Correlation of IIHS Moderate Overlap Offset Rating with 2011-Present NHTSA Injury Risk Score, Over Time
Note: I am not affiliated with the NHTSA or IIHS, or am even an engineer. I am a 22 year old with Asperger's Syndrome who is very interested in car safety.
This post is not intended to endorse one rating system over the other.
Explanations of tests and injury criteria:
The IIHS moderate overlap offset test, introduced in 1995, involves crashing a vehicle into a deformable offset barrier at 40 mph, with 40% of the front end on the driver's side, using a Hybrid III 50th percentile male dummy (about 5'10 and 180 lbs) belted into the driver's seat. Vehicles are rated on six categories; structure/safety cage (how much the occupant compartment collapsed), restraints/dummy kinematics (how well the seat belts, airbags, and other vehicle parts controlled dummy movement), and four injury subsections: head/neck, chest, left leg/foot, and right leg/foot. The entire body from head to toe is effectively covered. Each of these six categories gets a rating of good, acceptable, marginal, or poor, and these ratings are combined into a rating of good, acceptable, marginal, or poor.
NHTSA crash tests vehicles into a flat wall at 35 mph. Vehicles get a rating of 1-5 stars on each side, with the rating being strictly on injury risk to the head, neck, chest, and femur (in effect, the entire body except for the lower legs). The current system, introduced on October 5, 2010 for 2011 model vehicles, uses a formula to determine the risk of serious injury to a real life occupant. It's comprehensive, and it adds up. As an example, the 2011 Honda Accord got a 9.8% risk of driver injury in the NHTSA full frontal test, giving it 5 stars. But that 9.8% is divided into a 0.0% risk of head injury, a 6.6% risk of neck injury, a 2.7% risk of chest injury, and a 0.8% risk of femur injury. The total does not add up to 9.8% because there is a slight chance that multiple body regions could sustain serious injury.
In late 2010, when the first results for the new NHTSA test results came out, I downloaded a copy of a spreadsheet with the NHTSA tests up to that point. The dummies in the driver's seat of the new NHTSA tests are the same as the dummies in the IIHS offset tests, and all applicable injury measures are in the technical information. (HIC 15, Nij, Neck Tension in newtons, Chest Deflection in millimeters, and femur forces in newtons on each side). On the IIHS website, they give the newtons in kilonewtons; for instance, I put in 1.4 kilonewtons as 1,400 newtons. NHTSA also records neck compression, but it would take extreme values of neck compression to make a difference to the rating, and Nij (neck injury criterion) tends to take care of that factor, anyway.
Results are given for all tests done during a particular model year. For instance, the 2011 Honda Accord's crash test applies to 2008-2012 models, but is included under the 2008 model year because the tested car was a 2008. Similarly, the 1998 Ford F-150's crash test applies to 1997-2003 models, but is included under the 2001 model year as the tested truck was a 2001.
