Traffi c Safety Facts Research Note December 2003 DOT HS 809 692 Examination of Rollover Crash Mechanisms and Occupant Outcomes Ana María Eigen *

Objective In 2003, NHTSA has declared crush testing (Federal Highway In response to the growing the issue of passenger vehicle Administration, Test 03005 concern surrounding rollover rollover a priority, owing to A&B, 2003) in which each roof incidence and its consequences, its prominence in fatal crash strike appeared to weaken the the present analysis seeks to statistics. The agency has been loading capacity of the roof. answer the following questions: active in the implementation The examination of the most ! Which are the most of the Rollover Resistance frequently occurring types prevalent rollover crash Rating for new vehicles and in of rollover crashes and their attributes? the development of a dynamic associated injury outcomes ! What is the injury outcome rollover test. in relation to quarter turns, for rollover crash occupants? as reported in real-world data While, the number of quarter bases, is proposed to further Background turns is not always accepted as understand rollover crashes. A void in real-world rollover data a measure of rollover severity, compilation was fi lled in 2002 with it has been used in prior works Introduction the release of “Characteristics of such as Rains and Kanianthra From 1995 through 2001, an Fatal Rollover Crashes” (DOT HS (SAE 1995.) In this analysis, it estimated 17.1 million (weighted) 809 438.) Using the Fatality will be used as a comparative crashes, involving passenger Analysis Reporting System measure aimed at discerning vehicles towed from the scene (FARS), DOT HS 809 438 rollover crash attributes and due to damage sustained in the examined an important segment associated injury vulnerabilities. crashes, occurred on roadways of the rollover population. It was Based on previous reporting, in the United States. Of these shown in the referenced report the unbelted occupant is the 17.1 million crashes, 4.5 million that, generally, those without most vulnerable to ejection and were single vehicle events. restraint and subject to ejection fatality; however, even the belted Approximately 28 percent of are the most vulnerable. occupant is at risk because these single vehicle crashes some current seatbelts, and resulted in rollovers, as reported This Research Note seeks to most retractors, are primarily in the NASS – CDS. Among all expand on the fi ndings of the DOT designed to withstand the rollover crashes (1.6 million), HS 809 438 by using the National exigencies imposed by a planar 81 percent were single vehicle Automotive Sampling System crash. Further, the complexity encounters. (NASS) – Crashworthiness Data of the rollover crash suggests System (CDS.) Combining the that vehicle integrity, in particular Among rollover crashes reported crash investigation elements roof strength, in conjunction in NASS - CDS, approximately with the injury reporting inherent with, restraint use must provide 55 percent ended in one or two to this data base, a more exact adequate protection to minimize quarter turns contacting either description of the crash is occupant injury. The concern the near side of the roll or the possible. is accentuated by recent roof roof. Generally, these crashes

*Ana María Eigen is a Mechanical Engineer with the Mathematical Analysis Division of the National Center for Statistics and Analysis, NHTSA. National Center for Statistics and Analysis 400 Seventh St., S.W., Washington, D.C. 20590 occurred when the vehicle left Table 1 the roadway and encountered a Percentage of Vehicles in Single Vehicle Rollover by Body Type tripping mechanism such as soft dirt or loose gravel. In contrast, it and Type of Rollover (Sample sizes given in parentheses. Total column given in was noted that untripped rollover thousands, weighted) crashes comprised only a small fraction of all rollover crashes 1 2 3+ Unknown and were differentiated from the Vehicle Quarter Quarter Quarter End- Num. of (in tripped cases by the absence of Body Turn Turn Turn over- Quarter ‘000) any tripping mechanism. Type (Lateral) (Lateral) (Lateral) End Turns Total Pass. 13 % 42 % 40 % 1 % 4 % 632 Under the circumstances Car (n=144) (n=499) (n=719) (n=43) (n=107) 50 % encountered during a rollover SUV 13 % 41 % 40 % 0 % 6 % 330 crash, occupant vulnerability (n=108) (n=127) (n=384) (n=6) (n=58) 26 % increased with the omission of restraint use. For those Van 31 % 22 % 42 % 1 % 3 % 44 completely ejected, 97 percent (n=38) (n=25) (n=78) (n=2) (n=7) 