solving any problems. That's just a trade-off.

Now, I could be looking at this all wrong. I am, after all, not a traffic engineer, nor an expert in analyzing data.

As well (perhaps unrelated), please don't compare roundabouts with Michigan Lefts, as you have before. I love Michigan Lefts just as much as the next person, but the ROW required to install them is usually far more than most agencies can provide. I don't find see them in the same category as roundabouts or traditional signals (they're more like freeways -- a series of intersections working in tandem to filter traffic along a central artery).

[Tarkus]

I'm not sure there are any other effective ways to intersect five roads that wouldn't cause outrageous congestion.

I'll be upfront and say I'm typically not a fan of roundabouts, for various reasons, but one of the places they do make sense is situations with unusual geometry, that would be messy with a more "conventional" treatment.  There's one caveat to the particular Maine intersection being mentioned, however, that I'll note later.

Here's one that hasn't been discussed yet.  In 2002, the City of Bellingham constructed a roundabout at the intersection of Cordata Parkway & Kellogg Road near Whatcom Community College.  In 2008, after 6 years of operation, the roundabout had the highest accident rate in Bellingham with 17 total crashes.  In 2009 there were 18 crashes.  According to the Whatcom County traffic counts, the roundabout has an AADT of 15,200.  This equates to a crash rate of 3.15.
http://www.whatcomhorizon.com/2010/04/roundabout-confusion/

Now, if that's okay simply because more cars = more collisions, that's been my point all along. Multi-lane roundabouts are, obviously, capable of handling far more cars than single-lane roundabouts, simply because there are more circulating lanes, thus more cars in the intersection at any given point capable of whacking into each other. Considering that, I propose that three lane roundabouts are no more dangerous than two or one lane roundabouts. They just handle an enormous amount of cars relative to the smaller roundabouts, so they appear to be much more dangerous. If a traffic engineers' only method of making an intersection safer is by reducing the theoretical through-put of an intersection, simply to lessen the chances of a collision by decreasing the amount of cars in the intersection that could theoretically collide, you're not solving any problems. That's just a trade-off.

Now, I could be looking at this all wrong. I am, after all, not a traffic engineer, nor an expert in analyzing data.

The MEV abbreviation that tradephoric and others have used in this thread stands for "per million entering vehicles".  In short It's actually a way of getting a crash rate that is not dependent on the actual traffic volume, and allows engineers to better compare intersections.  An intersection with a crash rate of 3.15 per MEV need not necessarily have a higher volume of traffic than one with a rate of 0.90 per MEV, but it does mean statistically that the 3.15 per MEV intersection is going to be a bit more than three times as crash prone.

All these high MEV ratings for multi-lane roundabouts are essentially showing that once you get beyond 1x1 or 2x1 roundabouts, into roundabout treatments for higher-volume situations, the safety benefits quickly become null and void.  Cost also becomes a factor--the MLK/Hayden Bridge roundabout in Springfield cost over $9 million to construct.

[jakeroot]

The MEV abbreviation that tradephoric and others have used in this thread stands for "per million entering vehicles".  In short It's actually a way of getting a crash rate that is not dependent on the actual traffic volume, and allows engineers to better compare intersections.  An intersection with a crash rate of 3.15 per MEV need not necessarily have a higher volume of traffic than one with a rate of 0.90 per MEV, but it does mean statistically that the 3.15 per MEV intersection is going to be a bit more than three times as crash prone.

Thank you for clarifying that. I had a feeling I was missing a piece of the puzzle.

[lordsutch]

All these high MEV ratings for multi-lane roundabouts are essentially showing that once you get beyond 1x1 or 2x1 roundabouts, into roundabout treatments for higher-volume situations, the safety benefits quickly become null and void.  Cost also becomes a factor--the MLK/Hayden Bridge roundabout in Springfield cost over $9 million to construct.

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.

[Tarkus]

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.

In theory, yes--signalized intersections aren't "childproofed" with all those rounded edges, allowing the potential for those types of angle crashes--though theory and the actual real-world results don't always correlate, and that's one of the main arguments being discussed in this thread.  Many of the cases cited in this thread, where multi-lane roundabouts have replaced signals, have shown increases in injury accidents as well as total accidents, even over the longer term.  Curiously, I've yet to see a comparison involving a fatality on either side (pre- or post-multi-lane roundabout), however.

[jakeroot]

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.