9.9% or less = 5 stars
10.0%-14.9% = 4 stars
15.0%-19.9% = 3 stars
20.0%-39.9% = 2 stars
40.0% or more = 1 star
For more information, visit
IIHS technical info
NHTSA technical info
The formula for the injury risk calculation is at (It is an advanced calculus formula)
http://www-nrd.nhtsa.dot.gov/pdf/esv/esv21/09-0322.pdf
Data
Avg injury risk - all vehicles tested during the model year:
1995: 19.3%
1996: 23.7%
1997: 25.7%
1998: 19.7%
1999: 20.6%
2000: 19.8%
2001: 26.1%
2002: 18.3%
2003: 16.0%
2004: 15.0%
2005: 15.4%
2006: 16.2%
2007: 15.2%
2008: 17.4%
2009: 13.4%
2010: 14.9%
2011: 12.6%
2012: 13.3%
2013: 12.1%
2014: 11.7%
2015: 10.8%
Average risk, 1995-1997: 23.2%
1995-97 Poor: 26.7%
1995-97 Marginal: 21.0%
1995-97 Accept.: 23.9%
1995-97 Good: 17.3%
Poor-Good improvement: 36% lower risk
Average risk, 1998-2000: 20.1%
1998-2000 Poor: 27.8%
1998-2000 Marginal: 26.5%
1998-2000 Accept.: 18.9%
1998-2000 Good: 15.2%
Poor-Good improvement: 45% lower risk
Average risk, 2001-2003: 20.1%
2001-2003 Poor: 58.6%
2001-2003 Marg.: 28.5%
2001-2003 Accept.: 18.8%
2001-2003 Good: 16.1%
Poor-Good improvement: 73% lower risk
Average risk, 2004-2006: 15.6%
2004-2006 Accept.: 19.1%
2004-2006 Good: 14.8%
Conclusions
The correlation between IIHS rating and NHTSA injury risk score was strong. Good rated vehicles had a much lower risk than Poor rated vehicles, with Marginal and Acceptable being intermediate steps. Due to the low number of non-Good rated vehicles from 2004 and later model years, the comparison was for 1995-2003 vehicles. 94% of vehicles from 2004-present are rated Good. Some caveats to the comparison are the much higher risk on the 2001-2003 Poor rated vehicles than the 1995-2000 Poor rated vehicles. This is due to the small sample size (only 4 different models). However, these four models (the 1997-2003 Ford F-150, the 1994-2001 Dodge Ram, the 2001-2003 Hyundai Elantra, and the 2001 Dodge Caravan) were all high-selling models.
Also of note was the rather sudden improvement pattern seen from around 2010 on. On pre-2009 vehicles, the average risk for Good vehicles stayed around the 15-17% range. Improvements during these years were largely due to the improvements in the IIHS ratings. With Good ratings nearly ubiquitous since the late 2000s, the improvements seen in the post-2010 period represent major improvements in Good rated vehicles. The 2013-2015 vehicles (all of which have been rated Good thus far) have a 21% lower risk on average than the Good-rated 2004-2006 rated vehicles. (11.7% vs 14.8%). These recent improvements may be due to restraint changes stemming from the implementation of the new 2011-present NHTSA rating system, which is far more stringent than the 1979-2010 system.
While the correlation was strong, it was not total: the best Poor vehicle got an 11.0% risk, and the worst Good got a 32.1% risk. 23% of vehicles rated Poor got a risk under 20%. Part of this reason is that the NHTSA rating system does not cover tibia injuries, structure, or restraints. That 11.0% risk Poor vehicle (the 1997 Cadillac Seville) is a poster child of this. The structure of this vehicle collapsed, and dummy movement was poorly controlled, leading to Poor ratings on both of those factors. High forces on the right tibia lead to a Poor rating for the right leg. Many of the "low-risk" Poor rated vehicles suffered from a combination of tibia injury risk, structure, and restraint problems. Some vehicles had high head accelerations from the head hitting the steering wheel through the airbag or hitting the roof or pillar on rebound. In many cases, these high accelerations are of very short duration, so they do not result in a high HIC. Life threatening injuries would be unlikely, but these hard contacts can cause less severe head injuries including concussions.
Another factor is that IIHS does injury ratings "piecemeal"; there are certain force limits that are considered to be "likely" to cause certain levels of injury. If, for instance, a vehicle got Good on all injury ratings but the force levels were all "borderline", i.e. HIC, Nij, neck tension, chest compression, and both femur loads were just under the Good/Acceptable boundary, the injury risk would be high, over 45%. In practice, however, this would never occur. Also, theoretically, a vehicle could have a risk under 10% and get a Poor rating if structure and restraints were Poor and tibia loads were high, but head, neck, chest, and femur forces were all low. This nearly happened, but is extremely unlikely as structural collapse usually imparts higher upper body loads through the restraint system or airbag bottoming out, and dashboard movement often causes higher femur loads.
Some stock can be put to the risks, especially in Good rated vehicles. For instance, a Good rated vehicle with a risk of 10% is probably safer than one with 20%. But a low "risk" does not a safe vehicle make. Regardless of the risk, the IIHS rating is the best indicator of how well a vehicle will do in an offset crash. It was specifically designed for the crash scenario.
