2

CURRENT UNDERSTANDING OF MOTOR VEHICLE CRASH AVOIDANCE AND CRASHWORTHINESS

The characteristics and features of a motor vehicle that affect its safety can be classified into two broad categories: those helping the driver avoid a crash (crash avoidance) and those helping to protect vehicle occupants from harm during a crash (crashworthiness). Both aspects of vehicle safety are appropriate candidates for consumer information. In this chapter, the primary sources of information about vehicle-related factors affecting crash avoidance and crashworthiness are identified and the current state of knowledge about each is summarized. In the final section, a critical appraisal of the strengths and weaknesses of current knowledge is provided as a basis for developing meaningful consumer information about vehicle safety.

VEHICLE SAFETY DATA

The primary types of data available to analyze vehicle crash avoidance potential and vehicle crashworthiness are (a) engineering data and (b) crash data. The former, which include crash tests, are supported by research on injury mechanisms and human tolerance to impacts.

Engineering Data

Engineering data are physical measurements taken under static or dynamic conditions to predict vehicle crash avoidance potential and vehicle crashworthiness. Engineering data based on crash tests that measure vehicle crashworthiness are the most advanced. The National Highway Traffic Safety Administration (NHTSA) has more than 15 years of ex-



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 24
Shopping for Safety: Providing Consumer Automotive Safety Information 2 CURRENT UNDERSTANDING OF MOTOR VEHICLE CRASH AVOIDANCE AND CRASHWORTHINESS The characteristics and features of a motor vehicle that affect its safety can be classified into two broad categories: those helping the driver avoid a crash (crash avoidance) and those helping to protect vehicle occupants from harm during a crash (crashworthiness). Both aspects of vehicle safety are appropriate candidates for consumer information. In this chapter, the primary sources of information about vehicle-related factors affecting crash avoidance and crashworthiness are identified and the current state of knowledge about each is summarized. In the final section, a critical appraisal of the strengths and weaknesses of current knowledge is provided as a basis for developing meaningful consumer information about vehicle safety. VEHICLE SAFETY DATA The primary types of data available to analyze vehicle crash avoidance potential and vehicle crashworthiness are (a) engineering data and (b) crash data. The former, which include crash tests, are supported by research on injury mechanisms and human tolerance to impacts. Engineering Data Engineering data are physical measurements taken under static or dynamic conditions to predict vehicle crash avoidance potential and vehicle crashworthiness. Engineering data based on crash tests that measure vehicle crashworthiness are the most advanced. The National Highway Traffic Safety Administration (NHTSA) has more than 15 years of ex-

OCR for page 24
Shopping for Safety: Providing Consumer Automotive Safety Information perience with the New Car Assessment Program (NCAP), whose primary purpose is to provide consumers with information on the relative crashworthiness of passenger vehicles as measured in full-frontal crash tests. The tests are conducted at 56 km/hr (35 mph)—8 km/hr (5 mph) more and with 35 percent greater energy than the federal certification standard, Federal Motor Vehicle Safety Standard (FMVSS) 208, so that differences in frontal crashworthiness performance among vehicles can be more readily observed (NHTSA 1993, 2–4). NHTSA has also developed a dynamic crash test to measure vehicle crashworthiness in side-impact crashes in which the vehicle is struck in the side by a barrier traveling at 54.6 km/hr (33.5 mph) (DOT 1994). Agency requests to expand the NCAP to include side-impact crash testing, however, have not been funded. The Insurance Institute for Highway Safety (IIHS), a nonprofit scientific and educational organization supported by the insurance industry, has recently begun testing vehicle crashworthiness in frontal offset crashes.1 However, there is no federal safety standard relating to offset frontal crashes, and NHTSA does not conduct offset frontal crash tests itself. Very few engineering data are available on how vehicle characteristics affect the potential for a vehicle to be involved in a crash. As part of its recent rulemaking on rollover prevention (Federal Register 1994), NHTSA proposed two vehicle stability metrics,2 which agency analyses had shown to be significant predictors of vehicle propensity to roll over. The automobile industry, however, challenged NHTSA's finding that the stability metrics explain about half of the variability in rollover likelihood in single-vehicle accidents for the population of passenger vehicles and light trucks examined (Federal Register 1994, 33,260). Industry analysts maintained that the role of the vehicle, as measured by the stability metrics, is overstated relative to the role of environmental and driver factors in rollover crashes, and that vehicles with the same static stability measures have widely varying rollover crash experience (AAMA 1994, 1).3 Consumers Union, which publishes the popular Consumer Reports, has developed a simple dynamic test, which it believes better represents real-world emergency handling and vehicle rollover propensity. The test measures vehicle stability and handling in a double lane change maneuver typical of ordinary driving. Vehicles are driven at increasing speeds up to the point at which the vehicle is no longer controllable.

