Final Report
School Bus Safety Belt Study
Chapter 2

New Jersey Department of Law and Public Safety
Office of Highway Traffic Safety

New Jersey Institute of Technology
Center of Transportation Studies and Research
December 1989

Evidence on the desirability of using seat belts on school buses that can be derived from the actual crash testing of vehicles is rather limited. Over the past twenty years there have been only five tests that replicated a small sub-set of the total possible accidents that can occur in real situations on the road. Four of these tests are discussed in this chapter. The fifth, whose purpose was to evaluate alternative seat and restraining systems is presented in Chapter 7. Bus crash testing started with two series of tests that were conducted at the University of California at Los Angeles (UCLA) in 1967 and 1972. Transport Canada (the Canadian equivalent of our U.S. Department of Transportation) conducted bus crash tests in 1984, and the latest testing was done in 1985 in this country, and it actually involved one Thomas Built school bus. The only known sled crash test was conducted by the National Highway Traffic Safety Administration (NHTSA) in 1976. The purpose of this chapter is to give a brief account of all four tests, including a brief description of the tests, the conclusions drawn by the teams that conducted them, and the criticisms of the testing methods and results that appeared in the literature. Finally, a synthesis of all aspects of this body of knowledge is made in order to draw conclusions on whether or not seat belts should be required on school buses.

UCLA Crash Tests

Researchers at UCLA conducted two series of crash tests involving a variety of school bus collision modes. The purpose of the tests was to provide specific and practical recommendations and solutions to agencies responsible for school bus safety. The study had a variety of objectives. It was intended to evaluate a number of school bus seat types as well as passenger restraint systems and answer questions about their proper design, construction and installation.

Series I Tests - 1967
The 1967 collision experiments at UCLA duplicated three types of collisions, a head-on, a rear-end, and a right-angle collision. Although the same type school bus was struck in all three cases, the types of vehicles used to strike the school bus were different, and the passenger restraint types, seat types, and passenger sizes studied were numerous [5].

Experimental Features
For the head-on collision experiment, a 1965 GMC-Superior school bus (weighing 17,500 lbs.) travelling at 30 mph was struck squarely head-on by a 1944 Mac-Superior school bus (weighing 7,500 lbs.) travelling at the same speed. To simulate a rear-end collision, the same type (1965 GMC model) school bus, while stationary, was squarely impacted in the rear by a 1960 plymouth 4-door sedan (weighing 4,400 lbs.) travelling at 60 mph. In the third experiment, a 1966 chevrolet 4-door sedan of 4,500 lbs. Travelling at 60 mph struck head-on the side at the rear wheel position of the same type (1965 GMC model) school bus while at a stationary position.

The restraint types tested in the experiment were:

• i) no restraint
  

• ii) lap belt only
  

• iii) lap belt and diagonal shoulder strap
  

• iv) air bag
  

• v) restraint bar
  

A variety of different seats (11 types) were tested in the experiments. They ranged from the conventional standard seats used in school buses at the time, to other variations such as seats with high backs, seats with hand rails or armrests, inflated air bag seats, united airlines siesta seats, etc. To simulate passenger type variations, anthropometric dummies corresponding to ages of 3-, 6-, 13-year olds, and adult were used in the school bus.

The instrumentation involved a large number of transducers and specialized photographic systems strategically positioned, and other provisions were made for scientific observations. The categories of data recorded included a wide range of interacting factors. The experimental study collected data on the kinematics of passengers, forces sustained by passengers, loadings on restraint system, relative injury exposure for passengers in different seating arrangements under the same collision circumstances, vehicle collision dynamics, and vehicle structural performance.

Test Findings and Conclusions
The UCLA researchers pointed out that for the head-on collision test, the seat back height ahead of the passenger was the primary consideration because it generally was the object initially contacted by both the unrestrained and partially restrained passengers. For the rear-end collision test, the seat back height was the principal variable governing the occurrence of whiplash injury. Side impact collisions usually force passengers into direct contact with compartment structures or side window glass. Therefore, for the side-impact experiment, the presence or absence of restraint systems and arm rests represented the most important consideration. These observations are certainly in conformity with conventional wisdom. One has also to keep in mind, that these tests were conducted before the current standards on seat back height became effective. The standard school bus seat at the time had a low back with a metal rail along the entire length of the back's top.

