Effectiveness of Occupant Protection Systems and Their Use
The Intermodal Surface Transportation Efficiency Act (ISTEA) of 1991, enacted by Congress on December 18, 1991, directed the Secretary of Transportation to report on the effectiveness of occupant protection systems based on their actual use, and on lap and shoulder belt use by the public and various groups at both the state and national levels (Section 2508 (e)). This is the third of five biennial reports on the effectiveness of occupant protection systems and safety belt use.
With the substantial increase in available accident data during the past two years, it has become possible to investigate restraint system effectiveness across a broader range of issues and within particular subgroups of the driver and vehicle population. Thus, the current report presents a great deal more information than previous reports.
In addition, certain challenges regarding air bag deployment have materialized, which are discussed in this report. The first involves the increased risk of upper extremity injury associated with air bag deployment. The second, and more challenging issue, involves the child-passenger air bag interaction. On November 22, 1996, NHTSA announced a comprehensive approach to preserve the safety benefits of air bags while minimizing their danger to children and at-risk adults. Its approach centers on accelerating the development of "smart air bag" technology for future vehicles with the intent of having systems available for 1999 models.
More immediate measures include adopting enhanced warning labels, depowering of air bags, continuing to allow the use of on-off switches in vehicles without a rear seat to protect children, and allowing dealers to deactivate the air bags for any owner who requests it.
The major findings of this report are presented below.
Safety Belt Use
As of December 1995, state surveys indicated safety belt use rates ranging from 40 percent in
South Dakota to 86 percent in New Mexico. An estimate of national safety belt use is derived
through a population-weighted average of these state use rates. The national safety belt use
rate as of December 1995 was estimated to be 68 percent.
Belt use information is not routinely collected for the military, government employees, or law
enforcement personnel. However, based upon the existence of mandatory use policies, training
programs, and promotional campaigns, use among these groups is expected to be higher than in
the general population.
History of FMVSS 208 Requirements
Federal Motor Vehicle Safety Standard (FMVSS) 208 ("Occupant Crash Protection"), as
amended on July 17, 1984, required that automatic occupant protection, such as air bags or
automatic belts, be phased into passenger cars during 1987-1990. When the National Highway
Traffic Safety Administration (NHTSA) issued FMVSS 208, it also began a continuing,
nationwide effort to increase belt use through encouragement of state buckle-up laws,
enforcement and public education. Use of manual lap-shoulder belts reduces the risk of fatal
injury to front-seat occupants by 45 percent, but in 1983, only 14 percent of the general driver
population buckled up. Initially, automatic belts installed in response to FMVSS 208 helped
increase belt use. In the long run, however, NHTSA believed that the best protection would
come from air bags in combination with buckle-up laws in most of the states to ensure high rates
of belt use.
FMVSS 208's phase-in requirement for automatic occupant protection was: 10 percent of model
year 1987 passenger cars, 25 percent of model year 1988, 40 percent of model year 1989, and all
cars manufactured after September 1, 1989 (or model year 1990). FMVSS 208 was later
amended to allow an exclusion from the automatic protection requirement for the right-front
passenger position until September 1, 1993, if an air bag was installed for the driver. All vehicles
manufactured after September 1, 1993, are required to have automatic protection for the driver
and right-front passenger. The 1987-1990 phase-in schedule of the automatic occupant
protection requirement was met or exceeded.
The two components of NHTSA's occupant protection program have reinforced each other.
FMVSS 208 offered a "choice" between automatic protection and safety belt laws, and therefore,
became the catalyst for the adoption of state "buckle-up" laws. In 1983, prior to FMVSS 208 (as
amended), no state had a belt use law. Currently, 49 states plus the District of Columbia and
Puerto Rico have safety belt use laws. In addition, national safety belt use has risen dramatically
from 14 percent in 1983 to 68 percent as of December 1995.
The Intermodal Surface Transportation Efficiency Act (ISTEA), passed by the Congress in 1991,
required all passenger cars manufactured after September 1, 1997, and light trucks manufactured
after September 1, 1998, to have driver and passenger air bags, plus manual lap-shoulder belts.
Also in 1991, NHTSA extended the automatic occupant protection requirements to light trucks
and vans, on a phased-in basis for model years 1995, 1996, 1997, and 1998.
Descriptions of Occupant Protection Systems
Several distinct types of automatic occupant protection systems are available currently in the on-road fleet, including automatic safety belts, manual safety belts, air bags, and various
combinations of the air bag with either manual or automatic belts. For the purpose of conducting
the analysis of injury-reducing effectiveness, these systems were grouped as follows:
How Air Bags Work
Air bags, an automatic crash protection system that deploys quicker than the blink of an eye, are
the result of extensive research to provide maximum crash protection. Air bags by themselves
protect only in frontal crashes, and offer maximum protection when used in conjunction with
safety belts. Air bags should NOT be used as the only form of occupant protection; they are
intended to provide supplemental protection for belted front-seat occupants in frontal crashes.
Typical air bag systems consist of three components: an air bag module, crash sensor(s), and a
diagnostic unit. The air bag module, containing an inflator and a vented or porous, lightweight
fabric air bag, is located in the hub of the steering wheel on the driver side or in the instrument
panel on the passenger side. Crash sensor(s), located on the front of the vehicle or in the
passenger compartment, measure deceleration, the rate at which a vehicle slows down. When
these sensor(s) detect decelerations indicative of a crash severity that exposes the occupants to a
high risk of injury, they send an electronic signal to the inflator to trigger or deploy the bag. The
diagnostic unit is an electronic device that monitors the operational readiness of the air bag system
whenever the vehicle ignition is turned on and while the ignition is powered. The unit uses a
warning light to alert the driver if the air bag system needs service.
Air bags are designed to deploy (inflate) in moderate-to-severe frontal and near-frontal crashes.
They inflate when the crash forces are about equivalent to striking a brick wall head-on at 10-15
miles per hour or a similar sized vehicle head-on at 20-30 mph. Air bags are not designed to
deploy in side, rear, or rollover crashes. Rollover crashes can be particularly injurious to vehicle
occupants because of the unpredictable motion of the vehicle. In a rollover crash, unbelted
occupants can be thrown against the interior of the vehicle and strike hard surfaces such as
steering wheels, windows and other interior components. They also have a great risk of being
ejected, which usually results in very serious injuries. Ejected occupants also can be struck by
their own or other vehicles. Since air bags provide supplemental protection only in frontal
crashes, safety belts should always be used to provide maximum protection in rollovers and all
The bag inflates within about 1/20 of a second after impact. The inflated air bag creates a
protective cushion between the occupant and the vehicle's interior (i.e., steering wheel,
dashboard, and windshield). At 4/20 of a second following impact, the air bag begins to deflate.
The entire deployment, inflation, and deflation cycle is over in less than one second.
After deployment, the air bag deflates rapidly as the gas escapes through vent holes or through the
porous air bag fabric. Initial deflation enhances the cushioning effect of the air bag by maintaining
approximately the same internal pressure as the occupant strokes into the bag. Subsequent rapid
and total deflation enables the driver to maintain control if the vehicle is still moving after the
crash and ensures that the driver and/or the right-front passenger are not trapped by the inflated
Dust-like particles present during the inflation cycle primarily come from dry powder that is often
used to lubricate the tightly packed air bag to ease rapid unfolding during deployment. Small
amounts of particulate produced from combustion within the inflator also are released as gas is
vented from the air bag. These dust particles may produce minor throat and/or eye irritation.
Once an air bag is deployed, it cannot be reused. Air bag system parts must be replaced by an
authorized service dealer for the system to once again be operational.
To ensure that infants and children ride safely, with or without a passenger-side air bag, NHTSA
urges care givers to follow three "rules":
Market Shares of the Various Occupant Protection Systems
Americans are using safety belts more than ever before, and air bags are offered in virtually all
new cars, and in many light trucks, vans and sport utility vehicles in advance of the Federal
requirement. Their market share is nearly universal, with only 1.5 percent of new cars still
offering automatic belts. The dramatic increase in belt use over the past 10 years enables the
combination of air bags and manual belts to yield significant life-saving results.
|Driver Air Bags||17,718,575|
|Dual Air Bags||13,450,136|
As of September 1, 1995 more than 31 million air bag-equipped new cars had been registered
since 1987, 17.7 million of which were equipped with driver air bags and another 13.5 million
equipped with dual air bags. Exhibit 2 presents the distribution of automatic occupant
protection systems in new cars registered each year between September 1987 and September
During the phase-in (1987-1990) of automatic occupant protection, automatic belts were the most frequently sold automatic occupant protection system. During 1990 to 1993, manufacturers equipped more and more new cars with driver air bags, which became increasingly popular among new car buyers. In 1995, cars equipped with driver or dual air bag systems represented over 98 percent of new cars sold, with 84 percent of all model year 1995 cars sold being equipped with dual air bags.
The Analytical Challenge
Assessing the benefits of occupant protection systems, that is, their effectiveness in
reducing the chances of occupant mortality and morbidity, is not a simple task. While it
would be quite easy to compare the observed injury and fatality rates for occupants protected
by the various systems, differences in fatality- and injury-reducing effectiveness can be
masked by a multitude of factors not directly related to the air bag or automatic belt
For example, air bags originally were offered in the larger, more expensive and the
sporty, high-performance model lines. These vehicles may be driven much differently than,
for example, station wagons or family-size sedans. The different use patterns of these
vehicles will result in a different mix of crashes; for example, some make/models will
experience more single-vehicle crashes on rural, higher speed roads, which tend to be more
severe than the average two-vehicle crash in an urban, lower speed environment. Also,
scientific studies have demonstrated that heavier cars, on average, offer greater protection
to their occupants than do lighter cars (which were generally not included during the early
introduction of air bags). These and other factors need to be identified and accounted for
in any analysis of system effectiveness and are among the challenges of conducting an
For example, other challenges include the identification and availability of appropriate
data for conducting this assessment, in terms of relevance, sufficient numbers of cases for
study, and quality. Safety belt use laws were critical to increasing belt use to the levels
we have achieved. However, repeated analyses have demonstrated that self-reported safety belt
use, such as that contained in most police accident reports, overstates the level of safety
belt use in these crashes. This may be due to penalties for nonuse of safety belts, discounts
offered by some automobile insurance companies for a signed commitment that the policyholder
will always use his or her safety belt, or other reasons. What has been observed is a
tendency for surviving drivers and passengers (especially those receiving no or only minor
injury) to say that they were wearing their belts at the time of the crash when actually they
were unrestrained. This causes higher reported safety belt use rates for occupants in
police-reported crashes than for those in the general population. It is very unlikely that
crash-involved occupants could have higher safety belt use rates than the general public
since the very behavior which leads to increased risk of crash involvement is hypothesized to
be associated with an increased tolerance for risk, such as not wearing safety belts, or
driving after drinking too much alcohol.