OCR for page 24
Shopping for Safety: Providing Consumer Automotive Safety Information Biomechanics Research Research in biomechanics (the study of injury mechanisms and human tolerances to trauma) provides the basic knowledge to support the development of testing devices and, from these, performance standards for testing vehicle crashworthiness.4 To understand how injuries are sustained and how to prevent or minimize the severity of such injuries in crashes, biomechanical engineers examine the mechanism of a particular injury associated with a given type of impact and how a body region responds to that impact. Impact tests are typically performed on cadavers in laboratories equipped with impact sleds and pendulum-type impactors. Once the injury mechanism is understood, the next step is to determine how much of an impact a given body region can withstand before it becomes severely injured, that is, to establish a threshold of human tolerance. This threshold is typically set at a moderate level of injury for a given impact severity, because it would be too expensive to manufacture cars that allow the average occupant to walk away uninjured from a severe crash. Having determined these two crucial components of injury biomechanics—the injury mechanism and the human tolerance level—crashworthiness performance standards can be established. However, consumers should be made aware that there is a wide variation in human tolerance, which is primarily a function of age and gender. For example, many postmenopausal females suffer from osteoporosis, or bone loss, and may be injured in a crash from which a 20-year-old male can walk away. Testing occupant protection in controlled crash situations has also necessitated development of anthropomorphic test dummies, better known as crash dummies. The data to build the dummies were developed from the human response studies using cadavers. The goal was to make the dummies respond in as human a fashion as possible. Different dummies have been developed for different crash configurations (e.g., frontal impact, side impact) and for different ages, heights, and weights. Crash tests for certification and consumer information purposes are based on 50th percentile males, that is, a 5-ft 8-in. male weighing 170 lb (Wolkomir 1995, 31). Crash Data NHTSA, the states, and the insurance industry have developed data bases that provide information on the incidence and outcome of high-

OCR for page 24
Shopping for Safety: Providing Consumer Automotive Safety Information way crashes. NHTSA has developed the primary traffic accident data bases. The oldest is the Fatal Accident Reporting System (FARS), which has provided an annual census of fatal traffic crashes since 1975 (NHTSA 1994, 3). A companion data base to provide information on nonfatal crashes, the National Accident Sampling System (NASS), has been in operation since 1979. To reduce costs and tailor the system to meet user needs, NASS was recently restructured into two data collection systems: (a) a General Estimates System to provide national estimates of crashes by type from police accident reports without intensive follow-up crash investigations and (b) a Crashworthiness Data System (CDS) based on detailed information of a small sample of crashes to support research directly related to injury and crashworthiness of passenger cars and light truck vehicles. Other specialized data files, such as the Crash Avoidance Research Data File, have been developed to provide information for examining the relationships between vehicle design characteristics and crash likelihood in support of crash avoidance research. Most states have computerized accident data systems based on police accident reports, which include information on fatal and injury crashes. Criteria for reporting property-damage only accidents, however, vary from state to state, as do the overall consistency and level of detail of the information. The Highway Loss Data Institute (HLDI) is closely associated with IIHS. Both are funded by the automobile insurance industry. HLDI collects claims information from the major insurance companies and summarizes injury, collision, and theft losses of passenger cars and light truck vehicles. Injury losses are presented as the frequency of insurance claims made under personal injury protection insurance coverage in “no-fault” states (HLDI 1994).5 The use of no-fault states helps ensure that the claims associated with a vehicle are for the occupants of that vehicle only—under no-fault each covered vehicle is responsible for its own occupants regardless of “fault” in the collision. However, information is not included on the costs of injuries paid either from personal health insurance coverage or bodily injury liability from other vehicles (Council et al. 1995, 10), or on the actual injury or severity of the injury that occurred. Data Issues Each of the data sources just discussed has strengths and limitations as a basis for providing consumer safety information. Consumers are

OCR for page 24
Shopping for Safety: Providing Consumer Automotive Safety Information interested in the safety of motor vehicles as they are operated in the real world. Crash data for specific makes and models of vehicles can provide this type of information. However, real-world crashes are the product of many factors. Thus, isolating vehicle-related factors from the other factors, particularly human error and environmental conditions that affect crash likelihood and severity, is difficult. In addition, crash data are retrospective; the on-road crash performance of a particular car can only be evaluated after a few years' experience on the highway, leaving an information gap for purchasers of new car models. Finally, existing crash data bases are themselves limited by reliability issues and by differences in reporting criteria, variable definitions across reporting jurisdictions, and availability of critical variables of interest. Crash tests offer a controlled setting that can help isolate the vehicle-related variables of interest. Here, too, reliability issues are a concern. Repeated measurements under the same conditions using the same vehicle make and model should yield similar test results. In addition, the crash tests should be valid indicators of actual crash experience. As discussed subsequently, the reliability and validity of NHTSA's NCAP crash test results are an issue. One alternative for improving the predictive power of crash tests is to combine the crash test results for specific vehicle makes and models with occupant injury experience in prior crashes involving similar (i.e., “clone”) vehicles. The idea is that at least some of the nonvehicle factors affecting crash outcomes could be controlled by introducing real-world data for vehicles similar to the crash-tested vehicle. A recent study of the feasibility of this approach (Council et al. 1995) found that the addition of crash injury data did improve the ability to predict future crash experience of new cars. The head injury measurements from the crash test, the medical claims data for predecessor vehicles from HLDI, and the proportion of severe driver injury in crashes involving predecessor vehicles were statistically significant predictors of expected injury for the new vehicle. However, the overall predictive ability of the model was not high, reflecting the continuing difficulty of controlling for the nonvehicle factors that affect crash outcomes (Council et al. 1995, 17–19).6 In the absence of information about clones, the addition of information about vehicle weight and size—key variables affecting crashworthiness —should further improve the predictive capability of the model.7

