Frontal NCAP crash tests with rear-seat occupant

Abstract

Objectives

The initial frontal NCAP tests in 1979 included lap-shoulder belted driver and right-front passenger and lap belted 6-year-old (yo) in the rear. The 35 mph barrier tests were reviewed and analyzed for the restraint performance of the front occupants and child in the rear.

Methods

The initial 100 crash tests (#1–#100) in the NHTSA database were searched for frontal barrier impacts. Fifteen tests met the criteria. There were three tests with the 1980 Chevrolet Citation at 35, 40 and 48 mph. There were 12 other tests with different passenger vehicles at 35 mph into the rigid barrier. The tests included a lap-shoulder belted Hybrid II (Part 572) dummy in the driver and right-front passenger seat and a lap belted 6 yo child dummy (Alderson VIP 6 C) in the center or right rear seat. Vehicle dynamics and occupant kinematics were analyzed, and dummy responses were compared.

Results

Vehicle deformation was progressive with impact speed for the Citation tests, leading NHTSA to settle on a 35 mph NCAP speed. The thirteen 35 mph NCAP tests had an average driver HIC of 1099 ± 381 (95th CI 207) and 3 ms chest acceleration of 55.7 ± 16.1 g (95th CI 8.8) with 7 of 13 vehicles failing FMVSS 208 injury criteria. The average right-front passenger HIC was 1179 ± 555 (95th CI 302) and 3 ms chest acceleration was 47.2 ± 14.6 g (95th CI 7.9) with 7 of 13 failing injury criteria. Only four tests (30.8%) passed driver and right-front passenger injury criteria.

The responses in the rear seat were significantly worse. The average HIC was 2711 ± 1111 (95th CI 604) and 3 ms chest acceleration was 62.8 ± 10.6 g (95th CI 5.8). The films showed the child’s upper body moved forward and rotated downward around the lap belt resulting in severe head impacts on the front seatback, floor, dummy legs or interior. All vehicles failed injury criteria by large margins. Submarining the lap belt was noted in 6 tests. HIC for the rear child was 2.47-times greater than the driver (t = 4.72, p < 0.001) and 2.30-times greater than the right-front passenger (t = 3.64, p < 0.005).

Conclusions

In the 1979 NCAP tests, the child dummy experienced inadequate restraint by the lap belt in the rear seat. The child jackknifed around the lap belt, often submarined, with a severe head impact. No publication of the results has been found. NHTSA did not advise the public of the extremely poor restraint performance, even during the public discussions on the 1986 NTSB recommendation that U.S. vehicle manufacturers install lap-shoulder belts in rear outboard seats. None of the subsequent NCAP tests included a child or adult in the rear until nearly 25 years later.

Keywords:

• NCAP testing
• rear seat safety
• occupant protection
• child safety
• lap belt injury
• submarining

Introduction

NHTSA publications on NCAP

In the 1980s and 1990s, NHTSA regularly published studies at the ESV (Enhanced Safety Vehicle) Conference and SAE (Society of Automotive Engineers) on the frontal NCAP tests. Brownlee et al. (1980) provided one of the earliest analyses of the frontal NCAP tests. They noted the testing addressed two issues of importance for the driver and right-front passenger. Are there differences in the ability of competing small cars to protect their front occupants and what information would be helpful to the consumer. The NCAP testing showed significant difference in performance for the driver and right-front passenger with cars of similar weight. The tests showed the causes for the differing performance were related to different vehicle deformations that absorbed energy, steering column performance with driver head impact on the rim and hub and restraint performance of the lap-shoulder belts.

Hollowell and Arendt (1982) studied quality control issues with NHTSA barrier crash testing, including instrumentation, data filtering, data analysis and test dummies. Hackney and Quarles (1982) studied performance issues of seatbelt systems and steering column on differences in driver protection in NCAP testing. They addressed performance improvements of the 1981 Honda Civic in comparison to the poor NCAP results of the 1979 Civic. They also discussed consumer information on small car safety in the 1980 Car Book (NHTSA 1980). MacLaughlin and Saul (1982) conducted in-depth analysis of driver and right-front passenger performance in NCAP testing. Machey and Gauthier (1984) studied the repeatability of NCAP tests using 1982 Chevrolet Citations with a minimum of four repeat impacts at three different test sites, resulting in 14 tests. The Citations were manufactured consecutively to achieve uniformity. The results identified causes for variability and were used to improve NCAP test procedures.