On the basis of data collected during the 1967 crash tests, the conclusions drawn by the UCLA team pertaining to lap seat belts on school buses were the following:

The greatest single contribution to school bus collision safety is a high strength and high-back (28 in.) safety seat. High back seats (28 in. or more) greatly contribute to the compartmentalization of passengers, thereby reducing the chances of injuries sustained by passengers being hurled against one another, regardless of their size. Next in importance is the use of a three point belt, a lap belt or other form of effective restraint. Seat belts are recommended for use with safety seats.

Lap type safety belts would provide substantial additional protection to school bus passengers in high-back seats that have efficient padding on the rear panel of its back rests.

Lap belts should not be used for low seat back units because their use substantially increases the highly adverse forces to the spinal column resulting from whiplash and they virtually assure severe head or neck impacts with low backrests ahead.

In the absence of armrests, the lap belt does provide some hip restraint against sideward movement, thereby reducing forces that a displaced passenger may apply to a companion seated beside him during a side impact collision.

In summary, the first UCLA series of tests concluded that a high seat back (28 in.) is the most important safety feature that can be added to school buses, and seat belts would provide substantial additional safety when used with high-back padded seats. But, seat belts should not be used in combination with low-back seats. However, these recommendations should be carefully analyzed together with the results of the Series II UCLA experiments, which provided additional information.

Series II Tests - 1972
The second series of experiments at UCLA simulated two types of full scale collisions, a head-on and a right-angle, both involving the same type of school bus [6].

Experimental Features
For the head-on collision a 1969 Superior School bus (60 passenger) travelling at 30 mph was struck squarely head-on by a 1962 International 2-ton dump truck which was also travelling at 30 mph. For the second test involving a side impact collision, the same type school bus, while stationary, was impacted at its side by a 1967 Ford 4-door sedan travelling at 60 mph.

The school bus seat types, restraints, passengers simulated by anthropometric dummies, and data collection techniques were similar to those of the Series I experiments.

Test Findings and Conclusions
On the basis of data collected from the Series II crash tests, the major conclusion drawn by the UCLA team pertaining to lap seat belts on school buses was:

Seat belts are not recommended for school buses having conventional seats with hard surface, weakly structured frames, lack of side-force restraint (padded armrests at the aisle) and grossly inadequate backrest height.

In other words, the Series II experiments confirmed the findings of the Series I experiments to the extent that the safety performance of lap belts is unacceptable in conventional school buses (i.e., pre-1977 standard school buses). However, to analyze the overall safety performance of lap belts, the following additional findings of the Series II experiments are of significance:

The average size school child (13 years old) would sustain smaller head impact forces (44 g versus 67 g) if left unbelted than if lap-belted, provided that he was protected by a 28-inch high energy absorbing, UCLA-design seat back.

For side impact exposure, the UCLA padded armrest side restraint appeared to provide passenger protection as effectively as full use of lap belt restraints.

For buses provided with safety seats having a performance profile comparable to the UCLA design, seat belts with contribute significantly to improved safety, especially during severe upset collision exposures, provided that extrication procedures can be perfected to allow the rapid evacuation of a fully loaded, overturned bus, (i.e., removing 40 to 60 children hanging upside down suspended by their seat-belts). However, when safety seats are used, the researchers regarded further restraint measures, such as the installation of safety belts, of minor importance, because of the special protection afforded to school buses by their size and visibility.

The UCLA collision researchers have always advocated strongly the continuous use of lap-type safety belts in passenger vehicles on all occasions. However, they modified their views when school buses are concerned, because school bus seats are designed differently and are positioned close together making the use of lap belts highly inadvisable unless seat structures are designed, installed and spaced in a manner compatible with the use of lap belts.

In summary, the UCLA researches concluded after both Series of tests that taking into consideration the factor of special protection provided to the school bus by its size and visibility, the addition of seat-belts would be of minor importance when safety seats (28-in. high padded seat back) are used. Furthermore, if school buses are equipped with safety seats, the most likely contribution to safety that seat belts will provide during three of the most commonly occurring accidents will be as follows:

Head-On Collisions: Seat belts will not provide a significant safety enhancement. They will rather expose the passengers to a higher head acceleration level (67 g).

