Content uploaded by Eric R Teoh
Author content
All content in this area was uploaded by Eric R Teoh on Sep 14, 2017
Content may be subject to copyright.
Progress in teenage crash risk during the last decade
Susan A. Ferguson
a,
⁎, Eric R. Teoh
b
, Anne T. McCartt
b
a
Ferguson International LLC, 1328 Lancia Drive, McLean, VA 22102, USA
b
Insurance Institute for Highway Safety, 1005 North Glebe Road, Arlington, VA 22201, USA
Available online 28 March 2007
Abstract
Objective: The purpose of the present study was to examine the most recent data on teenagers' fatal and nonfatal crashes in the United States
to determine current crash rates as well as changes in crash rates during the past decade Methods: Data for calendar years 1996 and 2005 were
extracted for fatal crashes from the Fatality Analysis Reporting System and for police-reported crashes from the National Automotive
Sampling System/General Estimates System. To calculate crash rates, population data were obtained from the Census Bureau, and mileage
data were obtained from the 2001 National Household Travel Survey Results: During 2001–02, the latest year for which mileage data are
available, 16 year-old drivers had higher fatal and nonfatal crash rates per mile traveled than all but the very oldest drivers. However, fewer
16 year-olds typically are licensed to drive and they drive fewer miles per year than all but the oldest drivers. Thus, their fatal and nonfatal
crash rates per population in 2005 were lower than among other teenagers and among drivers 20–29. During the past decade the most
progress has been made in reducing crashes among the youngest drivers. Between 1996 and 2005 both fatal and police-reported crashes per
population declined about 40% for 16 year-old drivers, compared with about 25% for 17 year-old drivers and 15–19% for 18 year-old
drivers. The greatest reductions for 16 year-olds occurred in nighttime crashes, alcohol-related fatal crashes, and fatal crashes involving
multiple teenage passengers. Conclusions: Substantial progress has been made in reducing fatal and nonfatal crashes per population among
16 year-old drivers. Although this study was not designed to examine the role of graduated licensing, the results are consistent with the
increased presence of such laws, many of which restrict nighttime driving and driving with teenage passengers. Impact on Industry:
Restrictions on nighttime driving and driving with teenage passengers should be made a part of all states’graduated licensing systems.
Historically, 16 year-olds have had the highest crash risk per licensed driver and per mile traveled. Given the dramatic reductions in per
population crash rates among 16 year-olds, it is possible that their per mile and per licensed driver rates also have declined and may no longer
be as elevated relative to other ages. However, shortcomings in the licensed driver data and a lack of recent mileage data hamper our ability to
examine these issues. If we are to continue to provide a yardstick against which we can measure progress among the youngest drivers,
immediate steps need to be taken to restore the availability of reliable exposure data.
© 2007 National Safety Council and Elsevier Ltd. All rights reserved.
1. Introduction
Teenage drivers, especially 16 year-olds, have long been
considered to pose the greatest risk to themselves and other
road users, and for good reason. Historically, their fatal crashes
per mile traveled and per licensed driver have been higher than
for all but the very oldest drivers (Li, Braver, & Chen, 2003;
Lyman, Ferguson, Braver, & Williams, 2002). Beginning dri-
vers of all ages have higher crash rates than drivers with more
experience, but the youngest drivers are known to take addi-
tional risks because of youthful exuberance and immaturity
(Mayhew, Simpson, & Pak, 2003).
Teenage crash risk is particularly elevated when driving at
night and when carrying teenage passengers (Chen, Baker,
Braver, & Li, 2000; Doherty, Andrey, & MacGregor, 1998;
Preusser, Ferguson, & Williams, 1998; Ulmer, Williams, &
Preusser, 1997; Williams, 2003; Williams, Ferguson, & Wells,
2005). Legislators in many states have tried to address these
concerns through the adoption of graduated driver licensing
(GDL) laws. Starting with Florida in 1996, U.S. states have
modified their laws requiring beginning drivers to complete a
series of stages before obtaining full-privilege licenses. The
majority of states now have some form of graduated licensing.
Typically after a learner's permit period of at least 6 months,
Journal of Safety Research 38 (2007) 137 –145
www.nsc.org
⁎Corresponding author. Tel.: +1 703 847 5317.
E-mail address: fergsusan@gmail.com (S.A. Ferguson).
www.elsevier.com/locate/jsr
0022-4375/$ - see front matter © 2007 National Safety Council and Elsevier Ltd. All rights reserved.
doi:10.1016/j.jsr.2007.02.001
teenagers are given restricted licenses that often prohibit driving
unsupervised at night or carrying one or more teenage passen-
gers (Insurance Institute for Highway Safety [IIHS], 2006a).