OCR for page 24
Shopping for Safety: Providing Consumer Automotive Safety Information CURRENT STATE OF KNOWLEDGE FROM AVAILABLE DATA Vehicle characteristics and safety features affect the likelihood of being in a crash as well as the crashworthiness of the vehicle once a crash has occurred. However, the relationship between vehicle characteristics and crash likelihood is not as well understood. Crash Avoidance Crash avoidance research has traditionally taken a secondary role in NHTSA's motor vehicle safety research program, in part because of the difficulty of isolating the contribution of vehicle characteristics to crash potential (TRB 1990, 145). The probability of being in a crash is somewhat related to vehicle design components, such as rollover propensity, steering sensitivity, and braking performance, but it is perhaps most related to the characteristics and behavior of the driver (Council et al. 1995, 2). Nevertheless, NHTSA has conducted research on vehicle-related measures to reduce the likelihood of crash involvement. Avoidance of rollover crashes has been a priority of NHTSA's crash avoidance research program (TRB 1990, 145). The popularity of sport utility vehicles and the evidence that their rollover propensity is higher than that of some passenger vehicles put pressure on the agency to develop a vehicle stability standard in the late 1980s. Two petitions, one from Congressman Wirth in 1986 and the other from Consumers Union in 1988, provided the basis for initiation of a formal agency rulemaking process in 1992 (NHTSA 1992, 5–6).8 The agency proposed two static vehicle stability metrics, described earlier in this chapter, for a possible vehicle standard. Because of sharp differences of opinion over whether these measures are accurate predictors of real-world rollover crashes and cost-effectiveness considerations,9 NHTSA decided to terminate regulation of a vehicle rollover standard (Federal Register 1994, 33,258).10 Research and rulemaking are continuing, however, on such vehicle-related rollover crash avoidance features as antilock brakes11 and on provision of consumer information on vehicle rollover propensity. Other than research on avoidance of rollover crashes, the primary focus in NHTSA's crash avoidance research program has been on technology improvements—brake lighting, better tire and brake performance, and vehicle conspicuity. Studies of the effectiveness of many of

OCR for page 24
Shopping for Safety: Providing Consumer Automotive Safety Information these technology improvements have indicated only modest differences in crash likelihood. For example, antilock brakes have not provided as large a safety benefit as predicted.12 One reason may be the small number of circumstances in which antilock brakes can help; fewer crashes involve loss of control that could be prevented with antilock brakes than were supposed (IIHS 1995a, 4, 5). Another explanation may lie in improper use of the brakes. Some consumers apparently continue to pump the brakes.13 New vehicle technologies that may have greater potential for assisting the driver in avoiding crashes are being developed in conjunction with the Intelligent Transportation Systems (ITS) program. Such new technologies as enhanced night vision and collision avoidance systems could significantly improve the type of information given to drivers and the speed with which it is provided. Thus, although the driver-vehicle interaction in crash avoidance is still not well understood, crash avoidance technologies may yield significant payoffs as ITS safety-related vehicle technology improvements are introduced. Crashworthiness Driver and other vehicle occupant characteristics, such as age, belt use, and position at the time of a crash, affect injury outcome. Design features, such as vehicle size and weight and occupant compartment integrity, also play a dominant role in determining vehicle crashworthiness and occupant protection. Thus, NHTSA's motor vehicle safety research program has given top priority to research on measures to improve vehicle crashworthiness (TRB 1990, 41). Crash data provide a good indication of the key vehicle-related factors that affect crash outcomes as well as the types of crashes that lead to most of the deaths and injuries on the nation's highways. Vehicle Size and Weight Vehicle size and weight are consistent predictors of injury likelihood in all types of crashes. All else being equal, big and heavy cars offer more protection to their occupants than small and light cars (O'Neill 1995, 1). Although there is some question whether car size or car mass is the more important feature (Evans 1994; Evans and Frick 1992; O'Neill 1995), in practice there is a close relationship between the two. Typically, a large car is heavy, and a small car is light (Evans 1994, 14). The effect of weight or mass is large. On the basis of a relationship derived from FARS data on more than 40,000 driver fatalities (Evans

OCR for page 24
Shopping for Safety: Providing Consumer Automotive Safety Information and Frick 1993, 215), when two cars whose masses differ by a factor of 2 crash into each other, the driver of the lighter car is 12 times as likely to be killed as the driver of the heavier car (Figure 2-1). However, the weight differential between cars involved in most multiple-vehicle crashes typically is smaller,14 and when single-vehicle crashes are taken into account,15 the probability of fatality averaged over all crash types is 2 to 3 times greater for the driver of a lighter car than for a driver of a heavier car (Evans 1989, 1,152). Furthermore, larger cars are more likely to have additional crush space to protect the integrity of the occupant compartment in more severe crashes. Crash data indicate that the importance of vehicle weight and size holds across vehicle types. Occupants of small vehicles—whether cars, pickup trucks, or sport utility vehicles—are at much greater risk of fatal injury (all else being equal) taking into account the numbers of these vehicles on the road (Figure 2-2). Fatality rates are 1.5 to 2.5 times higher for occupants of the smallest vehicles than for occupants of the largest vehicles in each vehicle class (Figure 2-2)16 Despite the effect of vehicle weight and size on injury likelihood, the purchase of larger and heavier cars is not necessarily the best outcome from society's perspective. For example, occupants of a heavier car crashing with a lighter car are better protected, but at the expense of the occupants of the lighter car (O'Neill 1995, 5). Even if all car drivers choose to drive vehicles of the same mass, the range of other vehicles on the road—from bicycles to heavy trucks —is such that differences in mass could not be controlled (Evans 1994, 15). Moreover, mass is not a major factor in single-vehicle crashes, which account for 58 percent of all fatal crashes (NHTSA 1994, 43). In particular, vehicle size (i.e., vehicle length and track width) is the key factor in rollover crashes, which are a predominantly single-vehicle crash type (Evans 1994, 13). Finally, there are trade-offs between vehicle size and weight and other factors, such as fuel economy; smaller and lighter cars are more fuel efficient (Evans 1994, 16; Lave 1981). Consumers should be given information about the effect of vehicle size and weight on injury likelihood, but this is only one factor that must be considered in vehicle purchase decisions. Crash Types Crash data can also help identify crash types that lead to most of the deaths and injuries on the nation's highways. Data available from NHTSA's CDS17 indicate that vehicles involved in frontal crashes ac-