In theory, yes--signalized intersections aren't "childproofed" with all those rounded edges, allowing the potential for those types of angle crashes--though theory and the actual real-world results don't always correlate, and that's one of the main arguments being discussed in this thread.  Many of the cases cited in this thread, where multi-lane roundabouts have replaced signals, have shown increases in injury accidents as well as total accidents, even over the longer term.  Curiously, I've yet to see a comparison involving a fatality on either side (pre- or post-multi-lane roundabout), however.

It's a matter of high-speed impacts vs low-speed impacts. There can be injuries in both, of course, but it's more like whiplash vs broken back, or bump on the head vs broken legs. Unfortunately for us, both injuries get lumped together, even though 100% of people would prefer the former two over the latter two.

Carmel, Indiana, according to their mayor Jim Brainard, does not have a full Jaws of Life in their city because they don't really have severe collisions anymore (this was on the last page):

http://wishtv.com/2015/10/08/roundabouts-may-confuse-but-stats-back-up-their-safety/

And even if you make a mistake, because speeds are so slow, most wrecks don't involve serious injuries....[Carmel Mayor Jim Brainard] points out that because accidents and injuries are down in Carmel, insurance rates have dropped. He says the city doesn't have a full "jaws of life"  crew in its fire department anymore because they don't have the high speed impacts that they used to have and don't need them very often.

The issue that tradephoric likes to bring up is negligence. He states repeatedly that cities can be found negligent if they knowingly do nothing about intersections that have high crash rates. I would agree that there could be a negligence issue. The problem is that A) no one political body has power over a city's engineering department, thus there's no one that could single-handily bar any further roundabout construction, or more importantly B) while a citizen could privately sue the city for negligence after receiving a minor injury in the roundabout, any city which has invested millions of dollars into roundabouts would go to huge lengths to protect their investments in court, by providing evidence to a judge and jury, demonstrating that while there may be more collisions, the roundabouts are actually safer (in terms of both fatalities and severe injuries).

[tradephoric]

It's a matter of high-speed impacts vs low-speed impacts. There can be injuries in both, of course, but it's more like whiplash vs broken back, or bump on the head vs broken legs. Unfortunately for us, both injuries get lumped together, even though 100% of people would prefer the former two over the latter two.

There is a rating system known as FARS (Fatality Analysis Reporting System) that is used to code injury accidents by severity.  Here's a breakdown of the rating system:

0 No Injury (O)
1 Possible Injury (C)
2 Non-incapacitating Evident Injury (B)
3 Incapacitating Injury (A)
4 Fatal Injury (K)
5 Injured, Severity Unknown
6 Died Prior to Accident*
9 Unknown

Code 1 (Possible Injury). A possible injury is any injury reported or claimed which is not a fatal injury, incapacitating injury or non-incapacitating evident injury. This includes: momentary unconsciousness, claim of injuries not evident, limping, complaint of pain, nausea and hysteria.

Code 2 (Non-incapacitating Evident Injury). A non-incapacitating evident injury is any injury, other than a fatal injury or an incapacitating injury, which is evident to observers at the scene of the accident in which the injury occurred. This includes: lump on head, abrasions, bruises and minor lacerations. This does not include limping (the injury cannot be seen). (See code "1" ).

Code 3 (Incapacitating Injury). An incapacitating injury is any injury, other than a fatal injury, which prevents the injured person from walking, driving or normally continuing the activities the person was capable of performing before the injury occurred. This includes: severe lacerations, broken or distorted limbs, skull or chest injuries, abdominal injuries, unconsciousness at or when taken from the accident scene, and unable to leave the accident scene without assistance. This does not include momentary unconsciousness. (See code "1" ).

Code 4 (Fatal Injury), must only be used if the death occurred within thirty 24-hour time periods from the time of the accident. Every effort should be made to determine that the Death Date was within thirty 24-hour time periods from the Accident Time.

[jakeroot]

It's a matter of high-speed impacts vs low-speed impacts. There can be injuries in both, of course, but it's more like whiplash vs broken back, or bump on the head vs broken legs. Unfortunately for us, both injuries get lumped together, even though 100% of people would prefer the former two over the latter two.

There is a rating system known as FARS (Fatality Analysis Reporting System) that is used to code injury accidents by severity.  Here's a breakdown of the rating system:

Thank you for posting that. I figured there might be something, but I hadn't heard of anything until now. How often is crash data compiled with FARS' numbers?