Hackney and Ellyson (1985) extended the analysis of seatbelts, steering assemblies and vehicle structures with more NCAP testing of front occupants. Jones et al. (1985) provided analyses of driver head kinematics and impact on the steering wheel rim and hub. Hackney et al. (1989, 1996) conducted an in-depth analysis of trends in driver and passenger performance in NCAP tests with newer models performing better than older models. Cohen et al. (1989) conducted in-depth analysis of the right-front passenger performance in NCAP tests. Hackney (1991) compared NCAP with FMVSS 208 testing 1987 to 1991 MY vehicles. Hackney et el. (1994) provided a historic review of NCAP, vehicle performance trends and laid out future efforts to continue and expand the NCAP program by considering additional crash tests and measures of potential injuries to rear-seat passengers. This was 15 years after the initial 1979 NCAP tests with a rear occupant. None of the NHTSA publications mentioned the early NCAP tests with the 6 yo child lap belted in the rear seat.

NHTSA crash test database

NHTSA maintains a database of crash tests that has moved and upgraded as data collection and storage methods have changed (Digges K. Personal communications 2023; Summers S. Personal communications 2023). Initially, test films were stored at the NHTSA Headquarters Film Library and provided copies upon request. The high demand for NCAP films increased NHTSA’s in-house workload and in 1989, the Film Library was transferred under contract to the University of Virginia, where it was maintained with a full-time librarian and public access in nearby Merryfield, Virginia. In 1992 the film library was included as part of the NHTSA contract with the National Crash Analysis Center at George Washington University. The films were stored at the George Washington University, National Crash Analysis Center in Sterling Virginia where a full-time librarian provided public access. Before digital technology became available, George Washington University transferred crash test films to VHS, when requested by the public. The original films remained at George Washington University until the end of the contract. They were then sent to the National Archives, where they remain accessible to the public.

Test films were converted to digital videos that were transferred back to NHTSA and added to the database. NHTSA maintained the original written reports and data on magnetic tape. The reports were converted to electronic files in pdf. The transducer data was transferred from magnetic tape to digital files in various formats. Each of the crash tests was numbered in sequence. The tests became available online from the NHTSA website (https://www.nhtsa.gov/research-data/research-testing-databases#/) starting in 1998 with digital videos added as they became available after 2001.

Methods

Frontal NCAP tests

In 1978, NHTSA established NCAP (New Car Assessment Program) in response to Title II of the Motor Vehicle Information and Cost Savings Act (US Congress 1972). NCAP tests were intended to provide comparative information on the safety of new vehicles to assist consumers with purchasing decisions and encourage motor vehicle manufacturers to make safety improvements. NHTSA was advised that the severity of the NCAP test should not be so severe as to eliminate a class of vehicles because of performance (Digges K. Personal communications 2023). In 1979, NHTSA began rating passenger vehicles for frontal impact safety based on driver and right-front passenger injury responses. In 1997, NHTSA added crash tests and ratings for side impact safety. In 2001, NHTSA added a rollover resistance rating based on the vehicle SSF, static stability factor (Federal Register 2001).

In 1978, NHTSA setup protocols for frontal barrier crash testing at laboratories in the U.S. The testing was managed by the Office of Vehicle Safety Compliance. Four Federal Motor Vehicle Safety Standards (FMVSS) were evaluated in the NCAP tests, FMVSS 212 windshield mounting, FMVSS 219 windshield zone intrusion, FMVSS 301 fuel system integrity and FMVSS 208 injury criteria. The initial tests were conducted at Calspan Corporation, Advanced Technology Center, Buffalo, NY 14225.

The initial 100 crash tests (#1–#100) in the NHTSA database were searched for frontal barrier impacts. Fifteen tests met the criteria. There were three tests with the 1980 Chevrolet Citation at 40 (#1), 48 (#4) and 35 (#5) mph to establish the impact severity for subsequent NCAP testing. NHTSA chose the 1980–1985 MY Chevrolet Citation to develop front and side NCAP test procedures because they determined it was one of the best new vehicles on the market for safety (Digges K. Personal communications 2023). NHTSA also used the Citation to demonstrate improvements with airbags and modifications of vehicle structures and the interior. The Chevrolet Citation was built on the GM X-car platform. It was front-wheel drive with significant reduction in body size and weight from the larger, rear-wheel drive vehicles of the 1970s.