The effect of higher reported safety belt use, especially among the less seriously injured
and uninjured vehicle occupants, is to make safety belts appear safer than they actually
are. If some of the unbelted, uninjured people had been incorrectly reported as belted,
this would increase the computed injury rate among unbelted occupants (fewer uninjured
persons) and decrease the injury rate among the belted occupants, artificially increasing the
"gap" between belted and unbelted occupants and inflating the estimated advantage of using
Analysis Overview -- Fatality-Reducing Effectiveness
The database used to conduct the assessment of the fatality-reducing effectiveness of air
bags is NHTSA's Fatal Accident Reporting System (FARS). FARS is a census of all fatal
traffic crashes that occur in the U.S., on roads customarily open to the public, where at
least one person dies from crash-related causes within 30 days of the crash. Depending on
the state, the information entered into the FARS database by state analysts includes the
In previous reports, only the as-used fatality-reducing effectiveness of air bags
was estimated (as-used meaning that belted and unbelted occupants were combined into
one group, without regard to whether safety belts actually were used). With the substantial
increase in available fatal accident data during the past two years, it has become possible
to apply the same analytic methods used for overall air bag effectiveness to certain
subgroups of the driver population. One of the important ways to group drivers is by their
safety belt use. For the first time, separate estimates are presented for the
fatality-reducing effectiveness of air bags plus manual lap-shoulder belts
and air bags alone, relative to an unrestrained occupant, as requested in
the 1991 ISTEA legislation. At the same time, an updated estimate of the effectiveness of
manual lap-shoulder belts alone was not developed for this report -- it has been known for
some time that manual safety belts are 45 percent effective in reducing fatality risk. Also,
with the phasing out of automatic safety belts, no fatality-reducing effectiveness estimates
of these systems were computed.
In addition, while previous analyses of air bag effectiveness considered only
effectiveness in all crashes and frontal crashes, the data available today permit analysis of
other air bag-related issues, including fatality-reducing effectiveness for younger and older
persons, passenger air bags, driver air bags in light trucks, and driver air bags in small
and large cars.
As of July 1996, NHTSA's Fatal Accident Reporting System contained records of nearly
10,000 fatally injured front-seat occupants of cars and light trucks who were sitting at
positions equipped with air bags.
For the analysis of the fatality-reducing effects of air bags, the ability to control for
the various crash circumstances by comparing drivers and right-front passengers involved in
precisely the same crashes alleviated the need for detailed statistical modeling. A second
approach compared the driver fatality experience in air bag-equipped vehicles in frontal
fatal crashes (for which air bags were specifically designed) to their experience in
nonfrontal fatal crashes. This analysis assumes that air bags have no effect in nonfrontal
crashes, such as side and rear impacts and rollovers.
Analysis Overview -- Injury-Reducing Effectiveness
The database selected to conduct the assessment of injury-reducing effectiveness for
manual and automatic occupant protection systems is NHTSA's National Accident Sampling
System's (NASS) Crashworthiness Data System (CDS). This database is the most comprehensive,
representative crash investigation system available and has the most accurate safety belt use
reporting of any file available to NHTSA. Assessments of safety belt use in the current
analysis are based on the judgment of trained, professional accident investigators, after
having inspected the vehicles involved, reviewed the injury patterns based on hospital
records, and conducted interviews with crash-involved parties. Even so, occupant-reported
safety belt use in the NASS CDS is higher than the observed safety belt use of drivers on the
road. This is true for both manual and automatic safety belt systems. To a lesser extent,
this is also true for investigator-reported belt use, which was used in the injury-reduction
analysis. In the majority of crashes in the NASS CDS, safety belt use is determined by the
preponderance of evidence obtained from inspection of the vehicles involved, particularly in
moderate to high-severity crashes.
Overreporting of safety belt use is more likely in low damage severity and low occupant
injury crashes, which provide little in the way of hard evidence of safety belt use or
nonuse. These crashes are low frequency crashes in the NASS CDS sample but comprise a larger
proportion of the on-road crashes (NASS CDS oversamples more serious crashes, since these are
where the more serious injuries occur). Unlike other postcrash surveys, the NASS CDS
investigator does not rely primarily on self-reporting of safety belt use by the person
involved in the crash, which is generally the source for the information cited on police
accident reports. It is for this reason that the NASS CDS is believed to provide the most
reliable indications of the use of safety belts by crash-involved parties.
There continue to be too few cases of fatal injury in the NASS sample to conduct
an analysis of fatalities alone. Inasmuch as NASS consists of a sample of towaway passenger
vehicle crashes in the United States, the file contains detailed information on a broad range
of crash severities. In the NASS sample, as in the population of traffic crashes, less
serious injuries occur much more often than do fatalities. To achieve a sufficiently large
sample size, the analyses in the previous two reports to Congress estimated system
effectiveness in reducing the likelihood of moderate and greater injury (level 2 and greater
on the Abbreviated Injury Scale (AIS)(1)). Fatal injuries,
although relatively rare, are included in this category.
Additional exploratory analyses were conducted in the current report to investigate a
broader range of issues related to occupant injury. The presence of very detailed
information in the NASS CDS makes it possible to conduct these investigations. In addition
to effectiveness in preventing a maximum injury of severity AIS 2 and greater injury (MAIS
2+, or moderate and greater injury), analyses were conducted to estimate effectiveness in
preventing serious and greater injury (MAIS 3+), both overall and in frontal crashes.
For ease of discussion, these will be referred to as moderate injury and
serious injury in the remainder of this report. Analyses were also
conducted to investigate the moderate and serious injury-reducing effects for the following
body regions: head, chest, upper extremities and lower extremities. Separate estimates of
system effectiveness were investigated for male drivers and female drivers, for younger
(under age 22), middle age (22-49 years old) and older (50+ years) drivers; shorter (under 65
inches), moderate height (65-69 inches) and taller (70+ inches) drivers; and lighter weight
(under 135 pounds), moderate sized (135-179 pounds) and heavier (180+ pounds) drivers.
Analyses of these last issues were not possible at the serious injury level inasmuch as the
sample sizes decrease dramatically both from subsetting (to particular body regions injured,
age, sex, height and weight groups) and restricting to serious injuries, which are much less
frequent than are moderate injuries.
For the injury-reducing effectiveness, the analysis developed statistical models for
estimating system effectiveness that permit the analyst to account for the potentially
confounding effects of other factors, such as crash type and severity, occupant age, and sex,
etc. Two patterns of vehicle damage were studied. Effectiveness was measured for cars that
experienced damage to any part of the vehicle and those receiving damage to the front of the
vehicle. Frontal damage is the area where safety belts, and especially air bags, would prove
most beneficial at protecting vehicle occupants. For a more detailed discussion of the
specific statistical models employed to conduct these analyses, the reader is referred to
NHTSA's first report to the Congress(2) on this subject.
One of the more important variables determined and reported by the NASS investigators is
delta-v, a measure of crash severity representing the change in velocity to which vehicle
occupants are exposed. Delta-v has been shown to be a significant factor in determining
injury severity, with higher levels of delta-v indicating a greater likelihood of more
serious injury. The inclusion of delta-v in the statistical model yields a large improvement
in explaining why some vehicle occupants were injured.
The following section presents the estimates of effectiveness developed as a result of
these analyses. The injury-reducing effectiveness estimates presented cover four basic
groups of automatic occupant protection: air bag plus manual lap/shoulder belts, air bags
used alone, manual lap/shoulder belts used without the presence of an air bag, and all
automatic belts (without an air bag) considered as a single group.
Recent amendments to FMVSS 208 require the installation of air bags and manual
lap/shoulder belts in all new passenger cars manufactured in September 1997 and later (model
year 1998) and all new light trucks manufactured in September 1998 and later (model year
1999). Air bags have already proven to be the system of choice among consumers, so the trend
already had been established before this requirement was implemented. Automatic safety belts
have essentially been phased out of the new car market. Therefore, all the various 2-point
and 3-point automatic safety belt systems have been combined into a single group, automatic
belts. As stated earlier, no estimates of the fatality-reducing effectiveness of manual or
automatic safety belt were estimated.
For the exploratory injury analyses, the groupings for age, height and weight were
selected by ordering the data by the variable in question (e.g., height), dividing the
available data into quartiles (25 percent groupings) and selecting the lowest and highest
quartiles (e.g., the 25 percent of available drivers who were the shortest and the 25 percent
who were the tallest), and the remaining 50 percent in the moderate range (although for age,
only the highest quartile vs. the remaining 75 percent was studied to isolate effects on the
older population). This provides a reasonable number of cases for conducting the analyses
addressing the issues related to age, height and weight. Each estimate is accompanied by an
indication of its statistical significance to provide a measure of the reliability of the
estimate. However, many of these estimates failed to reach statistical significance due to
the reduced sample size resulting from selecting only those cases within an analysis
Results of Analysis of Fatality-Reducing Effectiveness of Air Bags
As of July 1996, NHTSA's Fatal Accident Reporting System contains records of nearly 10,000
fatally injured front-seat occupants of cars and light trucks who were sitting at positions
equipped with air bags. The agency now has enough data for preliminary estimates of the
fatality reduction by air bags plus safety belts, and air bags without safety belts, relative
to an unrestrained occupant, as requested in the 1991 ISTEA legislation. There are also
enough data to compare the fatality reduction by air bags in purely frontal crashes for car
drivers, car passengers, and drivers of light trucks and to compare the fatality-reducing
effectiveness of air bags for drivers of small and large cars, and for younger and older
The basic analyses of data from the Fatal Accident Reporting System are updates of the
findings in NHTSA's Second Report to Congress. These analyses estimate the fatality
reduction by air bags for drivers of passenger cars, and they are measured on the combined
population of belted and unbelted occupants of cars equipped with driver air bags. The
analyses studied the fatality experience of drivers in frontal crashes vs. two select
comparison groups: their right-front passengers and the drivers in nonfrontal crashes.
The first analysis considers crash-involved passenger cars equipped with air bags and
3-point belts at the driver's seat, but only 3-point belts at the right-front seat and with
both front outboard seats occupied. The ratio of driver fatalities (with the air bag) to
right-front passenger fatalities (without the air bag) is calculated, and it is compared to
the corresponding ratio in earlier cars of the same makes and models, equipped only with
3-point belts at both seating positions. The fatality-reducing effectiveness of air bags is
estimated by the relative difference in the two ratios. This analysis includes all drivers
and right-front passengers of the cars, both belted and unbelted.
In purely frontal crashes (principal impact of 12:00, excluding cases where the most harmful
event was a rollover), air bags are now associated with a statistically significant
fatality-reducing effectiveness of 34 percent. This result is essentially unchanged from the
35 percent cited in NHTSA's Second Report to Congress. In all frontal crashes
(including pure and partial frontals; principal and/or initial impact point of 10:00 to
2:00), the analysis now indicates a statistically significant 18 percent reduction in
fatality risk for air bags. For all types of crashes combined, including nonfrontals, the
fatality reduction is a statistically significant 10 percent.
The second analysis utilizes another distinctive characteristic of air bags: that they
are primarily designed for action in frontal crashes. With an inclusive definition of
"frontal and partially frontal" crashes (initial or principal impact location between 10:00
and 2:00), it can be assumed that air bags have little effect, relative to manual belt use
alone, in the remaining "nonfrontal" crashes. These nonfrontal fatalities represent an
exposure-based control group for the current analysis.
The ratio of frontal to nonfrontal driver fatalities in cars equipped with driver air bags
is compared to the corresponding ratio in earlier cars of the same makes and models, equipped
only with 3-point belts. The fatality-reducing effectiveness of air bags in frontal crashes
is estimated by the relative difference in the two ratios. This approach has the
disadvantage of relying on the unproven assumption of zero effectiveness for air bags in
nonfrontal crashes, but it allows a larger sample size than the preceding method, since it is
not limited to cases where the right-front seat was occupied.
When driver fatalities in purely frontal crashes (principal point of impact of 12:00,
excluding cases where the most injurious or damaging event was a rollover) are compared to
those in nonfrontal crashes, the fatality reduction for air bags is a statistically
significant 27 percent. For all frontal or partially frontal crashes (principal and/or
initial impact point of 10:00 to 2:00), there is an estimated 19 percent reduction of
fatality risk (statistically significant). With the assumption that air bags have negligible
effect in nonfrontal crashes, the combined fatality reduction for air bags in all types of
crashes is estimated at 12 percent, by this analysis method.