OCR for page 24
Shopping for Safety: Providing Consumer Automotive Safety Information FIGURE 2-1 The ratio, R, of driver fatalities in the lighter car to driver fatalities in the heavier car versus the ratio, µ, of the mass of the heavier car to the mass of the lighter car for frontal crashes, 1975–1989 FARS data. Modified and reprinted from Accident Analysis and Prevention, Vol. 25, No. 2, L. Evans and M.C. Frick, Mass Ratio and Relative Driver Fatality Risk in Two-Vehicle Crashes, pp. 213–224, copyright 1993, with kind permission from Elsevier Science Ltd., The Boulevard, Langford Lane, Kidlington 0X5 1GB, United Kingdom.

OCR for page 24
Shopping for Safety: Providing Consumer Automotive Safety Information FIGURE 2-2 Occupant deaths per million registered vehicles, 1990–1993 average (data compiled by Highway Loss Data Institute from FARS and R.L. Polk's National Vehicle Population Profile).

OCR for page 24
Shopping for Safety: Providing Consumer Automotive Safety Information FIGURE 2-3 Incidence of vehicle involvement in crashes by crash type and severity, 1990–1993 average (data compiled by NHTSA from NASS-CDS file). count for 51 percent of vehicles involved in crashes that result in fatalities and injuries (Figure 2-3). The next most common crash types from the standpoint of vehicle involvement are side-impact crashes, rollovers, and rear-impact crashes, which account for 25, 15, and 9 percent, respectively, of all vehicles involved in fatal and injury crashes (Figure 2-3). Vehicle involvement in frontal-, side-, and rear-impact crashes occurs primarily in multiple-vehicle collisions, whereas vehicle involvement in rollovers is more common in single-vehicle crashes. When the data are disaggregated by vehicle type, it appears that vehicles differ in their propensity to be involved in various types of crashes (Figure 2-4). The largest difference is for vehicles involved in rollover crashes. Light truck vehicles—sport utility vehicles, pickups, and vans—are more than twice as likely as passenger cars to be involved in rollover crashes involving fatalities and injuries (Figure 2-4). However, light truck vehicles are somewhat less likely to be involved in side-impact crashes.

OCR for page 24
Shopping for Safety: Providing Consumer Automotive Safety Information FIGURE 2-9 Rollover crash statistics, 1990–1993 average (data on vehicle involvement in crashes compiled by NHTSA from NASS-CDS file; data on fatality outcomes compiled by Highway Loss Data Institute from FARS and R.L. Polk's National Vehicle Population Profile).

OCR for page 24
Shopping for Safety: Providing Consumer Automotive Safety Information consumers about the crashworthiness of passenger vehicles in rollover crashes is premature. Rear-Impact Crashes Vehicles involved in rear-impact crashes account for the smallest fraction (9 percent) of vehicles involved in all fatal and injury crashes among all crash types (Figure 2-10). This finding is not unexpected, because one of the most common crash types is a two- FIGURE 2-10 Rear-impact crash statistics, 1990–1993 average (data on vehicle involvement in crashes compiled by NHTSA from NASS-CDS file; data on fatality outcomes compiled by Highway Loss Data Institute from FARS and R.L. Polk's National Vehicle Population Profile).

OCR for page 24
Shopping for Safety: Providing Consumer Automotive Safety Information vehicle fender bender, which does not result in life-threatening injuries. Rear-impact collisions predominantly involve passenger cars in multiple-vehicle crashes (Figure 2-10). Fatality levels are lower than in other crash types, and the rates are highest for passenger cars (Figure 2-10).32 The primary injury in rear-impact crashes is whiplash, but the mechanism of injury is not well understood.33 Research is being done and some tests have been conducted.34 There is evidence that head restraints reduce the probability of neck injury (Bourbeau et al. 1993; Kahane 1982; Ollson et al. 1990; Svenssan et al. 1993), and head restraints located behind and close to the back of the head are preferable. The level of knowledge is sufficient to provide consumers with information about head restraints but inadequate to provide general information about the crashworthiness of vehicles in rear-end collisions. SUMMARY OF THE STATE OF KNOWLEDGE Since the federal role in regulating motor vehicle safety was established in 1966, major advances have been made in understanding injury mechanisms and human tolerance levels in crashes. The advances have provided the basis for vehicle safety performance standards and design improvements. The state of knowledge is most advanced in the area of vehicle crashworthiness, where the link between design features, such as vehicle size and weight and occupant compartment integrity, is strongly associated with how well the occupants are protected in a crash. Vehicle weight and size are consistent predictors of injury likelihood in all types of crashes. The effect of vehicle weight or mass is large; the probability of a fatality, when averaged over all crash types, is 2 to 3 times greater for the driver of a lighter car than for the driver of a heavier car. Larger cars are more likely to have additional crush space to protect the integrity of the occupant compartment in more severe crashes. Knowledge about frontal crashes, the most common of the crash types that result in death and injury, is the best developed. Injury mechanisms and human tolerances are well understood, performance standards have been established, and frontal crash tests for certification and consumer information purposes have been conducted for more than 15 years. However, important gaps in understanding remain. Full-frontal crash tests reflect only a small fraction of the real-world variation in