Once these stages have been negotiated successfully, teenagers
can obtain unrestricted licenses. Research examining the effec-
tiveness of graduated licensing at the state level, as well as
comprehensive evaluations of national fatal crash data, have
demonstrated consistently that GDL programs are effective in
reducing crashes among the youngest drivers (see Shope, 2007).
Williams (2003) examined crash data from 2000 and crash
rates per mile traveled from 1995 to compare the crash risk of
teenage drivers with that of drivers of other ages. He
concluded, “On most measures, crash rates during the teenage
years are higher than for any other age, for both males and
females.”To the extent possible the present paper will answer
the question of whether this conclusion still is true.
Recent research suggests that we have seen significant
progress in the United States toward reducing fatal crashes
per population among 16 year-old drivers. Williams et al.
(2005) found that crashes per population declined more for
16 year-old drivers than for drivers of other ages. Between
1993 and 2003, fatal crashes per population declined 26%
for 16 year-old drivers, 11% for 17 year-old drivers, 6% for
18–19 year olds, and 7% for drivers ages 20–49. There were
few differences over time in the percentages of fatal crashes
that occurred at night, or involved single vehicles, driver
error, or speeding. However, the proportion of fatal crashes
in which 16-year-old drivers were transporting only teenage
passengers decreased from 53% in 1993 to 44% in 2003.
The present paper provides updated information on
teenage driver crash rates, including rates for fatal and
nonfatal crashes. Ideally, a comprehensive evaluation of
crash rates and crash rate trends would include calculations
per population, per licensed driver, and per mile traveled.
However, examination of annual state-by-state counts of
licensed drivers published by the Federal Highway Admin-
istration revealed that teenage drivers have been under-
counted in some states for the past several years (Insurance
Institute for Highway Safety, 2006b). Researchers found
large year-to-year differences in licensed driver counts in a
number of states, and these differences could not be
explained by fluctuations in population or changes in state
licensing laws. The IIHS concluded that these current and
historical data do not provide reliable measures of the
number of licensed drivers, especially the youngest ones.
2. Methods
Teenage drivers' crash risk, for both fatal and nonfatal
crashes, was examined using the latest available data from a
variety of sources. In-depth mileage data were obtained from
the National Household Travel Survey (NHTS), formerly the
Nationwide Personal Transportation Survey. NHTS collects
mileage data from a sample of drivers of all ages, including
data on travel at different times of the day. Samples can be
weighted to provide national estimates. Surveys typically are
conducted about every 5 years, but the latest survey was
conducted during April 2001–May 2002.
Data on fatal crashes were obtained from the Fatality
Analysis Reporting System (FARS). FARS is a virtual census
of fatal motor vehicle crashes that occur on public roads in the
United States in which vehicle occupants or other road users
die within 30 days of a crash. Data on police-reported crashes
were obtained from the National Automotive Sampling
System/General Estimates System (NASS/GES). NASS/
GES contains a representative sample of U.S. police-reported
crashes that can be weighted to provide national estimates.
Crashes were examined as a function of driver gender and age.
Fatal crash characteristics examined included driver error,
speeding, crash type, and drivers' blood alcohol concentra-
tions (BAC). The definition of driver error was based on
driver-related factors coded in FARS and was intended to
capture the physical movements of the vehicle that were
clearly indicative of driver mistakes (e.g., improper lane
change, passing on the wrong side, overcorrecting). Codes for
driver physical/mental conditions such as inattention, drows-
iness, or alcohol impairment were excluded. Speeding was
defined to include cases in which the driver was cited for
speeding or in which driver-related factors coded indicated
speed as a factor (e.g., driving too fast for conditions, racing).
2.1. Current Crash Rates
Fatal crash rates per 100 million miles traveled were
calculated using 2001–02 FARS data, and police reported
crash rates per 1 million miles traveled were calculated using
2001–02 NASS/GES data. Fatal and police-reported crashes
in 2005 per population were calculated using population
estimates from the U.S. Census Bureau. All rates were based
on passenger vehicle drivers' fatal and police-reported crash
involvement. Rates were examined by passenger presence
and by time of day —nighttime (9 p.m. to 5:59 a.m.) versus
daytime (6 a.m. to 8:59 p.m.).