There were 13 frontal NCAP tests with different passenger vehicles at 35 mph into the rigid barrier. The tests included a lap-shoulder belted Hybrid II (Part 572) dummy in the driver and right-front passenger seat and a lap belted 6 yo child dummy (Alderson VIP 6 C) in the center or right rear seat. Vehicle dynamics, occupant kinematics and dummy responses were evaluated. All the tests were conducted at Calspan Corporation except one (#35) conducted at Dynamic Science, Inc. 1850 West Pinnacle Peak Rd., Phoenix, Arizona 85027. The vehicle acceleration was measured at five or more locations: (1) engine compartment, (2,3) fire-wall, (4) right passenger door sill, (5) left B-pillar sill and (6) vehicle cg (center of gravity).

High-speed films

Each test involved 8–9 high-speed Photosonic cameras filming at 400–1,000 fps. There was a left and right side overall view of the vehicle dynamics, closeups of the driver and right-front passenger through the front windows, closeups of the child in the rear through the rear windows and other views through the windshield and vehicle underbody.

Occupants

The tests included a lap-shoulder belted Hybrid II (Part 572) dummy in the driver and right-front passenger seat and a lap belted 6 yo child dummy (Alderson VIP 6 C) in either the center or right rear seat. The dummies were normally seated. The dummies were instrumented to measure head and chest triaxial acceleration and left and right femur load. The triaxial acceleration of the head cg (center of gravity) was used to determine peak resultant head acceleration and HIC. The HIC was determined over the duration of the head acceleration; it was not capped at 36 ms or 15 ms duration. The triaxial acceleration of the chest was used to determine the 3 ms peak. The peak femur compression load was determined. Shoulder and lap belt loads were measured.

Instrumentation and filtering

Data acquisition at Calspan was provided by a Sangamo Mod 3500 FM Tape Transport with a recording speed of 60 in/s. Signal conditioning was provided by a Calspan designed and fabricated bridge amplifiers providing gain, balance. and shunt calibration of the sensor inputs. A real time reference was provided by Tektronix Time Mark Generator with 1 ms and 10 ms synchronously recorded with all data and reproduced at playback on all strip chart records. Calspan designed and fabricated second order active filters (6 channels) with selectable frequency responses consistent with SAE J2l1-1 Class 60, 180, 600 and 1,000 requirements (SAE 1995a, 1995b). Filters were compensated for 32:1 real time expansion resulting from data recording at 60 in/sec and playback at 1-7/8 in/sec. The data was reduced and plotted using a Brush Mark 260 6-channel recorder or Pace X-Y plotter and stored on magnetic tape.

Injury tolerances

The peak responses were compared to the current IARV (injury assessment reference value) from Mertz et al. (2016). The tolerances represent approximately 15% risk of serious injury (AIS 3).

Significance of response differences

The statistical significance of differences in HIC, peak head acceleration and 3 ms chest acceleration between the rear child and the driver or right-front passenger was determined using the paired sample t-test in Excel. Twelve (12) NCAP tests were available with front and rear occupant responses; the child dummy responses were lost in test #25. Differences were considered statistically significant with p < 0.05. The summary data is presented with the average ± standard deviation (95th confidence interval for the mean).

Results

Barrier impacts evaluating severity

Figure 1 shows a sequence of vehicle dynamics for the three Citation NCAP tests at 35, 40 and 48 mph (56.3, 64.4 and 77.3 km/h) into the barrier in tests #5, #1 and #4. Table 1 shows vehicle deformation and intrusion were progressive with impact speed for the three Citation tests. Front crush increased from 21.4, 29.0 to 40.3 in and firewall intrusion increased from 8.1, 10.8 to 21.2 in. The delta V was estimated as 39.2, 44.7 and 53.8 mph in tests #5, #1 and #4 based on 0.12 coefficient of restitution.

Figure 1. Vehicle dynamics in 35, 40 and 48 mph (56.3, 64.4 and 77.3 km/h) frontal barrier impacts, tests #5, #1 and #4.

Table 1. Initial NCAP testing with the 1980 Chevrolet Citation.