OCR for page 24
Shopping for Safety: Providing Consumer Automotive Safety Information crash speeds and configurations even for frontal crashes. Knowledge about potential variance in crash test scores is insufficient, which calls into question the repeatability of test results. Finally, these crash test results have a modest correlation with real-world crash performance. The correlation could be improved by combining crash test results with actual crash data. However, by their nature, crash tests are unlikely ever to have a high correlation with on-road crashes because they cannot reflect important driver and vehicle use characteristics. Knowledge about side-impact crashes, another common crash type, is also well advanced. Injury mechanisms and human tolerances are adequately understood, and performance standards and test criteria have been established. However, the U.S. test standards are not fully accepted by the biomechanical community, and much work remains to be done before consumers can be provided valid data about injury protection in side-impact crashes. Mechanisms for reducing injury when vehicles roll over are not well understood. Knowledge about injury in rear-impact crashes, which account for the smallest fraction of severe crashes, is also limited. Research is being conducted, but except for information on head restraints, the level of knowledge has not advanced to the stage that general information on the relative crashworthiness of vehicles in rollover or rear-end collisions can be made available to consumers. How vehicle design features affect the likelihood of being in a crash is less well understood, reflecting in part the importance of other factors, particularly driver behavior, in crash causation. Research has focused on technology enhancements, such as improvements in braking systems and vehicle conspicuity, but most of these advances have made only modest differences in crash likelihood. New technologies that have the potential to provide great assistance to drivers and reduce crash likelihood significantly are being introduced as part of the ITS program, such as enhanced night driving and collision avoidance systems. Thus, the role of vehicle factors in avoiding crashes is an important area for continued research. In summary, the current level of understanding about vehicle safety characteristics and features—both their effect on crash likelihood and the protection they afford to vehicle occupants—is not well enough advanced to provide consumers with a definitive assessment, based strictly on scientific grounds, of the highway safety performance of a new vehicle. By the same token, there is a good understanding of key

OCR for page 24
Shopping for Safety: Providing Consumer Automotive Safety Information vehicle-related factors that affect crash outcomes, such as vehicle size and weight, and of design features, such as energy absorption, that can reduce injury potential. There is also considerable experience with frontal crash test results that offer some modest correlation with realworld crash experience. With further work, plus expert judgment, this understanding can provide the foundation for giving consumers more predictive measures of the overall safety of new motor vehicles. NOTES 1. The 64-kph (40-mph) test speed, higher than the 56- to 60-kph (35- to 37.5-mph) speed being considered for the European standard, was selected for the same reason that the NCAP test is conducted at 8 kph (5 mph) above the compliance standards—high speeds magnify differences among cars. 2. The two metrics were critical sliding velocity and tilt table angle. Critical sliding velocity is a measure of the minimum lateral (sideways) vehicle velocity required to initiate rollover when the vehicle is tripped by something in the roadway (e.g., a curb). Tilt table angle is the angle at which the last uphill tire of the vehicle lifts off a platform as the platform is increasingly tilted (Federal Register 1994, 33,259). 3. The Transportation Analysis Institute, an independent company, has developed a rollover stability index—the K index, which is purported to have a good correlation with actual crash data. The index includes four factors—vehicle track width, vehicle height, vehicle weight, and vehicle weight above a certain height—but the specific calculations are proprietary (personal communication, Joseph Kimmel, Transportation Analysis Institute, July 13, 1995). 4. A more detailed description of research in impact biomechanics can be found elsewhere (NRC 1985; Viano et al. 1989). 5. Collision and theft losses are reported as average loss payments per insured vehicle year (i.e., the claim frequency times the average loss payment per claim) (HLDI 1994). 6. The researchers expected that, in such a multifactor situation, the variables that can be measured and be expected to remain constant for a given make and model, such as crash test variables and the clone vehicle crash and insurance information, would have a modest predictive ability. The results of the study support this hypothesis (Council et al. 1995, 18). 7. The model of Council et al. takes size and weight into account in that the crash-tested vehicle and its clones are approximately the same weight. Unlike crash-tested vehicles, the clone vehicles are involved in crashes with vehicles of different weights, representing different types of crashes (e.g., multiple-vehicle and single-vehicle crashes). Therefore, with this approach it is possible to predict crashworthiness across vehicles—a crash-tested vehicle and its clones will have one predicted injury outcome; a different vehicle and its clones will have another. Although not attempted in the study by