Side Impact Collisions: Seat belts, if fully used, will provide equal passenger protection as armrests.

Rollover Upset Collisions: Seat belts will be of substantial assistance, provided that appropriate evacuation procedures exist.

Transport Canada Crash Tests

Transport Canada conducted crash tests on three school buses in 1984 [3]. The purpose of the Transport Canada testing program was to determine the reaction of the belted and unbelted test-dummies in small and large school buses during frontal barrier collisions. The tests were intended to assess the effect that seat belts would have on school bus passengers, and to determine whether or not school bus standards provide the anticipated level of occupant protection. One has to keep in mind that in the time that elapsed between the UCLA tests and the Transport Canada tests, standards that became effective in the USA adopted the use of the safety seat recommended by the UCLA tests. In addition, Type II buses were also required to be equipped with seat belts.

Experimental Features
Bus types: Three different buses were used for the Transport Canada tests. Two of the buses consisted of a mid-sized 22-passenger Thomas Minotour, and a small 20-passenger Camp Wagon van conversion. The Gross Vehicle Weight Ratings (GVWR) of these two buses were 10,000 lbs. or less and they are equivalent to Type II buses. The third bus was a Type 1 bus, a 66-passenger, 1984 Blue Bird with a GVWR of 25,000 lbs.

Collision type: Each of the buses was subjected to a barrier frontal collision at 48 km/h (approximately 30 mph). The resulted impact forces were comparable to those resulting from a head-on-collision between the school bus and a car travelling at highway speed.

Anthropometric dummies: Each of the buses contained six 5th percentile female dummies representing large elementary school students. In addition two anthropometric dummies representing 6-year olds were used in the Type 1 bus only.

Seat spacing and instrumentation: There were 11 rows of seats in the bus. The 1st, 6th and 11th rows were used for the test. The seat spacing was 21 inches (maximum seat spacing allowed in Canada) for row 1, 27 inches (a spacing that would counter the compartmentalization concept for passive passenger protection) for row 6, and 24 inches for row 22 (the current standard in the USA). All six 5th percentile dummies were instrumented to record forces on the head and chest, and three of these were also instrumented with femur load cells.

Three dummies were lap belted and the remaining three were left unbelted. The dummies were placed in pairs, with each seat occupied by one restrained and one unrestrained dummy for each seat spacing arrangement. The two dummies representing 6-year olds that were used in the Type I bus only, were not instrumented or restrained. High speed cameras recorded the motion of dummies during collisions.

Head injury criteria (HIC) and chest acceleration rates were measured for the instrumented dummies. An HIC value of 1,000 and a chest acceleration rate of 60g were used as threshold values above which serious injury or death could be assumed to occur. Femur loads in excess of 2,250 lb. are unacceptable in U.S. and Canadian standards.

Test Findings and Conclusions
A Summary of the Transport Canada test results is presented in Table 2. Four basic observations can easily be made from the data of this Table.

Belted dummies experienced higher head and generally lower chest accelerations than did the unbelted ones.

In both small buses (Thomas Minotour and Camp Wagon) the lap belted dummies experienced HIC values in excess of 1,000, whereas the unrestrained dummies experienced HIC values less than 1,000. In the large (Blue Bird) bus all dummies (lap belted or unbelted) experienced HIC values much less than 1,000. But, the lap belted ones experienced HIC values approximately three times greater than those for the unrestrained ones. According to the study authors these differences can be attributed to "the fact that the restrained dummies' heads struck the seat backs in a manner that did not permit efficient energy absorption by those seat backs. Sharp peaks in head acceleration traces indicate that the dummy heads compressed the seat back padding to such a degree that they 'bottomed out' on the steel structure underlying the padding in the seat back".

In all cases (lap belted or unrestrained) except one, the chest accelerations were less than the life threatening maximum of 60 g. The single exception exceeded the criterion by only 0.4 g. For the Blue Bird (large) and Thomas (small) buses the unbelted dummies experienced higher chest acceleration than did the comparable belted dummies.