							Driver			Right-front passenger
DOT #	Date	Weight	Impact	Crush	Intrusion	Delta V^a	HIC	Head g	Chest 3ms	HIC	Head g	Chest 3ms
		lb	mph	in	in	mph		g	g		g	g
5	6/12/79	3,260	35.0	21.4	8.1	39.2	845	79.0	48.0	623	64.0	34.5
1	6/6/79	3,153	39.9	29.0	10.8	44.7	1039	100.0	60.0	1117	86.0	38.0
4	6/7/79	3,153	48.0	40.3	21.2	53.8	4167	315.0	94.5	2433	160.0	61.0
						IARV (Mertz et al. 2016)	1000	180	60	1000	180	60

^a Assuming e = 0.12 coefficient of restitution.

Figure 1 shows the 48 mph test (right images) resulted in severe distortion of the occupant compartment and the driver HIC was 4167. The deformation of the vehicle and occupant responses were severe. The 40 mph test (center) resulted in less deformation of the occupant compartment but there was windshield intrusion and failure of the fuel system. The driver HIC was 1039, above tolerance. The 35 mph test (left) passed FMVSS and injury criteria with a driver HIC of 845. The tests setup the 35 mph impact for subsequent NCAP testing with what NHTSA judged a safe vehicle design passing the test. The 35 mph impact speed involved 36% more kinetic energy into the barrier than the 30 mph impact in FMVSS 208.

35 mph frontal NCAP tests

Table 2 shows the 13 passenger cars in the initial 35 mph NCAP tests run in 1979. The average vehicle weight was 3,320 ± 731 lb (1,509 ± 332 kg). The front crush averaged 24.9 ± 2.8 in (633 ± 71 mm) and the intrusion of the firewall was 7.8 ± 1.1 in (198 ± 28 mm). Two of the vehicles had passive restraint seatbelts.

Table 2. 35 mph NCAP Cash tests with restrained occupants.

DOT				Weight	Impact	Crush	Intrusion
#	Date	MY	Vehicle	lb	mph	in	in
5	6/12/79	1980	Chevrolet Citation	3,260	35.0	21.4	8.1
7	7/10/79	1979	VW Rabbit^a	2,600	34.8	22.7	9.6
25	7/24/79	1979	Oldsmobile Cutlass	3,820	35.1	29.8	6.5
27	7/16/79	1979	Toyota Celica	3,025	34.8	24.2	6.9
35	5/24/79	1979	Nissan Datsun 210	2,425	35.2	25.1	--
63	8/28/79	1979	Plymoth Volare	3,836	35.0	23.9	7.3
64	9/7/79	1979	Chevrolet Monza	3,240	35.1	25.5	8.3
65	9/7/79	1979	Ford LTD	4,366	35.4	28.1	9.4
66	8/30/79	1979	Ford Granada	3,950	34.6	27.8	6.3
71	9/20/79	1979	Dodge Magnum	4,456	35.3	27.4	6.8
73	9/17/79	1979	Chevrolet Chevette^a	2,716	34.8	21.9	7.2
92	9/12/79	1979	Ford Fairmont	3,300	35.4	25.0	9.1
94	10/15/79	1979	Honda Civic	2,166	34.7	20.9	7.9
			avg	3,320	35.0	24.9	7.8
			sd	731	0.3	2.8	1.1
			95th CI for mean	397	0.1	1.5	0.6

^a Passive seatbelt.

Table 3 summarizes the responses for the driver, right-front passenger and rear child dummies. The HIC, peak head acceleration and peak 3 ms chest acceleration are listed. The femur and belt loads are provided in Supplementary Appendix A. A notation is added for the seating position of the rear child (C: center rear, R: right rear) and a notation of “yes” for roping, if the report indicated the child jackknifed around the lap belt causing the belt to rope and penetrate the abdomen.

Table 3. Occupant responses in the 35 mph NCAP Cash tests.

