(Please visit one of our advertisers)

No donations or subscriptions are required

   OR   
   
Subscription choices:
Board Karma = 40  (3452 positive of 3834 votes is 40 %pts higher than a neutral 50%)
All Things (Safety Oriented) Motorcycle   
Username:
Password:
Save Password
Forgot your Password?

You can the entire collection of Safety Tip articles in a 33 Megabyte PDF Portfolio

 All Forums
 Motorcycle Safety
 General Discussion
 Fatal Design Revisited
Member Previous Topic Discussion Topic Next Topic  

DataDan
Advanced Member
567 Posts
[Mentor]


Central Coast, CA
USA

Yamaha

FJR1300

Posted - 10/01/2018 :  2:38 PM                       Like
The title of this thread is an homage to Wendy Moon and her 2004 Motorcycle Consumer News article, "Fatal Design", which led me to explore this subject. But the work here--research, conclusions, graphics, words--is mine alone.



Summary

From 1990 to 2016, involvement of light trucks--pickups, SUVS, vans, minivans, crossovers--in 2-vehicle motorcycle crashes increased from 19% to 37%. At the same time, involvement of cars--sedans, coupes, convertibles, station wagons--fell from 75% to 56%. This is bad news for motorcyclists, because a crash with a light truck is 75% more likely to be deadly for a rider than a crash with a car. If light truck involvement had remained at the 1990 level, 250 fewer motorcyclists would have lost their lives in 2016.


Introduction

As I have mentioned in other posts recently (here, here, here), motorcycle crash lethality--the likelihood of rider death in the event of a crash--has gradually increased over the past 30 years. One reason is the aging motorcyclist population. Older riders are more vulnerable to serious and fatal injuries than younger riders. However, growth of the older age groups and their greater lethality isn't enough to account for the observed increase.

A tempting explanation for the remainder is repeal of helmet laws in 6 states, including 4 of the top-10 motorcycling states in the US. The greater lethality of unhelmeted crashes is well established, and decreasing helmet use would increase overall motorcycle crash lethality. But, surprisingly, the wave of helmet law repeals has not diminished the percentage of helmeted riders in crashes. Unhelmeted crashes have jumped in the short term a few times since the wave began in 1997, but over those 20 years, helmet use is up slightly. So helmet law repeal seems unable to explain increased lethality.

Another possibility is the growing popularity of vehicles that are more likely to be deadly in the event of a crash with a motorcycle--specifically, light trucks--and that is the subject of this thread. These are pickups, SUVs, vans, minivans, and crossovers, which now comprise half of the US passenger vehicle fleet. The late Wendy Moon brought this issue to the attention of motorcyclists in a July 2004 Motorcycle Consumer News article, "Fatal Design". Since then, I have tracked the relevant data, utilizing new resources as my knowledge of crash data expanded. I conclude that Moon was basically right: A significant portion of the motorcycle fatality increase over the past 30 years is due to deadlier vehicles on the road.


Since 1990, the basic family transportation appliance has morphed from a Taurus or Accord into an F-150, Odyssey, or Suburban. And the trend will surely continue. A few months ago, Ford Motor Company announced that by 2020, 90% of its North American sales will be trucks, SUVs and crossovers.

The graph below shows that, since 1990, light trucks have increased from less than one-third of registered passenger vehicles in the US to about half. (See footnote for an explanation of the discontinuity.)




Of course, as their presence on the road increased, so did their involvement in motorcycle crashes:



While light trucks collide with motorcycles less frequently than cars, the collisions more frequently result in motorcyclist death. That is, the red line remains below the blue line, but the red bars overtook the blue bars in 2004.


In my next post, I'll look at the proportions of crashes and deaths for each vehicle type and at crash lethality--motorcyclist deaths as a percentage of crashes by vehicle type.


Data for the registration graph is from NHTSA's Traffic Safety Facts Annual Report web page, Tables 7 and 8. The discontinuity at 2011 is explained on page 7 of the PDF of TSF 2016. This problem makes car/light truck registration and VMT data unsuitable for comparisons before and after 2011. In this report, I use registrations only to show the general trend over the past 30 years.