[tradephoric]

Thank you for posting that. I figured there might be something, but I hadn't heard of anything until now. How often is crash data compiled with FARS' numbers?

In Michigan, the FARS code is compiled for every reported traffic crash in the state.  Law enforcement agencies are required to complete a UD-10 which includes the FARS code.  Below is a link to a sample UD-10 which shows injury severity ranging from no injury (Unit 1 driver) to A-level injury (Unit 1 passenger):

http://semcog.org/Data-and-Maps/Crash-and-Road-Data/Crash_Id/8338533/view/UD10CrashReport

[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.

Compare the 2014 traffic fatalities that occurred in Sterling Heights, MI and Carmel, IN.  Sterling Heights has a population of 130,410 and Carmel has a population of 85,927.  Both communities experienced 2 fatal crashes in 2014 (based on google searches and UD-10 forms).  Ironically, the two fatal crashes that occurred in Carmel happened inside roundabouts.  Neither fatality in Sterling Heights occurred at an intersection (according to the UD-10 forms).  Keep in mind, Sterling Heights has a population nearly 45k higher than Carmel.   I thought roundabouts reduce fatal crashes by 90%.  How then did Sterling Heights, a city full of traffic signals, have zero fatal intersection crashes while Carmel, a city full of roundabouts,  have 2 fatal roundabout crashes?

Fatal crashes in Carmel:
http://fox59.com/2014/08/06/one-dead-in-carmel-accident-passenger-taken-to-hospital/
http://fox59.com/2014/09/26/person-dead-after-motorcycle-accident-in-carmel/

Fatal crashes in Sterling Heights:
http://semcog.org/Data-and-Maps/Crash-and-Road-Data/Crash_Id/8939474/view/UD10CrashReport
http://semcog.org/Data-and-Maps/Crash-and-Road-Data/Crash_Id/8986238/view/UD10CrashReport

[jakeroot]

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.

Compare the 2014 traffic fatalities that occurred in Sterling Heights, MI and Carmel, IN.  Sterling Heights has a population of 130,410 and Carmel has a population of 85,927.  Both communities experienced 2 fatal crashes in 2014 (based on google searches and UD-10 forms).  Ironically, the two fatal crashes that occurred in Carmel happened inside roundabouts.  Neither fatality in Sterling Heights occurred at an intersection (according to the UD-10 forms).  Keep in mind, Sterling Heights has a population nearly 45k higher than Carmel.   I thought roundabouts reduce fatal crashes by 90%.  How then did Sterling Heights, a city full of traffic signals, have zero fatal intersection crashes while Carmel, a city full of roundabouts,  have 2 fatal roundabout crashes?

Fatal crashes in Carmel:
http://fox59.com/2014/08/06/one-dead-in-carmel-accident-passenger-taken-to-hospital/
http://fox59.com/2014/09/26/person-dead-after-motorcycle-accident-in-carmel/

Fatal crashes in Sterling Heights:
http://semcog.org/Data-and-Maps/Crash-and-Road-Data/Crash_Id/8939474/view/UD10CrashReport
http://semcog.org/Data-and-Maps/Crash-and-Road-Data/Crash_Id/8986238/view/UD10CrashReport

Because stats vary between cities? Aren't you a traffic engineer? Anomalies are perfectly common. This is obviously the 10% (though, keep reading...)

The whole point of this thread was to compile data for hundreds of roundabouts to prove a point that multi-lane roundabouts have more collisions. You have done a great job illustrating this, and I am definitely a believer at this point. There is no doubt that multi-lane roundabouts have more collisions (at this point, it's just a matter of whether or not that's a big deal).

You do lose me, however, when you attempt to prove that roundabouts are deadlier than signals by grabbing random data from two random cities, over random time frames. You're one step away from generalizing roundabouts as being deadlier than signals. And that's obviously nuts...right? It doesn't take a genius to figure out that roundabouts, just by the very nature of their design, are safer than signals (taking into account FARS data).

Intersections should be designed to minimize risk. Stats show that, in almost all cases, roundabouts may have more collisions, but are far safer for the driver. There are going to be isolated incidents of drivers not paying attention to the circle ahead, and lose control. It's just a fact. The place where roundabouts succeed is by reducing the potential for dangerous two-party collisions like you might see at signals. Drivers can lose control anywhere along a road (including at signals) and kill themselves. That shouldn't be a strike against the roundabout.