This post is not intended to endorse one rating system over the other.
Explanations of tests and injury criteria:
The IIHS moderate overlap offset test, introduced in 1995, involves crashing a vehicle into a deformable offset barrier at 40 mph, with 40% of the front end on the driver's side, using a Hybrid III 50th percentile male dummy (about 5'10 and 180 lbs) belted into the driver's seat. Vehicles are rated on six categories; structure/safety cage (how much the occupant compartment collapsed), restraints/dummy kinematics (how well the seat belts, airbags, and other vehicle parts controlled dummy movement), and four injury subsections: head/neck, chest, left leg/foot, and right leg/foot. The entire body from head to toe is effectively covered. Each of these six categories gets a rating of good, acceptable, marginal, or poor, and these ratings are combined into a rating of good, acceptable, marginal, or poor.
NHTSA crash tests vehicles into a flat wall at 35 mph. Vehicles get a rating of 1-5 stars on each side, with the rating being strictly on injury risk to the head, neck, chest, and femur (in effect, the entire body except for the lower legs). The current system, introduced on October 5, 2010 for 2011 model vehicles, uses a formula to determine the risk of serious injury to a real life occupant. It's comprehensive, and it adds up. As an example, the 2011 Honda Accord got a 9.8% risk of driver injury in the NHTSA full frontal test, giving it 5 stars. But that 9.8% is divided into a 0.0% risk of head injury, a 6.6% risk of neck injury, a 2.7% risk of chest injury, and a 0.8% risk of femur injury. The total does not add up to 9.8% because there is a slight chance that multiple body regions could sustain serious injury.
In late 2010, when the first results for the new NHTSA test results came out, I downloaded a copy of a spreadsheet with the NHTSA tests up to that point. The dummies in the driver's seat of the new NHTSA tests are the same as the dummies in the IIHS offset tests, and all applicable injury measures are in the technical information. (HIC 15, Nij, Neck Tension in newtons, Chest Deflection in millimeters, and femur forces in newtons on each side). On the IIHS website, they give the newtons in kilonewtons; for instance, I put in 1.4 kilonewtons as 1,400 newtons. NHTSA also records neck compression, but it would take extreme values of neck compression to make a difference to the rating, and Nij (neck injury criterion) tends to take care of that factor, anyway.
Results are given for all tests done during a particular model year. For instance, the 2011 Honda Accord's crash test applies to 2008-2012 models, but is included under the 2008 model year because the tested car was a 2008. Similarly, the 1998 Ford F-150's crash test applies to 1997-2003 models, but is included under the 2001 model year as the tested truck was a 2001.
9.9% or less = 5 stars
10.0%-14.9% = 4 stars
15.0%-19.9% = 3 stars
20.0%-39.9% = 2 stars
40.0% or more = 1 star
For more information, visit
IIHS technical info
NHTSA technical info
The formula for the injury risk calculation is at (It is an advanced calculus formula)
http://www-nrd.nhtsa.dot.gov/pdf/esv/esv21/09-0322.pdf
Data
Avg injury risk - all vehicles tested during the model year:
1995: 19.3%
1996: 23.7%
1997: 25.7%
1998: 19.7%
1999: 20.6%
2000: 19.8%
2001: 26.1%
2002: 18.3%
2003: 16.0%
2004: 15.0%
2005: 15.4%
2006: 16.2%
2007: 15.2%
2008: 17.4%
2009: 13.4%
2010: 14.9%
2011: 12.6%
2012: 13.3%
2013: 12.1%
2014: 11.7%
2015: 10.8%
Average risk, 1995-1997: 23.2%
1995-97 Poor: 26.7%
1995-97 Marginal: 21.0%
1995-97 Accept.: 23.9%
1995-97 Good: 17.3%
Poor-Good improvement: 36% lower risk
Average risk, 1998-2000: 20.1%
1998-2000 Poor: 27.8%
1998-2000 Marginal: 26.5%
1998-2000 Accept.: 18.9%
1998-2000 Good: 15.2%
Poor-Good improvement: 45% lower risk
Average risk, 2001-2003: 20.1%
2001-2003 Poor: 58.6%
2001-2003 Marg.: 28.5%
2001-2003 Accept.: 18.8%
2001-2003 Good: 16.1%
Poor-Good improvement: 73% lower risk