3 % were unrestrained. Among Pickup 22 % 35 % 39 % 0 % 4 % 261 injured occupants, the proportion (n=95) (n=134) (n=328) (n=10) (n=32) 21 % of serious injuries was greater Total 16 % 39 % 40 % 1 % 5 % 1,268 for those involved in rollover (n=385) (n=785) (n=1,509) (n=61) (n=204) 100 % crashes. Note: Slight differences may exist in percentage calculations owing to rounding. Defi nitions and Constraints Source: NCSA, NHTSA, NASS-CDS,1995-2001 In order to amplify the scope of rollover crash understanding, experienced 50 percent of single if the rollover initiation type has the NASS – CDS was queried. vehicle rollovers. Polk vehicle been identifi ed as a fall-over, CDS is a nationally representative registration data aggregated over neither tripped nor untripped, sample of tow-away crashes 1995 through 2001, yielding or turn-over, an untripped based on Police Accident vehicle years, indicated that 65 situation, rollover initiation object Reports (PAR). Upon selection, a percent of the 1.54 billion vehicle aggregates these attributes. It comprehensive crash scene and years belonged to passenger should be noted that these are vehicle review combines with the cars. The remaining fi fty percent very different types of rollover occupant demography and injury of the single vehicle rollover crashes. Further, the rollover account to form the case fi le. crashes were attributable to light initiation type fall-over constitutes trucks; however, this vehicle class most of the entries found in this Passenger vehicle (passenger comprised 35 percent of the aggregated attribute. Please car, sport utility vehicle, van, and vehicle fl eet. The vehicle years reference the Section “Rollover pickup truck) crashes reported in were aggregated in response Initiation Type” for disaggregation CDS from 1995 through 2001 to the small CDS sample size. of these two attributes. were considered in this study. It was not dictated by the fl eet composition, which experienced Among multi-vehicle rollover Rollover Crash Mechanism rapid changes over this period. crashes, the vehicle is the Percentage of Rollover Crashes, by Crash predominant tripping mechanism, Confi guration and Fleet Composition Rollover Initiation Object accounting for 47 percent. Thirty- The predominant tripping four percent of multi-vehicle From 1995 through 2001, mechanism inducing rollovers in rollover crashes occur owing to single vehicle rollovers accounted single vehicle crashes tends to tripping on the ground. for 81 percent of all rollovers. As be ground (61 percent of single seen in Table 1, passenger cars vehicle rollover crashes.) In CDS,

National Center for Statistics and Analysis 400 Seventh St., S.W., Washington, D.C. 20590 Number of Quarter Turns Table 2 In the single vehicle rollover Percentage of Vehicles in Single Vehicle Rollover by Rollover Initiation crash, most vehicles come to Type and Type of Rollover rest at two quarter turns. This (Total column given in thousands, weighted) can be a promising sign once the rollover has occurred for several 1 2 3+ Unknown reasons. First, if the roof meets Rollover Quarter Quarter Quarter End- Num. of (in the required standard, it should Initiation Turn Turn Turn over- Quarter ‘000) withstand the mass of the vehicle. Type (Lateral) (Lateral) (Lateral) End Turns Total Second, the roof has not been debilitated with repeated strikes TRIP- 12% 43% 45% 0% 0% 562 since the vehicle undergoes two OVER 71% quarter turns and comes to rest. FLIP- 23% 47% 30% 0% 0% 44 Finally, since undue force has not OVER 6% been exerted during the fi rst roof TURN- 71% 19% 10% 0% 0% 27 strike the windshield will generally OVER 3% be intact. This is signifi cant CLIMB- 19% 29% 52% 0% 0% 9 because the windshield is an OVER 1% important structural element FALL- 19% 62% 19% 0% 0% 104 for roof integrity/strength, as OVER 13% shown by the work of Henderson BOUNCE- 32% 47% 21% 0% 0% 31 and Paine (1998.) Most of these OVER 4% benefi ts are severely curtailed COLLISION 6% 82% 12% 0% 0% 7 if the occupant omits restraint W/VEH 1% use. OTH ROLLOVER 2% 69% 28% 0% 0% 4 TYPE 0% Sixty-four percent of multi-vehicle END- 0% 0% 0% 0% 100% 4 rollovers last up to two quarter OVER-END 0% 128 357 304 0 4 793 turns, as compared to 55 Total percent in the single vehicle case. 16% 45% 38% 0% 0% 100% Since only 19 percent of rollover Note: Slight differences may exist in percentage calculations owing to rounding. crashes involve more than one Source: NCSA, NHTSA, NASS-CDS, 1998-2001 vehicle, fewer conclusions may be drawn than from the