Air bags did not have a statistically significant effect in partially frontal crashes
(principal and/or initial point of impact between 10:00 and 2:00, excluding purely frontal
crashes with 12:00 point of principal impact) by either analysis method. The first method
yielded a 2 percent fatality increase with air bags, the second, a 9 percent reduction.
Exhibit 3 summarizes the results of these two analytical approaches and presents the set
of estimates that result from computing the average of the two approaches, which is the final
estimate of effectiveness.
Fatality-Reducing Effectiveness of Driver Air Bags
The final estimates are presented graphically in Exhibit 4. The analyses indicate that
the life-saving benefits of air bags derive almost entirely from purely frontal crashes; that
their benefit in partially frontal crashes, if any, is quite limited; and that the fatality
reduction in all types of crashes is slightly more than one-third of the reduction in purely
One of the most important ways to group drivers is by their belt use. Belted drivers in
cars equipped with air bags experienced a statistically significant 21 percent fatality
reduction in purely frontal crashes, relative to belted drivers in comparable cars without
air bags. Unbelted drivers with air bags experienced a statistically significant 34 percent
fatality reduction in purely frontal crashes, relative to unbelted drivers without air bags.
In other words, air bags have significant life-saving benefits in purely frontal crashes for
belted and unbelted drivers; however, the benefit appears to be somewhat
larger, relatively speaking, for the unbelted driver.
The two preceding estimates need to be carefully interpreted. The 21 percent reduction
for the belted driver with an air bag is measured relative to the belted driver without an
air bag; it does not include the very substantial effect of belts, but represents the
increment of air bags plus belts over belts alone. Both estimates are for purely frontal
crashes; the fatality reduction in all types of crashes is slightly more than one-third of
the reduction in purely frontal crashes -- i.e., about 9 percent for the belted driver
(relative to a belted driver without air bags) and 13 percent for the unbelted driver. That
averages out to about 11 percent, for all drivers -- belted and unbelted, combined. NHTSA
estimates that safety belts alone reduce fatality risk by 45 percent. Thus, if an
unrestrained driver has a fatality risk of 100, a driver protected by both a safety belt and
an air bag will have a risk of:
Based on these considerations, NHTSA can now provide initial estimates of the
when-used fatality reduction for three types of occupant protection systems, for
passenger car drivers in crashes of all types, as requested by the 1991 ISTEA legislation,
and presented in Exhibit 5.
Estimated Effectiveness of Occupant Protection Systems
in Reducing Fatality Risk for Passenger Car Drivers
For example, if 100 drivers of cars not equipped with air bags were killed in crashes, 13
of them would have been saved (87 would still have been killed) had their cars been equipped
with a driver air bag. Had these fatally injured drivers been using their lap-shoulder belt
in their air bag-equipped car, 50 of them would have been saved.
While the estimates presented in Exhibit 5 are based on when safety belts
actually were or were not used, the fatality-reducing effectiveness estimates that follow
combine both belted and unbelted drivers, or passengers as noted, and are therefore
considered as used, in the same spirit as the overall effectiveness estimates
presented earlier in Exhibit 3.
Another important way to group drivers is by their age. Drivers age 29 or less
experienced a statistically significant 32 percent fatality reduction with air bags in purely
frontal crashes, relative to drivers of that age in comparable cars without air bags. The
fatality reduction for drivers age 30-55 was a statistically significant 35 percent. For
drivers age 56-69, the observed effectiveness dropped to a nonsignificant 25 percent, and for
drivers age 70 or older, it dropped to a nonsignificant 11 percent. These statistics,
although based on too few cases to be definitive, suggest that air bags are less effective
for older drivers than for young adults.
Cars can be grouped by their weight; specifically, the cars on the Fatal Accident Report
System were subdivided, by car weight, into three groups containing equal numbers of accident
records. The weight ranges for these three groups were: up to 2778 pounds, 2779-3119
pounds, 3120 pounds or greater. The observed fatality reduction by air bags in purely
frontal crashes is a statistically significant 30 percent in the light cars, a statistically
significant 24 percent in the medium-weight cars, and a statistically significant 39 percent
in the heavy cars. In other words, air bags are effective in cars of all sizes, and there
are not enough data, at this time, to reveal any trend in effectiveness by car size.
There are now enough accident data involving light trucks -- pickup trucks, vans or sport
utility vehicles -- for analyses of fatality reduction by driver air bags. Drivers of light
trucks equipped with air bags experienced a statistically significant 27 percent fatality
reduction in purely frontal crashes, relative to drivers of trucks of the same makes and
models, but without air bags. This is essentially the same effectiveness as for passenger
car drivers (31 percent), and it suggests that driver air bags may be as effective in light
trucks as they are in passenger cars.
There are also enough accidents involving passenger cars with dual air bags for analyses
of the effect of passenger air bags. However, it is important to distinguish between two
quite different groups of right-front passengers: pre-teen children (including infants and
toddlers), as opposed to teenagers and adult passengers. In-depth investigations of
individual crashes have shown that air bags of current design have fatally injured a number
of pre-teen children in frontal crashes, especially unrestrained children and infants in
rear-facing safety seats. The accident data involving child passengers age 0-12 on the Fatal
Accident Reporting System are sufficient for limited statistical analyses, and the results
support the unfavorable findings of the in-depth accident investigations. Every analysis
that includes frontal crashes shows a higher fatality risk for the children in cars with dual
air bags than for children in comparable cars without passenger air bags. Depending on the
analytical method, some, but not all, of the increases are statistically significant. In
other words, although a specific numerical value on the effect of passenger air bags on
children 0-12 years old cannot yet be determined, the results are consistent with the
conclusion, from special crash investigations, that child passengers are experiencing
problems with air bags.
On the other hand, for right-front passengers age 13 or older, passenger air bags have
significant benefits. Passengers of cars equipped with dual air bags experienced a
statistically significant 27 percent fatality reduction in purely frontal crashes, relative
to passengers of cars of the same makes and models, but equipped only with driver air bags.
This is essentially the same effectiveness as for car drivers (31 percent), and it suggests
that passenger and driver air bags may be equally effective for occupants age 13 or older.
Exhibit 6 summarizes the fatality-reducing effects of air bags for the subgroups discussed
Estimates of the Fatality-Reducing Effects of Air Bags
in Purely Frontal Crashes (12:00)
Analysis of Head-On Fatal Crashes
An update of the analysis presented in the previous report to Congress focuses on head-on
collisions between two passenger cars. Data from the Fatal Accident Reporting System for
1987 through 1995 included over 8,300 fatal head-on crashes involving two passenger cars in
which exactly one of the drivers was fatally injured and the other survived. In
this database, there were 370 crashes in which exactly one of the drivers had an air
bag-equipped vehicle. Fatality risk in this sample of crashes (which by definition was 50
percent overall, since the sample consisted of head-on crashes where exactly one of the two
drivers was fatally injured) was estimated as a function of the absence/presence of an air
bag, the difference in vehicle mass of the two cars colliding head-on, and the difference in
the ages of the two drivers. Safety belt use was not considered due to the aforementioned
problems with the reporting of safety belt use, especially among crash survivors.
The results of this analysis found that the mass differential was the most critical factor
in determining which of the two drivers involved was fatally injured, with the driver of the
heavier vehicle at reduced risk of fatal injury. Following the mass differential, the
difference in driver ages was a highly significant predictor of relative fatality risk, with
the younger of the two drivers at a lesser risk of fatal injury. Following this, the
presence of a driver air bag produced a statistically significant reduction in fatality risk
for drivers in these head-on fatal collisions. Remembering that in these crashes one of the
drivers was fatally injured and the other survived, the driver protected by an air bag
(without regard to whether the driver was using the available safety belt) experienced a
reduction in the chance that he/she was the fatally injured driver by a statistically
significant 27 percent.
To put the contribution of these three factors in perspective, one can translate the
fatality-reducing effects of the air bag into equivalent changes in the vehicle mass
differential and driver age differential. These equivalent effects are presented graphically
in Exhibit 7.
In this sample of head-on fatal crashes, the presence of a driver air bag reduced the
likelihood of that driver's being fatally injured by 27 percent, which is equivalent to
adding an additional 297 pounds of vehicle mass (greater mass provides greater protection),
or being 12 years younger (younger persons tend to be less fragile), all other things being
In conclusion, the analyses of fatal crashes discussed in this section demonstrate that
air bags are performing as they were designed, significantly reducing the likelihood of fatal
injury in the most severe types of crashes, both frontal and head-on collisions, with the
greatest benefits derived in crashes with a 12:00 point of impact.
Results of Analysis of Injury-Reducing Effectiveness
While most of the fatality-reducing effectiveness estimates presented in the previous
section were based on the combination of belted and unbelted occupants, without regard to
whether the available safety belts actually were used (that is, as-used), ALL of the
estimates of injury-reducing effectiveness presented in this report are based on
when safety belts actually were or were not used.
Overall and Frontal Crash Effectiveness Estimates for Moderate Injury
The results of statistical modeling of the NASS crash data are presented in Exhibit 8 and
graphically in Exhibit 9. Effectiveness was estimated for two vehicle damage populations: all
damage areas combined and damage to the front of the vehicle.
Effectiveness means that an occupant using the particular system will experience the cited
percentage reduction in the chance of injury, given that a crash has occurred, compared to an
As can be seen in Exhibit 8, both automatic and manual safety belt systems provide
significant overall injury-reducing benefits compared to being unrestrained. The estimates
for these two safety belt systems are very close, 43 percent for automatic belts and 49
percent for manual lap-shoulder belts. The air bag plus lap-shoulder belt system provides
the greatest injury protection, 60 percent (although the difference vs. safety belts alone is
not statistically significant). The injury-reducing effectiveness of the air bag alone,
without the use of safety belts, was estimated to be 18 percent, and was not significantly
better than being unrestrained.
For vehicles with frontal damage, the air bag plus manual lap-shoulder belts provides the
greatest protection against moderate injury (again, the differences compared to the other
systems are not significant). For air bags with manual belts, the effectiveness in reducing
the chance of moderate injury is about the same in frontal crashes as in all crashes (61
percent). The effectiveness of the air bag alone was lower in frontal crashes (6 percent)
than in all crashes. Frontal crashes are where the air bag is most likely to deploy,
suggesting that deployment may result in increased risk of moderate injury in some
situations. Analyses reported later in this section may shed light on this situation. The
above results are presented graphically in Exhibit 9.
In addition to the analysis of system effectiveness in preventing moderate injury,
investigations were conducted to assess system effectiveness in preventing serious injury,
which occurs less frequently. The results of this analysis are presented in Exhibits 10 and
Overall and Frontal Crash Effectiveness Estimates for Serious Injury
As can be seen in Exhibit 10, manual lap-shoulder belts, both with and without an air bag,
provide significant overall injury protection at the serious injury level, compared to being
unrestrained (about 60 percent reduction in serious injury risk for both systems). Thus, the
air bag did not appear to provide additional protection to a driver who was using his/her
belts for overall injury level. However, further analysis of specific body regions, reported
later, sheds more light on this situation. The serious injury protection afforded by
automatic belts (37 percent) also was statistically significant. The injury-reducing
effectiveness of the air bag alone, without the use of safety belts, is a nonsignificant 7
percent in all crashes and a nonsignificant -8 percent in frontal crashes. The results in
Exhibit 10 are presented graphically in Exhibit 11.