Femur loads are not shown in Table 2, but measurements on three dummies (one belted and two unbelted) were below the limit of 2,500 lbs.

Belted dummies experienced more severe neck extensions than did the unbelted ones due to the angle at which they struck the seat ahead of them. The neck extension of several belted dummies was judged to be life threatening.

The major conclusions of the Transport Canada tests pertaining to seat belt restraints and compartmentalization in frontal collisions are as follows:

The passive occupant protection of the seating system (known as compartmentalization), required by Federal regulations since 1980 (1977 in the U.S.A.) functions as intended during frontal impacts and provides excellent protection for occupants.

The use of lap belts in any of the three tested sizes of recent model school buses may result in more severe head and neck injuries for a belted occupant than for an unbelted one, in a severe frontal collision.

Criticisms of the Canadian Tests
University of Michigan and Rochester researchers, NHTSA, and professional groups criticized the validity of the Transport Canada tests on a variety of grounds.

The University of Michigan critics [7] concentrated on the following four aspects of the Transport Canada tests:

There are some questions "as to whether or not a HIC value of 1,000 is a conclusive measure of serious head injury, particularly for children". Although higher HIC values were measured on the belted dummies, the highest value of 731 recorded on the Type I bus is much less than the 1,000 limit.

The restrained dummy heads contacted the padded seat back (which could have been better padded) resulting in higher HIC values. But, the unrestrained dummies hit the top of the seat backs with their neck, and there were no load cells or accelerators mounted on the necks to measure the resulting load. Thus, no reliable injury prediction could be made.

There is no biomechanical justification to the Canadian tests' inference that the "neck extension of several restrained dummies was judged to be life threatening". Humans bend differently from dummies with stiff neck and rigid torso as used in the Canadian tests, and do not tend to suffer "life threatening" neck injuries in the situations that the tests attempted to model.

The large Blue Bird bus was occupied by six 5th percentile female dummies, representing 14-year old junior high school students, and two six-year old size dummies (unrestrained and uninstrumented). Thus, the conclusion of the test, if valid, is limited in real word applications. Smaller dummies representing younger children should have been used also. One of the 6-year old dummies should have been restrained for comparison purposes.

University of Rochester Professor John D. States [8] in another criticism disagreed with the Transport Canada findings on the following two grounds:

The dummies used in the Canadian tests do not accurately model the flexibility of human spines. The additional stiffness of dummy spine prevented load sharing contact of the chest, head and upper extremities with the seat in front. This contact would have reduced the head acceleration and HIC values for the belted dummies.

The HIC value of 1,000 is not applicable to children. Considering the results of experimental studies on arterial vessel walls in the brain of children, and the flexible and elastic characteristics of children skulls, the HIC value for children is greater than 1,500 and possibly 2,000.

Dr. Yeager of the National Coalition for Seat Belts on School Buses [9] criticized the selection of the 5th percentile (representing a 14-year old) female dummy because of its height, which during impact targets the head on the seat back where padding narrowly covers the metal bar, thus causing higher HIC values. Furthermore, he questioned the use of a type 572 dummy because of its stiffness that produces excessive HIC readings.

Explaining its position on seat belts on school buses, The National Highway Traffic Safety Administration (NHTSA) made the following observation on the Transport Canada tests:

"It should be recognized that "compartmentalization" countermeasures were specifically designed to protect the occupant in frontal barrier tests, similar to those conducted by the Canadians. The low head injury readings for the unbelted dummies is indicative that compartmentalization performs as well in production buses as it did in the research tests which perfected the concept." [1]

In the same report, however, NHTSA pointed out some of the limitation of the Transport Canada tests by stating:

"In examining the Canadian tests, several factors must be considered. A 30 mph barrier crash force for a large bus is an unlikely occurrence. For example, a head on crash between a large bus and a full-size car, both travelling at 55 mph would be less severe to bus occupants than the 30 mph barrier test. Also, only one size dummy was used which typically represents a junior high school student. The geometry for younger children would be significantly different with likely different results. Taken together, the results of the Canadian tests should be viewed with caution."