OCR for page 24
Shopping for Safety: Providing Consumer Automotive Safety Information Council et al., information about vehicle weight and size could be substituted for the clone data if clones for a particular new make or model vehicle are not available. 8. Congressman Wirth's petition, which proposed a specific minimum static stability factor among other requests, was denied because the approach was considered too narrow and inappropriate (NHTSA 1992, 5–6). Consumers Union's petition, which recommended “a minimum stability standard to protect against unreasonable risk of rollover,” was granted in September 1988 (NHTSA 1992, 6). NHTSA had issued an Advance Notice of Proposed Rulemaking on Rollover Resistance in 1973 to solicit comments on development of a test procedure, test conditions, and performance requirements to evaluate vehicle rollover propensities, but after reviewing comments to the docket and conducting several studies, the agency concluded the rulemaking until the factors contributing to rollover crashes could be better understood (NHTSA 1992, 5). 9. NHTSA analysis showed that setting a performance level high enough to affect passenger cars would require redesign of nearly all sport utility vehicles, vans, and pickup trucks. NHTSA concluded that the degree of redesign would have raised issues of public acceptance and possibly even the elimination of certain classes of vehicles as they are known today (Federal Register 1994, 33,258). 10. This ruling has been vigorously opposed by highway safety groups, among others, who maintain that NHTSA could have examined other options such as stability standards for certain vehicle classes (Advocates for Highway and Auto Safety 1994). NHTSA, however, found that its stability metrics, when applied to a specific class of vehicle rather than to all vehicles grouped together, lost much of their predictive power to explain rollover crash occurrence (Federal Register 1994, 33,258). 11. NHTSA published an advance notice of proposed rulemaking for antilock brake systems for light-duty vehicles, including cars, vans, pickup trucks, and sport utility vehicles on Jan. 4, 1994 (59 FR 281) (Federal Register 1994, 33,255). 12. Two studies by the Highway Loss Data Institute and one by NHTSA found that antilock brakes are not reducing the frequency or the cost of crashes as measured by insurance claims for vehicle damage (IIHS 1995a, 4, 5). 13. Yet another possibility is that drivers of vehicles with antilock brakes take more risks (e.g., drive faster on wet pavement than they would have), thus compensating for the risk reduction benefits of the new technology. This is an example of a phenomenon called risk compensation or risk homeostasis—the theory that drivers adjust their behavior in response to perceptions of the risks they face. Whether the phenomenon exists has been hotly debated and is the subject of a considerable literature (see Peltzman 1975; Wilde 1982; Graham 1982; Evans 1986; and Lund and O'Neill 1986). 14. Crashes with a 1.5 times mass ratio (i.e., one car is 50 percent heavier than the other) are more typical (personal communication, Leonard Evans, Gen-

OCR for page 24
Shopping for Safety: Providing Consumer Automotive Safety Information eral Motors R&D Center, Dec. 20, 1995). In this case, the risk of fatality is greater by a factor of 4 for the driver of the lighter car involved in a crash with the 50 percent heavier car (Evans and Frick 1993, 214). The mass differential, however, may be growing as light truck vehicles become a larger share of the total passenger fleet. For example, data from FARS indicate that the fatality risk is 7 times greater for drivers of light cars involved in crashes with pickup trucks (Evans and Frick 1993, 222). 15. Fatality risk estimates are less certain for single-car crashes because of the difficulty of estimating exposure in single-car crashes. 16. One anomaly is the slight increase in fatality rates for the largest cars (i.e., wheelbase 114 in.). The explanation may be that very large cars are driven primarily by the old, who are more likely to die in a crash, thereby increasing the numerator of the fatality rate calculation. Of all the car classes, the largest car size has the smallest number of registered vehicles, thereby decreasing the denominator of the fatality rate calculation (personal communication, Maria Penny, Highway Loss Data Institute, Oct. 4, 1995). 17. CDS data are obtained from a nationally representative probability sample selected from all police-reported, tow-away crashes involving light-duty vehicles. For purposes of this analysis, crashes were included if they resulted in a fatality or an injury of any type, that is, ranging from a minor injury (Abbreviated Injury Scale 1) to a critical or maximum injury type (Abbreviated Injury Scale 5 or 6). 18. One anomaly is the fatality rate figure for medium-sized cars, which is lower than for large cars. One would have expected lower fatality rates for large cars because of their weight and size advantage. 19. Information was also developed on how injury criteria relate to injury risk. Biomechanical experts from Ford Motor Company and General Motors Corporation developed injury risk functions that relate the probability of injury risk to various acceleration levels for the head and chest injury criteria and for various load levels for the upper leg criterion (Hackney and Kahane 1995, 1–2). 20. Kinetic energy is proportional to the square of the velocity. Thus, there is 36 percent more kinetic energy in a 56-km/hr (35-mph) crash than in a 48-km/hr (30-mph) crash (NHTSA 1995). 21. On the basis of vehicle involvement in crashes gathered from the NASS-CDS for 1990 to 1993, only 1 percent of vehicles involved in frontal crashes resulting in fatalities and injuries are directly comparable with the fully frontal crash tests used for certification purposes and for the NCAP. Crashes are defined as fully frontal if they meet the following three criteria: (a) direction of force is at 12 o'clock; (b) crush profile is zero, that is, the centerline of the vehicle is directly aligned with the center of the direct damage; and (c) the crush damage on the left front of the vehicle is equivalent to the crush damage on the right front of the vehicle within 20 percent (criteria were defined by Carl Ragland, Office of Crashworthiness Research, NHTSA).