2.2. Crash Rate Changes
Changes in crash rates were examined by comparing rates
for 2005 with those for 1996, the period during which the
majority of U.S. states adopted graduated licensing laws for
the youngest drivers. Because of limitations in the driver
licensing data and a lack of recent mileage data, it was not
possible to examine all of the measures of interest to
determine whether driver crash risk had changed during this
period. Thus the analyses examined changes in per
population crashes as well as changes in the underlying
crash characteristics such as presence of alcohol, time of day,
and passenger presence.
3. Results
In 2005, a total of 3,889 passenger vehicle occupants ages
16–19 were killed on U.S. roads, and an estimated
138 S.A. Ferguson et al. / Journal of Safety Research 38 (2007) 137–145
1.89 million were involved in police-reported crashes. This
is 8% fewer deaths (4212) and 20% fewer police reported
crashes (2.37 million) than occurred in 1996 for this age
group.
3.1. Crashes per Mile Traveled during 2001–02
Table 1 shows the estimated average number of miles driven
per year by driver age during 2001–02. Teenage drivers,
especially 16 year-olds, drove fewer miles per year than all but
the oldest drivers (85 and older). Sixteen year-olds drove an
average of about 7,000 miles per year, compared with almost
10,000 miles for 17 year-olds. Drivers ages 25–54 drove an
average of about 16,500 to 17,500 miles per year, more than
twice as many miles as 16 year-olds.
Crashes per mile traveled represent crash risk for a
quantifiable amount of exposure. Fig. 1 shows fatal crashes
per 100 million miles traveled by driver age and gender
during 2001–02, corresponding with the period during
which the mileage data were collected. Fatal crash rates were
highest for teenage drivers (ages 16–19) and drivers 80 and
older, and lowest for drivers ages 30–74 (rates among the
oldest drivers were based on relatively few respondents, so
estimates may be less reliable). However, factors contribut-
ing to the high fatal crash rates differ between the youngest
and oldest drivers. The high fatal crash rates among teenage
drivers are due primarily to their youthful age and inex-
perience. Older drivers' high fatal crash rates result primarily
from their physical fragility. That is, older drivers are more
likely to be injured in crashes and are more likely subse-
quently to die from their injuries (Li et al., 2003).
Among teenage drivers, 16 year-olds had the highest fatal
crash rate, about 6 times higher than the rate for the lowest risk
drivers. Among 16 year-olds, males had the higher fatal crash
rate (11 fatal crashes per 100 million miles traveled versus 7 per
100 million miles for 16 year-old females, and 2 per 100 million
milesfordriversages30–59). Male drivers had higher fatal
crash rates than females until age 65, although the gender
differences were most pronounced during the teenage years.
Rates of police-reported crashes per mile traveled during
2001–02 varied similarly with age (Fig. 2). Estimated
crash rates for 16-year-old drivers again were higher than
for all other age groups–26 crashes per million miles
traveled, versus 14 for drivers ages 18–19, and 4 for drivers
ages 30–59 (Fig. 2). Unlike the gender differences observed
for fatal crashes, male and female drivers in all age groups
had similar police-reported crash rates.
3.2. Crashes per Mile Traveled during 2001–02 by Time of Day
Fig. 3 shows mileage-based fatal crash rates during 2001–
02 by driver age and time of day. For all age groups, fatal
crashes per mile traveled were higher at night (9 p.m. to 5:59
a.m.) than during the day (6 a.m. to 8:59 p.m.), but the largest
differences were among teenage drivers. Daytime fatal crash
Table 1
Estimated average annual miles traveled by driver age, 2001–02 NHTS
Age Miles
16 7,179
17 9,843
18 12,704
19 15,570
20–24 15,860
25–29 17,277
30–34 16,539
35–39 17,462
40–44 17,158
45–49 16,677
50–54 16,514
55–59 14,300
60–64 14,357
65–69 12,641
70–74 10,375
75–79 8,786
80–84 7,186
≥85 5,806
Fig. 1. Fatal Passenger Vehicle Crashes per 100 Million Miles Traveled by Driver Age and Gender, 2001–02 FARS, NHTS.
139S.A. Ferguson et al. / Journal of Safety Research 38 (2007) 137–145
rates were highest for 16 year-olds, but 17 year-olds had the
highest nighttime rates, declining thereafter.
Nighttime and daytime police-reported crash rates followed
the same general pattern by age, with 16-year-old drivers
having the highest daytime rates and 17 year-old drivers
having the highest nighttime rates (Fig. 4). However, the
differences between daytime and nighttime police-reported
crash rates were much smaller during the teenage years than
the differences for fatal crash rates, and the differences essen-
tially disappeared after age 25.