	Driver			Right-front passenger			Rear 6-year-old
		Head g	Chest 3ms		Head g	Chest 3ms			Head g	Chest 3ms
DOT #	HIC	g	g	HIC	g	g	Position^a	HIC	g	g	Roping^b
5	845	79.0	48.0	623	64.0	34.5	C	1450	105.0	61.0	--
7	1024	90.0	67.0	429	47.0	33.0	R	2193	103.0	52.0	--
25	997	72.0	58.0	799	59.0	37.0	C	--	--	--	--
27	849	94.0	61.0	1862	158.0	59.0	R	2055	145.0	75.0	--
35	1358	92.0	76.0	1745	88.0	59.4	R	1908	147.8	70.2	--
63	724	74.0	40.0	1677	200.0	40.0	R	1816	170.0	65.0	Yes
64	1108	103.0	42.0	1033	87.0	41.0	R	3497	176.0	67.0	Yes
65	1452	143.0	38.0	616	93.0	34.0	C	4390	188.0	65.0	Yes
66	1442	110.0	61.0	1279	99.0	56.0	C	1589	89.0	54.0	Yes
71	646	58.0	39.0	731	58.0	44.0	C	4354	185.0	60.0	Yes
73	871	118.0	47.0	859	88.0	45.0	R	4182	215.0	84.0	--
92	939	102.0	54.0	1583	94.0	85.0	C	2215	118.0	55.0	--
94	2030	116.0	92.6	2093	142.0	46.0	R	2886	145.0	45.5	Yes
avg	1099	96.2	55.7	1179	98.2	47.2		2711	148.9	62.8
sd	381	22.8	16.1	555	44.0	14.6		1111	39.3	10.6
95th CI	207	12.4	8.8	302	23.9	7.9		604	21.4	5.8
Tolerance	1000	180	60	1000	180	60		1000	189	93

^a C: center rear seat, R: right rear seat. ^bPost test notes indicate the occupant jackknifed around the lap belt causing the lap belt to rope and penetrate the abdominal area.

Front seat adults

The average driver HIC was 1,099 ± 381 (95th CI 207) and 3 ms chest acceleration was 55.7 ± 16.1 g (95th CI 8.8) for the thirteen 35 mph NCAP tests. Seven of 13 vehicles failed injury criteria. For the driver. the average left femur load was 1,153 ± 711 lb, the right femur load was 801 ± 482 lb, the peak shoulder belt load was 1,504 ± 397 lb and the lap belt load was 1,200 ± 340 lb. Jones et al. (1985) provided analysis of the driver head impacts on the rim and hub of the steering wheel.

The average right-front passenger HIC was 1,179 ± 555 (95th CI 302) and 3 ms chest acceleration was 47.2 ± 14.6 g (95th CI 7.9). Seven of 13 vehicles failed injury criteria. For the right-front passenger, the average left femur load was 772 ± 388 lb, the right femur load was 666 ± 465 lb, the peak shoulder belt load was 1,646 ± 180 lb and the lap belt load was 1,371 ± 414 lb. Cohen et al. (1989) provided an analysis of the right-front passenger impacts on the interior and seatbelt loading.

Only four tests (30.8%) passed driver and right-front passenger injury criteria, three GM vehicles (#5, #25 and #73) and a Dodge (#71). The NCAP testing at 35 mph was a challenge for many vehicles. NHTSA identified performance issues for the driver and right-front passenger and worked with the industry to improve vehicle safety (Hackney et al. 1989, 1996).

Rear seat child

The responses in the rear seat were significantly worse. The average HIC was 2,711 ± 1,111 (95th CI 604) and 3 ms chest acceleration was 62.8 ± 10.6 g (95th CI 5.8). Figure 2 shows HIC for 12 vehicles with the child lap belted in the center or right rear seat. All HIC were well above tolerance. All vehicles failed injury criteria by large margins. In 1979, the HIC calculation did not have a limit on duration. The UDS data was downloaded and run through the NHTSA HIC calculator to determine HIC₃₆ and HIC₁₅. The HIC₁₅ calculation and time window are given in Supplementary Appendix B. The average HIC₃₆ was 2,488 ± 1,213 and HIC₁₅ was 1,943 ± 1,144. The 6 yo tolerance for HIC₁₅ is 723 (Mertz et al. 2016). The average HIC₁₅ was 2.69-times tolerance.

Figure 2. Head injury criterion (HIC) for lap-belted 6 yo dummy in rear seat of 35 mph NCAP crash tests.

The film showed the child’s upper body moved forward and rotated downward around the lap belt resulting in severe head impacts on the front seatback, floor, dummy legs or other components. Figure 3 shows kinematics of the rear child in three tests #65 (center rear seat), #73 (right rear seat) and #92 (center rear seat). The movement of the head forward and down around the lap belt involved head velocities above the 30–35 mph. Post-test notes indicated that the rear child jackknifed forward around the lap belt causing the belt to rope and penetrate the abdomen in 6 tests. This indicated submarining the lap belt and risk for abdominal injury. NHTSA did nothing to push the industry to improve rear seat restraints after the 1979 NCAP tests.

Figure 3. Child kinematics with lap belt in rear seat in 35 mph frontal NCAP tests #65, #73 and #92.