Estimated Effectiveness of Occupant Protection Systems
in Reducing the Likelihood of Serious Injury (MAIS 3+)
For this report, a number of exploratory analyses were conducted to estimate, in greater
detail, the effectiveness of occupant protection systems including separate estimates for
male and female drivers, for two driver age groups, for three driver height groups, and for
three driver weight groups. In addition, analyses focused on the effectiveness of these
systems in preventing injury to the following major body regions: head, chest, upper
extremity and lower extremity. It should be emphasized that these analyses required further
subsetting of the database, which considerably reduced the available number of cases for
analysis. As a result, many of these estimates are not statistically significant. However,
for the purposes of conducting exploratory analyses, this series of estimates may provide
suggestive evidence of patterns that might be indicative of the need for further
research and analysis. The focus of these analyses was to investigate the performance of air
bags plus lap-shoulder belts contrasted with lap-shoulder belt use only, and air bag alone
contrasted with totally unrestrained. Thus, no analysis of automatic safety belt performance
Effectiveness Estimates for Major Body Regions (Moderate and Serious Injury)
The effectiveness estimates in each column of Exhibit 12 represent the percentage
reduction (or increase in the case of a negative effectiveness estimate) in the risk of
moderate or serious injury to the respective major body region, without regard to other body
regions that may have been injured. Thus, the 83 percent effectiveness for the air bag plus
lap-shoulder belt in the column titled head means that drivers protected by
that system experienced an 83 percent reduction in the risk of moderate head injury, compared
to if that driver had been unrestrained, without an air bag.
Estimated Effectiveness of Occupant Protection Systems
in Reducing the Likelihood of Moderate and Serious Injury
to the Head, Chest, Upper Extremity and Lower Extremity
HEAD -- Manual lap-shoulder safety belts provide significant
injury-reducing benefits for moderate and serious head injury. The addition of an air bag to
a lap-shoulder belt appeared to result in increased head injury protection at both injury
levels. The estimated effectiveness of the air bag alone is 46 percent in reducing moderate
head injury. This 46 percent effectiveness for the air bag alone is much greater than the 18
percent effectiveness in reducing moderate injury to any body region
(Exhibit 8) and suggests that the air bag alone affords good protection in reducing head
injury. However, much can be gained from the addition of a lap-shoulder belt (effectiveness
increased from 46 percent to 83 percent). Further analysis indicates that the air bag alone
provides less protection for preventing serious head injury (a nonsignificant 16 percent),
while the protective benefits of the air bag plus lap-shoulder belt remains high.
CHEST -- The air bag plus lap-shoulder belt provided the only
statistically significant injury protection to drivers at the moderate injury level, and
significant benefits at the serious level. The estimated effectiveness of the manual
lap-shoulder belt system in reducing moderate chest injury was 14 percent, indicating that
drivers experienced slightly less risk of an moderate chest injury (but not significantly so)
if they wore their lap-shoulder belts compared to being unrestrained. However, at the
serious injury level, manual lap-shoulder belt effectiveness increased to a significant 54
percent. The air bag alone was associated with a (nonsignificant) increase in the risk of
moderate chest injury; at the serious level, the air bag provided some protection (18 percent
reduction in risk). One possible explanation for the large benefit of the air bag plus
lap-shoulder belt at the moderate level may be that the restraining effect of safety belt
systems, inhibiting drivers from moving forward in a crash, may result in fractured ribs or
sternum, bruised diaphragm, or minor bruises and lacerations to abdominal organs, all
classified as AIS 2 injuries. The addition of the air bag to a lap-shoulder belt may serve
to cushion the driver's forward movement, resulting in the driver "striking" the safety belt
with less force and thus, a lesser chance of these AIS 2 chest and abdominal injuries.
UPPER EXTREMITIES -- Manual safety belts provide statistically
significant protection against moderate injury. Drivers wearing manual lap-shoulder belts
experienced an estimated 54 percent reduction in the risk of moderate arm injury and a
(nonsignificant) 28 percent reduction of serious arm injury, compared to if they had been
unrestrained. However, the addition of an air bag reduced this benefit from 54 percent to 45
percent at the moderate level and from a nonsignificant 28 percent to an increased risk of
serious arm injury of 40 percent. From this, one could conclude that the air bag may be
causing an increased risk of arm injury to restrained drivers. The air bag alone appears to
provide some positive benefit in reducing moderate arm injury, but this benefit disappears at
the serious injury level. It appears that restrained drivers face a higher risk of moderate
and serious arm injury from the deploying air bag than does the unbelted driver. The
unrestrained driver in a crash experiences forward movement of his/her body into the bag,
while a belted driver's torso is held in place, possibly allowing his/her arms to flail
forward into the path of a deploying air bag. The expanding air bag may then injure the
driver's arm, or propel the arms upward or laterally into hard passenger compartment
surfaces. Another arm injury mechanism involves the positioning of the arm across the
steering wheel, directly in the path of a deploying air bag, while the vehicle is turning
left or right.
LOWER EXTREMITIES -- Manual lap-shoulder belts provided statistically
significant protection -- a 53 percent reduction in the risk of moderate injury. The air bag
plus lap-shoulder belt provided a nonsignificant 37 percent reduction in risk, while the air
bag alone provided a nonsignificant 21 percent reduction in the risk of moderate lower
extremity injury. For serious lower extremity injury (which is very high severity for lower
extremities, since AIS 4 is the highest possible coding for a lower extremity injury, e.g.,
severing of the femoral artery), manual lap-shoulder belts, with and without the air bag,
provide a very high degree of protection, almost 80 percent effectiveness (statistically
significant). At the same time, the effectiveness of the air bag alone is -5 percent,
suggesting the possibility of some increase in risk of serious lower extremity injury
(neither of these two estimates is statistically significant). For the air bag alone, some
degree of submarining, similar to unbelted drivers without an air bag, probably occurs.
From the analysis of injury to major body regions, the addition of an air bag to a
lap-shoulder belt system appears to involve a beneficial trade-off: reductions in the more
life-threatening moderate and serious injury to the head and chest, at the risk of increased
likelihood of upper extremity injury. The air bag system alone (without the use of a safety
belt) appears to be associated with increased risk of moderate injury to the chest, while
providing less protection to the head and upper extremity than any of the three safety belt
systems. This is further evidence of the need to always use safety belts, whether or not the
vehicle is equipped with air bags.
Effectiveness Estimates for Male and Female Drivers (Moderate Injury)
The estimates presented in Exhibit 13 represent the percentage reduction in the likelihood
of a moderate injury for male and female drivers. For example, the 64 percent estimated
effectiveness of the air bag plus lap-shoulder belt for male drivers means that males
protected by this system experienced a 64 percent reduction in the chance of a moderate
injury, compared to an unrestrained male driver.
For male drivers, the air bag plus lap-shoulder belt provided significant injury
protection (64 percent), while manual lap-shoulder belts alone reduced injury risk by a
significant 38 percent. The estimated effectiveness of the air bag alone for male drivers was
12 percent (nonsignificant).
All of the safety belt systems provided statistically significant injury reductions for
female drivers (compared to an unrestrained female driver). The air bag plus lap-shoulder
belt and lap-shoulder belt alone provided the same effectiveness (59 percent). The estimated
effectiveness of the air bag alone was a nonsignificant 25 percent.
Effectiveness Estimates for Two Driver Age Groups (Moderate Injury)
The estimates presented in Exhibit 14 represent the percentage reduction in the likelihood
of a moderate injury for a driver in the stated age group and protected by the particular
system, compared to a totally unrestrained driver in the stated age group.
For drivers age 15-49 years, the air bag plus lap-shoulder belt demonstrated a 62 percent
reduction in risk, followed by manual lap-shoulder belts alone (46 percent) and the air bag
alone (12 percent, nonsignificant). Thus, while the air bag alone appears to provide only
some injury reducing protection, the data suggest that the addition of an air bag to a manual
lap-shoulder belt provided additional injury-reducing effects at the moderate level (although
the difference was not statistically significant).
The same general pattern was found for drivers age 50 years and older, with both safety
belt systems demonstrating statistically significant injury-reducing benefits in the range of
54-57 percent. The 9 percent risk reduction for the air bag alone was not significantly
different from zero. Along with this, the addition of an air bag to the manual lap-shoulder
belt provided little or no increase in injury-reducing protection over the manual
lap-shoulder belt alone.
In summary, safety belt systems (manual lap-shoulder and manual lap-shoulder belts
supplemented by air bags) are providing significant injury-reducing protection for drivers in
age groups 15-49 years old and age 50 and older. In addition, the air bag system alone
appears to provide little protection to drivers in either age group at the moderate injury
Effectiveness Estimates for Three Driver Height Groups (Moderate Injury)
The estimates presented in Exhibit 15 represent the percentage reduction in the likelihood
of a moderate injury for a driver in the stated height group and protected by the particular
system, compared to a totally unrestrained driver in the stated height group. As stated
earlier, group definitions were determined by taking the upper and lower quartiles (25
percent) of the unweighted cases, with the remaining 50 percent making up the middle group.
Therefore, the group of drivers less than 65 inches tall has about the same number of cases
in the analysis as the group of drivers 71 inches and taller, each having about one-half the
number of cases available for analysis of the groups of drivers 65-70 inches tall.
For drivers less than 65 inches tall, both safety belt systems provided statistically
significant injury-reducing protection, with the air bag plus lap-shoulder belt having
somewhat lower protective benefits than the lap-shoulder belt alone. An interesting finding
is the 31 percent effectiveness for the air bag alone, which, while not statistically
significant, is relatively high for this small-statured population, comprised
disproportionately by women.
For drivers 65-70 inches tall, only the air bag plus lap-shoulder belt exhibited
statistically significant injury-reducing effectiveness, while the remaining systems
exhibited lower injury-reducing effectiveness. The 24 percent effectiveness for manual
lap-shoulder belts is surprisingly low.
Manual lap-shoulder belts provided significant injury-reducing protection for drivers over
70 inches tall. The addition of an air bag to the manual belt was associated with a
reduction in injury-reducing benefits, while the air bag alone provided only marginal
benefits to taller people (mostly males).
Effectiveness Estimates for Three Driver Weight Groups (Moderate Injury)
The estimates presented in Exhibit 16 represent the percentage reduction in the likelihood
of a moderate injury for a driver in the stated weight group and protected by the particular
system, compared to a totally unrestrained driver in the stated weight group.
Only manual lap-shoulder belts and air bag plus lap-shoulder belts provided statistically
significant injury-reducing benefits for drivers less than 135 pounds. The negative
effectiveness (-36 percent) for the air bag system alone indicates an increased risk of
injury (not statistically significant). These results do not agree with the previous
analysis by driver height group, where the air bag alone was associated with a
(nonsignificant) 31 percent reduction in risk. However, this may corroborate the allegations
that air bags pose an injury risk for small-statured persons, most notably when they are
unbelted. This weight category contains a disproportionately high share of females.
Only the manual lap-shoulder belt's effectiveness estimates for drivers 135-179 pounds was
statistically significant; although all of the systems appear to provide injury-reducing
benefits of similar magnitudes, compared to being unrestrained.
Both the air bag plus lap-shoulder belt and manual lap-shoulder belts alone provided
significant injury-reducing protection for drivers 180 pounds and heavier. The addition of
the air bag for belted persons appeared to provide further protection against the chance of a
moderate injury (although the difference was not statistically significant).