Recent data for the 2-vehicle crash and fatality graph can be found on the TSF web page, Table 30, but it goes back only to 2010. For the graph above, I used the GES and FARS databases. Here, and throughout this report I plot a 3-year rolling average of raw data to smooth out year-to-year variation. Thus, the value of 46,967 plotted for 1990 car crashes is the average of GES database counts 1988-1990.

James R. Davis
Male Administrator
17328 Posts
[Mentor]


Houston, TX
USA

Honda

GoldWing 1500

Posted - 10/01/2018 :  6:46 PM Follow poster on Twitter  Join poster on Facebook as Friend  
Really excited to see what comes next in your thread.

Since Wendy Moon and I engaged in some joint research together over the years, and because her writings, though admittedly often anti-MSF in nature, were extraordinarily useful and informative, I have placed five years of those writings on this site (with her approval).

Readers can find those writings here.

Wendy died in January of 2011 of a heart attack.

Three years later I was an expert witness in a case against the MSF. I was confronted outside the courtroom by Dr. Ray Ochs, A Director of the MSF, before I was called in to testify.

Ochs said: "So we're up against you today, huh?"

I responded: "It could have been worse, you might have been up against Wendy Moon."

Ochs nastily quipped: "Birds of a feather."

Best compliment I can remember having ever received.

She is missed.
Go to Top of Page

scottrnelson
Advanced Member
6915 Posts
[Mentor]


Meridian, ID
USA

Honda

XR650L, 1090 Adv R

Posted - 10/01/2018 :  7:34 PM
Thanks for posting this.

A few months ago I moved to the land of a lot more big trucks - you call them light trucks, but a jacked up 4-door F-150 looks pretty big to me. And there are a lot more trucks like that in Idaho than there were in California. I'll have to keep in mind that they're a bigger danger to me when I'm riding a motorcycle.
Go to Top of Page

DataDan
Advanced Member
567 Posts
[Mentor]


Central Coast, CA
USA

Yamaha

FJR1300

Posted - 10/02/2018 :  10:29 AM
Comparing Crash Lethality

The key point of my first post is: With the growing popularity of light trucks as passenger vehicles, they now equal cars in registrations, and their involvement in 2-vehicle motorcycle crashes increased similarly.

Here again is the second graph from that post:



Some features to note:

1. Car involvement in 2-vehicle motorcycle crashes has dropped by one-third since 1990, from 47,000 per year to 34,000 (though with a much larger intervening drop in the late 1990s).

2. Light truck involvement has doubled since 1990, from 11,000 to 22,000.

3. The combined number of crashes is down slightly, from 59,000 to 55,000. This comes in spite of the growth of motorcycling over the same period, and I would attribute it to improving rider safety (see the thread The Remarkable Decline in Motorcycle Crash Rate).

4. The combined number of deaths is up by 50%, from 1400 to 2100.

5. Light trucks now account for more motorcyclist deaths in 2-vehicle crashes than cars.


The relationship between car and light truck involvement can be seen in the next graph. The line represents the division of 2-vehicle crashes between cars and light trucks. The shaded areas represent the proportion of fatalities in those crashes.



Light truck involvement in 2-vehicle motorcycle crashes has increased from 20% of the combined car/light truck total to 40%, yet they now account for more than half of the deaths in these crashes.

The discrepancy between crash involvement and fatalities is the second key point of this thread: A crash with a light truck is much more likely to kill a motorcyclist than a crash with a car.

I call this "lethality", the number of motorcyclist deaths with another type of vehicle as a percentage of crashes with that type of vehicle. The next graph shows lethality of 2-vehicle motorcycle crashes with cars and light trucks.



In 2016, the lethality of light truck crashes was 5.1% compared to 3.1% for cars--two-thirds higher. One hundred motorcycle-car crashes would kill 3 riders, but one hundred motorcycle-light truck crashes would kill 5.

One interesting feature of this relationship is that it has closed somewhat over 25 years. In the first few years shown on the graph, light trucks were more than twice as deadly as cars, but in recent years they are down to 1.7 times as deadly. My speculation is that the growing number of crossovers, technically light trucks but smaller than the Land Crusher/Executioner class of SUV, tends to reduce overall lethality.

Another interesting feature is that both have increased somewhat over 25 years. This seems to reflect the overall increase in motorcycle crash lethality, which I attribute mainly to the aging population (see the thread Vulnerability of older riders in motorcycle crashes).