single are the vehicles that experience stopping and induced by contact vehicle scenario. Additionally, off-paved road excursion and with a curb, pothole, wheel rim/ single vehicle rollovers are very rejoin the paved roadway. Upon pavement contact, or pavement/ complex and unique events; the publication, CDS data sets are soil dug into by a vehicle’s wheels. presence of more vehicles, in not revised, therefore, the work At each of the quarter turn levels the multi-vehicle rollover case, of Hendriks (1998) and Hertz the trip-over is predominant. serves to further obfuscate the event and hamper reconstruction (1999) provide an update to The fall–over is the second most of the crash relevant to injury and the attributes relevant to this prevalent event occurring in 13 damage. variable. The summarization percent of the single vehicle below is limited to 1998 through rollover crashes. As defi ned by Rollover Initiation Type 2001. the CDS coding manual, in fall- over, the vehicle is tipped by a The results for this variable are included for completeness; The trip–over (tripped rollover) slope so that its center of gravity however, data prior to 1997 are is the rollover initiation type is outboard of its wheels. It should subject to recodifi cation based most commonly experienced (71 be noted that this defi nition is true on methodologies proposed by percent) during single vehicle for a vehicle not in motion. For a Hertz (1999). Of specifi c concern rollover crashes, per Table 2. moving vehicle, the dynamics and warranting additional study This scenario is enacted by a of the vehicle are controlled by sudden sideways slowing or gravity and centrifugal forces. National Center for Statistics and Analysis 400 Seventh St., S.W., Washington, D.C. 20590 Fifty-one percent of multi- turn (45 percent) indicating a Occupant Demography vehicle rollover crashes occur potentially lower severity crash. Owing to the small sample size as a result of the collision with The side plane refers to any side of occupants involved in rollover another vehicle. This indicates plane other than the wheels and crashes, the crash confi gurations that another vehicle rather tires that precipitates rollover. were collapsed into one data than a roadway element serves set. The attributes of interest to precipitate the event. This Behavioral Elements were on-scene disposition of the is followed by the trip-over in In reconstructing rollover crashes, occupant, restraint use, ejection, 38 percent of the multi-vehicle it is important to understand the and injury severity. rollovers. behavioral elements attributable to the event. Examination of the On-Scene Occupant Disposition Direction of Initial Rollover pre-event movement prior to The severity of rollover crashes Approximately, 45 percent of the recognition of critical event is suggested by the following vehicles roll to the right and and quarter turn frequency, comparisons. Fifty-six percent of 55 percent to the left in single with regard to driver inputs, as occupants receive some form vehicle crashes. A nearly equal described in CDS, are included. of treatment, as compared to proportion exists in multi- 40 percent of planar crash vehicular confi gurations. In single vehicle, non-planar occupants. Further, for rollovers events approximately 49 percent with more than two quarter Tripping Location on the Vehicle of the drivers sought to negotiate turns, the number of occupants For most rollover crashes, the a turn. Forty-one percent seem receiving no treatment wheels/tires are the point where to have failed to continue on their experiences a sharp decline from the initial tripping force occurs straight path. The remaining 24 to 11 percent. (67 percent for single vehicle pre-event movements were rollovers and 42 percent for seen to occur very infrequently Occupant Restraint Use multi-vehicle.) This also holds and included: an absence of The data supports the advocacy true regardless of the number driver, decelerating in traffi c for manual restraint use. of quarter turns. The non- lane, accelerating in traffi c lane, Although the Occupant Protection contact force is associated with starting in traffi c lane, stopped in Standard (FMVSS 208) does 17 percent of single vehicle traffi c lane, passing or overtaking not require a safety belt test rollovers including fall-over, which another vehicle, disabled or under rollover conditions, the occurs in 13 percent of single parked in travel lane, leaving safety belt does provide the vehicle rollover crashes. Non- a parking position, entering