Summary of Exploratory Analyses
A number of exploratory analyses were conducted to investigate issues of occupant
protection system effectiveness, with the objective of focusing on air bag performance, both
when the driver used available safety belts and when he/she did not. The driver-related
issues for which effectiveness estimates were developed included: body region injured, sex,
age, height, and weight. Many of these estimates failed to reach statistical significance
due to the need to analyze each subset of the data (e.g., males vs. female and driver age
groups). In addition, none of the differences among effectiveness estimates were
statistically significant. However, it is useful, for the purposes of exploratory analysis,
to be able to compare the estimates for various systems and various subpopulations, for the
purposes of future analysis and/or research. Several patterns emerged that may corroborate
the findings of others and may point to the need for further research and analysis.
There appears to be a beneficial trade-off in moderate and serious injury risk between
head injury reduction and an increased risk of upper extremity injury, especially for belted
drivers protected by air bags. The addition of an air bag for a belted driver resulted in
further reduction of moderate head injury but decreased prevention of upper extremity injury
and lower extremity injury, and increased the risk of serious upper extremity injury. At the
same time, the air bag plus safety belt combination provided a reduction in the chance of
moderate and serious chest injury, while the air bag alone appeared to increase the risk of
moderate chest injury, and provide some protection against serious chest injury.
Increased moderate injury prevention was associated with the addition of an air bag to a
manual lap-shoulder belted male driver (38 percent to 64 percent for males, although a
nonsignificant difference), while for female drivers, both the manual lap-shoulder belt and
air bag plus lap-shoulder belt yielded estimated 59 percent effectiveness. This suggests
that belted male drivers may derive greater benefit from the air bag than do belted female
drivers. At the same time, manual lap-shoulder belts appeared more effective in preventing
moderate injury for females (59 percent) than for males (38 percent) compared to each being
unrestrained. The air bag alone appeared to provide about twice the injury-reducing
protection to females (25 percent) as for males (12 percent), although neither estimate was
Belted drivers age 15-49 years appeared to derive greater benefits from the air bag than
did drivers age 50+ years, although the manual lap-shoulder belt was slightly more effective
for these older drivers. The air bag findings match what was found in the analysis of
fatality-reducing effectiveness of air bags. There, significant benefits were noted for
drivers less than 56 years old (32-35 percent), a nonsignificant 25 percent effectiveness for
drivers age 56-69 years, and a nonsignificant 11 percent for drivers age 70 years and older.
The findings for driver height and weight groups appeared somewhat contradictory, since
one should expect a high correlation between driver height and weight. Many of the estimates
in these analyses were not statistically significant. For driver weight, manual lap-shoulder
belts showed the greatest consistency in injury-reducing effectiveness, 42-44 percent for the
three groups, while the addition of an air bag to lap-shoulder belts exhibited additional
protection for the lighter and heavier groups. The air bag alone for lighter weight drivers
showed a nonsignificant increase in risk of moderate injury of 36 percent, which was not
observed for shorter drivers. The inconsistency of the results for the height-weight
analyses is puzzling.
As noted earlier in the section on estimating effectiveness, NHTSA has documented a number
of cases where deployment of the passenger-side air bag resulted in serious injury to a child
passenger. This section discusses, in some detail, the results of in-depth accident
investigation activities conducted by NHTSA as part of its Special Crash Investigation
As of October 1, 1996, NHTSA has identified 31 crashes where deployment of the
passenger-side air bag resulted in critical-to-fatal head or neck injuries to children.
Eleven involved infants in rear-facing child seats, including seven deaths. All but one of
the other 20 children (19 deaths) were determined to have been unrestrained or improperly
restrained at the time of the crashes, all of which involved pre-impact braking, placing the
child in proximity to the deploying air bag. The one exception was a fatal injury to a child
who was wearing the lap-shoulder belt, but it is unknown whether the child was seated in a
correct posture position and if the belts were snug. Pre-impact braking, coupled with
improper or no safety belt use, generally results in the child moving forward into proximity
with the passenger-side air bag prior to the actual crash and subsequent air bag deployment.
The Department of Transportation has formed a coalition with manufacturers, insurance
companies and other organizations to prevent injuries and fatalities which may be
inadvertently caused by air bags, especially to children. The May 21, 1996, press release
announcing the coalition offers the following safety guidelines for child passengers:
"Infants in rear-facing child safety seats should never be placed in the front seat if the vehicle has a passenger-side air bag. The safest place for children of all ages is the back seat. If riding in the back seat is not an option, toddlers and older children may ride in the front seat of a vehicle with a passenger-side air bag, but only if buckled up properly and with the seat moved as far back as possible."
This section of the report provides a discussion of the passenger kinematics and injury
mechanisms associated with the air bag-induced injuries. In all cases, the crash
investigators have identified the passenger air bag and/or cover flap as the source of the
critical-to-fatal injuries. Little to no intrusion of the occupant compartment was reported,
and the vast majority of the crashes would be considered minor or moderate in severity, as
estimated by delta-v (simply put, the crash-induced change in the vehicle's speed).
Delta-v's of less than 10 mph would be considered minor severity, and delta-v's of 11 to 20
mph would be considered moderate severity. Given the level of the crash severities involved,
one would not expect that these children would have sustained critical-to-fatal injuries had
there been no air bag deployment.
The child-air bag problem is most logically broken down into two distinct situations:
infants in rear-facing child seats and children facing forward in the right-front passenger
seat. A discussion of the injury mechanisms from each group are provided below.
All the infants, with one exception, were restrained in an appropriate infant seat, and
the seats were secured to the vehicle. However, this is not considered properly restrained
since one should never place a rear-facing child safety seat in the front seat of a vehicle
equipped with a passenger air bag. In some cases the investigators have pointed out some
departures from owner manual specifications, such as not using a locking clip. However,
these discrepancies probably had little effect on the outcome. In all cases, the vehicle's
driver and/or other adult passengers ignored the warning labels located on the sun visor and
owner's manual and placed the infant in the right-front seat.
The crash scenario for air bag involvement with rear-facing infant seats is similar for
all cases. Upon impact, the deploying passenger air bag interacts violently with the back of
the rear-facing infant seat, typically with sufficient force to crack or break the plastic
shell. The force and rapid acceleration of this impact are carried through the seat and into
the child's head causing skull fractures and associated brain injuries. Cervical spine
(neck) injuries are difficult to diagnose in infants due to the developmental immaturity of
their bones and joints. It is therefore possible that neck injuries existed in these cases
but were not reported.
The crash scenario for air bag involvement with forward-facing children focuses on the
fact that the children are either unrestrained or improperly restrained by the available
safety belt system. All of the cases involved pre-impact braking, which causes the child to
move forward into proximity of the stored air bag prior to deployment. Occupant contact with
the instrument panel prior to deployment has been confirmed for some of the cases by the
identification of tissue, fluid, and/or clothing transfers on the air bag cover flap and/or
Upon impact, the air bag deploys into the out-of-position child's chest, neck, and face.
As the air bag expands, it results in the rapid translation and rotation of the child's
skull, causing a number of injuries. These include fractures of the cervical spine, bruising
and laceration of the spinal cord, and brain stem injuries. Brain injuries are also commonly
reported, but skull fractures were typically not observed. These head injuries are
consistent with large and rapid rotations of the head produced by a large distributed force.
Mandibular (jaw bone) fractures and avulsed (knocked-out) teeth have also been reported as a
result of air bag or cover flap impact with the chin and face. Although not common in this
scenario, injuries to the lungs and heart have been reported.
On May 23, 1995, NHTSA published a final rule (60 FR 27233), effective June 22, 1995, that
allowed manufacturers the option of installing a manual device that motorists could use to
deactivate the front passenger-side air bag (that is, an on-off switch) in vehicles in which
infant restraints can be used in the front seat only. The affected vehicles were passenger
cars and light trucks without rear seats, and vehicles with rear seats that are too small to
accommodate typical rear-facing infant seats. The on-off switch is needed because when
rear-facing infant restraints are used in the front seats of dual air bag vehicles, they
extend forward to a point near the dashboard where they can be struck by a deploying air bag,
with the potential for serious injury or death to the infant.
On October 27, 1995, because of the incidence of several fatalities to improperly
restrained children in air bag-equipped positions, NHTSA issued a strong warning in a press
release, "SAFETY AGENCY ISSUES WARNING ON AIR BAG DANGER TO CHILDREN." It "warned that
children who are not protected by a safety belt could be seriously injured or killed by an
air bag, and in the strongest possible terms urged parents to insist that their children ride
belted in the back seat whenever possible." This release repeated prior agency warnings of
the dangers of placing a rear-facing seat in front of an air bag and
broadened the previous warnings to apply to older children and even adults
who may ride unrestrained. To ensure that infants and children ride safely, with or without
a passenger-side air bag, this warning and advisory urges care givers to follow three
On November 9, 1995, NHTSA published a request for comments to inform the public about its
efforts to reduce the adverse effects of air bags and to invite the public to share
information and views with the agency (60 FR 56554). The request for comments focused on
possible technological changes to air bags to reduce their adverse effects, including
possible regulatory changes.
Since publishing its October 1995 warning and November 1995 request for comments, NHTSA
has intensified its efforts to educate the public about air bag performance and the campaign
to properly restrain children. A large part of the agency's plan is to increase information
to the affected public through the traffic safety community throughout the country. With this
support, the agency will be able to extend the reach of its safety messages to a wider
A few of the agency's many activities include: press releases and media events focusing
on the air bag injury issue; development and distribution of a public service announcement
and video news release; development of articles in a large number of popular publications
such as Redbook magazine and Good Housekeeping; development of articles in
professional periodicals such as the Centers for Disease Control's Morbidity and
Mortality Weekly Report; notification and education of health and safety professionals
through a variety of organizations such as the American Academy of Pediatrics, Emergency
Nurses Association, and the International Association of Chiefs of Police; distribution of
educational materials through state highway safety offices and other traffic safety networks;
initiation and participation in national professional and scientific meetings; and support
for educational efforts conducted by industry and corporate groups such as Morton
International, Midas International and the Network of Employers for Traffic Safety.
NHTSA believes national safety belt use rates can be increased significantly beyond the
current national average of 68 percent. The agency knows, for example, from its own research
and demonstration efforts and the efforts of the insurance and automobile industries, that
three ingredients are essential to increasing safety belt use: (1) strengthening current
state safety belt use laws to allow for primary enforcement; (2) implementing periodic,
highly visible enforcement programs in the states so that the public will know these laws are
important and are being enforced; and (3) conducting public information and education
programs to reinforce these efforts and alert the public to the dangers of riding
unrestrained or improperly restrained.
On May 21, 1996, Secretary of Transportation Federico Peña announced the formation
of a coalition of automobile manufacturers, air bag suppliers, insurance companies, safety
organizations, and the Federal government to prevent injuries and fatalities which may be
inadvertently caused by air bags, especially to children. NHTSA served a central role in
uniting these private-sector partners to form the National Automotive Occupant Protection
Campaign. Coalition members pledged almost $10 million to pursue a three-point program:
NHTSA published a request for comments in November 1995 concerning the need to reduce the
adverse effects of air bags. The request for comments in particular sought information about
possible technological changes to air bags to reduce the adverse effects, including possible
The request for comments noted the agency's belief that, for vehicles manufactured far
enough in the future to incorporate significant design changes, there will be technological
enhancements available that could minimize the adverse effects of air bags. NHTSA noted that
the vehicle manufacturers and air bag suppliers are working on "smart bags," which could
include advanced technologies for occupant sensing, phased deployment of air bags, and so
forth. These technologies will be able to perform a number of functions, including
preventing air bag deployment when they sense that an occupant is too close to the point of
deployment, inflating the air bag at different speeds according to the severity of the crash,
and preventing the passenger-side air bag from deploying when that seat is not occupied.