Significance of differences

Figure 4 shows the percent of IARV for HIC for the rear child and front occupants. HIC for the rear child was significantly greater (2.47-times) than for the driver (t = 4.72, p < 0.001) and significantly greater (2.30-times) than for the right-front passenger (t = 3.64, p < 0.005). The peak head acceleration for the rear child was significantly greater than for the driver (t = 4.18, df = 11, p < 0.002) and the right-front passenger (t = 3.03, df = 11, p < 0.015). The peak 3 ms chest acceleration for the rear child was similar to the driver (t = 1.09, df = 11, NS) but significantly higher than the right-front passenger (t = 2.76, df = 11, p < 0.020). The responses for the rear 6 yo child were significantly worse than for the driver or right-front passenger in the 35 mph NCAP tests. The HIC in the rear seat was more than twice (2.30–2.47 times) that of the front occupants.

Figure 4. Percent of IARV for HIC in the driver, right-front passenger and rear 6 yo child in 35 mph NCAP tests with significance of differences between the rear child and the driver or right-front passenger.

Discussion

The 13 frontal NCAP tests in 1979 showed extremely poor restraint of the lap-belted child dummy in the rear seat. The kinematics demonstrated the adverse effects of no shoulder belt, including submarining and high-speed head contact on the vehicle interior or the dummy’s lower extremities. NHTSA publications in the 1980s-1990s focused solely on the driver and right-front passenger; they neglected to mention or publish a report on the poor, rear-seat lap belt performance. The following is a review of the literature on the effectiveness of rear seatbelts and provides context to the deliberation of rear seat restraints in the 1980s.

Lap belt effectiveness

Flamboe (1975) at NHTSA used MDAI (Multidisciplinary Accident Investigation) data to determine the effectiveness of lap belt use by rear seat occupants. He found the risk (±se) for severe injury (MAIS 4+) was 8.09 ± 0.90% (n = 890) for unrestrained and 3.75 ± 2.99% (n = 80) for lap belted rear occupants. This gave a 53.6% effectiveness for lap belts. The difference was statistically significant. He concluded unrestrained rear occupants were more likely to be severely injured than restrained occupants. He noted that about 47% rear passenger injuries were due to the lap belt, including abdominal injuries from submarining.

Huelke and Lawson (1975) evaluate field data collected near the University of Michigan. They found the risk (±se) for severe injury (MAIS 4+) to rear occupants was 10.92 ± 1.10% (n = 806) for unrestrained and 11.27 ± 3.71% (n = 71) for lap belted rear occupants. The effectiveness was −3.2%, disbenefit of lap belts, although the difference was not statistically significant.

Evans (1986, 1987, 1988) used 1975–1985 FARS (Fatal Accident Reporting System) data on adults to estimate the effectiveness of rear lap belts in preventing fatalities. There was not enough field data to study children. Evans (1986) used the "double pair comparison method" and estimated an overall fatality reduction of 18 ± 9% for rear lap belts. However, the effectiveness was −9 ± 17%, a disbenefit on average in frontal crashes.

Smith (1987) from NTSB summarized other field studies on the effectiveness of lap-belts for rear occupants. The NTSB investigated 22 predominantly frontal crashes involving unrestrained and lap belted rear occupants (NTSB 1986). The data lead the NTSB to recommend rear seat lap-shoulder belts for new vehicles and retrofit lap-shoulder assemblies for existing models (Goldman 1986). GAO (1987) concurred. Campbell (1986, 1987) claimed the NTSB study dismissed the existing evidence of lap belt effectiveness and the NTSB data had shortcomings.

Rear seat child restraints

In the late 1970s, there was considerable research on the methods to evaluate the performance of child safety seats used in rear seats (Backaitis et al. 1975; Kelleher and Walsh 1978, 1979). Frontal sled were conducted at 20, 25 and 30 mph with child dummies. Head excursion was used to assess the restraint provided by child seats. Some of the testing used a lap belt only to show the increased excursion of the head without torso restraint by the shoulder belts in the child seat.

Recent NCAP focus on rear-seat restraints

Kuppa et al. (2005) discussed NHTSA research on rear seats in frontal crashes. They noted a paucity of data on the performance of rear seat occupant restraint. They said “a research program was initiated to better understand rear seat restraint performance.” This was nearly 25 years after the 1979 NCAP tests with rear occupant. The research included examining real world data in NASS-CDS (National Automotive Sampling System-Crashworthiness Data System) and analyzing FARS (Fatality Analysis Reporting System) using a double-paired comparison. They conducted frontal crash tests with a 6 yo, 5th and 50th Hybrid III lap-shoulder belted in outboard rear seats and the center seat where the lap-shoulder belts were integrated in the seat.