a intended result of retaining most contact forces describe a roll parking position, turning right, restrained occupants within the precipitated by centrifugal or turning left, making a U-turn, vehicle. By virtue of restraint gravitational forces. At three or backing up (other than for use, these occupants generally more quarter turns, roughly 57 parking position,) changing lanes, walk away from the rollover percent of tripping is attributable merging, successful avoidance sequence. Additionally, these to wheel/tire engagements. maneuver to a previous critical occupants experience crashes of Wheel rim contact is associated event, other movement, or higher severity with regard to the with tripped rollover while the tire unknown. recorded quarter turns and are force only is related to untripped not ejected. It is found that most rollover. In multi-vehicle, non-planar events, of the occupants classifi ed at 60 percent of the drivers failed the various treatment levels are Although 42 percent of the to continue in a straight path. restrained. multi-vehicle rollover crashes Sixteen percent were attempting are initially tripped at the a left turn movement prior to the It has long been theorized that wheels/tires, 35 percent are rollover event. These were noted risk takers generally succumb attributable to the side plane. as the highest frequency multi- to rollover. This may be at least Further, the side-plane incidents vehicle pre-event movements. partially supported by elevated are concentrated at one quarter incidence of unrestrained occ-

National Center for Statistics and Analysis 400 Seventh St., S.W., Washington, D.C. 20590 Table 3 have sacrifi ced potential injury Percentage of Occupants Involved in Crashes by Restraint Use and Number mitigation afforded by restraint of Quarter Turns (Totals given in thousands, weighted. Sample sizes given in use. The data suggested that parentheses.) the restrained occupants were able to withstand rollover without Quarter Turns being ejected at a greater (Column Percentage) End- Safety over- Unk (in ‘000’s) frequency than those who were Belt Use 0 1 2 3+ End Num Total unrestrained. Of the partially and completely ejected occupants, No 27% 35% 34% 34% 67% 51% 7,277 81 percent were unrestrained, (n=17,465) (n=544) (n=888) (n=1,783) (n=71) (n=232) 28% as compared to the 18 percent Yes 71% 64% 65% 66% 32% 48% 18,340 who were protected by a safety (n=28,564) (n=684) (n=959) (n=1705) (n=38) (n=187) 70% belt. Among completely ejected Unknown 2% 1% 1% 1% 2% 1% 459 (n=1,053) (n=22) (n=50) (n=45) (n=5) (n=16) 2% occupants subjected to rollover, 97 percent were unrestrained. In Total 23,480 447 1053 983 16 102 26,080 (n=47,082) (n=1,250) (n=1,897) (n=3,533) (n=114) (n=435) 100% support of restraint use were the Note: Slight differences may exist in percentage calculations owing to rounding. 69 percent of all CDS-reported Source: NCSA, NHTSA, NASS-CDS, 1995-2001 rollover occupants who by virtue of their restraint use were upants in rollover crashes. Among restrained. Although testing has retained within the vehicle. The planar crash occupants, 27 not subjected restraint systems percentage of occupant retention percent failed to use a safety belt to the forces imparted during a during rollover was exceedingly (were unrestrained), as compared rollover crash, they have been high for a technology not directly to 35 percent of the rollover shown to retard the outward addressing this situation. (other than zero quarter turn) motion of the occupants. crash occupants, per Table 3. Maximum Injury Severity As seen in Table 4, those The Maximum Abbreviated Ejection unrestrained occupants who Injury Scale (MAIS) describes As noted previously, those who remained within the vehicle the highest abbreviated injury were belted fared better than (30 percent of the rollover severity score (AIS); based on those unrestrained occupants. population) will have benefi ted mortality risk, sustained by an Further, ejection incidence was from the integrity of the occupant. The scale ranges from found to be far lower among those occupant compartment but will minor (AIS 1) through maximum Table 4 (AIS 6) injuries and unknown (AIS Percentage and Number of Occupants Involved in Rollover Crashes by 7.) Ten percent of all occupants Ejection Status and Restraint Use (Total column given in thousands, weighted) injured in rollover were seriously, including fatally, injured, as shown Restraint Use in Table 5. Another 15 percent Ejection (Row Percentage) (in ‘000’s) of occupants sustained MAIS