3.3. Crashes per Population in 2005
A smaller percentage of the teenage population is licensed to
drive, and teenagers tend to drive fewer annual miles than all but
the oldest drivers. Among teenage drivers, 16 year-olds are less
likely to obtain a license than other teenagers and they drive
fewer miles. Their crash rates per population reflect these facts
and are not as elevated, relative to other ages, as rates per mile
traveled (Table 2). In fact, drivers ages 20–24 had slightly
higher per population fatal crash rates than drivers ages 16–19
(34 vs. 32 crashes per 100,000 population). Further, 16 year-
olds had the lowest fatal crash rate among 16–19 year-old
drivers; the crash rate for 16 year-olds was about half that for 18
and 19 year-olds (19 vs. 40 crashes per 100,000 population).
When police-reported crashes were examined, 18 year-old
drivers had the highest crash rates, substantially higher than
16 year-olds (92 vs. 56 crashes per 1000 population) (Table 2).
Crashes per population also differed by gender. Female
drivers had lower fatal crash rates than male drivers through-
out the teenage years and into adulthood. For police-reported
crashes, female drivers had slightly higher per population
crash rates at age 16, but by age 17 crash rates among males
eclipsed those of females and remained higher throughout
the age span.
3.4. Fatal Crashes Involving Alcohol in 2005
All U.S. states have zero tolerance laws making it illegal
for people younger than 21 to drive with any measurable
Fig. 2. Police-Reported Passenger Vehicle Crashes per Million Miles Traveled by Driver Age and Gender, 2001–02 NASS/GES, NHTS.
Fig. 3. Fatal Passenger Vehicle Crashes per 100 Million Miles Traveled by Driver Age and Time of Day, 2001–02 FARS, NHTS.
140 S.A. Ferguson et al. / Journal of Safety Research 38 (2007) 137–145
amount of alcohol in their bodies (IIHS, 2007). Fig. 5 shows
the percentage of fatally-injured passenger vehicle drivers
with positive BACs by driver age and gender during 2005.
Data on driver BACs in nonfatal crashes are not provided
because of their limited availability in police crash reports.
The percentage of fatally injured drivers with positive
BACs was lower for 16 year-olds than for other teenage
drivers; this was the case for both males and females.
The percentage increased through the teenage years and ages
20–29 and declined thereafter, with the lowest rates among the
oldest drivers. Fatally injured female drivers in all age groups
had consistently lower percentages of positive BACs
compared with male drivers, and even the highest percentage
for females was considerably lower than that for males of that
age group. Similar patterns were observed when examining
percentages of drivers with BACs at or above 0.08% (data not
shown).
3.5. Teenage Passenger Deaths in Passenger Vehicles in 2005
A large proportion of 16–19 year-olds killed in passenger
vehicles in 2005 were riding as passengers. Passengers
accounted for 40% of all passenger and driver deaths in 2005
(Table 3). The largest proportion of passenger deaths (51%)
occurred among 16 year-olds. The death rate per population
for 16–19-year-old passengers was higher than for passen-
gers of other age groups combined (9 vs. 3 per 100,000
population) (table not shown).
Most of the teenage passenger deaths occurred when
other teenagers were driving. This was particularly evident
for 16 year-old passengers, 70% of whom died when another
teenager was driving (table not shown).
Examination of fatal crash characteristics provides insight
into the factors involved in high death rates among teenage
drivers (Table 4). Teenage drivers in fatal crashes, especially
16 year-olds, were more likely than drivers 20 and older to
have committed a driver error, to have been speeding, and to
have been transporting three or more occupants in the
vehicle. However, 16 year-old drivers in fatal crashes were
less likely to have positive BACs compared with drivers of
other ages.
Table 5 shows the crash characteristics of 16–17-year-old
drivers in fatal crashes as a function of passenger presence in
the vehicle. As the number of teenage passengers increased,
fatal crashes among 16- and 17-year-old drivers were more
likely to have involved a single vehicle, speeding, and driver
error. With three or more teenage passengers, 85% of the
crashes involved driver error, almost half involved speeding,
and almost 70% involved a single vehicle.
3.6. Changes in Crash Rates between 1996 and 2005
3.6.1. Crashes per Population
Between 1996 and 2005, the 16 year-old driver fatal crash
rate per 100,000 people declined from 33 to 19 per 100,000
population, a 42% decrease (Table 6). By comparison,
among 17 year-olds, per population rates decreased 23%.