OCR for page 24
Shopping for Safety: Providing Consumer Automotive Safety Information 22. On the basis of the same NASS-CDS data for 1990 through 1993, 21 percent of vehicles involved in frontal crashes resulting in fatalities and injuries are classified as being in frontal offset crashes. 23. Deformable barriers were selected because they provide a reasonable approximation of how cars actually perform in offset crashes (IIHS 1994, 6). 24. For example, the HIC scores for vehicles in 1979 ranged from 521 to 4,513; in 1993 that range was between 273 and 1,459 (GAO 1995, 31). 25. By controlling for differences in driver age and sex as well as weight of the vehicle and examining only frontal crashes involving belted occupants, the researchers were able to isolate differences in vehicle crashworthiness. 26. It is conceivable that a combination of regulatory standards and voluntary efforts by the manufacturers could result in improvements in vehicle crashworthiness so large that, by the time a vehicle crashworthiness rating system is developed, the differences in crashworthiness between vehicles of equivalent weight would be less than could be predicted with statistical confidence by the rating. Should this occur, however, the process of developing the rating would have accomplished one of the primary objectives of a program to improve consumer vehicle safety information, that is, to provide additional incentives for the automobile manufacturers to enhance vehicle safety. 27. The vehicles are actually struck in the side by a barrier traveling at this speed (DOT 1994). 28. The federal standard uses the thoracic trauma index (TTI), which is based on the measured lateral accelerations on the spine and near-side rib of the SID. The industry maintained that TTI is not a good predictor of soft organ damage or overall injury probability in side-impact crashes. General Motors, in particular, developed an alternative criterion, known as the viscous criterion, which measures the extent of chest deflection and the speed of the chest wall, as well as a more humanlike dummy for side-impact testing, known as BIOSID. (BIOSID has a more flexible rib cage than SID.) 29. Crashes are defined here as rollover crashes if a vehicle rollover occurred either as the first harmful event or as a subsequent event (e.g., from a collision with another vehicle or object). This definition of rollover is used by NHTSA (NHTSA 1994, 185). 30. The forces causing severe neck injuries vary over a wide range depending on the extent of neck flexion or extension and on the configuration of the neck at the time of impact. 31. The effect of the use of shatterproof glass in side windows on head injuries in side-impact crashes must also be studied. 32. The scale for the fatality figures is the same as for the other crash impact directions—frontal, side, and rollover—to facilitate comparison. 33. One hypothesis of the injury mechanism is that the sudden forward acceleration of the struck vehicle causes a shearing effect to be developed between adjacent cervical vertebrae; this relative motion causes injury to the soft tissues connecting these vertebrae. Testing of this hypothesis is under

OCR for page 24
Shopping for Safety: Providing Consumer Automotive Safety Information way (personal communication, Albert King, Wayne State University Bioengineering Center, Sept. 6, 1995). 34. IIHS recently tested 164 head restraints for their ability to prevent whiplash (IIHS 1995c). Test results, however, have not yet been correlated with onroad crash injury rates. REFERENCES Abbreviations AAMA American Automobile Manufacturers Association DOT U.S. Department of Transportation GAO General Accounting Office HLDI Highway Loss Data Institute IIHS Insurance Institute for Highway Safety NHTSA National Highway Traffic Safety Administration NRC National Research Council TRB Transportation Research Board AAMA. 1994. Response to Docket No. 91-68, Notice 3, No. 47. Comments on Consumer information Regulations; Rollover Prevention. Detroit, Mich., Oct. 20, 71 pp. Advocates for Highway and Auto Safety. 1994. Response to Docket No. 91-68, Notice 3, No. 35. Petition for Reconsideration: Termination of Rulemaking on Light Passenger Vehicle Rollover Prevention. Washington, D.C., Aug. 30, 55 pp. Bourbeau, R., D. Desjardins, M. Maag, and C. Laberge-Nadeau. 1993. Neck Injuries Among Belted and Unbelted Occupants of the Front Seat of Cars. Journal of Trauma, Vol. 35, pp. 794–799. Council, F.M., J.R. Stewart, and C.L. Cox. 1995. Predicting Future Crashworthiness for New Cars: Exploration of a New Methodology. University of North Carolina, Chapel Hill, July, 23 pp. DOT. 1994. NHTSA Cites Safety Features in 1995 Cars, Trucks, and Vans. News, Office of the Assistant Secretary for Public Affairs, Washington, D.C., Oct. 6. Evans, L. 1986. Risk Homeostasis Theory and Traffic Accident Data. Risk Analysis, Vol. 6, pp. 81–94. Evans, L. 1989. Passive Compared to Active Approaches To Reducing Occupant Fatalities . Proc., Twelfth International Technical Conference on Experimental Safety Vehicles, Gothenburg, Sweden, May 29–June 1, NHTSA, U.S. Department of Transportation, Vol. 2, pp. 1,149–1,157. Evans, L. 1994. Small Cars, Big Cars: What Is the Safety Difference? Chance, Vol. 7, No. 3., pp. 9–16. Evans, L., and M.C. Frick. 1992. Car Size or Car Mass: Which Has Greater Influence on Fatality Risk? American Journal of Public Health, Vol. 82, No. 8, Aug., pp. 1,105–1,112.