Status No Yes Unk. Total injury in the moderate range. No Ejection 30 % 69 % 1 % 2,355 Since most rollover crashes 92% resulted in two quarter turns Complete Ejection 97 % 3 % 0 % 142 or less, thereby safeguarding 6% the integrity of the vehicle, it Partial Ejection 49 % 51 % 0 % 69 was not surprising that rollover 3% injuries were predominantly of Ejection - 87 % 12 % 1 % 2 lower injury severity. This did Unknown Degree 0% not mean that rollovers were Total 34 % 65 % 1 % 2,567 benign in nature. Most planar 100% crashes were also of low injury Note: Slight differences may exist in percentage calculations owing to rounding. severity, per Table 6. In rollover Source: NCSA, NHTSA, NASS-CDS, 1995-2001

National Center for Statistics and Analysis 400 Seventh St., S.W., Washington, D.C. 20590 crashes, however, 49 percent Table 6 of occupants experiencing three Percentage and Number of Rollover Occupants, by Maximum Injury or more quarter turns sustained Severity and Number of Quarter Turns minor injuries. When examining Quarter Turns the percentage of occupants (Column Percentage) End- involved in rollover crashes, over- Unknown (in ‘000’s) injury patterns indicated the MAIS 0 1 2 3+ End Number Total potential detriment to the occupant involved in a rollover 0 51 % 39 % 53 % 26 % 2 % 12 % 12,840 versus a planar crash. As a No Inj. 50 % percentage, MAIS scores for 1 39 % 47 % 34 % 49 % 54 % 38 % 10,060 Minor 39 % rollover occupants tended to be greater than those for planar 2 4 % 7 % 6 % 11 % 8 % 17 % 1,206 Moderate 5 % crash occupants. 3 – 6 2 % 4 % 3 % 9 % 32 % 10 % 669 Serious – Max. 3 % Table 5 7 4 % 3 % 4 % 5 % 5 % 23 % 1,088 Percentage of Injured Occupants, Unknown/Severe 4 % by Injury Severity and Crash Total 23,290 438 1,041 979 16 100 25,860 Attitude 100 % 100 % 100 % 100 % 100 % 100 % Note: Slight differences may exist in percentage calculation owing to rounding. Injury Severity Crash Attitude Source: NCSA, NHTSA, NASS-CDS, 1995-2001 (MAIS) Planar Rollover Minor 86 % 74 % Injuries by Body Region examinations may allow less Moderate 9 % 15 % Thirty-three percent of all severe injuries to remain Serious - Maximum 5 % 10 % reported injuries occurring in unidentifi ed until further All Known Injured 100 % 100 % non-planar crashes affected the examination or never become Note: Slight differences may exist in percentage head. Another one percent were documented for CDS. calculation owing to rounding. assessed as neck injuries. Rains Source: NCSA, NHTSA, NASS-CDS, 1995-2001 and Kanianthra (SAE 1995) Restrained occupants sustain suggested, that headroom nearly half of the injuries to Among CDS crashes, 70 percent reduction increased head injury extremities, as seen in Table 7. of occupants were restrained, risk, however, they did not The compression of the occupant as compared to 28 percent discount the risks associated compartment and fl ailing of who were unrestrained, per with vertical occupant excursion. members may partially explain Table 3. Upon disaggregation, It may be reasoned that the neck the head/face and extremity however, 64 percent of rollover may fi rst have loaded the roof, and injury frequency that is evident for occupants were restrained and rebounded off the roof thereby the unrestrained occupants. For 35 percent were unrestrained. creating a two-pronged insult. restrained occupants, the thorax Although a very high incidence Neck injuries of lower severity most frequently sustains serious of crashes resulted in low MAIS may have been overlooked or may injury. In contrast, the head and scores, this may have been a not have manifested themselves face of unrestrained occupants masking effect of the restrained until after examination. Similarly, are the most vulnerable to in relation to the unrestrained non-life threatening neck injuries serious injury. On average for occupant outcomes. Among may have been overlooked in these rollover data, three injuries rollover crashes, as compared to severely injured occupants. are recorded for unrestrained planar crashes, a slightly lower These considerations may be and six for restrained occupants. percentage of MAIS 1 injuries attributable, at least in part, Of injuries to unrestrained existed. As a percentage, MAIS to the necessity to immediately occupants, 23 percent are readings for rollover occupants treat more severe, potentially at least of AIS 2 severity, as tended to be greater than those life-threatening injuries in the compared to 10 percent for the for planar crash occupants. emergency room. The initial restrained occupants.