The overall lethality increase affects both single- and multiple-vehicle motorcycle crashes. But a comparison will show that multiple-vehicle crash lethality has increased more than single-vehicle lethality, which I claim is due to the growing proportion of light trucks. That's the subject of my next post.


Graphs here are derived from the same data as my first post, NHTSA's GES and FARS databases, and are 3-year moving averages.

Edited by - DataDan on 10/02/2018 10:54 AM
Go to Top of Page

DataDan
Advanced Member
567 Posts
[Mentor]


Central Coast, CA
USA

Yamaha

FJR1300

Posted - 10/08/2018 :  9:00 AM
The Effect of Light Trucks on Multiple-Vehicle Crashes

So far in this thread, I have attempted to show that: 1) As the popularity of light trucks grew over the past 25 years, they have been increasingly involved in motorcycle crashes; and 2) A crash between a motorcycle and a light truck is more likely to be deadly than a motorcycle-car crash. But have light trucks indeed contributed substantially to the overall motorcycle crash lethality increase?



The graph above shows the same plots as the last graph of my second post, the lethality of motorcycle-car crashes and of motorcycle-light truck crashes (dashed lines). In addition, I have plotted overall motorcycle crash lethality (the gray shaded area) and broken it down into single-vehicle (blue line) and multiple-vehicle crash lethality (red line).

As I said in the introduction to this thread, motorcycle crash lethality has increased over the past 25 years in part because of the aging rider population. Involvement of riders age 55+ in crashes has climbed from 3.1% to 21% of the total, while < 25 involvement has dropped from 47% to 21%. But crash lethality for the 55+ age group is much higher than for the < 25s. In effect, 26% of the motorcycling population shifted from the lowest lethality age group to higher ones, most into the highest. Inevitably, this increased overall motorcycle crash lethality.

Increased lethality due to the aging riding population presents a problem in estimating the lethality effect of light trucks in the vehicle mix. Both effects are in the same direction, but how much for each? Breaking down overall lethality into single- and multiple-vehicle crash components helps delineate the two effects. Single-vehicle crashes, affected by rider age but not by another vehicle, increased in lethality from 3.78% to 4.2% between 1990 and 2016, or about 13%. On the other hand, multiple-vehicle crashes, affected by both rider age and other vehicles, increased from 2.8% to 4.6%, or about 66%. Filtering out the age effect from the multiple-vehicle lethality increase, yields a rough estimate of lethality increase due to the changing vehicle mix of 46%. I am not confident in the accuracy of this estimate, dependent as it is on other estimates and on the start and end points. However, I am confident that it reflects a true relationship between vehicle mix and motorcycle crash lethality.

With that, I can state the third main point of this thread: The increased involvement of light trucks has substantially increased the lethality of multiple-vehicle motorcycle crashes.

To recap recent history shown in the graph above:
  • In 1990, riders age 55+ were only 3% of motorcycle riders in crashes and contributed little to overall lethality. Light trucks comprised 20% of passenger vehicles involved in 2-vehicle motorcycle crashes, in which they were twice as deadly as cars. Overall crash lethality was 3.1%, and single-vehicle crashes were deadlier than multis.

  • Over the next 13 years, the motorcycle population aged and light trucks became more popular as family vehicles. By 2003, the 55+ age group was involved in 10% of crashes, and light trucks were in 36% of two-vehicle crashes. Both of these factors pushed lethality higher, single-vehicle due to age alone, multi due to both age and other vehicles. Vehicle mix had lifted multiple-vehicle lethality equal to single-vehicle at 4.4%.

  • In the 13 years that followed, from 2003 to 2016, the two trends continued. The 55+ age group doubled to 21% of the crashers, but growth of the light truck segment eased, increasing only to 40%. In the most recent data, multiple-vehicle crash lethality is now greater than single-vehicle, 4.6% to 4.2%, but overall lethality is unchanged from 2003 at 4.4%.

I believe I have shown (to my own satisfaction, at least) that the popular pickups, SUVs, vans, and crossovers now in the passenger vehicle fleet have made crashing deadlier for motorcyclists. How many riders' lives has this trend cost? My final post in this thread will attempt to answer that question.