NHTSA stated that, based on discussions with suppliers and vehicle manufacturers, it
anticipates these types of smart bags will eventually be widely incorporated into production.
The agency indicated that it will step up its monitoring of manufacturer efforts to develop
and use smart bags, the technologies being explored, the practicability and reliability of
smart bag systems, and the timetables for availability of smart bag systems.
While NHTSA anticipates that these smart bag systems will substantially reduce adverse
effects of air bags in the relatively near future, there is still the question of what can be
done in addition to public education for the near future. NHTSA stated that manufacturers
may be able to make adjustments to existing air bag system designs and that the agency may
make temporary adjustments to its regulations if it is shown to be appropriate to enable
manufacturers to reduce any adverse effects during this period.
On November 22, 1996, NHTSA announced a comprehensive approach to preserve the safety
benefits of air bags while minimizing their danger to children and at-risk adults. Its
approach centers on accelerating the development of "smart air bag" technology for future
vehicles with the intent of having systems available for 1999 models. More immediate measures
include adopting enhanced warning labels, depowering of air bags and continuing to allow the
use of on-off switches in vehicles without a rear seat to protect children, and allowing
dealers to deactivate the air bags for any owner who requests it.
On November 27, 1996, NHTSA published a final rule requiring vehicles with air bags to
bear three new warning labels. Two of the labels replace existing labels on the sun visor.
The third is a temporary label on the dash These labels would not be required on vehicles
having a "smart" passenger-side air bag. This rule also requires rear-facing child seats to
bear a new, enhanced warning label.
Basis for Regulatory Actions
Almost all of the previous experience in evaluating air bag effectiveness has been based
on driver-side air bags. While the number of passenger-side air bags has been too small to
conduct statistically meaningful evaluations of their life-saving benefits, this report
provides the first evidence that passenger-side air bags may be as effective for passengers
age 13 years and older as are driver air bags. However, these analyses also demonstrated an
increased fatality risk for passengers age 0-12 years which, due to the limited sample size,
is not possible to accurately quantify. As the dual air bag fleet continues to grow, greater
reliability of findings will become possible. Currently, anecdotal information, located and
developed by NHTSA's Special Crash Investigation program, is available on passenger-side air
Although the safety benefits of air bags are documented, there are situations in which air bags can have adverse effects. As more vehicles have been equipped with air bags, these effects have become better known to researchers. The table below shows, in no particular order, the types of situations in which the agency has some information suggesting that there may be a risk of serious injury to vehicle occupants from the air bag.
As shown in this table, the risks of adverse effects from air bags primarily relate to
occupants who are very near the air bag at the time of deployment.
Current Rulemaking Proposals
The most direct solution to the problem of child fatalities from air bags is for children
to be properly belted and placed in the back seat. This necessitates increasing the
percentage of children who are properly restrained by child safety seats and improving the
current 68 percent rate of safety belt use by a combination of methods, including support for
State primary safety belt laws. The most direct technical solution to the problem of child
fatalities from air bags is the development and installation of smart passenger-side air bags
that automatically protect children from the adverse effects that can occur from proximity to
a deploying bag. However, until these smart air bags can be incorporated introduction
vehicles, behavioral changes based on improved information and communication of potential
hazards and simpler, manually operated technology appear to be the best means of addressing
child fatalities from air bags.
The agency proposed that vehicles without smart passenger-side air bags would be required
to have new, attention-getting warning labels and permitted to have a manual on-off switch
for the passenger-side air bag. NHTSA also proposed to require rear-facing child seats to
bear new, enhanced warning labels. Finally, this notice discussed the agency's research on
other air bag issues, such as technology to reduce arm and other injuries to drivers.
To act on these tentative conclusions, in August 1996, NHTSA published a notice of
proposed rulemaking, proposing amendments to NHTSA's occupant crash protection standard and
child restraint standard to reduce the adverse effects of air bags, especially those on
children. NHTSA proposed the following for passenger cars and light trucks whose
passenger-side air bag lacks smart capability: (1) to require new, enhanced warning labels,
and (2) to permit manual on-off switches for the passenger-side air bags (to accommodate
parents who need to place rear-facing child seats in the front seat). By limiting the
labeling requirement to vehicles without smart air bags, NHTSA hopes to encourage the
introduction of those air bags as soon as possible. NHTSA considers smart passenger-side air
bags to include ones designed so that they automatically avoid injuring the two groups of
children shown by experience to be at special risk from air bags: infants in rear-facing
child seats, and children who are out-of-position (because they are unbelted or improperly
belted) when the air bag deploys.
The agency is also proposing to require vehicles and rear-facing child seats to bear new,
enhanced warning labels. The proposed labels would warn that unbelted children and children
in rear-facing child seats may be seriously injured or killed by the passenger-side air bag.
Vehicle manufacturers and air bag suppliers are working on an array of systems that might
qualify as smart air bags. These systems fall into two categories: (1) ones which would
prevent the air bag from deploying in situations where it might have an adverse effect,
based, for example, on the weight, size and/or location of the occupant, and (2) ones
designed so that they would deploy in a manner that does not create a risk of serious injury
to occupants very near the bag, e.g., deploying at a slower speed when an occupant is very
near the air bag and/or deploying less aggressively as a result of being stowed with an
improved fold pattern.
The Notice of Proposed Rulemaking discusses other issues relating to the introduction of
smart passenger-side air bags and requests comments on whether to assure the timely
introduction of those air bags by requiring their installation, and if so, by what date. As
an alternative, the agency is also requesting comments on whether it should specify an
expiration date for the manual on-off switch option to encourage smart passenger-side air
NHTSA believes that the solution to these problems is smart air bags. The agency is vigorously pursuing the development and implementation of smart air bags for production vehicles. NHTSA is aware of one system that apparently would automatically protect children and that is in production now. This system uses a weight sensor that activates the air bag only if more than a specified amount of weight is on the passenger seat. While this technology is currently being used to prevent the unnecessary and costly deployment of a passenger air bag when no passenger is present, commenters have suggested that the same technology could be used to prevent deployment of the air bag either when no passenger or a child of less than a specified weight (e.g., 30 kilograms or 66 pounds) is present.
The proposed regulatory action should have the following results. Eventually, either
through market forces or government regulation, NHTSA expects that smart passenger-side air
bags will be installed in passenger cars and light trucks to mitigate the adverse effects of
air bags. As stated previously, the agency is proposing that vehicles without smart
passenger-side air bags would be required to have new, attention-getting warning labels and
permitted to have a manual on-off switch for the passenger-side air bag. By limiting the
labeling requirement to vehicles without smart air bags, NHTSA hopes to encourage the
introduction of the next generation of air bags as soon as possible.
Based upon the fatalities and injuries caused by air bags in accidents, the agency initiated an extensive test program at the Vehicle Research and Test Center. Testing has been underway since February of 1996. The program has investigated air bag deployment injuries to both out-of-position child passengers and to small-statured female drivers. Test dummies representative of twelve month, three and six year old children and a 5th percentile female were used for the testing. The improvement to child injury responses with de-powering of the air bag has been explored, as well as the change in adult occupant responses in 30 mile per hour crashes with de-powered air bags. The effects of various air bag/vehicle design features have been explored. The program was performed with industry participation in order to assure broader utility of the test results. The results of the program are being placed into publicly accessible storage to enable the findings to be of use not only to the agency, but to the industry and other interested parties.
In addition to reporting on the effectiveness of occupant protection systems, Section
2508(e) of ISTEA requests that the Secretary of Transportation, in consultation with the
Secretary of Labor and the Secretary of Defense, provide a biennial report of safety belt use
by Federal, state and local law enforcement officers, military personnel, Federal and state
employees other than law enforcement officers, and the public nationally and in each state.
This chapter presents currently available information on safety belt use by a number of these
Safety Belt Use by the Public
Of the groups for which safety belt use information has been requested (i.e., law
enforcement, military, government employees, and the public), the group for which the most
complete information is available is the public. Until late 1991, the NHTSA 19-city survey
had been used to track national safety belt use rate trends. This survey utilized
observations at locations in 19 cities in 16 states across the country to derive an index of
safety belt use which, while not representative of safety belt use nationwide, was useful for
assessing changes in safety belt use over time. With the implementation of NHTSA's National
"70% by '92" Safety Belt Program, the agency replaced the 19-city survey with a measure based
upon an aggregate of individual statewide surveys. This aggregate of state surveys is
weighted by each state's population to produce an estimate of national safety belt use and is
sensitive to changes in each of the 50 states, the District of Columbia, and Puerto Rico.
State safety belt surveys differ in design. Most states utilize probability-based designs
which allow the accuracy of the survey to be calculated. Others utilize convenience samples,
which may be reasonable indicators of belt use but do not have a known degree of accuracy or
Twenty-eight states (comprising about 73 percent of the U.S. population) conduct
probability-based surveys that have been reviewed by NHTSA and meet the minimum design
standards that were established by the Section 153 incentive grant program. These states
requested that their survey designs be reviewed either to be eligible for the incentive
grants or to ensure that their designs were meeting the latest standard of quality.
Another 11 states (comprising 19 percent of the U.S. population) conduct probability-based
surveys, but have not demonstrated their compliance with the established guidelines. Seven
states (comprising 5 percent of the U.S. population) conduct extensive convenience samples
and the remaining 5 states (3 percent of the population) either conduct less extensive
surveys or use crash data to estimate state safety belt use.
In addition to the statewide surveys, in late 1994 NHTSA conducted a National Occupant
Protection Use Survey, which provides additional information about occupant protection system
use by the public. The following discussions address the results of both the state surveys
and findings from NHTSA's national survey.
State-Based Survey Use Estimates
The statewide surveys vary in methodology and frequency of observation. However, all
except Wyoming are based upon direct observation of safety belt use. Wyoming's data are
based upon accident reports. Although many of these state surveys do not constitute
probability samples (that is, statistically designed surveys) of the state, they are
generally based on a large number of observations from representative sites and provide a
reasonable estimate of safety belt use.
ISTEA provided an opportunity to obtain better documentation of state surveys of safety
belt use and to improve their survey methodology. To be eligible to receive second-year
incentive grant money under Section 153 of the legislation, states were required to utilize
probability-based safety belt use survey designs. Safety belt survey guidelines were
developed by NHTSA and published in the Federal Register on June 29, 1992.
According to these guidelines, a state was required to measure safety belt use with an
observational survey wherein observation sites were a probability sample of sites in the
state. The sites had to represent all areas in the state, including urban and rural areas,
and the resulting safety belt use estimates must meet a specified level of statistical
precision. The survey designs and subsequent estimates were to be adequately documented and
submitted for approval by NHTSA. Approved surveys provide statewide representative estimates
of safety belt use with a known margin of error.
Exhibit 17 lists the most recent state use rates as of December 1995. Thirty-six of these
surveys were conducted in 1995, eleven were conducted in 1994 and five in 1993. The
population-weighted national average of these rates is 68 percent. Use rates range from 40
percent in South Dakota to 86 percent in New Mexico. Differences between state use rates are
possibly a reflection of variations in safety belt use laws, enforcement of these laws, and
demographics. States with laws requiring safety belt use, usually by drivers and front-seat
passengers, typically have substantially higher use rates than do states without such laws.