OCR for page 24
Shopping for Safety: Providing Consumer Automotive Safety Information Evans, L., and M.C. Frick. 1993. Mass Ratio and Relative Driver Fatality Risk in Two-Vehicle Crashes . Accident Analysis and Prevention, Vol. 25, No. 2, pp. 213–224. Federal Register. 1994. Consumer Information Regulations; Federal Motor Vehicle Safety Standards; Rollover Prevention. NHTSA, U.S. Department of Transportation. Vol. 59, No. 123, June 28, pp. 33,254–33,272. GAO. 1995. Highway Safety: Reliability and Validity of DOT Crash Tests. GAO/PEMD-95-5. May, 76 pp. Graham, J.D. 1982. On Wilde's Theory of Risk Homeostasis. Risk Analysis, Vol. 2, pp. 235–237. Hackney, J.R., and C.J. Kahane. 1995. The New Car Assessment Program: Five Star Rating System and Vehicle Safety Performance Characteristics. No. 950888. Presented at Society of Automotive Engineers International Congress and Exposition, Detroit, Mich., Feb. 27–Mar. 2, 16 pp. Haland, Y. 1994. Background, Description, and Evaluation of a New Side Airbag System . Proc., International Body Engineering Conference on Automotive Body Design and Engineering, Detroit, Mich., pp. 74–80. HLDI. 1994. Injury, Collision, and Theft Losses by Make and Model. Arlington, Va., Sept. IIHS. 1994. Future of Crashworthiness Research Includes Frontal Offset Tests into a Deformable Barrier. Status Report, Vol. 29, No. 7, pp. 1–7. IIHS. 1995a. Antilock Brakes Don't Reduce Crash Frequency or Cost, HLDI Reports. Status Report, Vol. 30, No. 2, Feb. 2, pp. 4–5. IIHS. 1995b. European Union Moves Toward New Safety Standards with Dynamic Tests . Status Report, Vol. 30, No. 7, Aug. 12, p. 5. IIHS. 1995c. IIHS Ranks Most Head Restraints “Poor” in Preventing Whiplash. Highway and Vehicle Safety Report, Vol. 22, No. 2, Oct. 2, pp. 5–6. Kahane, C.J. 1982. An Evaluation of Head Restraints, Federal Motor Vehicle Safety Standard 202. DOT HS-806 108. NHTSA, U.S. Department of Transportation, Feb., 308 pp. Kahane, C.J. 1994. Correlation of NCAP Performance with Fatality Risk in Actual Head-On Collisions. DOT-HS-808-061. National Highway Traffic Safety Administration, Jan., 164 pp. Kahane, C.J., J.R. Hackney, and A.M Berkowitz. 1994. Correlation of Vehicle Performance in the New Car Assessment Program with Fatality Risk in Actual Head-On Collisions. Paper No. 94-S8-O-11. National Highway Traffic Safety Administration, 17 pp. Lave, L.B. 1981. Conflicting Objectives in Regulating the Automobile. Science, Vol. 212, May 22, pp. 893–899. Lund, A.K., and B. O'Neill. 1986. Perceived Risks and Driving Behavior. Accident Analysis and Prevention, Vol. 18, No. 5, pp. 367–370. Machey, J.M., and C.L. Gauthier. 1984. Results, Analysis and Conclusions of NHTSA's 35 MPH Frontal Crash Test Repeatability Program. SAE Paper No. 840201. Society of Automotive Engineers, Warrendale, Pa., pp. 73–85. Malliaris, A.C., and K.H. Digges. 1987. Crash Protection Offered by Safety Belts. Proc., 11th International Conference on Safety Vehicles, Washington, D.C.

OCR for page 24
Shopping for Safety: Providing Consumer Automotive Safety Information McElhaney, J.H., and B.S. Myers. 1993. Biomechanical Aspects of Cervical Trauma. In Accidental Injury: Biomechanics and Prevention (A.M. Nahum and J.W. Melvin, eds.), Springer-Verlag, New York, pp. 311–361. NHTSA. 1992. Planning Document for Rollover Prevention and Injury Mitigation. Docket 91-68. No. 1. Office of Vehicle Safety Standards, Sept., 13 pp. NHTSA. 1993. New Car Assessment Program. Response to the NCAP FY 1992 Congressional Requirements. U.S. Department of Transportation, Dec., 125 pp. NHTSA. 1994. Traffic Safety Facts 1993. DOT HS 808 169. U.S. Department of Transportation, Oct., 190 pp. NHTSA. 1995. The New Car Assessment Program. Office of Market Incentives, U.S. Department of Transportation, 5 pp. NRC. 1985. Injury in America: A Continuing Public Health Problem. National Academy Press, Washington, D.C., 164 pp. Ollson, I., O. Bunketorp, G. Carlsson, C. Gustagfsson, I. Planath, H. Norin, and L. Yslander. 1990. An In-Depth Study of Neck Injuries in Rear End Collisions. Proc., 1990 International IRCOBI Conference on the Biomechanics of Impacts, Lyon, France, pp. 269–280. O'Neill, B. 1995. The Physics of Car Crashes and the Role of Vehicle Size and Weight in Occupant Protection. Insurance Institute for Highway Safety, Arlington, Va., July, 14 pp. Peltzman, S. 1975. The Effects of Automobile Safety Regulation. Journal of Political Economy, Vol. 83, Aug.–Dec., pp. 677–725. Svenssan, M., P. Lōvsund, Y. Háland, and S. Larsson. 1993. Rear-End Collisions: A Study of the Influence of Backrest Properties on Head-Neck Motion Using a New Dummy Neck. Paper 93-343. Society of Automotive Engineers, Warrendale, Pa. Syson, S.R. 1995. Occupant to Roof Contact: Rollovers and Drop Tests. Advances in Occupant Protection Technologies for the Mid-Nineties. SP1077. Society of Automotive Engineers, Warrendale, Pa., pp. 1–27. TRB. 1988. Special Report 218: Transportation in an Aging Society, Vol. 1. National Research Council, Washington, D.C., 125 pp. TRB. 1990. Special Report 229: Safety Research for a Changing Highway Environment . National Research Council, Washington, D.C., 166 pp. Viano, D.C. 1987. Evaluation of the SID Dummy and TTI Injury Criterion for Side Impact Testing. No. 872208. Proc., 31st Stapp Car Crash Conference, Society of Automotive Engineers. Viano, D.C., A. King, J.W. Melvin, and K. Weber. 1989. Injury Biomechanics Research: An Essential Element in the Prevention of Trauma. Journal of Biomechanics, Vol. 22, pp. 403–418. Wilde, G.J.S. 1982. The Theory of Risk Homeostasis: Implications for Safety and Health . Risk Analysis, Vol. 2, No. 4, pp. 209–225. Wolkomir, R. 1995. Sitting in Our Stead: Crash Dummies Take the Hard Knocks for All of Us. Smithsonian, Vol. 26, No. 4, July, pp. 30–41. Yoganandan, N., F.A. Pintar, A. Sances, J.M. Reinhartz, and S.J. Larson. 1991. Strength and Kinematic Response of Dynamic Cervical Spine Injuries . Spine, Vol. 16, pp. 511–517.