Data in this post again comes from NHTSA's GES and FARS databases.
Go to Top of Page

DataDan
Advanced Member
567 Posts
[Mentor]


Central Coast, CA
USA

Yamaha

FJR1300

Posted - 10/10/2018 :  9:16 AM
Estimating Lives Lost

Recall from my last post that "The increased involvement of light trucks has substantially increased the lethality of multiple-vehicle motorcycle crashes." That is, the percentage of crashes that take a motorcyclist's life is greater than it would have been without the rising popularity of light trucks. But what does that mean in terms of rider deaths? In this graph, I estimate the cost in motorcyclists' lives:



The dashed line is light truck involvement in 2-vehicle motorcycle crashes, climbing from 20% in 1990 to 40% in 2016. The shaded area is the number of deaths reported by NHTSA in multiple-vehicle motorcycle crashes. The red segment represents my estimate of lives that would not have been lost had the light truck crash involvement remained at 20%. In recent years, I project 251 fewer deaths per year, 9.2% of multiple-vehicle crash deaths, 5.1% of all motorcyclist deaths.

Conclusion

Motorcycle safety has come a long way in 25 years. Crash rate per motorcycle and per mile traveled is down considerably. Yet the fatality rate is, at best, flat compared to the low of the mid 1990s and perhaps up somewhat. In other words, crash lethality is up; motorcycle crashes have become more deadly. In part, that is due to the aging rider population. But--as I hope this thread has shown--it is also due to the growing number of light trucks in the passenger vehicle fleet.

The growing preference for light trucks as family vehicles has cost motorcyclists' lives. The choice of light trucks isn't going to change anytime soon and, in fact, could grow. But the presence of more dangerous vehicles on the road (to us, at least) is something the riding community needs to understand.



Calculating the Projection

I am including this section to document the procedure I used to project lives saved. It's an old school habit: "Show your work," as they say.

To calculate a plausible "What if?" estimate, as few parameters as possible should be changed from actual observations. Deviation of the hypothetical result from reality should be attributable only to the variable being tested. So, in estimating the number of riders who have been killed due to the growing number of light trucks on the road, I assume for each year the estimate covers:
  • The number of multiple-vehicle motorcycle crashes is unchanged from reported totals.

  • However, the proportions of two-vehicle crashes with cars and light trucks will remain at their 1990 levels. In 1990, there were 4 times as many car crashes as light truck crashes, or 20% light trucks. I assume the same proportion, 20% light trucks, for the projection.

  • In two-vehicle crashes between a motorcycle and either a car or light truck, lethality--the percentage of fatal crashes for each vehicle type--is as observed. This accommodates the aging rider population and changing vehicle design.

  • The percentage change in two-vehicle crash deaths will also be applied to motorcycle crashes involving three or more vehicles.

To illustrate, I will run through the calculation for 2016. As I have throughout this thread I use 3-year moving average to smooth out spurious year-to-year variation.

The total number of 2-vehicle crashes between a motorcycle and a passenger vehicle as estimated in NHTSA's Crash Report Sampling System:

....actual mc-car crashes:......... 33,194 (60%)
....actual mc-LT crashes:.......... 22,021 (40%)
....actual total crashes:.......... 55,215

The number of deaths in these crashes reported by NHTSA's Fatality Analysis Reporting System:

....mc-car deaths:................. 1,024
....mc-LT deaths:.................. 1,130
....total deaths:.................. 2,154

Lethality, deaths as a percentage of crashes:

....mc-car lethality:.............. 3.1%
....mc-LT lethality:............... 5.1%

Now, assume that the crash proportion is 80%-20% instead of the actual percentage and project the redistribution of crashes, but with the same total as above:

....mc-car crashes:................ 44,172 (80%)
....mc-LT crashes:................. 11,043 (20%)
....total crashes:................. 55,215

Apply lethality calculated from observation to the projected crash count to project deaths:

....mc-car deaths:................. 1,363
....mc-LT deaths:.................... 567
....total deaths:.................. 1,929

Projected deaths in 2-vehicle crashes are down by 225 from the same number of crashes.

Crashes with 3 or more vehicles are a problem because it isn't easy to tell who hit whom. So I will calculate the 2-vehicle lives saved of 225 as a percentage all two vehicle crash deaths--including crashes with trucks, buses, other motorcycles, etc.--then apply that same percentage reduction to 3+ vehicle crashes.