National Center for Statistics and Analysis 400 Seventh St., S.W., Washington, D.C. 20590 Table 7 leading. As the counterpoint, Percentage of Injuries Sustained by Occupants Involved in Rollover Crashes it was noted that the turn-over by Injured Body Region, Restraint Use, and Injury Severity (untripped) rollover type was the least frequent type of rollover Safety Injury Severity crash. Belt (Row Percentage) (in ‘000) Injured Body Region Use AIS 1 AIS 2 AIS 3-6 AIS 7 Total What is the Injury Outcome for the Rollover Occupants? Head/Face No 76% 13% 11% 1% 981 The occupant data were sorted Yes 89% 7% 3% 0% 882 All 82% 10% 7% 0% 1,872 in relation to restraint use and Upper & Lower No 82% 13% 5% 0% 1,017 crash confi guration. It was Extremities Yes 93% 6% 2% 0% 1,372 noted that among rollover crash All 88% 9% 3% 0% 2,398 occupants, the percentage that Spine No 71% 23% 6% 0% 283 were unrestrained was higher Yes 91% 7% 2% 0% 342 than the analogous percentage All 82% 14% 4% 0% 628 among planar crashes. Also, Thorax No 66% 6% 27% 1% 221 unrestrained occupants were at Yes 77% 7% 14% 1% 233 the greatest risk for injury and All 71% 7% 21% 1% 455 ejection. Further, MAIS readings Abdomen No 67% 20% 11% 1% 130 Yes 74% 18% 7% 1% 95 for rollover occupants were All 70% 19% 9% 1% 227 proportionally higher than those Unspecifi ed No 99% 0% 1% 0% 57 for planar crash occupants. Yes 98% 0% 2% 0% 37 The head and extremities were All 99% 0% 1% 0% 94 injured with greatest frequency. Neck No 91% 3% 3% 3% 23 It can be concluded that there Yes 99% 1% 0% 1% 37 would have been a larger number All 96% 2% 1% 1% 60 of severe injuries were it not for Total No 77% 14% 9% 0% 2,712 the benefi ts of safety belt use. Yes 90% 7% 3% 0% 2,998 All 84% 10% 6% 0% 5,734 The research is on-going and Note: Slight differences may exist in percentage calculation owing to rounding. The “All” category includes the unknown restraint category, however, it is not explicitly included in the table. subsequent phases will be Source: NCSA, NHTSA, NASS-CDS, 1995-2001 published in the form of research notes and technical reports. The Conclusions Based on the data examined in next phase will contemplate the In this review of NASS – CDS data this report, it possible to address issues discussed below. over the years 1995 through the questions posed in the Section 2001, rollover crash mechanisms “Objectives” regarding: prevalent Issues Identifi ed with Current and their related injury patterns rollover crash attributes and Parameters and Rationale for are summarized. Although the injury outcomes for occupants of Future Modifi cations single vehicle rollover is a highly rollover crashes. In examining the data for all complex crash mode, the multi- Which are the Prevalent Rollover occupants and their vehicles, issues were encountered. These vehicle rollover becomes less Crash Attributes? clear with the added dimension issues existed in comparing the The tripping mechanism generally of a planar component. For this injury outcomes and vehicle inducing rollover was the ground reason, only single vehicle crash damage and included: crash with initial tripping force exerted mechanisms are considered in the confi guration, impact orientation, at the wheels/tires. The trip- summation. Owing to the small occupant biomechanics, seating over type of rollover was found sample size, however, injuries of position, and clinical case to be the most prevalent. Most single and multiple vehicle crash analysis. rollovers ended in one or two confi gurations are aggregated quarter turns. Slightly more than Crash Confi guration within the study and for the half of the single vehicle crashes The crash confi guration should summarization that follows. rolled to the left, driver side be considered in this type of