Average risk, 2004-2006: 15.6%
2004-2006 Accept.: 19.1%
2004-2006 Good: 14.8%
Conclusions
The correlation between IIHS rating and NHTSA injury risk score was strong. Good rated vehicles had a much lower risk than Poor rated vehicles, with Marginal and Acceptable being intermediate steps. Due to the low number of non-Good rated vehicles from 2004 and later model years, the comparison was for 1995-2003 vehicles. 94% of vehicles from 2004-present are rated Good. Some caveats to the comparison are the much higher risk on the 2001-2003 Poor rated vehicles than the 1995-2000 Poor rated vehicles. This is due to the small sample size (only 4 different models). However, these four models (the 1997-2003 Ford F-150, the 1994-2001 Dodge Ram, the 2001-2003 Hyundai Elantra, and the 2001 Dodge Caravan) were all high-selling models.
Also of note was the rather sudden improvement pattern seen from around 2010 on. On pre-2009 vehicles, the average risk for Good vehicles stayed around the 15-17% range. Improvements during these years were largely due to the improvements in the IIHS ratings. With Good ratings nearly ubiquitous since the late 2000s, the improvements seen in the post-2010 period represent major improvements in Good rated vehicles. The 2013-2015 vehicles (all of which have been rated Good thus far) have a 21% lower risk on average than the Good-rated 2004-2006 rated vehicles. (11.7% vs 14.8%). These recent improvements may be due to restraint changes stemming from the implementation of the new 2011-present NHTSA rating system, which is far more stringent than the 1979-2010 system.
While the correlation was strong, it was not total: the best Poor vehicle got an 11.0% risk, and the worst Good got a 32.1% risk. 23% of vehicles rated Poor got a risk under 20%. Part of this reason is that the NHTSA rating system does not cover tibia injuries, structure, or restraints. That 11.0% risk Poor vehicle (the 1997 Cadillac Seville) is a poster child of this. The structure of this vehicle collapsed, and dummy movement was poorly controlled, leading to Poor ratings on both of those factors. High forces on the right tibia lead to a Poor rating for the right leg. Many of the "low-risk" Poor rated vehicles suffered from a combination of tibia injury risk, structure, and restraint problems. Some vehicles had high head accelerations from the head hitting the steering wheel through the airbag or hitting the roof or pillar on rebound. In many cases, these high accelerations are of very short duration, so they do not result in a high HIC. Life threatening injuries would be unlikely, but these hard contacts can cause less severe head injuries including concussions.
Another factor is that IIHS does injury ratings "piecemeal"; there are certain force limits that are considered to be "likely" to cause certain levels of injury. If, for instance, a vehicle got Good on all injury ratings but the force levels were all "borderline", i.e. HIC, Nij, neck tension, chest compression, and both femur loads were just under the Good/Acceptable boundary, the injury risk would be high, over 45%. In practice, however, this would never occur. Also, theoretically, a vehicle could have a risk under 10% and get a Poor rating if structure and restraints were Poor and tibia loads were high, but head, neck, chest, and femur forces were all low. This nearly happened, but is extremely unlikely as structural collapse usually imparts higher upper body loads through the restraint system or airbag bottoming out, and dashboard movement often causes higher femur loads.
Some stock can be put to the risks, especially in Good rated vehicles. For instance, a Good rated vehicle with a risk of 10% is probably safer than one with 20%. But a low "risk" does not a safe vehicle make. Regardless of the risk, the IIHS rating is the best indicator of how well a vehicle will do in an offset crash. It was specifically designed for the crash scenario.
Excellent Article.. Kindly visit us @ Best Construcion in Nagerkoil
ReplyDelete