Thomas Built Buses Crash Tests - 1985

Of the three tests conducted by Thomas Build Buses, Inc. only one (a right side impact) can be used for the purpose of comparison of both lap belted and unrestrained dummies. The other two tests (a head-on crash and a left side impact crash) either did not involve instrumented dummies or data were partially lost [10].

Experimental Features
The right side impact test involved a 16-passenger 1985 Thomas Minotour Bus (GWVR less than 10,000 lbs.). This vehicle, being a Type II school bus, is currently required to be equipped with lap belts. The bus was impacted from its right side by a barrier of 4,000 lbs. moving at 30.8 mph. There was a total of eight dummies in the vehicle. Six were instrumented 50th percentile dummies (two lap belted, and four unrestrained), the seventh was an uninstrumented 5th percentile dummy (lap belted), and the eighth an uninstrumented 5th percentile dummy (lap belted).

Test Findings and Conclusions
Of the six dummies that were instrumented, the two lap belted, and three of the four unrestrained dummies sustained non life threatening HIC (less than 1,000) and chest acceleration (less than 60 g) values. The remaining unrestrained dummy suffered a HIC of 67.5 and a chest acceleration of 97.5 g considered to be life threatening. Figure 2 contains a pictorial representation of the test.

On the basis of the obtained results, Thomas Built Buses, Inc. concluded:

"Compartmentalization works as it was designed to work in frontal or side impacts. These tests also indicate that in the case of side impact, there seems to be very little significant difference between the belted and unbelted dummies in these test conditions relating to head and chest injuries".

The National Transportation Safety Board (NTSB) reviewing the Thomas crash tests noted that ". the Thomas Built Buses crash tests provide an indication of what can be expected from a 30 mph side impact involving a school bus transporting both lap belted and unrestrained passengers. Since a belted dummy was seated next to an unbelted dummy during the test, the test results do not necessarily provide an indication of head or chest injuries to be expected if a small school bus transporting all lap belted passengers is involved in a side impact, nor for that matter, what to expect if all passengers are unrestrained". [11]

In a critique of Thomas Bus side impact crash tests, the National Coalition for Seat Belts on School Buses argued that "The unbelted dummies remained in the "compartment' because during the side collision they were thrown into the belted dummies. The belted dummies acted like padded side arms and helped keep the unbelted dummies in their seating area. Unfortunately, school buses don't have padded side arms to contain children during side collisions." [9] This statement refers to dummies in positions #4 and #6 on the bus (see Figure 2), which were thrown on their belted neighbors in positions #3 and #5.

NHTSA Sled Tests - 1978

The second best alternative to performing crash tests utilizing the entire vehicle, is the performance of crash tests utilizing only a group of seats mounted on a sled. Sled tests are conducted by fixing school bus passenger seats with specific spacing and restraint systems on an electro-mechanical movable sled mounted on a track. Anthropomorphic dummies (restrained and unrestrained) are placed on the seats which are fixed with respect to the sled. The sled is then rapidly accelerated or decelerated on the track, and the resulting forces exerted on various body parts of the dummies are measured. Obviously, lateral impacts cannot be model with sled testing since the vehicle body is not present and observations of dummies hitting the side walls or other interior features cannot be made.

Experimental Features
The NHTSA sled tests were conducted to determine the response of dummies in simulated frontal collisions with and without lap belts. Five tests (test nos. 37, 38, 39, 40, 41, NHTSA) are considered here as each of these tests involved both belted and unbelted school bus passenger dummies, thus providing an opportunity to compare the performance of both belted and unbelted school bus occupants. In these tests, standard school bus seats were subjected to simulated frontal collisions at 15 mph (for tests 37, 38, 39 and 40), and 20 mph (for test no.41). In each of these tests, the responses of four 50th percentile, male anthropomorphic dummies (two lap belted, two unrestrained) were compared. The school bus seat spacing remained constant at 20 inches. The front seat was empty, the center seat held two lap belted dummies, and the two dummies at the back seat were unrestrained [12].

Test Findings and Conclusions
The test data for all five NHTSA sled crashes are summarized in Table 3. On the basis of the measurements obtained during the sled tests on the 50th percentile dummies, the following major observations were made by the NHTSA researchers:

The use of lap belts do not reduce peak head accelerations but in fact, in most cases, actually cause an increase in peak accelerations. This may be probably due to the fact that the head contact point is moving higher up on the dummy head with the use of lap belts. It may also be due to the redirection of the head impact into the stiff axis of the seat back structure.