Fatal crash rates declined 15% for 18-year-old drivers and
7% for 19-year-old drivers. Among drivers ages 30–59, rates
declined 15%. Thus, 16 year-olds experienced the greatest
Fig. 4. Police-Reported Passenger Vehicle Crashes per 1 Million Miles Traveled by Driver Age and Time of Day, 2001–02 NASS/GES, NHTS.
Table 2
Fatal and police-reported crash involvements per population by driver age
and gender, 2005 FARS and 2005 NASS/GES
Age Fatal crashes per 100,000
population
Police-reported crashes per
1000 population
Males Females Total Males Females Total
16 23 15 19 55 58 56
17 38 21 30 85 76 80
18 55 24 40 106 78 92
19 54 24 40 95 76 86
16–19 43 21 32 85 72 79
20–24 49 19 34 77 65 71
25–29 35 15 25 66 49 58
30–59 24 11 17 46 35 40
60–69 22 9 15 32 20 26
70+ 26 9 16 27 15 20
141S.A. Ferguson et al. / Journal of Safety Research 38 (2007) 137–145
reduction in fatal crashes per population, followed by 17 and
then 18 year-olds. Reductions for 19 year-olds were
somewhat lower than for drivers ages 30–59.
During the same time period, police-reported crash
involvements per 1000 population also decreased the most
(41%) among 16 year-old drivers (Table 6). This compares
with reductions of 26% for 17 year-olds, 19% for 18 year-
olds, 12% for 19 year olds, and 26% for drivers ages 30–59.
Again 16 year-old drivers experienced the largest percentage
reduction in police-reported crashes per population com-
pared with drivers of other ages.
When changes in fatal crash rates per population were
examined by driver age and gender, reductions generally were
greater for males than for females. Among 16 year olds, fatal
crash rates per population decreased 44% for males versus
39% for females. Similar gender differences were found
among 17 and 18-year-old drivers (25 vs. 20% and 17 vs. 12%,
respectively). Nineteen year-old male drivers experienced a
10% reduction versus a 1% increase among female drivers.
Similar patterns occurred when examining police-reported
crashes per population by driver age and gender.
3.6.2. Fatal Crashes Involving Teenage Passengers
Many U.S. GDL laws restrict the number of teenage
passengers the youngest drivers may carry when first licensed.
The effects of these restrictions can be seen by examining the
fatal crash involvements of 16-year-old drivers as a function of
passenger presence (see Table 7). Between 1996 and 2005,
Fig. 5. Percentage of Fatally-Injured Passenger Vehicle Drivers with Positive BACs by Driver Age and Gender, 2005 FARS.
Table 3
Occupant deaths in passenger vehicles by driver age, 2005 FARS
Age Driver deaths Passenger deaths Total Percent passenger deaths
16 356 365 722 51
17 519 386 909 42
18 736 415 1,154 36
19 724 380 1,104 34
Total 2,335 1,546 3,889 40
Table 4
Characteristics of fatal crashes by driver age (Percent), 2005 FARS
Driver age
16 17 18 19 20–25 26–49
Driver error 74 73 71 68 64 51
Speeding 34 32 33 33 30 19
Single vehicle 49 47 44 46 45 38
3+ occupants 29 24 23 24 19 17
Driver killed with positive BACs 15 23 30 32 53 48
Table 5
Characteristics of fatal crashes among 16–17 year-olds when driving alone
or when carrying teenage passengers (Percent), 2005 FARS
Driver alone Driver and
1 teenage
passenger
Driver and
2 teenage
passengers
Driver and
3+ teenage
passengers
Driver error 71 75 78 85
Speeding 30 34 42 46
Single vehicle 41 45 57 69
Drivers killed
with positive
BACs
12 15 12 16
Table 6
Fatal crashes per 100,000 population by driver age, 1996 vs. 2005 FARS
Age 1996 2005 Percent reduction
16 33 19 42
17 39 30 23
18 47 40 15
19 43 40 7
30–59 20 17 15
Police-Reported Crashes per 1,000 Population by Driver Age, 1996 vs. 2005
NASS/GES
Age 1996 2005 Percent reduction
16 95 56 41
17 108 80 26
18 113 92 19
19 98 86 12
30–59 54 40 26
142 S.A. Ferguson et al. / Journal of Safety Research 38 (2007) 137–145
fatal crashes involving 16 year-old drivers decreased 37%
overall. The largest reductions occurred for fatal crashes
involving passengers. Fatal crashes involving teenage passen-
gers decreased 41%, and crashes involving passengers of other
ages decreased 49%. By comparison, crashes involving
16 year-olds driving alone decreased 24%. Furthermore, the
percent of fatal crashes involving 16 year-olds carrying 3 or
more passengers was reduced by half. Because of incon-
sistencies in NASS/GES for reporting passenger data, similar
analyses were not conducted for police-reported crashes.