Forty-nine states, the District of Columbia, and Puerto Rico had mandatory use laws in place
at the end of 1995. The population-weighted average use rate in "law" states was 68 percent.
The use rate in New Hampshire, the only remaining non-law state, was 57 percent.
The manner in which the law is enforced also affects use rates. A primary
enforcement safety belt law allows officers to make a traffic stop based solely upon
observation of a safety belt law violation. A law which prescribes secondary
enforcement may only be enforced after a traffic stop has been made for another purpose.
The average use rate in the eleven states which have laws that permit primary enforcement is
14 percentage points higher than the average of those states which permit only secondary
enforcement. The average use rate among primary law enforcement states in 1995 was 75
percent. The average use rate among secondary law enforcement states in 1995 was 61 percent.
|State||Effective Date||Enforcement||Fine||Seats|| Vehicle and Coverage by Law
*Latest Usage Rate (%)
|Alabama||July 18, 1992||Secondary||$25||Front||Passenger car from model year 1965.||52|
|Alaska||September 12, 1990||Secondary||$15||All||Motor vehicle. Over age 16.||69|
|Arizona||January 1, 1991||Secondary||$10||Front||Passenger car and van from model year 1972.||60|
|Arkansas||July 15, 1991||Secondary||$30||Front||Passenger car, truck, and van.||51|
|California||January 1, 1986||Primary||$20||All||Passenger car, van, and small truck.||85|
|Colorado||July 1, 1987||Secondary||$15||Front||Passenger car, van, taxi, ambulance, RV and small truck.||56|
|Connecticut||January 1, 1986||Primary||$37||Front||Passenger car, van, and truck.||72|
|Delaware||January 1, 1992||Secondary||$20||Front||Passenger car.||60|
|Dist. of Columbia||December 12, 1985||Secondary||$15||Front||Vehicle seating 8 or less people.||63|
|Florida||July 1, 1986||Secondary||$20||Front||Motor vehicle and pickup truck.||59|
|Georgia||September 1, 1988||Primary**||$15||Front||Passenger vehicle for under 10 people and pickup for under age 18||53|
|Hawaii||December 16, 1985||Primary||$20||Front||Vehicle registered in state.||80|
|Idaho||July 1, 1986||Secondary||$5||Front||Motor vehicle under 8 thousand pounds.||59|
|Illinois||July 1, 1985||Secondary||$25||Front||Motor vehicle to carry under 10 people and RV.||69|
|Indiana||July 1, 1987||Secondary||$25||Front||Passenger car, bus, and school bus.||64|
|Iowa||July 1, 1986||Primary||$10||Front||Passenger car, van, and truck 10 thousand pounds or less.||76|
|Kansas||July 1, 1986||Secondary||$10||Front||Passenger car and van.||54|
|Kentucky||July 13, 1994||Secondary||$25||All||Motor vehicles from model year 1965.||52|
|Louisiana||July 1, 1986||Primary||$25||Front||Passenger car, van, and truck under 6 thousand pounds.||59|
|Maine||December 27, 1995||Secondary||$25||All||Passenger vehicles.||50|
|Maryland||July 1, 1986||Secondary||$25||Front||Passenger/multi-purpose vehicle, truck, tractor, and bus.||70|
|Massachusetts||February 1, 1994||Secondary||$25||All||Passenger car, van, and truck.||53|
|Michigan||July 1, 1985||Secondary||$25||Front||Motor vehicle.||67|
|Minnesota||August 1, 1986||Secondary||$25||Front||Passenger car, pickup truck, van, and RV.||65|
|Mississippi||March 20, 1990||Secondary||$25||Front||Passenger car and van.||46|
|Missouri||September 28, 1985||Secondary||$10||Front||Passenger car to carry under 10 people.||71|
|Montana||October 1, 1987||Secondary||$20||All||Motor vehicle.||70|
|Nebraska||January 1, 1993||Secondary||$25||Front||Motor vehicle.||64|
|Nevada||July 1, 1987||Secondary||$25||All||Passenger car under 6 thousand pounds.||71|
|New Jersey||March 1, 1985||Secondary||$20||Front||Passenger car.||61|
|New Mexico||January 1, 1986||Primary||$25||Front||Motor vehicle under 10 thousand pounds.||86|
|New York||December 1, 1984||Primary||$50||Front||Passenger car.||72|
|North Carolina||October 1, 1985||Primary||$25||Front||Passenger motor vehicle to carry under 10 people.||81|
|North Dakota||July 14, 1994||Secondary||$20||Front||Motor vehicle.||42|
|Ohio||May 6, 1986||Secondary||$25||Front||Passenger/commercial car, van, tractor, and truck.||63|
|Oklahoma||February 1, 1987||Secondary||$10||Front||Passenger car, van, and pickup truck.||46|
|Oregon||December 7, 1990||Primary||$95||All||Motor vehicle.||80|
|Pennsylvania||November 23, 1987||Secondary||$10||Front||Passenger car, truck, and motor home.||71|
|Puerto Rico||January 19, 1975||Primary||$10||Front||Passenger car. Over age 4.||62|
|Rhode Island||June 18, 1991||Secondary||No||All||Passenger car. Over age 12.||58|
|South Carolina||July 1, 1989||Secondary||$10||Front||Passenger car, truck, van, RV, and taxi.||64|
|South Dakota||January 1, 1995||Secondary||$20||Front||Passenger car, truck, van, RV, and taxi.||40|
|Tennessee||April 21, 1986||Secondary||$25||Front||Vehicle under 8.5 thousand pounds.||64|
|Texas||September 1, 1985||Primary||$25||Front||Passenger car, van, and certain trucks.||72|
|Utah||April 28, 1986||Secondary||$10||Front||Motor vehicle.||56|
|Vermont||January 1, 1994||Secondary||$10||All||Passenger car.||67|
|Virginia||January 1, 1988||Secondary||$25||Front||Motor vehicle.||70|
|Washington||June 11, 1986||Secondary||$25||All||Passenger/multi-purpose vehicle, bus, and truck.||83|
|West Virginia||September 1, 1993||Secondary||$25||Front||Passenger car. Age 18 and under in rear seat.||58|
|Wisconsin||December 1, 1987||Secondary||$10||All||Motor vehicle.||64|
|Wyoming||June 8, 1989||Secondary||No||Front||Passenger car, van, and pickup truck.||NA|
| Total Use Laws: 49 States + D.C. and Puerto Rico|
*Reported December 1995
The national use rate estimate of 68 percent is considerably higher than the estimated 11
percent rate in 1980. The increase reflects the substantial emphasis placed on increasing
safety belt use among the general public over the past decade, as well as the implementation
of the amendment to Federal Motor Vehicle Safety Standard 208 requiring automatic occupant
protection systems on a phased-in schedule beginning in model year 1987. While voluntary
methods (i.e., public information and education alone) increased national use to about 15
percent by 1984, it was the enactment of state safety belt use laws that provided the
increase to about 50 percent by 1990. Without highly publicized enforcement, however, most
state use rates stabilized at about 50 percent. With the implementation of the National "70%
by '92" Safety Belt Program in early 1991, these rates again started to rise. This rise is
thought to be due primarily to the increased emphasis on enforcement and public
information/education by state safety belt programs.
The National "70% by '92" Safety Belt Program consisted of two primary components: (a) two
summer public awareness campaigns which were conducted during 1991 and 1992, with special
emphasis on enforcement during the three summer holiday periods (Memorial Day, July 4th, and
Labor Day); and (b) Operation Buckle Down, a program using peer-to-peer outreach to increase
law enforcement officer use of safety belts and to increase the effectiveness of safety belt
and child passenger safety law enforcement by officers.
Large increases in safety belt use occurred while the "70% by '92" program was in effect.
By the end of 1992, 23 states had increased their use rate by 10 percentage points or more,
compared with their 1990 level. Another 11 states increased their rate by 5-9 points. A
total of 14 states reported use rates of 70 percent or greater, compared to one state in
1990. The national use rate, as measured by the population-weighted average of state survey
results, increased to 62 percent by the end of 1992 (an 8-10 percentage point increase since
Many states maintained activities that were initiated during the "70% by '92" program
after the program was completed at the end of 1992. As a result of these efforts and of the
introduction of new and improved legislation in several states, the national average of
surveys reported by the states continued to increase, reaching 68 percent by the end of 1995.
Building on the success of the "70% by '92" program, NHTSA initiated a new combined safety
belt and alcohol program in the fall of 1994. This new initiative, known as Campaign Safe
& Sober, complements highly publicized enforcement, as used in the "70% by '92" program,
with legislative and public information and education (PI&E) strategies. The goals for
Campaign Safe & Sober are to achieve at least 80 percent safety belt use by the year 2000
and to reduce alcohol-related fatalities to 11,000 by 2005.
NHTSA's National Occupant Protection Use Survey (NOPUS)
NOPUS was conducted in the Fall of 1994. The previous report to Congress documented the
results of the 1994 survey. At the current time, data collection is underway for the 1996
NOPUS and therefore, no new information is yet available. However, the information presented
in the previous report has been repeated in this report for completeness.
NOPUS is composed of three separate studies: the moving traffic study
which provides information on overall shoulder belt use, the
controlled intersection study which provides more detailed
information about shoulder belt use by type of vehicle, characteristics of the belt users and
child restraint use, and the shopping center study which provides
information on rear-seat belt use and shoulder belt misuse. This note presents the results
from the moving traffic study. Results from the other studies will be released as they become
Shoulder belt use observed in the moving traffic study was 62.8 percent for passenger car occupants and 50.2 percent for light truck occupants.
Data collection from the moving traffic study was conducted at almost 4,000 randomly
selected sites across the country in October and November 1994. Pairs of observers were
stationed for 30 minutes at exit ramps, intersections with stop signs, and stop lights, and
uncontrolled intersections. One observer counted shoulder belt use for the drivers of
passenger cars and light trucks (vans, minivans, sport utility vehicles, and pick-up trucks).
The second observer counted shoulder belt use for the right front passengers of cars and
light trucks and helmet use for motorcycle riders and passengers. Every day of the week and
all daylight hours (8 a.m. to 6 p.m.) were covered by the study. Commercial and emergency
vehicles were excluded.
NOPUS was designed as a multi-stage probability sample to ensure that the results would
represent occupant protection use in the country. In the first stage, counties were grouped
by region (northeast, midwest, south, west), level of urbanization (metropolitan or not), and
level of belt use (high, medium, or low). Fifty counties or groups of counties were selected
based on the vehicle miles of travel in those locations. In the next stage, roadways were
selected from two categories: major roads and local roads. Finally, approximately 4,000
intersections or exit ramps were chosen on these roadways. Of the originally selected sites,
some were found to be ineligible during mapping and data collection, and at some sites no
vehicles were observed. A total of over 167,000 passenger cars, almost 84,000 light trucks,
and 997 motorcycles were observed.
Detailed results of the moving traffic study are presented on the following page. Each
estimate has been statistically weighted according to the sample design. Since these are
estimates from a sample, each has an associated margin of error or standard error. Two
standard errors are given in parentheses next to each estimate. By simply adding and
subtracting the standard errors from the estimates, an approximate 95 percent confidence
interval can be created. This means that you can be 95 percent sure that the true use rate
lies within this interval.