Kuppa et al. (2005) noted that prior to 1989, only lap belts were required in rear outboard seating positions. Rear seat outboard lap-shoulder belts were required in passenger cars after December 1989 and in convertible passenger cars, light trucks, vans and sport utility vehicles after September 1991. NHTSA published a final rule in December 2004, requiring lap-shoulder belts for each designated rear seating position in a passenger vehicle, pursuant to Anton’s Law passed by Congress in 2002. There have been a number of studies on rear seat restraints (Kawaguchi et al. 2003; Mizuno et al. 2007; Esfahani and Digges 2009).

Tylko and Dalmotas (2005) and Tylko et al. (2007) discussed Transport Canada research on occupant restraint in rear seats, including crash testing with the 6 yo, 10 yo and 5th Hybrid III restrained by lap-shoulder belts in outboard rear seats. Kent et al. (2007) worked with NHTSA and focused on load-limiting the shoulder belt pretensioning to lower responses in rear seat. Pretensioning of the shoulder belt in rear seats can cause submarining with a flexible webbing tether from the latch to the anchor on the floor. The upward pull of the shoulder belt can lift the latch causing submarining (Viano, Parenteau 2019). Park et al. (2011) conducted 95 frontal NCAP tests with child dummies in child restraint systems (CRS) in the rear seat. They used a 12 mo (month old), 3 yo and 6 yo restrained in a CRS or booster seat. Hong et al. (2008) conducted 28 frontal NCAP crash tests with a 10 yo Hybrid III in the rear seat. Various vehicles were tested to assess the effect of a booster seat and pretensioning. Tylko and Bussières (2012) conducted barrier crash tests at 40, 48 and 56 km/h to investigate the responses of belt restrained 10 yo Hybrid III in the rear seat with and without a booster seat. Today, frontal NCAP tests often include an occupant in the rear seat.

Rear seat performance in the 1979 NCAP tests

The child dummy in the 1979 NCAP tests experienced severe head impacts and often submarined the lap belt. The child dummy moved forward, jackknifed around the lap belt and their head impacted the vehicle floor, interior or their lower extremities. Clarke et al. (1970) demonstrated the risk of fatal head impact on the floor with lap-belt restraint of an anesthetized animal in frontal sled tests. They reported a tolerance (LD₅₀) of 32 ± 4 g deceleration. The most severe injuries were atlanto-occipital avulsion and cervical fracture-dislocation with spinal cord transection. The tests were sponsored, in part, by the Department of Transportation (DOT).

No publication of the rear seat NCAP test results has been found; NHTSA never mentioned the poor rear seat performance. NHTSA did not inform the public of the extremely poor restraint performance, even during the public discussions on the 1986 NTSB recommendation that U.S. vehicle manufacturers install lap-shoulder belts in rear outboard seats. After the 1979 NCAP tests, NHTSA removed the rear-seat dummy from frontal NCAP tests for nearly 25 years, except for some child-seat testing.

The 1979 NCAP tests showed the lap belt was not sufficient to restrain the child in any vehicle tested at 35 mph. It must have been obvious to NHTSA that poor performance would occur in subsequent NCAP tests, irrespective of the vehicle tested or the dummy used. By removing the dummy from the rear, NHTSA avoided the issue that rear lap belts were not adequate to restrain children in 35 mph frontal impacts. It would have been a false argument if NHTSA claimed they did not report the rear-seat results because of biofidelity issues with the child dummy. The Alderson VIP 6 C was adequate to demonstrate the extremely poor performance; field accidents investigated by NTSB found serious injury with rear lap belts in frontal crashes (NTSB 1986).

Acknowledgments

The author conducted occupant protection and biomechanics research at the General Motors Research Laboratories during the 1979 frontal NCAP testing. He was not aware of the 1979 NCAP tests with a rear child until 2023. In the 1980s, he managed research on occupant restraints and injury biomechanics in GM. In 1983, he was the general chairman of a SAE conference on Child Injury and Restraint (SAE Publication P-135). Nothing was presented that had the importance of these NCAP tests.