3.6.3. Fatal Crashes Involving Alcohol
Tab le 8 illustrates the percentage change between 1996 and
2005 in the percent of fatally injured drivers with positive
BACs, by driver age. The largest reductions occurred among
the youngest (age 16) and oldest (60 and older) drivers.
Reductions ranged from 16% for 16 year-old drivers to 5–9%
for 17–19 year-old drivers, and 14% for drivers 60 and older.
There was almost no change among drivers ages 20–29; the
groups with the highest percentage of positive BACs.
3.6.4. Daytime and Nighttime Fatal Crashes
Between 1996 and 2005, percentage reductions in the
number of fatal crashes per population among 16-year-old-
drivers were larger for crashes occurring at night than during
the day (Table 9). Nighttime and daytime fatal crash rates
decreased 48% and 40%, respectively, among this age group.
Reductions in per population fatal crash rates among 17- and
30–59 year-old drivers were not different for nighttime
and daytime crashes (24% at night and during the
day for 17 year-olds; 12 during the day vs. 11% at night
for 30–59 year-olds).
During the same period, the percentage reductions in
police-reported crashes per population for 16- and 17 year-old
drivers were larger at night than during the day (47% vs. 39%
respectively among 16 year-olds; 29 vs. 24% among 17 year-
olds) (see Table 10). Drivers ages 30–59 also experienced
greater reductions in their nighttime police-reported crash rates
per population (30% at night vs. 24% in the day). Thus,
although the pattern of reductions in per population fatal and
nonfatal crash rates by time of day varies as a function of driver
age, the reductions among 16 year-olds were consistently
higher at night than during the day.
4. Conclusions
The number of teenagers killed in motor vehicle crashes
in 2005 in the United States was the lowest since 1992,
despite the largest population of teenagers since 1977 (IIHS,
2006a). Similar progress has not been realized among drivers
Table 7
Fatal crashes of 16-year-old drivers by passenger presence, 1996 vs. 2005 FARS
Crashes Percent
1996 2005 reduction
No passengers 426 324 24
Teenage passengers only
One 309 206 33
Two 175 93 47
Three or more 144 72 50
Total 628 371 41
Other passenger combinations 215 110 49
Total 1,269 805 37
Table 8
Percent of fatally-injured passenger vehicle drivers with positive BACs by
driver age, 1996 vs. 2005 FARS
Age Percent Percent
reduction
1996 2005
16 18 15 16
17 25 23 7
18 31 30 5
19 35 32 9
20–24 53 52 2
25–29 56 55 2
30–59 46 42 8
60–69 21 18 14
70+ 9 8 14
Table 9
Daytime and nighttime fatal crashes per 100,000 population by driver age,
1996 vs. 2005 FARS
Age 1996 2005 Percent reduction
Daytime (6 a.m. to 8:59 p.m.)
16 22 14 40
17 25 19 24
18 28 24 14
19 24 23 3
30–59 14 12 12
Nighttime (9 p.m. to 5:59 a.m.)
16 11 6 48
17 14 10 24
18 19 16 17
19 19 17 12
30–59 6 5 11
Table 10
Daytime and nighttime police-reported crashes per 1,000 population by
driver age, 1996 vs. 2005 NASS/GES
Age 1996 2005 Percent reduction
Daytime (6 a.m. to 8:59 p.m.)
16 80 49 39
17 91 69 24
18 92 75 18
19 80 68 14
30–59 47 36 24
Nighttime (9 p.m. to 5:59 a.m.)
16 14 7 47
17 16 11 29
18 20 17 16
19 18 17 4
30–59 7 5 30
143S.A. Ferguson et al. / Journal of Safety Research 38 (2007) 137–145
of other ages. During the past decade there have been
dramatic reductions in per population rates for both fatal and
police-reported crashes among 16 year-olds, and the re-
ductions are of about the same magnitude. Reductions also
have been recorded for 17-year-old drivers, but these re-
ductions are not as dramatic.