National Center for Statistics and Analysis 400 Seventh St., S.W., Washington, D.C. 20590 analysis. The bulk of the rollovers presence of vehicles with many References are attributable to single vehicle occupants. The representation ! Federal Highway Administration, events. However, the presence of the vehicle fl eet suggests that “Vehicle Test Summary for Roof of other vehicles introduces there are many seating positions Crush Drop Test (Chevrolet S-10) issues beyond those within the other than the front that might be on Load Cell Plate (3x4), 03005 A&B”, Washington, D.C., 2003. traditional single vehicle rollover. occupied, however, consideration ! Henderson, Michael, Paine, for mass, human and cargo, It may also be useful to determine Michael, “Passenger Car Roof whether the case vehicle that distribution should be considered. Crush Strength Requirements”, does roll acts as the bullet or For normalization of the data Department of Transport and struck vehicle. analysis, it is suggested that the Regional Development, The Impact Orientation driver and right front passenger Federal Offi ce of Road Safety Additionally, the issue of near would serve to balance impact (Australia), Canberra, 1998. side versus far side leading orientation reporting. ! Hendriks, D., Scheiffl ee, T., rollovers must be considered in Clinical Case Analysis Page, J., Bellis, E., “Analysis of Untripped Rollovers, Final Report tandem with the quarter turn Finally, based upon this update, Technical Findings”, CALSPAN, CDS variables and attributes progression. Consideration to Buffalo, 1998. the occupant position and his should be revised to capture ! Hertz, Ellen, “Passenger Vehicles relation to the leading side may more specifi c elements of rollover in Untripped Rollovers”, NHTSA, be useful. First, for purposes of crashes. Of specifi c interest is the Washington, D.C., 1999. comparison, planar crashes of further defi nition of the rollover ! National Archives and Records near side and far side may be object to which an attribute Administration, Code of Federal used. Further, the occupant on might be ascribed. For instance, Regulations, 49 Parts 400 to the far side of the roll initiation the width of a curb or the type 999, generally sustains the most of ground could be aggregated ! Revised as of October 1, 2001, Transportation, Washington, severe injuries. in a statistical analysis, however, D.C., 2001. for an engineering analysis the Occupant - Biomechanics ! Rains, Glen C. and Kanianthra, Crash injury outcomes should be disaggregated elements may be Joseph N., “Determination of reported separately for children crucial. These descriptors could the Signifi cance of Roof Crush and adults. The biomechanical lead to better understanding of on Head and Neck Injury to responses of children are the true sequence of a rollover Passenger Vehicle Occupants in different from those of adults. crash and eventual test design Rollover Crashes”, SAE, 1995. First, vehicle hardware is based upon these concepts. ! United States Department of designed for adults. Secondly, Some of these descriptors Transportation, National Highway Traffi c Safety Administration, child safety seats are retrofi t might be found via clinical review for cases with copious “National Automotive Sampling items that are meant to work System, Crashworthiness Data supplementary information. in concert with a wide variety of System, 2000 Coding and Editing restraint hardware found in the Currently, however, the existing Manual”, NHTSA, Washington, vehicle fl eet. Finally, children lack CDS variable set does not require D.C., 2000. soft tissue and bone development. this type of reporting. They will not be fully developed, in general, until the child reaches 12 years of age. For additional copies of this research note, please call Occupant – Seating Confi guration 1 800 934 8517 or fax your request to 202 366 3189. Crash severities for front and For questions regarding the data reported in this research, rear seated occupants are contact Ana María Eigen [202 366 2903]. This research note different. Further, crash severity and other general information on highway traffi c safety may be estimates can be biased by accessed by internet users at: http://www-nrd.nhtsa.dot.gov/ departments/nrd-30/ncsa/AvailInf.html

National Center for Statistics and Analysis 400 Seventh St., S.W., Washington, D.C. 20590