In 1984 NTSB started investigating the crash performance of seatbelts that included occupants in the front and seat seats. A number of frontal crashes involved serious injury with a lap belted rear occupant (NTSB 1986). Goldman (1986) sent a letter to U.S. vehicle manufacturers summarizing 26 cases of lap belt injury where NTSB concluded the lap belt performance was poor. This started a flurry of discussions.

In 1987, the author and Mr. Stanely H Backaitis from NHTSA were the general chairman of a SAE conference on Restraint Technologies: Rear Seat Occupant Protection, (SAE Publication, SP-691). Several of the references in this study are to SAE papers presented at the 1987 conference (Campbell 1987; Evans 1987; Smith 1987). The focus was on effectiveness of rear seatbelts and designs of lap-shoulder belts for rear occupants. NHTSA did not mention the 1979 NCAP tests with a rear lap-belted child.

Dr. Claude H Tarriere from the Physiological and Biomechanical Laboratory, Renault-Peugeot Association, 92250 LaGarenne-Colombes, France showed the author movies of two frontal sled tests with child cadavers using a lap belt in the rear seat. The kinematics were extreme with the child moving forward, jackknifing around the lap belt and experiencing severe head impact on the floor or front edge of the rear seat. In retrospect, the kinematics were similar to what was seen in the 1979 NCAP tests.

The author was a proponent of outboard rear lap-shoulder belts. In the mid-1980s, discussions in GM were difficult with a number of engineers and lawyers pushing the status quo of rear lap belts. The field data generally showed lap belts were effective in preventing fatal injury compared to being unbelted. There was limited data from Europe on the effectiveness of rear lap-shoulder belts. The discussions in GM were compounded by a concern that changing to rear lap-shoulder belts would mean they were better than lap belts, which would be used in product litigation as an admission that lap belts were not adequate.

Evans (1986) reported in GM a field accident study estimating the effectiveness of rear lap belts in preventing fatalities. He used the “double pair comparison method" to estimate an overall fatality reduction of 18 ± 9% for rear lap belts. However, in frontal crashes, the effectiveness was −9 ± 17%.

The author’s efforts in GM were bolstered with a copy of Dr. Tarriere’s sled tests showing the extreme kinematics with high-speed head impact on the floor. He showed the movies and noted Evans’ study pointing to a disbenefit of lap belts in high-speed frontal crashes. GM approved the installation of lap-shoulder belts in outboard rear seats beginning with some 1987 MY vehicles and the entire fleet by the 1990 MY. Ford Motor Company made a similar decision. GM also provided retrofit kits to install lap-shoulder belts in rear seats (Smith 1987). The anchorages for lap-shoulder belts were already available in the outboard rear seats.

The debate for rear lap-shoulder belts was a painful experience for the author in GM. His efforts would have been much simpler if the 1979 NCAP tests had been publicized by NHTSA. The 6 yo child dummy responses and movies from the 1979 NCAP tests were self-explanatory. They showed the lap belt was not sufficient in 35 mph frontal barrier crashes.

NHTSA was irresponsible for not publicizing the extremely poor results with the lap-belted child in the 1979 NCAP tests. Essentially all of NHTSA’s efforts focused on airbags to address driver impacts on the steering wheel and right-front passenger impacts on the interior.

NHTSA ignored the extremely poor performance of the lap belt in the 1979 NCAP tests and failed to inform vehicle manufacturers, NTSB or the public, even during the public discussions on the 1986 NTSB recommendation that U.S. vehicle manufacturers install lap-shoulder belts in rear outboard seats. None of the subsequent NCAP tests included a belted child or adult in the rear seat until nearly 25 years later.

The author thanks Dr. Kennerly Digges, retired NHTSA, Dr. Thomas Hollowell, retired NHTSA, and Mr. Stephen Summers, NHTSA, for discussions on the background of NCAP and the NHTSA crash test database. Their perspective was helpful. The author also thanks the helpful comments on the manuscript by Matthew P Reed, UMTRI, University of Michigan.

The tests reported here were conducted by NHTSA in 1979. The reports and data used English units. For consistency, English units are used here as would have occurred if this data was reported in the 1980s. The reader can easily convert to metric units.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The crash test movies, data and reports are available from the NHTSA website (https://www.nhtsa.gov/research-data/research-testing-databases#/) using the test number (#) as a guide. Additional information on the data analysis can be requested of the author at [email protected].

Additional information

Funding

The author(s) reported there is no funding associated with the work featured in this article.