Graduated licensing is designed to reduce high-risk
driving situations among the youngest drivers. Most often
these are 16 year-olds, but 17 year-olds often are affected by
the requirements as well. Although the present study was not
intended to evaluate the effects of graduated licensing per se,
results suggest that graduated licensing laws have been suc-
cessful in reducing fatalities among the youngest drivers.
This is consistent with the findings of Chen, Baker, and Li
(2006). Compared with states that had no GDL provisions,
the authors found reductions of 16% to 21% in the per
population fatal crash rate among 16 year-old drivers in
states with graduated licensing provisions that included a
3 month or longer mandatory leaners' permit period and a
nighttime driving restriction, plus either 30 or more hours of
supervised driving or a passenger restriction.
There is some concern that delaying full-privilege
licensure until age 17 or 18 could result in higher crash
risks among 18 year-old drivers because of their reduced
exposure in the beginning years of driving. However, there is
little evidence from the present study that reduced fatal crash
rates among younger drivers have been at the cost of higher
fatal crash rates among older teenagers (see also Williams,
2007).
It would be gratifying to be able to point definitively to
the factors responsible for the declines in fatal crashes per
population. Is it because teenagers were licensed for less time
during the year when they turned 16? That is, were 16 year-
olds obtaining licenses later, or were fewer of them licensed
at all thus reducing driving exposure? Is it because the nature
of driving exposure changed with added restrictions on when
and with whom beginning drivers can drive? Or were
beginners driving more safely as a result of the longer learner
periods and potentially greater experience under a wider
range of conditions? Some or all of these factors could have
contributed to the decline, but the present study could not
identify which of these factors were important and by how
much. Williams (2007) examined the evidence on the
effectiveness of specific graduated licensing components to
try to disentangle some of these factors.
A closer examination of fatal crashes, for which compre-
hensive data are available, indicates where progress has been
made. For example, fatal crashes in 2005 involving 16 year-
olds were less likely than such crashes in 1996 to involve
multiple teenage passengers in the vehicle. Among 16 year-
olds, crashes such as these have been reduced by more than
half (see also Williams, Ferguson, & McCartt, 2006). Some
concern has been raised that progress in this area may be made
at the cost of additional crashes that could occur if more
teenagers drive alone, but reductions also were found in the
numbers of fatal crashes in which the 16 year-old drivers were
alone. Other studies designed to assess graduated licensing
effectiveness have found that passenger restrictions have been
beneficial in reducing crashes (Chen et al., 2006; Cooper,
Gillen, & Atkins, 2004; Highway Safety Research Center
Directions, 2006; Masten & Hagge, 2003; Morrisey, Grabow-
ski, Dee, & Campbell, 2006; Rice, Peek-Asa, & Kraus, 2004;
Williams et al., 2006; Zwicker, Williams, Chaudhary, &
Farmer, 2006). In addition, the percentage of fatal crashes at
night as well as those involving alcohol has declined more
among 16 year-old drivers than among older teenagers.
In summary, we have seen dramatic reductions in the per
population crash rates, both fatal and nonfatal, of 16 year-
old drivers. These gains do not appear to be at the expense of
older teenage drivers whose rates also have declined but by
smaller amounts. Fatal crashes among 16 year-olds are at a
historic low.
Acknowledgments
The authors wish to thank Allan Williams and Adrian
Lund for their helpful comments. This work was supported
by the Insurance Institute for Highway Safety.
References
Chen, L., Baker, S. P., Braver, E. R., & Li, G. (2000). Carrying passengers as
a risk factor for crashes fatal to 16–17-year-old driver. Journal of the
American Medical Association,283, 1578−1617.
Chen, L., Baker, S. P., & Li, G. (2006). Graduated driver licensing programs
and fatal crashes of 16-year-old drivers. Pediatrics,118 ,56−62.
Cooper, D., Gillen, D., & Atkins, F. (2004). Impacts of California's
graduated licensing law of 1998. Berkeley, CA: University of California
Institute of Transportation Economics.
Doherty, S. T., Andrey, J. C., & MacGregor, C. (1998). The situational risks
of young drivers: The influence of passengers, time of day, and day of
the week on accident rates. Accident Analysis and Prevention,30,
45−52.
Highway Safety Research Center Directions. (2006). Evaluating the effects
of NC's teen driver cell phone restriction. Chapel Hill, NC: University of
North Carolina.
Insurance Institute for Highway Safety [IIHS]. (2006). Fatality Facts 2005:
Teenagers.: Insurance Institute for Highway Safety [IIHS] Arlington,
VA, Available at http://www.iihs.org (accessed 7 December 2006).