The total number of 2-vehicle motorcycle crash deaths in 2016 with all vehicle types was 2,571. So projected fatalities are reduced by 8.75%. Now apply that same reduction to deaths in 3+ vehicle crashes:

actual 3+ vehicle deaths............ 376
projected 3+ vehicle deaths......... 343

That leaves us with the fatality reduction plotted above in this post:

....2-vehicle crash lives saved:.... 225
....3+ vehicle crash lives saved:.... 33
....total lives saved:.............. 258



Data is from NHTSA's GES and FARS databases

Edited by - DataDan on 10/10/2018 11:12 AM
Go to Top of Page

bachman1961
Male Advanced Member
2268 Posts
[Mentor]


colorado springs, co
USA

Honda

CB750 NightHawk

Posted - 11/23/2018 :  3:40 AM
Interesting trends.
I always appreciate the topics you take on and all the leg work associated. Somehow I feel safer reading this stuff as if I know what's going to try to get me. Well, definitely not WHEN but on the bike, it all helps defy mind drift and keep a tight rein on concentration.

I'd imagine in no small way, many of these bigger vehicles more popular on the roads these days (if not taking over) are more of a challenge in emergency maneuver's / accident avoidance or even stopping distance.
My newer car is just a tad over 3100#,and feels very fit and tidy on roads with nearly surgical response to input. My wife's daily ride and our 'family car' is a 4300# crossover or SUV (not sure), handles and rides very luxurious but certainly a far cry from responsive behaviors of smaller options.
To some degree, I suspect crashes will be less avoidable by this larger group on the roads just going by the physics involved.
Go to Top of Page

DataDan
Advanced Member
567 Posts
[Mentor]


Central Coast, CA
USA

Yamaha

FJR1300

Posted - 11/23/2018 :  12:30 PM
quote:
Originally posted by bachman1961

I'd imagine in no small way, many of these bigger vehicles more popular on the roads these days (if not taking over) are more of a challenge in emergency maneuver's / accident avoidance or even stopping distance.

My newer car is just a tad over 3100#,and feels very fit and tidy on roads with nearly surgical response to input. My wife's daily ride and our 'family car' is a 4300# crossover or SUV (not sure), handles and rides very luxurious but certainly a far cry from responsive behaviors of smaller options.

To some degree, I suspect crashes will be less avoidable by this larger group on the roads just going by the physics involved.


While it seems likely that light trucks are involved in more crashes due to their inferior maneuverability--published braking distances are much longer, for example--I wasn't able to build a quantitative case. The problem is that exposure measurements aren't very good. Crashes can easily be compared by vehicle type, but not exposure by vehicle type.

However, I did do one narrow comparison: rear-endER crashes--hitting the vehicle ahead--vs rear-endEE crashes--getting hit from behind. Using the ratio of the two kinds of crashes effectively takes exposure into account. Then, if light truck braking were a significant factor, they would be more likely to hit a motorcycle from behind than would a car. That's not how it worked out, though. In fact, the two ratios were very close (and the motorcycle was one-third more likely to be the striking vehicle).

Go to Top of Page

Baggsy
Male Advanced Member
727 Posts
[Mentor]


Ottawa, Ontario
Canada

Suzuki

09 Wee

Posted - 11/28/2018 :  8:40 PM
Is it possible that one of the factors of light trucks being involved in more accidents, is simply them being a bigger target? They also take up more of the lane, so chances of a motorcycle squeezing between two trucks is less than it squeezing between two small cars.

Go to Top of Page

DataDan
Advanced Member
567 Posts
[Mentor]


Central Coast, CA
USA

Yamaha

FJR1300

Posted - 11/29/2018 :  11:10 AM
quote:
Originally posted by Baggsy

Is it possible that one of the factors of light trucks being involved in more accidents, is simply them being a bigger target? They also take up more of the lane, so chances of a motorcycle squeezing between two trucks is less than it squeezing between two small cars.

Light trucks are actually involved in fewer motorcycle crashes than cars, but when they are in a crash, the motorcyclist is more likely to be killed. In 2016, they were involved in 33% fewer crashes but 10% more deaths.
Go to Top of Page
  Previous Topic Discussion Topic Next Topic  
Jump To:
All Things (Safety Oriented) Motorcycle © Master Strategy Group Go To Top Of Page
  This page was generated in 0.38 seconds. Powered By: Snitz Forums 2000 Version 3.4.05