1994 NOPUS: Moving Traffic Results by Region
(Estimates and 2 Standard Errors)
|Shoulder Belt Use (%)||58.0 (3.8)||55.1 (8.2)||59.1 (5.8)||55.4 (6.4)||63.3 (8.4)|
|Passenger Cars (%)||62.8 (3.8)||57.5 (8.4)||63.9 (5.6)||60.7 (7.0)||69.1 (7.6)|
|Car Drivers (%)||64.2 (3.6)||57.8 (7.8)||65.3 (5.8)||62.2 (6.8)||71.2 (7.0)|
|Car Passengers (%)||59.1 (4.4)||56.8 (10.8)||59.8 (5.4)||57.5 (7.2)||63.5 (9.8)|
|Light Trucks (%)||50.2 (3.6)||47.8 (9.4)||50.6 (6.6)||47.9 (5.4)||54.9 (9.2)|
|Truck Drivers (%)||50.7 (3.8)||46.8 (9.8)||51.4 (7.0)||47.6 (6.6)||56.8 (8.4)|
|Truck Passengers (%)||49.1 (3.6)||50.5 (9.0)||48.4 (6.0)||48.7 (4.0)||50.2 (11.2)|
1994 NOPUS: Moving Traffic Results by Day of Week and Time of Day
(Estimates and 2 Standard Errors)
|Weekday||Weekend||Rush Hour||Non-Rush Hr|
|Shoulder Belt Use (%)||58.4 (3.6)||57.0 (5.8)||58.0 (8.0)||58.0 (3.6)|
|Passenger Cars (%)||63.8 (3.2)||60.0 (6.6)||63.2 (7.0)||62.7 (4.0)|
|Car Drivers (%)||65.2 (3.2)||61.2 (6.2)||65.4 (5.8)||63.9 (3.6)|
|Car Passengers (%)||59.7 (3.6)||58.1 (7.0)||56.0 (11.0)||59.5 (4.6)|
|Light Trucks (%)||49.3 (4.2)||52.2 (4.0)||49.6 (8.8)||50.3 (3.2)|
|Truck Drivers (%)||50.0 (4.2)||52.3 (4.4)||51.9 (7.8)||50.4 (3.6)|
|Truck Passengers (%)||46.9 (5.0)||52.1 (4.0)||43.0 (12.0)||50.1 (2.8)|
|Helmet Use (%)||85.9 (7.8)||55.2 (9.0)||91.2 (10.4)||60.6 (15.0)|
Direct comparison of findings between the NOPUS and state surveys is difficult, primarily
because of the differences in vehicle and occupant coverage. However, a rough comparison of
overall use can be made between the state-based estimate for 1994 of 67 percent and the NOPUS
estimate for passenger car drivers and passengers of 63 percent. In this comparison, the
state based estimate falls within the 95 percent confidence interval of the NOPUS estimate.
The combination of surveys that have been used to measure safety belt usage over the past
several years also provides us with some insight with regard to change in usage rates. Until
1990, the 19 cities survey was used as the index of national use. In 1990, that index for
passenger car drivers was 49 percent. The NOPUS estimate of belt use among passenger car
drivers in 1994 is 64 percent. The difference of 15 percentage points between the 19 cities
index and the NOPUS estimate is consistent with the 14 percentage point change in usage
indicated by the aggregate of state surveys between 1990 and 1994 (i.e., 53 percent in 1990
and 67 percent in 1994).
State surveys provide an essential source of information for monitoring progress in the
states. The NOPUS provides a probability-based sample of national usage with the ability to
estimate error. In addition, the NOPUS provides a unique source of detailed information
concerning restraint use by vehicle type, age, gender, shoulder belt misuse, etc. Plans for
repeating the NOPUS survey will be based upon the frequency of need for this level of
analysis. Annual estimates of belt use progress will continue to be made with the
Safety Belt Use by Federal Employees
On September 26, 1986, President Reagan signed Executive Order 12566, "Safety Belt Use
Requirements for Federal Employees," requiring Federal employees to use safety belts when
traveling on official Government business. Since that time, NHTSA has been actively involved
in promoting safety belt programs with Federal agencies and the military.
Prior to 1989, NHTSA conducted briefing sessions with all departments and agencies and
provided promotional materials to support the safety belt program. Other activities included
addressing special meetings at Federal agencies and suggesting activities to motivate
employees and family members concerning the importance of safety belt and child safety seat
laws. A brief description of some of the more recent efforts follows.
In 1989, NHTSA sponsored a "Saved-By-The-Belt" program that recognized Federal employees
and their family members who had been survivors of a motor vehicle crash. NHTSA presented
the recipients with a "Saved-By-The-Belt" certificate.
In 1990, NHTSA redesigned the "Saved-By-The-Belt" program to include a plaque for "The
Right Choice Award" for making the right choice to buckle up. For participating agencies
receiving the Right Choice Award, a ceremony was held at the Department of Transportation
highlighting the recipients who were crash survivors.
In 1991, as part of the National "70% by '92" safety belt program, NHTSA urged Federal
departments and agencies, including military installations, to participate in the goal set by
the President to reach 70 percent safety belt use among Federal employees by the end of 1992.
NHTSA developed the "Federal Employees Buckle Up" kit and distributed over 4,000 copies of
the promotional kits to Federal agencies. Recognizing that safety belt use is mandatory
while traveling on official business, the objective of this effort was to extend the safety
belt use habit to personal travel.
Finally, NHTSA has encouraged Federal and military agencies to participate in awards
programs that have provided recognition to agencies, organizations, and cities for reaching
specified belt use rate goals. The "70% PLUS" program was established in 1990 to recognize
achievement of use rates in excess of 70 percent. By the end of the program in 1992, more
than 200 Federal departments and agencies, including military installations, were awarded
"70% Award Plaques" and a thank you letter from NHTSA for achieving 70 percent safety belt
In 1993, the "70% PLUS" program was superseded by the "National Safety Belt Honor Roll"
which provides bronze, silver and gold awards for achieving 70 percent, 80 percent, and 90
percent belt use, respectively. During the first year of this program, awards were made to
five Federal agencies and to 70 military organizations. The program is now managed by the
NHTSA Regional Offices with additional awards made each year.
Obviously, these awards cannot be interpreted as a probability-based or representative
sample of safety belt use rates in the Federal Government. However, they provide anecdotal
evidence of high rates of safety belt use at some Federal employment sites.
NHTSA intends to continue its educational and incentive programs for Federal employees.
It seems clear that, for many agencies, the program has been useful for increasing employee
safety belt use rates.
Safety Belt Use by State and Local Government Employees
Few data are currently available for estimating safety belt use among state and local
government employees other than law enforcement officers. Although this population is
included in surveys of the general population, specific surveys of use among this group have
not been routinely or consistently conducted. In lieu of consistent survey data, the
following anecdotal information presents examples of policies and awards which provide some
indication of belt use practices among this group.
In the northeast, the States of Connecticut, Maine, Massachusetts, Rhode Island, and
Vermont have safety belt use policies covering all state employees.
In New York, over 100 state agencies have received 70, 80, or 90 percent belt use awards
since 1992. The majority of these were in the mid-to upper-80 percent range.
In the mid-Atlantic area, the Governor of Delaware issued an Executive Order requiring all
state employees to use safety belts while on government business. The Virginia Department of
Alcohol Beverage Control received an award for reaching 80 percent belt use.
In the southeast, more than 20 state and local government agencies have qualified for 70
percent belt use awards. The Florida Department of Highway Safety and Motor Vehicles has
conducted safety belt educational programs in more than 100 government agencies. South
Carolina's Department of Highways and Public Transportation has established a similar
educational program. The Mississippi Department of Transportation conducted a belt use
program among employees, indicating a use rate of 87 percent.
In the Midwest, Illinois, Indiana, Minnesota, Michigan, Ohio, and Wisconsin have
established belt use policies covering all state employees.
In the southwest, New Mexico instituted a joint belt use program among state and Federal
employees, achieving use rates above 80 percent. The Texas Departments of Health and
Transportation conducted belt use campaigns achieving rates of 70 to 80 percent.
The western states of Alaska, Arizona, California, Idaho, Oregon, and Washington have
established safety belt use policies covering all state employees. In addition, many local
government agencies in these states have instituted similar policies.
Safety Belt Use by State and Local Police
NHTSA has worked with the International Association of Chiefs of Police (IACP) to develop
a model program that law enforcement agencies can follow to increase public and officer use
of occupant protection. This program has been promoted as part of "Operation Buckle Down"
(OBD) by IACP, the National Sheriff's Association, the Fraternal Order of Police, the
International Association of Directors of Law Enforcement Standards and Training (IADLEST),
the National Association of Governors' Highway Safety Representatives (NAGHSR), and the Law
Enforcement Television Network (LETN).
To help implement this program nationwide, NHTSA awarded grants to 46 states plus the
District of Columbia and Puerto Rico since late 1990 to assist them in conducting an OBD
program. Several other states have used Section 402 funds to support such programs.
Concurrently, NHTSA developed and promoted several law enforcement training courses and
numerous videotaped presentations suitable for officer roll call training. The training
courses include an Occupant Protection Usage and Enforcement (OPUE) Course for officers, an
OPUE course to train police trainers to conduct the officer course, and an Occupant
Protection Risk Management Course targeted to inform law enforcement executives and
administrators about the risks their officers and agencies face from traffic crashes--and how
safety belt use can reduce those risks. Hundreds of these courses have been conducted across
the country. To institutionalize the courses within ongoing law enforcement in-service and
academy training programs, NHTSA works with national and state level police training and
certification organizations to obtain certification status for OPUE training.
To support individual agency roll call training on occupant protection, NHTSA has
distributed thousands of the roll call video tapes to law enforcement agencies directly,
through the state highway safety offices and the state OBD programs. NHTSA has also
collaborated with the LETN to include segments of the training courses and the roll call
tapes within their programming schedule. As a result, these messages have been aired
hundreds of times to the over 3,000 subscribers to the service.
Finally, NHTSA is working with IACP to recognize the lifesaving benefits of officer safety
belt use and to provide information about all these programs to police chiefs and other law
enforcement executives. With NHTSA's support, IACP conducts its "National Law Enforcement
Saved By The Belt/Air Bag Awards Program" which has presented awards to 587 officers as of
May 1996. This program differs from the civilian equivalent in that it is based on actual
police accident or incident reports and the recipient is only eligible if he or she is not
judged to be "at fault" in the crash. Also with NHTSA's assistance, IACP is publishing and
distributing monthly newsletters, the Buckle Down Dispatch, to heads of agencies
across the country. These newsletters reach over 5,000 chiefs of police and other law
enforcement executives at the state and local levels with information about the importance of
safety belt use by the public and by their officers.
1. The AIS, first developed by the Association for the Advancement of Automotive Medicine in 1971, is a consensus-derived, anatomically based system that ranks individual injuries by body region on a scale of 1 to 6 as follows: 1=minor, 2=moderate, 3=serious, 4=severe, 5=critical, and 6=maximum/currently untreatable. The AIS is intended as a measure of the severity of the injury itself and not as a measure of impairments or disabilities that may result from the injury. It does not assess the combined effects of multiple injuries to a patient. The AIS was revised and updated several times, with the most recent revision in 1990. In this report, reference to an AIS level (i.e., AIS of 2 or greater) refers to the maximum AIS level for any injury suffered by a vehicle occupant; that is, the most severe injury. MAIS represents the maximum injury severity (AIS) of any injury received by a person, regardless of the nature or location of the injury.
2. First Report to Congress, Effectiveness of Occupant Protection Systems and Their Use, DOT-HS-808-019, January 1993.
3. Fatality Reduction by Air Bags -- Analyses of Accident Data through Early 1996, DOT-HS-808-470, August 1996.