Insurance Institute for Highway Safety [IIHS]. (2006). Laws and regulations,
regulatory comments.: Insurance Institute for Highway Safety [IIHS]
Arlington, VA, Available at http://www.iihs.org (accessed 7 December
2006).
Insurance Institute for Highway Safety [IIHS]. (2007). Q and A teenagers:
Underage drinking.: Insurance Institute for Highway Safety [IIHS]
Available at http://www.iihs.org (accessed 24 January 2007).
Li, G., Braver, E. R., & Chen, L. (2003). Fragility versus excessive crash
involvement as determinants of high death rates per vehicle-mile of travel
among older drivers. Accident Analysis and Prevention,35,227−235.
Lyman, S., Ferguson, S. A., Braver, E. R., & Williams, A. F. (2002). Older
driver involvements in police reported crashes and fatal crashes: trends
and projections. Injury Prevention,8,116−120.
Masten, S. V., & Hagge, R. A. (2003). Evaluation of California's graduated
driver licensing program. Report no. 205. Sacramento, CA: California
Department of Motor Vehicles.
Mayhew, D. R., Simpson, H. M., & Pak, A. (2003). Changes in collision
rates among novice drivers during the first months of driving. Accident
Analysis and Prevention,35, 638−691.
144 S.A. Ferguson et al. / Journal of Safety Research 38 (2007) 137–145
Morrisey, M. A., Grabowski, D. C., Dee, T. S., & Campbell, C. (2006). The
strength of graduated drivers license programs and fatalities among teen
drivers and passengers. Accident Analysis and Prevention,38(1),
135−141.
Preusser, D. F., Ferguson, S. A., & Williams, A. F. (1998). The effect of
teenage passengers on fatal crash risk of teenage drivers. Accident
Analysis and Prevention,30, 217−222.
Rice, T. M., Peek-Asa, C., &Kraus, J. F.(2004). Effectsof California graduated
driver licensing program. Journal of Safety Research,35,63−69.
Shope, J. T. (2007). Graduated Driver Licensing: Review of Evaluation
Results since 2002. Journal of Safety Research,37(2).
Ulmer, R. G., Williams, A. F., & Preusser, D. F. (1997). Crash involvements
of 16-year-old drivers. Journal of Safety Research,28,97−103.
Williams, A. F. (2003). Teenage drivers: patterns of risk. Journal of Safety
Research,34,5−15.
Williams, A. F. (2007). Contribution of the Components of Graduated
Licensing to Crash Reductions. Journal of Safety Research,37(2).
Williams, A. F., Ferguson, S. A., & McCartt, A. (2006). Passenger effects on
teenage driving and opportunities for reducing the risks of such travel.
Arlington, VA: Insurance Institute for Highway Safety.
Williams, A. F., Ferguson, S. A., & Wells, J. K. (2005). Sixteen-year-old
drivers in fatal crashes, United States, 2003. Traffic Injury Prevention,6,
202−206.
Zwicker, T. J., Williams, A. F., Chaudhary, N. K., & Farmer, C. M. (2006).
Evaluation of California's graduated licensing system. Arlington, VA:
Insurance Institute for Highway Safety.
Susan A. Ferguson, Ph.D. is President, Ferguson International, a highway
safety research consulting company. Dr. Ferguson received a B.A. in
psychology in 1980 and a Ph.D. in experimental psychology from the
George Washington University in 1991. Formerly senior vice president for
research at the Insurance Institute for Highway Safety Dr. Ferguson directed
research activities in a wide range of highway safety areas determining
priorities and overseeing their execution. During her 15 years at the Institute
(1991–2006).
Susan conducted research in many different highway safety areas with
emphasis on young and older drivers, vehicle safety issues, alcohol and
driving, and child occupant protection, and has published more than 100
scientific papers. She has been invited to present on highway safety
research to diverse audiences both in the U.S. and around the world and has
frequently been featured on national television, radio, and in print.
Dr. Ferguson serves on many committees and advisory boards both within
the U.S. and internationally, and chairs the Blue Ribbon Panel on Advanced
Airbags and the Blue Ribbon Panel for the Development of Advanced
Alcohol Detection Technology. She serves on the Executive Board of the
International Council on Alcohol, Drugs and Traffic Safety, and the
Transportation Research Board Committee on Alcohol, Other Drugs and
Transportation Committee.
145S.A. Ferguson et al. / Journal of Safety Research 38 (2007) 137–145