The effect of seat belts on head acceleration and torso response appear to be significant. However, on the basis of knee response evaluations, the use of lap belts has a slight to moderate influence on decreasing femur loads.

HIC and chest acceleration values for both belted and unbelted dummies were within acceptable limits (HIC less than 1,000, chest acceleration less than 60 g).

Passenger containment increases with increasing seat back height, and there are no significant additional benefits that can be obtained by using lap belts. However, during rebound, the use of lap belts seems to have a positive influence on containment.

All seats (with belted or unbelted dummies) appeared to satisfy the established injury criteria for femur loads (i.e., less than or equal to 1,700 lbs. For an impact speed of 20 mph.).

In effect the NHTSA study results suggest that for frontal collisions, lap belts do not appear to have a significant effect on the response characteristics of a 50th percentile adult dummy and as a consequence, they cannot make significant contributions to increased occupant safety. The use of lap belts increased peak head acceleration for most of the belted dummies, seemed to have a positive influence on containment during rebound, and a slight to moderate influence on decreasing femur loads.

NHTSA researchers also reported that they were unable to evaluate the potential for serious neck and spine injuries on the unbelted dummies. Film documentation of the sled tests revealed that there may be other potentially harmful body loadings that were not covered by the instrumentation of the tests. One such case involves unbelted dummies impacting the seat back on their throat. It was also observed that an unbelted child dummy having stiff knee padding was stopped abruptly allowing the torso to rotate until the head made contact with the seat back. Then a violent whipping was set in the dummy's spine as it attempted to "beam" the inertial loads of the torso to the knee and head contact points. It is not known if this "whipping action" is unique to the dummy structure or it represents evidence of a real serious injury problem. There are no existing injury criteria to cover these potential injury modes, noted the NHTSA researchers.

Review of Bus and Sled Crash Tests
The series I crash tests at UCLA recommended the use of seat belts in combination with padded high-back safety seats. The series II crash tests concluded that belted passengers were subjected to higher head impact forces than their unbelted counterparts, but no serious injuries were predicted for the belted passengers. Both series I and series II tests confirmed that seat belts would provide additional protection to passengers in side impact and rollover accidents. However, considering the special protection afforded to school buses by their visibility and size, the addition of seat belts would be of minor importance. The fact is that even with the additional benefit of their greater visibility and size, school buses do get involved in accidents in all impact modes. Therefore, the evidence derived from the UCLA crash tests that seat belts are beneficial in side impacts and rollovers is important in terms of the overall school bus passenger safety improvement.

The Transport Canada crash tests and the NHTSA sled tests reported higher HIC values for belted passengers. But, the results of those two tests are not consistent in terms of their prediction of potential injuries because of the presence of seat belts. Whereas the HIC values of the NHTSA sled tests did not even reach half the maximum acceptable limit of 1,000, the HIC values measured by the Transport Canada crash tests exceeded 2,000 for Type II buses, but were still well below 1,000 for Type I buses. This should not be surprising, since the Canadian buses were crashed at roughly double the speed of the NHTSA sleds. There has been a considerable amount of criticism of the high speeds, as well as the instrumentation and test dummies used in the Transport Canada tests. Researchers also argued that a maximum acceptable HIC value of much higher than 1,000 might be applicable for children. Experts' reviews of the Canadian tests noted that the results should be viewed with caution.

The side impact crash test conducted by Thomas Built Buses Inc. was criticized for the inappropriate positioning of belted and unbelted passengers. Thus, the validity of this crash test is also questionable.

On the basis of the conclusions and reviews of the bus and sled crash test results, the authors of this report have determined that seat belts may not be beneficial in frontal impacts, but in side impacts and rollover accidents they would provide significant additional protection to school bus passengers. Therefore, the decision to require the installation of seat belts on all school buses should consider both the increase of injury potential in frontal impacts and the reduction of injuries in side impact and rollover accidents. This trade-off will be quantified in Chapter 8 of this report.